a







LEGIBILITY, ATMOSPHERE-VALUE AND FORMS OF PRINTING TYPES



Printed in Holland
by A. W. Sijthoff’s Uitgeversmaatschappij N.V., Leiden in Monotype-“Times New Roman”, Series 327.



LEGIBILITY,
ATMOSPHERE-VALUE
AND FORMS of
PRINTING TYPES
PROEFSCH Rl FT
TER VERKRIJGING VAN DEN GRAAD VAN DOCTOR IN DE LETTEREN EN WIJSBEGEERTE AAN DE RIJKSUNIVERSITEIT TE UTRECHT OP GEZAG VAN DEN RECTOR MAGNIFICUS Dr. J. BOEKE, HOOGLEERAAR IN DE FACULTEIT DER GENEESKUNDE, VOLGENS BESLUIT VAN DEN SENAAT DER UNIVERSITEIT IN HET OPENBAAR TE VERDEDIGEN OP DONDERDAG 7 JULI 1938,
DES NAMIDDAGS TE 2 UUR,
DOOR
GERRIT WILLEM OVINK
GEBOREN’TE LEIDEN
LEIDEN 1938
A. W. SIJTHOFF’S UITGEVERSMAATSCHAPPIJ N.V.







Voor mijn Vader en Moeder







De beëindiging van dit proefschrift schenkt mij de gelegenheid (die zich gedurende de studie niet licht voordoet) om mijn dankbaarheid te uiten jegens U Hooggeleerde Roels, Hooggeachte Promotor, voor al hetgeen Gij voor mij zijt geweest, voor de nooit verflauwende belangstelling en .steun en de heldere critiek bij de bewerking van mijn proefschrift, voor den invloed op mijn persoonlijke en wetenschappelijke vorming — voor alles waardoor Gij het laboratorium tot een tweede thuis voor ons maakte, zeg ik U mijn diepgevoelden dank.
Hooggeleerde Vogelsang, het is mij een behoefte om het tegenover U uit te spreken, hoezeer Uw colleges voor mij door hun inhoud en vorm een wetenschappelijk en aesthetisch genot waren; moge ik het van U ontvangen onderwijs niet te zeer verloochend hebben waar ik mij in dit proefschrift op de paden der kunstbeschouwing begaf.
Hooggeleerde Franken, Uw kennistheoretische colleges zijn voor mij van zeer veel waarde geweest, niet minder het persoonlijk contact, dat ik gedurende mijn studietijd met U mocht hebben.
Hooggeleerde Rümke, Uw colleges hebben mijn blik op het terrein der specieele psychologie zeer verruimd.
Zeergeleerde Kraft, het is mij een voorrecht buiten het Academiegebouw Uw vriendschap te mogen genieten; het zij mij vergund U op deze plaats te danken voor al hetgeen Gij ook daarbinnen, als docent, aan mijn vorming hebt bijgedragen.
Zeergeleerde van der Spek, Zeergeleerde Tenhaeff, Uw beider boeiende colleges hebben mij bekend gemaakt met problemen, waarvoor ik de belangstelling steeds zal blijven behouden.
Zeergeleerde van Heuven, verscheidene malen is Uw kennis en kunde mij tot grooten steun bij het onderzoek geweest; voor de bijzonder welwillende wijze waarop Gij telkenmale Uw hulp hebt verleend, breng ik U mijn hartelijken dank.
Aan Mevrouw T. Osieck—Meereboer ben ik dank verschuldigd voor de zorgen, die zij aan de correctie van het manuscript besteedde.
Den Heer J. W. Stegink dank ik ten zeerste voor de voortreffelijke hulp, die ik gedurende mijn laboratoriumjaren steeds ondervond.
Al kan ik ook de beteekenis daarvan voor mij hier niet goed onder woorden brengen, toch kan ik niet nalaten mijn diepe dankbaarheid uit te spreken voor de geestelijke vorming, die mijn vader mij op zoo grondige en veelzijdige wijze heeft geschonken.







PREFACE.
In the present volume an attempt has been made to collect our psychological knowledge of the modern printing type, some new data included. The reader will find that this knowledge is far from being complete. As every investigation on this subject is conducted in behalf of printers, advertisers, publishers etc., the practical needs of these instances should determine the object and the method of investigation. Hence some technical knowledge on the part of the investigator is needed in order to avoid the compilation of results that have but a theoretical or heuristical value. The sources for such a technical knowledge are not easily accessible for the layman in “typophily”. The chapters concerning the problem of legibility are therefore preceded by a technical introduction on typefounding and type setting. A short historical review of the printing type since the year 1900, and a list of more than 600 type faces have been included too. They are intended to enable the reader to locate a design in its artistical surroundings. For a study on type faces that neglects their nature as objects of art, is bound to limit its own value, if it does not even hit beside the mark alltogether.
For the typographer this volume presents some material from which he can gain viewpoints for the handling of practical cases by means of combination and deduction.
For the psychologist it adds some points to the lengthy programme of investigations still to be done; at the same time it may call forth, so I hope, some methodological discussions.
I wish to express my gratitude to the numerous persons and institutions in various countries who have given me their kind assistance, either by printing the necessary material, or by providing me with informations; their interest in the problem of legibility and atmosphere-value has greatly encouraged me. I am very much indebted to the teachers and pupils of the advertising classes of the Academy of Art at The Hague, who prepared the alphabets used in the experiments with isolated letters, and, last not least, to the subjects of the experiments and enquiries, who devoted so many hours to this cause.
Psychological Laboratory University of Utrecht, May 1938.







CONTENTS.
Page
Technical Introduction 1
Part I — The legibility of printing types
Chapter 1. — The problem 7
Chapter 2. — Theoretical foundations of the legibility of isolated letters . 12
Chapter 3. — Variations of the shape in proper sense 22
Chapter 4. — Variations of the thickness of the constituent parts ... 38
Chapter 5. — Theoretical foundations of reading of letters in context. —
The difference between the reading of isolated characters, of single lines and of several lines \ 71
§ 1. The problem 71
§ 2. Recognition of single characters 72
§ 3. Recognition of single words 72
Gestaltfactors 77
Space and outline 73
Boldness 74
Experience. 74
Importance of upper half 76
Serifs ; . 77
Horizontal extension 78
Vertical extension 80
§ 4. Recognition of lines of words (reading in proper sense) .... 80
Chapter 6. — The measurement of reading-speed of a line of words . . 84
Chapter 7. — The judgment of the subjects on legibility . . . . . . 105
Chapter 8. — Summary and application of the results 107
§ 1. The studies on typographical factors influencing speed of reading
by Tinker & Paterson 107
§ 2. Style of type face 111
Fraktur versus Antiqua 115
§ 3. Size of Type and Length of Line 117
§ 4. Space between lines 121
§ 5. Legibility of Small Print 122
Newspapers 122
Time-Tables, Dictionaries etc 123
Continuous Texts 124
§ 6. Colour and surface of paper 124



Part II — The atmosphere-value oj type /aces
Page
Chapter 9 127
§ 1. Description and method of investigation 127
§ 2. Former studies 134
§ 3. Series Vila 135
§ 4. Series 139
The types J39
The questionnaire 141
The instruction 142
Conditions • J43
Results 14*
Description of the qualities 154
Description of the type faces 1”1
Chapter 10. — The influence of margin, interlineage, colour and quality
of paper 1"*
Questionnaire • 1"
Material 
Interlineage 
Margin 
Colour J°Y
Paper 
Proceeding 
Results 
Part III — Short historical review of the modern printing type
§ 1. Introduction 1®®
§ 2. United States 1^1
§ 3. England 200
§ 4. France 209
§ 5. Germany 214
§ 6. Other Countries 226
Spain 226
Holland 227
Kaly 228
List of the most important modern printing types 229
246
Selected references 



TECHNICAL INTRODUCTION.
TYPE-FOUNDING AND -SETTING.
Technique.
In former days the type-founder combined the functions of artist, designer, engraver and tradesman. When type-founding became an industry, these functions were specialised. To-day a new design is not made by the punchcutter himself, but by the art department or by an artist connected with the house.
The original design is mostly executed on a large scale. Before the foundry decides to cut a trial-grade, elaborate experiments are done, to see if the characters ‘sit’ well in all combinations; photographical reductions are made to study their behaviour in lower grades: When the result is satisfying, one or more grades are cut to judge the ultimate effect. During this work many alterations are made in steady collaboration of artist and crafts-men.
After that the matrices are cut from which the types will be cast. There are various methods for this. When the first and oldest way is used, the reversed design is transferred on the top of a polished steelbar; the surroundings are cut away, till the character stands out, similar to the future type. This punch is stamped in a block of softer material (mostly copper), which receives the image in the right way, deepened.
A second method for obtaining matrices is to cut the type in type-metal and to get a galvanoplastic nickel-cast of it,
wnicn ui course may serve as a matrix.
The third method consists in the engraving of a matrix, without the preliminary cutting of an original. Therefore a stencilplate is made with the contours of the character in it. These contours are mechanically engraved in the copper-block to any size desired.
Then the justifier comes. His task is to see to it that the image stands right in the matrix in both horizontal and vertical directioh; that all the characters of a grade are standing along a straight line, that the space surrounding the characters is chosen well, to ensure a regular appearance of the word composed of them. General! y the trial-grade is of medium height. The other grades cannot always be obtained by mechanical reduction or enlarging, because in small dimensions the type would look too narrow and frail, while in large dimensions it would look too heavy and wide. This is prevented by
some slight corresponding alterations.
The dimensions of printing types (and all composing material) are measured by an own Standard, the point system. In England and the U.S.A. a point is 1/72
Legibiiity, atmosphere-value and forms of printing types. 1



inch, 1 inch counting 6 units (so called “em” = 2 “en”) of 12 points, also called “pica”. In the other countries generally the French Didot-system is followed, which counts 2660 pointss) to 1 meter; the 12 point-unit is called ‘augustin’ or ‘cicero’. The difference is not big; the 12 point em measures 0.166 inch, the (didot) cicero 0.177 inch.
The size of the typebody is thus expressed in points. This means that not the size of the actual face is indicated, because thiscan vary with thesame body-height. Several types are available with a small or large face on the same body. This is often usual with small newspaper-types ; e.g. a 6 pt. face (i. e. the face usual of a 6 pt. body) is fount on a 7 point body. In a more simple way this serves the
LARGE FACE SMALL FACE
SOLID 6 Pt. 6 Pt.
6 Pt. CAST 1 Pt.
ON 7 Pt. LEAOED
FIGURE 2 FIGURE 3
same purpose, of leavingT pt. blank (“white”) between the lines, so-called “leading” (inscrting strips of metal between the lines). The most used type-heights for bookwork are: 8, 9, 10 to 12 point. Newspapers often use 6 or 7 pt.; sometimes 5 pt. Types to 3 point exist for special purposes; larger sizes than 12—14 pt.
FIGURE 4-
are used for display-work, i. e. for headings, text matter that has to call attention, etc Display sizes do not increase with one point, the usual heights being: 16, 18 (on the continent 20), 24, 30 (continent: 29 or 32), 36, sometimes 42, and further 48 60 72, 84, 96. Larger sizes, as those in use for posters and window-bills, are cut in wood, because the weight of the metal-type would be too heavy.
For handsetting type is purchased from a foundry by weight, but according to a special system of quantities for each letter or sign, expressed in the “fount-bill , based on the relative frequency of the letter in a given language.
The types are kept in cases *), wooden of steel trays, which are divided in small parititons, one for each character or digit. These partitions vary in size and posi-
*) The word em is used in two senses : 1) as square of type body-depth, e.g. 6 pt. em - 6 pt. square ; 2) specially, as 12 pt. square, to measure length of line etc.
*) Further abbreviated as ‘D point’. , . . xhev
■) One case contains the majuscules (or minuscules) of one typeface in “h® »‘*e- They are kept in frames ; those in use are mounted upon it. The case for
is tilted behind the minuscule case. Therefore the composer talks about upper- and “lower case”, instead of about ‘‘capitals and minuscules .



tion, according again to the frequency of the use of their occupants. One by one the types are taken up and put together in the “composing-stick”, a tooi in which the typesetter assembles the types to lines. Between the words spaces (thin pieces of type-metal, less high than the types) are inserted. To get lines of uniform length with a good abbreviation, these spaces are varied for each line in analmost invisible way. A good “filling-out” of the line is one of the chief difficulties of composition.
Another adjustment is needed to avoid the irregular appearance of a word. It is based upon the optical peculiarities of the various characters of the alphabet. When characters of a rectangular shape (E, N) are placed side by side, the space between them is also rectangular. Those characters, that have triangular (V, A) or round (O, C, D) forms, leave a large, white spot on the top- or base-line ; this is also the case with overhanging sorts, as f and j. In letter-drawing these uneven
AVHOO AVHOD
FIGURE 5
UNH A W AIA N UNHAWAIAN
FIGURE 6
spaces are prevented by allowing the two characters to overlap. This is impossible with the printing-type, because the square blocks of type metal cannot be made to overlap, save at considerable higher costs (see under “kerns”). However, there are some ways to overcome this difficulty.
Sparing.
The first method costs too much time to be of use in the regular practice. It can only serve for display sizes. When the rectangular and round characters cannot be set more close together, it follows that the other characters must be spaced, in order to obtain (optically) the same quantity of white between them.
Natural fitting.
Of course it is possible to minimize the annoying patches of white by squaring-up the triangular and round letters.
See ng. 7. Sometimes also a little flap is added at the top left side of A, J; and base-line strokes to V, W, etc.
The justifier can take care that the rectangular characters obtain a proper amount of white surrounding them; this, however,
musi aiways oe a compromise,
lest too much space is used. FIGURE 7
AOTWY



Logotypes.
The same trouble presents the f, when the loop is bent over. For the most used ;ombinations special “logotypes” are provided; e.g.: fi, ft, ffi, fa, fu, fe, fo. But landling logotypes slows down the setting-speed ; therefore (with a loss of grace ind clearness) mostly the curve is shortened to a mere “crochet-hook”. Making ogotypes for the j would be an endless task, thus the common form is without a :urve.
Kerns.
Overhanging sorts, as italics, cursives, “swash” (decorative) characters and roman f, g, j, y, can be made to fit well by undercutting the overhanging part, which :omes to rest on the non-printing part of the adjoining type. The founding and finishing of these types requires more labour and is therefore more expensive; a snug fitting is gained at the cost of a greater vulnerability during the printing or stereotyping. Ornamental capitals in use for headings, etc., are often mortised, i. e. a part of the body is cut away, so that the character can come closer to the next one.
Standard Lining System.
Many modern types have been cut according to the so called “lining system” (Universallinie) in order to make it possible to use different sizes in one line. We quote Updike’s description and criticism *) : “There are mechanical advantages inherent in this scheme, but they are attained at the expense of the correct proportion of certain letters. The convenience of being able to align different sizes of type by the use of 1-point leads or their multiples is no doubt appreciated by the compositor; but to make this possible, types below 18-point are arranged, so far as “line” is concerned, in but three groups : in the first are 5 and 6 -point types; in the second, 7 to 10-point types ; and in the third, 11 to 16-point. Taking the second group for illustration, the “descenders” of g, j, p, q and y can be no longer in 10-point than in a 7-point face; and as they are none too long in 7-point they become much too short in 10-point. This disregard for proportionate length of descenders of types of different sizes pervades the whole scheme; and consequently no sizeof type (excepting possibly 7-point and 11-point) which is cast on Standard Line can have descenders of adequate length; while in some sizes they are so “chopped off” as to produce real deformity.”
Typesetting.
As we saw above, handsetting means taking the separate types out of the case and putting them in the composing-stick. After printing they must be distributed again in the cases, which means that the same amount of work has to be done in reverse direction. Moreover, having been some time in use, the types become battered and worn, and give bad printing results. No wonder that in the kst century everywhere the mechanization of typesetting has been attempted. The first machines worked-up foundry type, which had to be distributed by hand. It was an idea of Dr. William Church (who took in 1822 the first patent out for a composing-machine) to melt the used type down instead of distributing it, and to found new type for each new job. New type has the advantage of requiring less “make-ready” on the press, and giving a better printing result. Besides, set-up matter can in this case be kept for reprinting. . ■
At that stage of the development of founding the idea was uneconomical and it was not taken up untill Ottmar Mergenthaler brought out his first machine in 1884, which was developed to the Linotype of 1890, the prototype of the well-known machine of today. lts principle (and also that of the similar “Intertype ) js to compose — no type — but matrices, from which a line of type, or “slug , is cast. The matrices are stored in magazines, from which they are released by an ordinary keyboard system ; they are assembled on whatin handsetting is called acomposing-
*) “Printing Types”, 3d ed. I, 35.



stick. The spaces are no strips of metal, but they have the form of two thinopposite wedges, clamped together. When the line is nearly full, these spaces are driven between the matrices, till the fixed length of line is reached. Type metal is then pumped out of the melting pot against the row of matrices, the slug is cast, trimmed and cooled. The matrices are then returned to the channels in the magazines where they belong. In the meantime another line is composed. The matrices of slugcasting machines bear the character on their side l), they vary in width to some fixed dimensions. This fact obviates a true rendering of the free design of a lettering artist; some characters have to be broadened or narrowed. The first Linotype matrices bore only one face; soon the accompanying heavy or italic of the character was added, to make a mixed use on the same slug possible. When another type size or -face was needed, another magazine with matrices had to be inserted in the machine, a difficulty met by the introduction of more magazines in the same machine, located above, later on also at the side of the first. To-day slug casting machines exist with three main- and three side magazines; by dividing the channels of the main magazines, sixteen alphabets to the maximum can be set and cast by one operator on the same machine. The matrices of special characters and signs are inserted by hand. However, these machines can only work at an economical rate, when they are in continuous operation upon a same, mixed job, especially newspaper advertisements. Slug casting machines are in general use for simple text matter, as cheap books, newspapers and periodicals. The slugs facilitate the assembling for the make-up of the printing forme. A handicap for the speed of correction is the necessity of setting a complete new line for one wrong character at the risk of making another error.
In the make-ready of the forme before the printing a greater evenness of impression can be obtained from separate type than from slugs, because the errors in printing-height can be better localized a).
For difficult work, broken off by illustrations, for tabular matter, formula’s etc., where lines of different length have to be used, the necessarily frequent changing of magazines and casting moulds would make the composition too expensive, if not impossible. There separate type-composition is economically the only way. This is achieved by the Monotype machine, or, more correctly : machines. For the Monotype includes a casting- and a separate setting-machine, thus requiring two operators. The keyboard of the composing-machine operates a set of punches, which perforate a paper-ribbon similar to that of a pianola. This ribbon, transferred to the casting-machine and unrolled, puts the matrices successively in position before the casting-mould by a pneumatical system. The width of the various letters and signs is a multiple of a same basic unit of width. This enables the operator to calculate by means of a special device the required amount of space in the line. The casting-machine is able to produce unlimited quantities of types, rules, borders, etc. for hand-composition; for display-sizes a special “super caster” is built.
The design of characters on the Monotype is subject to the same limitations as on the Linotype, only at a smaller degree. From the perforated paper-ribbon the same text can be recast (even in another type face) in case of another edition. In the separate types corrections can be easily made.
The range of machine-composition possibilities.
When, at the end of our investigation, we shall have reviewed the most important part of the existing type-material of foundries and composing-machines, it will be important to know to which degree the various printing-offices are relying on the products of any of these.
This depends on the quantity and nature of the jobs the office is usually handling.
*) When the matrices are old, and the edges damaged, thin hairlines appear between the letters.
.") Make-ready includes securing an equal printing-force on all parts of the forme. Slight difference in height are always left: the resuit is that some parts print too heaviiy, other too grey. Therefore the paper is brought closer to the type in the grey parts by gumming little strips of paper on the corresponding places of the platen or cylinder. In this way the paper will receive en equal pressure in afl parts.



Small offices, that cannot afford a composing-machine, have their large quantities of straight running text-matter (s. c. “body matter”) set up by firms specialising in type-setting. The rest of their work is done out of handset foundry-type. A somewhat larger office owns a ‘Typograph’-machine2) or a smal! Lino- or Intertype 3). Tabular and other mixed works are entirely done by hand. When there is question about much higher class book-, periodical-, or commercial printing, the Monotype enters. Slug-casting machines in large quantities are chiefly kept for newspaper printing, where sometimes a “Ludlow” is added to cast large displaysizes, or a similar slug-casting machine, working from hand-set matrices in a special composing-stick.
As we saw above, the same design is executed in different sizes often ranging from 6 to 84 points. When a printing office of medium size buys a font, from which mixed work is to be set, this implies, for reasons of purity of style of the ultimate product, the purchase of at least the middle sizes up to 36, and also of the italics, the heavy and sometimes narrow, wide or decorated editions of these, which costs many tenths of pounds. Such a type family will have to do the most varied jobs : catalogues, advertising-folders, advertisements, etc. Therefore it must be neutral in its appearance, not too thick, nor too thin ; not too graceful, nor too ‘shouting’. Usually no more than 3 to 6 different faces are run on the machines. Impulses to change to other type-design often meet a certain inertia, due to the difficulty of changing the appearance of the regularly returning jobs, set up out of the old ones.
Among the printers a personal interest in type faces is seldom fóund; mostly the once established relations with a certain foundry prevail over reasons of aesthetics and legibility. This regrettable situation is facilitated by the fact that a .successful type is soon imitated by many other foundries ‘) and then sold as “about 'the same”, though the alterations made (necessary to make the imitation less obvious) generally make just the difference between a good and a bad type. Still, one cannot always speak of imitation. A new direction in type design is often a long time “in the air”, e.g. commercial printing asks for a type for a newly arisen need ; or book-printing gets stuck in a certain style of type ahd requires a new variant. Such has been the case with the originating of the modern sans-serif style. However, also in this case much bad real imitation has been done.
Non-typographical lettering.
The term ‘non-typographical lettering’ includes different techniques, which each have their own possibilities. They are executed by people who received their training either at a school of arts and crafts or from teachers, who had been trained there ; often they are aided by copy-books. When a country enjoys the possession of a group of able lettering artists, such as is the case in England and Germany, the whole level of letter-drawing may be raised to a high Standard in the course of a few decades. On the other hand, a deficiënt technical education will show its effects everywhere. In regard to the relation between “handlettering” and typography, the former usually influences the latter, because the artists of the foundries are recruited from designers of commercial art or professional scribes. Painting and drawing imposes less restrictions on the freedom of the design; therefore it often serves as experimental field for new styles ; successful variations are adopted by the letter-foundries. A certain amount of influence in reverse, however, isquite natural and actually a fact; e.g. the sans-serif style, as interpreted by typography, has greatly influenced handlettering.
The principal territories of handlettering are: drawing (newspaper-ads, shopwindow advertising, showcards, etc., magazine headings, packaging, posters), painting (shop-windows, hoardings), enameled nameplates, electrical advertising (neon tubes, etc.). The nature of these various techniques makes drawing the best suitable to discover new ways.
•) A small slug-casting machine for simple work, cheaper than Lino-and Intertype.
8) Standardization makes it possible tö eniarge the basic machine with more main- and side magazines.
8) At best the matrices are bought by a co-operating foundry.



PART I.
THE LEGIBILITY OF PRINTING TYPES.
CHAPTER 1.
The problem.
Introduction.
In his admirable ‘Report on the legibility of print’ R. L. Pyke gathered together all that was known in the year 1926 about the legibility of printing types ; after sorting it appeared that the few remaining useful investigations could not teil us anything definite, still far less could these scattered fragments of doubtful knowledge cope with the immense variety of typographical products, the relative merits of which, in regard of legibility, were asked.
The entire lack of a systematical plan, and deficiënties of experimental technique are the reason of this failure. Many investigations were conducted with other purposes than of ascertaining the relative legibility of the type-face or type-faces in question, or they concerned special problems ; some were conducted by oculists with insufficiënt knowledge or at least insufficiënt precautions in respect of the psychological factors in reading. Therefore Pyke started with a systematization of the problem and of the few results, which he could take into consideration ; after that he arranged a series of various experiments in order to obtain a corroboration, or perhaps, some definite new facts. His efforts were not rewarded with any success. In view of the able and conscientious way, in which Pyke’s experiments were conducted, it is little promising to read the conclusion of the Committee, that ultimately directed the investigation the experimental
methods now available are not likely to provide clear and direct answers to the many practical problems remaining to be solved, and .... the whole subject can more profitably be approached by subordinating it as part of a more general and comprehensive scheme of research into the physiology of vision.”
The subject was far from being dropped. The Medical Research Council appointed instantly another Committee to continue the work of its predecessor under wider scope ; in America several



new investigations were carried through, that were partly crowned with apparent success. However, many problems remained still unanswered, and no psychologist, who is attracted by this important problem, and who notices the certainty with which we can indicate off-hand the superior legibility of one piece of print above another, shall ever be willing to consent to suspension of its experimental investigation. In this hurt pride of the psychologist the reader may find a justification for the experiments that are described in the present volume ; but there are other grounds too.
First, the legibility of isolated characters has been neglected hitherto, while it is at the same time highly important for the Science and art of printing and advertising, and experimentally more accessible than the legibility of continuous reading-matter.
Secondly, not every possible method of investigation has been used.
Thirdly, the theoretical (psychological) and the typographical under-structure of the experiments still shows large gaps.
As an attempt to fill some of these many hiatuses with other view-points and, may be, useful data, the present report was written. Technical data about typography and type-faces, that are not easily accessible for the average investigator, but that may have their use for him in adjusting his way of putting the problem to the needs and the possibilities of printing and advertising, were collected.
Definition.
Pyke has delivered a severe criticism on previous investigators, who got to work without a proper definition and a clear description of their criterion. He tried to frame an exact definition himself, which runs as follows : “Legibility refers to (1) letters, (2) words, and (3) continuous texts. lts essential and sufficiënt conditions are that it should be possible : (1) with letters, for the eye to define clearly the shape of a letter ; (2) with words, the collective shape of a word ; and (3) with continuous texts, to read a type both accurately and fast.”
In true logical sense this is not a definition, but an illustration, an enumeration of essential criteria. The difficulty of the circumscription of the term legibility has to be found in the multitude of its significances. It can be used to indicate a quality of a type or a piece of print, and then it means the ease, with which the subject can read that type or printed matter. The ease is a property not of the type etc. but of the subject, who has to spend some energy in order to reach his purpose of getting the know-



ledge contained in the print; the ease is either the quality of his feeling towards the obstruction presented on his way, or the amount of consumed energy, expressed in terms of changes in metabolism. One should not think in this respect of fatigue, because fatigue is an insufficiënt supply of new energy, which causes a reduction of output, whereas a difference in facility (either subjectiv? or objective) does not necessarily make a difference in output.
The subjective facility, or feeling of ease, has not got the place in the literature on our subject that it undoubtedly deserves. Mostly one looked for speed of reading, though it is often far more important for a publisher or advertiser to know whether the reading public dislikes a type or not, than to know the actual differences in speed, which are small anyhow.
The objective facility (consumption of energy) cannot be measured in a satisfactory way, and therefore the attention has been shifted to the measurement of output, which can be varying, owing to differences in ease of reading, in quantity (speed) or quality (errors).
For the reading of single characters or words the saving of time is not necessary, while the faultless apperception is ; on the contrary reading continuous texts needs speed in the first place, which includes the need for a minimum of errors, because the subject cannot pass on to a next word, if the foregoing one is not understood correctly. We have therefore confined ourselves to an observation of the shape of the single characters and on the other hand to a measurement of speed in reading continuous texts. The latter can never yield exact and complete results, but only indications, because the reading attitude of the subjects is never the same. In the observation of the single characters of the alphabet the subject is entirely concentrated upon their figural value ; they are all equally well known to him. Single words already produce complications : the subjects’ experience with them is unequal; moreover a pure description of the observed design seems more difficult, because any uncertainty about the single members of the word can immediately be cleared up by inference from the word-whole. If the subject is only asked to pronunciate the observed word, and if only actual misreadings are counted, the differences in objective legibility will not show up adequately. Therefore we did not arrange any experiments with single words.
The measurement of speed of reading is necessarily unexact, unless we become able sometime to measure the whole process of working-up the matter read by the subject, which seems very far-off. We want to know the typographical factors influencing



speed of reading, whereas we have to put up also with a group of largely uncontrollable factors, that influenóe this speed too.
Pyke was well aware of these influences ; he put up two lists of them, which we reproduce here. “The first contains all the factors, which can at present be supposed to influence errors only, and both errors and speed ; the second those which influence speed only.”
LIST 1 LIST 2
Speed and errors, errors only Speed only
Subjective Subjective
1 Association ) „ . . 1 Powers of comprehension ) t l
2 Imagery habits central 2 Natural alertness S
3 Visual acuity peripheral \ Vocalization > periphera
4 Eye-voice span ^ v 4 Eye-movement habits \
5 Knowledge of 1 5 Early training ) constant
foreign languages 1 6 Practice (
6 Word familiarity ' variahie 7 Age ) .
(or rather novelty) ( 8 Purpose of reading > vanable
7 Health I 9 Conscious attitude )
8 Fatigue
Objective Objective
9 Illumination 10 Nature of topic
11 Word-length
In respect of these lists some questions arise. E.g. the nature of topic and (partly) the word-length have influence in sofar as the subject is not familiar with them, for example owing to lack of practice, which can be due to little early (or later) training.
Age can exert influence via practice, powers of comprehension, natural alertness, or visual acuity, health and fatigue. The readingattitude (whether conscious or unconscious) covers the purpose of reading. We may conclude that such a distinction between subjective and objective influences, further separated in speed and error, cannot be carried through. Instead we propose the twe following lists of factors, influencing speed of silent reading. In spite of the relation between the reading subject and the read object being bilateral, we noted the factors of those relations, that consist rather of a reaction on the concemed subject, at the object-side, and of those, based on a preconceived purpose of reading on the subject-side.



Objective.
1 Typographical factors
a type 1 design (skeleton, serifs)
2 horizontal extension
3 vertical extension
4 amount and distribution of weight (contrast and boldness)
5 letter-space b word-space
c length of line d leading e margin ƒ paper and ink
2 Natural clearness of the single characters or figures and of the words or groups of figures
3 Reading conditions a illumination
b distance
c reading-poise (position of text sheet, e.g. at an angle with axis of vision, in rest or shaken) d distractions
4 Affective value
a reading act itself b contents of text c form of text
. 5 Familiarity of text
a knowledge of contents (special knowledge of or understanding for the concerned subject)
b knowledge of form (special knowledge of the language or terminology, of the location of important items etc.)
Subjective.
1 General faculties
a visual acuity )
b powers of comprehension / variable (health, fatigue) c alertness ;
d general practice of reading ) e general knowledge of language I constant ƒ eye-movement habits /
g (eye-voice span) ) . h (vocalization) \ m some cases
2 Attitude (intensity of effort)



3 Reading-purpose (way of working-up the meaning of the text) e.g. of copying, finding something interesting, new, unusual etc., learning by heart (auditive), understanding (by following the logical consequences either in behalf of a subsequent conclusion or for reproduction, etc.), aesthetical judgment or delight (either over form or contents of text), etc.
A discussion of the nature and relative importance of the influences, caused by these factors, will take place partly together with the discussion of the experimental methods, partly with the analysis of the experimental results.
CHAPTER 2.
Theoretical foundations of the legibility of isolated letters.
Every investigation, which does not stop at stating the superior legibility of a certain typeface above another, but goes into the causes of such a fact, ends in framing prescriptions for future typedesigning. Among these rules, given by previous authors, the greater part has not been founded upon an experimental basis ; they were off-hand constructions (though supported by experimental knowledge), intended to diminish the possibility of interchange between similar characters. They started their arguments also from a fictitious Standard, representing printing-type in general — which usually turned out to be a modern face of the late 19th century style ; at best an old style-revival of the same period was added for comparison. Today we have to put up with such a variety of forms, that we are obliged to take into consideration at least a number of the most important groups.
If the causes of illegibility of a certain booktype are searched in failures in the design of single characters only, the implicit assumption is being made that the reading process of cohesive lines of type does not fundamentally differ from that of isolated characters.
The psychology of reading teaches us that this is not true *). The study of the shapes of single characters only, means an arbitrary limitation of possible causes ; it has to be done side by side of research of assimilative reading. Nevertheless it remains interesting if we could ascertain the relative legibility of the different lettert of the alphabet, deduced from the obser-
>) The difference between the reading processes of isolated letters and coherent words will be dealt with in Chapter 5.



vations, made with continuous texts in assimilative reading. Pyke attempted such an operation ‘by counting the number of times each letter of the alphabet was read by each subject and dividing this figure into the number of his errors’, but he stopped at the quantity of difficult statistical calculations. Besides, it can only be done in experiments as Pyke’s, who counted errors in oral reading and not when amounts of texts in silent reading form the criterion. But it is improbable that we would find anything else than the fact that the most used letters contribute for the greater part to illegibility, or, at best, that some characters, which off-hand observation shows to cause easy interchanges, such as c, e, m, n, are in fact guilty in this respect, to which degree, however, we cannot calculate. It therefore seems far more promising to study the legibility of the various letters in isolation, which method may yield indications about subtle factors, useful to future designing. Moreover such a staccato reading by successive apperception of separate characters, with very limited assimilation, or altogether without it, is done frequently in practical life. Various authors have pointed out that a decrease in speed of reading is caused by an increasing number of eye-movements. Ultimately spelling is done. This can be the case if:
1. the subject is not familiar with reading itself (children) ;
2. the meaning of the text is not clear ;
3. too few characters cover the aera of distinct vision, which depends on the size of the characters ;
4. the object is not clear (shape of the letters or poor illumination), etc.
A good deal of what we are reading everyday is expressed in characters, that are large as to call attention (which is something quite different from being legible), or that appear in poor conditions of visibility ; we are reading many names and other words, with which we are not acquainted. This will often be the case in advertising, telephone directories, catalogues, lettering of posters, shopfronts, signboards, which are usually given under unfavourable conditions (minimal or excessive distance, size or illumination, minimal time of exposition).
Because we attach such a great value to this problem, we shall devote part of our study to it.
The shape of isolated letters varies in two respects :
1. variations of the prototype (shape in strict sense) ;
2. variations in thickness.



These possibilities have to be gone into separately.
By prototype we mean the abstract shape of a certain character of the alphabet. Every actual letter is a concretion of this idea. A letter consists of a basic form (alteration of prototype) plus additional thickening of certain parts. Under the influence of fashion the prototype may happen to change gradually (e.g. arabic figures).
In our study we shall come across errors in perception, not due to deficiencies of the sensory organs. These inadequacies of stimulus and perception have been explained in a satisfactory way and their laws formulated principally by the Gestaltpsychologists. We shall try to explain our results from this point of view.
Let us recapitulate the main points.
Against the theory of the associationists, the Gestaltpsychologists state*) : “The typical form of being of our experiences (both simultaneous and successive) is neither built up out of summative, real elements, nor can it be dissolved in such, but the experiences constitute usually (chaotic masses excepted) a distinct characteristic coexistence, i.e. circumscript entities, often subsumed under a centre, to which the other parts of the entity are subjoined in a hierarchical system. Such entities are to be called “Gestalten in strict sense. These “Gestalten” are in no way less direct than their parts ; often the whole is perceived before the simple parts become conscious at all. Therefore the description of experience cannot start from the sensations, but instead from the Gestalt and its properties.”
Following the associationist-theory, two identical stimuli must give the same sensation ; whereas different stimuli must give different sensations. Numerous phenomena on the other hand prove that this is not the case. The individual gets different sensations from identical stimuli under certain circumstances. The theory of a fixed relation between stimulus and sensation therefore, does not fit the facts and is substituted by a biological point of view : the relation between Man and his surroundings. The constancy of sensations is dropped and a plurivalence (Mehrdeutigkeit) is introduced. Experience depends on the complex factors of the stimulus situations, its meaning to the subject, or on the subjects attitude.
The Gestalt is influenced by two factors specially : the physical stimulus (y') and the complex (configuration, C) ; this can be
i) Koffka, Beitrage zur Psychologie der Gestalt Ba 1, Leipzig p.



expressed in the function : G = f (y. C), if, in the visual field, distance and surveyability are kept constant1). The physical stimulus itself, of course, is not plurivalent, it gives a fixed retinal image, etc.
The circumstances, under which the Gestalten make themselves observed particularly, and that lead to their being investigated, are those of the gross inadequacies of stimulus and sensation, as shown by the geometrical-optical illusions.
In physiological respect, we must not suppose (in case of these gestalt-experiences in a narrower sense) “ .. . . the stimulation of sensations, who are funding them and moreover another “ Gestalt-stimulation”, but the whole process is specifically another one ; it depends on the circumstances whether we are confronted with Gestalten or with “sensations” 2)’\
It is well known, that the modifications of form are governed by two principles, nivellation and precization, both serving the same purpose of making a “good”, pregnant form. A further determination of this term “good” is not given by the German Gestalt-psychologists. “One looks in vain”, Washburn justly remarks, “for a deeper lying principle, to explain why one figure should produce emphasis and another assimilation”3). Is the good form the aesthetical one? Brunswik called attention to the “kalotrope” tendency; Wulf4) mentions the problem in view of G. E. Müller’s hypothesis of affective modification, but he admits that it can only be looked at as a solution for rare occasions. Indeed, with affects we cannot explain the various phenomena.
Washburn introducés the motor factor. According to her the pregnant forms have their elements distributed in such order as to suggest a single response movement on the whole group. However, this seems to play but an indirect part in several phenomena (e.g. in the a, fi and y-illusory movements), viz. only in so far every perception contains a motor component. Besides, the application of the laws of the ‘totality’-viewpoint is not limited to the sensory field. We therefore need a more general principle than motor response. The modifications are only observed when the stimulus is weak or unusual, (to go on with teleological terms) they serve to make sense for us out of these unclear or queer data. Is it therefore unlogical to make use of our general experience as a possible means of explanation ? To quote Wulf
*) Koffka, Beitrdge zur Psychologie der Gestalt Bd I, Leipzig 1919, p. 249. a) Koffka, ibid. p. 251.
s) M. Washburn, Gestaltpsychology and motorps. Am. J. Ps. 37, 517.
■*) F. Wulf, Ps. Forschung I, 346.



again : “The organism in its reaction on the stimulus makes use of certain proceedings, to which it is accustomed, of structures” 1). Wulf applies this point of view only to the apprehension of a figure, according to a known structure (sample), but we may generalize it; indeed not in the sense of conscious comparison between the reproduction of a certain object and the present stimulus. On the other hand W. denies that the reproduction of the average of earlier perceptions could have any influence, because, to his opinion,- this would produce a nivellation in all cases ; consequently precization would not be explained. If a “good” form is the one that makes sense, that gives a solution for the problem presented to us by the weak stimulus, then out of the scala of residues of previous experiences, which are to be considered as possible Solutions, simply that solution will be taken, that comes nearest to the stimulus, giving either “less” (nivellation) or “more” (precization). If the stimulus is nearest to a “lower” unit of experience, but shows a difference in a confined space, which would bring it nearer to a “higher unit, the process, called “pointierung” by Wulf (exaggeration of a minor point in an otherwise normalized figure) will follow. Experience is not a factor among others, but the central factor 2). The “good curve”, equality, neamess, etc. have only influence inasmuch as they are founded in experience.
With the introduction of this supposition the way of events in visual perception would be the following :
Our everyday experience of the external world follows the rules, laid down by the Gestaltpsychologists, which we can pass over here. Our perceptions (of which the motor component has not to be neglected) leave tracés ; in every new perception the related ones are activated. In this way we are constantly building up a complex of experiences, structured by “total processes , constituting a tooi, which enables us to tackle almost any stimulus situation with immediate success. It has the task to maintain the continuity in our perceptions against the discontinuity of the physical facts (stimuli). Under widely different conditions we are able to recognize a known complex ; it can be “transposed , so that the greater part of its original local stimuli are absent. Of visual objects, which we are used to recognize by their shape and colour, thanks to the constancy of colour and size, one of these properties may become unrecognisable in itself, provided that the other remains sufficiently above a certain limit. If this
‘) Op. cit. 373. .
2) It has to be emphasized that experience in our sense is essentielly aitterent from the associationist conception, against which Köhler has argued.



condition is not fulfilled, we become subject to curious illusions. For example, we discover something unusual among the trees, surrounding a meadow, which we put down as a rag of blue cloth, until the thing detaches itself from the foliage and we recognize it as a white goat in the shadow. The blue colour changes in gray and we are unable to see blue anymore. The shape of the goat was covered by the leaves and therefore unrecognisable, consequently we had no reason to see white-in-shadow and the nearest possible solution was a rag of blue cloth.... In other words : we had no sufficiënt stimulus material to our disposition to make still a white goat of it. No ‘error of judgment’ was present, corrigible by critical observation ; the blue colour was given with perfect evidence; the rag was not seen horizontally, shortened by perspective, but hanging vertically.
Let us consider a few of the Gestalt phenomena in a more narrow sense from this point of view.
One will remember Wertheimer’s a-movements. Two overlapping lines of different length were exposed successively ; they were seen by the subjects as a single, contracting and expanding line. A physiological theory was presented by the Gestalt psychologists, making use of short-circuit processes in the brain.
Psychologically we must suppose a reproduction of all those cases, in which we saw an object moved quickly to and fro ; if we try to imitate the situation with our arm, functioning as a short and long line, we are obliged to pass the intervening stadia as well. Therefore, if the two lines mentioned are presented to us, we are completely satisfied with seeing a single moving line ; we have understood the situation and no critical observation can make us drop this view, because the “real” state of affairs is meaningless to us. If, however, the interval is reduced (the movement slowed down) and consequently the single phases of the line are convincing, we don’t have any more reason to explain the case to ourselves in another way, because we know it from experience with the same certainty as we are acquainted with the moving line. However we know anomalies, for which we can give only a physiological explanation at first sight, e.g. some figures of Ehrenstein 1).
This author gummed geometrical figures (some of them coloured) on a horizontally moving belt of paper, over which a fixationmark was fitted. The speed varied from 20—30 cm. per second. It is to be regretted that no extensive protocols are given by
') Z.f.Ps. 97, 161.
Legibility, atmosphere-value and forms of printing types. 2



more subjects ; Wulf’s experiments (see p. 19) make differences in the observation very probable. Ehrenstein exposed e.g 4 verticals, which were seen multiplied, though the subjects could not teil the exact number. Similar phenomena are presented bv the disk of Purkinje, of which the radii are multiplied. The author speaks in physiological terms, of shift of retinal images.
Psychologically speaking: only a unit of four lines would have passed-by intact under the fixation mark. The umty was apparently disturbed, the interlinear space probably being too large. Now not ‘a block of four lines’ was seen, but simply ‘some lmes . Perhaps no modifications would have happened if they had been grouped two by two.
Three vertical lines (that had more space between them) were each seen larger than the foregoing one. It is possible that they were experienced as three successive strokes (in proper sense), in a rhythm as usual with hammer-blows. This innervation with increasing force is also easier for the writing muscles if a few strokes are to be put down. For a greater number of strokes some other kind of dividing (rhythm) is chosen.
The other figures presented by Ehrenstein can be similarly dissolved into parts, either as figure and background, or as container and contents. Ground and container are fixed ; fagure and contents are “loose”, “we can take them off . If we should form Ehrenstein’s figures out of pieces of cardboard, or flexible copperwire and we should push or draw at the “fixed part, the “loose” part would remain lying, or the wire would be twisted in the same way as the figures of the experiment. A diagonaJ line was seen as curved as an integral-sign. This was possible because we can draw more easily a diagonal in this way than straight (cf. the “rounding-off” tendency observed in our expe-
The MÜLLER-LYER-illusion is explained motorically. The illusion is brought about when the hooks are long and not too dose on the shank. One could think of the arrow-form, which should however, produce the opposite effect. But arrows suggest speed only if the hooks at the top are short and close on the stem. Otherwise they loose their power and bear but a symbolic value. On the contrary, one line is a long, slender thmg, which we can draw out, the other is short and bulky, inviting us to press it together. Both are equiped with convement grips, so that we
will not hurt our hands , ... - _
We must regard the results of Kqffka and Wulf ) from
i) Ps. Forschung I, 333.



a similar view-point, connecting it with Werner’s paper on the problem of motorical form-givingx).
It will be remembered that Wulf presented a number of geometrical figures to his subjects, which he asked them to copy from memory a few times in the course of some weeks. The drawings, obtained in this way, showed interesting modifications from the original model. Either they were normalized, or exaggerated (precisized), or, in some cases, partly normalized, partly exaggerated (“pointierung”). The modifications became more and more conspicuous in the course of time.
Werner made his subjects draw the same figures several times, e.g. a simple right angle. The modifications appeared to depend on the subject’s attitude. The hook was either drawn in two movements, as two segments of a circle, with the chords at a right angle and on the outside, or nearing a quarter of a circle, done in one movement. The figures of Wulf were indifferent to the subject. We can say that they were truly determined per genus proximum (of easy writing-movement, either nivellated or emphasized) et difjerentiam specificam (deviation from one of these in a minor point, “pointierung”). Wulf denies the possibility of experience in our sense, because, e.g., a subject called one of his figures a water-bottle, which, as he says, should have produced emphasis in the copy, whereas the drawing on the contrary was normalized. But here two essentially different functions are present, drawing versus name-giving. If my hands choose to follow the shortest way (nivellation) and I was asked to explain this act in words, it would be easier to say “waterbottle”, which is an unmistakable shape, than “closed, curved line, round at the left, bending inwards, going on as a straight line after that, etc.” Wulf makes a dinstinction between the “comprehensive” type (which takes the drawings as “real” objects of our daily experience) and the “isolative” type, which takes them as abstract figures. This distinction has no more than a systematical significance, it cannot be allowed tö serve as a principle to make an essential difference between the ways in which the figures come about, e.g. at the assumption that memoryaids are specially active in the comprehensive type. In the isolative type they are only more of another kind, viz. motorical. That the deviations become more conspicuous in the course of time is indeed explained in the way described by Wulf (page 369).
“For the “pointierung”,” he writes, “one could venture the ïxplanation, basing on association-rules, that the deviation from
) Z.f.Ps. 94, 265.



the “normal” is impressed by way of judgment. lms juagmeni would be reproduced and would determine the construction of the copy, the particulars of the optical onginal benig forgotten. But this explanation fails. It is not explamed why the deviaüon is intensified from copy to copy, unless one supposes that the influence of the residues of the perceptions on the copies dimimshes in the course of time, so that it is to be taken as a mixed associative effect ” The argument neglects the motor element in our perception (and consequently in its residues). An extra judgment is superfluous, because we recogmze the movement directly. From the first we experience it from either the normalizing, emphasizing or “ pointierung” view-point, on the base of t e ^Ven stimuli, that fade away quicker than the older associaüons of our experience, according to Jost’s second law (page: 368). Wulf is therefore right when argumg against G. E. Muller, that normalization is not due to the wear and tear of time.
The application of this view to the Gestalt-problems of letters would lead to the following conclusions :
Every known object is a Gestalt. An object is not recogmzed if the Gestalt is disturbed in some way, which depends on the degree of its cohesion. In this respect we can discem various levels. The highest level (in other terms : the preferred form) is constituted by the simplest, easiest and best-practised figures, such as circle, straight line, square, triangle, a.s.o.
We spoke of the motor component in experience. Because we do not only read characters, but write them too, the motor element in the apperception of letters is very strong. Therefore we may assume that the most coherent Gestalt, m the greater part o cases is the one, that is most easily wntten, ï.e. with a minimum of interruptions and sudden changes of direction, wh!ch require careful innervations. In our expenments we observedadisunction or even a partly disappearance in figures, wntten with reversion or radical change of the wnting-movement, e.g. in
L’The Characters of the alphabet are made of a few elements, which are altered in various details. The old theonsts of the Renaissance have already developed their speculations from the fact that our alphabet is composed of straight lines and arjrs f^ a circle. Geoffroy Tory even dragged the unhappy Io of th Greek saga’s into it. Constructions of the past and the present bS this principle. Thus we are placed before the
task of constructing the forty different figures of the upper and lower case alphabet with the aid of these two ele^ents^only and those new structures have to show a good deal of difference



too, in order to avoid mistakes between them. Because with two elements the range of possibilities is soon exhausted, we try to bring more variation in the resulting forms by making slight alterations in similar parts, e.g. by cutting parts out of circles, squares and triangles, or by adding cross-strokes to verticals.
But our work threatens to be spoiled by the illusory movements mentioned above: entities, which we wanted to be disturbed, are either restored or further (but too far) disturbed. Even totally new, meaningless structures are built up, which bear no resemblance to known characters. For example : the L becomes a simple vertical again; the gap in the O that made the C is filled up. Our task must therefore be to reduce the form of letters to the most coherent Gestalten, which are not subject to further changes, or to make a new, strong structure, with openings, too large to allow completion and members, too firmly fastened to sulfer separation.
There is yet another aspect to this question : we do not alter the components of the letter or the basic letterform only to prevent mistakes between them, but also for aesthetical reasons, when we are not content with the usual shape. In that case the very coherence of the Gestalten, that counter-acted our manipulations in the former case, can be made useful to save the recognizability. It brings about a constancy of form that opposes the modifications for aesthetical purposes.
If our theory is correct, and if we, when placed before the enigma of an unclear character, accept the next possible solution, it becomes important to know the hierarchy of the different levels of coherence.
Belonging to the same level, and therefore mutually (not unilaterally) interchanged are the T, F, 7, Y and V. The cause of the alterations, observed with these characters, is to be found in the fact that these digits perform a connection between two points on the top-line with one point on the base-line ; consequently the horizontal, vertical and diagonal are equally characteristic for this shape. This is the same case with C and L. We know through Werner’s experiments (see p. 19) that a hook is either drawn as a C or as two concave lines in L-shape ; both are equal Solutions. Our L was seen as C, Tinker’s C as L. Such observations make it possible to learn the relative strength of the various figures and to predict the possible interchanges in newly constructed types.
It may be clear that the recognizability of a certain letter cannot be deduced from the quantity of area it has in common with



some other letter, and still less the nature of the occurnng mistake. Our perceptions of letters do not mind about quantitative differences in the first place. Such an operation, though, has been effectuated by Legros and Grant. They copied one of two characters of similar structure, e.g. h and b, one upon the
K-X .L-C.r-T.t-C t-H,4-A-<.N -N L - C , F - r - f -t X-K-I,L-l,K-k.T-Y B8-8.q-g-9.9S-QO H-8,X-X.i-K,e-896
ffy-yt-t, d-d
FIGURE 8
Typical modifications of form.
other, and measured the overlapping areas. From this figUI"e they computed a “specific legibility coëfficiënt” for the whole typeface. Such a formula may have some value, but it cannot teach us enough.
Some of the most typical modifications of form are reproduced in fig. 8.
CHAPTER 3.
Variations oj the shape in proper sense.
Series I. Lower case bookfaces.
Method.
Only a few investigators used isolated letters; others however, used words or lines. In both cases no different forms of the same character were compared, but rather the relative possibility of interchange was searched between the various characters oi the alphabet at one and the same typeface. In consequence the analysis of the results and following conclusions are unable to answer the question, to which degree the basic form of a character



may be modified without becoming unrecognisable. The reader will remember that we are pursuing this aim (to know at which costs of legibility the artist can realize his intentions) rather than searching after the only and unique form.
The central difficulty of every investigation into reading mechanisms is to be found in the circumstance that numerous compensating processes are tending to diminish the differences between the various things read. In order to obtain dependable results the conditions of reading have to be aggravated considerably, however without making them too unnatural. In previous experiments this has been done along four lines :
1. time of exposition
2. distance
3. illumination
4. fixation
A reduced illumination, without time-limit at normal distance, allows uncontrollable visual difficulties to enter with the subjects, so it was doubted whether any results would be obtained.
The focal variator (see Burtt and Basch, also Weiss, J. Exp. Ps. 1917, 2, 106) would probably have yielded good results ; we did, however, not possess such an apparatus.
At first an ordinary tachistoscopic method was tried out by using a Michotte projection-apparatus with rotating disc and sector. The characters were copied on a diapositive glass-plate and projected on a milk-glass screen. This tachistoscope has many advantages over others : fixation and accommodation are excellent, it works inaudibly ; no moving parts are in the field of vision. The differences in accommodation were slight (the screen was evenly lighted by a 100 Watt bulb overhead, the projected dark-grey letter appearing on it in a little, white patch). It proved to be impossible to obtain any notable differences in the types, even with a minimal opening of the sector, ± 50 o, height of character 3 mm., and with reduced light. So this method was rejected.
Next a combination of short exposition and false fixation was chosen. Instead of the Michotte apparatus, with its unpractical diapositives and unsharp images, an ordinary photographical focal plane shutter was used with an opening of 20 o. It was placed exactly midway between the headrest of the subjects and the exposed object (the distance being 46 cm.). A glass-plate, fixed before the black curtain of the shutter, gave a total reflexion of the subject’s head this Virtual image made an exact accommodation possible. In order to obtain a peripheral vision the



fixation-point had to be shifted for about 5 cm. at the side (ï.c. the left side) of the object, but of course it was placed on the shutter1). This was made possible by projection of a narrow pencil of light on the reflecting shutter-glass from a 6 volt bulb at the left hand of the subject (mounted in a closed tube, perforated at the end with a needle). The shutter was turned
n U/Mit 70° tliA nnpnino
ior duuui ju , liiv/
I remained sufficiently wide to allow a perfect view of the object. The mirror-reflex of the subject’s face took away the dark colour of the curtain of the shutter; behind it the exposed letter appeared on a matt, white paper in a square of 5 cm., cut out in a vertical, wooden board, covered with white paper. See figure 9. Though the adaptation was not perfect, this cannot have influenced the results considerably. The exposed object was lighted by a 100 Watt bulb above and slightly behind the subject’s head. No daylight was admitted. The subject himself released the shutter when fixation and accommodation were complete and the
attention reached its maximum. To our regret the snuttermaae a rattling noise and a clap when opening ; an observation of the subject’s eye-movements showed that sometimes a flickering of the eyelids followed the exposition rather late. Still, this fact has to be taken into consideration. On the other hand I do not know of a tachistoscopic apparatus, that has no drawbacks, either psychologically, in regard to fixation, accommodation, noise, etc., or technically, in regard to the necessity of making diapositives, reverse copies, etc., with their unavoidable costliness or slowness of manipulation.
The subject’s attention was not diverted in another way, because a greyish, hard paper, extending from both sides of the object,
i) Because the subjects discerned many details at the right side ot tne exposea character, a control series with the fixation point at the right side was thought unnecessary.



hid the rest of the room from the subject’s sight. The experiments (which lasted two sessions of about 50 minutes) were conducted as follows : the experimenter inserted a card with the character in the holder and said “yes”, the subject kept the shutter-release in his hand, waited till his eyes held the fixation point steadily, then pulled off. Next he described the shapes he had seen, drawing, if necessary, on a paper that was lying close at hand. The results were written down by the experimenter, during which time the subject got the shutter ready for the next exposition. When the described letter was not the right one, or more details about it were required, the subject was asked to give an exposition once more. This did not influence the subjects as much as one would expect, because no comments on the findings were made by the experimenter ; repeated expositions were asked also when the subject himself was sure to have recognized the character allright. The instruction asked expressly for a description rather than for a recognition. Tinker asked for a character, and so made way for guessing, as it seems. On account of our own experience at least it is highly improbable that a Q is really observed, when an O is given, as happened in T.’s experiments. Likewise I doubt whether the root-sign, frequently seen for F, V and Y, is really recognized as such, opposed to V and not rather as “V-shape”. A n was also seen too often.
If this should be the case, the reliability of Tinker’s results is diminished, though still yielding interesting data, which show a high correlation with our own, specially the misreadings with highest frequency. That a mere quantitative statistical operation is misleading, shows his record of lower-case ‘o’, which ranks under ‘a’ in legibility, because it was 79 times mistaken for either a (30%), O and n (10%), c (8%), u and %-sign (5 %), as against a 75 times for either n, o, s, n, d, u, z. (When the “no response” answers are taken into account, o ranks immediately above a). It is more important that instead of ‘a’ in 12% of the times ‘o’ was read, than that V in 30% was read for ‘o’.
The objects were given in random order, but care was taken that each time a different letter was exposed. Five subjects were college-undergraduates, the director of the laboratory being the sixth. Their eyesight was normal, or corrected adequately. The sessions were not tiring, the descriptions and drawing of the perceptions and a few moments of talking gave the necessary interruptions. The whole situation, giving a high attention and minimal fixation, was very suitable to study the y-movements.



Material.
The exposed material consisted of matt, white photographicpaper, with one character in the middle of it. See fig. 10 1). The characters were drawn on large scale and photographically reduced
to an x-height (without descenders and ascenders) of exactly 5 mm. and uniform thickness of 0.8 mm. This becomes of some importance for the types with a big face, because the characters with descenders or ascenders, having the form designed according to these dimensions, suffer relatively more from the reduction than the projector-characters of small-faced types. In consequence the exposed character can only be taken as a fictitious representative of a method of varying the basic form, not as a specimen of the type it is named after. The designs were based upon the most notable variants of the basic form, as used in different booktypes of today. Abstraction was made from thickness, not from the relative length of descenders and ascenders. With some types the serifs and the thinning of the curves could not be omitted or substituted by strokes of the adopted thickness (0.8 mm.), without changing the form itself, as these features constituted an integral part of the design. In the majority of cases, where this could not happen, some sans-serif type was
i) The original characters were destroyed in an accident; this figure has been drawn after the copies.



chosen. In many cases a modern-egyptian made it possible to tracé the influence of serifs ; but it should be kept in mind that these serifs are much heavier than the usual ones.
Several less important forms were added to obtain sufficiënt material for comparison (each character minimally in 3 different editions), or to study more minutely the occurring Gestalt phenomena.
The investigation dealt only with the lower case a, b, c, e, f, g, h, i, j, k, m, q, r, s, t, u ; totally 63 different forms. This was done first because the other characters and the upper case show very little variations (which can, however, be understood in analogy to the others), secondly to reduce the quantity of the studying material. Also figures and cursive (or italic) forms were excluded, because this would increase the task enormously. For this reason we shall only be able to make some suppositions probable; however, the question remains to be gone into thoroughly. In the series II upper case and figures were included.
Some of the results of Tinker’s first series will be cited for comparison. They were made on 8 point Linotype No. 16, a face, almost identical with the Nordische Antiqua of the German Linotype. A Wirth disc tachistoscope was used with an exposition-time varying from 0.017 to 0.0107 second.
Results (see fig. 10).
a.
1. Futura.
A pure circle with a vertical at the right side.
Two subjects saw an upright cursive a (short upturn at the base), others a flattened circle, with a distinct vertical bar. All subjects recognized it as a. Though on account of the protocol this bar might be less pronounced (one subject hesitated at the first exposition between a, d, q, another saw a normal a with an arch), the addition of weight (series II, pg. 59) makes the loop so predominant as to cause interchange with the o.
2. Nobel.
A flattened loop, fixed at two points to the vertical, the arch clearly abo\e it; “normal” shape.
All subjects recognized it as a, be it as an archless one in two cases, where the arch became fused with the loop. Though recognition as letter a is assured, which makes the question rather unimportant, we can take this quantity of space between arch and loop as a minimum and prefer, especially for the bookgrades, a loop as the Garamond shows : a tiny loop, hanging diagonally from the middle of the vertical bar.
3. Futura.
Unpublished edition. A vertical bar with a horizontal to the left, the curl beginning at the bottom and ending abruptly. Designed with the presumption that a hook to the left, with something round beneath, would be sufficiënt to give the idea “a”.
Three subjects described this figure as a normal, complete a; two failed to recognize it at all — they saw a circle with something attached to it somewhere — at any rate they saw some detached parts without sense; the 6th subject started in the same way, but ended with seeing a normal a with a closed loop.



This simplified “quintessence” of the letter “a” appears to be not successful in all cases as for suggesting the usual form.
4. Memphis.
Similar to 1, except the serifs at the ends of the vertical bar, extending to the
right side. , ,.
Here again, as in 1, the vertical bar suggests d or q, remarkably not hindered by the serifs ; they are rather the cause, being more usual on d and q than on a. Generally the vertical bar is seen as fixed rather loosely to the side of the full round circle. The recognition followed with some reluctance, for three subjects repetitions were necessary.
5. Bravour. , . ,, ,
A flat, bent arch, with serif at the right bottomside; a big, o val loop, close under the arch, fixed at two points on the arch. Extreme situation of a variant, popular with many refined display-types.
Three subjects saw this letter as a normal a with an arch, one did not see any other thing than a very flat, broad 8, the other two saw a flat, archess a, one of them only after 6 expositions. We may take this form as too broad and flat, the arch too close on the loop.
Concl. : The archless a appears to be apt to disjunction into circle and vertical, in consequence to be less legible than the old-style form, with an arch on a vertical and a little loop amply beneath it. For both kinds of design the vertical movement has to dominate the necessary round parts. Tinker’s subjects exaggerated every possible trait : round (o), circle with vertical at the right (d), diagonal from top left (s), idem from right (z), two verticals (n, n, u). This is due firstly to the aspect circle with horizontal” (cf. e), secondly to “two verticals with connecting strokes”. The small loop will preclude both possibilities. Tn conformity with Maire, Javal demands an inversed r, in other words : without an arch and the loop substituted by a curl. This seems so far removed from the usual shape, that new difficulties will certainly arise.
b.
To my regret the proportions of three exposed b’s did not quite reach the Standard of height and thickness fixed for the others. However, we will be able to learn something in comparing them with each other.
6. Futura.
A vertical bar with a fully round circle.
All subjects recognized it immediately as b.
7. Memphis.
As 6, but with serifs to the left. . , , , . *. .
Recognition followed not so easily as in the case of the 6, the subjects found something unusual at the bottom or at the(top of the vertical bar, not specihc to b.
8. Goudy. . , , , 
The shank is bent away a little at the base-left; the loop is of a diagonal ova
shape, fixed to the end of the shank. 
Three subjects started with seeing an h with the right leg bent inwards, while two subjects kept on doing so. The recognition was reluctant.



The reason for this is to be sought in the fixation of the loop — not above — but at the end of the stem. This makes the situation at the base-line unclear.
Concl. : The b should have a generous loop, fixed a little above the base of the stem, to avoid “Pointierung” of the lower left corner. It is important that Tinker’s s.s. mistook p, b and 9 for d, q, 9 and 6 and vice versa, which is corroborated in both our series I and II. Apparently the very feeble structure of these figures is due to the incompatibility of the dissimilar movements “straight” versus “round”.
c.
9. Nobel.
The circle cut out along the radii of 195° and 135°, usual form.
All subjects recognized it immediately.
10. Futura.
The circle cut out along a vertical line, very little to the right of the top and the bottom.
One subject required a repetition, another thought it a bad letter.
We may take it as too narrow.
11. Memphis.
Similar to 9, but with a vertical serif at the upper end instead of curl.
Perfectr ecognition.
Concl. : The main differences in forms of the c (C) are to be found in variations of thickness ; therefore the results of series II will be more instructive. Here we can only ask for a not too narrow form.
e.
12. Garamond.
A high-placed horizontal cross-bar.
Only two subjects did not hesitate in the recognition; three of them described “half a circle, with something high in the top inside, perhaps a c”, while one of them fancied to see an a with an arch close to the loop. The eye is far too high (in reality the Garamond e has a still higher placed cross-bar). This form is not recommendable.
13. Veronese.
A diagonal cross-bar.
As in the preceding case, interchanges with the c or general vagueness made several repetitions necessary, though the lower position of the left end of the cross-bar appears to help a little.
14. Futura.
A horizontal cross-bar exactly through the centre, the lower half of the fully round circle almost touching it.
Five subjects did not recognize it, even after 5 or 6 repetitions; they saw instead some winding movement either as s, 8, Futura-g or G.
The eye seems to be too big, the whole form too round, the opening at the right side too narrow.
Concl. : These results are not astonishing. The e is in the unanimous opinion of several investigators one of the least legible characters (Cattell, Sanford, Roethlein, Tinker). Two things



ire to be avoided : the restoration of the complete circle and a predominance of the horizontal, the winding movement being :aused specially by the latter. Something might be said in favour of an oblique crossbar, as Javal and Maire propose. In this way a part of the eye can be kept rather big, without disturbing the optical balance, while the winding movement from upper left to under right is counteracted. At the same time the opening at the right side is wide. When this is not possible for reasons of style, the cross piece should not be placed too high. A narrow (oval) shape is preferable. The addition of thickness (series II) leads to opposite observations and conclusions. Tinker : interch. with c, o, o.
f.
15. Memphis.
Extended under the base line ; very narrow, at the upper end only a smalt hook.
This type was never recognized at once, even by two subjects not at all.
There were interchanges with j or r. In some cases the cross-bar was not seen, neither the hook. The length seems to be too large in relation to the size of the hook, the cross-bar too short and high-placed, the hook itself insufficiently clear.
16. Futura.
A long ascender, broad cross-bar and generous hook.
One subject mistook the arch at first for an r, and then for a T with an oblique cross-bar, but the others recognized it immediately.
The width of the design' at the right side of the vertical takes up the room for an extended cross-bar to the left, because the printing-type f cannot be too broad.
17. Garamond.
An arch, starting from the cross-bar and hanging forward considerably.
The exposed character was executed in the original thickness, because the abstraction of it would alter the whole form too much. This has to be taken into account.
The big arch always called the attention, the cross-bar was overlooked a few times, especially by one subject, who assisted also to series II and was in the habit of describing as r everything vertical with something at the upper right side.
Concl. : As the f is an intrinsically legible character, the differences in design will not have great significance. Still we may ask for a generous curl or arch at the top and reject the variant with a simple bending, or the “crochet-hook” of the slug castingmachines1), because the vertical tries to straighten the loop. Tinker also observed interchanges with t, 1, 1, j, and r. The form 16 appears to be sufficiënt, though an overhanging loop increases the legibility.
g.
18. Old Style.
A circle measuring about 2/3 of x-height, with a small horizontal at the upper right. The loop starts in the lower left corner, crosses the base line and bends to an angular loop, which ends at the Crossing of the base line.
*) See Introduction, “logotypes’, pg. 4.



AH subjects recognized the figure at once, some needed repetitions to describe it more correctly. They all saw the two circles linked up by a little line, and the small flag at the upper circle. A good, clear form. (Also a proof that the fixation at the left side did not influence the clear perception of the right side of the object.) The g of Tinker’s type was more broad and vertically more compressed ; it caused all interchanges, observed in our two series : 8, 6, s, %, a, e.
19. Futura.
A circle of total x-height, fixed at a vertical bar at the right side, which ends in an arch under the circle.
This form too proved excellent, no mistakes.
20. Koch.
A circle to 2/3 of x-height, with a curved stroke, directly attached to it. The flag is treated as an unimportant appendix.
Four subjects called it a g at once, but were unable to describe the lower part correctly. Two ss. saw some winding s-movement, or a reserved c with something above it; they did not recognize the letter.
This form has to be rejected, though it will probably show up better if placed beside non-projecters, so that the curved stroke under the circle can be put down as a descender of the g (which does not alter the fact that the form is no speciflc g).
21. Futura (unpublished edition).
A full circle with a rectangular flag, resting upon (but not fixed to) a triangle, with the apex downwards. Experimental construction after analysis of the old-style form.
The results, obtained with this remarkable tentative, are not unsatisfactory and in any case instructive. All subjects hesitated and needed at least two expositions, but five of them described at once an unclear g, although the nature of the lower half was defined differently, either as a full loop, joining the upper circle in the usual way, or as a flat ellipse, touching it.
The horizontal line is overlooked and the triangle rounded off, except by one subject, who saw an S crossed horizontally, but who, on the other hand, identified this character with nr. 18 by memory, after exposition of the latter. The flag was clearly seen by all.
Concl. : The notion of two closed loops at some distance above each other, with a flag at the top one, appears to be sufficiënt to make a g, the lower loop may even be open.
22. Lo.
A broad, oval loop, with a flag on the top, a simple curved sling under it.
The situation above the circle and underneath was not clear to the subjects, they described an a or some intricate curved figure. Three subjects failed to recognize it at all, the others were slow. This form has to be rejected.
23. Memphis.
Principle of nr. 19, but with serif to the right at the top and a very short arch close under the circle.
The lower arch proved to be the greatest difficulty. In three cases it was not seen at all. Th^ recognitions were slow.
24. Belwe Antiqua.
!ow„r Part consists °f a sharp hook, horizontal line and a small curl at the end. Relatively small upper circle, flag placed obliquely. The represented letter was based on the g of the type mentioned above, but it stands for every type which has the upper line of the lower loop extraordinarily stressed.
There were only two direct recognitions. Again the lower part was the Principal obstacle; it was sometimes seen as the tail of a Q, or as a 2 with big :losed loop.



25. Schelter Antiqua.
Big vertical oval loop of nearly the total x-height, a simple horizontal stroke under it, fixed at the right side. Flag reduced to a mere attached point.
A cross-breed of the forms of nrs. 18 and 19.
Two subjects saw the upper part at the right side or below, and failed to recognize it. The others described the lower part differently : loop, bowl or hook. The join at the right is the probable cause of these uncertainties.
26. Bernhard.
Modification of nr. 18: flag on the top, upper circle big, the lower one subordinated.
The total subordination of the lower part, without the possibihty to make it appear as treated after the principle of nr. 19, seems to account for the difficulties. The figure is not clearly organized, the relation between upper and lower half incomprehensible. Three subjects saw the g quickly, two after several expositions, another one not at all.
Concl. : After the foregoing observations and those of Tinker we may conclude that the g must be designed as purely as possible to either of the two principles : nr. 18 or nr. 19. In the first case neither of the two parts should predominate, the separation must be clear, the fixation simple and at the left, the flag must be extending generously. In the second case the stroke under the upper circle has to be an unmistakable bowl-form, fixed at the right side and for the rest also clearly separated from the upper half. Javal is rather puzzling when he demands a flag at the left side of the upper circle “which has been out of use since Garamond”. Even if we read “right” instead of “left the fact remains that Garamond did not change anything at the form of flag, that had been for centuries a simple straight connecting stroke. The rejection of the “italicised form (nr. 19) is determined by his aim to avoid similarities in the upper part of the x-height (i.c. resemblance with 9).
h.
27. Memphis. . , , . . , 
Simple form with serifs at the base line and at the top to the left. Ascender.
rather short. ... , .
Recognized by all subjects at once; but one remarked that the ascender was
too short which makes it easily possible to mistake h for n.
Extremely long ascender. Serifs added abusively. Arch beginnmg obliquely. All subjects recognized it immediately.
29. Futura.
Without serifs. Arch fully round.
Perfect recognition.
Concl. : Intrinsically legible character. Heavy serifs give no interchange with b. A form similar to Benedictine might have been added, which has the right limb bent inwards. The results would probably have been unfavourable, for the round movement would be induced (interchange with b).



30. Lutetia.
Small dot placed high upwards and slightly to the right, diamond-shaned. Tiny serifs.
Three subjects needed a repetition, because they were uncertain in regard to the dot, which appears to be too small and in no direct relation to the stem.
31. Futura.
Without serifs, dot rather high-placed, slightly smaller than width of stem (with the original the dot is in reality a little bit broader).
One subject hesitated between j and i, another did not see the dot quite clearly
Excellent form, the dot might be heavier.
32. Memphis.
Stem with serifs, small round dot close to it.
For one subject the dot seemed grown together with the stem, another one saw a j twice.
The dot may be placed higher, at least the lower serif can be omitted.
33. Memphis.
Identical with the foregoing number, but with big square dot.
Same remarks as to nr. 32, a second subject persisted also in seeing a j.
Concl. : The dot should be big and rather high above the stem, the shape of it is irrelevant. Serifs are superfluous, or even misleading if they are heavy.
34. Memphis.
Rather short descender with short hook, big square dot.
Again the dot was too close to the stem. Three subjects saw no curve at the bottom to the left or only a tendency to a motion in that direction.
35. Memphis. ,
Identical with nr. 34, but with small, round dot.
Likewise the curve was thought too small, but the dot this time clearly visible • the square form is more subjected to growing together with the stem and i’ts upper serif.
36. Futura.
Simple, vertical stroke, without curve. To my regret the exposed character turned out to be a little less heavy than the foregoing numbers.
Four subjects persisted in seeing an i. The two others mentioned a deviation to the left. The curve proves to be an integrant part.
Twice the dot was not seen.
37. “Cochin”.
The exposed type is a fantasy, but my friendly cooperators made the dot too heavy. However the curve is the main feature, it represents every curve that is more generous than the foregoing numbers.
Every subject saw it clearly.
Concl. : The j is an intrinsically legible character. The curve must be wide and may begin at the base line. A heavy dot, not too close to the stem is preferable. Ja val asks for a short des:ender and Maire for a mere point to the left, instead of full 'Urve or tail-dot, with which I do not agree for isolated letters, see also f.
k.
38. Belwe Antiqua.
Based on the type mentioned above, which has in fact a still bigger loop and ihor ter right limb. Big, o val loop instead of upper diagonal.
-egibility, atmosphere-value and forms of printing types. 3



Four subjects saw clearly a b, two of them also an h. The remainmg two mbjects described it properly. Apparently the loop is so utterly dominating ;hat the short, right limb disappears.
39. Bell.
Curved tail instead of lower diagonal.
No mistakes: no curve is seen, but. a straight limb.
40. Futura. ... . ,. ,
Orthodox form, lower limb not fixed at vertical bar, but at the upper diagonal. Two times one subject did not see the upper diagonal. The only explanation 1
:an offer is, that the subject assimilates a part of the upper diagonal with the lower jne, which is made possible by the position of the join. The end of the upper diagonal in consequence detaches itself from the rest of the figure and is overlooked.
This would be a confirmation of the supposition that ‘ unexpected lines can be readily overlooked, as is the case with L, F, T — see also concl. r, series 11. Tinker : interch. with w, h, t, i, b, x, z.
Otherwise no mistakes.
Concl. : An intrinsically legible character. Still the design should not deviate too much from the pure form with two diagonal limbs, neither big loops, nor probably a vertical curved tail will fail to diminish this legibility.
m.
41. Modern. ’ , .. - 
Rather narrow, serif above at the left. broad senfs, that close the figure at
the base line.
There were no mistakes.
42. Futura, (unpublished edition).
Like a Capital E laid down, rectangular shape. Tnal-cut to see whether the
connection strokes have to be arches. .......
Only one subject saw this form angular, the others thought it round. One hesitated between m and n, otherwise no mistakes.
43. Futura.
No serifs, arches, wide form. , . ... „
Two subjects hesitated between n and m. This can only happen when a disjunction takes place, caused by a too broad design (or by the outline ot a narrow form, being perceived only, with neglect of the mternal organization which cannot have been the case here). Otherwise a perfect form. Tinker met also this mistake. His «. answered a, o, n, u for n. If an ordinary a was seen, perhaps a kind of short-cut between two verticals was responsible, similar to those we shall find in series II.
Concl. : The m should not be too broad, indeed like typewriter-m (Javal), perceptibly less than two joined n s. The nature of the joining is irrelevant, serifs do not add much to recognizability.
q-
44. Memphis. . . , , ., . ...
Fully round circle, very short descender, with cross-stroke and serit at tne
top to the right. ... , .
Three subjects hesitated between a, d, p, q, owing to uncertainty about the right half. Perhaps the lower serif is still adding to a growing together.
The loop is somewhat flattened at the top, the stem is bent back and ending in a long point. Representative for the forms, that have the upper right corner kept thin and light.



Two subjects hesitated about the shape of this spot (open or closed like figure g). Otherwise no mistakes.
46. Futura.
Fully round circle with long vertical stem at the right, no serifs. One subject saw a disjunction between loop and shank. Otherwise a good form.
Concl. : The descender should be generous, the loop not too loosely attached to the stem.
r.
47. Fantasy.
Flag in the form of a hook turned upwards. Standing for every form, of which the flag projects high above the middle serif-line turning the other way as usual. The design is not very satisfying.
No direct recognitions could be booked. Two subjects saw a definite arch, like flg. 48 (ignoring the highest vertical part), but wanted to give a repetition! The others described the figure correctly, not calling it r, in which they werë wholly justified. . . .
48. Futura.
Short bow without dot or curl.
One subject saw a straight diagonal line, otherwise no mistakes.
Perfect form.
49. Futura (unpublished edition).
Added to see whether there should happen any connection movement between stem and the separate dot.
Three subjects did not recognize it at all (only a stem or stem with something attached to it), the others made immediately some connection, an extending flag.
If the dot had been larger, probably the result would have been much better.
Concl. : The end of the flag — whether in form of a dot or chub — should not be too thin ; the flag itself not too close on the stem, not too far removed, as assimilation or disjunction is apt to take place. In accordance with Ja val and Maire it should not be placed too low.
s.
50. Futura.
Narrow; upper half small and bent round to half a circle, lower half bigger.
All subjects described this form as extremely narrow ; one did not recognize it, while another took it once for a very narrow a.
The main difficulty forms the width.
51. Memphis.
Very broad, with heavy serifs, almost touching the diagonal. The halves have a distinctly horizontal oval shape.
The record shows only one direct recognition ; the other subjects saw some form of superimposed circles (8 or g), which means that the winding movement, peculiar to the s, gets lost when the diagonal movement is insufificiently stressed' or the sides are not kept clearly open. This form has to be rejected.
52. Erbar.
Both ends are not bent round, but horizontal, thus making large openings. The curves have a comparatively large radius; but the narrowness of the character makes still a diagonal movement possible. Three subjects saw again a closed circle and a curved stroke or another circle under it, while the other three recognized a narrow s after one or two repetitions.
Concl.: Tinker : conf. with a, x, n, g, z, 8, %, 5. Admittedly the s is the worst of (lower case) letters, it requires features that



exclude each other. When the diagonal movement is aimed at, the curves should have small radii, but then the letter becomes too narrow. When it is made broader or the radii increased, the 8-shape appears, even when the openings are large (see fig. 52). In a desperate mood Slefrig asked for an abolition of the present form and a return to the old long f-like s. Ja val asks for an inverted z-shape, which would probably not improve things very much, because z ranks also low in legibility, according to several authors. Still, we might make the figure more angular, stressing the diagonal, by flattening the curves; no vertical serifs of considerable length should be added. Cf. concl. cap. S, series II.
t. -
53. Futura.
Straight stem with simple cross-bar, extending more to the right.
Novel form, copied by other sans-serif types.
Three subjects saw “some tendency to the right” at the base lme; four subjects believed that the cross-bar extended only at the right side. One subject was sure to see a Capital H. The recognition was in four cases inferential.. The cross-bar appears to be at the correct height to produce the cross effect (this was mentioned specially by one subject). Still we ask for a larger extending to the left and a bending to the right at the base line.
54. Memphis. . .
Straight shank with two short cross-strokes at middle- and base line, extending
equally to both sides. _ ...
This form led to remarkable mistakes. Two subjects saw at the first exposition a mere vertical bar; four professed to see a j or i (three even a dot upon it); at two expositions anorl was seen; movement to the right at the base line came hesitatingly. Two subjects stuck to their j or z.
The bottom serif extending to the left may be taken as the cause of this, together with the insufficiënt length of the cross-strokes to the right. The insufficiënt height above the upper cross-stroke accounts for the four retarded recognitions of the true situation there.
Stem sliced-off obliquely and joined at the left with the cross-stroke which is thin at the right side; round hook at the base line, ending in a thin upward stroke. Representative for most of the booktypes, save for the upward stroke
at the bottom. „ , . .. 
One subject required a repetition to be sure of the cross-stroke, the others
recognized it immediately. , . . .
Probably this result is owed to the unmistakable situation at the top, clearly the stem is ending, the crossbar setting in. Such an accentuation of a Crossing is unique among all the characters of the alphabet, but it seems to be worth while. The genereus bow at the bottom seems a considerable help.
Two cross-strokes (at middle- and baseline) to the stem, which is bent to the right (above the middle line) to reach the effect of the diagonal joint, cf. nr. 55.
Three subjects could not make a t of it: the resemblanee to f proves too streng. Three subjects described a hook to the right at the base line, the others did not see anything there.
This form has to be rejected.
Norma°form with curve at the base line, cross-bar extending equally to both sides.
Three subiects recognized it at once; another required four repetitions to



make sure, the other two did not recognize it at all, because these last three subjects saw no cross-bar. Comparison shows the cross-stroke of the Futura being not further to the left, though to the right it certainly is; moreover the amount of stem extending above the cross-stroke is larger there.
Concl. : The hook at the base line cannot be omitted or substituted by a cross-stroke ; it should be a full curve. The amount of the vertical stem, extending above the upper cross-stroke, should not be confined to a mere point, but rather kept approaching half way the top and middle line. If the style of type allows this, it should be joined with the left side of the cross-bar to avoid a disjunction, which makes it look like a dot close on the stem. The cross-bar has to be made fairly broad, but not exaggerated, to avoid the curious completion to ‘H’. Tinker : interch. with f, i, 1, 1. Javal based his remarks upon the modern style t, that shows a tail-curve ending in a little vertical, parallel to the stem, and which he is justified to discard. His demand for a cross-bar extending to the left only (in opposition to the f which should have it to the right), was based on the observation that the arch of much used and worn f-types often breaks off — which happens scarcely nowadays, because for large texts only new machine type is used. We therefore maintain our view. Extension to the left at the bottom should be avoided.
u.
The figures 58, 59 and 60 produced too few difficulties and differences to be of interest. Tinker concluded again (as with n) that it was mistaken for a and o.
, y.
61. Futura.
Narrow, no serifs* left diagonal joining the right at the base line. Two subjects saw a v at first exposition, two professed to see a turn to the left at the bottom. Two times also the main line was seen vertical from the junction downwards. The reason for this can be sought in the narrowness as well as in the low place of the junction.
62. Atlas.
With three cross-strokes, junction at the base line, broader form.
One subject hesitated between v and y, another saw again a vertical lower part. Otherwise direct recognitions.
63. Memphis.
Like the foregoing figure, junction above the base line, short vertical descender.
Five times a v was seen at first and many repetitions were necessary to correct this view. The shortness of the descender and its being vertical seem responsible for this.
Concl. : To avoid the principal difficulty, the interchange with v, the junction of the two diagonals should be not too low in order to allow an appreciable part of the right diagonal to project, that is, as the produced part, not vertically, because then (and also in case of a narrow form) the appearance of a concave v is promoted.



CHAPTER 4.
Variations of thickness of the constituent parts.
Series II. Upper and lower case display types.
Having got some knowledge of the dynamics of form, we can iroceed from the lower case letters with uniform thickness of he preceding series to the study of other changes of the basic 'orm. The dilferences in thickening of bookfaces are not so large hat the method of series I would have shown the effect of them idequately ; but in regard to display types these dilferences become mportant. We could have used the method there, but refrained 'rom doing so for practical reasons. The very fat faces would lave required a very short exposition, which our photographic shutter could not produce. Our objections against the usual tachistoscopic methods were formulated on page 24, Series I. Moreover, photographical reductions to the uniform height would igain have been necessary, requiring much time, painstaking labour and money, because about 550 different letters would have to be exposed.
Therefore the distance method was chosen. Because the heavy display-faces are suitable too for use on posters, etc., it was considered advisable to choose a corresponding distance, viz. 50 m., which was likely to bring out more subtle differences. The necessary movement of either subject or object could be easier effectuated by the subject, but then it was difficult to measure and control every little displacement. We therefore made a footstep of 75 cm. the unit of distance. In this way a displacement by a mechanical method was superfluous, while the subject had no other impediments to endure than the necessity to take steps of the said length, otherwise he was free. The tests were done in a large corridor, long about 55 m. and 1.00 m. wide, and passing through arches of 3.00 m. high, which gave it a tunnel-like appearance. Windows in the roof provided an even indirect illumination. One may object that the intensity of üght must have varied with the weather, which might be regarded as too much for an experiment. We considered this factor not of prime importance for the following reasons : The cards, bearing the drawn characters, were exposed in a box at the end of the corridor, standing against a dark background, and the illumination of the box was always much superior to that of the surroundings. The “tunnel” riveted the subjects eyes to this lighted area. In the course of the experiments we found the variations in intensity of illumination of the surroundings so



slight, that adaptation amply compensated them. Moreover, again our purpose was to learn rather about forxn dynamics than to get exact numbers of distance. The box, which stood on a table, was made of triplex, of a light, natural colour at the front side. Internally it was lined with rough white asbestos paper, that diffused evenly the light of two 75 Watt shopwindow tubes. The cards were inserted in a white cardboard frame, hanging against the back-wall of the box, the letter facing the point of intersection of the opening in the front-wall, 1.50 m. above the floor. The experimenter was sitting, invisible to the subject, at the side of the table. The floor of the corridor was marked over the entire length with chalk lines for every 75 cm. The subject started his walk towards the box from a distance of 52 m., calling the distances to the experimenter, together with every observation. He had been instructed to describe (and draw for explanation, if necessary) every observation and the changes in it; if necessary, he was summoned to do so again, or special questipns were asked to get a more fully substantiated account. In this way the subject could not suffice with the statement that, for instance, he saw an a ; sometimes he was also asked to go on, even if he had recognized the letter correctly. Usually the sessions comprised the exposition of about 28 characters ; in this v^ay 18 sessions of 1 to 1J hour were necessary. This varied for the three subjects. All were emmetropes, college undergraduates. During the changing of the cards the subjects walked back to the other end of the corridor, they had a chair there to rest for a few moments ; as a rule they drew their eyes away from the lighted box during the pauses.
Discussion of the method.
Three factors can have influenced the observation of the subjects. First a phenomenon described by Roels x), the inhibition following a false recognition. The forms, seen by the subjects, tended to persist, even when, after reduced distance, the present stimulus should have produced the adequate solution quite easily. Often the subject got suddenly the “Aha !-erlebnis” and gave the correct answer, or at least a relatively reasonable description in regard to the distance in question.
Secondly, when the subject was quite at a loss and “couldn’t make head or tail of it” he became a little fantastic, that is to say, he gave several Solutions in a “you never can tell”-mood.
1) F. Roels. On inhibition proceeding from a false recognition. Verslagen Kon. Acad. v. Wet., Wis- en Nat. Afd. Dl. 18, Nr. 9. Some further experiments on inhibition proc. from a false ree. Ibid. Dl. 19, Nr. 4.



These answers are not all valueless of course, but they give, reason to take them in consideration only in correlation with those of other subjects, further with marked tendencies and descriptions, made in cahn and long observation.
Thirdly, the subjects acquired after some experience an ability in recognizing certain letters and figures by secondary traits o’r logical conclusion. For instance, a triangle can mean an A or a 4. If after the answer “A” the subject was asked to go on, he was sometimes inclined to give the answer “4” befóre he actually saw it in the necessary details. In the same way the subject learned that a cracknel-form can mean 8, s, 6, 9, a and G; this possibility enlarged the tendency to guess. Such souijces of errors can only be detected by extensive and minute description.
The method was less apt for eliciting spontaneous obsertfations than that of series I, in consequence it was more likely to cause contortions of the real observations, owing to the intervention of controlling activities of the mind. In opposition to this, the* subjects were jnstructed to give accurate descriptions ofi every little change they believed to see.
Material. • ?
The stimulps# 'material was selected in order to represent:
1. gradual differences in contrast, ?
2. the most widely used styles,
3. some old and modem Solutions of the problem bf the “characteristic element”.
The following types were chosen : (see fig. 11)
1. Block, capitals and figures.
2. Futura, caps., l.c. and figures.
3. Reform Grotesk B, Heavy Narrow, caps., l.c. and fig.
4. Beton, lx.
5. Fanfare, caps. and fig.
6. Goudy Heavy, caps., l.c. and fig.
7. Poster Bodoni, caps., l.c. and fig.
8. Lo, caps.
9. Futura Black, caps., l.c. and fig.
10. Shadow Nobel, caps. and fig.
11. Bifur, caps. and fig.
Two other faces were tried, but discarded :
1. Ludlow Black, which closely resembles Goudy Heavy, introduced to study possible differences, which gave entirely negative results.



FIGURE 11
ABCDEFGHIJKLM NOPQRSTUVWXY Z123456789Oab cdefghijklmnopq rstuvwxyz
1 BLOCK
ABCDEFGHIJKLMNOPQR STUVWXYZ & £1234567890abcdefghij kim nopq rstuvwxyz
2 FUT U RA
ABCDEFGHIJKLMNOPQRSTUVWXYZ
abcdefghijklmnopqrstuvwxyzO
1234567890
3 REFORM GROTESK B



FIGURE 11
abcdefghijlclmnop
qrstuvwxyz
4 BETON
ABCDEFGHI3KLMN OPORSTUVWXyZS 1234569890 abcdeta hWdmnoiMirstuvwx
n
5 FANFARE
ABCDEFGHIJKLMNOPQRSTUVWXYZ
abcdefghijkimnopqrstuvwxyz 1234567890
6 GOUDY HEAVY
ABCDEFGHMKLMNOPQRST1JV $12345 WXYZ 67890
abcdefghijkimnopqrstuvwxyz
7 POSTER BODONI



FIGURE 11
ABCDEFüHIJKLl'lNOPQRSTUVW
XYI
8 LO
AKCIIGIT'HIJICLAINO l’OltSTIJVWX Y Z&fi f2345i;S90abcdef f|liijklmno|M|rstuvwx
Yi
9 FUTURA BLACK
ABCDËf.GH JKLMNOPQ R S 7 U VWXY2 12 3 4 567390
10 SHADOW NOBEL
ABC »[rfcUIJ(IUI
OPOtSTITWZTI
I2345ITII6
11 BIFUR



2. Kontrast, a clean cut version of Lo ; the considerable degree of contrast (about 1 : 10) produced abominable results.
The letters and digits were drawn in Chinese ink on white cardboard to a uniform height of 7 cm. (descenders and ascenders excepted), they were given in random order. The distances of the recognition are given only for the spot where the subject became certain in his recognition. Often he hesitated long before ; if important, such cases of a surmise are mentioned especially. If there are no remarks about a type-face in the following part, the letter was read from the end of the corridor.
A.
2. One subject did not see the white inner triangle in type 2. It measures 19 mm. high, 14 mm. base line, the black part above it is 27 mm. high; but this did not prevent recognition.
3. Type 3 gives no clear forms; the broad top is misleading, it causes interchange with other diagonal-containing letters, X, N, S or simple triangle. The low horizontal cross-stroke does not give a clear, white notch in the base line. Distances ranging from 24 to 45.
5. Two subjects saw a solid triangle, remarkably with a pointed apex. Obviously the general width of the letter prevents the conception of a rectangular form, like type 3, notwithstanding the top, which is as broad as type 3. Distance 52.
6. Distance for all subjects 52, white inner triangle invisible for one subject, no serifs seen; the triangle form is predominant.
7. Horizontal, serifs and left diagonal seen only at 36 by two subjects. One subject saw at first only the right diagonal.
8. Distances ranging from 27 to 39.5, interchanges with X, with I at first. Horizontal cross-stroke is only seen at rather short distance, probably placed too low.
9. Recognized by one subject at 52 as triangle and so as A; the upper part was seen solid by another subject; the third described as a hook (inversed V) with something between it, thus seeing limbs as the second subject did.
10. All recognitions at 52, but as an inversed V without the horizontal stroke, which was long 21 mm. thick 5 mm. and 21 mm. above the base line. The other strokes were equally thick, so we can explain this fact only by the predominating function of the diagonal limbs.
11. All recognitions at 52, the hatched part looking solid.
Concl. : An intrinsically legible character whenever the triangular form is kept clear, which is not the case when the letter is too narrow, the left diagonal too thin or the top too heavy and square. Tinker also recorded interch. with X and S.
B.
1. Distance for all subjects in the forties. The right side is too square, the indentation too small (interchanges with D, 9, R), but this is no serious defect in view of the distance.
3. Distance about 32, all sorts of fantastical misinterpretations. The right contour is utterly meaningless.
5. The triangular white notch at the right was 10 mm. wide and deep, which proved insufficiënt to prevent interchange with D. The internal disposition was a source of difficulties. One guess at 52, recognitions from 31.5 to 45.



6. Yielded exactly the same results as 7, viz. interchange with lower case n, though the horizontal strokes were 12 mm thick as against 3 mm. of type 7. Recognitions resp. at 43, 50 (after taking back the n), 37.5. The indentation is insufficiënt.
7. See 6, recognitions resp. at 38, 50, 37.5, the indentation is better, but the corresponding connecting stroke feeble.
8. Surprisingly good results in regard to the strong contrast and general narrowness; two recognitions at about 37.5 and one at 50, which we put down to the clear outline at the right side, the connecting strokes being feeble for that distance. The indentation is deeper, compared with the general weight, than that of type 5.
9. The right outline is good, but the inside unclear, giving tardy recognitions, hesitations between the 8 and the H-form.
10. All recognitions around 30, after confusions with S or 8-like shapes. Apparently the broken-up left outline is a hindrance for seeing a B, though aiding the necessary splitting.
11. Recognitions from 13.5 to 39, all subjects mistook it for 3, though the shading was good.
Concl. : The B should have a clear vertical at the left and a marked division at the right side, which can be obtained either by a strong central cross-stroke or by a generous indentation. A white vertical part should be avoided ; circular halves at the right are preferable. Cf. H 6, where the cross-stroke induced indentations too (8, two 3’s), and K 3.
These results agree with Tinker, who noted also interchanges with R, H, D, 8 ; Q and E were not seen by our ss.
c.
3. Two recognitions at 52, of whicli one saw a curl at the base, just as the third subject (at 36), who saw the usual winding movement of type 3, and moreover the vertical as half a circle.
8. One recognition at 52, the others at 33.5 and 37.5. Because the parts on the right are rather big in relation to the other half, this form is prone to interchanges with e. Type C 3 or C 4, that have the ends approaching in about the same way, do not suffer from this mistake; the round parts having a far bigger mass.
9. The drawings made by all subjects at 52 show the left half as perfectly curved; so the vertical follows the outer contour.
10. One subject saw the figure closed (O) till 34.5.
11. The same person was hindered by the overhanging bow of type 11 ; at first she saw an E with abnormally heavy upper horizontal, at 37 a C, the same horizontal remaining. Everything more than a pure half circle seems to be superfluous.
Concl. : Essential is only half a circle, which may be extended slightly over the half. If a narrow form is necessary, care has to be taken not to suggest an e by closing the figure too much, or by making the upper ending too heavy.
D.
1. One subject hesitated between D and O till 44.
2. Type 2 she saw as p; from earlier experiments she was known to mistake p for r, and therefore, perhaps, she took care to call this clear loop a p, afraid of making the same mistake again.



3. All recognitions succeeded only in the twenties. The roundings are :ntirely insufficiëntit is the worst character of this type, but, remaimng in ;his style, it cannot be improved, unless one would slice off the upper right
:orner diagonaUy. ,
5. The same subject, mentioned above, descnbed as b with extremely short iscênder till 37.5, which is accounted for by the very heavy upper left part. knother subject did not see any white in the figure, which is easy to understand. rhe curve at the right appears to be sufficiënt (cf. conclusion).
7. One subject described this figure as identical with 2, which means that
white amidst dominant black parts is seen as foUowing the outline in a regular iistance. We find this rule confirmed repeatedly. .
8. Though 8 follows about the same manner as 3, the contour at the right is more fortunate. One recognition at 52 and for the others tiU 30 and 42.75. The usual hesitations (B, H, N) of two vertical parts close to each other.
9 The exposed drawing of this type had a curve at the right somewhat fiatter than the original. The left part was 26 mm. broad, the right 30 which should have been more. This mistake accounts to a, _de8ree for the
recognition (by all ss.) as n (definite recognition at 30, 43.5, 34.5 resp.), though the fact is not entirely clear. We shall have to make a restnction to our rule, laid down above in regard to type 7, in this sense, that the contour has to be conspicuous, otherwise the adjacent white of another shape will counteract the effect of the outline all the same. In this case the white gives a vertical movement to the right half instead of following the curve As the n is; the only figure to make sense in the given problem: two verticals of which the one at
the right is rounded-off at the top”, a recognition as n is the result, the cu-cumstance of the lower right corner being rounded-off too, is overlooked. Cf. n. 9.
10. Like type 1, this form was seen as 0 till 45 by the same subject, remarkably the others recognized it at 52. It has the same proportions as 2,
3 nrC The^esults withthis8type are rather unsatisfactory. All subjects started with describing a reversed C or a 3; one guessed at a D (see'ng a very thin vertical) another saw the shading at 40.5, but not the specific D-form, the third described a full-grown D at 43. We may take the vertical as essential.
Concl. : This character shows clearly the tendenties of simplification towards a geometrically pure unit (i.c. an O). Vital parts are vertical and curve, not any horizontal strokes connecting those two. Therefore one should spend the available space rather to a curve with shorter radius than to horizontal parts. If aesthetical needs prohibit such design, the corners should be at least as round as possible. Tinker : interch. with B, Q, U.
E.
1 One subject saw the vertical stem much thicker than the cross-strokes.
2 The openings at the right seem to be a little too narrow: there were misreadings ffil about 40 with F and a rectangular G. This is probably due to ™ imprefsion of “a vertical with horizontals and white patches m.t’, which gives way for different interpretations. The opemngs measure each 9.5 mm. as
ag3in Oni subject hesitated1 a^long^ time (tiU 32.5, probably parüyov^ng to inhibition) between T and E. This can be explained by the jomed effect of
narrowness^and^ffie^large^oPext^ous disposition of white allows at the same
d7 fTheClsTmeTt7Tfor4eTyCS7ak7k7consider the central cross-stroke
assecondlry lt canbeshoner and less heavy than the others ; the white inlet clears the situation.



7. One misreading (K) till 43. The triangular ends are too thin (12 mm. maximally, as against the main stem 30 mm.), the connecting horizontals 3 mm. Nevertheless, the result is very satisfactory.
8. All recognitions at about 20. The horizontals are far too short, too thin, the endings paltry.
9. Ordinary horizontals were seen by all subjects at 52. A clear case of simplification and assimilation, also confirming our remarks to type 6.
10. Two recognitions at 42 and 45, otherwise a remarkable achievement, the thickness of the strokes being 5 mm.
11. One subject was uncertain till 42, otherwise no difficulties.
Concl. : The central cross-stroke has to be shortened if the type is not very narrow ; the total impression must remain that of a rectangular C with something in the centre.
F.
I. One subject hesitated between P and F till 45. The opening was 11 mm. between strokes of 14 mm.
3. This type was seen by one subject as r till 43.5. The lower cross-stroke was 10 mm. thick, which may be a reason for this. (It is interesting to bear in mind that type 10 has strokes of only 5 mm.). The large opening of 12 mm. explains the fact that no misreadings for P were made.
7. One subject recognized all right at 52, but saw the central part accentuated, the upper horizontal feeble.
Very interesting is the observation of another subject, who saw a V (also of type io, with heavy left diagonal) till 36, then a reversed upside down 4 (]/) till 33. This diagonal movement at the right is caused by the fact that the heavy terminals of the horizontal strokes stand out rather independently; the very thin strokes do not connect them with the vertical stem, which is unproportionally thicker. In this way the triangular endings mark a diagonal line which ends in the base. For the same reason the lower horizontal is seen more clearly than the upper one, because it has more relation with the shank.
8. The observations start with “vertical with something at the top right”, later on this is specialised in r, t or f (with visible curve), all recognitions in the thirties. The upper serif appears to be unable to check this tendency to a curve. The top right part is stressed too much.
9. Two ss. describe an F (at 52) with ordinary horizontal strokes, of which the lower is the shortest. The third subject makes a similar statement as did his colleague concerning type 7: he saw a V. Afterwards an r, then a vague second horizontal, recognition at 42.
10. Two recognitions at 52, the third at 45. They start with the vertical, and make an F immediately without the usual preliminary p or r.
II. In all cases the vertical is ultimately seen as thick as the horizontals. Two recognitions resp. at 40 and 47.25.
Concl. : The threatening interchange with P or r can be avoided in two ways : either by enlarging the distance between the horizontals or (when this is impossible) by making the lower one short. Care has to be taken, however, in the latter case, that no triangular ending or serif to this cross-stroke leads back to the upper stroke. In general no heavy terminals are allowed on the horizontals.
Tinker : Interchange with P, I, \/, T, E, Y, similar to our own findings.



G.
1. One subject described “a circle with something in the lower right corner” (possibility of Q), recognition at 35.75, the others at 52. This type is built by the form of its curve to stress the lower right corner (to distinguish it from C and 6). A hesitation of one subject between C and G and Q demonstrates the desirability of a larger opening and slightly longer horizontal.
2. To two subjects the situation in the “danger-corner” of this character (bottom right) was so unclear that their definitive recognitions followed at 40.5 resp. 36 (there were no interchanges with other characters). The introduction of a short vertical seems advisable.
3. All recognitions took place in the thirties, each after interchange with the figure 6. They described a rather bulky loop, clearly bending-in. Because the horizontal overhang is only 3.5 mm., as against 15 mm. of theadjacent limbs, and we know how easily such small parts get absorbed by large surrounding masses, it seems highly improbable that this should be the cause of the inward movement; rather — we should think — the relative height of this part is responsible.
5. Two hesitations between G and 6, of which one till 48. The long vertical at the right (32 mm.) is expressly stated by all ss., the horizontal by two. The oblique ending of the curve takes the place of a real opening. The same remarks as concerning type 3 are applicable in this case.
6. One subject described this figure as being of a moderate, but uniform thickness; also
7. type 7, save for a gradually diminishing thickness at the top, but with horizontal cross-stroke. Another subject professed to see a 6, consisting of a black dot at the base line, with a simple curve attached to it (recognition followed at 43.5). A remarkable misinterpretation in view of the heavyness of the constituent parts. Comparison shows why no similar statement was made about type 6: the heavy terminal of the overhanging bow prevents the conception of the idea ”6”. The vertical part is less square and therefore not so easily rounded-off to a dot; it is placed lower and farther to the right, and therefore the movement of the curve cannot take it in, as will happen with the figure 6.
8. The usual fantasies, due to a “jumping-over” between two parallel verticals excepted, only interchanges with 6 did occur, till recognitions followed at 33.5, 37.5, 43.5. Probably a half as long vertical similar to type 3 would be an improvement, because a 6 of those proportions is unusual.
9. The triangular terminal was seen by only one subject; he saw an ordinary
C with a uniformly thick curve at first. The others described a curve that, beginning thin at exactly the middle of the top, gradually increased and decreased to a kind of three-quarter moon-sickle, seen perforated by a white dot by one subject, as ending in a triangular dot by the other ; recognitions resp. at 28.5, 32 and 52. These observations agree in so far as the vertical side is completely hollowed-out; the thin, white dividing-line (4.5 mm.) answers its function without being recognized as such. Cf. with D 9, where also the white followed the black surrounding. The recognitions were based on guess-work, for the characteristical G movement (curve to the right and then upwards or back to the left) was absent. _
10. One interchange with S till 38.5, another with 6 till 44.25. See also type 2: the lower right corner is too much compressed.
11. This type leads to all sorts of fantasies, because the subjects could not make head or tail to it. They saw all right a more or less closed flat loop with a thin vertical at the upper left corner; all recognitions about 34.5.
Concl. : The chief danger for this character forms the interchange with the figure 6, which can be avoided by making a conspicuous vertical at the right or a horizontal, that should not, however, close the loop. Tinker : interch. with 6, C, O, Q.



H.
1. Opening measures 24.5 mm. One interchange with M (built-up of slanting, not vertical strokes) till 46.
2. Opening of 26 mm. One hesitation between N and M (left diagonal thickened) at 52, an interchange with M till 44.5 (only the left vertical thin; the right vertical and diagonal compressed).
3. One interchange with N and A with flat top. Recognitions resp. at 31.5, 42.5, 34.25.
5. The heavy square block with its narrow white middle-line lead to conceptions of a square O or n. One subject saw an M and an H with broad serifs at the top (accounted for by their greater width at the top). Recognitions at 35 and two at 45.
6. One subject saw a W (much the same as the M of type 1) till 44. Another got the impression of an 8, and later of two 3’s, turned to each other, till 37.5, for which the small openings (10 mm.) may be the cause.
7. The white vertical was broad 10 mm. (1 mm. between the serifs), the horizontal cross-stroke 2 mm., the black verticals 30 mm. Recognitions followed at 22.5, 28.5, 34.5, after interchanges with U, n, II, 8.
8. Again the usual guesses after the initial conception of two parallel verticals. These guesses make the scored distances less reliable. Recognitions about 30 and 40.
9. The same is true in regard to this type. Recognitions about 30.
10. Two subjects saw anM till 40, with a marked V either down to the middle, or to the base line.
11. The hatched lower part of this type causes additional interchanges with U, O. Recognitions (or more precisely: accurate descriptions) at 25 and 34.5.
Concl. : It is a curious problem why the H, one of the simpliest characters in the alphabet, should make such a poor score. With some other characters (M, N, O, U) it has the parallel verticals in common. If its horizontal cross-stroke is of equal thickness as the verticals, interchanges with M and N will occur, which have their centre of gravity in the middle too, facilitating therewith a short cut between vertical and horizontal; if, on the other hand, the verticals are accentuated, to the neglect of the cross-stroke, similarity with O and U arises. Tinker : interchanges with B, U, R, N, P.
I.
The I does not give much occasion to different interpretations. In regard to the visibility it may be noted, however, that type 10 (5 mm. thick) was seen by all subjects at 52.
6. Type 6 was seen by one subject at 52. The others, deceived by the heavy serifs, saw a T. Also a narrow X, an N, till about 24; see about this L 5. •
J.
3. After hesitations between I, 1, V, recognitions followed at 34.5, 48, 42.75. We may take the curve as too short and square.
5. This character was 78 mm. high (the original is a descender). The horizontal at the top and the cross-stroke proved a difliculty; interchanges with the figures 2, and later on with 3, were registered. Recognitions at 30 and 24, one at 45.
6. Was 93 mm. high. No difficulties.
Legibiüty, atmosphere-value and forms of printing types. 4



7. One subject was hindered by the left top-serü (whicn was b.d mm. iong and 3 mm. thick), so much that he saw a 3, or a very thick J with prolonged serif, recognition at 45, but certainty at 31.5.
8. The curl appears to be a little too thin; two subjects saw it at 33.5 resp. 39 75 and described it as tiny. Still it was max. 9 mm. thick, the white notch separating it from the shank 7 mm.; so this was rather unexpected. Probably
the curve is also too flat. . . .. ,
9. No curl was seen by two subjects, but a horizontal bending of the stem.
All recognitions at 52.
10. One subject recognized at 34.5.
Concl. : No protruding below the base-line is necessary; heavy serifs or cross-strokes at the top are superfluous, if not a hindrance. The curve at the base should begin as high as possible, it may end simply in a curl. Tinker : interchanges with J, 3, 7.
K.
1. One subject mistook E for K till 44.5, probably due to the very heavy
join in the middle of the shank. ... , , , .. . ,
3 Because the diagonals keep almost m line and only touch the vertical supêrficiaUy, the shank suffers a similar reformation: an X was seen, composed of lightly touching arcs of a circle, see concl. B. It is ïmpossible to squeeze this character into the Standard width of the two-vertical-characters (B, C, H etc.) of this type. Recognitions at 34.5 and about 25.
5 One subject saw an E and N (with unclear diagonal) at first probably as a result of the heavy join and top right corner. Recognitions at 25.5, 50, 37.5 If the upper diagonal were not bent back to the left, the ambiguity, which caused the E and N, would be precluded. . . .
7. The upper diagonal was seen unclear or not at all. One recognition at
8 The high-placed join was responsible for the h, seen by all subjects. The dot-shaped upper diagonal was not expected and therefore overlooked. Interchanees with N did also happen. The recognitions ranged from 20 till 30.
9 8 Recognitions about 40. One subject was deceived by the almost solid topand base-line, and described an E with unclear middle horizontal; anothersaw an ordinary K with the diagonals as heavy as theverticalwhilethethirdgavea description of a vertical, touching a C at the right, which meansthattherght half of the character is modelled by the outline into a simple curved strokeofumform thickness. Another subject saw a reversed X tdl 48 : the upper half of the vertical was not seen, its lower half brought in line with the upper diagonal. Cf.
’bne Interchange with S till 37. For intricate figures with diagonal parts S is often given as solution. The reason should in this case be sought for m the feebleness of the whole figure, rather than in misshaping. 7 Th.
11. All subjects described a flat curved C, two of them later on a 7. The hatching was not quite as heavy as that of the other charactersofth.s type^But still, the angularity of the hook is too little conspicuous for this purpose. Recog nitions from 18 till 25.5.
Concl. : Only the upper diagonal may be a little thinner than the other strokes, though the disposition has to be such as to allow a marked triangular outline at the right. Because of the intricacy of the figure, the distribution of weight should be as equal as possible, avoiding a heavy join or terminals. Linker • interchanges with R, Z and with 2 and % (see type 10).



L.
1. One subject did not see the horizontal at first, then saw a tendency to the right, and finally discovered a second one at the top as well, making a C. The first recognition took place at 43.5, of the other ss. at 50. The horizontal could be longer.
2. The same subject as above saw I till 40, which seems no shor tcoming of the design.
3. The horizontal visible for one subject at 39 ; interchange with C (by another) till 41.25.
5. One subject saw an X at first (a similar case of splitting up as I 6), then a reversed J, afterwards a wedge-shaped vertical (apex downwards) with round dot next to it: recognition at about 40, faultless description at 24. Another subject saw an F till 35.5. Similar phenomena occurred with type 8. The L, it appears, is an incomplete, “unnatural” figure. The subjects try to counterbalance it (C, t, E, topserifs), to round-off the corner (reversed J, t). They make the horizontal independent, or see a mere tendency to the right. All this happens especially with short horizontals. Cf. Z 3.
7. One subject saw a 1 till 39, but the others mentioned the terminal of the horizontal expressly at 52.
8. Recognitions at 33.5 and two at 27, an E, t, 1, 1 and a wedge with dot were seen (see type 5).
9. Two subjects saw anordinary horizontal. All recognitions at 52.
10. One recognition at 40.5.
11. One description as two triangles, with the tops joined in the corner.
Concl.: A broad and solid horizontal is essential, otherwise misinterpretations will happen. Heavy serifs at the top, added to fill the white gap, must be condemned. Tinker : interchanges with I, square, \J, Z, U.
M.
1. One subject saw the V extraordinary heavy, another described a thickening similar to type 7. Remarkably no interchange with H occurred, though, as we told before, the contrary was often the case, and moreover a horizontal black stroke could be drawn across the figure (see ill.). The V-shape appears to be clear, perhaps because it reaches only half-way.
2. One interchange with N till 34.5. The answer N was often given instead of M, also when the right diagonal was equally thick. This can be due to both guessing between M and N (“two verticals with something between them, which is no simple horizontal”) and to a greater susceptibility for diagonals from top left.
3. Recognitions at 37.5 and two at 42, after interchange with N.
5. Though one subject recognized at 52 (probably judging after the relative width), the V-shape appears to be insufficiënt, for another s. after supposing M at 52, interchanged with N, H, even described a coupled pq till 20. The third s. described V, H, N till 27, the indentations apparently being invisible.
7. The right diagonal was invisible for all, which accounts for one interchange with N till 33.
8. The left vertical, though equally thick as type 11 (5 mm.), was invisible for all ss., causing the description “vertical with flag at the left”. One recognition at 37.
9. Two descriptions as type 2, of which one at 34.5 after seeing an A. One subject saw a solid trapezium at 52, then anM with thick left diagonal, till 44.25. This result is not unsatisfactory; probably the relative width and the diagonal movement do the work ; that the little white notch in the top should be the stimulus for the right diagonal, seen by all subjects, is highly improbable. A remarkable illusion !
10. The V became visible for one s. at 36.5, otherwise direct recognitions.



11. One s. saw a thick V hanging at thin verticals; another saw, apart from the thick strokes, the right diagonal only ; both recognized at 52. A third subject saw only the thick strokes till 39. This time we may call type 11 successful.
Concl. : The M is an intrinsically legible character, if only sufficiënt space is allowed to unfold its V ; for this purpose it should reach only halfway down too, unless only the left diagonal ind right vertical are accentuated, or ample space is available. riNKER : interch. with W, N.
N.
3. Two recognitions at 39, after the usual fantasies of this type; onesubject was not certain about the diagonal till 25. It is impossible to press an N in this space.
5. One recognition at 52 ; the descriptions of the other subjëcts were not reliable, swing to external reasons, but they showed the tendency to interchanges with K, H, 2, etc. I should regard it also as too narrow.
7. Only the diagonal was clearly visible, which caused uncertainty. Recognitions at 37.5 and two at 42. One subject saw the figure as belonging to cursive type, turned clockwise; for which I know no explanation, perhaps it was a try to “put the diagonal right”.
8. The verticals are too thin, giving recognitions at about 35, 52, 47, of which the first after interchange with an upside down V and with S.
9. One recognition at 48 after interchange with K, another at 38 after descriptiori of a solid circle and an n, (which I cannot explain). This form seems recognizable, but not clear.
10. One subject (after seeing an H) stated expressly at 36, that the diagonal
ended at about one third of the height of the right vertical. No improvements are suggested. , . _.
11. Two ss. supposed an N at 52, but became certain later on (at 40). One subject started with the description of a mere vertical (ree. at 30) ; two ss. described the left vertical as slanting, which points to the same phenomenon as the cursive of type 7, a desire to counterbalance. The solid block of hatching was probably responsible for the 8 or S, seen by one j.
Concl. : Like M, the N is naturally a broad character. In order to avoid interchange with S, 8, K (that contain diagonals of the same direction) the verticals should not be neglected too much. Tikker : interch. with S, X, %, M, triangle.
o.
3. Not so much interchange as uncertainty caused the scores 25, 45, 37.5. We suggest an oval shape also of the white centre, though not quite in style.
5. The unequal distribution of weight originated uncertainties, though the initial supposition was always an O, so that the final decision was made later on
in two cases. . , n .... .,
7. The vertical white induced one subject to hesitdte between D and O tul 4i.
8. As type 3, the thin opening lowered the score to 37, 41.25, 45.
9. The thin white vertical proved to be no hindrance whatsoever in this case. Probably it was too thin (3.5 mm.) thus making the figure appear like a solid circle; all ss. stated a small hole in the centre: the only rest of the white, or a fiction. This favourable score would not be made by a narrower edition, I suppose.
10 One subject described this form as thin at the top and flat at the left, after having mistaken it for the figure 2 till 38.5, probably owing to the poor visibility. The diagonal movement must have been caused by the lower sickle.



11. One subject suspicious after previous cases of overlooking of a ta.il, hesitated between O and Q till 46.5. The hatching was not noticed by any subject and therefore may be omitted as well.
Concl. : The opening can only be allowed to have a linear or drawn-out oval shape, if the total form is clearly oval or round. Otherwise no mistakes can happen. Tinker recorded interchange with O, a guessing of a nivellated square.
p.
3. Two interchanges with r (F) till 30 and 37.5.
5. One interchange „ „ „ 44.25.
6- „ ... „ „ „ 45.
7. Two interchanges „ „ „ 36 and 48.
8. One s. hesitated between b, p and r till 44.25, probably the curve comes down too low.
9. All subjects saw an r, till resp. 37.5, 41.5, 48.
10. The figure became visible to one subject at 40.5, no mistakes.
11. One interchange with r till 44.25.
Concl. : Only the types 1, 2 and 10 did not suffer interchange with r, owing, no doubt, to their large quantity of white (equal to the thickness of the strokes or more). Tinker : interch. with F.
Q.
Unless specially stated, all mistakes concerned O; each distance mentioned concerns one subject only.
1. 48 : one hesitation between Q and 8 till 33.
2. 34.5 and 39.75.
3. 28, 37.5, 38.25.
5. 16, 21, 18.
6. 33.
7. 37, 45, 37.5.
8. 26, 40.5, 33.75.
9. 23.5, 30, 27.
10. One subject saw at 33 a figure 2 (a not quite closed circle with horizontal at the right), ree. at 20, another supposed a Q at first, then changed to a I, ree. at 18.
11. 39 ; one subject saw the thickened base all right at 43.5, but refused tocall it a Q.
Concl. : A large flourish like type 6, or a big cross-stroke, extending on both sides of the circle, are necessary ; a mere “something at the base” is insufficiënt.
R.
1. One subject mistook it for D till 42, another confirmed this possibility.
2. Again, an O or D was seen till 40, while another saw the diagonal clearly at 44.25 only. These diflficulties seem to be inherent to this character ; I can suggest no improvements that would not open the possibility of other movements in the design.
3. Interchanges with K, N, recognitions at 34, 48, 39. The diagonal appears to be very clear (in relation to the general legibility of this type and the results of this stroke with other types), the width of the character is larger than usual to get a broader wedge in the base-line. The loop is still rather ambiguous.



5 One confusion with D till 36.75; another one hesitated till 48, while a P and later on an N were seen till 30 by the third s. The character should be made a little broader, allowing for the join to be placed higher up; this would yield a bulkier bowl and a clearer diagonal.
6. One subject mistook it for n till 39, another s. was not sure about the loop (K) till 43.5. To make the indentation deeper (in view of the n) would hardly be in harmony with the style.
7 Two interchanges with n (till 31 and 42.75), one subject saw a straight diagonal starting from the middle of the shank (at 50). This type is very poor ; everything essential of an R (loop and diagonal) is nivellated. The join is in no way prepared. The point of application of the diagonal has to move over to the lett, the contour at the right can then be made more conspicuous.
8 The curly tail is unclear. Hesitations or interchanges with P, D, B were
the result. Recognitions at 32.5, 40.5, 39. As type 5, the join has to be placed higher up, the whole character must be made broader. . . . _
9. One subject saw an ordinary B like type 2 at 52. Two ss. mistook it for E till 39 and 32, which means that the very clear right outlme has not succeeded alone in converging the idea “R”. The figure is split-up horizontally from the right, which would make a kind of C, if the figure were not too complicated ; K was seen once, also an e. So here again was the clear outlme spoiled by the too large and shapeless black mass. Therefore more white inside will be of help, (but not in the form of a broader white vertical!).
10. One interchange with 2, presumably because of lack of visibility, till 31.
11. Recognitions (or rather guesses) at about 30. The thick lines were seen all right, but in different interpretations (e. g. reversed S, ?, P, 2 upsidedown, e with diagonal tail etc.).
Concl. : One cannot tamper with this figure without risking a serious loss of legibility. It is advisable to keep faithfully to the traditional proportions of lines and weight, like type 2, 10. The join should be not too much from the middle, the diagonal conspicuous. Tinker’s records show interchanges with B, H, 8, P and therefore do not confirm the D and O, observed by our ss., due to the heaviness of our types.
S.
1. Two interchanges with 5 ; ree. at about 45. The S needs a vigorous winding movement, which can only be brought about by a broad form ; therefore we should
like this type to be less narrow. ...
2. One interchange with 3 till 43. This 3 also occurred several t.mes with other types. Because the types of this experimental series are heavyfandtherefore extended), less interchanges with 8 were noted as in series I, where this was not
the case. . , ...
3. Recognitions at 37.5, two at 48. This is a very good result.
5. 3 and 5 were favourites. Ree. at 22.5, 31.5, 34.75.
6. One hesitation (5 with vertical missing) till 41.
7. All recognitions at 52.
8 „ „ about 35. . . . . .n
9. Two interchanges with N, R, till 29.5, 34.5, one (without terminals) ree. at 50. The curves to the diagonal appear to be insufficiënt.
10. One subject was sure at once, though he called the character unclear another recognized at 38, while the third s. became certain only at 18. we eau attention to the fact that the diagonal attains the thickness of 5 mm. only for about
10ll!m BeSdes 5 or a mere diagonal curve, a 4 was seen by all subjects.withthe horizontal put aslant. Though an S was supposed in two cases at 52, the detinitive remenition ca me off at about 30.



Concl. : It is difficult to make conclusions out of these data. Good scores were attained in different, partly contradictory ways.
The diagonal swing is the essential part; but this alone is not sufficiënt (type 11), extension over the middle-line is necessary. That curving is required too, is proved by type 5 and 9. Type 7 was the most successful: it gives a broad, vigorous swing ; the upper and lower curves and terminals are faintly indicated, but quite sufficiently. If this is done with heavier strokes, the 3, 5, 8-possibilities arise (type 2, 6, 10), at the same time the curves must begin earlier. The narrow editions (1, 3, 6) do not seem to suffer much from these obstacles ; they are so thin as to leave enough white space, but the necessary early-starting turn at the top leads to interchange with 5.
The light edition of type 2 (Futura Heavy) is based on an entirely other principle (see p. 35) and constructed with a pair of compasses, like type 11. In this way the upper and lower half become independent entities, and no smoothly running diagonal is obtained (cf. type 11). Tinker : interch. with 8, %, N.
T.
I. One interchange with 7 (with cross-stroke) till 41.25.
3. One hesitation till 39.
5. One subject saw stem upwards growing out into a triangle with the apex down.
6. The terminals ot' the cross-stroke were not mentioned.
7. The cross-stroke became visible for two ss. at about 41.
8. The cross-stroke was just visible, but not clear, the ss. described it as a tendency, resp. at 37.5, 52, 32.75.
9. The cross-stroke was seen as a bar (heavier than the stem) by two subjects, by one of them at 46.5.
10. One hesitation till 48 ; an interchange with 7 till 33 ; an interchange with t or f till 27. The latter is partly due to the short cross-stroke, partly to the crack in it.
II. One subject saw a Y, V (with heavy serifs), also the stem running into a triangle, till 42.
Concl. : Though a broad T is not beautiful in context (leaving conspicuous white area’s at its sides), we may regard it as a necessity. The tendency to blot out the sharp Crossing at the top (7, Y, V etc.) should be noted ; we can call it also a tendency to fold together all, that extends from the simple verticql. See also p. 21. Tinker : interch. with I, Y, F, 7.
u.
3. One interchange with 9 (jumping-over effect) till 40.5. One subject saw the curve at the base vaguely (at 52), another described it as pointed like a scutcheon till 28.5.
5. Two recognitions at 52, one at 32.5 after N- and H-interchanges. The vertical at the right seems rather a nuisance with this narrow type.



7. The right vertical became visible to one subject at 40, to another at 45.75.
8. Added to the narrowness (comparable with type 3) is the thinness of the right vertical. A bad score is the result: 33, 39, 23.25.
9. One recognition at 52 (seen as type 2), two at 31 and 37.5 resp. with pointed base. There the white vertical is clearly seen, even over-estimated, because the outline suits well to its direction ; one will remember that in the opposite case the white lines were overlooked or differently interpreted.
10. One hesitation (O, V) till 32.
11. One hesitation till 42.
Concl. : This character can only suffer from the “two close verticals-effect”, which also makes the join at the base unclear. If contrast is introduced, the curve has to be prepared high-up, to make the observer at least suspicious (if the right vertical is scarcely visible) that the left is joined to something at the right. Serifs have to be rather light or short lest a closure (O) should be seen. Tinker : interch. with I, L, C, G, T.
v.
3. One interchange with Y till 40.5 in one case. Contrast (like type 7) was seen by another subject at 39. A simple triangle, then a V or 7 was seen by the third (ree. at 44.25) ; the narrowness seems for a large part responsible.
5. One hesitation between V and Y (at 52).
7. For two subjects the right diagonal became visible at 37, resp. 48.75. No
interchange with N. . „
8. The narrowness of this type lead to mistakes with vanous possibilities ot the type “vertical with something at the top” (P, 1, T, Y etc.), so that recognitions took place at 41.5, 33, 44.25. The minimal quantity of white adds to the con-
ception of a vertical. ,
9. One hesitation between Y (“with very short vertical) and V till 45 ; another between V and 7 at52. An interchange with 2 (“vertical with something at the top”) till 39. This type seems rather narrow in relation to its weight; moreover the white line turned out'rather thin (3.5 mm.), which may account for the difficulties. The outline, though, is not so eloquent as it might be.
10. A hesitation between V and Y, at 52 and another one till 34. Of course, the left diagonal is not directed to the base, it ends 10 mm. above it.
11. One recognition at 38. One subject saw the hatching, the others a thin diagonal. The pointed base and the large angle (width) seem to do well.
Concl. : Everything is allowed with this character, if only the triangular form remains. One must take care that narrow (steep) forms keep enough white, to avoid easy interchange with Y, etc. Tinker : interch. with \4 Y.
w.
1. Two subjects saw the lower part as a solid black. Though all ss. recognized at 52, it seems advisable to introducé more white at the base.
3. One hesitation till 45. Again the base was seen solid by twois. No ïmprove-
ments suggested. . , , ^ .
5. The small amount of white, both at the top and the base, accounts for recognitions at 28.5 and 48. .
7. One subject saw the thick diagonals only, and supposed a W at 52, but left the possibility open of a kind of slanted N till 40.75.
8. The indentation of the base was quite sufficiënt, but it made the required



trapeziform shape impossible. In consequence an N was seen by all subjects, though a W was presumed in two cases (at 52, 48). The closed top added to the difficulties. Score: 31.5, 43.5, 24. A narrower base is advisable, if the total width is to be kept.
9. Seen as two crossed V’s in one case ; as a V till 40.5 by one subject, who described the descending diagonals as clearly accentuated.
10. One hesitation till 40.5 (V).
11. ,, ,, ,, 42 „ . Another s. was uncertain about the right
diagonal, at 52.
Concl. : Before all, the trapeziform shape has to be aimed at. Narrow editions must introducé a generous amount of white at the border, in order to avoid interchange with V. The interior disposition is of secondary importance then ; it may be fully written out as a zig-zag, or crossed, or the left ascending diagonal suppressed.
x.
I. One interchange with 5 till 43, the right diagonal described as thin (like type 7).
3. All subjects described at first a vertical with something attached (t, a nail), recognitions about 40. One subject saw then two juxtaposed hooks, perhaps owing to the long area of intersection, which blurs the diagonal movement.
5. One subject saw an I till 39.5, then two juxtaposed halves of a circle. The small amounts of white at top and base seem responsible.
6. One subject saw a Z till 43.
7. Two subjects saw an N (with unclear verticals) till 40 and 38.25.
8. Again interchanges with I, N ; two recognitions at 38.5, resp. 42.5.
9. One interchange with Z till 43.
10. One interchange with A till 41.75 ; the other recognitions at 42 and 45. All ss. described the Crossing as situated high at the top. This is the result of the top-ends which are sliced off obliquely, being invisible, in consequence the whole part above the Crossing becomes shorter.
II. One subject mistook X for I, K till 35.5. Another saw at first a narrow C (cf. K), then a K, or a hook with tendencies to extend itself to the left. Recognition at 22.5.
Though one subject recognized or rather got the impression “X”, at 52, this form does not seem to be so evident.
Concl. : Owing to the “tendency to fold together”, all narrow editions suffer from interchange with I, K etc. The right diagonal cannot be wholly ignored in the design, to avoid the N-impression. Tinker : interch. with N, %, Y.
Y.
3. After interchanges with r, 7, T, “nail” two recognitions at 34, 48.
5. One interchange with X till 42, another with 7 (horizontal bent inwards) till 41.25. The curl was seen and understood by all ss.
6. One hesitation between V and Y.
7. > Two ss. saw the thick line only (as a kind of reversed C) till 39, 41.25.
8. The right diagonal was seen by all ss. apparently, but the high-placed bifurcation gave the impression of a slender V. Score: 32.5, 39.5, 34.5 though each supposed a Y 5 m. earlier. So we see that a narrow V is mistaken for Y, the little space between the limbs causing consolidation, and that, on the contrary, a Y with high-placed bifurcation gives the impression “V”, because the splitting is clear, and supposed to reach far down.



9. Hesitations (T, V) till 40.5, 48. The bilurcation is seen low-piacea.
10 Score : 43.5, 48, 34.5, after the same mistakes.
11. One subject mistook it for T (P with loop extended to the left at the top) till 31.5 ; the others got sufficiënt certainty about the right diagonal at 45, 33.
Concl. : The bifurcation should not be too far from the middle ; impending dangers are either V or the nail-shape (T) ; if the right diagonal is neglected, the rest is seen as a straight left diagonal (N, X). Tinker : interch. with \/,Y,X.
z.
3. AU three ss. saw a 2 at first then a Z with waved horizontals (29.5, 34.5, 40.5). This undulation did not occur with type 8. Probably, the short horizontals are too feeble to suggest the straight horizontal, consequently the natural, undulated writing movement gets a chance. On the contrary, the strokes of type 8 are so scarcely visible, that they are completed by the subject.
5. Interchange with 1, X; till 43, 40.5. One subject saw the lower horizontal vaguely at 52 (as a J with something attached). The horizontals seem too heavy, the diagonal too steep.
6. One interchange with X till 47.5. , . ... »
7. One subject saw a diagonal with vertical at the right (and cothing more) till 33.5, another was not sure about the upper horizontal. •
8. The lower horizontal was invisible for two subjects till 34.5, 27.75. ine
diagonal was seen vertical at first, as usual. ... f
9 One subject hesitated tiU 43.5, he saw a square block with white of an unknown form in it. Then he described an ordinary Z of equal thickness, with long vertical serifs to the horizontals. No interchange with X was made, probably owing to the long verticals and the conspicuous diagonality.
10. One interchange with I till 41.5. , 7 ix. q
11. One subject saw only the diagonal at first, then an X, then a Z like type 9, though with more white in it (smaller triangles), recogmtion at 30 ; another at 46.5.
Concl. : The diagonal is the principal feature. Care should be taken with narrow types, because there is a marked tendency (in general) to see steep diagonals as vertical. Tinker : interch. with 7, X.
Lower-case.
For the study of the lower-case characters the types 2, 3, 4, 6, 7 and 9 were chosen. One has to keep in mind that the l.c. characters are relatively more enlarged than the capitals; in consequence they are heavier, e.g. the Capital Z of type 2 had the following dimensions : width 58 mm., horizontals thick 17 mm., diagonal 18.5, while the corresponding figures for l.c. are 75, 19.5, 21.5. Comparison with series I shows that several characters make a better score when they are supplied with thickness ; e.g. b 6 did not present the difficulties of nr. 8 in series 1: The internal shape of the loop is quite normal in the former case ; in this way the effect of the pointed end is counteracted. On the other hand the conclusions of the two series can be different, e.g. in case of the e ; the question of weight is therefore of prime importance.



a.
2. One interchange with o, at 42, seen with the right side flattened (no separate stroke), ree. at 36.5. One subject described the loop as an oval hanging diagonally from the top of the vertical i.e. the vertical was not assimilated. No interchanges with d or q were noted.
3. All subjects started with seeing a 3 or a reversed E, then came to a reversed 6 ; recognitions at 36, 45, 37.8. Because there is no tendency to the right at the base, the lower end of the vertical is lead over to the left in the loop, which makes the figure symmetrical and simple. The upper-end of the loop becomes a crossstroke (rev. E). A serif to the right should be added.
4. One interchange with 6 till 42 ; otherwise the arch was not absorbed by the loop. This form is comparable with series I, nr. 5, of which the centre of gravity is lying higher, making assimilation possible between arch and loop.
6, 7. Though the distance between the end of the arch and the loop of these types was small enough to call forth a “jumping-over”, no such thing happened. The typical a-movement appears to be present, owing to the large amount of white internally.
9. No interchanges with 3 did occur, one subject even saw a little curl to the right at the base. The large circle-half, the total width and the clear white divisionline could not make a plausible 3. The dot at the top was seen as a thick arch.
Concl. : The movement “arch-vertical-upturn with attached loop” must be kept. Of narrow faces the loop must not become too subordinate; if necessary, the arch has to be shortened as to raise the importance of the loop.
b.
3. AH types were recognized at 52, save 3 which was seen as a simple vertical (at the top bent to the left) by one subject till 42.
c.
3. Described as oval.
4. Two subjects saw a dot or curl at the top.
7. All subjects saw an e at first; recognitions at 40, 42, 45.75. The dot is no simple upstroke, such as 4 and 6 show ; it a is separate ornament, designed to counterbalance the left. By its position and weight, it is destined to cause inter changes with e. Cf. C 8 ; here also an inward movement was seen, after judging from the relative weight of the vertical; the same is happening here with the dot.
9. The triangle was not seen as such, but as a dot or club.
d.
3. One interchange with 4 till 41.5. The loop was connected with the top of the vertical, and the latter extended downwards.
e.
3. One subject recognized at 32, at first seeing a vertical with something at the top, then also a tendency to the right at the base, finally describing, as the others, an oval e.
4. One subject hesitated about o, c, 6 till 42, probably owing to inhibition following false recognition.
6. All ss. saw a c till resp. 39, 48 (two). No accént is laid upon the horizontal, but, opposed to c 7, of the angle of the loop a dot or an upturn is made as with the c. The large opening does not suggest an e; the white spot (in the loop) strives for reunion with this opening. The thick cross-bar (10.5 mm.) even cannot prevent this. The angle of shading should be more vertical, the opening smaller.



7. One subject actually saw a c till 44.5, but the others too mentioned the missing of any horizontal; one s. described the upper half as solid. Essential for the e is not the quarter circle north-east, but the horizontal.
9. Two JJ. did not see the white dividing line, and accordingly described as c or the upper part as solid; recognitions at 42, 37.5.
Concl. : If one chooses to design an e as type 7, a fair amount of white has to be introduced, the horizontal must be kept low (about the middle) ; the remaining reinforcement of the loop will not be seen as the dot of a c then. In other cases the crossbar has to be stressed. The difference with the results of series I is entirely due to the absence of a heavy corner in the eye in forms of the latter.
f.
3. The ss. could not describe exactly the situation at the top, and hesitated about v, t, f. Definitive ree. at 31 and two at 34.5. The space, allowed for the curve, is far too small; it is too angular in any case. A lower placed cross-stroke
would be a help. .... , ,
4. One subject described this type as a vertical with dot, the others as a curved
vertical. ...
9. Though following the same principle as 4, this type is narrower. One subject did see no deviation to the right till 41.25, another became sure at 48. The dot was not seen as detached by anyone, but as a knob.
Concl. : The legibility of this character varies directly with the extension of the arch or dot to the right.
g-
3 No subject could make sense of this character at once. The true proportions seem to be wholly disturbed ; the exact place of the opening could not be located. Definitive recognition followed at 36, 39, 33. If a character, that is naturally round and broad, like the g, is made angular and narrow, it follows that troubles must be the result; if not mistakes, then at least hesitations.
6. One subject, though recognizing at 52, called it unclear and overloaded; another could not make head or tail of it, mistook it for Z, E till 42. Every detail must have been visible, physiologically speaking (the figure was 10 cm. high, the thinnest strokes 12 mm. thick); still the total impression was not clear sit once
7. This figure, 103 mm. high, proved also too colossal. Though the flag had a diameter of 18 mm., and the lowest stroke was 35 mm. thick minimally, two subjects saw only the upper circle with a horizontal under it; one of them recognized at 41. The complete subordination of the details to the whole is again
clearly demonstrated by this fact. -nu -ah.
9. The rendering of the original was not above ïeproach, especially the wiatn of the circle-half was too small; notwithstanding, two ss. saw a complete circle, attached at the vertical; the third hesitated till 40.5.
Concl. : For heavy faces the g form with vertical at the right (as type 2) is preferable. The old-style form, with its four horizontals above each other, makes a complicated figure, which is extremely difficult to model, to organize.



h.
3. Two subjects mistook it for b, till 35 and 45.75.
4. One subject hesitated at 50, another actually mistook it lor b till 45.
9. The protocol was not entirely reliable, but one subject was noted to mistake h for b till 41.25.
Concl. : The interchange with b is entirely a matter of the amount of white between the limbs at the base-line : though type 4 has a heavy serif to the right at the right vertical, it has less white at the base-line than type 7, which did not suffer interchange with b. Type 27 of series I has a sufficiënt amount of white.
i.
3. One subject saw a J without dot at first, then the dot was added, whereas
the curve shrinked to a mere thickening; ree. at 35.5. Another saw the dot as
separate at 43.5. Cf. series I: the square dot is sooner assimilated than a round one.
4. Again the j and thickened i by the same j till 40. Dot plainly visible.
6. Tbe dot became visible later on to two ss.: 36, 42.
7. One subject saw the shank split as a K till 43.8. We came across these
illusory movements in broad verticals several times.
9. The dot was seen as separate by two subjects at 34.5.
Concl. : see Series I.
j-
2. Only one subject supposed a j, judging from the relative length. If used separate, a straight and square terminal, without anything more, is absolutely insufficiënt.
3. The dot was assimilated, the turn at the base also became visible later on; score (faultless description) 42, 40.5, 34.5.
6. Dot visible to one s. at 44.
7. Turn at the base plainly visible, no dot mentioned (cf. f7).
9. One subject saw the desired turn at 40, the others at once; the dot became visible to another one at 45. This result is not altogether unsatisfactory, but a real deviation to the left at the base is much safer.
Concl. : see Series I.
k.
2. The difficulty of series I, 40 was not repeated.
3. The total impression was unclear: one subject hesitated between b and k till 43.5 ; another saw something like an N, then an h till 36 ; the third described it as a Capital K with a long upper diagonal, the lower diagonal nearly parallel to the shank (at 52). Indeed both views are possible : if the observer is concentrated on the lower half of the character and on the join, a similar effect as described with series I, 40 arises: but now the end of the upper diagonal is stressed. If on the other hand the general outline is regarded, the two diagonals are seen in line, and so as h.
7. No difficulties, because the end of the upper diagonal is stressed.
9. Description as a normal k. The triangles were broader, in consequence the angle of the diagonals was 90°, the white vertical besides the stem was also broader ; therefore the difficulties of the Capital of this type were not repeated.
Concl. : The legibility varies partly with the angle of the diagonals (broad or narrow form) ; the omission of serifs at the base (type 4 and 6) is justified.



1.
. The results of the experiments with 1 will not be given here ; there were too many sources of error. The serifs of type 6 only were seen at 52 by all ss.
m.
3. Two subjects saw a block of vertical stripes, open at the top, onebwame sure at 43 5. The third saw a square D till 35.5, then suddenly an m with extended vertical at the left. No interchanges with n were recorded (cf. series I, 43). No ïmpro vemen ts suggested.
n.
3 One subject saw a D not quite closed at the base till 40.5. Another saw two verticals and between them (successively) a diagonal (cap. N), an unclear connecdon at the top, a diagonal to the middle of the right. Accurate descnpüon at 24 The cause of the error is not to besearchedin theeonnectingstrokelbecause the relatively equally thick ones of 4, 6, 7 were seen as thick as the verticals) but
a89m AlUubjectffone of them at 44.25) described as a broad n, uniformly thick everywhere; the right vertical slowly bending to the left and joining the other at a rieht angle • the white as broad as the verticals. This ïllusory movement is the most extreme one that we have met with this type, and that while the roun cornm were a iktle too sharp, but identical with u 9. The question m.ght be raised why thrtop right corner proved sufficiënt to induce the rounding movement with all subiects* whereas the two corners at the base of u 9 were unable to do so. It is better to ask why two parallel verticals are called an u. A possible exPlanaïion might bea?he fact that the discriminating parts ofa characterarem^ny si tuated in the upper half; therefore any details in the lower part are less likeiy to draw the attention and, in this case, to counteract this predisposition for n, caused by its more frequent use. See also D 9.
o.
3.' The usual troubles with the “two verticals”: interchange with D, U, hesita
tI09 S TherebeautiTul4score8made by the capital was by no means repeated. The renroduction was not quite faithful (total width 55.7 mm.; white lme over 4 mm. thtek being for the Capital resp. 61 and scarcely 4). One subjectrecogn.zedat 52 (white seen as a hole, just as the capitals), the others sawavertical white and also a horizontal movement from the centre to the nght.makmgaG or 6. IJJab y the white vertical caused a more oval movement and m this way the 6 snap^, which is made by the same writing-movement as the O but with a^etched^g nning. One recognition as ordinary O, the other as a solid disc, at . ,
For conclusion see Capital O.
P-
No mistakes or hesitations, not even with type 3.
q-
The same as p, but for one subject, who hesitated between q and 9 with type 3 till 40.5. No disjunction of loop and stem or short-cut or assimilation were
recorded.
r.
2. One subject saw the flag as a flat stroke till 46. recoeni-
3. Two subjects saw at first a vertical with somethmgatthetopaXrecogn
tions at 41, 30.75. Here we see again that, if there is undoubtedly something at



the top” which cannot be determined further, the tendency exists to make it symmetrical, balanced.
4. One subject saw an F at first, then a Z, finally an r with separate dot (the same as type 9). Tink er mentions the opposite case: r seen for z. The lower horizontal of Z is readily overlooked, which causes the interchanges with 7, F described before.
7. The serifs, almost 5 mm. thick, were invisible to one subject, who was specially asked about it.
9. One subject saw the dot as a simple stroke, the others saw a dot with connecting stroke.
Concl. : The flag, being alien, unorganical to the stem, is apt to be overlooked if it is not made extra-large. It is the same effect as the horizontal to L, F, I, series I, nr. 40 etc.
s.
A careful comparison with the Capital forms is very instructïve. Owing to the proportional enlargement, the lower-case characters are slightly heavier, which just makes the difference.
2. Two subjects described as a rather steep curve without prolonging further than the middle-line.
3. All recognitions at 52 ; the lower case is broader than the Capital.
4. One subject saw a 3 till 44.5, another one described an 8 at first, later on an 8 with opening above, but recognized at 52 too.
6. One subject mistook it for 8 at 52 but corrected himself. Only the diagonal movement was seen. One recognition at 40.5. On the whole an unclear character.
7. One subject saw an O (which means an “ignoramus”) till 42. The second mistook it for 8 till 48, the third saw a diagonal only, then an N till 37.5.
9. Very different interpretations were given. One subject: K, N, R, 8, reversed Z (curves seen at 25.5); another, X, N, reversed Z (ree. at 39). Again one recognition at 52, “without terminals”. The long verticals seem to be a serious trouble, though.
Concl. : To the conclusions on pp. 35 and 55 we may add the following remarks :
If the diagonal is made too steep, there is not enough room for a clear development of the arches ; if its angle is well chosen, the curves have to be clear in order to avoid interchange with N ; if, however, these curves cover a long vertical stretch, interchange with 5 or 8 may happen too. The diagonal cannot be induced by the join of two ending circlehalves for this reason and also because the halves are too dominating to allow a nivellation of the curves down to a straight diagonal line; on the contrary, they try to complete themselves. A little over-accentuation of the arch or its ends brings about the 8-shape.
t.
2. After seeing an r, one subject described a t with turn to the right at 40.5. Another saw the cross allright, but called it a t at 39 when seeing a turn to the right. The third s described a rather broad vertical with a triangle in the shape of a roof on it, he saw the cross at 42, but would not call it a t. The triangle is clearly a product of egalization of the cross. Though the phantasy of some



subjects adds a turn to the right, its actual absence proves to be adefinitive
obstacle for others to recognize. ... . .
3 Two ss. did not see a turn (at 52) but recognized by the cross stroke, the third j. mistook it for 1 and definitively recognized at 39, the turn was seen by
him at 48. The cross-stroke appears to be sufficiënt.
4 One subject saw a squarely built J tül 42 ; the oblique top and lower left
serif were notseen. Another saw the actual figure at first w>thastroke to the riaht at the top (making an f), that disappeared at 45 ; s. would not call it a t. The third 5. saw at first a cross above the cross-stroke gradual y thinning to the top (no serifs at the base) ; at 48 an f with tnangular base, not calling it a t because nf the svmmetrical base, ree. at 20.25. , ,
6 One 5. saw an r tül 41.5 (overlooking the turn), the ®thf.r^^tha square C (Darenthesis-sign) tül 42, 44.25. The oblique top and the little knob at the left are completely absorbed by the stem. Indeed, Mr. McMurtri e is justified to ask . “Why do modernType designers object so strongly to giving any height to a lower
'one subject seems to have been inhibited, when describing successively: f 1 K tül 27. No improvements suggested.
’ 9.’ Interchanges with 1, with f. Score: 32.5, 40.5, 46.5.
Concl. : The absence of the customary turn with types 2, 4, 9 proved to be a serious obstacle every time. A cross simply is no t. Those, who wish to remove the upturn (considenng it a superfluous rehc of handwriting-style), have to keep this m mind ; cross-stroke plus slanted top cannot do the work alone, they induce an f or 1 too. The interchange of 4 with t, observed by Tinker and once by us (see p. 67), shows the importance of clear projection at the left.
U.
3 The usual “two verticals” mistake : G, 6, 8, n, II, O. Score . 29.5,40.5,41.25.
4 One subject saw the top-ends sliced-off obliquely for which I know no explanation. ^Another'one'described the figure as almost ctad.'"*** «1 without th. cprif to the rieht at the base. The arrangement of the serifs at tne top gave no se^ous trouble with the n either, but with h it did ; the inner one should be omit-
te6 3^ïrif at the base (invisible to the others) was not quite understood by
rrrds ss
thl'top serifs; he saw also 11 (see type 4) and w^is not sure that the figure was closed, tül 37.5. The prolonged vertical wasseen by oduction was not
9 All subiects saw an n, one of them a H and 8 too. 1 ne reprouueuu.i v. faultlei: the tound edges’too sharp, but identical with those of the n.
Concl -Anarrow design will produce the two verticals-trouble serifs are in that case no improvement. The extended vertical at the right is superfluous. The curve should be conspicuous.
V.
3. One interchange with Y till 43J. heavy> causing mistakes
tilf45Tn‘one case. S^recognition'as v*already
white vertical, giving also a T-like ïmpression {im j , &
3t952 One interchange with Y till 46, the white diagonal almost halves the figure. For conclusions see Capital.



w.
9. Though in the exposed character the white diagonals were rather broad (4 mm.), the results can be taken as valid. All ss. saw the right diagonals complete (one of them saw two crossed V’s). Surprisingly little stimulus material is available to support this impression !
3. One interchange with 5 till 40.
4. „ „ „ z (thin diagonal) till 43.
7. One hesitation (N) at 52.
9. One interchange with 3 (thin ends) till 40.
For conclusion also compare conclusion Capital S (interchange with 3 and 5). The greater thickness and width of the lower case precluded the mistakes with the capitals to a large extent.
y.
3. No interchanges with V were recorded. Probably the relative length (93 mm.) was a help.
7. One subject could see no dot.
Concl.: Though type 9 came off well, I suppose that the prolonged right diagonal (supported by the resulting sharper angle at the left) was a great help : the movement of “folding together” (see concl. cap. Y and series I, p. 37) is obstructed in this way.
z.
3. One interchange with I till 40.5.
4. „ „ „ „ „ 46.
7. „ „ „ X „ 46.
9. „ „ „ X, K „ 34 (white diagonals too thick). For con-
clusions see Capital.
All these errors were made by the same subject.
Figures.
At this the type 1, 2, 3, 5, 6 (narrow edition), 7, 9, 10, 11 participated. Again the results with the figure 1 will not be given. We can only say that the 1 needs a conspicuous little stroke at the top, only to the left, and only downwards (cf. type 2 and 6).
2.
1. Two subjects saw a J at first, which means that a short-cut has taken place between horizontal and the right side of the arch. The diagonal is nearing the horizontal under a too sharp angle (see conclusion). Recognitions at 36, 52, 48.
2. One interchange with Z till 45.
3. One interchange with I till 37, a hesitation between 7 and 2 till 46.5. Rather good result for such a narrow type.
5. One interchange with Z till 45; again the lower horizontal doubtful till 44.25.
6. One subject described (at 52) as a reversed C on a horizontal (underlined). Another subject too saw only half a circle, then a 3, at 38 also a rev.Conhoriz., not connected. The third saw a Z till 32.25. The “underlined reversed C” can have two causes: the thin joint (the accents are directedaway from the joint)
Legibility, atmosphere-value and forms of printing types. 5



jnd the curve, which is prepared to be finished above the horizontal, and by no means to turn into a steep diagonal.
7. A hesitation between Z and 7 till 46. One recognition at 40.
9. Two interchanges with 3. One recognized at 35.5 as normal 2 with diagonal, the other at 43.5 after another interchange with Z. The third s. described at once a 2 as type 7 without dot. Just like G 9 and D 7, the circle-half is hollowed out; when its contour is followed up till the left of the horizontal, the diagonal movement can be seen. But is also possible to see this movement directed towards the centre of the figure, with the horizontal as a third movement from right to left, in which case the 3 is plausible.
10. One subject saw the horizontal at 39 only, another at 36 (after seeing an S, owing to lack of visibility).
11. One subject linked circle and horizontal in a wrong way : she saw a vertical at the right, makinga g (till 28.5). A second subject also saw a vertical, that suddenly turned to the left, as a kind of 9 (ree. at 42, certain at 34.5); ree. of the third s. at 48 (describing a horiz., diag. and big circle almost closed).
Concl. : The difficulties of the 2 are much the same as those of the 5, due to the heterogeneity of writing movements. Both turnings of the diagonal must come out fully. Now if the upper one is to be executed as a large curve, it ends low at the right and in consequence not much space remains to prepare an equal large angle for the base. In this way the bowl-shape, open to the left, originates interchange with J, reversed C, 9, whereas the 3 is seen in the way, described with type 9.
If, on the contrary, the designer starts making a large angle at the base, the diagonal must finish high-up, leaving room only for a short, sharp curve. In this case, the possibility of mistaking for Z arises, because the observer cannot decide between a compressed arch and a thick horizontal. Exactly the opposite happens, but with the same result, when the curve is too flat and broad, and therefore too much nearing a horizontal. In both cases a prolonged curve (with large dot or otherwise) can help out. The results with type 11 and (to a somewhat less degree) with type 9 show that a diagonal movement is important. Tinker : interch. with 3, Z, g, J, 2.
3.
2. One interchange with 5 till 44, one hesitation (“reversed C with something
in the middle, probably 3”) till 42. .
3. One hesitation (upper half seen as almost closed) till 43.5. One subject saw the terminal as two dots.
5. All recognitions at 52, but two after hesitation (S or 5). Total length 82 mm.
6. All ree. at 52, one as reversed square C with something in it, another description as large lower circle with a very little upper one.
7. One recognition as half a circle with something in it, afterwards correct description. One interchange with successively j (general form as j 5, cf. 2p 9,
Y, tül 37.5. „
10. Recognitions at 41, 37.5, 35.25, after considering the possibility of 2, 5, 8, 9.
11. One subject described a j, then a 3 of the general form of type 10 at 41. Another subject hesitated between 3 and 5, then decided on 5, describing the black lines correctly. The oblique end was taken for the usual flag of the 5. Recognition at 22.5. The third s. saw a reversed E, later on an S, till 36.



Concl. : The best results were scored by type 1, 7 and 6. They have in common an accentuation of the terminals, each about half-way the middle, and an accentuation of the middle itself (a good indentation or an extra stroke). Type 2 has a too compact upper half, but still it is a good character ; 3 made a satisfactory score too. The exact way, in which a horizontal at the top is linked with the curve (whether as 5 or as 3), appears to be difficult to fix. It is advisable to move the accentuation of the upper part of the 3 to the right and to reserve the stress on the left half to the 5 ; this has not been done with type 5. Type 9 shows the bowl-shape, the indentation is insufficiënt.
The bowl-shape, open to the left, which was mentioned in the conclusion to the 2, was again present here. Among the characters the J (with actual or imagined serifs to the left), S, Z, to a less degree also 9, y and modern g show it, moreover the figures 5 and 7. Accordingly, we have met interchanges with each of these. The X, it appears, does not make such an enclosing movement. Tinker : interch. with 2, 9, 8, 5.
4.
1. One interchange with d till 42. The horizontal seems too low.
2. The same mistake by the same subject till 40. Horizontal seen running down to the right by another.
3. Again the same mistake, later on also seen for A till 34.5. Another subject saw the triangle at 48, while the third recognized at 41.76.
5. One interchange with S till 45, another with 8 till 46.5 ; the third s. became sure at 49.
7. One interchange with t till 31.5 ; the others saw the triangle at about 45.
9. One subject saw a knob at the top left of the vertical as a reversed solid P, ree. at 41.5.
10. One interchange with A till 28.
11. One subject saw a hook (as the ma thema tical sign “less than”) ; obviously a short-cut was made from the bottom of the diagonal to the solid vertical. Recognition at 23, the others at 52.
Concl. : Type 6 (77 mm. high) turns out well. lts principal features are : the generous amount of vertical extending below the horizontal (to avoid the A or d), its general width, the large diagonal, obtained by cutting the vertical obliquely. Type 5 shows that two heavy masses, at the top left and at the right below, succeed in giving the diagonal movement, necessary for S, 8, etc. The diagonal movement in the other direction is absent here. Type 9 does not need bother to get a white triangle; the space, available therefore, is used to extend the vertical. The results with type 10 and 11 are good. Again the phenomenon of symmetrization or “putting right” is observed : the 4 becomes an A. Tinker : interch. with t, i, f, j, 1. The triangle at the left was very light, therefore the vertical predominated.



5.
1. One subject hesitated between S and g or 9; another saw tnis * too mi «ti.
The whole upper part is seen as an oval loop. Therefore the horizontal must be shorter, and the left upper half more stressed. ..
2. All subjects mistook it for s (one s. described also a g, circle with flourish
under it), ree. at 46.5, 48, 40.25.
3. One interchange with S till 34.5, another with g (9-shape) tul 36.75.
5. One interchange with S or 9 till 41.25 (length 84 mm.).
6. One interchange with 3 and S till 32.5, another with S till 45. Vertical invisible to another at 50, recognized by horizontal.
9. Two interchanges with 3 (one with J too) till 27 and 19.5, explanation was
given in similar cases previously.
10. Two interchanges with S till 36.75 (5 supposed at 44.25) and 4 ' .
11. One subject recognized at 50, calling the vertical the c.earestfeature. Another mistook it for 3 till 36, but became sure at 18.75 only. The third j saw the S-movement, described an 8 with the north-east part vague, then saw the solid strokes correctly at 40.5, ree. at 34.5.
Concl. : This figure adds to the solution. of the complex of problems surging from the mistakes of S, 3, 5, 9, 8. In regard to this figure, interchange is not avoided by the msertion oi a long and heavy vertical: the results of type 2 are bad. On the contrary when this stroke is almost omitted, the recogmtions are instantaneous ! (type 7). It seems that the mere fact of something being there in the dangerous corner increases the chance of interchange. To put the question in this way only, would lead to misinterpretations. For instance, type 9 has the vertical neglected too, but it is a poor type all the same. The difference with type 7 is to be found in the accentuation, in the distnbution of weight. The lower half is equal of both. If S, 3, 5, 9, 8 were constructed in the same rectangle, the only point of the 5, not overlapped by the others, would be exactly the top corner at the left (and c.q. the little upturn at the right). This corner is stressed by type 7, but type 9 gives equal weight to the whole upper part. Tinker : interch. with 8, 3, 0, a, 6, /0, b, a, y, o, S, Z, 2.
6.
I One interchange with G (with vertical and horizontal) till 34.5, a hesitation nf eaual kind till 43.5. The arch is hanging over too far.
3 One interchange with S till 39, a hesitation tül 45. One would expect
a G here too; perhaps the loop was too round and small. diaaonal is
5. One interchange with 4 (diagonal extending) till 32.5. The diagonal
accentuated too much.
5; Two interchangesawith G till 40.5, 42.5. Another with Q and with C (lower
ending half-way up) till 34.5.
Q T ikewise interchanges with C or G till 34.j, Zo.j, ***—.
10 One subject saw an O and then an S, ree. at 24. A triangle s®6” anod!®( till 39.75. The third subject saw exactly the left half of anA fd Oa| with horizontal cross-stroke) and hesitated accordmglybetween^ Arndfi.
II r»np «ihiect saw at first a diagonal from top left to the rignt, men an upcn 4 till'25.5. Another subject was hindered by the unclear base, and became certain
at 34.5.



Concl.: No interchanges with G, etc. were made with types 2 and 6, that have both a clear diagonal. If, however, this diagonal is too short, the pear-shape is seen (type 10). Dangerous is a full curve in the upper left region; presumably its absence saved type 3 from the G-impression. Tinker : interch. with 0, 8, 9, d, g.
7.
2. One subject mistook it for y, or a kind of cursive 4, till recognition followed at 39.
3. One subject suspected a 7 at 52, but then hesitated about the horizontal (y), finally described it curved (as handwritten) at 43.5. Another s. saw a I (with left serif only, general design as I 5) till 42, which can be explained by the joint effect of the tendencies to symmatrize and to put a slanted line right again. The swinged horizontal also occurred with Z 3, one will remember.
5. One subject mistook it for T, P, F (with upper horizontal extending to the left) till 30, owing to the general heaviness and the interrupting cross-stroke. On this small area and in view of the thickness it is difficult to decide whether one is confronted with a diagonal or a vertical. The others recognized at 52.
7. The horizontal was seen straight by two ss., they saw a cross-stroke through the diagonal (owing to the club-form).
10. One interchange with T (serifs at the foot) till 38, when a 7 with crossstroke was seen.
11. One subject mistook it for T till 38. Another described successively: T, T with cross stroke,horizontal with a point under it. Ree. at 20. The third s. saw a y, then a Y with hatched triangle till 43.5. These observations may be due for a small part to the hatching, that was on the heavy side. But it is chiefly, I think, the conspicuous width of the upper horizontal which suggests the T, because it is not counteracted by a diagonal extending to just below its left end, as type 2 shows. In this case too, the cross-stroke adds to the difficulties, because an undue triangular impression is given.
Concl. : Mistakes arise, if a short-circuit occurs between the horizontal and the diagonal; then the y can be imagined. If the diagonal is to steep balancing (T) becomes possible. Either a short horizontal (type 3) or a sharp angle will open this possibility ; it will be counteracted by a thin joint (type 6, 7, 9). A cross-stroke should be omitted. Tinker : interch. with : ?, y, i, 2.
8.
1. One subject described an A, then an 8 with heavy middle part (34). Another mistook it for 6, a, s, suspected an 8 at 46.5, but became certain 5 m. further. One interchange with B till 36.75. Indentation and white inner circles seem insufficiënt.
2. Interchanges with d, q, s, 5, a, D, B. Score : 25.5, 37, 33.75.
3. Interchanges with S, N, B, 9, 6, O, ree. at. 24, 37.5, 35.
5. Two interchanges with O till 37 and 43.5. Indentations good. Height 73 mm.
6. One interchange with V, which I cannot explain. Recognitions at 34, 43.5 (certain at 34.5), 35.25.
7. One interchange with 3 (cutoff at the left along a vertical) till 38.5, two with a till 43.5, 42.75. The vertical movement is brought about in the same way as the n of B 7 and R 7.
9. An a seen by all subjects. One of them saw a horizontal with half a circle under it too (as the black lines of 5n); another described as four square blocks. Ree. at 41. 40.5. 30.



10. Interchanges with D, D with cross-stroke, as B, 9, y, score: 27, 30, 21.
11. One subject saw an O, an A, then described the figure correctly with the
upper circle solid, the lower circle gradually thinning to the base (28). Another saw a triangle, then the upper part of the figure at about 44, at 30 described as “8 without lower part”. The third s. saw successively: e, c, arch (open downwards), broad, round A (arch with cross-stroke); recognition at 34 : lower circle thin, upper circle very small. f
The hatching of the exposed figure was on the light side, but I don t think the results could have been better anyhow.
Concl. : There is not much to be done in order to improve this character. It can be taken in two ways : as two superimposed circles, and as a closed S (diagonal movement). The latter kind is not present here in pure form ; type 7 can be regarded as nearest to it. Variations of type 7 exist, which are modelled on the principle of the S. Among the types shown here, 5 seems the best. lts width and the disposition of white make the diagonal movement difficult. No mistakes of importance were made with 7 either ; the indentations are good. In general, the hour-glass shape must be kept at, the junction in the middle thin. Tinker : interch. with g, a, s, 6, 3, 5, R, 9.
9.
1. One interchange with P till 37.5, which seems rather unfounded. Certainly
the lower left part is heavy. .
2. One subject saw a V with inscribed circle, then a Q till 45, also a misinterpretation which is not supported by the corresponding 6.
3. One interchange with S, 3, till 39. Another with V, reversed E, S, till 29. All ss. described it as g (modern form), with square upper right corner.
5. One interchange with P till 42.
6. One interchange with n (serif at the right limb) till 39; another with a,
o till 32.25. In contradiction with 6«. . ...
7. One interchange with n, later on with D till 33 (cf 87).The dot was invisible
(or abs’orbed) at 52. . . .
9. One subject saw only a diagonal at first, then described: a circle with ver-
tical at the right, that turns at a right angle to the left (40.5). Another subject hesitated till 37.5. Dot seen absorbed by all ss. .
10. One subject saw a 7 till 45, then a 9 open at the left (closed at 39), another
also saw an opening till 33.5. . ... . ,
11. One subject suspected a 9 from the first, became sure at 41.25 (the wnoie circle seen vaguely) ; another saw a cursive 4 till 43.5 ; the third saw a mere hook (the circle-half was apparently straightened) then a 4, ree. at 30 (circle seen luiiyj.
Concl.: The general results are worse than those of the 6. Probably, the important diagonal of both is more conspicuous with the 6. Where it is placed high, the tendency to doublé the diagonal (V) has to be noted. In analogy to the 6, the form of type 2 seems the best. The apparent difficulty of the exact location of the stretched part (interch. with p) was mentioned on occasion of b, series I. The oval shape is responsible for the “ two-vertical”-idea (interch. with n), which is related to the observed V.



Final remarks.
The fact that each type of series II represents a feature in type design, makes it possible to apply the results, by comparing and combining, on most of the other existing designs. The conclusions drawn from the results were meant to serve that purpose, i.e. to present some data to the type-designer, that will enable him to consider how much the recognizability will suffer, if he realizes his decorative aims. Very often the reading-conditions are so favourable (when the type face is read in a large size on a short distance), that he does not need to bother much over questions of legibility. When, however, such a surplus of legibility is absent, and consequently the determination of the optimal shape becomes important, the following conclusions may serve as a basis :
1. The distances of recognition indicate a range for each type, beyond which too many misreadings occur. On the score of these distances the rankorder of legibility would roughly be the following : 1 : type nr. 2 and 7 ; 2 : 1 and 6;3:4;4:9;5: 3, 5, 10 and 11 ; 6 : 6.
2. The optimal thickness of the lines will be about one fifth of the height for types without contrast and about as wide as type nr. 2. Below this limit the recognizability will decrease with the visibility 1), i.e. no changes in the proper subjective form will happen. If, on the other hand, the designer wants to make a type with thicker lines, he will have to modify the original proportions of most of the characters. He can introducé contrast or increase the width, or he must find totally new outlines and internal organizations if the same recognizability is to be kept.
CHAPTER 5.
Theoretical foundations of reading of letters in context.
The difference between the reading of isolated characters, of single words, of single lines and of several lines.
§ 1. The problem.
Since the first experiments of Erdmann and Dodge the consecutive investigations only confirmed and amplified these first principles : reading is not the successive apperception of single characters, but a combined process of apperception of collective
l) White using the term “visibility” we do not make abstraction of irradiation and imperfect illumination (reflecting black), as Javal does; hence we make visibility of lines dependent on distance too.



shapes of words or word-parts and of assimilation, i.e. completion of the partly apperceived words by inferring the omitted parts out of the already guessed meaning of the words. The term ‘reading’ can therefore properly only be applied to the perception of groups of words that are connected logically.
The description given of the process of assimilation should not lead to the conclusion that we can confine ourselves to the investigation of apperception, because assimilation is built upon the apperceived fragments by some constant intellectual process. It is the relative quantity of apperceived and of completed matter that makes up legibility, and these amounts are not only depending on the sensory qualities of the objects, but also on the attitude of the subject. It will be clear that therefore the reading process is not the same as the recognition process of single characters, transposed on larger scale, but that totally new and different factors are coming in. In the following lines we shall try to build up the complex process called ‘reading’ ont of its various factors.
§ 2. Recognition of single characters.
It was pointed out in Part I that it is important for the recognition of single characters that they have a good Gestalt. The characters of the alphabet were divided in classes, according to their conformity with the most coherent, the strongest Gestalten, which lead to the distinction between intrinsically legible characters (e.g. O) and illegible ones (e.g. S). We saw that the actual type faces are variations on the basic form of the character , that those type faces are the most legible ones, which come nearest to the basic form and which do not relapse into other or more simple and coherent Gestalten.
§ 3. Recognition of single words.
Gestaltfactors.
If a character is assembled with other characters into a word, it looses its individuality so completely that only wilful visual abstraction can isolate it again. Standing isolated, the surrounding white space being back-ground, its lines form a confined whole that cannot be brought into relation with its surroundings. But if it stands close to other characters, the space between the characters loses its back-ground quality to some extent. For example : the right-hand outline of an isolated character has its Principal meaning only in relation with the inside or the left-



hand outline ; standing next to the left-hand outline of another character, it can be regarded as the left-hand outline of the white space next to it. Under circumstances the outline as such is seen only in opposition to another figure beside it. Hence both the nature of the letterspace and the outline of the characters are of prime importance to the degree to which the singlecharacter Gestalten are suited for subordination under the new word-gestalt.
We observed in series I en II some letters falling to pieces, because a part of them had such a strong Gestalt, that the other parts could not be seen in relation to it. We wanted to make easily recognizable characters, i.e. strong Gestalten, but for that purpose we had to use parts of the strongest existing Gestalten, which consequently tried to restore themselves. Anthropomorphically speaking : they lived an own life before we forced them to play a subordinated role in new surroundings and now they try to become independent and to live their own old life again. Exactly the same viewpoint can be applied in regard of words.
The single character is a Gestalt of feebler structure than its components before their subordination, so again the word, which has a less strong Gestalt than its components (the single characters) originally had. And a very strong letter-gestalt does not want to be subordinated to the word-gestalt. Consequently the best form for isolated characters is the worst for use in wordrelation. These is the key to the problem of the relative legibility of roman versus sans-serif and also of roman versus Fraktur (see Chap. 8).
Space and outline.
Let us go back to letterspace and outline for a further substantiation.
Among the conditions promoting the creation of new Gestalten out of independent visual objects, have been found nearness and equality. So a group of scratches amidst other, but more widely scattered ones, will stand out as a Gestalt, and if in the same field some equally large circles are strewn, they will also detach themselves from the scratches as a pattern of their own.
Applied to our problem this means that no good word-gestalt is obtained if the characters are not well-fitted. Therefore spacing forces us to spell, which means : to read slowly and carefully.
Too little space ultimately leads also to illegibility, because the characters must retain a certain degree of individuality.
Similarity of the characters is promoted by similarities in the outline ; if all the characters are built in a rectangle, such as the old Gothic ‘Textur’, they lose their individuality, and the



result is illegibility. But maximal individuality is equally bad : the single characters cannot be subsumed under a whole. Those types are the best, that keep the juste milieu between individuality md equality.
Boldness.
There is another instance where the single characters remain too independent, and here we have bold types in mind. These are more visible, they call attention to themselves and that is reason enough to make them unsuitable for subordination. Moreover they have often less clear forms, because the necessary internal white space is consumed for the addition of weight. The lines come nearer to each other, which makes the occurrence of‘short circuits’, desintegration into simpler and stronger structures more frequent, as we saw in part I. Moreover irrelevant parts are more likely to be stressed.
The ordinary Bodoni editions have generally been found to be less legible, probably because of the combined effect of boldness and exclusivily vertical stress ; which does not lead on to the adjoining character. The general aspect of Bodoni can be much the same as of a ‘Textur’ page : the individual forms are lost in the multitude of verticals, because the equality makes a connection between all the latter ones.
Too thin characters often fail on the score of visibility, but if this is not the case, we must suppose that the white space in and around the thin lines lose their back-ground character or, which is the same, that the lines of the figure become mere outlines, seen only in relation to the white planes of which they mark the borders. Here again, as with bold types, an approach to strokes of more uniform thickness is made ; the result is that parts, which contribute little or nothing to recognition, receive relatively too much emphasis.
Experience.
In the word-gestalt not all parts stand on the same level of strength. Some of them (combinations of two or more characters, syllables, or even words, if they are used separately has) are parts of many other words too. It will be clear that these parts have a tendency to separate themselves from the word-gestalt, if one keeps in mind what has been said of the Gestalt in relation with experience. On the other hand, if they are too strongly connected with the other parts in the new complex, they are apperceived more easily. This means that word-endings, such as -ably, -ously, -ingly, etc., receive but little attention because we



focus directly on the stem 1), a circumstance which is the cause of many misreadings. For, if we give a word merely a cursory glance, because we recognize some parts of it and therefore take the meaning of the word for granted, it can easily happen that the unseen part is not the one expected. It takes some familiarity with the classical languages to be able not only to understand, but also to read quickly such composite words as incommensurability, irreducibility etc. The difficulty here is due, first, to the recognition überhaupt of these additional words as such. This recognition is partly founded upon knowledge of the entire addition, partly on conversance with the sterns (e.g. -duci-, -versa-, -voca-, etc.) because the syllables of the sterns mostly consist of a consonant plus vocal, which makes them look equivalent to ir-, re-, -bi, li, ty, whereas they should be taken together. Secondly, (but that is another problem) to the quantity of different vocals, which are not easily said in the silent pronunciation by which reading is accompanied. Where such a cumulation of the same or similar characters occurs, the attention is shifted from the total impression to the details, and that is where clear single characters are required. Always lack of knowledge of a word makes attention focus itself on detail, whereas command of the language makes it turn towards the whole.
Nothing could be more common than this, that we have a good look at something we must use without knowing it well. Another aspect gives the opposite ; that we should not be continually forced to look intently at something we know quite well. Translated for our problem : for any text, that contains many common parts (i.e. well-known = good Gestalten) we should use a type that does not force itself upon us, but that reproduces the text in a simple conversational tone ; for any text (be it half a page, or one word, or a line, or a row of figures) that does
*) At the time that all books were written with the hand, this practice. and reasons of saving energy lead to the use of abbreviations. Also after the invention of printing all these figures were cut as types together with the ordinary alphabet. Later-on it proved to be more economical to compose with as few characters as possible ; to-day an abbreviation sign would produce just the opposite effect, it would draw attention instead of making it superfluous. Still one cannot say that the problem is solved. Newspapers, notably in the headlines, omit all words that are not strictly necessary for understanding of the sense ; simplified orthography has the same purpose towards single words ; but otherwise all other suggestions for simplification are stranding on the almost unsurmountable difficulty that only a good deal of practice can make abbreviations really labour-saving instead of hindering. Stenography is there to prove the possibility of an even far more simplified orthography, that nearly everyone can handle. It is one of the advantages of the Fraktur that it can represent e.g. the three characters s c h, which are frequently occuring in the German language, in about half the space that is required by the roman alphabet.



ïot contain parts that we can expect on the score of the logica! ievelopment, or that we know well, we should use a type that presents us on its own initiative the whole thing cut and dried, dearly pronunciated. It is equally annoying to hear triviaüties being presented as important novelties as to hear messages carelessly and casually bold, which one has eagerly awaited.
The problem of the relative legibility of printing-types can only be well understood after analysis of the reading attitude. We saw that a character, that is eminently clear if isolated, cannot be subsumed easily in the word-complex. One next conclusion is, that apperceptive reading (interest directed on detail) requires another kind of type than assimilative reading, which needs an unobtrusive medium.
Importance of upper half.
The remarks about the differences in level of strength of the secundary complexes, (constituent parts) of the word-gestalt, can be substantiated in yet another observation. Javal was the first, as far as I know, to point to the far greater importance of the upper half of a word in regard to recognition than of the lower half; which can be easily proved by covering either of the half of a printed line by a piece of paper. We can read a line even if more than the lower half is invisible. If this effect is agreed upon, the obvious course is to make the upper half of the letter as characteristical as possible. Javal began with reducing the descenders to a fraction of their usual length ; the space obtained in this way came to the benefit of the letter-height.
This has become general practice, and as far legibility concerns, with succes. But in larger sizes than 9 point, these types become unsupportable if they are not leaded, though it is against the purpose of saving space (see also Chap. 8). From an aesthetical point of view they are decidedly dull, the mutilation has taken the zest out of them, making them top-heavy too.
In part I the improvement of recognizability by stressing the characteristical features was dealt with. We shall recapitulate the main points, with special consideration for the upper half of the characters.
Lowercase a has its top part in common with c, e and o. The loop should therefore be flat rather than curved, at the left it should end soon, at the right it should soon pass into the vertical.
b has its loop in common with h. This loop must therefore join the stem at a sharp angle ; that of h at an obtuse angle.
c has to avoid interchange with e ; the curve should end as high as possible ; see further part I, also for e.



g, with its little flag is not easily to mistake ; the g-form has its descender to distinguish it from q.
Capital I and lower case 1 have to rely on their serifs.
n should have a clear connecting-bow, whereas u ought to have short serifs.
r has the white gap to make it stand out; therefore the dot should not be placed too close on the stem ; neither should it approach the connecting bow of the n.
If these prescriptions are observed, one has not to fear that these characters are favoured at the cost of their aptitude for subsumption under the word-gestalt : it does not make them look bizarre. Moreover, the relative thickness of the lines, the contrast, and the width of the characters contribute far more to lack of coherence than these minor alterations of forms (at least if the alphabet fundamentally follows the classical example).
A comparison with Fraktur shows that the latter has some of its characteristical features not only in the upper half of x-height, but even in the ascenders, e.g. lower-case d, k, long and short s. On the other hand, some characters that have no resemblance at all in the roman alphabet, show little or no difference in Fraktur.
Conceming Fraktur Ehrenstein pointed out (basing himself on a theory of Ernst Mach) that the pointed endings to the limbs of Fraktur-characters should have an emanating effect, according to experimental results of Gestalt psychology; he calls it “Spitzenwirkung” (point-effect). He supposes that in this way Fraktur is able to produce the effect of a real line, where only the beginning is indicated ; that it therefore establishes Gestalten of singular character by simple means, such as is impossible for the roman alphabet.
We do not want to deny the existence of the mentioned effect; we have even proved it several times in part I and without extra pointed ends at that. But we doubt that the emanating lines or protuberances, suggested by those points, could exert a favourable influence on recognizability. A reconstruction of these lines gives sufficiënt evidence for this doubt : they would run criss-cross through the design, doing nothing but a lot of harm ; if they fall in line with a real ending of movement, at best nothing is added to the working that the ending would have had anyhow.
Serifs.
For some time serifs were regarded as relics of the stoneengraving and the subsequent manuscript period of the history of the alphabet. One could have wondered why during all these centuries no one ever attempted to abolish those strokes, if they



had no use anymore. But the sans-seni propaganaists aia noi bother about this fact and they acted accordingly. One knows the course of events : the sans-serifs proved to be unsuitable for book-work and a compromise was made in this sense, that the sans-serifs got serifs. So the “egyptians” reappeared.
No one can seriously dispute the important function of the serifs in the constitution of the word-gestalt.
We do not only have the practical proof, we can infer it theoretically too from what we know of the Gestalt.
The serifs fill up the gaps between the characters ; they mark the top and the base of the x-height and tie-up the bulk of the word in horizontal direction ; they constitute the identical element (beside the verticals) in all lowercase characters, with the exception of c, e, g, o and s, whereas t and z have naturally horizontal
cross-strokes. . , , •
If these serifs will fulfil their functions effectively their extension and general form cannot be restricted to mere indications (relying on the point-effect or Spitzenwirkung), because they would not be able to fill gaps in that case, nor should they be too thick in relation to the other strokes, and consequently be obtrusive.
Some display-types of the last century, which were taken in use again lately by French typographers, showed overstressed serifs, and produced a strong ornamental effect in that way, but their legibility was poor. Bracketed serifs were applied in some instances where it was considered necessary, apart from aesthetical reasons, to reduce the mass of black at the ends of strokes, which gradually flow-out in triangular shape. Indeed, such heavy endings can have a disturbing effect, but care should be taken to avoid too hollow serifs, which are almost parts of circles. ror serifs treated in this way cannot possibly do the work of tne normal ones, they show against the rest of the design and therefore
call attention.
Horizontal extension.
The horizontal extension of the word is an important factor in the relative facility of apperception. If the distance from the eyes is kept constant, this horizontal extension depends on 1. the number of characters, 2. the width of the characters, which again depends on the basic form (e.g. w is always wider than 1) and on the actual form, which varies with the size of the body (the a—z-length of a 10 point type is shorter than that of a 24 pomt type) and with the forms chosen by the designer.
We can neglect the extension due to the body-size.
Ceteris paribus, the more constituent parts a Gestalt is made



of, the more difficult it is grasped. This fact is not of great importance, because the condition of ceteris paribus is seldom fulfilled, i.e. our experience of the words of a text is seldom equally large ; if it is, then it is probably maximal and therefore again unlikely to influence the speed of apperception. Very long words are not much used, and consequently, the factor of experience coming in, we cannot isolate the quantity-factor.
Some languages, e.g. English, have short words compared with others, such as German. This fact has sometimes been neglected by English critics in the discussion of the problem roman versus Fraktur. They forgot that the narrowness of Fraktur has a hearing on this question (see Chap. 8).
The difference in width due to the basic form has to be mentioned for the saké of completeness, though it is of no importance for our view on the problem. In experiments with single words it must not be neglected.
Difference in width due to the design, on the contrary, influences legibility to a considerable degree. For a clear understanding we have to distinguish here between a kind of design that accidentally makes some characters narrower (e.g. the modern sansserif, that deprived i, j, 1, r of the little width they had and so increased their disadvantages) and the design that makes every character narrower on purpose, such as Fraktur or Cheltenham condensed do in the extreme.
Condensed types have been considered from the standpoint of typographical economy and of legibility of the single characters. The latter is indeed important, but it seems more promising to start our reasoning from reading.
Reading a same text covering a smaller area, means apperceiving more sense per unit of space. Assuming for a moment that observing a larger area causes a larger consumption of energy, the lecture of condensed print would only present advantages, if it were not a fact that condensed types are less legible. As it was said, one looked for the cause in peripheral functions, viz. in the clearness of the characters, whereas the cause can be found too in central functions, notably in the scope of involuntary attention. We mentioned the influence of experience upon the speed of recognition. It depends on this experience whether an increase of the amount of text perceived in one act of attention will mean an economy or not. If the text does not present any difflculties for the reader and if the factor of assimilation is accordingly large, such an increase can produce an increase in the rate of reading too. There is furthermore a physiological factor that complicates matters : the relation between the area



of distinct vision and the quantity of units which can De apperceived in one act of attention. However, we shall speak of that in the next paragraph, dealing with reading context.
Vertical extension.
A word can either have a greater height because the size in typographical points is larger, or because the characters are elongated, compared with the height of the face usual for that body. But this difference is a theoretical one, for the latter kind can be regarded as (horizontally) condensed.
The influence, due to the size of type, depends again on the experience of the reader with the word and on the amount of the word complex covering the area of distinct vision, which, in its turn, depends on the distance from the eye. A Lectern Bible must be printed in a larger size of type than a newspaper, if the actual retinal image will be the same. The question is therefore not essentially different from that of the lateral extension of a word.
§ 4. Recognition OF LINES of words (reading in proper sense).
The apperception of strings of words is radically different fro™ the apperception of single characters or words. We mentioned the fact that it is a combined process of apperception and completion, i.e. the reading subject does not make the same use of all parts of the object; the more he is conversant with the object or skilled in reading, the more he can confine his apperception to single points de repère, who possess, therefore, a higher egree of consciousness. As a rule these points that are thus standing in the focus of attention, are also standing in the focus of Vision, whereas the periphery of this field is not apperceived but completed. This state of affaire is not the only possible one : the coincidence of apperception and fixation is not a necessary law, but an efficiënt rule (Wundt) ; it is far more economical to focus attention and the eye on the same spot.
Theoretically there are three possible directions of attention, when the eye is resting on the same place :
1. focussing on the spot that falls on the fovea centralis,
2. focussing on the periphery of the visual field, i.c. either on the left, on the right or successively on both sides and the centre of the visual axis,
3. distribution of attention over the whole field.
Regarding the second possibility, we have said that it was



an inefficiënt method ; the ordinary subject usually does not success in doing it without some practice.
In respect of the third possibility we recall the investigations of Mittenzwey on the level of consciousness of various parts of the visual field at different ways of directing attention. It was found that a dilatation of attention over the whole field does not bring about a higher level of consciousness for the peripheral parts, which means that our first possibility, coincidence of apperception and fixation, guarantees the highest level of consciousness for all parts of the field, even for those parts of the field that did not stand in the focus. Though generally the concentration of attention causes a reduction of the size of the field, with a corresponding lower level of the non-focussed stimuli, this appears not to be the case in the visual field. Hence it has become possible to standardize the theoretically possible methods of reading on a single one, which is accordingly automaticized to a high degree.
Only the duration and number of eye-pauses are ever varying, within the same subject and from one subject to another. During these pauses of the eye wandering along the line the apperception is done.
The number of eye-pauses has often been used as a criterion of legibility, under the silent assumption that the duration of the pauses is constant for the same subject.
The relationship, however, between difficulty of the text and the way of reading, as expressed in the number and the duration of the pauses, is a very complex one, which does not give much support to the theorist.
Going into the causes that might induce the subject to change his way of reading at an increasing difficulty of the text, one meets in the first place with the contrast between the peripheral and central factors of the process of apperception. Now the hypothesis could be put forward, that the peripheral factors (viz. the number and figural nature of the objects to be found on the area of clearest vision and highest level of consciousness) exert their influence merely on the number of pauses, whereas the duration could be considered to depend on the central processes (“powers of comprehension” in general).
For if both attention and fixation remain directed on the same spot, a longer duration of the perception (given the short times with which we are concerned here) cannot make the latter more clear ; in the cases that a longer duration is actually found, the latter has to be ascribed to central processes such as distractions, fluctuations in attention etc. Only the groups 1, 2 and 3 of
Legibility, atmosphere-value and forms of printing types.



our list of objective factors are governing the peripheral processes, and we would be able to isolate easily the typographical factors from the other ones.
The proof of the correctness of such a hypothesis would be delivered, if the same subject were to read the same text with about the same number of eye-movements, but this is not the •case. We cannot even find any relation whatsoever between the number and the duration ; hence we have to take the total time as the principal criterion in the experiment, followed by the number of pauses.
Owing to these circumstances we cannot make a clear picture of the course of the reading-process.
One might imagine it to be somewhat as follows :
The subject starts to read with a certain speed and concentration of attention agreeable or usual to him. If it appears that the sense of the text does not penetrate to him sufficiently, he can :
1. Accept extra-strain in order to be able to maintain this (inadequate) speed ; his total time and the number of his eye-pauses will not show any differences compared to a more legible text.
2. Avoid extra-strain while maintaining his speed, which causes skimming, and therefore inaccuracy. Whipple and Curtis distinguish five ways of skimming : (a) fast reading ; (b) “trading”, which is “perceiving without apperceiving”, catching impressions, that may or may not be meaningful for context probably not in most cases ; (c) “covering”, like trading, except that it may be either up or down or right and left, while trading in the latter ; (d) omitting logically ; (e) omitting arbitrardy.
Evidently, ‘fast reading’ is not necessarily ‘skinjming, unless a kind of ‘trading’ is done, by which a low level of consciousness of the text is kept, without there being question of “catching up impressions, etc.” (which, by the way, includes omitting arbi-
trarily”). .
3. avoid extra-strain by lowering his speed to a suitable degree. The lower speed can be the result of either a longer duration of the pauses, or of a greater number of them, or of a combination of these two factors. An increase of the number of eye-pauses always causes a longer total time, as the length of the pauses cannot be reduced below a certain minimum, which is not much smader than the ‘normal’ length (owing to the automatization).
Therefore the original total time cannot be maintained by a reduction of duration proportional to the increase of the number.



Taking this into account (see 7), we can draw-up the following scheme of possibilities :
Eye-pauses
Number Duration Total Time
1. equal longer longer
2. equal shorter shorter
3. larger equal longer
4. smaller equal shorter
5. larger longer longer
6. smaller shorter shorter
7. larger shorter longer
8. smaller longer equal
The number of pauses, we can say that much, depends on the amount of sense perceived per area of vision (direct + indirect vision). The eye will be shifted from the apperceived group of letters to another group, which is so far removed from the first that the subject can pursue the course of thoughts contained in the text by making use of the following data :
1. the parts he actually reads in the first group,
2. the following words, which he expects on the score of the meaning of the first group,
3. the parts actually read in the second group.
The letters or words between the first and the second group need not be read (thought they may be) if the meaning which the subjects expects them to have, is confirmed by the words of the second group. If the subject has shifted his eye too far, he will experience a gap in the course of thoughts, which can only be filled up by spending an extra look at the words that thus became insufficiently conscious to him. In other words : if a text is difficult, i.e. if the course of thoughts cannot be predicted by the subject in order to pass over any parts of the sentence that possess a lower degree of consciousness, the subject has to reduce the length of his eye-movements in such a way, that peripheries of the areas of vision are overlapping, so that all the parts of the sentence can be apperceived, instead of being reconstructed (completed).
This difiiculty can be due to different causes. It may be the meaning of the text itself, or the illegibility of the letters, the frequent interruptions of short lines, or the abundance of stimuü. The latter case will be present where too small a size of type or too little interlineage is used, which makes a reduction of the



ield of attention necessary, so that the distraction by the other itimuli is avoided.
If, on the contrary, the area of vision presents too few stimuli large size of type) the number of eye-movements has to be ncreased too (and also length of the eye-movements), because here is here no possibility for reconstruction (completion) either; ;he reading of the first group of words does not give enough iata for that. In practice, by the way, the reader will bring nore letters into his area of vision by holding the text at a greater listance from the eye, but if he does not think of doing that, ie will have to make more eye-movements per word.
We have to mention here Kirschmann’s theory, stating that i typographically easier text enables the subject to adapt the length of his eye-movements to the text, avoiding therewith a uniformity of the contractions of the eye-muscles. This supposition is correct, with this restriction however, that not only typographical factors will lead up to an increase, and therefore to a greater uniformity. The substitution of the legibility-criterion inequality of eye-movements’ by ‘number of eye-movements’, which was rejected by Kirschmann, is, we agree to that with Lobsien, entirely justified, provided the acknowledgement of the fact that not the number, but the uniformity is the cause of ocular fatigue.
The central principle governing the way of reading remains the relation between the momentaneous ability and willingness of the subject to apperceive, and the amount of meaningful signs offered to him in one act of attention.
This relation is experimentally inaccessible as long as we are unable to neutralize the influence of the capricious intensity of nttpntinn.
CHAPTER 6.
The measurement of reading-speed of a line of words.
Series III.
Method.
The purpose of this series of experiments was the measurement of any differences in the rate of reading due to differences in the design of the type face, contrast, boldness, width and serifs included. Consequently, the influence of interlineage had to be kept constant. But because each type face has its own optimal interlineage, given a certain size and length of line, and because



therefore the same interlineage might have unequal results, it was decided upon to make use of one line of words only. This enabled us at the same time to reduce the differences in speed due to central factors, which necessarily occur during the reading of a longer text; moreover, by following this method, we could ask the subjects to reproduce the text verbally, thus giving a method to control the “intensity” and correctness of their reading. Coupled with the neutrality of the subject of the text (newspapercuttings) and its shortness, which reduced the possibility of any interfering associations, it assured fairly well a uniform attitude of all subjects and of the same subject towards the single text.
Of the two possible methods of measurement (time needed for given amount, or amount read in a given time) we chose the latter.
The standardization of the length of the text would have caused lines of very different typographical length and some schematism in the construction of the sentences, which would have to be composed specially for the occasion. It was feared that, owing to the closed form and the closed sense of such Standard lines, personal differences would show up more, than if the reading subject were left in suspense regarding the logical development of the sentence. The latter could be brought about by leaving the sentences unfinished and of equal typographical length; measuring, in that case, the amount read in a given time, not long enough to finish the line.
Material.
Six type faces were chosen, which could be called representative for the principal kinds of design, and, which had approximatingly the same height of face.
With each of them twelve lines of text were composed, and these lines, printed in black on good white paper, were gummed on white cards, for insertion in the quick-exposure apparatus.
A few of these lines, which contained no capitals, interpunctions or other typographical signs, are following here (translated from the original Dutch ones), to give an idea of their nature :
the new house of representatives will be chosen for a period of four years
it is therefore highly probable that violent thunderstorms which were accompanied with num
the significance of this change for the following days can as yet not be
during the war the necessity was repeatedly stressed of



the president mentioned a report in which it is said that the accus
she has a considerable sense of duty and to a certain extent she would not hesitate
The length of the lines was 28 didot augustins (12.55 cm.). The wordspaces were of uniform width (IJ mm.) and the ends of the lines were made exactly as it turned out, mostly in the middle of a word. A hst of the types follows together with their actual size, measured by the stem of the lower case i, and by the distance from the top of the 1 to end of the j.
Name point official height height
size designation of i-stem of 1—j in mm. in mm.
1 Monotype-Modern Wide 11 25—11 1.64 3.38
2 Monotype-Old Style 11 2—11 1.72 3.46
, 3 Intertype-Ideal News 10 1764 1.78 3.36
, 4 Intertype-Mediaeval 10 813 1.58 3.56
Intertype-Bodoni 10 800 1.5 3.34
6 Futura Buchschrift 9 1-66 3.32
The next larger size available of Bodoni (nr. 908, 12 point) measured 1.66 and 3.8 mm. resp., which comes nearer to the i-size of the other type faces, but is otherwise larger in design. The lines in Futura were composed by hand, with type from the Bauer Foundry at Frankfurt am Main.
Apparatus.
The exposure apparatus (see fig. 16, under A) consisted of a vertical square plate, large 40 x 40 cm. or 16 x 16 inches, covered with matt white cardboard, with a horizontal and rectangular opening in the middle of 10 inches high. Behind this plate a metal slide with the same kind of opening could move up-anddown ; it could be kept at different heights (closing or opening the opening in the front-plate) by way of two studs, fitted in indentations at its sides, and which were commanded electromagnetically. The electrical circuit including these magnets was opened and shut two times in rapid succession by means of a rotating contact, the constant turning speed of which was secured by a highspeed electromotor with the aid of a gearing system. Before every exposure the slide was put in its highest position by the experimenter ; it then covered the text, which was placed vertically behind the opening of the front plate and of the slide. The first electrical contact made the slide fall till its opening corresponded exactly with that of the frontplate, thus exposing the text; the second contact made it fall on the rubber cushion on



the bottom, simultaneously shutting-off the opening. The slide was also painted white ; the space between slide and text-sheet was 5 mm., in order to prevent differences in accomodation of the eyes. The subjects were seated with the head in a head-rest at a distance of 50 cm., or 19J inches from the text to the eyes ; their attention was directed on the beginning of the printed line by an indicator fixed at the head-rest, and which pointed immediately below the spot in question. The illumination consisted of a 100 Watt bülb immediately behind and above the subject. The time of exposure was 1 second. The experimenter, in order to make attention maximal on the moment of exposure, warned the subjects IJ sec. before the exposure, judging on the position of the rotating contact. Immediately after the exposure the subjects mentioned the last word or character they had read, which was recorded by the experimenter, who next restored the slide in its proper position, and changed the text. The 72 lines of text were given at random order, though never two in the same type after each other. The sessions lasted about an hour and were frequently interrupted by chats between the subjects and the experimenter.
Four subjects were employed, viz. three undergraduates and the director of the laboratory ; they received the instruction to read quickly but accurately, in order to be able to reproduce the text after exposure. A few lines were exposed to acquaint them with the proceedings.
Results.
A first series of readings was held, in which Futura Light was used as sixth type face (Intertype-154-9 ; the Medium was considered too bold), but it had to be discarded owing to the poor results ; apparently it was too light. In a second series the Futura Bookface was procured.
The total number of letters of each of the 72 lines, actually read by each of the four subjects, was counted ; out of these the number of letters, read by each subject in each type face, was computed. It appeared, however, that the nature of the words on that spot of the line, read at the moment of ending the exposure, influenced the counted number of letters considerably ; even to such a degree, that the mean deviation of the scores of the single lines from the total sum of scores of the given type face was too great to make this final sum reliable. The interruption of the reading act did not meet any difficulties if some short words, such as in, is, on, the, by etc., were found at that place, but if a long word stood there, the reading (or



rather the reproduction) shopped either before or after that word. rhe length of such words equalled the differences in length due to differences in speed of reading. Also the difficulties of the sentences in general proved to be very unequal, e.g. a sentence as “the enormous and unsound enlargement of various leather factories in the” is running more smoothly and does not bear so much stress on each word as : “much attention was devoted to the announcement by the”, and it is clear that such reading texts have to be standardized, like the Chapman-Cook Speed of Reading Tests, used by Tinker and Paterson.
Several of the recordings had therefore to be discarded. An increase (and partly revision) of the text material and of the number of subjects might have corrected some of the short-comings, but would not have met the principal objection : that the influence of the text is predominant over the typographical one, especially in such a short line, and that this influence cannot be controlled. We regret that we were not able to standardize a suitable text, for the method to measure single lines possesses several advantages above the measurement of longer texts, as was pointed out above. Because a continuation with more subjects would not have brought about the necessary improvement, it was decided upon to stop the experiments.
Below we reproduce the scores of the four subjects ; they read the text of the types 1—5 two times (with an interval of at least a month) and that of type 6 only once.
TABLE I.
The results indicate a minor legibility of type 6, and a higher one of type 4 and 5, which cannot be due to their width, as the number of letters per cm. (spaces included) show.
Series IV. Registration of Eye-movements.
Method.
In view of the apparent impossibility to measure speed of reading with unstandardized texts by ordinary methods, the measurement of eye-movements was considered. Several authors (e.g. Galley, Lobsien) supposed a direct relation between legibility and number of eye-movements, others (Kirschmann)
Type nr. J 2 3 4 5 6
number of letters 52 51.8 53.5 54 55 50.3
mean deviation 4.5 5.9 5.8 4.9 6.1 5.0
letters per cm. 5.48 5.71 5.64 5.95 6.03 6.14



rejected the latter view and regarded the regularity of the movements as a criterion.
If the legibility of a type face influenced the speed so little, that the same reader could finish the same text in totally different times, entirely according to his “moods”, then we thought it interesting to know at least how he read it, with the same number of eye-movements but shorter pauses, or with less movements and pauses, by taking more letters into his span of vision.
There are various methods for recording. First direct observation by the experimenter, who counts the movements, by pressing every time a Morse key connected with a registrating apparatus, or in another way. Hindrance of the subject is avoided by the use of mirrors or special closed and screened observation stands. In this way only the number of pauses can be counted.
The most subtle method makes use of a beam of light, directed by means of a small mirror upon the subjects’ eye and reflected by the cornea upon a photographic film moving at a constant speed. An interruption of the beam of light at e.g. 50 times per second makes the traced line on the film appear as a series of dots, suitable for a very accurate calculation of the time. The technique of this method however, is very difficult and it establishes reading-conditions that are far from being natural. Eurich, who used a camera constructed at the University of Minnesota for this purpose, does not give a detailed description 1). A criticism may therefore not be applicable to him ; but usually such a photographical method requires an absolutely dark room, in which the lighting of the object cannot penetrate in the camera, and a rather high intensity of the recording beam of light. This strong light, directed on the subject’s eye and the strong contrast between the lighting of the object and the rest of the room, cannot fail to be highly disturbing to the subject. The possibilities of direct observation being limited (owing to the number, the speed and the smallness of the observed eye-movements) such recording methods are the only ones, by which backward movements (refixations) or in general the kind of movements of the eye can be recorded. But the almost unavoidable artificial setting of the experiments seems to outdo these advances. Needless to say that similar methods which make it necessary to fix a small mirror or a disc connected with a wire (Delabarre) on the desensitized cornea are even more objectionable.
A third way makes use of the flattening of the bulbus, some lever system being operated by this eccentric form. Lamare, at
l) J. appl. Ps. XVII, 1933, 604.



instigation of Javal, fitted a light peg externally on the eye-lid, that touched the membrane of a microphone, connected by tubes with his own ears ; in this way he was able to hear movements of his eyes. A far more practical application of this principle was devised by Schackwitz to serve as nystagmographx) and used by him and by Lobsien for the investigation of legibility. This apparatus was adopted by us and constructed for this special purpose by Messrs. F. Kooyman, opticians in Utrecht (see fig. 12).
It consists of two parallel bars, that can be shifted in respect of each other ; one of them is fixed to a nose-rest and both of them end in the ear-hooks customary for spectacles. So the frame can be adjusted to every shape of head of the subjects. One of these bars holds a cross bar, that can be fixed in various positions by means of a screw, and that is directed towards the eye. It carries at the end a disc of 12 mm. diameter, in the centre of which a hole is made ; this hole stands in connection with a short tube as side-outlet. On the disc an air-tight rubber pad is braced (the top of a thin rubber condome can serve well for that purpose). The side-outlet of this pad was connected with a small and sensitive Marey drum by means of a narrow rubber tube (as used for bicycle-tire valves). The rubber pad was adjusted to one of the eye lids, sideways and on the upper half of the eye bulb ; hence it was pressed-in according to the
*) A. Schackwitz, Apparat zur Aufzeichnung von Augenbewegungen beun Zusammenhangenden Lesen (Nystagmograph). ZfPs 63, 442.
FIGURE 12



movements of the bulb. Extra-pressure proved necessary and was regularly added by means of a rubber balloon connected to the tubes by a fork junction with a tap. Too little pressure makes the smaller details of the curve disappear ; too much pressure however, has the same effect, moreover the pad has to be pressed too firmly to the eye in that case. If the pressure is right, one can produce very clear curves without any hindrance for the eyes of the subject (see fig. 13).
The capacity of the tubes and drums has to be kept low, therefore we used a drum of 17 mm. diameter, covered by thin rubber
FIGURE 13
cloth. The depth of the drum and the diameter of the holeconnecting with the rubber tube measured both 3 mm. A larger drum is less sensitive and the elasticity of the larger area of rubber cover can produce undesirable vibrations, hence also the point of support of the recording needle resting upon the centre of the drum should not be executed as a little disc, which almost covers the whole surface of the drum, but as a small knob of only a few milümeters diameter. All turning points should work as easily as possible. The needle consisted of a thin straw ending in an aluminium point; for greater distinctness of the curve it was made as long as the weight allowed.
An ordinary kymographion with sootened cylinder was used ; attempts to use a Boulitte recorder did not meet with any success, because the drums and recording pens were not sensitive enough. Before the extra sensitive Marey recorder was developed, attempts were made to fix a concave mirror to the drum (with a diameter of only a few millimeters), instead of the usual recording needle, for photographical recording, in order to enlarge the small amplitudo of the drum-skin in the resulting curve. This method appeared to be possible, but, when the successful drum was found, it was discarded on the score of the same objections| that were raised against the other methods * of photographical recording. The time was recorded by a tuningfork in fifths of seconds. The beginning of the readings were indicated by the subject pressing a Morse key, connected with a recorder ; the



endings could be read more accurately in the steps of the curve.
For each subject the right place on the eye has to be found, which produces the clearest curves without causing eye-strain and which does not require a partial or entire closing of the eye-lid.
The readings were done in daylight during the summer ; the subjects (two students and the author) were sated close to the window in a normal reading poise, holding the reading object in the hands at convenient distance. No direct sunlight feil into the room.
Material.
In accordance with the preliminary character of these experiments no special reading material was used ; it was thought sufficiënt for the achievement of our purpose — the investigation of factors influencing the number and duration of eye pauses — to employ type samples of foundries and printing offices, that consist of a same text printed in different ways.
The repeated reading of a same text has obvious disadvantages : the subject, knowing the contents almost by heart, falls easily and unconsciously into skimming. This effect however, can be counteracted to a certain extent by reading with inner speech, though it should be kept in mind that this is a special kind of reading ; hence the results obtained in this way cannot be generalized. Otherwise there is no other control of the reading method than the subject’s conviction that he spent the same amount of attention. The necessary equivalence is of course maximal in case of identic texts.
Another objection can be raised to the shortness of the texts. In view of the intensive method such an objection seems less founded, because the best material hitherto used — the ChapmanCook Speed of Reading Tests — takes IJ minutes and even then proved to be rather short. Hence at least several pages of print would be necessary to meet the need for a larger material; the standardization of such material presents considerable difficulties, whereas the costs of printing would be rather high if e.g. 6 different types, and also different lengths of line, leading, kinds of paper etc., were to be employed. Hence we confined ourselves to the following material :
1. “American Types”, sample book of the American Type Founders Company, Elizabeth, NJ, 1934, page 192—197.
2. “Intertype Faces” sample book of the Intertype Corporation, Brooklyn N.Y.



3. “Linotype Faces” sample book of the Mergenthaler Linotype Company, Brooklyn N.Y.
4. “Linotype” sample book of the Mergenthaler Setzmaschinen Fabrik, Berlin, Germany.
5. The sample book of Messrs. J. van Boekhoven, printers at Utrecht, Holland.
Results.
The principal difficulty for the subjects was to keep an equal amount of attention and a uniform reading attitude without opposing any possibilities for greater speed, i.c. without counteracting their own impulse to read legible texts more quickly. The different scores made by the same subject on identical objects made it necessary to repeat the readings for control. The divergence of the results gave us reason to reject the whole method in principle and to refrain from continuation with a larger number of subjects and improved material. We attach some value, though, to several findings, either as negative evidence, or as indications that need corroboration in order to become evidence. And it is for this reason that we present part of the results hereafter.
a. Dutch text of 150 words in sample book nr. 5. Length of line 24 Didot augustins (109 mm.), first line in capitals. See Table II.
What first meets the eye is the difference in speed of reading of the three subjects.
Subject nr. 2 reads about 60 %, nr. 3 about 20 % slower than subject nr. 1.
A comparison between the columns 19 and 20 (type 2, 8, 10, 14 !) also shows large differences for the same subject.
The pause-length is varying likewise ; cf. the figures for type 8 and 9 in column 13 and 14. These scores raise the question whether any causal relation exists between the duration and number of eye-pauses and interlineage — a question which cannot be answered in a positive sense on account of the present material.
The total values seem to be in favour of the types 2, 3 and 5 in the first place, all three small faces, and leaded. The types 8, 9, 13, 15, 16, 17 were read comparatively slowly, but why e.g. nr. 15 should cause such a retardation is not clear ; probably its length was too great. The optimal length for this text, read in this poise etc. is lying between 1400 and 1550 mm.
For a better illustration of the entire lack of relation between the types and the pausefactors table III presents a selection of identical values out of scores of subject nr. 1.



TABLE II.
Scores for text nr. Ia.
a m Subject nr. 1. Subject nr. 2 Subject
g g g H . ■ NR. 3
g * jO ®£ 3 5 £ h fc !2i £ W K 5 R
m og 3s 3g a g ai pa . «S " h ■ . «S - « ®„
1 name of type face ós sg „g g *£ gs gs i* g| I ei £* *g b gg g; g»
z s2 °s i« ■ i r & a r 5 r 1» a y « s- hb ^
co ** S? H H Eh H
j - 2 3 4 5 0 7 8 9 10 11 12 13 14 16 16 17 18 19 20
1 Neo Didot M 65—6 6 12 1383 115 81 5 4 223 110 16 1 233 127 150 161
2 Neo Didot M 65-6 6 + 2 12 1383 122 83 8 6 197 118 4 3 44 60
3 Garamond M 156-8 8 f 1 13 1430 108 76 1 1 202 124 5 5 49 52
4 Ideal News I 1579 6 13 1516 113 81 3 3 208 126 6 6 170 173
5 BaskerTle M 169-8 8 + 2 13 1516 115 76 4 2 190 122 2 4 195 134 51 55
6 Mediaeval I 713 8 14 1540 111 85 2 7 211 28 7 1 59 59
7 Bodoni I 687 8 14 1540 116 86 6 8 213 131 8 12 62 67
8 Futura Semi Bld. I 154—8 8 15 1657 130 96 12 16 229 115 7 2 61 83
9 Futura Semi Bld. I 154-8 8 + 2 15 1657 133 93 13 13 218 127 12 9 56 63
10 Cheltenham M 99-11 10 + 1 15 1659 139 98 15 17 188 128 1 10 35 67
11 Old Style Ant. M 3—10 9 + 1 15 1670 139 92 14 12 191 127 3 7 56 162
12 Garamond M 156—10 9 + 2 15 1699 124 86 9 9 214 138 9 4 48 48
13 Mediaeval I 740 9 15 1722 125 96 10 15 220 131 13 13 173 76
14 Ideal News I 1728 9 16 1805 121 87 7 10 117 215 127 11 8 159 175
15 Baskerville M 169—10 9 +2 16 1805 142 90 16 11 214 143 10 17
16 Baskerville M 169—10 9 16 1805 148 94 17 14 220 139 14 15 225 141
17 Bodoni I 800 10 + 1 16 1832 129 82 11 5 220 142 15 16 154 160
0\



TABLE III.
Selection out of scores of subject nr. 1 for comparison of identical
values.
S g 2 S S é o* M M
* £ §§ ■ J S 5- 2S S
-S Si *2 2- 2® 2P* rt §2 1“ S
33 - s* Ss *5 2 -a -s g-
15 g£ 3 a 5 es e" 2
1 2 34 567 8 9 10
14 Ideal News 16 1805 121 87 20.6 1.39 0.69 0.48
16 Baskerville 16 1805 148 94 19.1 1.57 0.82 0.52
12 Garamond 15 1699 124 86 19.6 1.44 0.63 0.50
13 Mediaeval 15 1722 125 96 17.8 1.30 0.72 0.56
1 NeoDidot 12 1383 115 81 17.0 1.42 0.83 0.59
5 Baskerville 13 1516 115 76 19.9 1.51 0.76 0.50
1 Neo Didot 12 1383 115 81 17.0 1.42 0.83 0.59
17 Bodoni 16 1832 129 82 22.2 1.57 0.70 0.45
5 Baskerville 13 1516 115 76 19.9 1.51 0.76 0.50
15 Baskerville 16 1805 142 90 20.0 1.57 0.77 0.50
6 Mediaeval 14 1540 111 85 18.0 1.30 0.72 0.55
13 Mediaeval 15 1722 125 96 17.8 1.30 0.72 0.56
14 Ideal News 16 1805 121 87 20.6 1.39 0.69 0.48
17 Bodoni 16 1832 129 82 22.2 1..57 0.70 0.45
b. English text out of sample book nr. 1, page 193. 7 lines,
the first line in capitals (except the types marked with * in table IV), the last line in italics. Length of line 13 pica’s (55 mm.) ; size of type 8 points. See table IV and figure 14.
The types are shown in descending order of speed per 10 letters (column 6). Even on these short texts the intensity of the used method can bring forward considerable differences in the speed of reading. For better comparison the speed for 60 words of 5 letters was computed out of the results obtained on the 7 lines (totally 385 mm.).
The differences in speed are considerable : 1000 words of 5 letters would take several minutes more of the readers’ time when set-up in type nr. 24, than if type nr. 1 were used.
Comparing the types that produced the best and those that produced the worst results, the eye is struck by their heterogeneity. Piranesi and Century Catalogue are very dissimilar, but so are Stymie and Bemhard Booklet. Wide and narrow faces are found at both ends of the list.



TABLE IV.
Scores of text b.
The types marked with* have the first line printed in lower-case type.
The causes for the high speed of the small and narrow types 1, 2, 3 and 5 are different from those, that governed the large and wide types 4 and 6. Apparently, the former are legible because the eye does not need to dweil a long time upon the lines (average speed per cm., small length), the latter because the words are very clear (high speed per cm., considerable length).
A type, that combines these features, being small and still clear, will produce the best results.
The common feature of these “good” types appears to be a rather small and evenly distributed weight, tight fitting and clear forms of the single letters : avoiding of stress on the single letters, like it was pointed out in Chapter 5. A survey of iigure 14 will show that the things that first strike the eyes in the less legible types, are either single characters, or unusually shaped parts of them ; a dazzling mass of meaningless strokes, dots or curves,
GS fa tz fa o cq to co fa o O Mm H
fa O i-i O rH ^ Ss
£ hh. on H SfaO S
S rfS So- rfs * s 5E gss rfSg g|s-
« 8§ bs sa- *** " *51
^ h ^ o* ^ ■< „ S o® b®°'
■<* 2 <j • fa gba 1-1 H
* S 2^5° s »B° JU Bp! a6 5®
0 •< O o o SS Ja f{"g^2£o
55 Z B fn Eh Eh Z pi p» SaF P5
1 2 3456789 10 11
1 Cloister Lightface 323 42 35 1.30 1.08 39 33 362 5
2 Baskerville 339 45 35 1.32 1.03 39.5 32 348 11
3 Piranesi* 362 49 37 1.35 1.02 40.5 31 328 20
4 Century Catalogue 289 41 33 1.41 1.14 42 34 400 3
5 Caslon O.S. 471 344 49 34 1.42 1.00' 42.5 30 342 21
6 Roman 510 294 42.5 33 1.44 1.12 43 33 392 7
7 Century Old style 288 42.5 33 1.47 1.14 44 34 400 8
8 Century Schoolbook 264 39 32 1.47 1.21 44 36 435 1
9 Bulmer 294 44 32 1.49 1.08 44.5 32 392 9
10 Goudy Old Style 293 44 33 1.50 1.12 45 33 392 10
11 Bernhard Goth. Lt.* 326 49.5 36 1.52 1.10 45.5 33 362 22
12 Caslon 540 301 46 35 1.52 1.16 45.5 35 385 15
13 Garamond 296 45 35 1.52 1.18 45.5 35 390 12
14 Cloister Old Style 316 48.5 36 1.53 1.11 46 35 375 17
15 Scotch 294 46 34 1.56 1.15 46.5 34 392 16
16 Schoolbook O.S. 259 40.5 34 1.56 1.31 46.5 39 445 2
17 Bodoni Book 307 48.5 32 1.58 1.03 47.5 32 385 18
17 Bookman* 288 45.5 35 1.58 1.21 47.5 36 400 14
19 Goudy Catal.* 282 45 33 1.59 1.17 47.5 35 407 13
20 Century Expand. 268 42.5 37 1.58 1.31 47.5 41 430 6
21 Modern Rom. 64 259 42 32 1.62 1.23 48.5 37 445 4
22 Bodoni 288 48 38 1.66 1.31 49.5 39 400 19
23 Stymie Light* 291 50 38 1.71 1.30 51 39 395 23
24 Bernhard Booklet* 321 57 37 1.77 1.13 53 35 365 24



17 2 1
Eigijt Point Bodoni Book and Italic. The Eight Point Baskerville Roman and Itauc. Eight Point Cloister Lightface and Italic.
body type faces shown on this and accompany- The body type feces shown on this and accompany- The body type feces shown on this and accom-
n™tW histn S°tl °f 1116 tyP6S °c !ngPage£,fesome of th.e femous types of pnnting panying pages are some of the femous types of
P”,"* ï”t y' T^y cover a wide range of history. They cover a wide range of styles making printing history. They cover a wide range of
selfectl.°"oftlLe correct easy the selection of the correct body type for any styles making easy the selection of the correct
£°dy !yP f° 5 y;P f Pi.mtl1g' Tre lln.es plece ,of P™t,ng- The lines have been made long body type for any piece of printing. The lines
h been made long erwugh to be of service enough to be of service in calculating copy space. AU hare heen made long enough lo be of service in calcu-
12 21 6
SCH.T P°INT Casl°n No. 540 and Italic. Eight Point Modern Roman No. 64 Eight Point Roman No. 510 and Italic
I he body type faces shown on this and accom- AND ITALIC. The body type faces shown The body tvpe faces shown on this and ac-
panying pages are some pf the famous types on this and accompanying pages are companyingPpages are some of the famous
of printing history. They cover a wide range some of the famous types of printing types of printml history Thev cover a wide
of Sty es makmg easy the selection of the history. They cover a wide range o! range of^tyles^naking^easydi^^lecüon of
Theses haVlZn 7V/ P'eCe ° StyIe* makin? easy the selection of the the forrect body type for any piece ofprint-
The lines have been made long enough to be of correct body type for any piece of printing. ing. The lines have been made long enough to
5 14 10
rnTT. Caslon öldstyle No 471 and Eight Point Cloister Oldstyle aNd Italic. Eight Point Goudy Oldstyle and Italic
ic. e o y type faces shown on this and ac- The body type feces shown on this and accom- The body type faces shown on this and ac-
“™?“y!ng Pa8cs are some of femous types of panying pages are some of the femous types of companying pages are some of the famous
printing history. They cover a w.de range of styles printing history. They cover a wide range of types of printinl historv Thev cover aw!de
m ing easy the selection of the correct body type for styles making easy the selection of the correct range of styles Making easy the selection
ny P,ece of printing. The lines have been made body type for any piece of printing. The lines of the coSct body Tpe for any nieceo"
ng cnoug o e of service m calculating copy space. have been made long enough to be of service in printing. The lines havenen made long enough
9 13 15
TlfHK jOI,NT BFLMERRoman and Italic. Eight Point Garamond and Itauc. The Eight Point Scotch Roman and Italic
The body type faces shown on this and ac- body type feces shown on this and accom- The body type faces shown on this and ae-
companying pages are some of the famous panying pages are some of the femous types eompanjdng pages are some of the famous
types of printing history. They cover a wide of printing history. They cover a wide range types of PrinUnfhistorv?Thev cover aw°de
range of styles making easy the selechon of of styles making easy the selection of the range of styles making easy the selection of
ingCT/7ïi,tshl^!7n!je7a7e«(pT)nf' TTT ^ ^ for «“V Piece of printing. the correct body type for any piece of print-
g. u Lines have heen made long enough to The lines have been made long enough to be of ing. The lines have been made long mimgisMs—
8 4 16
Eight Point Century Schoolbook and Eight Point Century Catalogue and Eight Point Schoolbook Oldstyle
716 body type faces shown on Italic. The body type faces shown on this with Century Schoolbook Italic
oTth|nfamou°sIwr?fgD?toti^ehisto™e f^ acc°“Panying Pa8e,s. are so™e o{ the The body type faces shown on this and
They cov^ rv^e rangeTf ^lPs m^k' famous types of pnnting history. They cover accompanying pages are some of the
W1 ^ange «Styles mak- a wide range of styles makmg easy the selee- famous tvnes of orinting historv Thev
f y rhe selec.tlon the correct tion of the correct body typt for any piece coverawiderangfof s^lesmaktogeasv
y ype for any piece of pnnting. The of pnnting. The lines have been made long the selection of the correct body type for
v n 7 22 19
Tt1? ,C®NTUEY Oldstyle and Eight Point Bodoni and Italic. The body Eight Point Goudy Catalogue and Italic
and acCImpanJngypLesarfMme0onf ‘the faCC8 9hownf°t?t¥8 and accomPanying The body type faces shown on this and
famous typSrfndnfhistotv T^v rover pages .are some of the famous types of print- accompanying pages are some of the fa-
a '"g, hl9t0fy- They ,?.OVer, 3 Wlde rraage of mous tyPes of Printing history. They cover
OI maKing easy tne se- styles making easy the selection of the cor- a wide ranse of stvles makina rhe
lection of thecorrectbodytypeforanyp ece reet body type for any piece of printing. seletdon of the coSe« b“dvTvoefor »v
fp mtmg. The lines have been made long The lines have been made long enough to piece of printing. The lines hlsTbeen madl
20 18 3
|ALrC.PThér body"type tocetThown on Ldy typ^face^Thow^ontolsandtc:
of tht famouTtvptsof p^ftlng hlsto^ hfeto^They covt/t w°de i&ZZ *Tlf C°Ver f
They cover a wide range of styles making range of styles making easy the selection Tk J T" Ju8 t* select,°” of
easy the selection of the correct body of the correct body type for any piece of mTCe WT J V P °k °/ TT8'
type for any piece of vnntlna The lines 1 UV u y P ce OI lm«lh«ve been made long enough to te ol service
J F J p 1 ne unes pnnting. The lines have been made long in calcuUng copy space. Jlll the face, of each ,Ue have
11 23 2
Tifbody’^pe fechesrsh®wnhón thfe end"!^' ^ ^ The Point Berntard Bcoklet and Italic. The
panying pages are some of the Lol t body type faces shown on this and accom- hody type face, shown on this and accompany-
printing h^^Th.^rattlerang^ o&° ^pSTsto"6 They c°over rS '1 «T "tT“ °f ^ ^7’ 1 ^
“ *• •>-*•«snïïÊSisSï’ïSïS'SS Kr. br«‘r. ic
FIGURE 14
See Table IV, page 96. From the sample book of the American Type Founders Company.







a multitude of rolling or scrawly lines, but in all these cases no words. When we actually read the pieces of text, we do not “feel” these differences, because we are able to adapt ourselves to considerable obstructions, but the reading-speed is reduced, as the recorded times prove.
If we are to put down these findings in rules, we could say :
1. the x-height can be fairly small, provided the single characters are clear and comparatively wide and exaggerated projectors are absent;
2. the lower limit of thickness approaches the limit of typecutting possibilities, provided for perfect visual acuity of the reader, excellent impression and illumination ; as these are exceptional, we would ask for a thickness like that of Cloister Lightface, Garamond etc. in fig. 14;
3. the contrast must be small, but not absent; vertical stress should be avoided, Baskerville is exemplary ;
4. moderate width is advisable, both because the word-length is smaller and because the shape of the single characters is more typical; cf. Goudy and Stymie, where the round characters are so full, that the text gets a distinctly“ rolling appearance” (Updike).
c. German text of 50 words out of sample-book nr. 4 ; length of line 28 augustins (126 mm.), and 19 augustins (86 mm.). See Table V.
These texts were included because of the differences in size. Table V shows that the total time and the number of eye-pauses is steadily increasing with increasing size for the types 1—12 and 17—19. The “Schmale Antiqua”, a special narrow face for space-saving, makes an interesting exception : nrs. 15 and 16 were read in less time and with less eye-pauses than nr. 13 and 14. Apparently the lack of clearness, due to the narrowness, has been compensated by the larger sizes of the former only. The absolute values of this type cannot be fully compared with those of the other types, because the length of the lines was different.
The number of eye-pauses per cm. and per second in this table does not follow any rules, though the figures in column 10 are on the whole decreasing with increasing size.
d. English text of 88 words out of sample-book nr. 2. See Table VI.
• '
The composition of this text varied in length of line, size, width and boldness of the types. The Cheltenham-series was used.
Legibility, atmosphere-value and forms of printing types.



TABLE V.
Scores for text c. Differences in size of type.
ra cj M Ö5 PM OO
H V Ba P=t „ >-l S.o H p3
* 2 o ï' o £ n o o OD H
t» g £ * g . H . « «i 00 P * *
b „ q m * 5 a £ go p5® J H ◄©
ü w5 ^ ^ H co g <m ^ p ^ <® §
« o •< 2 H ^ ïï j ü faS
g pgfe m o 2 2 8 ö 3 5 3 ” H ° 3 c2 °S ë
3 3 ** as gg 2 2° 3s * „• 3
B -I S fc o O 2M O « fc
! 2 3456789 10 11
1 Nordische Antiqua 6 126 3 494 44 0 89 30 0.60 3.40
2 8 126 4 583 50 0.85 36 0.61 3.60
3 9 s. 126 5 651 55.5 0.85 37 0.57 3.30
4 lOs. 126 5 707 57 0 80 39 0.55 3.50
5 91. 126 5 707 57 0.80 40 0.57 3.90
6 101. 126 5 732 62 0.84 42 0.57 3.35
7 12 126 6 847 64 0.75 44 0.52 3.40
8 Garamond 6 126 3 451 43 0.95 32 0.71 3.70
9 8 126 4 566 47 0.83 37 0.65 3.90
10 9 126 4 610 56 0.91 47 0.77 4.20
11 10 126 5 688 58 0.84 48 0.69 4.10
12 12 126 6 800 65 0.81 48 0.60 3.65
13 Schmale Antiqua 8 86 4 404 47 1.16 39 0.96 4.15
14 10 s. 86 5 460 50 1.08 39 0.84 3.90
15 101. 86 5 478 44 0.90 35 0.73 3.95
16 12 86 7 622 42 0.67 32 .0.51 3.80
17 Cheltenham 8 126 3 409 43 1.02 33 0.80 3.80
18 9 126 3 467 46 0.98 32 0.68 3.45
19 10 126 3 496 48 0.96 33 0.66 3.40
A comparison between the texts in 14 pica lines with the
corresponding texts in 23 or 25 pica-lines shows a decrease in
total time and total number of eye-pauses of the latter for both subjects, cf. nr. 1—9, 2—8, 3—10, 4—11, 5—13, 6—12, 7 14. That this fact is entirely due to the length of line, and not to the shorter total length, is proved by the greater length of nr. 13 and 14. The time per cm. is shorter in each case, which means that the eye can get on more quickly on the longer line. The figures in the columns 10, 11, 15 and 16 do not follow any rules.
BC is the narrowest version, followed by N, W and B. In the different categories of point-size and length of line, N was the fastest except in two cases :
1. Of the numbers 5, 6 and 7, W was the fastest, which can be explained as follows. The short line-length hinders the 10 point size still more than the 8 point. Every effort to cram more into this small compass will make things worse. Hence BC, which is the blackest and least clear of the four, takes by far the



TABLE VI.
Scores for text d. Differences in width and boldness of type and in length of line. Intertype-Cheltonian ^
Normal (N), Wide (fV), Bold (B) and Bold Condensed (BC).
S “ § S S 5 Subject Ne. l Subject Ne. 2 5 g * «r . »»
“o tn g te — O rt O N Ohn
8g ai SS i. g* °-l Sb |1 s. “S s0. Ss 5| ga.* ga.g g5g' -|g'
g «5 g£ °i -s 3g s§ g; ms a- s§ gg j a- sas s*g §Sg §«g
3 Sj 5* g^ * ** *2 otk 2° Sr S? o51 g° ?Sg ?2g ?3«
S «a » M S g h 2 o 2 g 2 (. . g sg B 2 * S H° « h2 gMg«gH«aHp|
p w r; 2 Eh t» 5 g gr! aj £ y ho h !* a fc • w £ 5 p 2 p
* " °° a ^ 8 HH s* gg g g* fi" g* gg « 03 « 50 5s® «~®
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
\ Z9a 8 + 1 BC 10 633 82 62 1.22 10.2 1.32 150 84 2.37 7.5 1.78 15 13.5 15 12
2 (14) N 11 653 80 56 1.29 11.6 1.43 132 74 2.02 8.8 1.78 14 8 13 4 5
3 W 12 749 83 62 1.10 12.0 1.33 133 76 1.77 9.8 1.75 16 9.5 16 10
4 B 13 789 85 68 1.07 11.6 1.25 154 91 1.97 8.6 1.69 17 17 18 17.5
5 10 + 1 N 12 734 87 62 1.18 11.8 1.40 154 89 2.09 8.2 1.73 18 17 17 15 5
6 BC 13 799 90 70 1.12 11.4 1.20 178 99 2.22 8.0 1.79 19 19 19 19’
J_ w 14 855 76 64 0.92 13.3 1.18 151 102 1.76 8.3 1.48 12.5 15 14 17.5
8 97 8 + 2 N 6 627 65 49 1.03 12.8 1.32 125 74 +99 8+ +69 2 2 ï ï
® (23) BC 6 627 68 50 1.08 12.5 1.36 140 82 2.23 7.6 1.70 6 11 4 11
JV w 7 743 73 55 0.98 13.5 1.33 126 80 1.69 9.1 +57 10 3 9 7 5
11 B 8 776 66 55 0.70 14.1 +20 143 87 +84 8.9 +64 3.5 12 6 13
P 10 + 2 BC 8 789 69 56 0.87 14.0 +23 133 79 +68 93 +68 +5 9Ü 7 9~
13 105.5 10 + 2 N 7 744 66 54 0.90 13.7 +22 Hl 76 +75 «Ti +72 +5 ë+5 5 77
14 (2S) W 8 868 69 58 0.79 14.9 +18 123 81 +41 10.7 +52 7.5 1 11 6
15 *^7 10 B 7 942 73 55 0.71 17.1 +33 154 89 +63 ÏÖTi +73 lö +7 9 Ï77
12 BC 6 850 64 51 0.75 16.6 +25 131 68 +54 12.5 +92 ï 65 2 ~1~
N 6 859 67 50 0.78 17.1 +34 127 76 +47 1+3 +69 5 4 3 3
W 7 972 76 54 0.79 17.6 +40 129 76 +32 12.8 +70 12.5 5 9 4.5
_ B 8 1023 73 56 0.71 22.0 +30 150 86 +46 14.3 +74 10 13.5 12 14



largest amount of time, N ranks after it. N is also blacker than W, the most open of them all. Only the latter can cause the eye to develop sufficiënt speed. On the longer üne (13 and 14)
N is again the superior.
2. Of the number 16, 17, 18 and 19, BC is the fastest. It is narrower than the others, and its 12 point-size is sufficiënt to counterbalance the drawbacks of that.
e. The same text as d. Intertype-Ideal News. One subject.
The average total time of the sizes 5|, 6, 6|, 6f, 7, 1\, 8, 9 and 10, both solid and 1 point leaded in lines of 14 picas (59 mm.) against the average of the same sizes on long lines (5|, 6, 6J,
6| solid and one point leaded in lines of 21 picas (88.5 mm.);
7, 1 point leaded, 23 picas wide ; 1\ and 8, 2 point leaded, 23 picas wide ; 8 and 9 point 2 point leaded, 25 picas wide ; 10 point,
4 point leaded, 25 picas wide amounts to 82.2 against 74.0; the average number of eye-pauses 63.5 against 55.1. This confirms our findings on text d in respect of length of line.
Series V. The measurement of reading-speed on a mixed text.
The legibility of a text based upon the mixed use of normal, bold and italic types (e.g. dictionaries, catalogues, etc.) constitutes a problem of even greater importance than the legibility of straight “body matter”, for differences in legibility will have a more serious effect in a mixed composition, that prevents the reader from adjusting himself to extra-strain.
The methodological difficulty in the investigation of this side of the legibility-problem is due to the fact that not the speed of reading, but the speed of choosing is important. We are not concerned with the question how much time is required to read through e.g. a whole dictionary, but how much time we need to single out'one word among a thousand others. This circumstance makes it necessary to build the experimental situation upon the process of selection.
No earlier investigations on this subject have come to my knowledge. Baird and Lyon measured the time needed for looking up names in a telephone directory, but that is a different situation, as such names are usually arranged one below the other at the left side of columns in a bold type, hence the speed of looking-up is chiefly influenced by such factors as size of type and leading. In fact, to know these factors was the purpose ^ of both the authors. In a dictionary, on the other hand, the location of the required word or meaning is entirely unknown ;



the reader is not only forced to look along the left side of the column among words in bold type, but throughout the column among light, bold and italic types ; often he is uncertain how a certain idea is formulated. Accordingly he cannot always go through the columns with the visual image of the word in his mind, simply comparing, but his attitude has to be expectant until he meets a sentence that has a bearing on the vague complex of possibilities and probabilities, connected with the word in question ; then he has to verify the adequacy of the translation or definition.
The typography of dictionaries follows fixed rules in the notation of the sex, plural forms, terminations, definition, translation and practical application in common expressions etc. The selection is facilitated by the use of contrasting kinds of type, which enables the reader to disregard at once those words or sentences which he does not need.
This peculiar reading attitude makes it necessary to employ a kind of material for the investigation, which has the same characteristics, provoking the same attitude. But this very circumstance also makes strong individual differences between the readers almost unavoidable and that is the reason for the experimental difficulty of this problem, and, incidentally, for the failure of our own experiments.
In view of keeping as close as possible to the attitude and situation of the dictionary-reader, we decided to ask questions about subjects out of an explanatory wordbook and to measure the time taken to look up the answers.
Three kinds of text were selected, printed together as three separate columns on one sheet. The first consisted of one entry (viz. the word dead, in the Dutch language also meaning death), with all its meanings, ways of occurring in expressions etc. The second consisted of several entries, beginning with the same two characters (“ex-”), each entry being placed at the left of the column. The third consisted of several composite words, derived from the same basis word, arranged either as a new entry on the left, or taken up under one of these, the separate terminations (preceded by a hyphen) being scattered over the whole column. Each column presented a peculiar kind of difficulty to the reader.
The sheets were printed in four kinds of Monotype-faces :
A Series nr. 18 and 74 (Special Dictionary-type)
B Series nr. 327 and 360 (Times New Roman)
C Series nr. 262 and 275 (Gill Sans)
D Series nr. 110 and 194 (Plantin).



Because each question could only be put once to the same subject, whereas we needed the scores of all four types for each
• I 1
1,1 l •
u y
{ \ : \ ( i n—fi i
i MiUi a» !
i ! pi|~i :-= |
! ; m ABCDEFGHl/ <Z> j j
j j Hi.L ! |
^T—l'7 / / V-y'
' ] ' 1
SSV' ^ I r.Ti
! i_>u
question, only four subjects could give us one set of complete data about one question. The differences in difficulty of the



(totally 39) questions and the personal differences in speed of the subjects made the recorded times highly divergent.
In order to prevent location of the answers from memory the questions were put in such order, that the subjects did not read the same kind of type or the same column immediately after the other. Each question required a new exposition of the sheet; when the answer was found, the sheet was immediately withdrawn from the sight of the ss.
Chance factors appeared to have influenced the results to such a degree that they were entirely unreliable.
The apparatus ïtself yielded satisfactory results. It is sketched in fig. 16. Instead of the exposure apparatus A, another one was used, illustrated in fig. 15. It consisted of a vertical frontplate, painted white, with a horizontal rectangular slit in the centre, behind which the sheets were made visible. The illumination consisted of two vertical 75 watt shop-window tubes, fixed at the front behind curved vertical reflectors, also painted white at the outside. The slit was covered by a piece of ground glass, that turned from P, being kept in position by the electromagnet M ; with this arrangement the subject could see the text through the glass without being able to read it, thus perfect fixation, accomodation and adaptation being possible. When the experi-



menter pressed the Morse-key C, he started the Hipp chronoscope (B), at the same time interrupting the current of the magnet which released the groundglass screen. The latter feil into a clip R, while the text-sheet became visible. The subject started to look at once for the answer on the question put on to him beforehand; as soon as he had found the suitable word, he pressed the Morse-key F, thus stopping the chronoscope, and answered the question. Then the expenmenter tumed back the sheet to make it disappear, noted the time, changed the sheets, asked the new question and brought the relays D and E in the initial position, while the subject himself tumed the strip of glass before the slit.
In this experiment the nature of the object material proved again to be the chief obstacle. One can look for it in several directions, but one can hardly think of any solution that will do justice to this peculiar situation.
It is possible to design a material comparable to the ChapmanCook tests (short paragraphs with a checking method), e.g. sentences like they occur in intelligence tests, which must be completed by a word that can be selected out of several possibilities. These possibilities should be hidden in something like a section of a dictionary. But then there might be the chance, as was the case in our experiment, that the subject hits on the right word
by accident. .
One can ask the subjects to cross out all bold types or itahcs etc. in a page full of separate letters. But it is doubtful whether that would bring out the dilferences adequately, whereas the results obtained with reading of isolated letters cannot be generalized, as we pointed out before. _
It is possible to construct a kind of puzzle pictures, in which the reader must find the word in bold or itahcs etc. But then the most dominant type would be favoured in every category, whereas in a dictionary only one of the three or four kinds of type used should be dominant, and the others unobtrusive, though clear and contrasting.
By compiling data obtained in several ways we might still be able to learn something about this subject, which is as important for the typographer as it is difficult for the investigator.
For further consideration of this subject see Chap. 8, § 5»



CHAPTER 7.
Series VI. The judgment of the subjects on legibility.
However important the objective legibility may be, the publisher and advertiser will have to reckon in the first place with the subjective judgment of the reading public. For the latter will turn the scale if there should happen to be a controversy between objective and subjective legibility.
In respect of the style of type face there is little or no disagreement between these two, but our knowledge of the optimal size of type, length of line and leading does not confirm the general preference of the public. Neither Pyke nor Hovde did find a positive correlation. The latter writes1): “The tendencies of the readers is to accept the larger type faces and leading as easiest to read, though this fact does not agree with the actual number
of words read Undoubtedly the columns in which the
most words were read on the average, show less readers’ choices than they actually deserve, because the readers thought they did not cover as much ground.”
All other authors on this subject had the same experience : the public (to quote Pyke) dislikes ‘heaviness’, ‘paleness’, ‘crampedness’, ‘dazzle’ and condensation ; it prefers a simple, clear, open, large type. Accordingly, sans-serifs get high ranks together with the good old-style forms, though these two kinds are by no means legible in the same way, as we pointed out before.
Webster and Tinker (in a study on type faces mentioned before) give the following description — which is incorrect as far as booktypes (context) are concemed — : “Analysis has shown that the legibility of letters is increased when the size of letters is increased, when the lines in the letter are widened, when the area of white space around or within the outline of the letter is increased, when the contrast of shading and hairlines is lessened, and when the outline of the letter is made simpler.” A comparison with Table VII shows that a type can comply with these demands without finding favour in the eyes of the public.
The 71 resp. 68 subjects who determinated the atmospherevalue of the book- resp. display types in our series VII were canvassed for their opinions on the clearness of the types (see fig. ibid.); they could choose out of three terms : clear, moderately clear, unclear. The results are shown in Table VII; the figures behind the names of the types are those used in series VII.
1) Howard T. Hovde, The relative Effects of Size of Type, Leading and Context. J. Appl. Ps. Vol. XHI, 1929, 601.



TABLE VII.
Rank-order of legibility according to the opinion of the subjects. See Part II, Series VII.
M pj 5 ï
£ . oi * % è *5 g ^
« « 3 „• g g * 3 rf S
5 pjWoa* «Ss16
pj Name of Sebies z Q S P (4 Name of Series Z o S P
1 Saeculum 5 | 62 8 1 1 Columbia 29 I 66 2
2 Bembo 10 61 9 1 2 Mediaeval 17 61 7 —
3 Mediaeval 8 | 52 8 1 3 de Vinne 26 I 61 6 1
4 Bodoni 3 43 21 7 4 Light Gothic Exp. 30 43 23 2
5 Beton 7 43 19 9 5 Arpke 28 I 38 28 2
6 Thannhauser 2 36 30 5 6 Narciss 18 I 28 38 2
7 Grasset 13 28 34 9 7 Locarno 24 29 35 4
8 Romaansch 12 30 26 15 8 Omega 20 | 23 31 14
9 Orpheus 6 25 36 10 9 Metropolis 23 18 37 13
10 Modern Wide 9 17 36 18 10 Kontrast 16 12 34 22
11 Lutetia 4 11 46 14 11 Excelsior 19 13 32 23
12 Bernhard Roman 1 17 29 25 12 Shadow Nobel 22 ! 13 31 24
13 Locarno 11 6 19 46 13 Trafton Script 27 Jf 30 26
14 Bernhard Cursive 14 11 28 29
15 Eckmann 21 9 29 30
16 Heavy Gothic Cond. 25 4 20 44
17 Gallia 15 1 8 59
The rank-order of legibility of the 8 types used by Pyke, computed from the judgments of 60 subjects, was the following : 1. Monotype-Series 2 (Old Style) ; 2. Series 7 (Modern Extended) ; 3. Series 101 (Imprint) ; 4. Caslon-Series 23 (Modem) ; 5. Stephenson & Blake-Series 10 (Lining Grotesque) ; 6. MonotypeSeries 161 (Old Style Antique) ; 7. Series 17 (Cushing) ; 8. Series 39 (Modern Condensed).
The 42 subjects used in our series V were also asked to give their opinion on the four kinds of type, judging after their legibility and suitability for dictionaries. This question caused the opinions to be more reasoned, for the subjects compared the contrast, the conspicuousness etc. accurately1). The results follow here :
TABLE VIII.
Rank-order of appropriateness for dictionaries according to the opinion of the subjects.
Kind of Type lst pos. 2nd pos. 3d pos. 4th pos. Final order
A Special Dict. Type 28 6 4 4 1
B Times New Roman — 14 14 14 4
C Gill Sans 11 16 6 9
D Plantin 4 11 21 6 3
!) See also Chapter 8, § 5.



CHAPTER 8.
Summary and application of the results.
§ 1. The studies on typographical factors influencing speed
OF READING BY TlNKER AND PATERSON.
In the following pages our present knowledge of the influence of printing types on the speed of reading will be summarized. We owe much of this knowledge to the efforts of Tinker and Paterson.
They were the first investigators, so it seemed, to find an entirely reliable method for the measurement of legibility of context; but their critics have still reason to feel not entirely satisfied, as will be pointed out hereafter. The Chapman-Cook Speed of Reading Texts, that were used in nearly all their studies, consist of sections of 30 words out of Thorndike’s Teachers Wordbook and therefore all equally well known. In the second part of these sections one word is concealed, which does not make sense with the first part; the reader has to put a checking mark under that word. Consequently, he is forced to read all the sections entirely and with the same amount of attention. The sections are grouped in paragraphs of 5 sections, 6 paragraphs constituting a test form. So little time is allowed to the readers, that they cannot finish the whole test form, i.e. lf minute for adults and 2i min. for the elementary schoolgrades. An example of the sections :
“There was a fire last night, and five houses were burned to the ground. It all happened because someone was careless and threw a nail into the waste-paper basket.”
Tinker and Paterson (further abbreviated : tp) made use of these test forms in the following way. They printed one of the forms in a Standard way of composition, while varying in the other form the feature they wanted to know about. In the difference between the mean scores (number of checked paragraphs) of the Standard and the varied form, a measure for the relative legibility is given.
Determining the equivalence of the Standard form (A) and the varied form (B), which is administered after the former, tp found that the efïects of practice nullify the greater difficulty of form B.
Nevertheless, as each new study of tp showed lower scores for form B, whatever the varied feature was, doubts were expressed as to the reliability of the method, and it was suggested that “mental set” might be responsible : . . .. “the “set” developed in reading



the Standard form A arrangement is disturbed when subjects are required to shift to a changed typography in form B.
In a special study, called “the röle of set in typographical studies”, tp tried to meet this objection. In a previous study concerning the differences in legibility due to the style of type face, they compared 9 different type faces with the Standard form A in Scotch Roman. All the scores of form B were lower, but only those of American Typewriter and Cloister Black were statistically significant. Now Cloister Black, a very heavy, intricate Old English type for decorative purposes, is so much less legible than Scotch Roman, that if it had not produced these results, the method of investigation would have been totally valueless. This very face was used by tp to prove the absence of “mental set” : in the study mentioned above they printed form A in Cloister Black, and form B once in C.B. too and once in Scotch Roman. The scores show again an inferiority of Cloister Black.
Evidently, this is no proof. The influence of “set”, if any, will make no greater difference than a few per cents, and it will be totally superseded by such large differences as those existing between Scotch and Cloister Black. A really conclusive proof would have been delivered if at least three of the types that appeared to be about equally legible as Scotch in the first study, had been used instead of Cloister Black, e.g. Garamont, Old Style, Caslon.
Until the proof of the contrary is delivered with such types, we have still to suppose that the small differences between the scores of types, that are about the same as those of the Standard form A, and that would be subject to chance factors in other kinds of experiments, are govemed by “mental set in this method.
The results obtained by Stanton and Burtt1) seemed also to give reason for accepting the figures of tp with some reserve.
As the equivalence of the forms A and B is fundamentally important, the mean difference in the scores of these forms (when printed in the same way) for one test group (control group) is also an important figure. It was used as a correction on the differences obtained by the other test groups by Stanton and Burtt, and also by tp in their later studies. In the first studies equivalence of the test forms was assumed, and therefore no correction was made.
The difference in question tumed out to be rather great in the results of Stanton and Burtt ; they write : “the magnitude
i) The influence of surface and tint of paper on the speed of reading. F. N. Stanton and H. E. Burtt. J. Appl. Ps. XIX, 1935, 683.



of the correction in the present experiment is rather disquieting. .... It seems that chance errors may play a serious part in a test as short as 1| min.”
The authors made a further control on 149 additional students, who constituted three special control groups of resp. 73, 57 and 19 subjects. These groups showed a difference between the forms A and B respectively of 0.78, 0.16 and 1.07, concerning which s. and b. remark : “the lack of uniformity in these special test groups is noticeable. The cause of such divergence is difficult to isolate and it seems more than a mere sampling error.” It must be noted, however, that tp used only groups of 80 or 100 subjects, which reduces the amount of chance errors.
Again a special study was devoted by tp to prove the reliability of the results1). In an experiment concerning the effect of sampling errors on the equivalence of the test forms, the scores of 17 test groups of 80 students each were compared. The differences between the means of the forms A and B ranged from — 0.45 to + 0.50, or maximally one half a paragraph, which is "... . well within the limits of variation to be expected from the operation of a large number of uncorrelated chance factors. .... This being so it becomes possible to introducé typographical changes in Form B as compared with Form A, and obtain a direct measure of the effect of the typographical change on speed of reading so far as sampling errors are concerned.”
The arrangement of the test forms in paragraphs of 5 sections, as used by Stanton and Burtt, was introduced in the later studies only. The equivalence of this arrangement with the former, with separate sections, was determined too, and it proved to be satisfactory.
Special measures were taken to ensure uniform testing procedure. The reliability coefficients, which should be at least + 0.40, were found to vary from + 0.75 to + 0.90. A special study 2) showed that the two methods of administering the reading tests, time-limit (counting amount of paragraphs read) and work-limit (counting total time taken) could be used interchangeably.
Yet another series of tests was devised to check up the results with longer periods of reading.
Instead of the comparatively short and easy Chapman-Cook tests the authors employed E. B. Greene’s Michigan Speed of Reading Test, which consists of 100 sections of 30 words after
9 Studies etc. XIII. Methodological considerations. J. Appl. Ps. Vol. XX, 1936, 132.
!) D. G. Paterson and M. A. Tinker, Time-Limit vs. Work-Limit Methods. Am. J. Ps. 1930. 42. 102.



the Chapman-Cook pattem, but more difficult, so that a timelimit of 10 minutes is given to college students. The Standard form A was in Scotch Roman, the varied form B in Cloister Black. Three time-limits were used : lf min. (as in all previous studies), 5f min. and 10 min. The differences in percents between the (corrected) means of Scotch and Cloister Black appeared to increase with increasing time-Hmit. The authors conclude . This objective evidence leads us to emphasize the significance of whatever differences are disclosed in our various experiments.
These results are highly important. Still, they have to be confirmed by similar experiments with other types and arrangements. For here again we are not justified to generalize these findings obtained with Cloister Black as compared type, the latter being exceptionally illegible. It requires an uncommon amount of attention for detail and so produces at once a statistically significant reduction of the reading rate. See also under “Style of Type Face.”
The authors assumed that whenever in their results the ratio D : P.E. diff. was equal to or greater than 4.00 equivalence between Forms A and B had been upset by some uncontrolled factor or factors in the experimental situation. Among the results of 24 special test groups (totally 2272 subjects) this value appeared to exceed 4.00 in seven instances. Hence the lack of equivalence found in the control groups of the various tp studies « . . .. is not due to mere sampling errors, but is due to some lack of uniformity in testing conditions. Thus the utilization of control groups in each of the experiments is shown to have been essential.” It appeared further that the applied correction on the results (by means of the difference between the control groups) was sufficiënt to ensure reliability.
Another objection had been raised against the considerable variations in the mean scores on form A obtained by the different sub-groups. tp believe these are due merely to sampling errors. “Evidence exists, however, showing that a given difference between two typographical arrangements will appear regardless of these minor variations in the level of mean performance of the subgroups. Even striking differences in level of mean performance will not change the results.”
Thus tp have surrounded their experiments by almost every possible precaution against statistical and methodical errors. It seems that only the choice of the compared arrangements of Form B and the composition of the test-groups in respect of reading-skill of the subjects can make the results questionable,



if once any influence of “mental set” is eliminated and the influence of the time-limit determined. Some final critical remarks are made in the next paragraph.
§ 2. Style of type face.
In the chapter conceming the theory of reading of contest it was pointed out that the legibility of isolated letters follows other rules than the legibility of cohesive texts.
Hence the judgments of previous investigators, based on studies with isolated letters and generalized for a whole style of type in every kind of reading, are acceptable only with some restrictions.
The rank-order of legibility according to Roethlein, with News Gothic and Bullfinch at the top, and Caslon O.S. together with Cushing at the end, can have no importance for cohesive reading, even if we do not take into consideration Pyke’s justified criticism on her method.
Likewise the scores obtained by Burtt and Basch on Bodoni, Baskerville and Cheltenham do not agree with the evidence conceming legibility in strict sense, though they are valuable for the recognizability of isolated letters. These authors measured the distance at which the letters could be thrown out of focus on the “focal variator” (expressed in units of 1/16 inch on the bench of this apparatus) without becoming unrecognizable. The Jj K, L, Q, U, X, Y, Z were not used. The average scores follow here :
Cheltenham Upper-case 16.9 Baskerville „ „ 14.9
Bodoni „ „ 13.2
Cheltenham Lower-case 13.0 Baskerville „ „ 11.9
Bodoni „ „ 11.1
The higher score of Cheltenham is due to its boldness, which is, on the other hand, a drawback in cohesive reading ; the lower score of Bodoni is due to the really less legible forms. Hence we would put Baskerville for cohesive reading in the first place, Cheltenham in the second and Bodoni again in the third.
In an unpublished report on the “Futura” sans serif type Moede found a superiority of the latter over old style types. He employed an apparatus for shaking the object to aggravate the reading-conditions, which required spelling instead of reading. As pointed out before, the sans serif is very suitable for that way of reading, but results obtained in that way should not be generalized.



The “relative legibility of linotyped and typewritten matenal was measured by Greene. He used his own Michigan Speed of Reading Tests (10 min.), printed in Linotype Ionic and American Typewriter, both 10 pt., 61 mm. line and 7 pt., 51 mm. line.” The linotyped samples were read a little faster than the typewritten on 7 pt. material, but the reverse was true for the 10 pt material.” None of the differences noted, however, were
significant,” so one must conclude that the tendencies noted
are so small as to be easily reversed by chance factors in the test situation, such as differences in fatigue, practise, motivations, and distractions of the groups compared. The mam conclusion, therefore, is that the two fonts are nearly equal in legibility lor these students on these particular ten minute tests. A conclusive experiment of this sort would have to include measures of an individual’s energy expenditure. Such measures are at present
not available for this kind of work. .
The differences observed by Greene are at present inexphcable to me • the further conclusions are supported by those of tp. The American Typewriter type of the tp study, by the way, was considerably less heavy than that of the Greene study.
From the results of Pyke’s experiments no definite conclusions can be drawn concerning the legibility of type faces, though they are highly enhghtening in other respects.
Paterson and Tinker compared Scotch Roman, Garamont, Antique, Bodoni, Old Style, Caslon O.S., Kabel Lite, Cheltenham, American Typewriter and Cloister Black. Only the last two faces produced significant differences. _ «. ,
If it were possible to add to the excellent series of studies by Paterson and Tinker another control experiment, m which a better choice of types x) is used for both a short and a long period, e.g. l3/4 min. and at least 20 minute, we wouid have at last a reliable fundament of knowledge of the objective leg^ihty as far as the general public is concemed. If the increased difference of Cloister Black is confirmed, then the hypothesis is correct that there is only an initial compensation by extra-effort, opposed to the suggestion of several authors that the reader adapts himself to a more difficult type after an initial reduction in the reading rate.
lust now we spoke of the legibility for the Zeneral public, because we still do not know about the objective legibihty for the individual. The tp experiments have been built upon the silent assumption that legibihty factors are supramdividual,
nee Scotch Bodoni, Bodoni Book, Baskerville, Cheltenham laeai wews (or^chooS b.S etc.); Stymie (or Memphis etc.), Intertype-Med.aeval, and crimp nrdinarv modern.



whereas it has been our aim to demonstrate that the abilities and the condition of the reader play an important part; his “powers of comprehension” (and, of course, the condition of his eyes) determine the amount of letters he is able to apperceive in one eye-pause, and accordingly the kind of type that is best legible to him. The rapid reader needs a small, thin, close and narrow type (prototype : Baskerville, Fraktur in general), the slow reader needs a large, thick, wide type (prototype : Schoolbook O.S., Bookman, Ionic). It will be clear that the proportional composition of the best groups in experiments like tp’s is not irrelevant in this respect; neither is a very “fast” type the most suitable for the Chapman-Cook tests.
A difficult text in a “fast” type, read by slow readers will produce considerably lower scores than an easy text in a “fast” type read by fast readers. If a test group contains both slow and fast readers, and if the material consists of both slow and fast types, the scores will stand on the same level. These differences in reading-skill are not to be neglected. tp found a difference of about 40 words in 13/4 min. between freshmen and sophomores — who are both on a certain level of education.
It would be a large step forward if we had reliable and detailed results with a few subjects next to the already available material of tp with large groups. Only in that case every risk is avoided of accepting conclusions “statistically plausible but psychologically false” (Pyke).
The equal scores of the common types in the tp study induced Tinker and Webster (in order to get more comparable data) “ .... to check the Paterson and Tinker study, which used speed of reading as the measure of legibility, by employing the distance at which the material could be read correctly as the criterion of legibility. The identical type faces, size of type, paper and ink employed by tp were used in (this) study.”
55 subjects were used : 10 as a control group and 5 for each test group (9 combinations of series A in Scotch Roman with series B in another type). To quote the authors : “The direction of the differences between the average scores of individuals for Scotch Roman and the other type faces used was found to be very consistent with the exception of Kabel Lite and Cloister Black, where two subjects showed differences in one direction and three in the other.” (Roethlein and Pyke also found a low correlation between the subjects specially for the less legible types).
“The rank order for legibility, as determined by speed of reading in the study of tp, shows little relation to the rank order in this
Legibility, atmosphere-value and forms of printing types. 8



experiment. The correlation between the two series of ranks is — 0.07 ± 0.22. The difference is even greater than these figures indicate. In the tp experiment the difference between the legibility of the Scotch Roman and 7 of the other type faces used were too slight to be of statistical significance. In our study 7 cases showed the variations of type face to be significantly more legible than the Scotch Roman.”
The opposite results obtained with “American Typewriter” induced the authors to draw the conclusion that the factors, which make a type legible in this study, produce the opposite result for cohesive reading, the total word-form being more important in the latter case.
This conclusion agrees entirely with our own theory, based on other facts and considerations.
Webster and Tinker further conclude that the two criteria of legibility (speed versus distance) cannot be used interchangeably. Hence the results of the original tp study remain vaüd : “For rapid reading under ordinary conditions, it may be accepted that the type faces in common use are about equally good.”
On the score of our own results (theoretical and practical) we cannot accept this conclusion in this way, but would rather formulate it as follows :
For reading connected material under ordinary conditions during comparatively short spells the type faces in common use will be read at about equal speed, a little extra-effort being able to neu-
tralize the difference. n
A distinction has to be made between being “equally legible and “read at equal speed”, for the difference in real legibility will manifest itself in the long run; then (ceteris paribus) the type read in the shortest time will be of such design that it is “flowing and continuous in words”, to use a term of the British Association Committee ; in short, it will be conform the rules described on p. 97.
For unconnected material, words that are not familiar to the reader, etc. (dictionaries, time-tables, telephone-directories, addressbooks, catalogues), where single words or even single letters are read, the single letters may be more obtrusive and emphasized. For the extreme cases, that come near to spelling, the sans serif is advisable, on the condition, however, that the round characters are made süghtly oval and not so full as e.g. Futura.
There are numerous circumstances that promote isolation of the single letters and consequently hinder the formation of wordforms, e.g. :



1. too much letter space,
2. inadequately large size,
3. too much or too little contrast,
4. too much or too little thickness,
5. difficult single forms (queer shapes, excessive width or narrowness, too full forms, uniform shapes).
Fraktur versus Antiqua.
Concerning the question of the relative legibility of Fraktur versus Roman (Antiqua), all that has been written in the foregoing pages can give but one answer : for the reader who is conversant with its basic forms, Fraktur is as legible as Antiqua. Probably even more, if only the text is in the German language and with logical and well-connected contents. If, however, these conditions are not fulfilled, i.e. if the reader is not practised in this kind of type, if the text is in another language (with another composition and length of words), and if the text requires attention for parts of the words, Fraktur will be decidedly inferior to Antiqua. The true “ Gotisch” or “ Textur” is less legible under all circumstances.
The characters of the Fraktur alphabet are meant for use in word-form. The lower case characters are constructed in narrow rectangles, which makes them extremely close-fitting, but within this scheme they possess highly characteristical individual shapes. In other words : they constitute very strong structural wholes {Gestalten), but all these Gestalten possess one and the same feature, viz. the rectangular basic form, which makes it very easy to build up a new complex with them. If once the characters have lost their individual existence and act as a part of a strong other complex, they are very difficult to isolate again. Standing absolutely alone, the shapes in themselves are on the whole more characteristical than those of Antiqua, though some of them are decidedly less ; the narrowness is often an obstruction too.
The capitals of Fraktur are inferior, being subject to dangerous deviations of form because of their intricacy ; they do not possess the advantages of tight fitting combined with individuality like the lower case. Considered as linear forms, the capitals and lower case of Fraktur are infinitely richer than Antiqua. Schwabacher stands in every respect between Fraktur and Antiqua.
The Roman alphabet is not only dissimilar to Fraktur, it is essentially different. Overstating a little, we may say that the Roman alphabet consists of essentially dissimilar characters that are made more suitable for a connection into words, whereas



:he Fraktur alphabet consists of essentially similar characters hat are made more individual.
This fundamental difference cannot be entirely nullified, to vhatever treatment they are subjected, and therefore the specific jsefulness of these two styles cannot be the same.
Apart from ethnographical limits, the field of optimal usefulness 3f Fraktur is more confined than that of Antiqua *), because t covers only texts composed of those words, with which the reader is fully conversant and which are to a high degree logically :onnected, in short, texts that do not present any difficulties, ïither by their meaning, or visually (insufficiënt illumination, small size etc.). These conditions require a very simple choice af words, without any uncommon words, names, foreign expressions etc. or else a high level of education.
When, however, the difficulties of the text are forcing the reader to focus his attention on parts of words, Antiqua is easier, because it leaves the single units relatively more independent. A word in Fraktur can almost only be seen as such, its proper structure is superseded by the total impression. This strong assimilative effect explains why Fraktur, notwithstanding its typical single forms, fails (against Antiqua) in isolated short words : if the word (and hence the visual word-complex) is unfamiliar to the reader, the single letters are difficult to reconstruct. Moreover, but that is no question of recognition, such a word will lack emphasis compared with equally thick Roman characters. It has no hitting force.
Finally some remarks should be devoted to two German studies on this subject.
In the first, by Kirschmann, the distance from the fixationpoint was measured at which the reader first recognized a single character that was moved from the periphery of the visual field towards the area of direct vision. Kirschmann found a considerably higher distance for Fraktur than for Antiqua. We can accept his results entirely, save for his conclusion that Fraktur is superior in legibility to Antiqua in general.
Lobsien (the second study) counted the total time and the number of eye-pauses (in direct observation by means of a mirror) of his subjects, lower-grade school-children, who were tired after a whole day of lessons 2). He had found a superiority of Fraktur, but these results are not so reliable, for several reasons :
!) Hence it follows that in practice Fraktur is used in cases where Antiqua would be preferable, as far as legibility is concerned.
2) The fatigue factor was introduced by Lobsien in order to prevent extra-ettort on less legible types.



1. The method of recording does not guarantee that the smallest shifts of the eye were correctly counted.
2. The lines of the texts in Fraktur used by this author were longer than those in Antiqua. For the same amount of words more lines of the latter kind were necessary. Now we know that, ceteris paribus, lines of slightly different length will take the same time and the same number of eye-pauses ; in other words : it is far more economical to add words to existing lines than to compose a separate new line out of them.
This circumstance may have influenced the results.
3. Lobsien does not give the mean deviations of his average scores. They can be about as high as the differences shown here :
Average total time in 0.2 sec. for Gewöhnliche Fraktur, Schwabacher and Senatsfraktur on sense material 19.2, on nonsense material 43.3 ; for Gewöhnliche Antiqua and Romanische Antiqua 20.5 resp. 45.4.
Therefore we must make some reserve.
4. Tired children are no ideal subjects for experiments in reading, especially when nonsense material is included. Lobsien asserts that they were equally conversant with both kinds of types.
§ 3. Size of Type and Length of Line.
All authors up till now are holding the view that the optimal size of printing types is about 9 or 10 point, i.e. at least 1 y2 mm. x-height, preferably a little more. There is less agreement on the amount of retardation caused by sizes that are smaller or larger than this optimal value. Judd asserts that the rate of reading and the span of recognition (he recorded eye-movements) are only slightly altered by doubling or halving the body size (11 pt. compared with 5% and 22 pt.). Gilliland writes: “The reading of the average adult is not greatly affected by changes in size of type between the limits of 36 point and 6 point type. Above 36 point type or below 6 point type the rate of reading begins to decrease for a majority of the subjects.” Neither did Greene (with typewritten material) find “a reliable difference in the speed and accuracy of reading four sizes of type, 7, 10, 12 and 14 point.” The results of Gilliland, however, have been questioned by tp on various considerations, and not without reason. Beside other objections they deny the possibility of applying his results to typographical practice, because the photographical enlargings and reductions used by Gilliland are proportional, whereas an increase in point size of type does not result in a proportional increase of the height and width of letters



and spaces. Nevertheless ït remains very mteresung mai itr themselves did find a decrease of only 0.45 words per second, when their subjects read a photographical reduction^of 50 % of a newspaper text. The study in question, devoted to “Reductions in Size of Newspaper Print” in behalf of library filing, measured the decrease in reading rate due to a reduction down to resp. 80, 50 and 30 % of the Standard Form A in Linotype Ionic, 63/4 on 7 points, 12y2 pica (53 mm.) line length. The differences in number of words read per second amounted to resp. 0.05, 0.45 and 3.16.
The comparatively slight difference of the 30 and 50 % reduction (the actual point size of the latter will have been about 3 points !) does not fit quite well within the conclusions drawn by tp in their study nr. II (Size of Type). The values found there are following :
Differences between Form A (Scotch, 10 pt., 19 pica) and Form B,
6 point: 0.31 words/sec., retardation : 6.2 %
8 point: 0.27 „ „ , » 5.2 %
12 point: 0.30 „ ,, , » 5.8 %
14 point: 0.34 „ „ , „ 6.9 %
The low score of the 6 point size is confirmed by Study V (Simultaneous Variation of Type Size and Line Length), in which the 10 point. size of the Standard arrangement (Scotch Roman, 19 pica or 80 mm. long), was enlarged and reduced with a proportional variation of the line length.
The results are shown here :
6 point, 68 mm. or 16 picas, retardation : 0.32 words/sec., 7.4 %
8 „ , 72 mm. or 17 „ , » : 0.07 „ „ , 1.6 /o
12 „ , 97 mm. or 23 „ , „ • 0.19 „ » > 4.1 /o
14 „ ,115 mm. or27 „ , „ :0.20 „ » »4.4^
Evidently, the proportional reduction of line length makes the 8 point size equally legible as 10 point, but 6 point remains inferior. It is interesting to compare the mean scores of the 6 point in this study (15.14) and of the 8 point (15.59) with those of the reductions of newspaper print:
Test Group I, Form B, 7 point, 12* pica : 14.73 II „ , reduced to 80 % : 15.20
;; m, „ , .. „ 50%. 13.71.
These figures point to a proportional relation between the 6 and the 8 point sizes, which is more complicated than the conclusions of tp suggest.



Whatever may be the case, it is certain that each x-height has its own length of line (with a tolerance of a few picas or augustins) in which it makes optimal performance possible. Our series V, text e, showed marked tendencies towards lower scores for large types on inadequately short lines. On the other hand, the 97 mm. yielded the best results (together with 127 mm. for the 10 and 12 point types), though it lowered the scores in the tp study.
The psychological basis of the apparent relation between type size and line length has as yet not become entirely clear. Paterson and Tinker venture a somewhat surprising explanation by referring to “vertical cues favoring premonitions of meaning,” (originally an idea of Dearborn). It seems highly improbable that such vertical cues (i.e. advance recognitions of words in the next line) should have any other but a disturbing influence. For, according to this theory, the vertical cues would have to operate in the following way : the cues in the line to be read next would effect a certain selection (Bahnung) among the various possibilities of continuation of the logical sense, whereas the cues in the line just fïnished would serve as a confirmation of its being finished. This situation will occur only under the most favourable circumstances in respect of the logical connection of the sentence and the “powers of comprehension” on the part of the reader. In 95 per cent of the cases, however, the vertical cues will cause confusion and this fact is the sole reason for the existence of leading. It is true that cues favouring premonitions of meaning are present, but only in horizontal direction. We may illustrate this by the following sketch :
A is the fixation-point, the triangle represents the field of attention, which is directed to the right. Cues on the left, therefore, do not exist. What we need is a premonition of the words on the same line as A, i.e. part C, but not B on the last line, or D on the next. If B and D are close to C we will get premonitions of irrelevant meanings mixed with the only important ones, viz. of the words following immediately in the sentence.



B and D can be separated from C by employing a larger point size of type, a smaller face on the same size, or leading.
A different explanation of the relation between type size and line length is suggested by the viewpoints that guided us in the previous chapters.
The beginning and the end of a sentence are the least important for understanding. The most rapid understanding (and therefore the greatest reading speed) is made possible by that length of line, that allows the largest part of a sentence to be read without interruptions by shifts of the eye to the next line. The most important part of a sentence of, let us say, 190 mm. will have more chances of being read without interruptions on a page of 80 mm. line length than on a page of 50 mm. lines. A line has its highest level of consciousness also in the centre. Therefore a short sentence will have more chances of falling beyond the focus of attention on a long line than on a short one. The long sentence with long words, as usual in scientific language, (in turn more frequent e.g. in the German than in English language) and which is meant for intelligent (and consequently rapid) readers, should have a long line length ; the short sentence with short words, as usual in texts meant for the average public (e.g. newspapers), should have a short line length. The rhythm of the sentence and the rhythm of the line should be coinciding as much as possible.
The following rules can be given :
1. Unskilled readers (children, people of moderate intelhgence or culture), who cannot make use of premonitions of meaning, should have large, wide and thick types (prototype : Schoolbook O.S.), short lines and short, simple sentences.
2. Skilled readers (adults of high intelligence), who can make a profitable use of horizontal cues, should have small, narrow and thin types (prototype : Baskerville) and long lines. They can bear long sentences.
3. Small types lessen the difficulty of long logical connections, but they are a real hindrance when slow, word after word-reading has to be done x).
4. Large types are necessary in the latter case, but they obstruct the easy conception in the former case.
i) xhe “Galgenlieder” of the German poet Christian Morgenstern (Bruno Cassirer Verlag, Berlin) offer a striking illustration of this fact. The effect of these splendid humoristical poems is based on sudden changes of meaning, logical jumps, unexpected images etc. They are printed in an old-fashioned, very simple and extra-bold type. A good deal of the effect would have gone lost when a ‘ fast type, giving “premonitions of meaning” had been used instead.



The Chapman-Cook tests utilized by tp consisted of very short, simple sentences ; they were properly meant for school-children. It is possible that this fact, perhaps coupled with the relatively too high intelligence of the subjects, has influenced the results in respect of the style and size of type and length of line. The low scores of the large lengths (22—27 picas) would then be explained, at any> rate.
§ 4. SPACE BETWEEN LINES.
The distance between the non-projector letters of two lines increases in absolute sense with increasing size, decreasing x-height (longer projectors) and insertion of leading, in a relative sense with the last two only.
The purpose of interlinear space is to prevent the eye to wander off the line or to include parts of other lines into the observation of the focussed line. An increase in size of type will not be sufficiënt to produce the necessary separation, because the focus of attention is extended too ; when the projectors touch each other the lines will be still mixed up. Only in the larger sizes (about 11—12 point), where the focussed area is covered by the x-height for a great part, a visual contact between the lines is less likely to occur.
A smaller x-height on the same body will be a help as long as it remains clear itself, in other words : as long as the x-height is not reduced below the limit of easy recognizability.
Leading separates the lines in any case and it facilitates reading in most cases, according to nearly all the authors who studied the subject.
The usefulness of the 1 point lead has been doubted. Baird, who measured the speed and accuracy with which telephone directories can be consulted, found an increased efficiency, but these results were not confirmed by Lyon in a similar investigation, nor by Hovde. The most important study on this subject, again by tp, did not reveal an increase in the rate of reading for the
1 point lead either, but it confirmed the favourable effect of the
2 point lead (7.5 per cent faster) and the 4 point lead (5 per cent faster) found by previous authors. Bentley’s results point in the same direction. More experiments should be conducted (with different sizes and styles of type) before we can take it for granted that a 1 point lead does not promote objective legibility. The reader himself always prefers leaded texts.
The amount of leading necessary to achieve actual differences in legibility depends on the size and style of type and length of



line. A 6 point text will need some leading rather than a 12 pomt text, but a 2 point lead will be sufficiënt. Extra-effort, however, during short spells of reading (in experiments) probably reduces the improvement, compared with a solid 6 point text, below the limit of measurability.
A 12 point text requires a larger amount of interlineage before it is measurably more legible than a solid one.
A type with a relatively large face, and/or poor word-forms (owing to the shape or the width of the characters) requires considerably more interlineage than a type with a small face and good word-forms (see also the next paragraph).
The tendency of the lines to get mixed up immediately beyond the focussed area will have less serious effects on short lines than on long ones. A text printed in short lines contains more lineends ; hence it is read with more fixations during which all the text is falling in the focussed area, in other words : with less chances of mixing up the lines. Consequently leading will be more necessary and comparatively more effective on long lines. Greene, in experiments with typewritten material, actually found “a slight tendency for the leaded samples to be read more rapidly than the solid, and this tendency is more noticeable in the longer line than the shorter. The mean saving was approximatingly 3.5 per cent during 10 minute periods.”
§ 5. Legibility of Small Print.
The rules described in the foregoing pages can be clearly illustrated by applying them to the problem of legibility of small print.
The surplus of legibility in the larger sizes (from 8 point upwards) makes it possible to allow a freedom in the design that serves other purposes than that of legibiüty. But under the unfavourable conditions of sizes below 8 point every other intention must give way to the improvement of recognizability.
If we must sum up the properties of the best type for a given text that has to be printed in a limited space, we need an answer on two preliminary questions. The first conceming the relative importance of space saving, the second concerning the difficulties of the text for the public that is to read it.
The following examples may represent the most common cases.
Newspapers.
The usual sizes are 7 to 8 point for the most important editorial matter. With regard to the choice of type one has to take into



account defective visual acuity of the readers, poor illumination and a non-optimal printing result (grey ink) ; moderate understanding by the reader (either actually moderate intelligence and/or many unfamiliar words and names).
Two fundamentally different Solutions have been found :
1. Starting from the Ionic-Century-Schoolbook group Linotype Ionic Nr. 5 and Intertype Ideal News were developed. These types are extremely simple ; the loops, bowls, dots and upturns have been kept clear of the other parts, the usual hairlines have been considerably reinforced. The projectors are extremely short and the x-height is accordingly large. The shape is wide and round, the lines are heavy. Though being a decided improvement on former types, they made a page extremely heavy and crowded. Hence some lighter versions with slightly longer projectors were issued (e.g. Linotype-Paragon), and these series are very near to the ideal newspaper-type. They are fairly narrow and small and do not possess a superfluous thickness or stress in unimportant parts of the letter. Their extremely slight contrast is feit as a hindrance in the largest sizes only 1).
2. Stanley Morison and his cooperators rejected the reinforcement of serifs and hairlines in their design for the newspaper “ The Times'''2), they adopted contrast as a basic feature. Hence the serifs are thin, but real hairlines, as we know them in the true “modern” face, are skilfully avoided ; in general it is a fairly heavy and narrow face, which combines clear single forms with good word-forms. It is sufficiently bold for poor reading conditions without having too much emphasis for continuous reading ; in fact, it is extremely legible in all sizes for all kinds of informatory texts. It only needs excellent stereotyping and printing when utilized in newspapers, because the sharp impression of the thin serifs and connecting strokes is vitally important.
Both the improved Ionics and the “Times New Roman” should be leaded, at least half a point.
Time-Tables, Dictionaries etc.
The speed with which one line can be singled out among a group of similar but undesired lines is most important in this case. Moreover the contrast between normal, bold and italics
*) The Linotype issued a “Legibility Group” of types, consisting of Ionic No. 5 and Opticon (heavy), Excelsior (medium weight), Paragon (light), and Textype (smaÜ face).
2) The 11 and 8 point sizes of this series have been used in the present volume.



ihould be as large as possible ; the single characters have to be /ery clear because the text is unfamiliar to the reader, no continuous reading is done.
A 1 point lead to a 6 point type can be regarded as the minimum, rhe projectors should extend beyond the x-height for at least the thickness of the limbs of the n. The subjects of our series V appreciated the contrast of type A, but called this type scrawly ; the clearness of type C was universally recognized, but the contrast of normal, bold and italics was found to be insufficiënt.
The types B and D could not compete with A in respect of contrast and size, nor with C in respect of the clearness of the single words.
If in the first place contrast is desired, the body type (normal fount) has to be light, large and fairly wide, very much like the improved Ionics, for a small, narrow, fairly heavy type cannot serve well as a base for the construction of a clear bold version. So the “Times New Roman” is less suitable in this case.
If clearness is needed before striking contrast, the normal version may be more emphasized. In that case simple and fairly heavy types like Monotype-Plantin enter into consideration.
For the shape of the figures we refer to the conclusions of our series with isolated letters.
Continuous Texts.
Pocket editions of long texts (e.g. the Bible) usually need a space-saving type. The reading public comprises all social classes and all ages ; defective eye-sight and poor illumination must be taken into account.
Since their appearance more than a century ago, the “Ionic” types are utilized for this kind of work, and they fulfill these needs very well. When such a face is not too heavy, leading can be dropped in favour of a larger size. These pocket-editions are intended to enable the reader to follow the text (which is fairly well known to him) during the services in the church, where lighting is all but optimal. The numbering of the verses is sufficiënt to find a certain verse easily. As, moreover, the readingspeed is low, we can safely assume that maximal size in the available space is preferable. A page set-up in this way, however, is rather tiring for the skilled reader.
§ 6. COLOUR AND SURFACE OF PAPER.
In an article on the influence of surface and tint of paper on the speed of reading Stanton and Burtt made the remark that



there is a popular belief among typographical authors that a coarse, rough or “antique” finish of a moderate ‘india’ or ‘natural’ tint facilitates legibility. On the other hand, a glossy finish and a white tint are considered to produce eye-strain.
Differences in aesthetical appearance and in feeling-tone are certainly present (cf. Part II of this volume), but they should not be taken for differences in legibility.
Stanton and Burtt, who compared glossy, dull and rough surfaces both in white and ivory (india) tints, could not find a significant difference in reading-rate between these three surfaces. They made use of the Chapman-Cook tests in the same way as Tinker and Paterson. The last named authors also devoted a study to this problem ; they compared three surfaces with different degree of gloss, viz. Egg-shell (22.9 per cent of glare), Artisan Enamel (85.8 per cent) and Flint Enamel (95.1 per cent) ; no differences in speed of reading were found. In the discussion of the results tp have pointed out that longer periods of reading are necessary to confirm these results with glossy papers. They think it possible, however, that the harmful effect of glare is compensated by the advantages of sharper impression and greater purity of the white colour.
We must make two critical remarks in this respect. A sharp impression does not improve legibility in all cases ; MonotypeGaramond for example is very difficult to read on a coated paper, whereas it ranks among the most legible faces when printed upon a rough surface. In the second place an intensily black impression on the purest white, though aesthetically more satisfactory, is likely to become soon “dazzling”. This point has not been elucidated sufficiently ; further studies on the effect of irradiation on different kinds of type should be made, preferably with continuous reading.
On the problem of the relative legibility of black type on white paper versus white type on black paper the last word has not yet been spoken. The inferiority of the latter (± 15 per cent) has been proved by two studies, one by tp and the other by Holmes. The theoretical foundations of this fact, however, are not complete. Kirschmann x) expected a superiority of the white type on black paper on the score of a calculation based on a brightness-difference of 1 : 50, which seems exceptional; he actually found a 16 per cent superiority. Neither tp nor Holmes could give a satisfactory explanation for these findings. Hollingworth pointed out that we do only regard black types as objects
9 Antiqua oder Fraktur ? 3d ed. p. 13.



standing against a back-ground, whereas white types appear as ioles cut in the object. If this lack of object-character of the svhite type were actually the cause of its inferior legibiüty, some training of the reader will remove this difficulty ; experiments will settle this question.
We have to keep in mind, at any rate, that when the reflectionfactor of the object is low (as in the case of black paper), the sensitivity of the eye to brightness-dilferences is also low unless the intensity of illumination is high, about 100 footcandles (cf. Luckiesh and Moss, “Seeing”, p. 65). The intensity of illumination of the exposure apparatus employed by Holmes amounted to 35 footcandles.
Brightness-difference also determines the legibility of texts where the paper and ink are coloured. tp compared ten colour combinations, among them the 4 best and the 3 least legible of a hst reported by Luckiesh, viz. 1. dark green ink on white paper, 2. dark blue on white, 3. black on yellow, 4. light red on yellow, 5. light red on white, 6. dark grayish green on red (fairly dark), 7. dark yellow on black, 8. light orange on white, 9. dark brown on dark green, 10. black on dark purple. The colours are in descending order of legibility compared with black on white. Nr. 9 was read 4.8 per cent slower, which is sufficiënt to advise against its application. This percentage increases through the following numbers, till finally nr. 10 shows a retardation of 51.5 per cent. Needless to say that the advertiser should avoid by any means the use of such colour combinations.
Sumner measured in an outdoor experiment the maximum distance at which legibility of six units of letters or digits became possible, the copy being stencilled in various colours on cardboard also in various colours, together 42 different combinations of colour. His results confirmed the findings of previous authors, that legibility depends on brightness-difference between colour of lettering and that of background. It appeared further that
1. “dark colored lettering on a light colored background is more legible than the reverse in daylight;
2. gray forms the best background for the legibility of colored
lettering;
3. legibility and affective preference of color-combinations show a fairly high positive correlation (rho. 54) ;
4. affective preference of the color-combinations obeys the law of brightness-difference more strikingly than does legibility.”



PART n.
THE ATMOSPHERE VALUE OF TYPE FACES.
CHAPTER 9.
§ 1. Description and method of investigation.
Atmosphere-value of a type face we call those properties by which it excites feelings within the reader.
The reader receives an impression by the mere aspect of the printed type, which is quite distinct from any judgment on the beauty or legibility. The nature and vividness of this impression is expressed in the atmosphere-value of the type face.
It is important that we should know this “feeling-tone” (as we can call it too) of the printing types that are in use to-day. If a type, while fulfilling its original function, i.e. the conveying of a message, at the same time creates a certain mood or feeling, then we should suit this created feeling to the general tendency of feeling of the message. For by doing so, we make the reader more accessible for that kind of communication. Whenever it is important to secure the appropriate feeling of the reader, e.g. in advertising or in the reading of poetry, one cannot neglect this effect, and it is already being made useful in practice, whether consciously or unconsciously.
Another aspect to this is presented by the proved fact that affects (like or dislike) have influence on the speed of apperception (and speed of reading *), and on the strength of associations (and therefore on the memory value of an object). So an intense feeling of dislike can reduce the speed of reading ; in advertising it can impress the advertised object that aroused the dislike very deeply in the memory of the reader — but with this antipathy adhering to it almost indissolubly. In any case we can agree that it is preferable to use a type of which the feeling-tone supports the text, than one, that more or less counteracts it.
French typographers have been pioneers in the adaptation of
*) See A. Prandtl : Experimente über den Einfluss von gefühlsbetonten Bewusstseinslagen auf Lesezeit und Betonung. Z. f. Ps. 60, 1911, 26.



the type face to the character of the text. One of their nrst and best bibliophile pubhshers, Edouard Pelletan, who did already much splendid work in the dark period before 1900, was the first to bring this idea into practice consciously. Quoting Raymond Hesse x): “The text has its personality, since Buffon”, Pelletan wrote in his second letter to bibliophiles: “The black-on-white must have its own personaüty too, independently of the illustration. I have seldom seen a more beautiful text than that of the Traité du Sublime of Longin in the Bodoni edition published in Parma in 1795, that one admires in the show-cases of the Plantin Museum.... For each work a type face has to be found. One could not allow Mathurin Régnier being printed in the same type as Verlaine or Flaubert. It is even clear that Verlaine and Flaubert must be composed differently, given the divergence of their temperaments”.
Another French author, Jean Bruller 1 2), speaking of the difference between the Old Style famiües, writes : “Each of them has such a personal physiognomy, which has such influence on the aspect of a text, that anyone reading Rabelais composed in bodoni or Valéry in old style looses a good deal of his pleasure. For example : Carlègle has designed a bastard character, the iDorique\ which name one will understand if one knows that the limbs are thicker at the base than at the top. With this character an edition of the Histoires Naturelles of Jules Renard has been made. One can imagine the effect of a phrase like this one : “La puce: un grain de tabac a ressort”, composed in a type face, that seems to have been shaped out of the very columns of the Parthenon ....”
Stephane Mallarmé brought this idea into practice. Paul Valéry (quoted by Jean Selz in “Publicité 1936”) describes this very clearly : “Mallarmé avait étudié trés soigneusement (même sur affiches, les journaux) 1’efficace des distributions de blanc et noir, 1’intensité comparée des types .... Une page, dans son système, doit, s’adressant au coup d’ceil qui précède et enveloppe la lecture, “intimer” le mouvement de la composition; faire pressentir, par une sorte d’intuition matérielle, par une harmonie préétablie entre nos divers modes de perception ou entre les différences de marche de nos sens, ce qui va se produire a l’intelligence. II introduit une lecture superficielle, qu’il enchaine a la lecture linéaire.”
Guillaume Apollinaire and those of his spirit, also in other
1) Le Livre d’Art du XIX® Siècle h nos Jours, Paris 1927.
2) in ie Livre d’Art International, Special Number 26 of Les Arts et Métiers Graphiques, Paris 1931, p. 44.



countries, e.g. Paul van Ostayen in Belgium, have made many experiments, in which they tried to realize their intentions by the direct support of lay-out and face of type. The influence of these typographical ideograms on current typography has been considerable, and can be traced in many advertisings of to-day.
The intensity of the feeling-tone varies from type to type. Printing types are objects of applied art. In the design the sense of beauty of the artist or of the master craftsman has expressed itself. In the shape of the object one idea is realized, that the artist has deliberately chosen out of many other possible ideas, but that he could only realize in that peculiar way, to which he was forced by his whole personality.
Every object of art reflects the personality of its originator.
According to Klages (whose principles we follow in this part of our study) “the way by which every voluntary movement passes by, contains something that is lastingly characteristic for the executor of that movement” 1). Only the Soul (opposed to Mind and Body) expresses itself. “The expression {Ausdruck) realizes (in its intensity, duration and the course of its directions) the form (Gestalt) of a stir (Regung) of the soul” 2).
We can read thus the expression of the soul in every tracé it has left inasmuch the motion was not bound to a fixed form. Now there is one easily accessible group of tracés of voluntary movements, consisting of variations upon a simple theme (which makes them comparable and apt for systematization), whereas enough freedom is granted to allow personal differences to manifest themselves clearly. We mean handwriting. Not everybody is an artist who can express himself in decorations, printings, sculpture etc., but everybody can write. In the handwriting we find the pure precipitation of the voluntary movement and therefore of character-traits.
To some extent a type design can be compared with handwriting, with the restriction that it is executed consciously, controlled and retouched on the ground of aesthetical and technical considerations. This fact does not exclude the possibility of a systematical analysis of the expression-value of such art-hands ; in fact, Klages dedicated a chapter to this in his standard-work on graphology, under the title “ Die erworbene Handschrift”. He points out that the result of the analysis is necessarily poorer
*) Grundlegung der Wissensch. vom Ausdruck, p. 30. s) Ibid. p. 156.
Legibility, atmosphere-value and forms of printing types. 9



than that of the personal handwriting, but the significance is not fundamentally modified; hence the application of graphology on type-designs can be taken into consideration.
In order to be able to answer the question whether a scientifically reliable analysis of letter designs can be carried through, we have to consider the method of graphology, i.c. the system of Ludwig Klages, which I consider the only acceptable one.
Whereas, according to the former French school of the Abbé Michon, each feature of a handwriting had its own unchanging significance (theory of the ‘ signes fixes'), these features are, according to Klages, bi-valent, they can be explained in positive or negative sense (theory of ‘ Doppeldeutigkeit') ; they get their full significance only after confrontation with a judgment on the level of form (Formniveau) of the whole text, which contains an appreciation of the specific relation in the person of the writer between Mind (Geist) and Soul (Seele).
An analysis based on Klages’ system is therefore all but an automatic application of fixed rules ; a vast characterological experience is necessary if one is to judge about the significance and the relative importance of the various traits in relation to the total character of the concemed writer. Especially the appropriate location of the writing on the scale of form-level is not everybody’s work, because it supposes on the part of the analyst those very properties that make for a high level of form, and which Klages calls, in pregnant terms, ‘ durchgeistigtes Leben', or the expression of the ‘ Lebendigkeit geistbetönter Persönlichkeiten', and which is peculiar to those persons, who unite in themselves a perfect harmony of Mind with its antagonist, the Soul. Klages defines these two idea’s as follows : “ ... der Ermöglichungsgrund des Zusammenhangens aller Erlebnisse einunddesselben Wesens, mogen sie im Wachen oder im Schlafen stattfinden, gleichzeitig oder nacheinander, heiBt die Seele, der Ermöglichungsgrund ihres BewuBtwerdens heiBt der Geist."
We must refrain from going deeper into this matter, as this would lead up to regions, that fall beyond the scope of this study. But it may have become clear that the fixation of the level of form is an irrational factor, which has a function comparable with that of the key to the stave in music : it settles the true value of the single factors that are subjected to it. It consists of a peculiar blend of intuition and analysis, and here is the common part of “Sciences” which infer the character from the bodily appearance and the effects of the actions of the individual, such as graphology, chirology, physiognomy etc. It is a custom of those who practise such Sciences, to deny this intuitonistic



base of their system, and they emphasize the analytical-systematical construction, which they declare to use as sole instrument. This is the cause of the impression of incongruity that comes to the unbiassed reader, who is accosted by sentences about an alleged relation, that present themselves under the claims of the infalhble correctness proper to intuition, while being wrapped in the form of scientifically inadmissible generalization. Psychologically such frantical negations are easy to explain. The big public has been educated in distrust of unusual phenomena and powers, that the official Science cannot explain. Artists who show conversance with some of these things are forgiven, because they are queer people anyhow, and the rest is assigned as a harmless play to spiritualists or as a despicable swindle to gypsies. If therefore anyone is so sure of his powers of such a strange kind that he feels himself justified to use them at pay for the educated public, then he takes care to avoid an appropriate admission or even a mentioning of the intuitional part of them.
The psychology of the unconscious has placed us during the last years before so many inexplicable facts, that this self-conceit is recognized as such and that it is no longer considered as the hall-mark of the truly scientific attitude.
The systematical constructions of the character-finding theories mentioned above, are serving as guides to the analysis of the intuitional knowledge, which is clear and detailed, but not systematical and intellectual (Begrijfsmafiig); they make the analysis complete and fixable.
They can be only of use to such people, who are in possession of the primary gift of intuition ; there are no little people among the educated who possess these intuitional powers, of those however there are not many who should like to practise these Sciences. Nevertheless, when we do not venture to project a theoretical system for the graphology of printing types and decorative handwriting, this has another reason than a doubt in respect of the intuitional powers of the reader. It is rather the smallness of the problem, which does not seem to award the considerable amount of necessary labour, in view of the purpose of the present study. Taking this course of action, we make the results of the experimental investigation (who would have served otherwise only as a control for the theoretically developed judgment) the main material of our study, which will be analyzed and, if possible, to some extent generalized with the aid of Klages’ system.
The application of this system (see Table XXI), which is valid only for handwriting (i.c. the direct product of the writer’s vital impulses), upon a type face, that was designed in long hours of



;onscious shaping and reshaping, is therefore only justified by the restricted and provisional use we make of it. The classifi:ations will serve as a guide in the analysis of the results and :hey can be transposed, therefore, without many alterations.
The characterological significance of the various traits has Deen taken over with some slight corrections, which seemed iirectly necessary. For the theoretical explication of these signiicances and for a further correction upon the use we make of them we refer to the original works dealing with graphology of Klages himself. It may be sufficiënt here to point to the foliowing restrictions to be made at our tentative explanations :
1. A high level of form and rhythm *) can only be reached if the soul of the writer is not obstructed in its expression by unsurmountable difficulties from the side of the material, such as is the case in the design of printing types. The latter are therefore necessarily poorer in contents of Life (Lebensgeha.lt), to use Klages’ terms. In most of the cases the designer tries to realize a special omamental pattern or to express a certain mood, attitude, or frame-of-mind. Then the decorative aims and the conscious section of the personality supersede the total personality. Such an intentional expression has to be judged by its own standards, viz. by the consistency of the elaboration of the premeditated plan, which we call “ductus”, in this special circumstance. To the “ductus” we have given the same five degrees as Klages assigned to his ‘level of form’. A clear example of what we mean, is to be found in Bernhard Roman and Cursive, of which the fundamental idea has been developed in a very good way, deserving a ductus grade nr. lf, whereas the personal contents of that very idea might be honoured with a level of form of about 3.
On the other hand, Omega and Metropolis are examples of a lower ductus, because there is a contradiction between their boldness and the features of refinement (long ascenders, delicate serifs of Metropolis). Such inadequate traits lower the degree of ductus. Likewise, the semi-bold edition of Locamo (nr. 24) has a somewhat lower ductus than the light version (nr. 11), for this very boldness is not quite compatible with the other features of the design.
In the analysis of the results features and (or) their significances have been thrown out
a. when they were irrelevant for the analysis in question,
b. when they could not be determined with sufficiënt certainty.
i) Rhythm, in Klages’ terms, is the feature of the natural movement, appearing as regular, periodical variations (Schwankungen) in the movementsof a human being.



2. In respect of rhythm nothing can be said, as each character of the alphabet has but one form. The corresponding degree of affectivity has to be inferred from other features, as far as possible.
3. The degree and nature of connection cannot be circumscribed with certainty either, though an interrupted, resp. uninterrupted connection may be assumed for the gothics resp. the script types used in our study.
4. “Pastiness” supposes thick lines with round contours, “sharpness” thin lines with sharp, angular contours. A real pasty type is Columbia (nr. 29), a sharp specimen is Bernhard Roman (nr. 1). Many types with much contrast, however, cannot be put in either of these categories, so e.g. Bodoni, Omega, Metropolis ; they combine features o:- both, because the pressure, necessary to achieve such thickness with such a sharp writinginstrument as has been used, would have to be so excessive as to be improbable. Therefore we cannot call these types sharp, and though they are bold, they are too sharp to be called pasty.
5. Long ascenders of type faces are not drawn in an upward movement, from which Klages inferred their significance as “capacity for illusion, enthousiasm”, for “mobility of the mind” in general, but in a downward movement, which requires a prolonged, careful innervation, terminated by a little movement for that part of the character, which covers the x-height. The significance of long ascenders (on sharp types) is therefore rather that of delicacy, precision and grace (cf. Bernhard Roman, Locarno (Eve), Nicolas Cochin, Ehmcke Antiqua).
As a rule, short projectors have the significance which Klages attachés to D > A. A pronounced case of D > A does not exist for printing-types, as far as I know.
6. Speed and inclination can be determined for script types only; in some cases the design allows to speculate upon the probable form of the corresponding handwriting, so e.g. steepness for the gothics, backward inclination for Gallia.
7. Where an analysis of a whole type face has to be made (not its appearance in a practical application) its ‘natural size’ has to be found out. As a rule this size is lying somewhere between 12 and 28 point for the modern book types, and not in the usual book sizes up to 12 point. The old engravers’ types, newspaper faces etc. have to be judged by their small sizes ; for display types this depends on the purpose they were made for.
A highly remarkable use of this ‘natural size’ was made by E. Schiffrin (Editions de la Pléiade) in his Dostojevsky-editions,



where an ordinary kind of didot (essentially a small face) was chosen in a large size. The result is a curious pathological atmosphere, of incongruity and exaltation.
8. Some of the subjects for advertizing of the questionnaire (“feelings”) do not have a characterological foundation ; they are founded upon spatial associations (see under the report of the results).
§ 2. Former studies.
In 1920 Anna Berliner published an investigation into the atmosphere-value of 18 type faces in respect of four commodities, viz. 1. Fish, 2. Pork and Beans, 3. Pancake Flour, 4. Orange Marmalade. 20 female subjects made up a rank-order of appropriateness of these 18 types for the 4 commodities. The conclusion of Berliner was as follows : “Das auswahlende Prinzip der Atmosphare eines Artikels ist spezifisch insofern, als nicht jeder Artikel dieselbe Auslese bedingt wie jede andere. Dieser spezifische Character ist jedoch dadurch ein geschrankt, dass es Gruppen von Artikeln gibt, deren auswahlendes Prinzip sich mehr oder weniger ahnelt.”
The types were designed by a New-York advertising agency, they do not bear a resemblance to any printing types. Moreover, the unclear reproductions given by Berliner do not allow to draw any conclusions. Hence we could not take her results in account for our own work.
In 1923 Poffenberger and the auctor intellectualis of Berliner’s study, R. B. Franken, determined the appropriateness of 29 printing types for five qualities, viz. cheapness, dignity, economy, luxury and strength, and five commodities, viz. automobiles, building material, coffee, jewelry and perfume. They used also the rank-order method. Though many of their type faces are scarcely used nowadays, the results have some value for purposes of comparison. Other types, but the same quaüties and commodities, were used by Gwendolyn Schiller in a later study, which followed the same method and dealt with colourcombinations in advertising too.
The rank-order method is very simple, both for the investigator and for the subjects. The absence of absolute values, however, is a serious drawback in this case, because the specificity of the atmospherevalue and the questions can only be inferred from the correlations, i.e. indirectly. In our own enquiry we have made use, therefore, of a valuation with figures that are absolute as far as possible.



A fourth study was made by Da vis and Smith, in order:
1. “to determinate the effect of the different variations of mechanical characteristics in the same type face on the feeling tone : boldness, condensation, expansion, italics and size ;
2. to investigate the similarity of choice of type faces of the same family, particularly Caslon and Cheltenham ;
6. to investigate the relative importance of the different characteristics in the selection of a type face to express a feeling-tone ;
4. to compare two methods of approach to the problem — select a type to fit a product or feeling (as the typographical worker has to do) or selecting a product or feeling to fit a type face — and to make a study of the relative merits of the two methods.”
13 types were employed ; they certainly make it possible to group the results into the four categories of boldness, italics, condensation-expansion and size, bpt, on the other hand, the range of possibilities of expression is rather limited, compared with the choice of types which the typographer has at his command. In fact, this classification leaves so much room for fundamental differences in design (and hence for differences in feelingtone) that we could not use these results for our purpose.
Da vis and Smith employed 23 products and 24 feelings, each containing but three of the products resp. feelings used in the two studies mentioned above. They have found, that the method, “where the product or feeling is chosen to fit the type face is the more justifiable of the two, psychologically and statistically.” The products showed a minor uniformity of ranking than the feelings ; the former are more likely to arouse different sensations indifferent persons than the latter. As, moreover, the conclusions in respect of products are more difficult to generalize than those of the feelings, we have only made use of feelings in our own enquiry, or (in blank “A”) of sufficiently general descriptions.
An extract of the results of Poffenberger and Franken, Da vis and Smith, and Schiller will be found in Table IX and X.
§ 3. Series Vila.
The studies mentioned in § 2 dealt with American types, that did not show much variation in design and that could not be said to be representative for the enormous range of advertizing types, out of which the European typographer is able to make his choice, and that include English, German, French and some



TABLE IX.
List of Type Faces used by Poffenberger and Franken (PF), Schiller (S), Davis and Smith (DS).
POFFENBERGER AND FRANKEN
1 Della Robbia 16 Masterman
2 Bodoni Bold 17 Typo Slope
3 John Hancock 18 Century Bold
4 Roycroft Tinted 19 Post Monotone
5 New Caslon Italic 20 Caslon O. S.
6 Blair 21 Cheltenham Bold Outline
7 Caslon Old Style 471 22 Tiffany Text
8 Caslon O. S. 471 Italic, SwashCaps 23 Boorman O. S.
9 New York Gothic 24 Cheltenham Bold
10 Century O. S. 25 Globe Gothic Bold
11 Old English 26 Engravers Roman
12 Goudy O. S. 27 Roycroft
13 Scotch Roman 28 Bullfinch
14 Circular Gothic 29 Priory Text
15 Antique Bold
SCHILLER DAVIS AND SMITH
1 Bernhard Cursive 36 pt. 1 Caslon Bold 36 pt.
2 Bernhard Cursive Bold 36 pt. 2 Caslon Bold 24 pt.
3 Broadway 36 pt. 3 Condensed Gothic 36 pt. (Bold)
4 Caslon O. S. 471 Italic 36 pt. 4 Cheltenham Condensed Bold 24 i
5 Caslon Text 36 pt. 5 Cheltenham Cond. Extra-Bold 24 i
6 Cheltenham O. S. 36 pt. 6 Caslon Bold Condensed 24 pt.
7 Cheltenham Bold 36 pt. 7 Caslon Bold Expanded 24 pt.
8 Clear Cut Initials 36 pt. 8 Goudy Bold Italic 24 pt.
9 Eve (Locarno) 24 pt. 9 Old Style Italic 24 pt.
10 Futura Black 30 pt. 10 Caslon Italic 24 pt.
11 Goudy (O. S.?) 36 pt. 11 Century Italic 12 pt.
12 Parisian 36 pt. 12 Caslon 24 pt.
13 Scotch Roman 36 pt. 13 Century 12 pt.
14 Washington Text Shaded 30 pt.
15 Heavy Gothic 8 Ot.
Italian, Spanish and Dutch designs. Moreover, no book faces were used in these experiments. It was therefore thought possible, if not necessary, to add another series of experiments to the two American works.
The main purpose being to obtain more differentiated results for comparison, a somewhat other method was chosen than the one used in the previous investigations. First, a greater number of type samples were procured, viz. 38 book types and 48 display types.
In order to avoid any possible influence of the atmosphere of the text, a nonsense text was chosen. Secondly, the subjects were asked to describe the observed atmosphere with the aid of a series of adjectives (which were printed upon a sheet) as



TABLE X.
Comparison of the results of PF, S and DS for the same or related qualities and commodities.
O
ü Quality Kind of Type Commodity Kind of Type
o <
PF Cheapness 15,3,25,27,16 Automobiles 13,16,15,24,20
S 7, 6, 3, 13, 10 4, 6, 13, 12, 11
DS Bold
Normal Width Expanded
Large
PF Dignity 29,22,26,11,17 Building Material 15, 16,3,18,25
S 14, 5, 1, 2, 12 7, 3, 13, 10, 11
DS Non-Bold Hardware Tools, Coal Bold
Italic Non-I talie
Normal Width Condensed
Small Large
PF Luxury 22,26,17,11,29 Perfume 17,8,14,26,22
S 1, 14, 2, 8, 12 1, 2, 9, 12, 14
DS Beauty, Delicate Non-Bold Non-Bold
Italic Italic
Normal Width Normal Width
Small-Medium Small
PF Strength 25, 15,3,16, 18 Jewelry 26, 17, 8, 22, 14
S 7, 3, 10, 13, 11 1, 2, 9, 14, 12
DS Bold Non-Bold
Non-I talie Italic
Expanded Normal Width
Large Small
PF Economy 23, 10, 13, 9, 1 Coffee 25, 3, 15, 27, 24
S 11, 6, 12, 13, 15 11, 6, 12, 13, 7
completely as possible, or, if necessary, with other adjectives.
It must be kept in mind, when reading the following translation of these adjectives, that they were computed at the idea’s of the average civilized Dutch people, and also, that the translation can be but a defective rendering of the meaning of the original Dutch words.
1. French, German, English, American, Dutch, neutral.
2. Agreeable, disagreeable or neutral.
3. Legible, illegible or neutral.
4. Elegant, unelegant, ungainly (unwieldy), stately, dignified, noisy, insolent, matter-of-fact, romantic, pithy (to the point), loutish, reliable (solid), whimsical, airy (light-hearted), spiritual (witty), delicate, courteous, cheap, comfortable, modern, oldfashioned, precise.



Thirdly, it was thought advisable to check-up these descriptions by practical applications, like the other investigators did. For book types a number of book titles were chosen and the subjects were asked to write down the title of the book for which they should like best to use the type in question. An attempt was made to give only titles of books with a particular and pronounced atmosphere, experienced by all subjects in the same way, but it is doubtful whether we succeeded in this respect. The list included the following foreign books : The Holy Bible, Satires by Horace, Grausame Geschichten by H. H. Ewers, Also sprach Zarathustra by Nietzsche, le Mariage de Figaro by Beaumarchais, Mémoires of Casanova, Nibelungenüed, Hauptsatze der Differential- und Integralrechnung, Tales of Hans Andersen, der Arzt und seine Sendung by Liek, Nonsense Novels by Stephen Leacock, Geschichte der Caesaren by Mommsen, the Crimson Circle by Edgar Wallace, Pensées of Pascal, II Principe by Macchiavelli, De Bello Gallico by Julius Caesar ; Maximes of de la Rochefoucauld, Balladen by Schiller ; further : the Official Parliamentary Reports; works on Napoleon, on steel-construction, on the modern country-house ; finally some Dutch or Flemish novels : Pallieter by Felix Timmermans (glorification of the abundance and fertility of the earth and of its joys), Jordaan by Querido (multicoloured, impressionistic description of a convulsive quarter of Amsterdam), Eline Vereby Couperus (description of the languid and refined society-circles of the Hague), Het wassende Water, by de Man (description of the rough and sturdy peasant life) ; In den tuin van Eros by Engelman (graceful and playful poems), Poems by Leopold (sensitive and finely strung), and Erasmus by Huizinga (clear analysis in splendid form of the figure of this humanist).
The subjects were asked to write down the name of an article which they should like best to advertise with the display type in question. These articles were the following : very old gin (in old fashioned terms with suggested authenticity), precisioninstruments, Jaeger underwear, bankruptcy-sales, central heating, tinned food ; advices in tax-questions, jewelry, perfume, Aspirine, school of dancing, camival, real silk, armoured concrete, pipetobacco, lubricating oil, steel fumiture, trucks, non-splintering safety-glass, typewriters, leather goods, dairy-fresh, exposition of paintings by old masters, super sport cars, Oriënt cruises with luxury liners, invitation for a royal court-ball.
The samples were composed of 10 point type with the same length of line, they were glued on sheets of white paper in two columns of six. Because it was feit that the task was a heavy



one, requiring much time and devotion, the subjects were allowed to take the samples and application blanks home. When a dozen subjects had worked at his questionnaire, a further continuation was suspended, because the imposed task proved too heavy for the subjects, while moreover the choice both of types and possible descriptions was embarrassing. It was decided to reduce the number of types and to simplify the questionnaire, the results of the former inquiry serving for substantiation of the eventual results of the new series.
§ 4. Series VII b.
The types.
It was decided upon to use but 30 types in the abridged text, though such a small number of specimina cannot be possibly representative for the hundreds of different types that are in daily use on a large scale. We hoped to retain at least the principal features of the most important classes of design in these few samples, so that we should be able to compute the significance of these features by careful comparison. We let here follow the names of these types such as they are known in the catalogues of the original foundry or composing-machine, with the English name in parenthesis. The list of modern printing types on page 229 contains further information about them.
A. Book types, series VII b.
1. Zarte Bemhard Antiqua (Bernhard Roman),
2. Thannhauser,
3. Intertype-Bodoni,
4. Lutetia,
5. Linotype-Saeculum,
6. Orpheus,
7. Beton,
8. Intertype-Mediaeval,
9. Monotype-Modem Wide 25,
10. Bembo,
11. Koch Antiqua (Locamo),
12. Romaansch (Intertype-Lorimer),
13. Grasset.
Beside these thirteen book types, series a had included also : Tiemann Mediaeval (Wren), Pastonchi, Eckmann, Auriol, DeutschRömisch, Caslon, Imprint, Monotype-Garamond, Perpetua, Mono-



;ype-Baskerville, Sorbonne (Cheltenham), Monotype-Italian Old Style 108, Centaur, Monotype Old Style 2, Monotype-Bell, Weiss \ntiqua, Bodoni of the Bauer Foundry, Genzsch Antiqua (Staniiope Old Style), 1’Astrée (Mazarin), Behrens Antiqua, Behrens Mediaeval, Intertype-Futura 154, Parcival, Belwe Antiqua, Linotype-Excelsior (in 8-point size).
B. Display types, Series VII b.
14. Bernhard Schönschrift (Bernhard Cursive),
15. Monotype-Gallia,
16. Kontrast,
17. Hollandsche Mediaeval (Intertype-Mediaeval),
18. Narziss,
19. Excelsior (similar to Palace Script etc.),
20. Omega,
21. Eckmann Fett (-Bold),
22. Schaduw Nobel (Nobel Shadow),
23. Metropolis Fett (Metropolis Bold),
24. Koch Antiqua Fett (Locarno Bold),
25. Smalle vette grotesk (narrow heavy gothic),
26. De Vinne,
27. Quick (Trafton Script),
28. Arpke Antiqua,
29. Columbia (comparable with “Morland” and “Hawarden”),
30. Breede magere antieke (light gothic expanded).
Series a included also : Block, Erbar light, Goudy Heavy, Meridian, Fanfare, Futura Semi-Bold, Futura Bold, Futura Black, Bodoni (Bauer-edition), Pragefest (Samson Script), Lucian (Bernhard Bold), Lo, Genzsch Antiqua Fett (Stanhope Old Style Bold), Klingertype, Stadion, Koloss, Atrax, Bernhard Negro, Elegant Grotesk Licht (Elegant Outline), Naudin, MonotypeAshley, Lilith, Ella Bold, Neuland, Metropolis Licht (- Inline), Bristol, Lichtfette Grotesk (Phosphor Bold), Bifur, Maximilian, Beton Bold, Tauperle, Ludlow Black, Wieland Lichtx).
The book types were printed in 10 pt. type, the display types, where possible, in 20 point type. Of some display types, which are in use exclusively for head-lines, only two lines were reproduced. Owing to misunderstandings of the printers, some other types, that belonged to this category, are represented with more
!) These types can be either found in the works of Bastien and Jtreshwater, Waite, the Hartdbuch der Schrijtarten, or in the specimen books of the Mono-, Lino- or Intertype.



than two lines ; it is possible that they have a slightly blacker appearance than the two-line faces.
The texts were again gummed on sheets of white paper. The lines were 26 Didot augustins long, without interlineage.
The questionnaire.
After i: had become clear that a qualitative method, such as was used in series a, was less suitable for the investigation of the problem in question, the second series was planned on more quantitative lines. About 70 subjects were employed, and in order to make their work and the elaboration of it easier, we decided upon a method, by which the subjects had to assign marks proportionately to the appropriateness of the type faces for the printing of certain books or for the advertising of certain goods.
We had learned from series a that the subject’s feelings, though positive, are not very differentiated if they are to be defined; that the subject often only can define them after imaginary application to a practical case, and therefore only such applications were asked. Three groups of questions were framed. The questions were arranged in a column on the left of the blank on horizontal lines, one beneath the other, whereas the narrow columns for the types were placed on the righthand side of the blank in vertical direction, one beside the other. At the bottom of the blank some lines were left open for remarks. The first group of questions asked for a judgment on the legibility ; the subjects had to put ‘yes’ behind either ‘clear’, ‘moderately clear’ or ‘unclear’. The second group consisted of eight subjects of books, for which the appropriateness of the 13 book types had to be considered. They were the foliowing :
1. romantic adventures,
2. modern technics and architecture,
3. frank and rough farmers’-life,
4. stately, classic tragedy,
5. witty, spiritual satire,
6. clear analysis of culture-problems,
7. matter-of-fact report,
8. delicate, sensitive poems.
The third group consisted of eight goods or ideas, for which the appropriateness (on advertizing-purpose) of both the thirteen book types and the seventeen display types had to be considered. They were :



1. force, strength,
2. economy, soberness,
6. luxury, refinement,
4. distinction, dignity,
5. precision, fineness,
6. hygiene, freshness,
7. reliability, solidity,
8. warmth, comfort.
These ideas offer sufficiënt points for comparison with those >f Poffenberger and Franken, and of Schiller, who used :
Cheapness, Dignity, Economy, Luxury, Strength, Automobiles, iuilding Material, Coffee, Jewelry, Perfume.
We expected that both these subjects for books and those for idvertising would bring about rather distinct complexes of feeüng, iuited to feelings that the subjects had proved to be able to define n series a.
The instruction.
The subjects received a printed instruction, which runs as 'ollows (some unimportant indications omitted) :
Instruction.
The purpose of this inquiry is : to get an answer to the question vhether a type face is surrounded by a certain atmosphere, and, if this should be the case, by which one. On the blank is reserved i column fore each type face that occurs on the sheets that are lying before you, and its number is equal to that of the corresponding type on the sheets.
You are to look carefully at each type face, one by one, from a large and a short distance ; then you must form an opinion of it (if you like, by comparison) and after that you have to give an answer to questions put in the left column.
Blank ‘A’ (13 types).
The group of [questions nr. 1 have to be answered by yes behind the judgment you agree with.
In group nr. 2 you have to assign a rank-figure of preference, the more appropriate you think the type face to print a book with it, of which the general subject is indicated in the left column. If you think a type face very appropriate to print a book of sensitive poetry with it, then you have to put a 5 behind sensitive poems’ in the column of that type. Perhaps you think it unfit



for another kind of books ; in that case you put a 1 behind that kind of books. But for a third kind it could perhaps be used, if not by preference ; then you put behind that some figure between 1 and 5. And so you are to go on, assigning a mark for each subject.
Under group nr. 3 you will find some notions, that you have to advertise. In an advertisement for perfume, for jewelry etc. there has to be an atmosphere of refinement, which, however, is not entirely equal to that of ‘distinction and dignity’, though 'they are related. Here again you are to give rankmarks from 1 to 5, proportionally to the appropriateness of the type face for the advertising of that notion.
Blank B (17 types).
Answer only the groups of questions nr. 1 and nr. 2.
You can dispose of 5 figures, of which the value is indicated on the blank. (These values were : 5 = very appropriate ; 4 = can also serve for it; 3 = (the Dutch expression can only be translated approximately) just sufficiënt; 2 = rather not; 1 = wholly unsuitable). It is possible that you think a type face is rather neutral; you will, of course, not give the extreme marks in that case. You must imagine that you are a printer, who owns only the type face you are looking at just now, and that you want to know for which kind of work you must buy first a new, more suitable type face, for which kind of work after that, and so on.
And then there is still something important. When you have judged a certain type face as suitable for printing e.g. a ‘matterof-fact report’, but also for advertising ‘strength’, then do not cross-out one of the two because a matter-of-fact report has nothing to do with strength. So you must not mutually compare between the books, between the advertising subjects, and between these two, whether they are corresponding or not, but only between the type faces and these subjects.
Conditions.
For blank ‘A’ 71 subjects were employed (31 men and 40 women), for blank ‘B’ 68 subjects (31 men and 37 women) ; they were seated in a room, which was artificially lighted by four 125 Watt bulbs, immediately above the tables. Each of them had sheets with the type samples, a printed instruction and an application-blank. Some of the subjects finished the task in about 75 minutes totally, whereas others needed almost twice that time.



rABLE XI.
Average scores of the 13 booktypes for the subjects for books and
advertising.
The mean deviations are printed in light type beneath the scores.
B
S P
GQ J ®
PP g > g
« ^ . P CO >• ►_
◄ _ B ^ p P B ~ _ 53 5 3
W fc fc fc H p S B <1 2 P4 ^ oa
fc < fc O « p B O £ -r £ ◄ 2 w
tf £ -4 P H w Pj ^ g O
sp b o p •< 05 3 ® 2 S
pqp3 fn k >j cc o w s s w m w o
1 2 3 4 5 6 7 8 9 10 11 12 13
Romantic Adventures 2.5 2.7 2.6 2.7 2.9 2.5 2.0 3.1 2.5 2.5 2.3 2.3 3.0
1.0 1.0 1.1 1.0 1.0 0.9 1.0 0.8 1.0 0.9 1.3 1.0 1.1
Mod. Technics and 1.3 3.0 3.2 2.3 3.2 2.6 2.6 3.1 2.8 3.3 1.3 2.9 2.2
Architecture o.5 1.0 ï.o 0.9 o.9 0.8 1.2 0.8 0.9 o.8 o.4 o.9 0.9
Frank and rough 1.1 3.1 3.1 1.8 2.6 2.2 4.0 2.8 2.3 3.0 1.0 3.1 3.3
Farmers’ Life 0.2 0.9 1.0 0.6 1.0 1.0 0.8 o.9 0.9 0.8 0.0 1.0 1.1
Stately, classic Tragedy 2.3 2.5 2.3 2.7 2.3 2.5 2.1 2.6 2.3 2.8 2.1 2.7 2.7
1.0 0.9 1.0 1.0 0.9 0.9 1.0 0.9 0.8 0.8 1.0 1.0 1.1
Witty, spiritual Satire 4.0 1.6 1.5 2.9 2.5 2.3 1.1 2.1 2.2 2.2 3.6 1.4 1.6
0.8 0.7 0.7 0.9 1.0 0.9 0.!( 0.7 0.8 0.9 1.1 0.5 0.6
"lear Analysis of Pro- 1.5 2.8 3.1 2.5 3.5 2.9 2.2 3.3 2.8 3.5 1.2 2.8 2.1
blems of Culture o.6 ï.o ï.o 0.9 0.9 1.0 0.9 0.9 0.9 0.8 0.3 1.0 0.7
Matter-of-fact Report 1.1 2.3 2.7 2.3 3.6 2.6 2.0 3.2 2.8 3.2 1.0 2.7 2.0
0.2 1.0 1.0 0.8 1.0 1.1 0.8 1.1 1.0 0.8 0.0 1.0' 0.9
Delicate, sensitive Poems 4.1 1.6 1.5 2.9 2.0 2.2 1.2 2.1 1.8 1.8 3.8 1.3 1.7
0.8 0.7 0.6 1.1 0.9 0.9 0.4 0.7 0.7 0.7 0.9 0.4 0.8
Force, Strength 1.1 2.9 3.5 1.7 2.5 2.0 4.5 2.6 2.3 2.9 1.1 3.5 3.6
0.2 0.8 1.0 0.7 0.9 0.7 0.6 0.8 0.9 0.9 0.2 0.8 0.8
Economy, Soberness 1.5 2.5 3.1 2.4 3.7 2.5 2.4 3.1 3.0 3.0 1.2 3.1 2.3
0.6 1.0 0.9 0.9 0.7 0.9 1.0 0.9 0.8 0.8 0.3 0.9 0.9
Luxury, Refinement 4.3 1.9 1.9 2.2 2.0 2.5 1.4 2.0 2.0 2.0 4.5 1.7 1.9
0.6 0.7 0.9 0.9 0.3 0.8 0.6 0.5 0.9 0.7 0.6 0.7 0.8
Distinction, Dignity 3.5 2.5 2.5 3.0 2.3 2.7 2.0 2.5 2.2 2.3 3.1 2.4 2.6
0.9 1.0 1.0 0.8 0.9 1.0 0.8 0.8 0.8 0.8 0.9 1.1 1.1
Precision, Fineness 3.7 1.9 2.0 2.9 2.4 2.7 1.4 2.2 2.2 2.4 3.5 1.9 1.7
1.0 0.7 0.7 0.9 0.9 0.8 0.5 0.8 0.8 0.9 1.0 0.6 0.7
Hygiene, Freshness 2.2 2.9 3.0 2.6 3.1 3.0 2.9 3.2 2.4 3.3 2.0 3.0 2.2
0.9 0.9 0.8 0.8 0.9 0.9 0.7 0.8 0.8 0.7 0.8 0.8 0.9
Reliability, Solidity 1.3 3.1 3.9 2.2 3.4 2.7 3.8 3.1 3.2 3.2 1.1 3.6 2.9
0.5 1.0 0.7 0.8 0.9 1.0 0.8 0.8 0.7 0.8 0.2 0.8 0.9
Warmth, Comfort 2.0 2.9 2.6 2.1 2.4 2.3 2.7 2.5 2.5 2.9 1.9 2.4 3.2
n 8 o o n» 0 7 0.9 0.7 1.0 0.8 1.0 0.6 0.8 0.9 1.0



A (13 book types).
The average scores of each type face for each of the subjecte for books or advertizing are given in Table XI. They were obtained by adding the various number of points assigned by the 71 ss. and dividing this total amount by 71. So e.g. in the case of type nr. 1 for romantic adventures : 23 votes for 1, 13 for 2, 11 for 3, 20 for 4, 4 for 5, totally 71 votes for 182 points ; average : 2.5, with a mean deviation of 1.0.
The value of these average scores cannot be interpreted by means of the original Standard employed by the subjects, so e.g. 3 points = just sufficiently suitable, but they have to be regarded as relative values. For that reason a rank-order of merit for each of the subjects for books or advertising was developed from the absolute values, to be found in Table XII, which served as a basis for the calculation of the coefficients of correlation between these subjects (Table XV), by means of the Bravais-Spearman rank-order of correlation formula
Another rank-order (for each of the type faces) is shown in table XVI ; from the latter the coefficients of correlation between the type faces (Table XVIII) were computed by means of the same formula.
B (17 display types).
The absolute values are found in Table XIII, the rank-orders of the type faces for the feelings in Table XIV, those of the feelings for the type faces in Table XVII; the coefficients of correlation between the type faces in Table XX.
The results are analyzed, first, in respect of the subjects for books and advertizing, secondly, in respect of the type faces.
The (modified) system of Klages has been used as far as possible (see Table XXI). The subjects for books or advertising (hereafter called quaüties) can be determined with this system, wherever they can be “translated” into a mood, frame-of-mind or way of conduct of a living being. Some of the quaüties, however, are characterologically more or less indifferent, e.g. hygiene and warmth. Like the “commodities” of the authors mentioned above, their feeüng-tone must be expressed in spatial analogies : the motor habitus of a warm man, the nature of the
Results.
Legibility, atmosphere-value and forms of printing types. 10



TABLE XII.
Rank-orders of the 13 booktypes for both kinds of subjects.
H
« . . Q
» >3
OP m «3 *r o
p3 M , 5 CQ t> . o£cj
◄ _ » M - .2 3 p w £ _ fc ^ a
«5 h 3 H fc -4 g g OS ^ CQ
fc ^ fc O H o W O g g ” *< J ®
P$ S ◄ ft Eh S P< H g g 2 O 5 ^
H O a o P M w W O O o o Ph
m« h » j m o n g g ffl • >4 « o
1 2 3 4 5 6 7 8 9 10 11 12 13
Adventures 84 5 6 4 3 8J 13 1 8J 8J 11 12 2
Technics 12 5 2J 10 24 84 84 4 ' 7 1 13 6 11
Farmers’ Life 12 3 4J 11 8 10 1 7 9 6 13 4} 2
Tragedy 9 6} 9 3J 11 64 12 5 9 1 13 3J 2
Satire 1 9 11 3 4 5 13 8 6J 6J 2 12 10
Analysis 12 7 4 9 1 5 10 3 7 2 13 7 11
Report 12 84 5 84 1 7 10 2J 4 24 13 6 11
Poems 1 10 11 3 6 4 13 5 8 7 2 12 9
Strength 124 5£ 3$ 11 8 10 1 7 9 54 124 34 2
Economy 12 7J 3i 10 1 7J 9 2 6 5 13 34 ' 11
Luxury 2 10 10 4 6 3 13 6 8 6 1 12 10
Dignity 1 6J 6J 3 10J 4 13 8 12 104 2 9 5
Precision 1 10J 9 3 5J 4 13 8 7 54 2 104 12
Hygiene 12 7 5 9 3 5 8 2 10 1 13 5 11
Reliability 12 74 1 11 4 10 2 74 54 54 13 3 9
Comfort 12 24 5 11 9 10 4 64 64 24 13 8 1
lines of a hygienic, fresh thing, the visual and tactile equivalent of the black, aromatic substance, called “coffee”, etc.
Schiller called attention to shift in feeling-tone of automobiles. Poffenberger and Franken found a high negative correlation between automobiles and luxury, which became positive in Schiller’s experiments. Indeed, an article as automobiles can be appreciated in several ways : as an efficiënt machinery, as a graceful luxury etc. The graphic expression of these different appreciations will vary accordingly from long, full flowing lines to short, quick, sturdy ones.
The fundamental principle of explication of the feeling-tone by way of visual analogies is essentially not different from the characterological method, because the personality of the writing subject is also found in his outward appearance. For the character of a man can be read as well from his stature and physiognomy, as from the graphic residues of his motor habitus.
Hence there can be no objection if we speak at a breath of the feeling-tones of poems, hygiene and coffee. In the analyses on the following pages the figures before the lines correspond with those of Table XXI. These traits are not mentioned or the significance is not added when the determination was impos-



TABLE XIII.
Average scores of the 17 display-types for the subjects for advertising.
The mean deviations are printed in light type beneath the scores.
* „ § a B B
8S S 2 . § g 5 § ® te§ g * ao
«6; a i 3 g 3 « * I o got-ggHpqS
Kto j H M o ca O g Q pd a ca >. kT t; q] 2 * h
««3 9 S o w m S h o 3 ◄ ^ 53 a £ 9 w q*
o W 3 M M o n M 0*Spa 3 3 3
pqö C5 W ^ £ {=] O H CQ ^ fi Hoq <1 O
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Force, 1.0 1.5 3.3 3.2 2.7 1.3 4.3 3.1 1.7 3.7 2.0 2.3 4.3 1.2 4.3 3.7 2.5
«TRENGHT 0.0 0.6 1.0 0.7 0.8 0.5 0.6 0.9 0.6 0.8 0.7 0.7 0.6 0.3 0.6 0.7 0.8
Economy, 1.3 1.0 2.6 3.7 1.9 2.0 2.3 1.6 2.0 2.1 1.9 3.0 3.1 1.7 3.0 3.3 3.7
SOBERNESS 0.4 0.0 1.0 0.7 0.6 0.8 0.8 0.6 0.9 0.8 0.6 0.9 0.9 0.8 0,8 1.1 1.0
Luxury, 4.4 3.7 2.0 1.5 2.9 3.2 1.4 2.2 3.3 1.7 3.8 1.3 1.2 4.2 1.5 1.6 1.9
«EFINEMENT 0.6 0.9 0.8 0.6 0.8 1.2 '-0.6 0.7 0.9 0.8 0.6 0.5 0.4 0.6 0.5 0.6 0.7
Distinction, 3.4 3.0 2.6 1.9 3.1 3.3 1.7 2.9 2.9 2.2 3.5 1.9 1.8 3.6 1.9 2.1 2.3
DIGNITY 0.8 1.1 1.1 0.6 0.9 0.9 0.7 0.7 1.0 0.8 0.8 0.8 0.8 0.9 0.7 0.8 1.0
Precision, 3.5 2.2 1.9 1.9 2.5 3.3 1.3 1.8 2.9 1.7 3.1 1.7 1.5 3.0 1.7 1.8 2.5
FINENESS 0.8 1.0 0.7 0.7 0.9 1.1 0.4 0.6 0.9 0.6 0.8 0.7 0.6 0.9 0.7 0.7 1.1
Hygiene, 2.2 1.5 2.6 3.7 2.6 2.1 2.1 2.0 3.0 2.2 2.5 1.9 3.2 2.2 2.7 3.6 3.7
FRESHNESS 0.7 0.7 1.0 0.7 1.0 0.9 0.7 0.8 1.2 1.0 0.9 0.8 0.9 1.2 0.9 0.8 0.8
Reliability, 1.2 1.1 2.5 3.9 2.4 2.0 2.9 1.9 1.7 2.7 1.8 2.7 3.9 1.5 3.4 3.7 3.4
SOLIDITY 0.3 0.2 1.0 0.6 0.8 1.0 0.9 0.7 0.7 0.9 0.6 0.8 0.5 0.6 0.8 0.8 1.0
kvARMTH, 2.2 2.0 2.1 2.2 2.9 1.9 2.4 2.9 1.8 2.6 2.8 1.4 2.5 2.3 2.7 2.7 2.1
■IMFORT 0.9 0.9 0.8 0.8 1.0 0.8 0.9 0.9 0.7 1.0 0.9 0.7 1.0 1.0 0.9 0.9 0.9
TABLE XIV.
Rank-orders of the 17 display-types for the advertising-subjects.
rf
Sm' §
_ j o ES s SJ
PS CG S O te ''H o O <5 o
<1 te '«Ira'a» je ® O w te te 2 rK
•W 3 g 3 2 3 ◄ ◄ o o § S O « °
5“ t-4 £ « ° w 0 S o § 5 te > £ Ü U b 5 **
0g 5 § § 5 8 S s 3 ü I 33 t §s S § gft'
fpo o w a z h o w " a S zn § Ss. 3 8
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
■trength 17 14 6 7 9 15 3 8 13 4 12 11 2 16 1 5 10
fcoNOMY 16 17 7 2 12$ 11 8 15 10 9 12$ 5 4 14 6 3 1
Luxury 1 4 9 13 7 6 15 8 5 11 3 16 17 2 14 12 10
Dignity 3 6 9 15 5 4 17 8 7 11 2 14 16 1 13 12 10
Precision 1 8 9 10 6$ 2 17 11$ 5 14$ 3 14$ 16 4 13 11$ 6$
Hygiene 12 17 7 2 8 13$ 13$ 15 5 11 9 16 4 10 6 3 1
Reliability 16 17 9 1 10 11 6 12 14 8 13 7 2 15 4$ 3 4$
Comfort 11 14 13 10 1 15 8 2 16 6 3 17 7 9 5 4 12



TABLE XV.
Specificity of the subjects.
Correlations between the subjects for books and advertising, computed from the rank-orders of Table XII and Table XIV (the latter in the upper half of the scale, in the lower right corner). The figures represent hundredths.
Pm k.
oq H m E
5 s ^ S IE „ a
BffSK-.gg., * £ g s g « 5
£ 2 " 3 3 * fi P P Ë E £ 5 £
Romantic Adventures Mod. Technics and Architecture 26 Frank and rough Farmers’ Life 46 40
Stately, classic Tragedy 37 24 26
WlTTY, SPIRITUAL SATIRE 14 —37 91 —13
Clear Analysis Probl. Culture 37 93 18 25 —11
Matter-of-fact Report 34 89 14 24 11 96
Delicate, sensitive Poems 22 —43 —90 — 7 95 13 17
Force, stoength 12 45 99 32 -97 23 19 -87 60 -82 -88 -89 33 79 _46
Economy, soberness 29 88 29 17 —27 91 94 35 32 . 
Luxury, refinement 14 -19 -90 -11 95 -14 -17 98 -87 -30 96 89 -38 -88 80
Distinction, dignity 17 -57 -82 5 62 -50 -62 68 58 _61 68 26 _J3
Precision, fineness — 1 —29 —95 —15 95 9 iu ^ Q 48 n
Hygiene, freshness 15 88 40 42 20 88 78 24 46 8 
Reliability, solidity —32 71 69 — 4 75 54 58 84 71 70
Warmth, comfort 23 47 88 42 -73 28 24 -74 90 26 -73 -53 -80 38 53



TABLE XVI.
Rank-order of appropriateness for each of the booktypes separately.
BS g
oq j £ B
« & y < £ O
ÈS<= 2 ë g
«2 K 2 s oWosI2-<»2
H *4 ^ O H
HO B O O -ajP^H^030OP?
n« h n j
1 2 3 4 5 6 7 8 U 10 11 12 13
Adventures 6 5 4 6.5 5 4.5 3.5 6 5 3 6 3 7
Technics 3 7 8 2.5 6 6.5 7 5 7 7 4 7 5
Farmers’ Life 1.5 8 6.5 1 4 1.5 8 4 3.5 5 1.5 8 8
Tragedy 5 4 3 6.5 2 4.5 5 3 3.5 4 5 4.5 6
Satire 7 1.5 1.5 7.5 3 3 1 1.5 2 2 7 2 1
Analysis 4 6 6.5 4 786878364
Report 1.5 3 5 2.5 8 6.5 3.5 7 7 6 1.5 4.5 3
Poems 8 1.5 1.5 7.5 1 1.5 2 1.5 1 1 8 1 2
Strength 167151854 4.5 1.5 7 8
Economy 3 3.5 6 5 8 3.5 4 6.5 7 6 3 6 4
Luxury 8 1.5 1 3.5 1 3.5 1.5 1 1 1 8 1 2
Dignity 6 3.5 3 8 2 6 3 3.5 2.5 2 6 3.5 5
Precision 7 1.5 2 7 3.5 6 1.5 2 2.5 3 7 2 1
Hygiene 5 6 5 6 686858553
Reliability 2 8 8 3.5 7 6 7 6.5 8 7 1.5 8 6
Comfort 4 6 4 2 3.5 2 5 3.5 6 4.5 4 3.5 7
TABLE XVII.
Rank-order of appropriateness of each of the display-types separately.
» M*
g s *■
. rt a ® m o M
§ o S»
S «3 S S g t* S g* « 3 Hs
Hl ^ ö P$ O H M << EH q H ◄ ^ ^ B p« o 3 EH
H p ◄ © ® ^ * * O H W O H tf o O 2 O
fQü fi Eh cq O h3C
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Strength 1 3.5 8 5 5 1 8 8 1.5 8 3 6 8 1 8 7.5 4.5
Economy 3 1 6 6.5 1 3.5 5 1 4 3 2 8 5 3 6 5 7.5
Luxury 8 8 2 1 6.5 6 2 5 8 1.5 8 1 1 8 1 1 1
Dignïty 6 7 6 2.5 8 7 3 6.5 5.5 4.5 7 4.5 3 7 3 3 3
Precision 7 6 1 2.5 3 8 1 2 5.5 1.5 6 3 2 6 2 2 4.5
Hygiene 4.5 3.5 6 6.5 4 5 4 4 7 4.5 4 4.5 6 4 4.5 6 7.5
Reliability 2 2 4 8 2 3.5 7 3 1.5 7 1 7 7 2 7 7.5 6
Comfort 4.5 5 3 4 6.5 2 6 6.5 3 6 5 2 4 5 4.5 4 2



sible or too uncertain. The analysis itself is an attempt to bring the quaüty into a System which allows a comparison with that af the type faces. Each time it is followed by the names of the types that were selected by the subjects ; the types are compared with other designs, that possess about the same characteristics, though they may be quite different in structure and artistical quaüty.
TABLE XVIII.
Specificity of the Type Faces.
Correlations between the rank-orders of the 13 booktypes for the subjects for books. The figures represent hmdredths.
1 2 3 4 5 6 7 8 9 10 11 12 1!
Bernhard Roman 1
DlANNHAUSER 2 65
Bodoni 3 —77 92
Lutetia 4 95 —78 —95
Saeculum 5 —63 41 72 —46
Drpheus 6 —35 27 56 — 4 81
Beton 7 —72 96 87 —88 35 28
Mediaeval 8 —63 41 55 —60 81 70 28
Modern Wide 9 —63 55 81 —66 93 91 53 83
Bembo 10 —77 68 90 —85 81 81 73 61 90
Locarno 11 98 —63 —73 92 —71 —39 —70 —70 —63 —79
Romaansch 12 —84 92 91 —97 50 36 97 31 61 81 —82
Grasset 13 —46 85 53 —53 10 — 2 74 30 27 29 —44 64
TABLE XIX.
Specificity of the Type Faces.
Correlations between the rank-orders of the 13 booktypes for the subjects for advertising. The figures represent hundredths.
1 2 3 4 5 6 7 8 9 10 11 12 13
Bernhard Roman 1
Thannhauser 2 —72
Bodoni 3 —95 85
Lutetia 4 58 —40 —41
Saeculum 5 —72 59 84 —11
Orpheus 6 40 5 — 8 79 15
Beton 7 —89 92 90 —56 64 —15
Mediaeval 8 —64 73 81 — 1 87 34 73
Modern Wide 9 —74 77 83 —27 86 4 65 77
Bembo 10 —63 75 78 —11 89 33 72 95 84
Locarno 11 99 —74 —96 55 —73 35 —86 —65 —78 —65
Romaansch 12 —92 83 99 —33 83 — 2 89 81 79 75 —92
Grasset 13 —81 75 67 —64 24 —52 81 33 48 25 —78 65



TABLE XX.
Specificity of the Type Faees.
Correlations between the rank-orders of the 17 display-types for the subjects for advertising.
The figures represent hundredths.
, 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Bernhard Cursive 14 Gallia IS 84
Kontrast 16 —65 —48
Mediaeval 17 —64 —90 36
Narciss 18 53 80 2 —66
Excelsior 19 83 60 —47 —34 — 2
Omega 20 —92 —66 62 70 —19 —92
Eckmann 21 —10 42 38 —29 78 —52 35
Shadow Nobel 22 85 62 —30 —53 31 74 —83 —15
Metropolis 23 —79 —43 62 58 8 —74 91 56 —76
Locarno 24 90 97 —51 —92 75 64 —76 31 74 —57
Narrow Heavy Gothic 25 —77 —78 67 80 —64 —35 56 —50 —54 43 —84
de Vinne 26 —95 —70 74 85 —39 —73 89 16 —70 85 —85 73
Trafton Script 27 97 87 —63 —82 59 75 —85 —22 80 —71 93 —71 —95
Arpke 28 —97 —82 71 83 —43 —79 93 12 —82 84 —89 75 96 —96
Columbia 29 —92 —79 70 89 —41 —70 87 11 —70 84 —88 75 99 —94 96
Light Gothic Exp. 30 —52 —81 48 83 —74 —19 28 —56 —17 17 —73 80 62 —63 56 65



TABLE XXI.
Graphological scheme for printing types and decorative lettering. After Klages (modified).
L L?vel ?f form 1- Very high; 2. Fairly high; 3. Medium; 4. Low; 5. Very low.
(Formmveau)
2. Ductus 1. Very high; 2. Fairly high; 3. Medium; 4. Low; 5. Very low.
3' reu”™ . Rhythmic: 1 + Depth, Harmony Unrhythmic: l + Openness, Delicacy
enmass) ow affectivity j _ Bluntness, Apathy High affectivity / _ Flatness, Irritability
4' R;GLTARITY i Regular-Preponderance S + Will-power Irregular-Preponderance k + Liveliness of sentiment
(Regelmassigkeit) of Vohtion /- Lack of Sentiment of Sentiment ) - Lack of willpower
e Fm i v„c ( + phantasy, „Anschauungsvermögen“, , T . „ . , r
(Völlel Full < Imagination, Impressionability Thin \ 1 ^n^c ectua Eorce
( — Intellectual weakness ^ Lack of Phantasy
6. Enrichment Creatiye in |se S + Sense of beauty Simplification \ + Efficiency’ Sense of order
(Bere.cherung) p ) _ Exaggeration Simplihcation j _ ^ Qf sense Qf bgauty
a) Volition. High ) + Wüi-power Low S + Susceptibility
7. Pressure ' — Heavy on hand ( — Lack of will-power
(Druckstarke) Sentiment Hioh + Impulsiveness S Stub- \ + Excitability ( ..
b) Sentiment. H,gh ^ _ Vehemence ( Borflness Low < __ Lability S Subjectivity to d.stractions
8. Pastiness S + 1“'InstSci8ina,ity’ Attachment t0 the $ + Spirituality, Sensitivity, Self-command
(Teieiekeif) ) ...... Sharpness . — Intellectualism
( Lack of spirituality and selfconstraint ( Incapacity for sensual pleasure
9. Vv'idtii Wide \ ^ Unconstraint, Activity k + Self-constraint
1 Swpex&ciaVVlv» Loaccuracv, \ m\'.r-.veY\ee arrow } —\_/aek oï diiectness, \nx\ousness
r ( + Ambition I + Content, Discretion
Long ) —Discontent t — Indifference, Apathy
10. Length of / Capacity for enthousiasm D >A: ( + Realism, Accuracy
descenders (D) and A > D: \ Abstract thinking Lack of ) Powers of observation
ascenders (A) Mobility of < _ olin„rfiria|jtv Mobility of ) — Materialism, Hard-headedness
the Mind f Volatüity the Mind ( Lack of intellectual skill
. k + Activity , X + Tranquillity
a) Volition. High ^ Unrest °W f — Inactivity
11. Speed t -(- Liveliness T S 1 Equanimity
b) Sentiment. High } _ Excitement Low ) -Bluntness
1 + Preponderance of the social sentiments \ + Preponderance of Intellect,
1? TiMriiNATioN < Capacity for self-delivery and enthousiasm Steepness Reserve
12. ( — Rashness < -Coldness, Heartlessness
/ + Powers of combination, Versati- / _i_ Wealth of idea’s, Independence of
( lity of the mind, Systematic and l 1 judgment, Quick-wittedness, Realism,
13. Connection Undisturbed speculative intellect Gaps in the ' Powers of observation
(Verbundenheits- connection \ — Poverty of thoughts, Dependence connection 1—Lack of logical sense, Jumpiness,
grad) / of judgment, Flight of thoughts, f Lack of adaptability and abstraction
\ Incapacity for observation.
/ + Versatility, Adaptability ( "f Resistance, Constancy
Wire ’ — Suggestibility Angle / — Uncompromisingness, Lack of harmony,
/ Lack of character. Mimicry ' Irritability
t4. Nature ofconnec- 0 .
TION i Goodness, Sympathy, \ + Reserve, Distinction
Garland , Unconstraint, Indulgence Arcade ^ Lack of frankness, Conventionalism
( — Dependence, Instability, Suggestibility
p.THn. t + Capacity for enthousiasm , + Realism, Thoroughness
PATHOS OF , ’ Depth of sentiment nru„Kt cPveri,v
sentiment ) ( — Dryness, Severity
I { — Exaltation
15. S.ZE PATHOS OF l . \ + Enterprise, Energy ( + Powers of concentration, Precision
volition l Earge , _ Lack of powers of concentration, smaii . _ jqarrow-miruledness
l i Absent-mindedness
Pathos of S + Pnde, Dignity \ + Humility, Unpretentiousness
self-esteem I l — Haughtiness, Arrogance \ —Lack of faith> Anxiousness



DESCRIPTION OF THE QUALITIES.
a. Romantic Adventures.
Associations with Byron, Robinson Crusoe. Slightly oldfashioned; can be narrative. No positive correlations with other qualities. The slight ditference in the absolute values (1.0) makes it probable that this quality was found too vague.
1. Level of form 1—2\ 7. Rather high pressure
2. Ductus 1—3£ 8. Pasty
3. Unrhythmic 9. Wide
4. Rather irregular 10. Long projectors, A >D
5. Full 11. High speed
6. Enrichment IS. Large
1. Mediaeval, which is more appropriate for two other qualities.
2. Grasset, which would be my own choice ; it ranks higher for “farmers’ life”.
3. Saeculum, probably owing to the points 4, 5, 6, 8, 9. Beton is too orderly and efficiënt.
b. Modem Technics and Architecture.
More or less the spirit of le Corbusier. Positive correlation with analysis, report, economy, hygiene, reliability.
1. Level of form can be low (one-sided 8. Not too sharp (architecture sup-
spiritual) poses “ Anschauungsfreude").
2. Ductus 1—4 12. Steep
6. Simplification 13. Interrupted connection
7. Fairly high pressure 15. Smalt
1. Bembo, 2. Saeculum-Bodoni, 3. Thannhauser, 4. Romaansch. Beton is probably too large : too little realism, precision, concentration.
c. Frank and rough Farmers’ Life.
This kind of literature, by Dutch and Scandinavian authors, is very popular in Holland. Positive correlations with strength, reliability and warmth.
1. Level of form 1—3 8. Pasty
2. Ductus 2—3i 9. Wide
3. Rhythmic 10. Rather short projectors, D > A
5. Full 11. Low speed
7. High pressure 15. Large



1. Beton, 2. Grasset, 3. Thannhauser-Bodoni-Romaansch,
4. Bembo.
Saeculum is probably a little too spiritual. Bembo is spiritual too, but has the necessary “swing”. The position of Bodoni is due to its strength.
d. Stately, classic Tragedy.
No positive correlation with any other quality, extremely small difference in valuation (0.7).
The character belonging to this subject must be thought as harmonious, combining a strong will and a deep sentiment in a perfect balance. The properties, constituting an antithesis in Klages’ system, have to be present to about the same extent,. which makes a description extremely difficult.
1. Level of form 1—2 7. High pressure
2. Ductus 1—2 8. Not too sharp
3. Rhythmic 10. Long and equal projectors
5. Fairly full 15. Fairly large
1. Bembo, which ranks considerably higher for other qualities, but fits the description excellently; 2. Lutetia, Romaansch, Grasset. The latter two contain some appropriate features (portliness, heaviness), but do not seem very adequate in the other respects. Mediaeval shows too much “Sinnenfreudë'.
e. Witty, spiritual Satire.
About the spirit of Yoltaire, Heine etc. Positive correlation with poems, luxury, precision, dignity. The correlation with luxury and dignity is due to the sensitivity and fineness of Bernhard Roman and Locarno.
1. Level of form 1—3 8. Sharp
2. Ductus 1—2 9. Wide
4. Rather irregular 10. Long projectors, A > D
5. Fairly full 11. High speed
7. Low pressure 13. Interrupted connection
1. Bernhard Roman, 2. Locarno, 3. Lutetia.
Grasset and Beton have their lowest scores for this quality.
f. Clear Analysis of Problems of Culture.
The spirit of Ortega y Gasset, Huizinga etc. Positive correlation with report, economy, hygiene, technics and reliabiüty.
1. Level of form 1—2£ 6. Rather simplified
2. Ductus 1—2t 8. Not too sharp
3. Rhythmic 9. Not too wide
4. Regular 12. Steep
5. Not too thin



1. Bembo and Saeculum, 2. Mediaeval (highest score of this type), 3. Bodoni, 4. Orpheus (highest score).
Though Saeculum, to my opinion, is a type very much inferior to Bembo, it ranks here equally with this face, though the latter has its highest score for this quality. The high correlation with matter-of-fact report shows that my subjects underrated the properties of spirit and character needed for this quality.
g. Matter-of-fact Report.
Positive correlation with analysis, economy, hygiene, reüability, technics.
1. Level of form 3—4 9. Rather narrow
2. Ductus 2—4 10. Short projectors, D > A
3. Rhythmic 11. Low speed
4. Regular 12. Steep
5. Thin 13. Uninterrupted connection
6. Simplified 15. Small
7. Rather high pressure
1. Saeculum, with a remarkable predilection, 2. Mediaeval and Bembo.
Modem wide shows again an average score. Romaansch is too large (too much pathos).
h. Delicate, sensitive Poems.
Character of the lyrical poet. Positive correlation with luxury, dignity (cf. satire), precision, satire.
1. Level of form 1—2 8. Not too sharp
2. Ductus 1—2 9. Wide
3. Unrhythmic 10. Long projector, A > D
4. Irregular 11. High speed
5. Full 12. Rather sloping
6. Enrichment 14. Garland
7. Low pressure 15. Large
1. Bernhard Roman, 2. Locarno, 3. Lutetia, 4. Orpheus.
i. Force, Strength.
Characterological equivalent: will-power, calm, unsentimental; high, broad-shouldered stature (military and policemen). Analogy with strong construction, capacity for bearing a heavy load.
Positive correlation with reüability, warmth, farmers’ life, economy; Poffenberger and Franken, Schiller : with automobiles, building-material, cheapness, coffee.
1. Level of form 1—3i 4. Regular
2. Ductus 1—4 7. High pressure
3. Rhvthmic



1. Beton (highest score), 2. Grasset, 3. Romaansch (higher score for reliability) and Bodoni (id.).
1. Omega, de Vinne, Arpke (highest score), 2. Metropolis and Columbia.
Poffenberger and Franken (pf) : 1. Globe Gothic Bold, 2. Antique Bold, 3. John Hancock, 4. Masterman.
Schiller (s) : 1. Cheltenham Bold, 2. Broadway, 3. Futura Black, 4. Scotch.
Da vis and Smith (ds) : Bold, non-italic, expanded, large.
Evidently, boldness alone cannot make up strength ; a certain tension must be present too. For example : Metropolis is blacker than de Vinne, but lacks the decisive force ; Columbia likewise. Schiller’s Futura Black is immediately followed by the far less heavy Scotch. (Her “Heavy Gothic” was probably too small (8 pt.) to compete with the others).
j. Economy, Soberness.
Spirit of the saving, middle-class, small business-man orfunctionary. Positive correlation with analysis, report, technics, hygiene, reliability, strength (display types only). s : with automobiles, building material, cheapness, coffee.
1. Level of form 2—4
2. Ductus 1—4
3. Rhythmic
4. Regular 6. Simplified
7. Rather high pressure
8. Rather narrow
10. Short projectors, D > A
11. Rather steep 15. Small
1. Saeculum (highest score), 2. Bodoni, Mediaeval, Romaansch, 3. Modem Wide, Bembo.
1. Mediaeval, Light Gothic Expanded, 2. Columbia, 3. de Vinne, 4. Narrow Heavy Gothic, Arpke.
pf : 1. Bookman, 2. Century, 3. Scotch, 4. New York Gothic, 5. Della Robbia, 6. Caslon 471. s : 1. Goudy, 2. Cheltenham, Parisian, Scotch. ds : Normal Width.
Apparently this quality needs a simple, resolute, light type. My somewhat unfavourable characterological appreciation is not entirely shared by the subjects, though this time’s higher score of Modern Wide points in the same direction.
k. Luxury, Refinement.
The spirit of idle, spoiled, rich people, of aesthetic attitude. Positive correlation with dignity, precision, satire, poems. pf, s : with dignity, jewelry, perfume.



1. Level of form 2—4 7. Low pressure
2. Ductus 1—3 8. Not too sharp
4. Rather irregular 9. Wide
5. Full 10. Long projectors
6. Enriched 15. Large
1. Locarno, 2. Bernhard Roman.
1. Bernhard Cursive, 2. Trafton Script, 3. Locarno, 4. Gallia, 5. Shadow Nobel (highest score for all these types).
■ pf : 1. Tiffany Text, 2. Engravers Roman, 3. Typo Slope, 4. Old English, 5. Priory Text.
s. : Bernhard Cursive, 2. Washington Text Shaded, 3. Bernhard Cursive Bold, 4. Clear Cut Initials. ds : Non-bold, italic, normal width, small.
If pf had included in their hst such types as Bernhard Cursive (which did not exist at that time), they would have superseded the old-fashioned script types without doubt. We call attention to the fact that Eve (Locarno) did not reach the highest scores in Schiller’s inquiry. The results of ds show the disadvantages of their one-sided choice of types.
1. Dignity, Distinction.
The spirit of the cultured, noble family. Positive correlation with luxury, satire, poems. pf, s : with luxury, jewelry, perfume.
1. Level of form 1—3£ 8. Rather sharp
2. Ductus 1—2£ 9. Fairly narrow
3. Rhythmic 11. Long projectors A >D
4. Rather regular 12. Rather low speed
5. Rather thin 13. Undisturbed connection
6. Not too exuberant 15. Large
7. Medium pressure
1. Bernhard Roman, 2. Locarno, 3. Lutetia, 4. Orpheus.
1. Trafton Script, 2. Locarno, 3. Bernhard Cursive, 4. Excelsior, 5. Narciss.
pf : 1. Priory Text, 2. Tiffany Text, 3. Engravers Roman, 4. Old Engüsh, 5. Typo Slope.
s : Washington Text Shaded, 2. Caslon Text, 3. Bernhard Cursive Bold, 4. Bernhard Cursive. ds : Italic, Normal width, Medium size.
It has to be noted that this quality needs more reserve than ‘luxury” did (cf. Bernhard Cursive). The Anglo-Saxon predilection for Text (Gothical) types on formal occasions, also apparent here, is unusual in other countries.
m. Precision, Fineness.
The spirit of the craftsman, whose job it is to make fine instru-



ments, clockworks etc. Positive correlation with satire, poems, luxury, dignity.
1. Level of form 3—4 7. Rather low
2. Ductus 2—4 8. Rather sharp
3. Rhythmic 9. Rather narrow
4. Regular 11. Low speed •
5. Rather thin 12. Undisturbed connection
6. Simplified 15. Small
1. Bernhard Roman, 2. Locarno, 3. Lutetia, 4. Orpheus.
1. Bernhard Cursive, 2. Excelsior, 3. Locarno, 4. Trafton Script. ds : Non-italic, condensed, small.
The choice of the subjects was apparently based upon sharpness, low pressure, small size, and long projectors (for delicacy). My own favourite for this quality would be Corvinus. Long projectors are not necessary.
n. Hygiene, Freshness.
Analogy with the lines of modern factory-buildings, bathrooms etc., clear, airy, sharp. Positive correlation with reliability, economy, technics, analysis, report.
2. Ductus 1—3 8. Sharp
5. Rather thin 9. Wide
6. Simplified 10. Rather short projectors
7. Fairly high pressure 15. Small
1. Bembo (highest score), 2. Mediaeval (highest score), 3. Saeculum, 4. Bodoni, Orpheus (highest score), Romaansch.
1. Mediaeval, Light Gothic Expanded (highest score), 2. Columbia, 3. de Vinne, 4. Shadow Nobel. ds : Non-bold, expanded, small.
The values are comparable to those of “economy”, except Orpheus, which ranks higher here, and Modem wide, which ranks lower.
The generous amount of white around Columbia has been to its advantage ; the lack of it lowered the score of Narciss. We would formulate the necessary features for this quality as resolute, precise, thin, clear, simple, not too graceful.
o. Reliability, Solidity.
Very much the same spirit as required for “strength”, but more with the characterological consequences of absence of whims, “moods”, lability, phantasy, and without its heaviness and agressivity. Positive correlation with strength, economy, technics, farmers’ life, analysis, report, hygiene (booktypes only).



1. Level of form 1—4 7. Fairly high pressure
2. Ductus 1 4 10. Short projectors D = or > A
3. Rhythmic 11- Rather low speed
4 Regular 12. Undisturbed connection
6. Simplified
1. Bodoni (highest score), 2. Beton, 3. Romaansch (highest score), 4. Saeculum, 5. Modem wide, Bembo, 6. Mediaeval, Thannhauser.
1. Mediaeval (highest score), 2. de Vinne, 3. Columbia (highest score), 3. Arpke, Light Gothic Expanded. ds : Bold, non-italic, condensed, medium size.
p. Warmth, Comfort.
Analogy with warm and comfortable clothes, fumiture etc. : thick, soft, round, flowing, slow lines. Positive correlation with farmers’ life, strength.
2. Ductus 1—4 9. Wide
5. Full 11. Low speed
6. Enriched 14- Garland
8. Pasty
1. Grasset, 2. Thannhauser, Bembo.
1. Narciss, Eckmann, 2. Locamo, 3. Arpke, Columbia. ds : Non-bold, normal width, small.
The small diflference between the values shows that the subjects found this quality difficult to describe.



DESCRIPTION OF THE TYPE FACES
1. Bernard Roman
Eê ia Jute fiocholu minegiten egiak, kutsean Jenari kankat gaicktoago keren falta, Jute eta kalere gure gain eraan naki zintzur kontra iretsi kekarko, kalinkeJute ere ir ets Jezatela egia okoniek okoreak Jituz
2. Ductus 2 8. Sharp: spiritual, sensitive
4. Irregular: Liveliness of sentiment 9. Medium width
5. Full: Empressionability, “ An- 10. Long projectors : refined
schauungsvermögen” 11, Fairly slow : inactivity
6. Enriched: sense of beauty 12. Steep (e.g.: g) : reserve
7. Moderate pressure: refinement, labil- 15. Small (actually) : precision
ity, lack of will-power (naturally fairly large : haughtiness)
Positive correlation with Locamo, Lutetia. Comparable : Announcement Roman, Bernhard Fashion, Egmont, Ehmcke Antiqua, Feenhaar, Klingertype, Lipsia Antiqua, Metropolis Mager, Mona Lisa, Moreau le Jeune, Naudin, Nic. Cochin, Nova, Olympia, Pabst and related designs, Paganini, Piranesi, Stahlstich Antiqua, Stellar Light.
2. Thannhauser
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra eretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek oho-
2. Ductus 3 8. Pasty : sensual, sense for colour
4. Moderately regular : sentiment, will- 9. Fairly wide; broad-mindedness
power ^ 10. Rather short projectors: lack of
5. Full: “Sinnenfreude”. ambition, materialism
6. Simplified (though still too full) : 11. Low speed : inactivity
orderly 12. Rather steep : intellect, reserve
7. Fair pressure: heavy on hand 15. Large: broad-mindedness
Positive correlation with Bodoni, Saeculum (advertising), Beton, Mediaeval (advertising), Modern Wide, Bembo, Romaansch, Grasset, negative corr. with Bernhard Roman, Lutetia, Locamo. Comparable : the modem sans-serif types, egyptians, in general every simple, somewhat pasty roman, e.g. Mundus Antiqua.
Legibility, atmosphere-value and forms of printing types. 11



3. Bodoni
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nabi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek
2. Ductus 2 9. Fairly narrow: accuracy, self-com-
4. Regular : will-power mand
5. Moderately full: “ Anschauungs- 10. Long projectors: interest for cul-
vermögen” and intellect ture, refinement
6. Simplified : efficiency, sense of order 11. Low speed : quiet, level headed-
7. High pressure : will-power, stubborn- ness
ness 12. Steep: preponderance of intellect
8. Bold and sharp : sensitivity, “Sin- 15. Small: thoroughness, precision nenfreude”
Positive correlation with Saeculum, Beton, Mediaeval, Modem Wide, Bembo, Romaansch, Grasset, Thannhauser. Comparable : all types of the Didot origin with pronounced contrast. About the same emphasis and precision, but not the delicacy and reserve, can be found in New Caslon (“Plantin”), Série 18—21 of Deberny & Peignot (“Baskerville O.S.” of H. W. Caslon & Co.).
4. Lutetia
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbadute ere, ir ets dezatela egia ohoinek ohoreak dituz eno izanen da
2. Ductus 2 8. Sharp : sensitivity, intellectualism,
4. Fairly regular i not giving way to self-command
emotions 9. Narrow on the whole: “attitude”
5. Moderately full: impressionability, 10. Long projectors, D > A : interest
aesthetic attitude for culture, sense of beauty
6. Mostly simplified : orderly, precise 11. Moderate speed : quiet, contempla-
7. Moderate pressure : susceptible tive
15. Small: precision
Positive correlation with Bernhard Roman, Orpheus (advertising only), Locarno. Comparable : the sharp, regular, spiritual roman, e.g. Centaur, Egmont, Erasmus, Tiemann Mediaeval (Wren); the latter lacks already the reserve of Lutetia.
5. Saeculum
Egia dute pocholu minegiten egiak hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek ohoreak dituz eno



1. Level of form 3—4 8. Rather pasty : some longing for
2. Ductus 3 sensual pleasure, but not unspiri-
4. Fairly irregular: fair liveliness of tual
sentiment 10. Short projectors: lack of culture,
5. Moderately full: “Anschauungsver- materialism
mogen”, intellect 14. Angular : some obstinacy
6. Simple forms: efficiency, sense of 15. Smalt: dryness, narrow-minded-
order, lack of sense of beauty ness, unpretentiousness, thorough-
7. Fair pression : action ness, precision, powers of concen-
tration
Positive correlation with Bodoni, Orpheus, Mediaeval, Modem Wide, Bembo, Romaansch, for advertising with Thannhauser and Beton. Comparable : every clean-cut, dry and somewhat pasty roman, e.g. Ionic and its derivates (modern newspapertypes), Cheltenham O. S., Monotype Veronese, Montaigne, Century.
The judgment of the subjects has been more favourable than I expected on the score of my own appreciation. The small size shows this type to its advantage.
6. Orpheus
Egia dute pocholu minegiten egiak, hutsean denari ham bat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek ohoreak dituz eno
2. Ductus 2 9. Fairly wide: unconstraint, easy
4. Fairly regular: not giving way to social manners
emotions 10. Rather short projectors, A > D :
5. Moderately full: “ Anschaunngs- content, joy in earthly pleasures,
vermogen” interest in culture
6. Mostly simplified: sense of order, 11. Fairly low speed: quiet, little
tidiness activity
7. Fairly low pressure: susceptible, 15. Actually small, naturally large:
little will-power unoffensive, quiet dignity, depth
8. Not without sharpness: culture, of sentiment
sense of beauty
Positive correlation with Bodoni, Mediaeval, Modern Wide, Bembo (books only), Lutetia. Comparable: the regularized revivals of old style types, with little contrast, like they appeared on the line-composing machines, e.g. Linotype-Baskerville, Cloister Lightface, Kenntonian etc. The light, even colour and moderately full forms are essential.
The rather low scores of this type, which it does not quite deserve, are due to the small size.



7. Beton.
Egia dute pocholu minegiten egiak, hutsean denari hanbal gaichtoago beren falla, dute eta halere gure gain eman nahi zintzur konlra iretsi beharko, balinbedule ere, ir ets deza-
2. Ductus 3 (it is 2 in the heavier grades) 7. Fairly high pressure: will-power
t. Moderately regular : more will than 8. Pasty : lack of spirituality and
sentiment sensitivity
5 Fairly full: “ Anschauungsvermö- 9. Wide: activity, directness
gen”, not intellectual 10. Fairly long projectors: ambition,
6. Simplified, tendencies to enrichment: enthousiasm, not vulgar
sense of beauty not absent, predo- 15. Large: enterprise minant efficiency, sense of order
Positive correlation with Thannhauser, Bodoni, Saeculum (advertising), Modem Wide, Bembo, Romaansch, Grasset, Mediaeval (advertising only). Comparable : all modem egyptians and sans-serif types; all the full, round, pasty, clear romans, e.g. Bookman.
8. Mediaeval
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia
2. Ductus 2 (the original foundry edi- 8. Moderately sharp : spiritual, sensition ranks higher) tive, sense of beauty
4. Fairly irregular: lively sentiment 9. Fairly wide: activity
5. Fairly full: phantasy, “ Anschau- 10. Moderately long projectors: relia-
ungsvermögen” bility, realism
6. Enriched: sense of beauty 11. Fairly high speed: activity, lively
7. Medium pressure sentiment
15. Large: energy, enterprise
Positive correlation with Thannhauser (advertising only), Bodoni, Saeculum, Orpheus (books only), Beton (advertising), Modem Wide, Bembo, Romaansch (advertising). Comparable : all the wide, open “old-style” types of recent design, with fairly even colour, e.g. Weiss Antiqua, Jost Mediaeval, Genzsch Antiqua, Monotype-Veronese, Italian O.S., Horley, Goudy O.S., Kennerley, Verona, Dolphin O.S., Monotype-Leysbourne, Littleworth ; to a certain extent also Cloister Lightface, Monotype-Plantin Light 113, Linotype-Baskerville.



9. Modem Wide
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek ohoreak dituz
1. Level of form 4 7. Moderately high pressure: heavy
2. Ductus 4—5 on hand
4. Regular : lack of sentiment 8. Sharp : dryness, lack of sentiment
5. Thin : lack of imagination and im- 9. Narrow : constraint, lack of direct-
pressionability ness
6. Traditional forms: pedantic sense 10. Short projectors, D>A: mate-
of order, lack of sense of beauty rialism, lack of intelligence, accu-
racy
Positive correlation with : Thannhauser, Bodoni, Saeculum, Orpheus (books only), Beton, Mediaeval, Bembo, Romaansch, Comparable : nearly all the anaemic 19th century types, both modem and so called “old style” or “old face”, also Cushing, Canterbury. In respect of the more favourable features : Scotch, Bulmer, Bodoni Book etc. Cf. Bodoni and Saeculum.
10. Bembo
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek ohoreak dituz
1. Level of form 1 8. Sharp: spiritual
2. Ductus 1 9. Medium width
4. Fairly irregular: lively sentiment 10 Fairly long projectors, D > A:
5. Moderately full, intellectual force, ambition, realism, accuracy, obser-
imagination vation
6. Enriched: sense of beauty 15. Medium size
7. Fairly high pressure: will-power
Positive correlation with : Thannhauser, Bodoni, Saeculum, Orpheus (books only), Beton, Mediaeval, Modem Wide, Romaansch. Comparable : all high-class old style types between Lutetia and Mediaeval, e.g. Garamond, Baskerville, Granjon, Estienne, Janson, Bell, van Dyck, Foumier, Times New Roman.
11. Locamo
Esia dute pocholu minesiten egiak, hutsean denari han= bat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur, kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek ohoreak dituz eno izanen da



2. Ductus 2—2i 8. Sharp : spirituality
4. Irregular : lively sentiment 9. Wide: activity
5. Full: imagination, capacity for 10. Long ascenders, > D: refinement,
sensual impressions lively interest in everything
6. Enriched : sense of beauty 14. Angular : constancy
7. Fairly low pressure: susceptibility, 15. Actually small, naturally fairly
excitability large
Positive correlation with Bernhard Roman, Lutetia. Comparable : Rivoli, Egmont, Nic. Cochin, Metropolis Mager, to some extent also Centaur, Erasmus ; see further Bernhard Roman.
12. Romaansch
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia ohoniek ohoreak
2. Ductus 3—3i 9. Fairly narrow : constraint
4. Regular : will-power 10. Short projectors, D > A : mate-
5. Thin: intellectual force, lack of rialism, lack of cultural interest
phantasy 12. Steep : coldness
6. Traditional forms: efficiency, sense 15. Large: swollen
of order, lack of sense of beauty Naturally rather (small: realism,
7. Fairly high pressure: heavy on narrow-mindedness, impretentious-
hand, stubbornness ness
8. Somewhat pasty: lack of spirituality
Positive correlation with Thannhauser, Bodoni, Saeculum (advertising only), Beton, Mediaeval (advertising), Modern Wide, Bembo, Grasset. Comparable : see Modern Wide, the sturdier of the quoted types.
13. Grasset
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi zintzur kontra iretsi beharko, balinbedute ere, ir ets dezatela egia
2. Ductus 3 9. Wide: activity, inaccuracy
4. Irregular : lively sentiment 10. Moderately long projectors D > A:
5. Full: phantasy, impressionability, absence of theoretical interests,
intellectual weakness lack of refinement
6. Enriched : sense of beauty 14. Garland : unconstraint, instability
7. Fairly high pressure: impulsiveness 15. Medium height: enthousiasm, lack
8. Pasty : sense of color, originality, of concentration
lack of self-command
Positive correlation with Thannhauser, Bodoni, Beton, Romaansch. Comparable: bold, lively romans of little contrast,



e.g. Goudy Bold, Cloister Bold, Cheltenham Bold, Richmond O.S. Bold, Gayton, Weiss Roman Bold, Ehmcke Mediaeval Bold, Deutsch-Römisch Bold, Clearface Bold, Jenson, Genzsch-Antiqua Narrow Semi-bold, Renaissance Semi-bold, Columbia, Hollandsche Mediaeval Bold, Grotius, Cooper, Minister Antiqua Semibold, Monotype-Imprint and -Plantin Bold etc.
14. Bernhard Cursive
COdta dute pocholu wunegiten egtafc, liuisean denari hanbai ócuchioago Leren falia, dute eia lialere gure gain
2. Ductus 11 9. Wide : superficial, easy social man-
4. Fairly irregular: some liveliness of ners
sentiment 10. Long projectors, A > D: refinement
5. Full: impressionability 11. Low speed: inactivity
6. Enriched : sense of beauty, sultry 12. Steep : reserve
7. Fairly low pressure: “spoiled” 15. Actually small: precision
8. Sharp : sensitive
Positive correlation with Gallia, Narciss, Excelsior, Shadow Nobel, Locarno, Trafton Script. Comparable : Every delicate cursive, e.g. Kleukens Scriptura, Troubadour, Bernhard Tango, Liberty, Piranesi Italic, Ariston, Gladiola, Arkona, Junior, Rheingold, Euphorion, Deutsch-Römisch Italic, Divina, Mirakula, Miami etc. Though totally different in design, many open and decorated faces show the same graphological features and have at least a related feeling-tone, e.g. Moreau le Jeune, Fournier le Jeune, June, Delphian, Fortuna.
Bernhard Cursive is too frivolous to reach for dignity the same high score as for luxury.
15. Gallia
BGIA DUTEo POCHOBII MINBCMTEN ÊGIAK, HXIT~
2. Ductus nearly 4 9. Wide: superficiality
4. Irregular: feeble will-power 11. Low speed: inactivity
5. Full : intellectual weakness 12. Backward inclination : affectation
6. Enrichment: tendency to exaggera- 15. Large : lack of powers ofconcentra-
tion, “would-be” tion, arrogance
Positive correlation with Bernhard Cursive, Narciss, Excelsior, Shadow Nobel, Locarno, Trafton Script. Comparable : Chic, Modernistic, Carlton, Boul’ Mich’, Trio B, C, Fatima, Vesta,



the heavier of the open or shaded capitals of many book and display types.
The low absolute values are probably due to its illegibility.
16. Kontrast
E6IA BETE FBCECLE MINE* EITEN EGIAK/EBTSEAN BE*
2. Ductus 3 8. Sharp : sensitivity, intellectualism
4. Regular : predominant will-power 9. Narrow: self-command, refinement
5. Thin: intellectual force 11. Low speed, equanimity
6. Simplified : precision, sense of order, 12. Steep : reserve
efficiency 15. Small: realism, dryness, precision
7. High pressure: will-power, stubbornness
Positive correlation with Omega, Metropolis, Narrow Heavy Gothic, de Vinne, Arpke, Columbia, Light Gothic Expanded. Comparable : Corvinus Bold, Eden Bold, Bodoni Bold Condensed ; further (less narrow) : Broadway, Bristol, Trio A, Ultra-Modem Bold, Stygian Black, Meridian, Modernique, in general all faces related with Ultra-Bodoni. This kind of type coinbines contradictory features : strength and stylishness ; the absolute values are equally high for economy, dignity and hygiene. The reason why “refinement” was inferred from narrowness, is found in the same careful and delicate innervation, which is also required for long ascenders. The high score for economy-soberness explains the figure for dignity, which is much higher than for luxury.
17. Mediaeval
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago
2. Ductus 1 9. Wide: activity, unconstraint
4. Fairly irregular: liveliness of senti- 10. Short projectors, A > D : reliabi-
ment lity, realism
5. Full: imagination, sense of colour 11. Fairly high speed: activity, lively
6. Enriched : sense of beauty sentiment
7. Fairly high pressure: will-power 15. Large: energy, entreprise
8. Moderately sharp: sensitivity, intellectual capacities
Positive correlation with Omega, Metropolis, Narrow Heavy Gothic, de Vinne, Arpke, Columbia, Light Gothic Expanded. Comparable : see 8, but only the sturdier kinds, for the correlations prove that strength and hygiene are the most important



features of this type. Because it is the only book face among the display types, the values are possibly not specific for Mediaeval, though its scores in the small size (nr. 8) makes us believe that it really has outstanding merits.
18. Narciss
Egia dute pocholm mmegitem egiak,hut* seam deeari hambat gaichtoago beren faU
2. Ductus 2
4. Moderately irregular: lively sentiment
5. Fairly full: imagination, sense of colour
6. Some enrichment, sense of beauty and of order
7. Fairly high pressure: will-power
8. Fairly pasty (total impression) : culture, sense of beauty
9. Wide: activity, unconstraint
10. Short projectors, A >D: contact with the world, interest in culture 15. Medium size
Positive correlation with Bernhard Cursive, Gallia, Eckmann, Locamo, Trafton Script. Apart from Metropolis, Narciss is the only type of these 17 that shows a correlation with Eckmann ; the common feature of these two is dignity and warmth, based on resolute, full, swinging lines and curves and pastiness. Note the appropriateness for strength, precision and reliability. Not all open, “handtooled” and outline faces possess these features, but most of them are comparable with Narciss in respect of grace and freshness. Ludlow-Cameo, Gravure (ATF), Cloister Cursive Handtooled, Goudy Handtooled, Nicolas Cochin Bold Open etc. are still on the side of “warmth and comfort”, whereas Caslon Shaded, Bodoni Open, Imprint Shadow, Dominus, Adastra, Colonna are different examples of a chiefly fresh and refïned atmosphere.
19. Excelsior
2. Ductus almost 4
4. Regular: conventionality
5. Moderately full: intellectual weakness
6. Traditional forms : lack of directness, conventionality
7. Medium pressure
8. Sharp
9. Narrow: lack of directness
10. Long projectors: delicacy, refinement
11. Slow : inactivity, bluntness
12. Fairly sloping: obliging
13. Undisturbed connection: dependance of judgment
14. School-model connection: lack of frankness
15. Small: unpretentiousness, precision, dryness, thouroughness, narrow-mindedness



Positive correlation with Bernhard Cursive, Gallia, Shadow Nobel, Locarno, Trafton Script. Comparable : all old-fashioned or stiff script and cursive types, inasmuch they are cold, delicate and precise.
Some ss. mentioned their disliking of this type. One of them defined : fine force (figure : 4).
20. Omega
Eg ia dtrte pocholu minegiien egiak, huisean denari hanbai gaichioago
2. Ductus 3—3£ 10. Long ascenders, > D : ambition,
4. Regular: will-power aversion from vulgarity •
6. Simplified : efficiënt, sense of order, 15. Medium size: thoroughness, enter-
lack of sense of beauty prise
7. High pressure: will-power, stubbornness
Positive correlation with Kontrast, Mediaeval, Metropolis, Narrow Heavy Gothic, de Vinne, Arpke, Columbia. As the clean way of cutting, the long ascenders and the narrow capitals do not appear to have succeeded in conveying the atmosphere of dignity etc. to a considerable degree, it will be very often possible to substitute any other regular bold display type in stead of this design.
21. Eckmann
Sgia dute pocholu minegifen egi ah, hulsean denari hanbai gaich
2. Ductus 1—1£
4. Irregular: lively sentiment
5. Full: imagination, sense of colour
6. Enriched : sense of beauty
7. High pressure : impulsiveness
8. Pasty : “Sinnenfreude”
9. Medium width
10. Short projectors: unintellectual, untamed impulse of expression 15. Large : depth of sentiment, enthousiasm, energy
Positive correlation with Narciss, Metropolis. Comparable : Cooper Black, Ludlow Black, Monotype-Goudy Heavy, Bravour Bold, Bernhard Heavy Antique (Extrafette Bernhard Antiqua), Fette Lo, Ella Bold, Stabil, Salut, Sensation ; in general every bold display type that is decidedly not angular and sharp.



22. Shadow Nobel
HGJA DUTH POCHOLU MIN£GJTEN EGIAK-WUÏSEAN DEN ARP
2. Ductus 2 8. Sharp : spirituality, sensitivity
4. Regular: will-power 9. Medium width; activity, self-
5. Fairly full constraint
6. Simplified : efficiency 15. Fairly small: realism, precision,
7. Fairly high pressure : will-power thoroughness
Positive correlation with Bernhard Cursive, Gallia, Excelsior, Locarno, Trafton Script. The delicacy of this type is formuled upon its thin and neat lines, not on the shadowing. Comparable are therefore those shadow, inline and open faces, or light sans types etc., that do not belong to the category of Narciss, with its smooth and round lines, e.g. Orplid, Luxor-Versalien, Pharos, Futura Inline, Erbar Inline, Gresham, Delphian, Neon (Nebiolo), Empire, Huxley Vertical, Banjo etc.
23. Metropolis
Egia dfute pocliolu minegiten egiak, nutsean denari banbat
2. Ductus 3 9. Wide: lack of self-command
4. Irregular : lively sentiment 10. Long projectors, A > D : refine-
5. Full: impressionability ment, ambition, aversion from
6. Simplified, some enrichments : sense vulgarity
of beauty and of order 15. Medium size: thoroughness
7. High pressure: heavy on hand, vehemence, stubbornness
Positive correlation with Kontrast, Narciss, Omega, Eckmann, de Yinne, Arpke, Columbia. We mentioned the inner contradiction of almost brutal force and refinement in this type. Comparable are, therefore, in the first line those types, which show this same incongruity. Omega and Arpke belong to this category too; further: Corvinus Bold, Kingston Bold, Modernique, Parcival, Rundfunk ; with less reserve : Ultra-Bodoni etc.
Several ss. called Metropolis a disagreeable, obtrusive type.
24. Locarno
Esia dute pocbolu minegiten egiak, butsean denari banbat gaicbtoago beren falta, dute eta



2. Ductus 2—3 9. Medium width : activity, self-com-
4. Irregular: lively sentiment mand
5. Full: impressionability 10. Long ascenders, >D: refinement,
6. Enriched : sense of beauty aversion from vulgarity
7. High pressure: will-power 15. Fairly small: concentration, preci-
8. Sharp : spirituality sion
Positive correlation with Bernhard Cursive, Gallia, Narciss, Excelsior, Shadow Nobel, Trafton Script. Comparable: see under nr. 11, bold versions. As Locarno is properly a text type, the remarks made in regard of nr. 17 are also valid here, to a smaller extent.
25. Narrow Heavy Gothic
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren falta, dute eta halere gure gain eman nahi
1. Level of form 3£—4 8. A little pasty : intellectualism, lack
2. Ductus 3—3i of spirituality
4. Regular : lack of sentiment 9. Narrow : constraint, lack of direct-
5. Thin : intellect, lack of phantasy ness
6. Simplified : sense of order 10. Short projectors : hard-headedness
7. High pressure: stubbornness 15. Fairly large
Positive correlation with Kontrast, Mediaeval, Omega, de Vinne, Arpke, Columbia, Light Gothic Expanded. Comparable : all bold, extra-condensed display types. Illegibility influences the values again. The narrowness still produces an average of 1.9. for dignity.
26. de Vinne
Egia dute pocholu minegiten egiak, hutsean denari hanbat
2. Ductus 3—31 9. Medium width : entreprise, self-
4. Fairly regular: will-power, lack of command
sentiment 10. Short projectors: realism, hard-
5. Moderately full headedness
6. Fairly simplified : efficiency, lack 14. Garland : friendly attitude
of sense of beauty 15. Large : enthousiasm, energy
7. High pressure : will-power, vehemence
Positive correlation with Kontrast, Mediaeval, Omega, Metropolis, Narrow Heavy Gothic, Arpke, Columbia, Light Gothic Expanded. Comparable : Goudy Bold, Caslon Bold, Century Bold, Cheltenham Bold, Clearface Extra Bold, Holbom, Whitefriars Bold, Chatsworth, Haddon etc.



27. Trafton Script
(^aici diite ^)&cli&Li wiLuec^cteu e^uiLy Itut&eau de— natL lianlat c^(iLcLt&a.^&
2. Ductus 2—21 10. Long projectors, A > D : refine-
4. Fairly irregular : lively sentiment ment, ambition
5. Thin: intellect 11. Relatively high speed: activity,
6. Simplified on the whole : efficiency, lively sentiment
sense of beauty 12. Rather steep : reserve, intellect
7. Fair pressure: will-power, suscepti- 13. Undisturbed connection: systema-
bility not absent tic intellect
8. Sharp: spirituality 15. Small: concentration, precision,
9. Fairly narrow: self-command, pre- thoroughness
cision
Positive correlation with Bernhard Cursive, Gallia, Narciss, Excelsior, Shadow Nobel, Locarno. Comparable, for delicacy, refinement: see nr. 14 ; for aler , personal handwriting-features : Gillies Gothic, Park Avenue, Legende, Raleigh Cursive, Jiujitsu, Skizze, Arabella, Forelle, Ariston, Holla, Signal, DrescherEilschrift, Energos, Fanal, Appell, Graphik.
The greater speed and steepness make this type less appropriate for precision than Excelsior.
28. Arpke
Egia dute pocholu minegiten egiak, hutsean denari hanbat gaichtoago
2. Ductus 2£ 7. High pressure: will-power, stub-
4. Fairly regular: will-power bornness
5. Fairly full 9. Fairly narrow: accuracy
6. Simplified, some tendencies to en- 10. Short projectors, A > D : hard-
richm.: sense of order, some sense headedness, realism, ambition, aver-
of beauty sion from vulgarity
15. Fairly large: energy
Positive correlation with Kontrast, Mediaeval, Omega, Metropolis, Narrow Heavy Gothic, de Vinne. Comparable : see under these numbers ; further : Locarno Bold, the narrow version of Fanfare and related designs (Triumpf, Alarm).
Note that Arpke, with its good values for hygiene and warmthcomfort, is used for advertisements of Jaeger Ltd.



29. Columbia.
Egia dute pocholu minegiten egiak, hutsean denari hanbat
2. Ductus 3J 8. Pasty: sensuality, lack of spiri-
4. Fairly regular tuality and sensitivity
5. Full: impressionability 9. Wide: activity, inaccuracy
6. Simplified: efficiency 10. Short projectors Naturhafügkeit”
7. Fairly high pressure: stubbornness 15. Moderately large: energy
Positive correlation with Kontrast, Mediaeval, Omega, Metropolis, Narrow Heavy Gothic, de Vinne, Arpke, Light Gothic Expanded. Comparable : Hawarden, Morland, Nelson O.S., Kolonial, further the round, smooth bold romans mentioned under nr. 21 and 26.
30. Light Gothic Expanded
Egiadute pocholu minegiten egiak, hutsean denari hanbat gaichtoago beren
1. Level of form 4 8. Sharp: intellectualism, dryness
2. Ductus 3 3i 9. Fairly wide : unconstraint
4. Regular : lack of sentiment 10. Rather short projectors: hard-
5. Fairly full headedness, materialism
6. Simplified : sense of order, efficiency, 15. Fairly large: some entreprise and
lack of sense of beauty activity
7. Moderate pressure: lability, lack of will-power
Positive correlation with Mediaeval, Narrow Heavy Gothic, de Vinne, Arpke, Columbia. Comparable : all related sansserif, egyptian or common and quiet roman faces.
Final remarks.
A. survey of the results shows that the atmospheres of the type faces can be brought under three categories chiefly, which might be called after the qualities luxury-refinement, economyprecision, strength. This division corresponds with the differences in weight, and therefore, as we can say in this case, of pressure. The light and the bold type represent extremes in the constitution of character : susceptibility without will-power versus will-power without sentiment (“ rücksichtslose Tatkraft”).
An average thickness of the lines can have two causes . 1. the writing subject does not have any quality in a large degree (small



TABLE XXII.
Comparative Scale of the various Atmospheres.
Characteroi. and Grapbol. ,, •
Features Book-types Dispiay-types Bboks Advertisins Commodities
123 4 i5 6 7 8
§ Irregular Gallia Perfume
0 H Full Bernhard Roman Bernhard Cursive Luxury
H z Hnriched
a § Low PRESSURE
£ b Sharp Locarno Excelsior Poems Jewelry
Sï | projectors Lutetia Trafton Script Distinction Silk stockings
p S m Satire
g w a Regular
S § o Thin Orpheus Shadow Nobel Analysis > Fineness
3M 3 Simplified o § Precision
ril H Low PRESSURE 5 Light Gothic Exp. § Freshness Breakfast Food
£ “ w Sharp Modern Wide pq Narrow Heavy Gothic Report f2 Vegetables
< » d n „ Refrigerators
» 3 F Regular Romaansch f de Vinne J Economy
|go « Thin Saeculum > Kontrast > Automobiles
w §« Simplified Bodoni < Metropolis <
High pressure Beton g Arpke g Technics Reliability Building Material
«gS d Pasty S Omega 5 Strength
g® ^ Irregular
Full Grasset Columbia Farmers’ Life
B Enriched Coal
1 Low PRESSURE ECKMANN
■ Pasty
o Short projectors Thannhauser Narciss Romantic Advent. Warmth, Comfort



capacity of the personality, shallowness, level of form low) ; 2. the writing subject is in the possession of all qualities to a large degree, but they are in perfect balance (large capacity of the personality, harmony, level of form high). The corresponding writings are represented by Modern Wide versus Bembo, the corresponding “books” by report versus tragedy. Though such types are both regular, thin etc. and though they show a significant positive correlation, they are based upon contrarious principles.
The difference in weight can also be due to difïerences in pastiness. In the types used in our investigation, these differences were difficult to distinguish from differences in pressure. Without going deeper into the theoretical foundations of this question, we shall assign to this degree of medium thickness the term “intellect”, which stands for “spirit” (Geist) or “consciousness”. The three terms “sentiment”, “intellect”, “will” constitute the base of comparative scale of the various atmospheres, as given in Table XXII. In the middle of this scale one can expect “ durchgeistigtes LeberT (highest level of form), or “Lebensferne” (entire absense of the vital forces) ; at the ends of the scale an absence of controlling forces. Therefore will-power cannot be placed at the (lower) end of the scale, because it supposes the controlling influence of the spirit : the lowest degree passes into “the Earth”, instinct, “Triebhaftigkeit”, the origin, with the attributes warmth and comfort, graphologically : pastiness. The types “Thannhauser” and “Narciss” are placed at the bottom beneath a horizontal line, which indicates a continuation towards the top of the scale, because these types combine the qualities of both the ends of the scale.
As far as possible the corresponding units have been placed upon the same height, but the columns should be read from the top to the bottom in the first place, not horizontally.
The principle of this scale is not strictly tenable, of course. The linear division according to will and sentiment cannot coincide with a division in respect of “Spirit” and “Life”, apart from the question whether such a scheme alone is defensible at all. But in general this graduation may render some services in visualizing the transitions of one atmosphere into another, taking into account the unavoidable vagueness and impurity in the determination of the terms and feelings.
Some of the “ commodities” used by Poffenberger and Franken, Schiller, and Da vis and Smith have been arranged in column 8.
It is impossible to distillate more detailed prescriptions from the experimental material available at present. Further investi-



gations are needed to verify and to complete the present data; much detailwork on the descriptions of the qualities and on other type faces are necessary.
The results of this enquiry into the atmosphere-value of type faces has shown that, for the average public, these atmospheres are not quite distinct, i.e. that about the same face has about the same feeling-tone, or — we think of a remark by Frederick Horn, quoted on page 178 — that a clever lay-out can make such types appropriate for a certain quality that are properly unsuitable. But, on the other hand, the absolute values show often such sudden and marked predilections or aversions that we are justified in drawing the conclusion that each distinct and circumscript quality has at least one type that is very appropriate to it, whereas, in a second choice, many other type faces of generally the same characteristic can serve. In other words : it might be advisable under certain circumstances to use a “beautiful” type (i.e. a type that is highly effective in its own peculiar atmosphere) for quite another kind of feeling-tone, in stead of a mediocre one which properly belongs to that feeling-tone. It will be clear, however, that this fact cannot be turned into a general criticism that “all this talk about a specific atmosphere-value rests upon nothing”, for this fact only means that a good type is always better than a poor one.
The typographer therefore, who did not hit upon the specially appropriate type, will not have done actual harm to the transmission of the meaning of the text, but he missed an opportunity to intensify the force of impression of the text in a considerable degree.
Legibility, atmosphere-value and forms of printing types. 12



CHAPTER 10.
The influence of margin, interlineage, colour and quality of paper.
Series VIII.
It is a common fact that the aspect of a type face changes radically (often almost into irrecognizability) with different sizes and environment (leading, margin etc.). A good typographer, who knows to make use of this, can produce with any moderately good type almost any effect he desires. So Frederick A. Horn writes : “ .. .. too long have typographers laboured under the fond delusion that one type suggests ‘dignity’, another ‘strength’ and so on ; good lay-out is a far more subtle thing than that; and it is as feasible to suggest delicacy, for instance, with Beton Extra-Bold used cunningly, at it is to suggest the same quality by any other type”.
However, the delusion is not so fond as Mr. Horn makes us believe. His example is very nicely chosen, because the atmosphere of delicacy is founded upon a slow movement of carefully and precisely putting down some tender and fragile things. It is always possible to produce these effects with heavy types (provided that the types are drawn precisely, not in a round and flowing manner) by reducing the size and by introducing a large amount of white (spacing and leading). But Mr. Horn would find himself in considerable difficulties if he was to create an atmosphere of strength and punch (hitting-force) with only such faces at his disposal as Kennerley, Verona, Goudy Old Style etc., because these types lack the vigour of movement necessary for that purpose.
The cunning lay-out, of which Mr. Horn speaks, supposes a knowledge of the factors that make a type suggest ‘delicacy’, and that he tries to imitate with other types in a different way. Where an appropriate type face needs only a moderately good lay-out, there will an unsuitable type need a brilliant lay-out.
If we therefore still think that the investigation of the atmos phere-value of printing types is worth while, we cannot confine ourselves to that only, leaving unobserved the influence of lay-out.
It is impossible to bring under experimental conditions the whole intricate complex of relations, which is called ‘lay-out’,



in such a way that we can draw conclusions out of the results for application in each single problem. We did not- try to do so, but decided upon an inquiry into the influence of four factors, which covers but a small area of the problems, and should be completed and corroborated by many other studies.
It would have been theoretically possible to use the questionnaire of series b again on a systematical variation of lay-out of one or a few type faces. We expected, however, that such a method would not produce satisfactory results, because the differences between the feeling-tones of the book- and advertising-subjects were made rather considerable on purpose, whereas it was not likely that e.g. adding a wider margin to a type would make enough difference that it would pass from one of these eight subjects into another one. We therefore preferred to analyze the differences themselves and to ask our subjects to define those differences, with the aid of a hst of adjectives.
Questionnaire.
A list of 21 adjectives was accordingly prepared, divided in four groups. The first group of adjectives had a hearing on the standing of the paper, on the quantity of money and care spended to embellish its appearance ; the second group concerned the ‘temperature’ ; the third its ‘attitude’ towards the reader, its complaisance ; the fourth, finally, tried to describe the degree of directness at which the paper conveyed its message to the reader. The adjectives were the following :
1. luxurious, generous,
2. very well got up, without being luxurious,
3. well got up,
4. decent,
5. decent, but modest,
6. on the poor side,
7. shabby.
8. warm,
9. fresh,
10. lukewarm (can be said in Dutch of moods)
11. cold.
12. cordial, cosy (can be said in Dutch of relations between people)
13. friendly,
14. keeping at a distance,
15. repulsive.



16. matter-of-fact — quiet,
17. matter-of-fact — busy,
18. matter-of-fact — rapid,
19. digressive — quiet,
20. digressive — tedious,
21. digressive — busy.
Material.
The same text of series VII a, b was used, printed in a neutral sans-serif, called Nobel, which is very popular in Holland, in both the 10 point and the 20 point size ; length of line 20 Didot lugustins.
Interlineage.
Group A was varied in interlineage, and included : 10 point solid, 4 point leaded, 8 point leaded ; 20 point solid, 12 point leaded, 24 point leaded. The texts were printed on a goodquality, ordinary white stock, 25 x 30 cm. large ; the margins were therefore as generous as to be practically equal for all the kinds of print; moreover we could compute any influence in this respect out of the results of group B.
Margin.
Group B was varied in margins, and included : the 10 point text, 6 lines 2 point leaded, measuring about 9 x 2i cm., surrounded by 1, 2\ and 5 cm. margin ; the 20 point text, 4 lines 4 point leaded, measuring about 9 X 3.3 cm., surrounded by 1.25, 3 and 6 cm. margin.
Colour.
Group C was varied in colour. According to the simple and schematical planning of this experiment, the colours were chosen as near as possible to spectral colours, with the aim of facilitating comparison and of making the differences as great as possible. The text was the 10 point text of group B, margin 2\ cm., a little less at the top and a little more at the bottom; size of paper 8 x 14 cm.
The paper itself was strong, thin ; the surface matt and smooth, though not supercalandered1). It was supplied by Messrs G. H. Bührmann, who call it Inferno 814 (the yellow is of another
i) Almost identical with the specimen of art paper in the ‘Dictionary ot Advertising and Printing by Freshwater and Bastien (Pitman 1930).



series called Diavolo 3814, but it has the same quality). The colours were according to Prase-Baumann’s Colour Charts :
13. purple, between 1301 7 Vp—Pv and 187 6 P
14. red , . „ 279 8 R and 279a 8| R
15. orange, „ 352 7 Ro—Or and 321 5 RO
16. yellow , „ 655 2 C and 654a lf C
17. green , „ 926a and 896 6 Gc—Geel
18. blue , „ 1186 8 Bv and 1143 8 B
Paper.
Group D was varied in quality of paper. Size and lay-out of imprint and size of paper were equal to group C. The sorts of paper were : art paper, rather glossyx); newspaper-siodt, matt, a very little brownish in colour ; calandered, good quality white stock, such as used in group A and B; offset-cartridge 2).
Proceeding.
The general course of the experiment was equal to that of series II. The various sheets of paper were administered to the subjects in four cardboard portfolio’s in such a way, that the subjects, regarding at the same time only the contents of one case, could not compare sheets of the same group, printed with the same size of type, but only e.g. a 10 point sheet of group A, a 20 point sheet A and a 10 point sheet B ; it was feared that otherwise the subjects would take a too strong dislike against some forms, that were aesthetically less acceptable, though the general public accepts them without much antipathy in daily hfe when no comparative material is available. The groups C and D were each placed completely in one portfolio, and the subjects were explicitly allowed to compare these groups in themselves, because we thought that a similar division as the one carried through for the groups A and B, would not be sufficiënt to obtain an (in some respect) unbiassed judgment, and therefore we dropped it altogether, so that the conditions for the single members of the group were the same.
The instruction was given verbally and ran as follows (all subjects had previously accomplished series YII b) : “ The purpose of this experiment is to know the impression you get from the sheets of paper contained in these portfolio’s. This time we are not concerned with the feeling-tone of the type face (which is
*) Almost identical with the specimen of art paper in the ‘Dictionary of Advertising and Printing by Freshwater and Bastien (Pitman 1930).
2) See specimen ibid.



the same on all sheets) but with the total impression that you get from the way the imprint stands on the paper, of the combined impression of paper and imprint. You are to describe this impression with the aid of the list of adjectives you have got, with the restriction that the groups of adjectives 1 — 7 and 16 — 21 are not applicable to the sheets marked C, while 16 — 21 are not applicable to the sheets marked D.
You have to work only at one portfolio at the same time ; so do not fetch sheets out of the other portfolio’s for comparison. Take care to look at the coloured sheets only immediately beneath the üghts, when no shadows fall across them, and put them on a large sheet of white paper as back-ground. Keep them apart from the other coloured sheets, and if you compare them mutually, take care to vary the order in which you compare them. For example, purple looks different besides yellow or dark blue or red. Your impression of the colour is the first thing that matters ; so do not take objection, e.g. to the fact that the imprint on the blue sheet is almost illegible.
The sheets marked D are of different quaüty of paper. This may or may not make a difference, you can judge about that for yourself.”
The subjects, 34 women and 30 men, had to write down their judgment on a blank, out of which the number of times were computed that each adjective was assigned to each sheet.
Results.
The results are contained in table XXIII and table XXIV, that reproduce the total number of tunes the various adjectives were assigned to the sheets. In some groups the full number of 64 answers was not given, or they included incomplete ones, which had to be discarded. Nevertheless we reproduce the absolute figures and no percentages, because the lacking answers are but few.
In view of the difficulty of the task imposed on the subjects the results are fairly conclusive for the first three groups of adjectives, whereas the results of the fourth group do not give clear indications. The latter can be due to the abstract nature of the description asked for, or to the imperfect formulation.
Therefore the chief purpose of this series has been reached : the proof that actual differences in feeling-tone owing to differences in “lay-out” exist and that they can be investigated by suitable interrogation of the public.
The results may seem too schematical and trivial compared with the infinitely multiform problems of daily practice and the



differentiated use the skilled artist makes of these instruments for varying the feeling-tone of a printed text.
It is not our aim to present the results as an aid for the mediocre craftsman in becoming a great artist (for the judgment of the public cannot replace the conception of the artist) ; but rather as evidence in the solution of elementary practical problems, e.g. whether a crowded page affects the public really as repellent or merely as ‘a little bit less beautiful’.
Such evidence can only be gathered in a slow progression of the methods of investigation towards a more pregnant formulation and a wider scope of the investigated material. One should not expect to leam more from the present results than they are able to offer ; no definite figures, but only clear tendencies. Hence they are presented as the foundations of future, more comprehensive investigations.
a. Interlineage (Table XXIII).
An increase in the interlinear space lends a quiet, decisive force to the movement by which the lines are put down on the sheets ; only this force becomes almost too great in sheet 6. There is an increase from the mere decency to luxury ; an increase also to freshness for the 10 point and even to coldness in the 20 point sheets (the latter show, also relatively, more white). Sheet 2 and 3 are more inviting than sheet 1, but, as we said, 6 appears to be overdone, compared with a slight increase in friendliness of 5. Altogether the more quiet 10 point texts have a less extreme character than the 20 point texts in respect of luxury, warmth and friendliness ; they are moderate in expression.
The figures of the fourth group of adjectives again point to the préference of the sheets 3 and 5 for quiet, efficiënt communication. The 20 point sheets together show a higher total figure for digressiveness.
Evidently, leading is suitable for introducing or increasing seVeral features that have a high value in advertising. The desired effects will appear only if the leading measures at least about half of the point size of the types. The feature of higher standing and refinement, however, can turn into haughtiness (or even intolerant brutality in case of large-size types).
The lay-out with generous leading has been practised of later years notably in Germany, and with excellent results. In this way one has succeeded in attaining a sphere of distinction with rather bold and unelegant “egyptian” and sans faces. At the same time the resulting transparent grille of horizontal lines is very easy to manipulate for the designer ; it can be superimposed



TABLE XXIII.
The influence of interlineage and margin on the feeling-tone of a printed text.
& B. o
PM eb pq O .. pj r. cj 0
gw g PJ H m t, «
Sg 'fc B; ® | S H « 2 J g 1 I I g
,•5 f5 p « g g 8 ^ g h 3» i. ia « g « §2
g§ ëg§ I e e-| I „ ga I i s sS § eg £§ II Iê §„ 11 I ga
gfi sB§ a g Si M a 1 6 a 3» S S Si ! M (? S? 3| SS SsM
g* aBi I I g» g I I3 g g s s ë° 8 § sfls r i1 3' 51 s"r
* 2 3 4 5 6 7 8 9 10 IX l* I* 11 15 16 17 18 19 20 21
«/ 1 , 10 Dt. solid 9 9 16 8 9 7 5 63 14 12 17 19 62 7 16 29 12 64 7 20 5 15 8 8 63
Sl 2 4 pt. leaded 12 14 19 12 2 2 1 63 15 16 21 12 64 4 26 25 9 64 17 4 7 9 7 16 60
Z 1 3 8 17 17 15 16 8 1 — 64 12 25 22 5 64 4 33 24 — 61 18 4 15 13 7 4 61
3 4 iÓpt.’solid ” 28 6 8 11 1 2 8 64 7 4 11 32 64 4 6 37 16 63 18 8 1 19 10 5 61
nj 5 12pt.leaded 24 10 13 9 1 5 2 64 6 27 5 26 64 3 13 40 6 62 29 2 4 20 1 5 61
ë ( 6 ” ”’ 24 „ „ 39 4 9 5 — 4 2 63 8 17 3 36 64 6 10 24 23 63 18 5 3 21 9 7 63
/ 7 10 point + 1 cm. — 2 4 15 17 5 11 64 14 14 26 10 64 12 21 16 15 64 10 15 15 4 7 10 61
1 8 +21,. 1 10 23 14 12 4 — 64 17 13 26 7 63 8 35 14 5 62 28 12 8 11 1 3 63
|)9” ” + 5 „ 7 11 16 13 16 1 — 64 14 13 24 12 63 8 23 29 4 64 13 11 14 8 6 12 64
£ ) 10 20 ” + II 6 1 14 10 4 14 14 63 8 14 17 25 64 8 9 17 30 64 11 20 7 5 8 12 63
iin ” +3 ’ 14 12 14 17 3 3 1 64 10 25 10 19 64 6 14 31 11 62 24 8 10 10 4 7 63
(12 ” ” + 6 „ 23 9 13 12 1 3 3 64 13 22 4 25 | 64 | 11 11 33 9 64 21 5 7 11 6 11 61



on pictures and it is very useful for linking parts of the page together. Moreover, it is not as heavy in colour as the blocks of text formerly were. Only some more advertising space is needed
b. Margin (Table XXIII).
The results point to the existence of an optimal width of margin. Though both the 10 point and the 20 point sheets show a steady increase towards luxury, the highest sum of the columns 12, 13 (friendliness) and of 16, 19 (quiet) is reached by sheet 8 for the three IQ point sheets, and not by sheet 9. The corresponding figures for the 20 point sheets show no regression, but a lack of progression of sheet 12 compared with sheet 11. In general the 10 point sheets are warmer than the 20 point sheets ; the latter show an increase towards coldness (columns 9, 11).
The conclusion can be drawn, that a moderate use of surrounding white space can promote the effects of luxury, friendliness and quiet efficiency, whereas a surplus of margin leads up to coldness and haughtiness. But it is remarkable that the very small margins of sheet 7 and 10 did not call forth a clearly unfavourable judgment. An increase of interlineage and margin will improve the standing of a text, but solid lines with minimal surrounding white are not necessarily disagreeable to the readers’ eye.
c. Colour (Table XXIV).
Only the second and third group of adjectives were used for describing the colours. The latter were so much saturated that these papers can only serve for covers. Further investigations are necessary, that make use of pale colours. The results of the present series do not produce new or deviating judgments.
Purple. The highest figure for lukewarm of all colours, in coldness only surpassed by blue. Decidedly repulsive.
Red. The warmest of all colours, foremost in cordiality.
Orange. Warmth is predominant, but the yellow has introduced a good deal of freshness and taken away a part of the cordiality. So yellow is the friendliest colour of the six, together with green. The fourteen answers of column 14 and 15 for red have been reduced to 5 in column 14.
Yellow. The freshest colour, even somewhat cold. It is neither cordial nor repulsive ; the figures are about equally divided between column 13 and 14.
Green. This colour combines various properties : there are as many votes for freshness as for warm and lukewarm combined.



Though it is certainly friendly, there are several subjects who wrote ‘cordial’ or ‘keeping at distance’.
Blue. Decidedly cold and repulsive, but again the judgment is not as unanimous as in case of orange and red.
The feeling-tone of colours is the most “subjective” ; there are strong personal difïerences if the subject is asked whether he likes them, though the practical application on articles does not present any difficulties, as the experiments of Collins and Gwendolyn Schiller showx).
TABLE XXIV.
The influence of colour and surf ace of paper on the feeling-tone of a printed text.
m> ra Cu Pd ft
Q O O O
» t, SB ft Ej CO ra m 0Q „ Ph
Bg gag * * 8* « H gg B « gj § S 5§ SS
2o°w2 o *5 .SS ° m 9 § s g £ o § g „ q » 3 g h
M « 3 H « § g gg g « S 9 £ O 9 § «2 B
j® inghi r « h 3 ^ § S K a
«E B O _ h g g w g
Bo o'-' p
> O E* E*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/13 PURPLE 14 4 26 20 64 5 9 12 38 64
*(l4RED 45 7 10 2 64 37 13 10 4 64
§ 315 orange 32 22 8 2 64 26 32 5 — 63
§116 yellow 3 37 7 17 64 6 25 22 11 64
Ö # 17 green 13 28 16 7 64 12 31 16 3 62
(is blue 10 10 16 28 64 5 8 18 33 64
W119 art 8 18 20 12 4 2 — 64 6 37 8 13 64 6 30 26 2 64
< )20 newspaper — 1 1 1 4 17 40 64 13 1 30 17 61 8 15 11 30 64
eè )21 calandered — 1 8 14 26 12 3 64 3 12 45 4 64 3 27 29 4 63
Si [22 offset 9 22 16 13 3 1 — 64 40 10 9 5 64 35 16 10 — 61
d. Surface (Table XXIV).
The results of this part of the series must be ascribed not only to the surface of the papers, but also to the differences in thickness.
The offset cartridge was the thickest; it was followed by the art paper, which had only a little more weight than the calandered paper ; the newspaper sheet was the thinnest.
Art. The columns 3, 2 and 4 are favoured. The paper was too small and thin to be called luxurious. The glossy finish and chilly white colour are responsible for the freshness ; the rather low figure for coldness is surprising and the high totals
!) N. Collins, The appropriateness of certain colour combinations in advertising, 1924.
G. Schiller, An experimental study of the appropriateness of colour and type in advertising. J. Appl. Ps. Vol. 19, 652.



in column 13 and 14 confirm this indication that the art paper (at least the variety used in this series) is not too haughty and refined.
Newspaper. This sheet made a poor impression on the readers ; the temperature balances about lukewarm ; there are many votes for the adjective ‘repulsive’, though as many subjects do not go so far.
Calandered. Though this paper was of a good quality, only on the thin side, the judgments stress this latter feature. There is a high figure for column 10 (lukewarm) ; the votes are almost equally divided between ‘friendly’ and ‘keeping at a distance’. All these properties are unsuitable for advertising purposes; that is to say, in cases where the advertiser is free to choose his paper to support the atmospheric effects of the lay-out, he should not take kinds of paper possessing the said properties.
Offset cartridge. This specimen was described as almost luxurious, decidedly warm and cordial. The combined qualities of soft, even texture, thickness and somewhat ivory colour have brought about these favourable comments.



PART III.
SHORT HISTORICAL REVIEW OF THE MODERN PRINTING-TYPE.
§ 1. Introduction.
The development of printing art from the period towards the end of the 19th century up to our time cannot be understood independently of the other branches of art, because their connection has become entirely different.
We scarcely need to picture the deplorable decadence of the arts in the 19th century : it is well known how both, free and applied arts, had got stuck in an impotent repetition of ancient styles and that moreover manual skill had come down.
Everywhere a revolution against the empty academie art was prepared ; the attitude of the artist towards his subject altered thoroughly.
Some artists fled out of the oppressing narrow reality into a serene arcadian world, filled with a noble race of faultless men and women ; their style was monumental, static and decorative, constantly menaced by formalizing mannerism, but highly inspiring for the applied arts (as a matter of fact its last acolytes were mainly occupied in that direction). Among this group we reckon in England the praeraphaelites, Morris, Walter Crane, Aubrey Beardsley ; in France Puvis de Chavannes and Grasset ; in Germany Feuerbach, Marées, Klinger, Leistikow, Stuck, Hoffmann, Eckmann ; in Austria Klimt ; in Switzerland Böcklin, Hodler. In the decorative arts this style led to a very peculiar ornamentation of slowly winding movements, lazily twisting sterns and foliage, with sudden curdling in bulky lobes ; no sharp lines and hard colours. The human faces keep the serene and unearthy beauty of a Jane Burden (Morris’s wife and model of the praeraphaelites) and the similar ideals of the German and French very often wear a flat and indolent expression. This same expression we find more or less uniformly in each country (Walter Crane, Aubrey Beardsley, Mucha, Grasset, Eckmann, and in America : Bradley, Leyendecker). This “art nouveau” or “style bois-courbé” (in Germany also



called “Jugendstil” after the magazine, to which also Eckmann and Pankok co-operated) found a well known exponent in the Belgian architect and decorative artist Henry van de Velde 1). On the other hand, opposite to this movement, we find an attitude, that eagerly accepts reality. No idealizing, but a deliberate tuming towards nature and mankind, often with a conscious socialistic element. It aroused a new respect for labour. The discovery of the “beauty in everything” plus the respect for manual labour formed the base for the work of Morris and his “Arts and Crafts” movement. It was his aim to pervade as much as possible the objects of our daily surroundings with beauty and to make them at the same time sound and useful. The scope of activity of this versatile artist and artisan was extraordinarily great; besides the books of the Kelmscott Press his workshop issued furniture, stained glass Windows, tiles, chintzes, wallpapers, etc. Nevertheless he exerted influence rather by his consistent striving for beauty than by his style (at least on the continent) because the sombre world of his tales of heroism and chivalry could not be a source of inspiration for the modern man of the twentieth century.
In the same period the political rise of Japan aroused a new interest in Japanese art. lts woodcuts, embroidery, pottery, lacquered ware, etc. became more generally known either in natura or in effigie (big illustrated works on Japanese art appeared). The composition and the bright colour scheme of the graphical works were gladly adopted by impressionism (in poster art it is easily traceable in the work of Toulouse Lautrec (see Frenzel, Penrose’s Annual 1935, p. 1); the carefulness of fabrication and design and the richness of forms gave a strong impulse, parallel to Morris’s. The need of new forms for the objects of daily use, created by the modern ways of living, could not be satisfied by mediaeval and renaissance elements, nor by these oriental motives only. The decorative art, mentioned above, supplied the necessary incentives instead.
The next period saw a complete invasion of young artists in the former domain of simple craftsmen ; not contented with functioning as advisory designer, they went to live in “colonies”, in “ Arbeitsgemeinschaften”, devoted to private manufacture, where economical considerations were sacrified to idealistic motives
’) These remarks are not entirely correct in so far as: 1. The röle of other periodicals (Pan, der Bunte Vogel, die Insel) is hardly less important than that of the “Jugend”. 2. The style of the co-operating artists was not very einheitlich. 3. Van de Velde and his pupils propagated a “free” line-ornament, whereas the German artists used vegetable and animal motives. But the similarity of ductus is large enough to justify the generalization.



af both social and artistic nature. Much theorizing was done, the material for which was mostly provided by the old controyersy between the need of omamentation (which often led to ïxuberance or over-modelling) and functionaüsm. Later on the ‘Wiener Werkstatte”, under Olbrich and Paul, worked for long years with better commercial results ; they exerted considerable artistical influence. With the rise of neo-impressionism and expressionism a more abstract, introvert and individualistic movement was created. This was the end of the “new art”. Perhaps never a finished, circumscript style passed away so soon as this one. It is true that in many countries popular national motifs had been absorbed, which strengthened its vitality, but still the position of the ornament shifted from ‘fundamental’ to ’admitted’. In book production it was realized that a printed text is not made to look at, but to read ; that the proper nature of print is not the presentation of an agglomeration of single characters in a rectangle, but rather the aspect of a number of lines, built up of words. Consequently the solid block of black with white margins, suggesting a simultaneous perception, was abandoned in favour of an open composition, resting upon the paper as background and which can only be perceived clearly by a successive act. Display-composition also became less decorative and more effective. The artistical needs were not neglected because of these utilitarian factors, but the connection remained an exterior one. Aesthetical formgiving was little tuned in to the nature of the text. The futurists were the first to make typography itself a method of expression, but their art was too abstract to have much influence in broad circles. Gradually these factors were brought more in harmony with each other, especially as in advertising it became a first need to aid the effect of the text as much as possible, also to give a general impression, even if the text is not or not yet read. Whereas formerly the extemal form pursued independent decorative aims, regardless of the contents of the text, it became more and more an effective co-operator, that can have two functions : to promote reading and to create atmosphere. In respect of the relation between the text and its form it is irrelevant which of the two is emphasized.
The lawless phantasy of the first years after the decorative period has been continually purified into an extreme simplicity, that has been accused, not without reason, of cold correctness. But it was necessary to go through the “typographie pure” and to master the essence of it before the new forms could be found, required by the greater technical possibilities. Photo-engraving and printing processes were highly improved, and whereas formerly



the printing of text and illustration each had their own needs (different processes and sorts of paper, which were scarcely to harmonize), these impediments are slowly being removed. These techniques, however, which are in steady development, do not allow as yet a both economically and artistically satisfying application. If a sharp fresh impression of the type is wanted, only the usual relief-printing can be taken into consideration. The planographic and intaglio processes yield either faded, chalky or unsharp, ragged letterprint. But commercial printing e.g. is impossible without half-tone illustrations (photo’s) which, however, require a smooth printing surface (‘coated’ paper). The glossy, cold appearance of these papers is very impersonal. For the planographic processes (offset, lithography) rough surfaces are preferred, but here the weak watercolour impression, mentioned before, is unsatisfactory. It is not improbable that the nature of the white paper will become more neutral in the future, that a standardization to a few types of surfaces will take place, viz. either a very little rough or a very even texture (as the present “ offset-cartridges”), or matt and smooth (something between the present “matt art” and “super-calandered” papers).
Many possibilities are hidden here, but a new orientation between the various ways of expression is necessary and only possible by a sound co-operation of artist and technician. A good deal has been done in this respect. The artist as a lonely pioneer in the first period of the revival of printing art has become the artist as head of the studio—the aesthetical laboratory of the printing-office.
§ 2. United States x).
The American type production is of prime importance, because it supplies both, the European foundries and composing-machines, with designs. Only in later years the export to America from these European countries, notably from Germany, has reached a considerable quantity.
The revival of printing art in England towards the end of the last century called forth an immediate and notable response in the United States 2). Though the enthousiasm partly took regrettable forms — Morris’ style was copied slovenly and without taste3)
*) Because only a few of the American foundry-types are used on the European continent, our review in many cases could only be founded on specimen books, not on observations of these types in practical application.
2) 'William Morris, His Influence on American Printing, by F. W. Goudy. Philobiblon 1934, 4.
s) e.g. the Jenson Type by Phinney, which has, alas!, been imported to many European foundries.



— some artists produced work of a high Standard, at first keeping close to Morris’ fashion, but soon developing a style of their own, just as was the case in England. Among these artists were Bruce Rogers, who founded the Riverside Press ; Bertram Grosvenor Goodhue, who designed the Merrymount type, and, about 7 years later, the Cheltenham type ; Daniël Berkeley Updike, the founder of the Merrymount Press J), and Frederic W. Goudy, who directed the Village Press.
They all followed, like Morris with his Golden Type, the Jenson model, though in a less sombre and therefore more trueto-life version ; the ruggedness, which is characteristic for this fashion, has been followed in later years by a far more polished style, eventually accepting true modern faces, even Bodoni. Beside the Merrymount type, already mentioned, we must note Goudy’s Village type (1903), Rogers’ Montaigne (1901), and the Montallegro (1905), which was designed and cut by two Englishmen : Herbert P. Horne and Edward Prince, once punchcutter to the Kelmscott Press 1 2).
Among the better-class types, that appeared in those years, there are a few, that show the features of the older foundrytypes, combined with the softer and rounder manner of the later years, e.g. Cushing (1897, by J. Stearns Cushing), and Bookman (1902). Century (1898, by L. B. Benton) has to be mentioned as one of the first examples of the modern Ionic style, which is remarkable for its legibility in the smaller sizes. To this model several features of another well-known design can be traced, viz. of the Cheltenham. Cheltenham became, and is partly still, the most popular type in the world, though, to be exact, not in its original form ; but it became the cauchemar of nearly every type lover. It appeared in 1902, designed by Bertram Grosvenor Goodhue in collaboration with Ingalls Kimball (who, in 1891, edited the “Chap Book” with Herbert Stone
— a notable document of the new style at that time). The Cheltenham is more firm, clean and personal than its predecessors, and very plain, but it lacks just the individuality to protect it against such terrible contortions as being edited in four grades of heaviness, each of these grades subdivided in four grades of
1) The Merrymount Press of Boston. An Account of the Work of Daniël Berkeley Updike, by Dr. G. P. Winship, Cambridge, Mass. Herbert Reichner,
Wien. . „ _ ...
Notes on the Merrymount Press and its Work, by D. B. Updike. Cambridge, Harvard University Press, 1934.
2) As the execution of a design is often delayed for years, no great value can be attached for purposes of historical and artistical interest to the year of release. The information that I could get hold of is given in the list on page 00.



width, in either solid, hatched or outlined finish, plus the same thing all over again for the various body-sizes (long and short ascenders) and for the itaüc. Century and Cheltenham were the first types on which a whole “family” was built. This typefamily system has become of prime importance. The larger grades of the older types looked feeble and irregular and could therefore not be used in display-composition : the Cheltenham was the first type to combine the true properties for both book- and advertising-composition. It opened the possibility for a unity and dignity of style, also in the more spectacular jobs, which were hitherto unknown. This all-round usefulness made the Cheltenham a most profitable proposition, but being made foolproof it lost the qualities for any better work.
Della Robbia is drawn on rather weak lines, it has little contrast and very short serifs, which makes it comparable to Cushing. Thomas Cleland, who designed this type about the year 1900, also belongs to the artists, who came to a brighter conception of the Italian Renaissance. He also did much decorative work in that style, festoons and other flower-ornaments, cherubs, cartouches, etc. A few years later another type appeared, done in the same manner, viz. Bullfinch (1904), partly by M. F. Benton.
Though I do not know of any generally available printingtype design of his, we cannot omit to mention Will H. Bradley, one of the most versatile and productive artists in those days. He underwent the influence of Morris like the others, and later became a follower of the contemporary English illustrators and of the “art nouveau”. He designed a large amount of decorative material for the American Typefounders Company. He used it lavishly himself for his compositions (e.g. the “ Chap Books”) at the Wayside Press.
In apparent relation to the poster-technique of about 1900 (we mentioned Bradley, Penfield and Leyendecker before) stands Blanchard, formerly of the Inland Type Foundry, which is still on sale in Europe in several versions (Morland and Nelson Old Style in England, Columbia in Holland, Kolonial in Germany), an excellent display-face, done in a simple and habile brushtechnique, that had many successors in the following years.
Side by side with the flowery, soft and round manner of this last group of types goes the development of the pure, classical style, and together with it the reappearance or revival of the best types in the history of printing 1).
*) See “Type Design” by Douglas McMurtrie, Bridgman, Pelham, New York 1927.
Legibility, atmosphere-value and forms of printing types. 13



Bradley detected strikes from the original Caslon matrices, md worked with these types almost exclusively ; later on they were generally for sale at the American Typefounders Company under the name Caslon 471. The Monotype cut the design in side-hole matrices in 1903 for the “Vogue” magazine of Arthur Turnure.
In the same year Updike imported a handsome mediaeval of the Stempel Foundry of Frankfurt, cut by A. Janson at about 1670 (which was put on the market in the twenties) and also got hold of an old-timer, which he called the Mountjoye type, but it was identified by Stanley Morison as Bell’s type (on Monotype since 1932).
For the next twenty years almost the entire type-production was in the hands of two men : Morris F. Benton, of the American Typefounders Company and Frederic W. Goudy, art director of the American Monotype, and owner of the Village Press.
Proceeding chronologically we mention : Clearface (1909) by Benton, a nondescript display-type, that became popular for newspaper work. Artistically more important was a design by the same artist, based upon the types of Bodoni, and named after him. We possess Bodoni’s types e.g. in his Manuale Tipografico, issued by his widow in 1818, six years after the death of the great printer. Following at first modem French types, Bodoni later on designed a series of letters, that show a more pronounced contrast and more square dimensions. Though the single forms vary in design, the general style is quite salient, and Benton has captured it excellently. The principle of such “interpretations” has been severely criticized by Mr. Updike, but I cannot share his contempt for such “composite afïairs”.
This Bodoni appeared in 1911 and was followed by a less heavy grade in 1912, called Bodoni Book. By now this revival and its numerous imitations have become as popular as Cheltenham and Caslon. Mr. Goudy, who has always displayed a lively admiration for Roman inscriptional lettering, as it is found on the Trajan column of the Forum in Rome, published two types in 1911, viz. Forum Title and Kennerley ; the first consisting of pleasing, but somewhat wobbly capitals, the second a good roman face x), but like many of Goudy’s faces, a little too glib. Goudy has a dangerous ease of design ; he produced over hundred faces in about fifty years, which, of course, do not all have equal value.
i) it has been copied in America; this copy is known as L aclede Old Style, later on as Munder; in England it is called Verona.



In fact but few of them show outstanding merit, as e.g. Goudy Old Style (1916), to which a heavy version by Benton was added in the next year. This is a fine, commercial type and it is used to a large extent in .America and England ; its lines flow in a broad manner, it is open, large and simple, but lively.
At that time Benton created a new revival, which Mr. McMurtrie has called the best rendering of Jenson. lts name is Cloister. Undoubtedly we have here a very fine face; very legible, quiet and of even colour. For use on rough stock a lighter edition was edited in 1926.
We owe yet another important recutting to this brilüant craftsman and artist: Garamond, edited during the years 1917—1919 and based upon the so called Garamond of the Imprimerie Nationale, which was identified by Mrs. Warde as having been cut by Jean Jannon in 1615. This series includes a beautiful italic and — of course an unhistorical — heavy grade. The American and English Monotype adopted it in 1923, several Continental foundries and the Intertype followed. This version is easily recognized by the upturned beginning strokes of m, n, etc. Like Cloister, the contrast is scarce ; the freedom of the lines tend to make a whole page look scrawly, but on the whole it has not many shortcomings.
A far less successful revival has been edited in 1915 by the Linotype under the name of Benedictine. It is an Italian Renaissance imitation of the worst kind ; heavy, plump and badly designed. The year 1918 saw the appearance of a very original type by Mr. Goudy, called Goudy Modern, a round, graceful fantasy on the Bodoni theme. We shall passr here Mr. Goudy’s alphabets, that are not generally for sale, e.g. Goudy New Style, Collier Old Style.
His Hadriano capitals (1918), according to Mr. McMurtrie “of magnificent drawing”, affect me in quite the opposite way : I think it is a poor design. Mr. Goudy also produced several rather neutral and quiet types in the classical style, Goudy Antique (1912), National Old Style (1923), Goudy Light, and Italian Old Style (1924). Goudytype is an experiment, and not a successful one. On the other hand Marlborough Old Style (1925) is again a very fine face ; its general proportions are indeed much the same as those of the Lutetia by J. van Krimpen, but the contrast is less striking than that of the latter and it keeps more to the Doves and the Montaigne type.
In 1914 Bruce Rogers designed a series of capitals, that were used for the Metropolitan Museum of Art by its secretary, Henry W. Kent. A lower case alphabet was added, and with the whole



iount a translation of Maurice de Guérin’s “ Centaur” has been :omposed ; accordingly this design was baptised Centaur. The English Monotype Corporation recognized its superb qualities ind reissued it. Like every design by the Anglo-American school, it is said to be based upon the types of Jenson’s Eusebius, but it scarcely needs such an introduction, which does not seem justiQed either. It is a small-faced type and only the sizes from pica apwards show the wonderful grace and crispness adequately.
A highly successful series for publicity purposes, named after its originator, Oswald Cooper, appeared in 1919. It is executed in firm brush-strokes, introducing a liveliness, that makes this type excellently suited for minor solid texts. The fount includes a heavy display-grade, called Cooper Black, which ranks among the best of the heavy weights, and was much imitated.
Beside these American designs some classics of English origin were imported, viz. Franklin Old Style, Baskerville, Bulmer and Scotch, all four modem faces, that met the growing demand for sharper, blacker faces.
In 1920 the Monotype edited a character by its typographical manager, Solomon Hess, and named after him ; a quiet, not very distinguished heavy roman. Electra (1935), by the productive and imaginative typographer and decorator W. A. Dwiggins, does not depart from the classical model, though its contrast lends it a more lively aspect than the true old-style classics. In 1934 the American Type Founders Company showed in their specimen-book the first fount of a new type, designed by and named after Morris Benton, which is in one word, brilliant. The experience of a life, dedicated to type-design of the highest Standard, has mastered every difficulty in a splendid way. The Benton is extremely plain, in fact one can scarcely point out the elements that account for its outstanding quality; every single character is as normal as it can be, the proportions and the contrast are not unusual either, but the total effect is striking, although in no way obtrusive. Indeed, it is not a type of grande allure, being too neat and precise, though again, it is far from being common. Technical perfection, excellent readability and simple grace are its most marked features.
After the war the productivity and interest was focussed on display-types. More and more differentiated needs of advertising asked for a vast range of special designs. The trend in the design of display-types cannot be understood without keeping in mind the needs of advertising and the way by which hand-lettering covered them. When advertising gradually was devoting more attention to the aesthetical appearance, it made use of the type



families (e.g. Cheltenham, Kennerley, Goudy, Bodoni) instead of mixing an incongruous lot of display-types. However, the distinction between headline, sub-head, text, name etc. was not clear. The problem of mise-en-page (Flachenteilung) was not recognized, which problem, by the way, does not deserve all the contempt Paul Renner has poured out upon it1). Truly, an aesthetical scheme for grouping the page may have undesirable results, but a well-considered optical division may guide the eyes in a sequence that suits the relative importance of the different parts and that establishes a higher attention value for each of them separately. So the contrast between the components was increased : whereas formerly the text was composed of largesized type, in a wide and open arrangement, this method has been deserted in later years, because, it was argued, the text is either not read, however clearly composed, owing to a lack of time or interest on the side of the reader, or it is read in any case, and then one must use the costly advertising-space for another purpose, e.g. granting more space to the headline or to the product’s name, to make it stand out more clearly. A study of advertisements of to-day will show that for the text of the greater part either Caslon, Baskerville, Bodoni, Cloister or Garamond is chosen ; being highly legible and economical in space, but also plain and rather neutral, their only function is to convey the meaning of the text as efficiently as possible. Any other effect, notably eye-catching and the creation of atmosphere and features, peculiar to the advertised product in question, is obtained by means of the lettering for headline, subheads and name. In order to counteract the easily arising danger of the page becoming cold and dead, livelier types were brought on the market : fanciful scripts and also adomed variations of the eye-catchers already in use. In the development of these types Germany took the lead ; it exported on fairly large scale to England and America. The Creative phantasy of the German type-designers is far greater than that of their colleagues of other nationalities, even of the French. They are brought up partly in the Fraktur tradition, and this fact enables them to stand more freely towards the traditional roman form ; they have nothing of the respect of the Anglo-American school for the classics, and it is this very circumstance, which establishes the proper conditions for the creation of fancy types, meant for the functions described above. Of course, between the amazing quantity of new display-designs many uninspired, ephemeral and even ugly ones are found, but this
) Gegen den Schematismus in der Typographie; Klimsch’s Jahrbuch 1933.



;an not alter our admiration for the many really original inno/ations. Because the advertising lay-out in Europe followed ;ssentially different lines than in America, the need of the fluent, ;ool and businesslike script (as opposed to the refined, soft and :urly European tradition, incorporated in Bernhard Schönschrift) was as yet not feit. In return to its export the Bauer foundry imported two excellent American designs, respectively by Howard Allen Trafton and by W. S. Gillies. Apart from the scripts the American designers keep close to the extra-heavy modern type, of which the prototype has been very aptly named Ultra Bodoni. For our review we shall take them together in groups.
Bernhard Booklet, Louvaine and Piranesi are indeed meant to be used for booklets ; their aimless mannerism put them among the ephemeral novelties. Rivoli is a shameless imitation of Koch Antiqua, but the former is lacking the very features, which made the latter one of the finest designs ever produced ; the external similarity of the copy cannot convey anything of the spirit of the original.
The sans-serif vogue called forth response also in America. Futura, Cable and Erbar were imported. The German letteringartist, Lucian Bernhard, emigrated to the United States and got to work for the American Typefounders Company, which produced his Bernhard Gothic and Bernhard Fashion, both modern sans types of free and graceful shape and therefore much more limited in use. The Linotype issued a sans face by W. A. Dwiggins, which tries to retain the naturalness and fluency of the written, not constructed character. This principle, however, is irreconcilable with the nature of the gothic; at best one can try to retain the general shape of the roman character, as Eric Gill did. This Metro-family, as it was called, lacks the punch of the true modem sans and is not often used in advertising. Huxley Vertical consists of very thin, narrow capitals, suggesting refinement; Pericles, by Rudolf Foster, is a series of sturdy, hand-drawn capitals, matching with Bernhard Gothic, including an extravagant and ugly alternative E and R ; it should be very useful for titlepages and such like ; Phenix and Empire (also Gothic capitals) derive their value from their narrowness ; Stellar and the related Delphian Open Title Capitals, edited by the Ludlow, again strike the tone of refinement and daintiness — a riice series for headlines. We pass such second rate designs as Intertype-Vogue, Ludlow“ Tempo" and -Karnak, Stymie, Tower, etc., that are mere copies of successful European types.
Of the new heavy gothics we mention Modernique (Willard Sniffen 1930), Eagle Bold and Novel Gothic (H. Becker 1930),



free interpretations of the usual extrabold sans, meant to avoid monotony. They are rather noisy.
The Ultra Bodoni-theme can be varied endlessly. The solid parts can be decorated, the serifs cut off, the proportions altered. The effect can be either of greater refinement, freshness, grace, or of greater firmness. To the first category belong : Gallia, Boul'Mich', Bradley Ultra Modern Initials (the latter being noisy and affected monstrosities), Chic, Parisian, Modernistic and Onyx, whereas we reckon among the other group : Nubian, Broadway and Modernique. Stencil, the foundry says, “is a type, whose letterforms have been made familiar by a century’s flour sacks, sugar barrels and packing cases. Simple, straight-forward, easy to read”. It should prove useful if the application is limited to names, marks, etc.
As we pointed out before, American cursives and script-types reflect the hastily jotted-down remarks of the modem businessman. They do not attempt as much to strive for calligraphic beauty as to express efficiency and energy. Liberty, on the other hand, is a copy of Bernhard Schönschrift (which we mentioned as a representative of the European calligraphic and delicate script), altered just enough to avoid infringement of copyright. For the very copyist, the A.T.F., Lucian Bernhard designed Bernhard Tango, another delicate script; some characters raise strong reminiscences of the German Fraktur-inspired varieties. A nice, graceful work, though not on the high level of his Schönschrift.
Of the original American designs we mentioned Trof ton Script, really a master-design, sharp, smooth, graceful and fresh ; Gillies Gothic, which consists of equally thick lines as the name indicates (Kaufmann Script is a similar product of the A.T.F.). It originates from the same spirit as Trafton Script, but it is even more smooth and emphatic. Park Avenue should be named next, as it can be regarded as a blend of these two. Like Keynote and Raleigh it is rather unruly and dazzling. These latter succeed certainly in the task of calling the attention by breaking up the monotony of the page. But it should be kept in mind that a too personal expression — which is easy to occur in scripts of this kind —- is often strongly disliked by a great part of the readers.
The Ludlow issued a few series, called Mandate, Mayjair and Laureate, which follow European models from a safe distance, whereas their Hauser Script is original and really a nice invention.
The line-composing-machines (Lino- and Intertype) limited the production of “beautiful” types to the adoption of successful foundry type-designs, but spent considerable energy on the research



af good newspaper types. The solution of the problem was sought in the direction of the so-called “ Ionics”. The first editions proved too bold, so subsequently less heavy versions were issued.
5 3. England.
Morris’ example of working with a private press, types, paper, stc. was soon followed by others. But, just like in America, the future course of English typography was originated by those, who first turned away from Morris’ sombre, Nordish style. This is no depreciation of Morris’ merits, but a mere stating of the too particular character of his taste, which was bound to lose its grip on anyone who did not stand under the compelling influence of his personality 1). The ambition of several of his followers was greater than their abilities, though their consistent striving for beauty in printing during that period is already sufficiënt to earn our gratitude.
Hence we mention only the Doves Press, conducted by Morris’ former co-operator Emery Walker and by Cobden Sanderson. The Doves Type is comparable to Bruce Rogers’ Montaigne (that also appeared in 1901) in respect of the lucidity of these types, the entire absence of cloggyness such as the other Jensonians show. Both types and the way they were used created something of a sensation in the typographical world of that time, and definitely set up the classical Anglo-American style which still dominates typography, in other countries too.
The leaders of this movement use only old style and transitional types, and from the modem faces only those, that have a distinct Angücan flavour, such as Scotch Roman. According to their point of view Caslon remains the alpha and omega of type design. As an exception Bodoni (in its lightest version) is allowed for “very Italian affaire”, but with apologies. Any 20th century type, that does not claim an immediate descent from the good, old classics, is refused on principle. Mr. Updike’s brilliant Standard work on Printing Types is even dangerously misleading, when dealing with 20th century designs, by its incompleteness and the disproportionate attention paid to types that are second-rate, judging to Mr. Updike’s own standards.
In its kind, the style of this school has been brought to an absolute perfection. But the Continental mind is affected in about the same way, as expressed by a French critic, Bertrand Guégan, in his exclamation on American typography : “On
i) The excellent appreciation of Morris and his work by D. B. Updike (Printing Types, 1937, II 207) is strongly recommended to the reader.



pardonnerait toutes ses erreurs au typographe en échange d’une audace, qui jamais ne se manifeste. On en vient a souhaiter — comme il nous arriva de le faire devant des livres anglais ou hollandais d’une beauté glaciale — qu’une faute, une grave faute, rende un peu de vie et d’humanité a des corps sans ames.”
No wonder there is little understanding or none at all for French and German types and printing. Stanley Morison, though, showed his appreciation, very much to Mr. Updike’s painful astonishment. It must indeed have looked as hightreason to the latter (who thinks “Tudor Black” and the lettre batarde by E. Mouchon delightful) when Mr. Morison wrote : “black letter is in design more homogeneous, more picturesque, more lively a type than the grey, round roman we use”, though the end of the quotation will have soothed his hurt feelings : “but I do not now expect people to read a book in it.”
As the German character is truly reflected in the Fraktur (and the Textur and Schwabacher), with “das Faustische” in it, and with its dualism of the cold, rigid system versus sensitiveness and almost overflowering phantasy, only those, who carry something of it in themselves (due either to their race or to personal disposition) are able to understand and appreciate it to the full.
But this circumstance does not entitle those, who are aliens to this world, to pronounce a judgment on it. The black letter is neither an undue continuation or revival of an alphabet that is bound to a foregone period of style, nor can it be reduced to the expression of bad taste. It is neither old-fashioned, nor inferior in quality.
Perhaps here the objection will be made that someone who is non-conversant with the Anglican spirit, in his turn is not entitled to criticize that black letter which is the expression of that spirit. The answer would be, that the black letter stands or falls with the vigour and freedom of its curlicues (Schnörkel) — as the adversaries call it —, and that the types, praised e.g. by Mr. Updike, lack these very features 1).
The fundamental difference between the Anglo-American group and its adversaries is not of purely aesthetic nature, it is finally not a superficial question of taste. Taking as representatives
!) Of course, it is an unadmissible limitation of the rights of the historian, if he is not allowed to give an appreciation of his subject on the score of “lack of understanding of the spirit” of the subject in question; but I just doubt the scientific character of Mr. Updike’s judgment. A scientific historian can attach a higher value to one style above another, but he cannot “reject” it altogether, if it is an original one, with clear and particular features, and not an imitation.



for each group the writing masters Edward Johnston, resp. Rudolf Koch, we are able to illustrate this point clearly.
Properly these two artists have much in common. Both honoured the traditional forms, but did not hesitate to design a sans-serif type ; both introduced a personal note in their renderings of the classical model, hut were far removed from taking such liberties as Rudolf von Larisch did ; both gathered a large group of enthousiastic followers, who showed much sympathy and interest for each other’s work. Of Johnston, however, never can be said what Georg Haupt wrote about Koch1) : “Texte, die ihn plagten und nicht loslassen wollten, hat er sich von der Seele geschrieben.” Johnston’s manuscripts are beautiful, Koch’s in the first place expressive ; he introduced this element consciously. To quote Haupt again : “Das Wort will sich zu plastischem Erleben gestalten.”
We experience Johnston’s personality from his manuscripts only insofar as every writing, however closely following a fixed model, retains some characteristics of the writer ; Koch, on the contrary, laid his soul open in every letter. This difference is founded upon a difference in temperament between the English and the German race. If we are allowed to exaggerate a little, we should say that the Englishman is hurt in his social feelings by such indiscreet display of personal emotions. Hence his types keep also on general terms and they do not reflect much more than the general spirit of the nation of that period. To the Continental mind they are formal, cold and all very much alike. They are used for the most divergent kinds of texts. The Continental type, on the other hand, has strongly personal features ; it can be used only for suitable texts, but possesses, in that case, a higher expression-value than the other ones. For displaywork types with “special effects” being needed, the Continental designers can fulfil these requirements better.
Contrary to the demand of uniformity with a certain model, which does not leave room for appreciation of almost the entire type production outside the Anglo-American group, we seek the criterion of the artistical value of a typeface in the degree, in which the designer has succeeded in realizing his intentions without neglecting the reading function of his material, and in the value of these intentions themselves.
Obstructions in the realization of the artist’s intentions are presented by the limits of his abilities and by the nature of the material itself. To illustrate this point we remind of the attempt
i) “Schrift und Handwerk”. Offenbacher Sonderheft of Philobiblon, 1934.



of Hugh J. Schonfield x) to write Hebrew characters with the features of Cochin, Cable etc., which showed, though it was very cleverly done, the fundamental incongruity between the natures of the material and its decoration. The Jewish culture of nowadays may have adapted itself to Western culture, but there is a yawning abyss between the spirits of the Hebrew alphabet and the Cochin.
More examples are mentioned in other paragraphs of this chapter, where the irreconcilability has been stressed, e.g. ofpersonal handwriting with printing types ; of the spirit of the Offenbach circle (Rudolf Koch) with the modern sans-serif style ; of the spirit of the Anglo-American school with blatant advertising types, etc.
The artistical value of the designer’s intentions constitutes another important factor. It may be possible that the observer does not sympathize with, or cares for the emotions of Eckmann, or Behrens, Koch, Renner, Naudin or Cassandre, and he is free to reject their types, but he cannot deny that the personalities, which are thus addressing him, are standing on some level, whereas it is the triviality and narrow-mindedness of the spirit, expressed in such types as e.g. the modern faces of the late eighteenth century, or the Monotype-Scripts, that can and should be sentenced by everyone.
The Anglo-American school, as I said before, has found its ideal in the original types of Caslon and Jenson. It is very difficult to describe the features of these and similar designs, because they consist rather of a complex of relations between contrast, relative size, way of thickening, etc. than of peculiarities of the basic form. In fact, their forms are so plain and seemingly common-place that, according to Morison, a new type is probably good if the amateur does not notice its novelty. Even if one takes this remark as a pleasant exaggeration, the underlying thought, that the conformity with the traditional forms has to be considered as the hall-mark of quality, reflects truly the norm of appreciation of any future type. It explains the fact, that, whenever the bearers of this ideal were forced by circumstances to design, to use or to criticize so-called “fancy types”, their choice reveals a curiously erring norm. The favourite is either stiff and awkward, or so extravagant and unruly that it has no equal, except in Germany. It looks as if the designer has thought: “if I must drop the good old tradition, everything is allowed”.
x) The New Hebrew Typography. Hugh J. Schonfield. London, Dennis Archer.



This controversy between design of the highest possible Standard, if shaped after the classical model, and of the lowest Standard, if it is a free invention, governs the British production.
Of the classical style three types by Herbert P. Horne, viz. the Florence, Montallegro and Riccardi, have been designed for private use. They proved to be milestones on the way to classical types for the professional printer1). It was, however, not before 1912 that the first of these became available for machine composition, viz. Monotype-Zmpraf, designed in collaboration by Edward Johnston, Gerald Meynell and A. H. Mason. A long series of beautiful types of the past was to follow, either copies or close imitations of historical models.
The Monotype, with its splendid staff of artists, has led the way here, and the results of its untiring efforts met with a warm approval by the users of this machine, who are spread all over the world. The share of this company in the upheaval of typography from the depths of thirty years ago toward its present height cannot easily be overrated.
Of its classical repertoire we mention : Bruce Rogers’ Centaur, to which a fine, chancery italic, called Arrighi, has been added by Frederic Warde ; Veronese, another Jenson-interpretation, but more in the line of Morris’ Golden Type ; Poliphilus and its italic Blado, based on Aldus Manutius’ Hypnerotomachia Poliphili, portly, heavy and pure, for use on rough stock ; Bembo, also with a fine italic, more regular and delicate than Poliphilus ; Garamond, in the version of Morris F. Benton ; Plantin, a free rendering of the French 16th century types, an all-round useful commercial type ; the venerable Caslon, simple, honest and brilliant. This old-style group alone opens a multitude of possibilities, because every weight and every degree of refinement is represented. But the transitional and true modern faces are also fairly complete : Fournier, small, narrow and delicate, almost anaemic ; Baskerville, perhaps the most successful member of the whole group, at the same time beautiful, distinguished and legible, and therefore highly versatile ; Bell, fresh, racy and tender ;
l) The direct influence of private founts upon commercial type-design has been doubted by nobody less than Updike. In this respect he attachés only a value to these founts insofar as they “stimulate the eye and remind printers of standards set by the past”. Though indeed, the cases are not so many that a commercial type can be traced down directly to a private type, some are to be found, and I am convinced that without the private presses, the present Standard of typography would not have been reached by a long chalk. Now that their task has been taken over by the better-class publishers and printing-offices, we witness a decrease of their number and of their activity.



Scotch, plain, vigorous and unpolished ; finally three Continental types : Pierre Didot, Walbaum and Bodoni (the Benton edition). It is difficult to bring out the proper significance of this oeuvre. Whereas the design of the“individualistic” types is an achievement in itself, which can be called important regardless of the extent of its practical application (not seldom the samplebook of the foundry, in lay-out by the designer, is at the same time the first and the best presentation of such faces), the “community” types of the Anglo-American school, on the other hand, get their true value only by the interaction between producer and userx). Because these types are not meant for a few virtuosi, who can handle them in the proper way, but for any good composer ; their perfection and variety would have been wasted if not a large group of typographers had been ready, sufficiently prepared to make a good use of them. Their importance for the entire printing-art of this century would not have surpassed that of a series of historical curio’s but for the example, in words or practice, given by such men as e.g. Morison, Updike, Rogers, Meynell, Simon, Curwen, by the Oxford and Cambridge University Presses, the Chiswick Press and many other printing houses.
The Linotype too enlarged their range of classical types. Edward Prince’s Venezia, Benton’s Cloister, the Stempel-Garamond, Caslon, the Stempel-Baskerville, Scotch and Bodoni were taken over, whereas two excellent new designs, by G. W. Jones, were produced. The first of these two, called Granjon, is regarded as one of the best renderings of Garamond, so, Updike says, it is “strangely misnamed”. Well-balanced, strong, clean and simple, it is a type for both producer and user to be proud of. The second was baptised Estienne, an equally pure face, but somehow lacking the warmth of its older companion.
The types of English and Scottish foundries of the year 1800 still strongly appeal to the British taste. Some of the original matrices have been preserved, and from these matrices types are cast again nowadays. In other cases recuttings have been made. We mentioned Scotch, a collective name for various related designs, one of which goes back to Alexander Wilson of Glasgow ;
0 The admirers of these series will make the sarcastic remark that this subtle distinction between “individualistic” and “community” types is a long way for confessing that Caslon, Baskerville, etc. are good and therefore popular, and that Eckmann, Koch Antiqua, etc. are bad and therefore avoided, or that the former are real printing-types and the latter a kind of still life pictures in disguise.
The fundamental mistake in such an objection is due to an unjustifïed emphasis of the importance of reading-speed. The chapter dealing with the atmospherevalue of type-faces substantiates this counter-argument.



Bulmer is the name of a series, produced by the American Typefounders Company after types, made by William Martin for W. Bulmer, printer of Shakespeare’s works. Likewise several Products of the Fry, Thorowgood, Thorne and Figgins foundries have reappeared.
The production of entirely original designs, however, has not been neglected. In comparison with other countries the English show a lively interest in letter design ; their schools of arts and crafts keep it to a high Standard and calligraphy is practised on a large scale. Edward Johnston, the great writing-master, has exerted influence in broad circles, also beyond the British Isles. One of the best examples of his style is presented by the Ganton series of capitals. Next to him his pupil, the sculptor Eric Gill, has to be mentioned. His best alphabet, laid down in MonotypePerpetua, is more severe than e.g. the Ganton series, it has more contrast too. This sharpness and contrast reduces its legibility in the long run, but in itself the forms of the lower case — and especially of the capitals — make it a masterpiece. Another design by Gill for the Monotype, Solus, has more colour; it is a sound and able design, and pleasing, but it does not reach the Standard of Perpetua. Four other types of his (for private use) demonstrate his versatility and inventiveness, at least, for English relations.
Dolphin Old Style, by G. W. Jones, is an early example of a sound, modern interpretation of the classical theme. MonotypeHorley is less sound, its lower case g is particularly objectionable.
The Röyal Academician F. Landseer Griggs designed Leysbourne for the Monotype, a well-drawn, honest, open face, in a pen-technique seldom used by the English in this way. Another type of his, Littleworth, shows the same sound ductus, but it is a little too broad in my opinion.
The reviewer is inclined to doubt the value of such types and their numerous colleagues, that have not been cut directly (üke Caslon, etc.), but after a design that imitates the forms of this former category, though any weaknesses in the single characters cannot easily be demonstrated. It is rather puzzling why these types, which are, artistically speaking, indefinitely better than 90 % of the material which is in daily use, should raise such doubts. The only possible answer is to be found in their lack of unintentionality. They have neither the legibility of the “community” types, nor the expression value of the “ individualistic” ones. On the score of this consideration the chances are greater that we can call a member of the latter group successful, for, if the intentions of the designer went in the direction of promoting a maximal speed of reading, the ways by which they can be realised



have been explored by a multitude of able artists and craftsmen, whereas personal expression cannot follow the beaten track ipso facto. On equal amounts of entirely new types in both categories, therefore, an expressive type is more likely to have realized its purpose.
Of course, there is no simple inverse proportionality between legibility and expressiveness or beauty, though e.g. the distance of recognition will be considerably larger for Century Schoolbook than for Koch Antiqua (Locamo, Eve). In practice such a safety-margin is taken, in respect of the size and the length of the spells, that no reasons are left for an absolute rejection of the expressive, individualistic types on the score of legibility. Hence two standards of value can be applied, according to the properties which the designer has chosen to emphasize: either legibility or individualistic expression. Where we are confronted with types, that follow closely the classical model, we are justified to ask also for a maximum of legibility, with a slight allowance for their aiming at new tone-values, etc.
Returning to the starting-point of our reasoning, i.c. the designs of F. L. Griggs and of the numerous artists of, in this regard, fundamentally similar spirit, we come to the conclusion that the norms of Updike, Morison, c.s. are absolutely valid here, and that the lack of appreciable advantages over the masterpieces of the past does not justify the cutting of many of them.
An example of a highly important new type is presented by Times New Roman, for which Stanley Morison himself was largely responsible. In the paragraph concerning the composition of newspapers we had the occasion to deal with it circumstantially ; let it be said here that this face seems to be the only satisfactory alternative to the Ionics, with regard to legibility in the smallest sizes and space-saving.
Next some types will be mentioned, which I have only seen in the form of a 36 point alphabet, but that seem to belong to the class of Kleukens Antiqua, Behrens Mediaeval, Belwe Antiqua etc., and therefore do not make a more substantial account necessary. We note : Gayton, Healy, and, more free, Richmond, Trajan, Worcester; unforgiveable deformities have to be noted in the Bloomsbury series.
Of the display-types Gill Sans comes in the first place. It is the British answer to the German modern sans; its remarkable success proves that both designer and producers have struck the right note.



In the course of time something of a legend has been woven iround the genealogy of modern sans types. English authors naintain that they are to ’be reduced to a type, which Edward Iohnston designed in 1916 for the London Underground. German irtists, it is said, have been inspired by that type on the occasion af a visit to England.
The incorrectness of this statement, however, must be clear to anyone, who compares the spirit of the English and the German designs ; a direct artistical relation cannot be laid. If e.g. the ideal English sans, that of Johnston’s pupil Eric Gill, is opposed to the best German interpretation, the Futura of Paul Renner, the fundamental difference is salient enough. The Gill sans is drawn as a pure roman, with equally thick strokes and the serifs stripped off. The Futura, on the other hand, is constructed ‘with a pair of compasses and a straight-edge” (words of Rudolf Koch, quoted by Waite), not after the roman model — as Koch did too — but after a new model, obtained by analysis of the former. The main difference finally has not to be sought in the nature of the technique, but in the model. Renner is an adversary of both Koch and Gill, when he thinks that: “Schreiber sind pseudo-mittelalterlich’’, and “ohne Bedeutung für die Graphik der Zukunft".
To summarize : Gill Sans has its basic form, the roman model, in common with Cable (or Kabel, i.e. Koch’s sans), whereas its technique, which settles the details of the form too, is a feature of its own.
In the course of years the Gill Sans series has been completed in all sizes and weights, decorated versions included, e.g. Kayo, which is so extremely bold, that some alterations in the shape have been necessary, which makes it comparable to such stencil types as Futura Black and Monotype-Braggadocio ; further : Cameo and Cameo-Ruled, Shadow, Outline and various condensed series.
For the festival occasion of the jubilee of King George V a bastard script has been designed, also by Gill, called Jubilee. It is a less successful work than Mr. Gill’s usual book types.
A former member of the Offenbach Circle, Berthold Wolpe, designed the Albertus capitals for the Monotype, which are plain, sturdy and with just that personal, unmechanical, unpolished touch, which is uncommon in England.
The Felix titling capitals, kind and regular, and Jocunda decorative sans capitals were issued by the Monotype, which firm is paying considerable attention to the improvement of Capital titling series. Their Colonna type has been made for headlines in the refined tone; it consists of parallel strokes ( open face ),



that are separated a little too far in respect of the legibility.
Basuto is a British contribution to the extra-heavy kind of display capitals ; the Ultra-Bodoni species is represented by the original model, Elephant, cut by Vincent Figgins more than a century ago, while Chatsworth falls in the line of Cheltenham Bold, Glenmoy in that of the slow, upright scripts.
Resuming the main points of this paragraph, we find in the British type production a marked superiority in the classical style, together with a marked inferiority where entirely new forms were needed. When we did not share the opinions of the British (or related American) authors, or even attacked them, it has not been done because we do not admire their work for the upheaval of the art, on the contrary, we did so because of their very authority in the whole world and the unassailable value of their work. For if everybody agrees with that, the possibility exists (and it is sometimes an actual fact) that one will follow these convincing theorists in opinions, which are questionable because their fundaments depend too much on time and place, and which therefore invite opposition. A typographical author, who ignores modern French printing and to whom fine printing in Germany ends with the year 1914, ought to meet opposition, however great his merits are as historian and as practical typographer.
§ 4. France.
In France extremes are found side by side : next to a number of foundries, which do not take any notice of what is going on in the rest of the world, the firm of Deberny & Peignot is always busy with fruitful and inspiring experimentation. This fouridry has been the first to produce a type of the “art nouveau”, designed by Eugène Grasset, a heavy face of original proportions and ductus, being quite a revolution against the usual didots and elzévirs, and more typographical than Eckmann’s contemporary design in Germany. The soft, full contours were practised in the new poster art in which (beside Grasset) Toulouse Lautrec, Chéret, Steinlen and Mucha were masters.
They raised the poster from being a large illustration to its proper function, i.c. to draw attention and to convey a message in a simple and effectual way. Their lettering had to make use of other forms than the printing-types offered at that time, with their thin hair-lines and anaemic, monotonous rigidity. Therefore they used intentionally shapeless daubs, easy to distribute over the sheet; serving as well as patches of colour as for making words.
Legibility, atmosphere-value and forms of printing types. 14



Some years later the Grasset was followed by a character of jeorges Auriol, consisting of loose round brush strokes, obviously jetraying the influence of the chinese-japanese lettering. A heavy iisplay-face (Robur) was added later on 1). Jules Hénaffe cut :he characters, theoretically developed by Jaugeon in his treatise }f 1704 for the Académie royale des Sciences, which probably nfluenced Grandjean in his romain du roi. It is regrettable that his design drew so little attention. The Deberny foundry2) ilso edited a série 18, used by Javal in his Physiologie of 1905, in England known by the deceiving name of Baskerville O ld Style, i clean cut type, rich in contrast, with a peculiar italic. In 1912 some remarkable revivals of the Louis XVI period were issued Linder the name of Cochin. These were no strict copies, but they were rather inspired by the lettering of some engraving masters of that time. Cochin exists in two versions of different lesign, although both versions show the same style. One of these is a nice bookface with much contrast, a little too round ind broad to be easily legible, but possessing a peculiar, delicate itmosphere. The other series is hardly a bookface, but very suitable for smaller texts and stylish publication. With its very original forms, inventive pen-work, and extremely long descenders and ascenders, it set up a new Standard in design aiming at refinement. lts popularity is still very great, its influence on other designs enormous. In the course of the following years some smaller series, Moreau le Jeune, an open version, and Fournier le Jeune, decorated italic capitals, were added, so that we are now in possession of a group of excellent types and many carefully selected ornaments, vignettes, etc., which faithfully reflect the late 18th century style. All these designs were either adopted or imitated by foundries in the principal countries.
At the same time the well known illustrator Bernard Naudin produced an alphabet named after him ; it is done with a few firm and fluent strokes and consequently looks as if casually sketched. The design lacks accuracy and precision and has a too personally confined style than that there could be a possibility in using it together with other types or illustrations not showing the same features, e.g. photographies, lino- or woodcuts, etc. Dorique is a private experiment of the printer Léon Pichon in collaboration with the illustrator Carlègle. The design betrays the amateur, but should be noted as a return to the heavy
1) The German foundry Genzsch & Heyse acquired both types.
2) The fusion with Peignot did not take place until 1921.



modem face, though less severe. Garamont1) is an adaptation of a type, used in France in the 16th century, that was either cut by Garamont himself or by an imitator. The Deberny fount is a very satisfactory one, though a little too narrow.
Robert Girard, one of the directors of the foundry, designed l'Astrée, a rather heavy, small and narrowfaced type, without the slightest platitude. It is surprising that both Garamont and Cochin are traceable, the former specially in the capitals. Some years later a kind of “Normande” was cut, the Sphinx, having rounder forms and thicker serifs and hair-lines than its model.
In 1929 a work (one can almost say : an invention) by the poster-artist A. M. Cassandre appeared. It is called Bifur, abbreviation of bifurcation, also the name of a periodical of internationally orientated young artists. And indeed—this name fits the principle of the design excellently, for here a conscious division is made between the “important” and “unimportant” parts of a character. Very much the same idea was followed in the heavy, modern faces (“Normandes”), but less consequently, because vertical stress has been applied, e.g. in the cap. E, where, however, the vertical is relatively unimportant. Moreover, the unity of the character has always been preserved. The Bifur is available in two series, one with the unimportant part hatched, the other (Doublé Bifur), consisting of two separate types for each character, one bearing the face of the important part, the other bearing the hatched part of the first series solid for printing in colour. The principal significance of the design consists of the recognition of the fundamental importance of the internal surfaces of a character for recognition. Paul Renner realised this too in his Futura Black. If a designer wants to make use of this fact, he does not need to make these area’s visible (by filling them out, as in the two types mentioned) ; he has only to reckon with them in the planning of the contours, i.e. in the position of the limbs. In other words : he must learn to see the strokes, out of which a letter is built up, not as strokes only, but as the contours of the area’s enclosed therein. The effect of the Bifur is planned just as that of three superimposed sheets : first the background (paper), then the hatched or coloured part, with the main strokes on the top. In his lettering of posters this tendency of Cassandre’s is noticeable : that he works with planes and regards the characters as cut out in ribbons. Likewise Bifur is
9 The historically correct spelling with t has only been adopted by the Deberny & Peignot Foundry, the Amsterdam Foundry and by the American Monotype.



tot meant as the impression of a stamp by typographical means, >ut as painted with a brush within constructed contours, through i stencil plate, or cut with scissors ; appearing as addition, as >verprint. A later type of his, Ader, did not reach by far the ligh level of Bifur ; it is rather a “Spielerei” with sans-capitals. rwo other Capital series of this foundry (Éclair and Film) could )e useful on single occasions.
Influenced by metal lettering on shop-fronts, a display-type :alled Vlndépendcmt has been designed by two Belgian artists : j. Collette and J. Dufour. The characters are shaped by ;utting little pieces out of solid planes. It has a too affected ind irregular appearance and is extremely illegible.
The World Exhibition at Paris in 1937 saw the first appearance af a new type by Cassandre, designed for Deberny & Peignot, malled Peignot. It was extensively used at that occasion for /arious typographical jobs, but also for inscriptional purposes n wood, plastics, etc.
The foundry has issued a circumstantial commentary to this type, making an attempt to prove that the Peignot means nothing less than a fundamental innovation of type design x). The basic Form of our alphabet, it is argued, could be developed as long is it was exclusively written; the present form, on the contrary, is almost the same as the one, which accidentally existed at the time when the first typographical alphabets were cut. Nobody will sincerely maintain that this form cannot be improved, for several letters, especially of the lower case, have a poor legibility. The Peignot takes up the line of development, reckoning with the special typographical requirements. The weak letters of the lower case are substituted by the corresponding Capital forms and the serifs are omitted. Apparently, the modern sans-serif has served as a model for the Peignot because of its legibility. We do not need to point out any longer that for bookfaces no legibility can be achieved in such a way. For display work, however, it is a very nice type, and the Capital forms in its lower case give surprisingly little trouble to the skilled reader, though just enough to make it unacceptable for the average printer, who cannot afford to take a risk. This leads us to the principal objection against this and other possible reforms of the alphabet, viz. that no deviation from the usual form, however small, will open possibilities in respect of greater legibility, if the public is not practically forced to read no other type during a time which is sufficiently long for training. The circumstance that at present
') See also: Arts et Métiers Graphiques, 1937, No. Spécial 59.



no new forms are introduced, but only equivalencies, cannot alter this fact.
The makers of this design are likely to forget that their own way of reading is superior to that of the multitude, which reads far more slowly and therefore with considerable more attention for each single character. Moreover, the inertia of the public in becoming used to new forms cannot soon be overrated. If therefore this public is free to reject any unusual piece of print on its first impression, just because it is annoying, the chances for carrying through a reform of the alphabet become very small. Only a persistent use by influential instances, e.g. by newspapers or large-scale advertisers, could meet with success.
Let us return to the Peignot. It has been said that this design is a further development of our alphabet, which has remained fundamentally unchanged since the invention of the art of printing. The question arises which norm has to be applied to such a creation. For the factor of easy writing, which influenced the forms of the alphabet at the time of the scribes, can have no actual significance any more ; we have to reckon with the technical requirements of typography nowadays. The Deberny foundry pretends that the Peignot is the only possible solution, but why this should be the case, is not further proved. We should say, on the other hand, that a solution only can be found after laborious investigations by the co-operation of typefounders, printers, specialists in the psychology and physiology of reading and artists, such as has been the case in the development of the American newspaper types in the Ionic style, and with the “Times New Roman. The types mentioned above are fundamentally and radically different, thus proving that an inevitably logical solution has as yet not been found. And even, if one starts from a purely aesthetical point of view (which is not done by the Deberny foundry), not one, but many possibilities appear to exist. In all cases, however, the advantages above our present alphabet are at least dubious.
The conservatism of the Anglo-American school, therefore, (as it has been put forward by Mr. Stanley Morison) has many points of defence. That it will be attacked as long as the art of printing is alive, will not be regretted by anyone, who keeps an open mind. Neither should it be rejected a priori, as has been done e.g. in comparison with the thoroughly unsound repetition of past styles in the last century. For a letter is not only an object of art, but also a reading-tool, the usefulness of which is based uoon the work and the experience of thousands of



people, gathered during several hundreds of years. Such a comparison is therefore not justified.
The French are masters in black and white. Perhaps owing to financial reasons, compelling to get the maximum profit out of a few types, they reached a high level in the art of “mise-enpage”. This technique is very difficult; it can hardly be acquired, but is founded in artistical talents to a large degree. In other countries a composer is soon inclined to recur to types with a special sphere, to another colour, paper, etc., if special effects are required. In France one is able to manage with “old-fashioned” types in black printing on all kinds of paper. Nowhere else it is realised so clearly that the principal means of expression is “ mise-en-page”, provided that paper and types are good in their kind. This, however, is'no ideal state of affairs. If printers of less typographical feeling are equiped with these relatively poor tools only, the result is the worst imaginable. But it can do no harm to those in possession of more subtle instruments to undergo the influence of Gallic phantasy and airiness and to leam their method of eliciting new elfects from old material. In this respect we should specially mention the pioneering done by the direction of the firm Deberny & Peignot, notably by Charles Peignot, whose influence is comparable with that of Klingspor and the Bauer foundry in Germany and of the Monotype Corp. in England.
§ 5. Germany.
In Germany the upheaval of printing art called forth response in broad circles. Occasionally there had been attempts towards a simpler, cleaner style. The editors Velhagen & Klasing issued fairly good bibliophile works as early as 1875 ; the “Romanisch” type of Genzsch & Heyse has been completed by Heinz König in 1888, but still the new form searched for had not been found, and many problems were waiting for a solution. The controversy about the German and the Latin alphabet stood in the focus of interest, though it was not to reach official instances before ca. 1910 (debates in Parliament, experiments by Kirschmann). The national feeling required a type of German characteristics. This demand was provided for by Otto Eckmann in his type, designed for Gebr. Klingspor (then Rudhardsche Giesserei, cut 1899, release 1900). The design is of such importance, that we have to go further into its history.
In 1895 the periodical “Pan” was founded, a year later the “Jugend”. Both magazines served the same purpose : to present



the literary work of young artists in the best possible form. Many capable designers of illustration and decoration were associated ; the printing was done in artistically prominent offices.
Eckmann co-operated too. In 1898 he designed the famous 7, which adomed the cover of the periodical “die Woche” during a long period.
By that time Karl Klingspor had bought an insignificant provincial type-foundry, the Rudhardsche Giesserei in Offenbach am Main, in the neighbourhood of the old centre of type-founding : Frankfurt. Together with his younger brother Wilhelm, who died in 1925, he set up an example that all other foundries had to follow : he promoted the type-founder from being a commercial furnisher of technical supplies to a co-operator and artistical pioneer. They were the first to install a printing-office to the foundry, solely devoted to experimenting and the production of the finest models.
Eckmann called his type a “decorative Schrift”. In the introduction to a specimen-book he attacks those, who cling to the old forms of the incunabula or even the mediaeval scribes as well as those, who forsake the traditions, in vain desire for originality, by adding all sorts of “curlicues”, which diminish the legibility. This refers to Morris and his followers, to Rudolf von Larisch in Vienna (who experimented rather wildly) and to the American types, which had come to Europe by the dozen, and were of an abominable design at that.
Eckmann wishes “to express artistical ideas on the base of the well-tried practice, which confines itself to the composition of the single letter and its total character”. It is quite clearly shown here that the form of each character has to be considered as a separate unit. This again is addressed to the theory that the printing-type has to keep the characteristics of the handwritten letter. “We have had time to think it over during four centuries and a half, and we must come to the result that with regard to the manufacturing methods of nowadays, such a theory brings about an unnecessary handicap to the artistical shaping of a type, while it gives no value whatsoever in exchange”.
Eckmann’s type was a fresh start in design, one of the most radical changes in the history of printing. lts influence on German type-designing has been enormous ; many features of productions of nowadays can be put down to it, though often only in the form of an adopted principle. Although the type was not explicitly intended to be a compromis between “German” and “Roman”, the influence of Textur and Fraktur is undeniable. The letters are treated to fill a rectangle, while keeping a minimum



of rigidity ; each is “statically balanced” (the tight fit of the rectangles is supposed to give good word-forms), every single line or curve is moulded, according to the style principle.
Perhaps never a type reflected more truly a style of art, but also never has a type distorted more violently the traditional forms. If each type-face is subject to the controversy “decoration versus reading-tool”, Eckmann’s type presents the extreme case of the former. German designs have shown the utmost individuality since ; their pioneering work presented us with daring designs, it is true, of very divergent merits.
The architect Peter Behrens, who remained a Champion of the writing-theory, presented a type in 1901 at the same foundry. This type too has a distinct German flavour, though of an entirely different kind than the Eckmann type. One feels the severe and rigid theorist behind its cool, consequent ductus ; it raised the letter-writing technique to a height, since a löng time unknown.
Both types were very completely equiped with corresponding decorative material. In numerous specimina various applications were shown, with new colour combinations (grey and chocolatebrown, olive, wine-red). In 1907 several new interpretations of the pure roman form were given. Among the lettering artists a punch-cutter of origin — and a very capable one — named Friedrich Bauer, designed the Genzsch Antiqua, a crisp, vigorous character, that was the first of a long series of new mediaevals to come. Saeculum by H. Hoffmeister follows the “Jenson” style. Georg Schiller, who had designed a Neudeutsch in 1900 too, edited the Rühlsche Antiqua Venetia, with handwriting features, looking as if written with a quill rather than with a steel pen. It influenced the Behrens Mediaeval of 1914. In 1905 the König Antiqua and the Schelter Antiqua had prepared this special style.
The development of poster art brought about a quiet, clear lettering, which consequently has been transferred into printingtype. In 1908 the Block appeared, a heavy, square grotesque of clean structure, without sharp edges or abrupt tums. The best masters of poster technique subsequently designed display-types : Lucian Bernhard, Louis Oppenheim (who probably lent a hand to the Block too), Martin Jacoby-Boy, F. P. Glass, Julius Gipkens.
In 1908 F. H. Ehmcke designed his Antiqua, a new turning point in type design, which still stands unrivalled in many respects. It represents the first attempt to introducé an atmosphere of a refined haughtiness. The technique of its delicate lines and the extremely long ascenders has become a Standard for this kind of types. Ehmcke reached this height never again. His former co-operator



at the Steglitzer Werkstatt, F. W. Kleukens, designed a sober, vigorous Antiqua in the same year, and Behrens did the same. His way of holding the pen at a right angle with the line influenced Kleukens and some others in a later design. By that time the well-known printer and book-lover, Carl Ernst Poeschel, founded the Janus Press with Walter Tiemann. The type for this press has been developed by the latter into the Tiemann Mediaeval, which appeared the next year at the Klingspor foundry. It shows a much lighter touch and a more playful phantasy than its predecessors mentioned above. We must not forget that since a decade Edward Johnston worked for a pure Roman form by means of a highly developed penmanship. His work had a great influence on the continent, where his pupil, Prof. Anna Simons, served as intermediary. She translated his “Writing and Illumination and Lettering” (1906) into German, co-operated at the Bremer Presse, but unfortunately never worked for a foundry.
In 1913 Georg Belwe added his interpretation of the roman to this series. Some characters of this face clearly demonstrate the limitations of this kind of pen-drawing, just as the mediaeval of Peter Behrens of 1914, mentioned above. The latter is a weak design. Two other faces in the same technique came out after the war : Schneidler-Latein in 1919 and an Antiqua by Julius Klinger of Vienna in 1920. The world-war caused the retardation of the appearance of the Maximilian capitals by Rudolf Koch, the first roman by this eminent artist, whose Deutsche Schrift had been issued in 1910. The Maximilian is no more than an announcement of his later achievements. Four years later, in 1921, the same foundry edited Tiemann’s Narziss, the finest open face ever produced. lts name is well-chosen, for it is both fresh and graceful.
Koch, who certainly had the Fraktur in his finger-ends, doubted his own ability to design a roman, but he was encouraged by the Klingspor foundry to have a try all the same. The result, the Koch Antiqua of 1922, was brilliant, and it gained a wide popularity in England and America too. This type seems to vibrate like the air in springtime; the delicacy of its lines is not enervated, but fresh.
A new roman by Tiemann followed the year after and was of the same sound ductus as Narziss and his later romans.
Karl Klingspor has allowed both, Koch and Tiemann, to unfold the full range of their abilities, each in their own style. It is very difficult to describe adequately the character of their work, roman and Fraktur types, graphical work and other products of appüed art. An extremely felicitous expression was found by Julius Rodenberg, who compared Koch with Bach and Tiemann with Mozart. The former forceful, terse ; simple by the humble



devotion of the Christian faith behind it, though with a wealth of forms, sometimes delicate and playful, but never gay in the same way as the latter, whose charming grace keeps aloof from any serious theoretical considerations, whose pure and joyous forms betray no struggle, but a peace of mind, to which an ardent passion is strange.
Directly opposed to this kind of work were the ideals of the post-war generation. Nearly all the types, mentioned hitherto, were designed by advocates of the writing theory. Though cleverly done, they were encumbered by a mass of details and refinements, which set the whole page squirming and took the punch entirely out of it. The anti-decorative functionalism of those years opposed against them. Paul Renner, who was typographical advisor of the publisher Georg Muller of Munich and had always used a clear, simple, black type, made the first design for his Futura by that time too. It has been said that the modem sans has been derived from Edw. Johnstons’ London Underground type. A trial-sheet of Renner’s type, dated April 1924, in my possession, shows on the contrary an entirely different manner, a thoroughly experimental, constructive mind. The design was too advanced for its time and therefore the edition has been delayed for commercial reasons till 1927 (q.v.).
In 1925 two types were put on the market, based on previous models, but abandoning the flowing “artistical” ductus in favour of cool, hard lines.
Klinger and two collaborators, Messrs. Frey and Willrab, joined their efforts to produce the Klingertype; its sharp, üght lines recall cuneiform writing. Lucian Bernhard added a delicate version to his roman of 1912, but its ductus is so entirely different from that of the latter, which was done with a soft brush, that in reality it is quite a new design. A beautiful cursive and several heavier grades completed the series in the following years. The soft and easy curves render a sultry grace to it, which made the whole family the most widely used for prestige work.
The need of simpler book-types was met by the cutting of “classical” types. The American Typefounders Co. had done admirable work by rendering Bodoni and Jenson (Cloister) before the war and Garamond at the end of it; European foundries followed in the years about 1925. Stempel published three revivals : Garamond, Baskerville and Janson. The latter is a very fresh, racy face of the Caslon group ; the two others are slightly regularized, but otherwise faithful copies. Mrs. Warde (in the Fleuron No. 5, 1926) has pointed out that only the specimen



sheet in the possession of this foundry can be said to show the real types of Garamond and that the font of the Imprimerie Nationale (assumed to be the charactères de 1’Université of Garamond) is about 80 years younger ; it has to be ascribed to Jean Jannon, a printer of Sedan. Linotype-(7ra«/'ow by G. W. Jones is also a true copy of the Garamond. The other types of this name are all based on the Jannon face, viz. of Deberny & Peignot (partly), the American Typefounding Cy., the Intertype, the Amsterdam Foundry, the American Linotype and of the Monotype. The Stempel-Garamond was edited on German Linotype, the Baskerville on both German and American Linotype. We mentioned the Tiemann Antiqua of 1923 ; the following year Kleukens created his Ratio Latein, a severe, almost dry kind of Didot. The Bauer-Bodoni, though prepared before the War, has been completed in 1923. The Bauer Foundry also edited an allround useful, quiet type (the Tages Antiqua) by its best punch-cutter, Louis Hoell, who had cut many famous “Künstlerschriften”, first for Klingspor, then for Flinsch (which firm was amalgamated with Bauer about 1915). In 1926 again a type of classical purity appeared at the same foundry, a product of E. R. Weiss, an artist who had a long career as a book-artist. Fifteen years before he had designed a Fraktur that had grown popular. After wild years at the beginning of the twenties (work for the Marées Gesellschaft) he prepared his pure mediaeval, which was accompanied by a charming cancellaresca. This sturdy, quiet, clear type is a worthy companion of the Genzsch Antiqua and the Hollandsche Mediaeval, which stand on the same level. Again the theories of the artist are not as good as the type itself. A series of bad initials and a funny alternative Capital A are claimed to be drawn also after careful study of ancient examples. A series of ornaments by Weiss ranks with the best work of Fournier and Caslon ; it is fresh, graceful and gay. When will the man come, who recognizes the value of this splendid material and who has the taste to handle it in the proper way ?
In the course of the last decade the technique of advertising lay-out has been developed, renewed and differentiated. New display-types had to be made to meet the growing demand for eye-catchers and special “feeling-tones”. A good deal of the older designs could not satisfy these requirements ; they were too intricate and scrawly, but specially their soft, flowing lines, due to handwriting or brush-technique, did not suit the sharp pendrawings or foto’s of the new lay-out. We mentioned the functionalistic movement, that had its head-quarters in the Bauhaus of Dessau, afterwards in Weimar, where a group of like-minded



young artists, besides other activities, explored the domain of typography, linked up with photography. Their principles were laid down in “die neue Typographie” 1), a writing by Jan Tschichold ; it appeared in 1927 and became the catechismus of all young designers. The basic principle is efficiency with a socialistic tinge. Not the anti-industrial socialism of William Morris, which retumed to handwork, to produce fine books beyond the reach of the multitude as well intellectually as financially, but a confessed industrialism, which rejects special paper, ink, type and binding. Printing is done to convey a thought in the quickest and cheapest way to the mind of the greatest possible number of people. Hence the psychological laws of attention, perception and memory have to be obeyed and not the traditional schemata, which are accidental or at best caused by decorative needs of a forgotten period. The disposition of the text has to be varied according to its peculiarities ; the text cannot be squeezed into a preconceived form, e.g. along the central axis of the page, the funnel shape, the F shape, etc. ; it must be found at eye-catching places. The range of possible Solutions is very limited and coupled with economical considerations. This is a reason why blanks, letter headings, sizes of paper, etc. are standardized. The ingredients and tools must be of a kind that can be made, supplied and handled by everyone so to speak, to achieve cheapness to reach the mass. No handpress, hand-made paper, hand-composition, hand-binding ; no other colours than black and spectral red, yellow and blue ; no lino- or woodcuts or any other drawings, but photo’s and blocks.
In this way no risk is run of misinterpretation, caused by the addition of “subjective” imponderabilities in the shape of broken colours, decorative types and drawings, etc., which are only a hindrance for a smoothly running conception. Everything that has to be told, can be expressed in the shortest way by the simplest means, which have the same significance and suggestive force for everyone. If an illustration is needed, the true-to-life photography shows everything clearly and in full detail; if a symbol or sign is needed, the primitive forms (arrow, cross, circle, triangle etc.) should be used, because they possess the highest attentionand memory-value.
This point of view has obvious consequences for the types to be used. The characters of the alphabet are considered as mere signs. Because signs are reading-tools and no objects for decoration or direct conveyors of a meaning (i.c. atmosphere), the traditional
!) Now out of print. lts principles can be found in “Eine Stunde Druckgestaltung”, Stuttgart 1930.



types must be stripped of those parts, that were added for these purposes or for reasons of easy writing, till the skeleton remains. The grotesque (sans-serif, gothic) answered this description. Unfortunately in the ordinary editions the pure geometrical forms had been spoiled : squares had become rectangles, circles were flattened to ellipses, etc. In 1926 Jacob Erbar, who had drawn a mediaeval in 1914, edited his grotesk, followed in the next years by Paul Renner’s Futura-type.
The Futura was a “big hit” everywhere, a success, due to its excellent proportions, the general balance, the carefully chosen weight in all grades, the beautiful cutting and justifying. Although the various sans-types are naturally very much alike to the casual observer, the Futura stands out by its quiet and even aspect.
Constructions of quite another kind were made by Ru dolf Koch for the Kabel-type. They show again an aversion of everything portly and heavy ; where no serifs and differences are present to counteract this liveliness, an unquiet, scrawly appearance is the result, though least, of course, in the heavy grade. The fundamental cause of this fault lies in the impossibility to reconcile the ideas of the Offenbach group with those of the Bauhaus people.
The new typography exerted a wholesome influence. Now, after ten years, we can say that it cleared up many abuses, simplified lay-out and guided into new ways, but it had to waive many claims.
The application of psychological laws is not the job of every composer or even designer. The composer must rely in practice on the examples of those, who are able to fix the optically most favourable spot of the page, the graduation of attention-value of the various parts of the text, and who can make a photo or a sign, that is more than an empty symbol. But in doing so they violate the prime law of varying the disposition according to the needs of the job ; consequently they relapse into schematism. It is a more serious indication against these theories that even the leaders did not escape this danger. Tschichold emphasizes the difference between architecture and the new (s.c. “constructivist”) typography. The former is entirely determined by the functions of the constituent parts, while the latter contains an aesthetical element, which makes it comparable with “free” painting. Hence the fact that the pioneers of the constructivist typography belonged to the group of “abstract” painters. The aesthetical element, as Tschichold explains, is represented by the movements between the elements of the page, the conscious use of which is one of the means of the designer. Here is the crucial point, for the functionalistic mise-en-page is often incon-



»ruent with an aesthetical one. If a large amount of mixed text is to be arranged on a small, fixed area, aesthetical appearance md surveyability being required, a subordination of the less important elements is necessary. The new typography, which feels itself “at home” with this kind of jobs, does not achieve brilliant results in practice. lts adepts claim conspicuousness and legibility for every detail and are quite at a loss in which way a graduation is to be obtained. Consequently they are focussing on every point in succession ; in doing so, they are obliged to increase more and more the intensity and extensiveness of their means of expression to obtain the necessary fore- and background, also because the aesthetical movements (balance) need some clearance, if they are to come into being at all. They want to make it too clear for us ; heavy arrows, encirclings, underlinings, heavy types, circumstantial photographies, various colours are all shouting together at the top of their voices, till we finally do not know where to direct our attention. In practice the new typography (quite contrary to its principles) is very costly, which would not have been necessary if its advocates did not overrate our ability to keep our attention strained to the maximum for a long time, whilst underrating our power to fix ourselves the suitable amount and direction of attention and to discern and read small pieces of print.
From the functionalist point of view the sans-serif type agrees best with the taste of our time. Some of the more neutral historical types, like Garamond, Bodoni, Walbaum, are being admitted reluctantly and only because they are plain and legible, but this already is a concession.
If those theorists had stopped to ask why the grotesque is the type of our period, they would have answered : “because it is cool and businesslike”, not because of the legibility alone, for the poor legibility as booktype does not justify its use as such. This implies the acknowledgement of a certain atmosphere-value. The discussion of this problem could be continued into etemity, if the real motives of both parties are not clear. Form-giving in the arts and crafts is governed by two Systems of value : beauty versus utility. The latter can be proved objectively, the former cannot. Consequently it is more convincing to argue for utility. But (explicite or implicite) the ulterior motives of the artist’s pleading for type as a reading-tool are of aesthetical nature. For these theorists refuse to measure the utility of a type by other qualities than legibility, even if experiments prove that the reader feels clearly the atmosphere of a type, and that above a certain



limit no differences in legibility (expressed in speed of reading) exist, especially with display-types, and that simplicity does not always mean legibility. If they therefore ding to their doctrine that faces as Bodoni should be the utmost concession to decorative needs, this happens for other reasons than utility; büt then no practical argument is left for the abolition of designs based on other principles. Altogether there is no reason why we should limit the use of type to the transmission of thoughts by way of understanding, if it is able to convey the purport of these thoughts by means of direct impression at the same time, provided an (easily reached) minimum of legibility is guaranteed1).
The modern sans-serif families were soon very popular. But where the new fount had to be used for a great variety of jobs, it was feit that the ways of expression were limited. Quite contrary to the original principle, decorative series were edited : shadowed, outline, negative capitals appeared ; in some of them a part of the character consisted of thin, parallel lines (comb-written). Even swash-(decorative) cursive capitals appeared, grace lacking badly.
An interesting experiment has been made also by Paul Renner, who designed a stencil-plate type. It is very heavy, eccentric looking and little legible, but a good eye-catcher. Similar types by other foundries are considerably worse.
The lack of serifs was supposed to be the cause of the deficiënties of the grotesque for continuous texts and consequently serifs were added 2). Beton has been inspired by the old variety, called “clarendon”, which had its serifs connected with the stem by a triangle ; it is, to my opinion,/aa7e princeps in its kind. Memphis, by Rudolf Wolf, is a true egyptian.
The German designs of both, grotesque and egyptian, spread over the world immediately; they even reached France, that is using them in an own style. Since their introduction the composition technique has become more daring ; more crisp and loud too. The plainness of this style has been a wholesome tonic against the exuberance of the artistic types of the foregoing period in all countries.
') A century ago the reverse had happened: the sans-serif was developed out of the “egyptian”, as it was called then.
2) Cf the excellent remark by Dr. Konrad F. Bauer (“Schrift und Technik”, in Buchkunst 1, Leipzig 1931) in respect of “efficiency”: “Es ist fiir die eigenartige geistige Lage unsrer Kunsttechnik sehr bezeichnend, dass in allen Proklamationen der “neuen Typographie” dieser Begriff nicht ein einziges Mal ernsthaft verwendet wurde, obwohl er eine vortreffliche Grundlage eines “wissenschaftlich-exacten” Programms abgeben würde. Bessere Efficiency der Drukschrift würde bedeuten: bessere Lesbarkeit, bessere Anpassung an die Sprache, bessere Wirtschaftlichkeit und — was für die Reklameschrift gewiss wichtig ware — starkere Wirkung im psychologischen Sinne.”



The constructivist “engineering” mind was leading. Only iesigns of a very simple and clean structure were appreciated in he next years ; the underlying principle had to be easily con:eivable, quiet and consistently worked out. In consequence :he pure Roman model, after the English school (Johnson and lis pupils), acted as an example next to the most severe repre.entatives of the Didot group. Variations on both themes were leveloped with considerable skill and fantasy. As belonging to :he former family we must mention Deutsch-Römisch by Ernst Schneidler, still rather free in design, graceful, with a very -emarkable italic; Minister by Fahrenwaldt ; Orpheus, with the •elated italic Euphorion, by Walter Tiemann, charming and -efined, the drawing careful and precise, after Narziss the artist’s Dest type and hitherto one of the most important designs of this :entury ; Mundus by Willy Schwerdtner, a strong, clean design vith, however, some weak spots in the lower case ; Zabel Antiqua by Lucian Zabel ; Jost Antiqua by Heinrich Jost, a small and pleasant face, that lacks the vigorous, broad movement of Atrax ind Beton by the same artist.
Among the latter family several revivals of the 18th century types occur : Bodoni, Firmin-Didot, Walbaum, Hartel and Haenel ire represented. Herbert Thannhauser edited Parcival, comparable with Bodoni Heavy, Tiemann Antiqua, etc. ; Arpke Antiqua, named after its designer, is a middle-weight display-type, that fulfils its purpose excellently, it matches well with many text-types and illustration techniques, a clear, original and distinguished face. Metropolis, also by W. Schwerdtner, is the heaviest “perfume-type” available, an eye-catcher, but all the same refined. Corvinus by Imre Reiner, Bayertype by Herbert Bayer and Mondial by Hans Bohn, are all exaggerations of the Didot principle, severe, sharp and cold.
The development of the technique of advertising lay-out required a greater choice in types for special effects. There was a need for cursives of various kinds : forceful script, bold or firm, quick handwriting and dainty cürves.
A rather stiff, but passably elegant design of the type-cutting firm Wagner & Schmidt has been adopted by 7 foundries under different names ; swash capitals were added to it by W. Schumann {Butterfly) and by Georg Belwe (Mozart). Lucian Bernhard and F. W. Kleukens each designed a “ Schönschrift” with charming features. In this kind of work Albert Auspurg developed considerable activity, more notable for the matter of quantity than of quality. Martin Jacoby-Boy (of “ Bravour” fame) should be mentioned for his Jacobea, which shows an able and practised hand.



Quick and Flott are American import; a few remarks concerning these types will be found in the corresponding paragraph. Ariston by M. Wilke also follows the American style ; it is a very clever work. Another type of his, Wilke Kursiv, is simple, clear and pleasant.
Some sensation was made by Legende, a product of F. H. E. Schneidler. Perhaps it can be best described as a Fraktur version of Civilité, the sixteenth century French handwriting, which e.g. Robert Granjon cut into type. lts expressive and original features render very useful services in special cases.
In the heavy class Adastra, a fresh open italic by Herbert Thannhauser, is a good representative of a mixture between austerity and grace that is useful for society-appeal. For a less refined kind of work E. Lautenbach produced Pragefest and Lucian Bernhard a heavy script which bears his name, both too carefully drawn to suggest a quick, jotted-down handwriting. The demand for these would-be spontaneous scrawlings was met by several faces ; but they are real bastards. It is to the credit of Prof. Koch that his version, Holla, is about the poorest one. No handwriting fulfils solely aesthetical purposes without being empty; for type faces however, the contrary serves. The controversy is unbridgeable : every effort to attain to a combination of both principles either spoils the aesthetical or the characterological unity.
Allegro is the name of another heavy-weight italic, which retains the characteristics of handwriting ; it does particularly well in colour.
Bernhard also produced an extrabold cursive, which is too omamented for its weight. Other famous designers also tried their forces at heavy display-types. As early as 1923 Rudolf Koch produced Neuland, just like his other works notable for individuality. It had been his aim to retain the unsmoothed spontaneity of handwork in founded type. Though the translation of his intentions by Gustav Eichenauer, his special punch-cutter at the Klingspor foundry, always had been an exemplary one, this time he wanted to attain a maximal force of expression in a personal struggle with the material, just as it had been the custom before the industriaüzation of the craft. Therefore he filed the punches with his own hands. The Neuland style has also been followed in some of his most impressive, religious texts, which are almost fanatical in their barren simplicity. This hitting-force serves commercial purposes as well.... The figures are weak.
Of equally unmistakably German origin, betraying strong
Legibility, atmosphere-value and forms of printing types. 15



Fraktur-influences, is Fanfare by Louis Oppenheim, a clever design, very legible for its weight. Beside Ratio Latein Bold, which is a “Normande” of the usual form, Prof. F. W. Kleukens designed Omega, a sound, narrow, heavy face. Germany has produced several of these quiet, clean-cut heavy faces, which are an improvement compared to the older display-types. Koloss by Jacob Erbar is another one. Kontrast by Prof. F. H. E. Schneidler requires careful handling, but it is able to give a curious impression of precision and nobility. Atrax by Heinrich Jost is an alphabet of bold, open capitals, drawn in a broad and firm manner.
The more national political orientation was followed by a rise in Fraktur and related types. Also a tendency to more simple forms is noticeable ; it could even be said that in this way, by giving Fraktur-forms with Roman simplicity, a new attempt has been made to get at a compromise between the two styles, were it not that all motives of legibility, international usefulness, unity of lettersigns, etc. were entirely absent. Wallau by Prof. Koch has to be mentioned in this respect, but we must abstain from further commentary because it is more of a Fraktur type (in the sense of “ broken hand”). Several types appeared, that could be called Fraktur-sans serif — a thoroughly unsound principle, because the essence of Fraktur is thus violated. These designs are entirely lacking the charming liveliness of the broken hand and the transparant purity of the good sans type, but they have a dull, hard appearance.
If we are to summarize our review of the German production, we do not hesitate to say that this country, more than any other, has contributed to the raise of the art of type designing. More experimenting has been done than elsewhere. It may be true that among the enormous production deplorable failures are to be found, but on the other hand many new ways have been explored, valuable novelties detected, the application differentiated, the richness of forms enlarged. All German designers have tried to develop an own style, without looking back to the masters of the past. Those display-types, for which one cannot fall back upon historical examples, gained profit by this circumstance : they spread all over the world.
§ 6. Other Countries.
Spain.
The Spanish foundry of R. Gans in Madrid issues its own versions of the current sorts of type. Several of its products came to



England, e.g. Rosart, decorated open capitals. The branch office of the Bauer foundry has also put its mark on Spanish typography.
Holland.
Holland enjoys the possession of two able type designers. The older of the two, S. H. de Roos, who is art director of the Amsterdam Foundry, belongs to the generation of artists, who started their career in the first years after Morris’ death. His position in Holland is a dominant one, but his types have found their way also to Scandinavia, Belgium, America and Germany. Influenced in his earlier years by English and American typographers, he soon developed a very original and personal style, which truly reflects the national character of his country. During the thirty years of his connection with the Amsterdam Foundry he exerted an enormous influence on the Dutch typography and every printing-office is in the possession of at least one of his five type designs.
The first of these, the Hollandsche Mediaeval, appeared in 1912. It is a quiet, portly, but at the same time graceful face of excellent design. Like his other types it is very carefully drawn, without taking any liberties and with master technique, which allows the original idea to come out to the full. Some years later a flowing, round, cursive type of de Roos was issued, called Ella. The Erasmus and Grotius show the same style as the “Hollandsche Mediaeval”, but they are more freely and sharply cut, even in a little too elaborate way.
Egmont was produced in 1933. It is comparable with Koch Antiqua and Zarte Bernhard Antiqua and, to a certain extent, with Klingertype too, but the design is worked out into an original unity. It is thoroughly sound, as only can be the result of long and serious preparation (the absence of which is often so traceable in designs released by an artistically less scrupulous foundry). Too much vertical stress and austerity are skillfully avoided. The light edition is rather scrawly, like those of Corvinus and Metropolis. It was a surprise to the numerous typophiles, who followed de Roos in his work as a faithful but original adherent to the Anglo-American school, when he produced this real and full-bodied modern face.
Both the “Hollandsche Mediaeval” (as “Mediaeval”) and the “Egmont” appeared on the Intertype ; “Ella”, as “Card Italië" on the Linotype.
De Roos’ younger colleague, J. van Krimpen, who works for the old house of Enschedé, gained quickly fame by his Lutetia, also on Monotype. This design has a very original, personal



ind delicate style ; it has to be placed in the same class as Centaur 'or style and quality. The accompanying chancery italic betrays ;he skilled hand of the calligrapher. Van Krimpen also produced i fine Greek type (Antigone), which shows the way for future design )f Greek alphabets, and quite a new way it is. Recently a new :ancellaresca bastarda made its first appearance and it should be welcomed as a valuable contribution to the present “types poétiques” ; a new book-face is in preparation.
Jtaly.
In Italy the upheaval of printing art kept pace with the other European countries owing to the untiring energy of R. Bertieri, who lead the movementwithhisperiodical“RisorgimentoGrafico”. He designed several types for the principal Italian foundry, the Societa Nebiolo of Torino. His Inkunabula (1911) was something of a revolution against the current Italian types of that time; it is one of the strongest examples of the handwriting-theory. Ruano (1925) and Sinibaldi (1928), of the same artist, retain these features too, but Paganini (1927) comes nearer to the usual drawingtechnique. The latter is a refined, thin face, a kind of regularized Bernhard Roman.
A type by the poet Francesco Pastonchi appeared also on the Monotype ; its tender lines betray the sensitiveness of the designer.
The Societa Nebiolo issued good versions of the various categories of display-types; a fine script, called Fluidum, deserves special mentioning.
Czechoslovakia.
Typography in Czechoslovakia stands on a high level; the art of the book is remarkably flourishing. One of the leading men is Karel Dyrynk, director of the Government printing-office, who designed several types for the Grégr foundry, all modern variations of the old style model, not so daring as most of the bibliophile typography.
Type-design is made very difficult for the Czechoslovakian artist by the multitude of accents required by the language : they have to be worked up with the character itself into a unity of form. Beside types of Vojtèch Preissig, Slavoboj Tusar and Karel Svolinsky, which are not for sale, two designs of Oldrich Menhart gained wider fame, because they were brought upon the world-market. The first was issued by the Bauer foundry in 1932 ; its crisp, yet sensitive lines give it more character than the later type for the Monotype displays. The latter, which appeared in 1934, goes on the verge of weakness.



LIST OF THE MOST
IMPORTANT MODERN PRINTING TYPES.
The second column presents the names of original designs that are generally for sale; the names used in other countries or by foundries which offer the same design, are placed in that column after an indentation. Of type-families only the name and year of the basic design are mentioned.
This list does not claim to be complete, of course, as there are at least twenty times as many series on the market. Artistical merits, frequency of use and recency of design have been applied as a Standard in the selection of the names.
In the third column the names of the original foundries are given. The designs that are not sold by other foundries than the original producers, are in many cases also obtainable through the foreign agents of these producers. The names of these agencies (such as Soldans and Meerloo in England, which represent German, French and Dutch foundries, or Messrs. Shortt, who act for the A.T.F.) have not been mentioned especially. The full names of the foundries, which were abbreviated in the third column, are following here :
American TF: American Type Founders SalesCorp., Elisabeth, N.J., U.S.A. Amsterdam: Lettergieterij “Amsterdam” v/h N. Tetterode, Amsterdam, Holland. Bauer : Bauersche Giesserei, Frankfurt am Main, Germany.
Berthold : H. Berthold, A.-G., Berlin, Germany.
Blackfriars: Blackfriars Type Foundry Ltd., London, England.
Butter: Schriftguss K.-G. vorm. Brüder Butter, Dresden, Germany.
Caslon : H. W. Caslon & Co, Ltd., London, England.
Continental TF: Continental Typefounders Association, New York, U.S.A. Deberny & Peignot: Deberny & Peignot, Paris, France.
Enschedé: Joh. Enschedé & Zonen, Haarlem, Holland.
Fonderie Typ. Fr.: Fonderie Typographique Frangaise, Paris, France.
Gans: R. Gans, Madrid, Spain.
Genzsch & Heyse: Genzsch & Heyse A.-G., Hamburg, Germany.
Goudy : F. W. Goudy, The Village Letter-Foundery, Forest Hill Gardens, N.Y., U.S.A.
Grégr : A. Grégr, Praha, Czechoslovakia.
Haas: Haas’sche Schriftgiesserei, Münchenstein, Switzerland.
Haddon : John Haddon & Co., London, England.
Idékowsky : Jan Idékowsky & Co., Warsawa, Poland.
Intertype : Intertype Corp., Brooklyn, N.Y., U.S.A.
John: J. John Söhne, Hamburg, Germany.
Krebs: Benjamin Krebs Nachf., Frankfurt am Main, Germany.
Klingspor: Gebr. Klingspor, Offenbach am Main, Germany.
Linotype: Mergenthaler Linotype Company, Brooklyn, N.Y., U.S.A.
Mergenthaler Setzmaschinén-Fabrik, Berlin, Germany.
Ludlow: Ludlow Typograph Company, Chicago, 111., U.S.A.
Ludwig & Mayer: Ludwig & Mayer, Frankfurt am Main, Germany.
Miller & Richard : Miller & Richard, Edinburgh, Great Britain.
Monotype : The Monotype Corporation, London, England.
The Lanston Monotype Machine Company, Philadelphia, Pa. U.S.A. Nebiolo: Societa Nebiolo, Torino, Italia.



'forddeutsche Schriftg.: Norddeutsche Schriftgiesserei, G.m.b.H., Berlin, Germany. Pavyers & Builens : Pavyers & Builens, Ltd., London, England.
Plantin: Etablissements “Plantin”, Bruxelles, Belgium.
Poppelbaum : Schriftgiesserei Póppelbaum, Wien, Germany.
Schelter & Giesecke: J. G. Schelter & Giesecke, Leipzig, Germany.
Stempel: D. Stempel, A.-G., Frankfurt am Main, Germany.
Stephenson Blake: Stephenson, Blake & Co., Ltd., London, England. Stevens-Shanks: Stevens, Shanks & Sons, Ltd., London, England.
Irennert: J. D. Trennert & Sohn, Altona-Elbe, Germany. fypograph : Typograph G.m.b.H., Berlin, Germany.
Wagner : Ludwig Wagner, A.-G., Leipzig, Germany.
Weber : C. E. Weber, Stuttgart,Germany.
Weisert: Otto Weisert, Stuttgart, Germany.
Woellmer: W. Woellmer’s Schriftgiesserei, Berlin, Germany.



i EAR OF
Nr. Name of Series Editing Firm Issue Designer
I 1 Achtung Ludwig & Mayer 1932 Erhard Grundeis
I 2 Acier Deberny & Peignot 1935 A. Mouron Cassandre
( 3 Adastra Stempel 1928 Herbert Thannhauser
4 Aigrette Amsterdam see Bernhard Tango
I 5 Aktuell Butter 1935 Walter Schnippering
Penflow
I 6 Alarm Trennert 1928 Heinz König
I 7 Albertus 324 Monotype 1935 Berthold Wolpe
I 8 Albingia Genzsch & Heyse 1906
I 9 Allegro Ludwig & Mayer 1936
; 10 Alt Mediaeval Berthold 1915 M. Hertwig
. 11 Ambra Schetter & Giesecke 1924 Schulze
12 Amstel Amsterdam see Lipsia Antiqua
; 13 Amts Antiqua Stempel 1909
14 Antike Elzevir Stempel see Jenson
i 15 Appell Butter 1933 Arno Drescher
I 16 Arabella Wagner 1936
117 Ariston Berthold 1933 M. Wilke
18 Arkona Genzsch & Heyse 1935
| 19 Arpke Antiqua Butter 1928 Otto Arpke
20 Arrighi 252 Monotype 1929 Frederic Warde
21 Arsis Amsterdam see Onyx A.T.F.
ï 22 Artcraft American T. F. 1912 or. Barnhart & Spindler
! 23 Artcraft Ludlow
24 Artista Schelter & Giesecke 1937 Rudolf Sternberg
25 Ashley-Crawford 238—279 Monotype 1930 Ashley Havinden
\ 26 Astrée Deberny & Peignot 1925 Robert Girard
Mazarin Stephenson Blake
Astrée Enschedé
127 Atlantis Woellmer 1931
: 28 Atlas Amsterdam 1932
29 Atrax Bauer 1926 Heinrich Jost
30 Attraktion Woellmer 1934
31 Augenheil Antiqua Ludwig & Mayer 1908
Augenheil Antiqua Linotype
| 32 Auriol Deberny & Peignot 1903 Georges Auriol
Auriol Genzsch & Heyse
l 33 Austria Berthold see Ver Sacrum
34 Ballé Initialen Bauer 1929 Maria Ballé
: 35 Banjo Deberny & Peignot 1932
I 36 Basalt Genzsch & Heyse 1927-’34
[ 37 Baskerville 169 Monotype 1924 (J. Baskerville 1750-’4)
38 Baskerville Stephenson Blake
Baskerville American T. F.
j 39 Baskerville Stempel 1924
Baskerville Linotype
i 40 Baskerville Caslon see Série 18
41 Basuto Stephenson Blake 1927
[ 42 Batarde Bauer 1913 Hugo Steiner-Prag
I 43 Baustein Grotesk Schelter & Giesecke 1928
I 44 Bayer Type Berthold 1935 Herbert Bayer
I 45 Beatrice Ludwig & Mayer 1934
146 Behrens Antiqua Klingspor 1908 Peter Behrens
I 47 Behrens Kursiv Klingspor 1907 Peter Behrens
148 Behrens Mediaeval Klingspor 1914 Peter Behrens
■ 49 Behrens Schrift Klingspor 1901 Peter Behrens



Nr. Name of Series Editing Firm Y^je°F Designer
50 Bell 341 Monotype 1932 (John Bell, 1788)
Bell Stephenson Blake
51 Bellery Deberny & Peignot 1912 Bellery-Desfontaines
52 Belwe Antiqua Schelter & Giesecke 1913 Georg Belwe
53 Bembo 270 Monotype 1931 (Francesco Griffo) j
54 Benedictine Linotype 1915 J. E. Hill & E. E. Bartlett; or. 14871
55 Benton American T. F. 1934 Morris F. Benton
56 Bernhard Antiqua Bauer 1912 Lucian Bernhard
Bernhard Antiqua Typograph
57 Bernhard Booklet American T. F. Lucian Bernhard
58 Bernhard Buchschrift Bauer 1932 Lucian Bernhard
59 Bernhard Extrafette Kursiv Bauer 1927 Lucian Bernhard
60 Bernhard Fashion American T. F. 1929 Lucian Bernhard
61 Bernhard Gothic American T. F. 1929 Lucian Bernhard
62 Bernhard Handschrift Bauer 1928 Lucian Bernhard
63 Bernhard Negro Bauer 1930 Lucian Bernhard
64 Bernhard Schönschrift Bauer 1924 Lucian Bernhard
(Liberty) (American T. F.)
65 Bernhard Tango American T. F. 1934 Lucian Bernhard
Aigrette Amsterdam
66 Bernhard Antiqua, Zarte Bauer 1925 Lucian Bernhard
67 Berthold Grotesk Berthold 1929 Georg Trump
Berthold Grotesk Intertype
68 Beton Bauer 1930 Heinrich Jost
Beton Intertype
69 Bifur Deberny & Peignot 1929 A. M. Cassandre
70 Blado Italic 119 Monotype 1931 (Antonio Blado)
71 Block Berthold 1908 Hermann Berthold & Louis
Oppenheim
72 Bloomsbury Stevens-Shanks Alexander Shanks
73 Bodoni Bauer 1925 (Giambattista Bodom)
74 Bodoni American T. F. 1911 Morris F. Benton
Bodoni Amsterdam
Bodoni Berthold
Bodoni Butter
Bodoni Caslon
Bodoni Genzsch & Heyse
Bodoni Haas
Bodoni Intertype
Bodoni Linotype
Bodoni Monotype
Bodoni Nebiolo
Bodoni Norddeutsche Schriftg.
Bodoni Stephenson Blake
Bodoni Typograph
75 Bodonia Nebiolo see Cheltenham
76 Bombe Ludwig & Mayer 1908 O. L. NSgele
77 Bookface Intertype see Bookman
78 Bookman American T. F. 1902 or. Bruce Typel.
Bookface Intertype
79 Boul’ Mich’ American T. F.
80 Bradley Ultra Mod. Init. American T. F. Will Bradley
81 Braggadocio 278 Monotype 1929
82 Bravour Stempel 1912 Martin Jacoby-Boy
Paladin Caslon
83 Bristol Stevens-Shanks
Vulcan Bold Linotype



Nr. Name of Series Editing Firm ^Issue°F Designer
I 84 Bristol Amsterdam 1929
Bristol Berthold
I 85 Broadway American T. F: 1928 Morris F. Benton
Broadway A 306 Monotype
| 86 Bücherfreund Antiqua Ludwig & Mayer see Cochin
I 87 Bullünch American T. F. 1904 Morris F. Benton (partly)
j 88 Bulmer American T. F. 1927 (William Martin, 1795)
I 89 Butterfly Butter see Troubadour
| 90 Butterfly Initialen Butter 1928 Willy Schumann
I 91 Byron see Watteau
fc 92 Cable see Kabel
t 93 Cairo Intertype 1935
| 94 Candida Ludwig & Mayer 1936 Jakob Erbar
Candida Linotype
95 Canterbury 197 Monotype 1925
r 96 Capitol Butter 1931 K. H. Schaefer
I 97 Card Italic Linotype see Ella Cursief
I 98 Carlton Amsterdam 1929
Carlton Berthold
| 99 Carlton Caslon see Ehmcke Antiqua
100 Cartoon see Fresko
101 Caslon Caslon 1722 W. Caslon, 1722
Caslon American T. F.
Caslon Enschedé
Caslon Linotype
Caslon 128 Monotype
s 102 Centaur 252 Monotype 1915 Bruce Rogers
103 Century American T. F. 1898 Linn Boyd Benton
Century Intertype
Century Linotype
Century Monotype
104 Charnwood see Watteau (Bold)
105 Chatsworth Stephenson Blake 1908
' 106 Cheltenham American T. F. 1902 Bertram Grosvenor Goodhue &
Ingalls Kimball
Cheltenham Amsterdam
Cheltenham Caslon
Cheltenham Linotype
Cheltenham Weisert
Cheltonian Intertype
Bodonia Nebiolo
Gloucester 99 Monotype
Katalog Antiqua Krebs
Lessing Antiqua Woellmer
Pfeil Antiqua Butter
Roosevelt Bauer
Sorbonne Berthold
Toskana Berthold
Union Stempel
[ 107 Cheltonian Intertype see Cheltenham
| 108 Chic American T. F. 1928 Morris F. Benton
| 109 Cissarz Latein Ludwig & Mayer Johann Vincenz Cissarz
I 110 Cito Versalien Butter 1930 K. H. Schaefer
1111 City Berthold 1930 Georg Trump
Métro Amsterdam
I 112 Clearcut Oldstyle American T. F. or. Barnhart & Spindler



Sr. Name of Series Editing Firm ^Issue ? Designer
113 Clearface American T. F. 1909 Morris F. Benton
Clearface Linotype
Clearface Ludlow
Clearface Monotype
Clearface Stephenson Blake
Dominus (open) Stephenson Blake
114 Cloister American T. F. 1914 Morris F. Benton
Cloister Intertype
Cloister Linotype
Cloister 272 Monotype
115 Cochin Deberny & Peignot 1912
Cochin 165 Monotype
Bücherfreund Antiqua Intertype Bücherfreund Antiqua Ludwig & Mayer Gravure Serie I Amsterdam
116 Cochin, Nicolas Deberny & Peignot 1912
Cochin, Nicolas Caslon
Sonderdruck Antiqua Ludwig & Mayer Gravure II Amsterdam
117 Colonna 225 Monotype
118 Columbia Amsterdam
119 Comtess Lola Norddeutsche Schriftg. see Troubadour
120 Cooper American T. F. 1919 Oswald Cooper
Cooper Butter
121 Cooper Modern American T. F. Oswald Cooper
122 Corvinus Bauer 1934 Imre Reiner
123 Cosmo Stevens-Shanks
124 Cushing American T. F. bef. 1897 J. Stearns Cushing
Cushing 17 Monotype
125 Czeschka Genzsch & Heyse 1914 C. O. Czeschka
126 Danziger Antiqua Stempel 1903 H. Cohn
127 Daphnis Klingspor 1928 Walter Tiemann
128 Das Antieke Amsterdam Peter Das
129 Deepdene Caslon F. W. Goudy
Deepdene Monotype
130 Delitsch Antiqua Berthold 1911 Hermann Delitsch
131 Della Robbia American T. F. about 1900 Thomas Maitland Cleland
Firenze Amsterdam
Della Robbia Deberny & Peignot
132 Delphian Ludlow
133 Deutsch Römisch Weber 1926 F. H Ernst Schneidler
Deutsch-Römisch 298 Monotype
134 Didot, Pierre Berthold 1913 (Pierre Didot, 1784)
Didot, Pierre 411 Monotype
135 Diva Butter 1936
136 Divina Butter 1930 Wilhelm Berg
137 Dolmen Schelter & Giesecke 1921 M. Salzmann
138 Dolphin Old Style Stevens-Shanks 1914 George W. Jones & Edward P.
Prince
139 Dominus Stephenson Blake see Clearface
140 Donatello Nebiolo see Troubadour
141 Dorique Deberny & Peignot 1917 Léon Pichon & Edouard Carlègle
142 Drescher Eilschrift Woellmer 1934 Arno Drescher
143 Van Dyck 203 Monotype 1938 (Chr. van Dyck)-J. van Krimpen
144 Dynamit Wagner 1936
145 Dynamo Ludwig & Mayer 1930
146 Dyrynk Latin Grégr 1928 Karei Dyrynk



Nr. Name of Series Editing Firm ^Issue** Designer
147 Eagle Bold American T. F.
148 Eckmann Klingspor 1900 Otto Eckmann
149 Éclair, 1’ Deberny & Peignot 1934
I 150 Effect Amsterdam 1935
: 151 Ege Schrift Genzsch & Heyse 1923' E. Ege
152 Egmont Amsterdam 1933 S. H. de Roos
Egmont Intertype
153 Ehmcke Antiqua Bauer 1908 Fritz Hellmuth Ehmcke
Carlton Caslon
154 Ehmcke Elzevir Ludwig & Mayer 1927 F. H. Ehmcke
155 Ehmcke Kursiv Bauer 1911 F. H. Ehmcke
156 Ehmcke Latein Ludwig & Mayer 1925 F. H. Ehmcke
157 Ehmcke Mediaeval Stempel 1922 F. H. Ehmcke
Ehmcke Mediaeval Caslon
158 Ehmcke Rustika Stempel 1914 F. H. Ehmcke
Ehmcke Rustika Caslon
159 Electra Linotype 1935 W. A. Dwiggins
160 Elegant Weisert see Troubadour
161 Elegant Grotesk Stempel 1928 Hans Möhring
Guildford Caslon
162 Elementar Versalien Norddeutsche Schriftg. 1930
163 Elephant Stevens-Shanks ca. 1820 Vincent Figgins
164 Elisabeth Bauer 1933 Elisabeth Friedlander
165 Ella Cursief Amsterdam 1915 S. H. de Roos
Card Italic Linotype
Ella Kursiv Wagner
Ella Kursiv Poppelbaum
Elvira Kursiv Weber
Handelscursief (bold) Amsterdam
Juno Haas
Kurrenta Kursiv Berthold
Merkantil Kursiv Trennert
Radio. Sport (bold) Weisert
Saxonia Butter
Siena (bold) Nebiolo
1166 Elvira Weber see Ella
1167 Elzévir Deberny & Peignot 1857 Théophile Beaudoire
168 Elzevir Bauer see Jenson
169 Elzévir Ancien Deberny & Peignot Aubert Bros. & Huchet
170 Energos Butter 1932 Arno Drescher
171 Erasmus Amsterdam 1923 S. H. de Roos
Oranien Mediaeval Berthold
Pilger Latein Norddeutsche Schriftg.
172 Erbar Grotesk Ludwig & Mayer 1926 Jakob Erbar
Erbar Grotesk Linotype
Erbar Grotesk Typograph
173 Erbar Mediaeval Ludwig & Mayer 1914 Jakob Erbar
Monastic (open caps.)
Luminous (open)
174 Escrivatura Decorativa Gans
1175 Estienne Linotype 1928 George W. Jones
1176 Etoile Deberny & Peignot see Trafton Script
1177 Euphorion Klingspor 1935 Walter Tiemann
1178 Europe Deberny & Peignot see Futura
1179 Eve see Koch Antiqua
1180 Excelsior Linotype 1933
1181 Expression Berthold 1921 Albert Auspurg



Mr. Name of Series Editino Firm ^Issue^ Designer
182 Fanal Schelter & Giesecke 1933
183 Fanfare Berthold 1927 Louis Oppenheim
184 Fatima Butter 1933 K. H. Schaefer
185 Feder Grotesk Ludwig & Mayer 1910 Jakob Erbar
186 Feenhaar Schelter & Giesecke 1913 Albert Auspurg
187 Felix Titling Capitals 399 Monotype 1934
188 Femina Bauer 1912 Julius Gipkens
189 Film, le Deberny & Peignot 1934 Marcel lacno
190 Firenze Amsterdam see Della Robbia
191 Flamingo Berthold 1928 Louis Oppenheim
192 Flamme Schelter & Giesecke 1932
193 Fleischmann Antiqua Wagner 1927 Georg Belwe
194 Flex Amsterdam 1937 Georg Salter
195 Florida Woellmer 1932
196 Flott Bauer see Gillies Gothic
197 Fluidum Nebiolo
198 Forelle Weber 1936 Erich Mollowitz
Rheingold Trennert
199 Fortuna Trennert 1930
200 Forum Title Caslon 1911 Frederic W. Goudy
201 Fournier 185 Monotype ca. 1740 (Pierre Simon Fournier)
202 Fournier, geperlte Butter 1922 P. A. Demeter
203 Fournier le Jeune Deberny & Peignot 1914
204 Franklin Old Style Linotype 1929 (Alexander Phemister 1863)
205 Frappant Genzsch & Heyse 1916 Heinz König
206 Freehand American T. F. 1926
207 Fresko Bauer 1936
Cartoon
208 Fridericus Genzsch & Heyse 1891
Fridericus Monotype
209 Friedrich Bauer Grotesk Trennert 1933 Friedrich Bauer
210 Futura Bauer 1927 Paul Renner
Futura Intertype
Europe Deberny & Peignot
211 Futura Black Bauer 1929 Paul Renner
Futura Black Haddon
212 Futura Schlagzeilen Bauer 1932 Paul Renner
213 Gallia American T. F. 1927 Morris F. Benton
Gallia 313 Monotype
214 Ganton Stephenson Blake
215 Garamond Stempel 1924 (Claude Garamont 1540)
Garamond 1925 Linotype
216 Garamond American T. F. 1917 Morris F. Benton
Garamont Amsterdam
Garamond Berthold
Garamond Intertype
Garamond Monotype
217 Garamont Deberny & Peignot
218 Gayton Caslon 1927 Frederic Gayton
219 Genzsch Antiqua Genzsch & Heyse 1907 Friedrich Bauer
Genzsch Antiqua Intertype
Genzsch Antiqua Typograph
Stanhope O.S. Stevens-Shanks
220 Germanische Antiqua Genzsch & Heyse 1912 Friedrich Bauer
221 Gigant Genzsch & Heyse 1926 H. Schmidt
222 Gill Sans Serif 262 Monotype 1929 Eric Gill



YeaR of
Nr. Name of Series Editing Firm Issue Designer
223 Gillies Gothic Bauer 1935 W. S. Gillies
Flott Bauer
224 Giraldon Deberny & Peignot 1905 A. Giraldon
225 Girder see Memphis
226 Gladiola Stempel 1936 M. Wilke
227 Glass Antiqua Genzsch & Heyse 1912 F. P. Glass
228 Glenmoy Stephenson Blake 1934
229 Globus Stempel 1932 Friedrich Wobst
230 Gloucester Monotype see Cheltenham
231 Gnom Schelter & Giesecke 1914 Albert Auspurg
232 Golf Butter 1936
233 Gothische Antiqua Stempel 1916 F. W. Kleukens
234 Goudy Antique Goudy 1912 Frederic W. Goudy
Ratdolt Roman Caslon
235 Goudy Heavy 214 Monotype 1925 F. W. Goudy
236 Goudy Modern 249 Monotype 1921 F. W. Goudy
Goudy Modern Caslon
237 Goudy Old Style American T. F. 1916 F. W. Goudy
Goudy Old Style 291 Monotype
11B Ludlow
238 Goudy Text 292 Monotype F. W. Goudy
239 Goudytype American T. F. F. W. Goudy
240 Gracia Amsterdam see Troubadour
241 Granby Stephenson Blake
242 Granjon Linotype 1924 George W. Jones
Granjon Stevens-Shanks
243 Graphik Weisert 1929 F. H. Ernst Schneic
244 Grasset Deberny & Peignot 1898 Eugène Grasset
Grasset Genzsch & Heyse
245 Gravira Schelter & Giesecke 1936
246 Gravure American T. F. 1927 Morris F. Benton
247 Gravure I Amsterdam see Cochin
248 Gravure II Amsterdam see Cochin. Nicolas
249 Grégrova romana Grégr 1930 Karei Dyrynk
250 Grock 388 Monotype 1935
251 Grotius Amsterdam 1926 S. H. de Roos
Grotius Berthold
252 Guildford Caslon see Elegant Grotesk
253 Haenel Antiqua Klingspor 1928
254 Haiduk Antiqua Bauer 1908 August Haiduk
Haiduk Antiqua Linotype
Haiduk Antiqua Stempel
255 Handelscursief (bold) Amsterdam see Ella
256 Harquil Deberny & Peignot
Harquil Haddon
257 Hartel Antiqua Butter 1928
Hfirtel Antiqua Weber
258 Hauser Script Ludlow 1936 Georg Hauser
259 Hawarden Haddon
260 Healy Haddon
261 Helga Stempel 1912 F. W. Kleukens
262 Helion Butter 1935
[263 Heraldisch Genzsch & Heyse 1910 Otto Hupp
i 264 Hercules Amsterdam 1926
Titanic Wagner
265 Herold Berthold 1901 Hermann Berthold
'266 Hess Old Style Lanston Monotype 1920 Solomon Hess



Nr. Name of Series Editing Firm Issue°F Designer
267 Heyse Antiqua Genzsch & Heyse 1926 Friedrich Bauer
268 Holla Klingspor 1932 Rudolf Koch
269 Hollandsche Mediaeval Amsterdam 1912 S. H. de Roos
Hollandische Media val Berthold Mediaeval Intertype
270 Hollandische Mediaval Berthold see Hollandsche Mediaeval
271 Hollywood American T. F.
272 Holzhausen Antiqua Stempel 1916 Holzhausen
Holzhausen Antiqua Linotype
273 Horley Old Face 199 Monotype
274 Hotspur Haddon
275 Hupp Antiqua Klingspor 1910 Otto Hupp
276 Huxley Vertical American T. F. 1935 Walter Huxley
277 Ideaal Amsterdam see Ideal News
278 Ideal News Intertype
Ideaal Amsterdam
279 Imprint Old Face 101 Monotype 1912 Edward Johnston, Gerard Mey-
nell, A. H. Mason
280 Indépendant Plantin 1930 G. Collette & J. Dufour
Indépendant Amsterdam
281 Ingeborg Stempel 1909 F. W. Kleukens
Ingeborg Linotype . .
282 Inkunabula Nebiolo 1911 Raffaello Bertien
283 Intermezzo Butter 1933
284 Ionic 265—342 Monotype (Vincent Figgins 1821)
Ionic Linotype
Ideal News Intertype
285 Italian Oldstyle American Monotype 1924 F. W. Goudy
286 Jacobea Berthold 1928 Martin Jacoby-Boy
287 Janson Stempel (A. Janson 1670)
Janson Linotype
288 Jenson American T. F. Phinney
Jenson Caslon
Jenson Linotype
Jenson Nebiolo
Antike Elzevir Stempel
Elzevir Bauer
Kloosterschrift Amsterdam
289 Jiu-Jitsu Berthold 1936
290 Jocunda 369 Monotype 1932 Stanley Baxter
291 Jost Mediaeval Ludwig & Mayer 1927 Heinnch Jost
Jost Mediaeval Intertype .
292 Jubilee Stephenson Blake 1935 Ene Gul
293 June Stephenson Blake 1927
294 Junior Schelter & Giesecke 1936
295 Juno Haas see Ella Cursief
296 Kaatskill 311 Monotype 1934 Fredenc W. Goudy
297 Kabel Klingspor 1927 Rudolf Koch
Cable
298 Kamak Ludlow 193:3
299 Katalog Antiqua Krebs see Cheltenham ^
300 Kaufmann Script American T. F. 1936 M. R. Kaufmann
301 Kennerley Caslon 1911 F. W. Goudy
Kennerley Monotype
Kenntonian Intertype



Nr. Name of Series Editing Firm Issue Designer
302 Kenntonian Inteftype see Kennerley
303 Keynote American T. F.
304 Kingston Stephenson Blake 1934
305 Kino 305 Monotype
306 Kleukens Antiqua Bauer 1908 F. W. Kleukens
Trieste Nebiolo
307 Kleukens Scriptura Stempel 1926 F. W. Kleukens
308 Klinger Antiqua Berthold 1920 Julius Klinger
309 Klingertype Butter 1925 Klinger, Frey. Willrab
310 Kloosterschrift Amsterdam see Jenson
311 Knock-out Berthold 1936
312 Koch Antiqua Klingspor 1922 Rudolf Koch
Locarno
Eve
313 Kolibri Schelter & Giesecke 1915 Albert Auspurg
314 Koloss Ludwig & Mayer 1923 Jakob Erbar
315 Kombi Versalien Ludwig & Mayer
316 König Antiqua Klingspor 1905 Heinz König
317 Kontrast Weber 1930 F. H. Ernst Schneidler
318 Koralle Schelter & Giesecke 1913
319 Kress Versalien Butter 1927 Oskar Kress v. Kressenstein
320 Knmhilde Ludwig & Mayer 1933 Albert Auspurg
321 Kurrenta Kursiv Berthold see Ella
322 Landi Nebiolo
323 Lang Schrift Bauer 1905 Paul Lang
324 Laureate Ludlow
325 Lautenbach Kursiv Ludwig & Mayer 1926 E. Lautenbach
326 Lautsprecher Ludwig & Mayer 1931 E. Lautenbach
327 Legende Bauer 1937 F. H. Ernst Schneidler
328 Lessing Antiqua Woellmer see Cheltenham
329 Leucht Grotesk Versalien Butter 1932
330 Lexington American T. F. 1932
331 Leysbourne 348 Monotype 1934 Frederic Landseer Griggs
332 Liberty American T. F. see Bernhard Schönschrift
333 Lichtfette Grotesk Ludwig & Mayer 1923
Phosphor
334 Lilith Bauer 1930 Lucian Bernhard
335 Lipsia Antiqua Berthold 1911 Wagner & Schmidt
Amstel Amsterdam
Neuklassisch Stempel
Noblesse Butter
Noblesse Trennert
Radio Nebiolo
Regent Antiqua Weber
Rekord Antiqua John
Renaissance Haas
336 Littleworth 284 Monotype Frederic Landseer Griggs
337 Lo Berthold 1914 Louis Oppenheim
Nero Amsterdam
338 Locarno see Koch Antiqua
339 Louvaine American T. F. 1929
340 Lucian Bauer 1928 Lucian Bernhard
341 Ludlow Black Ludlow
342 Lumina Versalien Ludwig & Mayer 1928 Jakob Erbar
343 Luminous see Erbar Mediaeval
344 Luna Amsterdam see Tango Antiqua



'Jr. Name of Series Editing Firm YIssue°F Designer
545 Lutetia Enschedé 1926 J. van Krimpen
Lutetia 255 Monotype
546 Luxor Gans
547 Luxor (Soldans) see Welt Antiqua
548 Luxor Versalien Butter 1934
549 Majestic Bauer 1915 Julius Gipkens
550 Mammut Wagner 1932
351 Mandate Ludlow 
352 Marggraff Kursiv Butter 1928 Gerhardt Marggraff
353 Maria Theresia Schelter & Giesecke 1903 H. Keune
354 Marlborough Old Style Continental T. F. 1925 F. W. Goudy
355 Maximilian Klingspor 1917 Rudolf Koch
356 Mayfair Cursive Ludlow 
357 Mazarin Stephenson Blake see 1 Astree
358 Mediaeval Intertype see Hollandsche Mediaeval
359 Memphis Stempel 1929 Rudolf Woir
Memphis Klingspor
Memphis Linotype
360 Mendelssohn Grotesk Wagner 1930 Georg Mendelssohn
361 Mendelssohn Type Butter 1921 Georg Mendelssohn
362 Menhart Antiqua Bauer 1932 Oldrich Menhart
363 Menhart 397 Monotype 1934 Oldïich Menhart
364 Mercado Amsterdam 1910
365 Meridian Klingspor i93° ,mre Reiner
366 Merkantil Trennert ... , A
367 Messe Grotesk Berthold 1921 Albert Auspurg
368 Metro Linotype 1929 W. A. Dwiggins
369 Métro Amsterdam see City (Berthold)
370 Metropolis Stempel 1926 Willy Schwerdtner
Metropolis Caslon .,, 
371 Miami Butter 1934 Albert Auspurg
372 Minister Butter 1929 Fahrenwaldt
173 Minster Stevens-Shanks
374 Mirakula Ludwig & Mayer 1931 Albert Auspurg
•575 Mohil Ludwig & Mayer 
376 Modernique American T. F. 1930 Willard Sniffen
377 Modemistic American T. F.
Modernistic 297 Monotype
378 Mona Lisa Ludwig & Mayer 1930 Albert Auspurg
379 Monastic
380 Mondial Stempel 1936 Hans Bohn
381 Montaigne Deberny & Peignot
382 Moreau le Jeune Deberny & Peignot „ „ „ , , , _ c
383 Morland Caslon bef. 1900 or. lnland Type F.
384 Motor Ludwig & Mayer 1930
385 Motto American T. F. . . _ ,
386 Möwe Genzsch & Heyse 1929 Hemz Beek
387 Mozart Wagner see Troubadour
388 Mozart Initialen Wagner 1927 Georg Belwe
389 Mundus Antiqua Stempel 1929 Willy Schwerdtner
Mundus Antiqua Caslon
390 Narziss Klingspor 1921 Walter Tiemann
Narciss Linotype _ ... „ ,
391 National Oldstyle American T. F. 1923 F. W. Goudy



Nr. Name of Series Editing Firm f^.R_°F Designer
. Issue
392 Naudin Deberny & Peignot 1913 Bernard Naudin
Naudin Stephenson Blake
393 Neon Nebiolo
394 Nero Amsterdam see Lo
395 Nero Kursiv Genzsch & Heyse 1908 O. L. Naegele
396 Neuklassisch Stempel see Lipsia Antiqua
397 Neuland Klingspor 1923 Rudolf Koch
398 Neuzeit Grotesk Stempel 1930
Neuzeit Grotesk Linotype
399 New Caslon American T. F. bef. 1905
Plantin O. S. Roman Stevens-Shanks
Plantijn Amsterdam
400 Nicolas Jenson Ludlow
401 Nobel Amsterdam 1930
Nobel Intertype
Umbra (Shadow) Berthold
402 Noblesse Butter see Lipsia Antiqua
403 Noblesse ' Trennert see Lipsia Antiqua
404 Nova Berthold 1924 Hermann Berthold &
Wohlfeld
405 Novel Gothic American T. F. 1930 H. Becker
406 Nubian American T. F. 1928 Willard Sniffen
407 Offenbacher Reform Latein Berthold 1911 Albert Auspurg
408 Ohio Butter 1912 E. Lautenbach
Ohio Trennert
409 Olympia Genzsch & Heyse 1929 C. O. Czeschka
410 Omega Stempel 1926 F. W. Kleukens ,
Omega Caslon
411 Ondina Versalien Butter 1935 K. Kranke
412 Onyx Butter 1936
413 Onyx American T. F. 1936
Arsis Amsterdam
414 Oranien Mediaeval Berthold see Erasmus
415 Orchidea Butter 1937 K. H. Schaefer
416 Original Old Style Linotype
417 Originell Butter 1935 Walter Schnippering
Pentape
418 Orpheus Klingspor 1928 Walter Tiemann
419 Orplid Klingspor 1929 Hans Bohn
420 Othello American T. F. 1934
421 Othello 246 Monotype
422 Pabst American T. F. 1905
423 Paganini Nebiolo 1927 Raffaello Bertieri
424 Paladin Caslon see Bravour
425 Paragon Linotype 1936
426 Paramount American T. F.
427 Parcival Schelter & Giesecke 1930 Herbert Thannhduser
428 Parisian American T. F. 1928 Morris F. Benton
429 Park Avenue American T. F.
430 Pastonchi Nebiolo 1928 Francesco Pastonchi
Pastonchi 206 Monotype
431 Peignot Deberny & Peignot 1937 A. M. Cassandre
432 Penflow see Aktuell
433 Pentape see Originell
434 Pergamon Wagner 1937 Alfons Schneider
435 Pericles American T. F. 1934 Robert Foster
Legibility, atmosphere-value and forms of printlng types. 16
Paul



Nr. Name of Series Editing Firm Issue°F Designer
436 Perkeo Schelter & Giesecke 1914 Albert Auspurg
437 Perpetua 239 Monotype 1925 Eric Gill
Perpetua Stephenson Blake
438 Pfeil Antiqua Butter see Cheltenham
439 Phalanx Genzsch & Heyse 1930 Hans Möhring
Phalanx Typograph
440 Pharaon Deberny & Peignot
441 Phanomen Krebs 1927 Franz Riedinger
442 Phosphor (Soldans) see Lichtfette Grotesk
443 Pilger Latein Norddeutsche Schriftg. see Erasmus
444 Piranesi American T. F.
445 Plak Stempel 1929 Paul Renner
446 Plantin 110 Monotype 1914
447 Plantin O. S. Roman Stevens-Shanks see New Caslon
448 Plantijn Amsterdam see New Caslon
449 Polar Grotesk John 1931
450 Polifilus 170 Monotype (Fr. Griffo)
451 Poltawski Idzkowski 1934 A. Poltawski
Poltawski 398 Monotype
452 Pragefest Ludwig & Mayer 1927 E. Lautenbach
Samson Script
453 Prisma Klingspor 1928 Rudolf Koch
454 Prius Antiqua Woellmer 1937
455 Progreso Gans
456 Quick Bauer see Trafton Script
457 Radio Nebiolo see Lipsia Antiqua
458 Radio Weisert see Ella
459 Raleigh American T. F. 1929 Willard Sniffen
460 Ramses Berthold 1912 Hermann Delitsch
461 Rasse Ludwig & Mayer 1924
462 Ratdolt Roman Caslon see Goudy Antique
463 Ratio Latein Stempel 1924 F. W. Kleukens
Ratio Lateina Linotype
464 Razionale Nebiolo da Milano
465 Recta Kursiv Berthold 1926
466 Reform Grotesk Stempel 1908
467 Regal Intertype 1935
468 Regatta Butter 1936
469 Regent Antiqua Weber see Lipsia Antiqua
470 Rekord John see Lipsia Antiqua
471 Rembrandt Bauer 1914 Scheffel
472 Renaissance Haas see Lipsia Antiqua
473 Rheingold Trennert see Forelle
474 Rhythmus Schelter & Giesecke 1932
475 Richmond Blackfriars T. F.
476 Richter Antiqua Ludwig & Mayer 1926 Marcel Richter
477 Riedinger Mediaeval Krebs 1929 Franz Riedinger
478 Robur Deberny & Peignot Georges Auriol
479 Rockwell 371 Monotype 1934
480 Romain XVIIe Siècle Fonderie Typ. Frangaise 1883 Gustave Mayeur
481 Romany American T. F. 1934
482 Römische Antiqua Genzsch & Heyse 1888 Heinz Kömg
483 Ronaldson American T. F. 1884 or. Binney & Ronaldson
Ronaldson Linotype
Ronaldson 10 Monotype



Nr. Name of Series Editing Firm ^ssue°F Designer
484 Roosevelt Bauer see Cheltenham
485 Rosart Gans
Rosart Stevens-Shanks
486 Rosetti American T. F.
487 Rousseau Schelter & Giesecke 1899
488 Ruano Nebiolo 1925 Raffaello Bertieri
489 Rundfunk Berthold 1928 Heinrich Behrmann
490 Rund Grotesk Weber 1931
491 Runnymede Stephenson Blake
Runnymede Monotype
492 Saeculum Stempel 1907 H. HofFmeister
Saeculum Linotype
493 Salut Klingspor 1930 Heinrich Maehler
Salut Stempel
494 Salzmann Antiqua Schelter & Giesecke 1910 M. Salzmann
495 Samson Script see PrSgefest
496 Saskia Schelter & Giesecke 1932 Jan Tschichold
497 Savoy Amsterdam 1936
498 Saxonia Berthold Richard Grimm-Sachsenberg
499 Saxonia Butter see Ella
500 Scarab Stephenson Blake
501 Schablone Weisert 1931
502 Schaefer Versalien Butter 1933 K. H. Schaefer
503 Schelter Antiqua Schelter & Giesecke 1905 M. Salzmann
504 Schneidler Latein Schelter & Giesecke 1919 F. H. Ernst Schneidler
505 Schneidler Schragschrift Schelter & Giesecke 1920 F. H. Ernst Schneidler
506 Schraffierte Antiqua Klingspor 1919 Karl Michel
507 Schraffierte Demeter Butter 1922 P. A. Demeter
508 Scotch Linotype 1919
Scotch American T. F.
Scotch Monotype
509 Script 322 Monotype
510 Script 351 Monotype
511 Script 385 Monotype
512 Select Amsterdam 1929
Select Berthold
513 Semplicita Nebiolo
514 Sensation Stempel 1913
515 Série 18—21 Deberny & Peignot 1905
Baskerville Caslon
516 Shadow American T. F. 1934
517 Shakespeare Mediaeval Schelter & Giesecke 1928 Georg Belwe
518 Siena Nebiolo see Ella Bold
519 Signal Berthold 1931 W. Wege
Sirene Amsterdam
; 520 Sinibaldi Nebiolo 1928 Raffaello Bertieri
521 Sirene Amsterdam see Signal
;522 Skizze Ludwig & Mayer 1936
E523 Solus 276 Monotype 1934 Eric Gill
524 Sonderdruck Antiqua Ludwig & Mayer see Cochin, Nicolas
?525 Sorbonne Berthold see Cheltenham
[526 Sphinx Deberny & Peignot 1927
>'527 Splendor Butter 1937 Wilhelm Berg
1528 Sport Weisert see Ella Bold
[529 Stabil Woellmer 1926
f530 Stadion Butter 1929 Erhard Grundeis
[531 Stahlstich Antiqua Ludwig & Mayer 1913



Nr. Name of Series Editing Firm 1ssue Designer
532 Standard Latein Stempel 1929 Willy Schwerdtner
533 Stanhope Old Style Stevens-Shanks see Genzsch Antiqua
534 Stellar Ludlow
535 Strand Stevens-Shanks
536 Stymie American T. F. 1931
537 Succes Amsterdam
538 Suggestion Bauer
539 Super Grotesk Butter 1930 Arno Drescher
540 Supremo Versalien Butter 1932
541 Tages Antiqua Bauer 1923 Louis Hoell
542 Tango Antiqua Berthold 1913 Ernst Deutsch
Luna Amsterdam
543 Tauperle Schelter & Giesecke 1921 Albert Auspurg
544 Tempo Ludlow 1932 .
545 Tempo Ludwig & Mayer 1930 Walter Hoehmsch
546 Textype Linotype 1930
547 Thannhauser Butter 1929 Herbert Thannhausei
548 Thor Schelter & Giesecke 1920 Georg Belwe
549 Tiemann Antiqua Klingspor 1923 Walter Tiemann
Tiemann Antiqua Typograph
550 Tiemann Mediaeval Klingspor 1909 Walter Tiemann
Wren Pavyers & Builens
551 Times New Roman 327 Monotype 1932
Toscana Linotype
552 Titania Gans
553 Titanic Wagner see Hercules
554 Toscana Linotype see Times New Roman
555 Toskana Berthold see Cheltenham
556 Trafton Script Bauer 1933 Howard Allen Trafti
Quick Bauer
Etoile Deberny & Peignot
557 Trajan Miller & Richard
558 Trajan Krebs 1928
559 Transito Amsterdam 1931 Jan Tschichold
560 Trennert Antiqua Trennert 1926 Friedrich Bauer
561 Trennert Latein Trennert 1932 Friedrich Bauer
562 Trianon Bauer 1905 Heinnch Wieynk
563 Trieste Nebiolo see Kleukens Antiqua
564 Trio Butter 1936
565 Triumf Wagner Albert Auspurg
566 Trocadero Trennert 1927 Albert Auspurg
567 Troubadour Stempel 1928 Wagner & Schmidt
Troubadour Berthold
Troubadour Haas
Butterfly Butter
Comtess Lola Norddeutsche Schriftg.
Donatello Nebiolo
Elegant Weisert
Gracia Amsterdam
Mozart Wagner
568 Ultra Butter see Welt Antiqua
569 Ultra Bodoni American T. F. 1928
570 Ultra Modern Ludlow
571 Umbra Ludlow 1933
572 Umbra Berthold
Nobel „Shadow” Amsterdam



Nr. Name of Series Editing Firm Issue Designer
573 Union Stempel see Cheltenham
574 Unziale Klingspor 1909 Ötto Hupp
575 Vanity Fair Capitals Continental T. F. 1923 Douglas McMurtrie
576 Veltro Nebiolo . „ „ .
577 Venezia Stephenson Blake 1929 Edward P. Prince
Venezia Linotype
578 Venezia Berthold 1907 Georg Schiller
579 Venus Grotesk Bauer 1907
580 Verona Stephenson Blake 1922 (or. Laclede Old Style)
581 Veronese 59 Monotype 1911
582 Ver Sacrum Poppelbaum
Austria Berthold
583 Vesta Berthold 1926 Albert Auspurg
584 Vinne, de American T. F. bef. 1890 Gustave Schroeder
Vinne, de Amsterdam
Vinne, de Caslon
585 Vittorio Nebiolo
586 Vogue Intertype
587 Vogue Stephenson Blake
588 Voitëch Preissig Grégr Voitéch Preissig
589 Vulcan Bold Linotype see Bristol
590 Walbaum Antiqua Berthold 1926 (E. Walbaum 1805)
Walbaum Antiqua Intertype
Walbaum Antiqua 374 Monotype
591 Walgunde Schelter & Giesecke 1906 E. Lautenbach
592 Watteau Schelter & Giesecke 1910 A. Kayser
Byron
Charnwood (bold)
593 Weiss Antiqua Bauer 1926 Emil Rudolf Weiss
Weiss Antiqua Intertype
594 Welt Antiqua Ludwig & Mayer 1932 Hans Wagner
Welt Antiqua Linotype
Luxor (Soldans)
Ultra Butter
595 Wieland Schelter & Giesecke 1926 Georg Belwe
596 Wiener Grotesk Berthold 1912 R. Geyer
597 Wieynk Kursiv Bauer 1911 Heinrich Wieynk
598 Wieynk Mediaeval Weisert 1928 Heinrich Wieynk
599 Wilke Kursiv Woellmer 1933 M. Wilke
600 Wilke Versalien Woellmer 1933 M. Wilke
601 Windisch Kursiv Klingspor 1917 A. Windisch
602 Wren Pavyers & Builens see Tiemann Mediaeval
603 Xylo Krebs
604 Zabel Antiqua Woellmer 1925 Lucian Zabel
605 Zeus Butter 1931 Jan Tschichold



5ELECTED REFERENCES.
rHEORY OF READING IN GENERAL.
Becher, E. : Experimentelle und kritische Beitrage zur Psychologie des Lesens bei kurzen Expositionszeiten. Z. f. Ps. 36, 1904.
Beer, M. : Die Abhangigkeit der Lesezeit von psychologischen und sprachlichen Faktoren. Z. f. Ps. 56, 1910, 264.
Berman, I. R. and Bird, Ch. : Sex difjerences in speed of reading. J. Appl. Ps. 17, 1933, 221.
Bird, Ch. and Beers, F. S. : Maximum and minimum inner speech in reading. J. Appl. Ps. 17, 1933, 182.
Burgess, M. A. : Measurement of silent reading. Col. Univ. Stud. in Philos. 1921, 163.
Cohn, H. L. : Die Hygiene des Auges in den Schalen. Leipzig-Wien, Urban & Schwarzenberg, 1883.
Dearborn, W. F. : The psychology of reading. Arch. Philos. Psychol. Nr. 4.
Col. Univ. Contrib. to Phil. and Psychol. 8, Nr. 1.
Dodge, R. : Recent studies in the relation of eye-movement and Visual perception. Psych. Buil. 3, 1906, 85.
Erdmann, B. and Dodge, R. : Psychologische Untersuchungen über das Lesen, auf experimenteller Grundlage. Halle, 1898.
Eurich, A. E. : The photographical eye-movement records of successful and unsuccessful college students. J. Appl. Ps. 17, 1933, 604.
Finzi, J. : Zur Untersuchung des Auffassungsfahigkeit und Merkfahigkeit. Psych. Arbeiten 3, 1901, 41.
Gilbert, L. C. : An experimental investigation of eye-movements in learning to spell words. Ps. Monogr. Princeton, Ps. Review, 1932.
Re view of the above by M. A. Tinker, J. Appl. Ps. 17, 1933, 490. Goldscheider, A. and Muller, E. : Zur Psychologie und Pathologie des Lesens. Z. f. klin. Medizin 23-11, 1893, 131.
Griffing, H. and Franz, S. I. : On the conditions of fatigue in reading. Ps. Revue 3, 1896, 513.
Huey, E. B. : Psychology and pedagogy of reading. New York, 1910.
Javal, E. : Physiologie de la lecture et de Vécriture. Paris, Félix Alcan, 1905. No longer obtainable.
— : Physiologie des Lesens und Schreibens. Leipzig, Wilh. Engelmann, 1907
(Transl. of the above).
Kirschmann, A. : Ueber die Erkennbarkeit geometrischer Figuren und Schriftzeichen im indirekten Sehen. Arch. f. d. ges. Ps. 13, 352.
— '. Die Augenbewegung beim Lesen. Deutsche Optische Wochenschr. 1917, 8. Korte, A. : Die Gestaltauffassung im indirekten Sehen. Z. f. Ps. 53, 1923, 93. Kutzner, O. : Kritische und experimentelle Beitrage zur Psychologie des Lesens.
Arch. f. d. ges. Ps. 1921, 41.
Litterer, O. F. : An experimental study of visual apprehension in reading. J. Appl. Ps. 17, 1933, 266.
Luckiesh, M., Taylor, A. H. and Sinden, B. H. : The hearing of illumination intensity upon the efficiency of visual operations. Electric World, 78, 1921, 668. Luckiesh, M. and Moss, F. K. : Seeing. A partnership of lighting and Vision. Baltimore, Williams & Wilkins, 1931.



McDougall, R. : On the relation of eye-movements to limiting visual stimuli. Am. J. Physiol. 9, 1903, 122.
Messmer, O. : Zur Psychologie des Lesens bei Kindern und Erwachsenen. Arch. f. d. ges. Ps. 2, 259.
O’Brien, J. A. : Silent reading. New York, MacMillan, 1921.
Parsons, J. H. : Hygiene of reading and of near Vision. Ophth. Rev. 31, 1914, 321.
Peters, C. C. : The influence of speed drills upon the rate and the effectiveness of silent reading. J. Educ. Ps. 8, 1933, 350.
Richtmyer, F. K. and Howes, H. L. : A method of studying the behaviour of the eye under different conditions of illumination. Tr. 111. Eng. Soc. 11, 1916, 100.
Schwender, J. : Die wichtigsten Ergebnisse der experimentellen Untersuchungen iiber das Lesen. Z. f. exp. Paedag. 9, 1909, 169.
Slefrig, S. : The normal school hygiene. London, The Normal Press, 1905.
Starch, D. : Measurement of efficiency in reading. J. Educ. Ps. 6, 1915, 1.
Thorndike, E. L. : Reading as reasoning : a study of mistakes in paragraph reading. J. Educ. Ps. 8, 1917, 323.
Trimble, O. C. and Asperger, D. H. : The “recognition spans” of “good” and “poor” readers; a comparative study. J. Appl. Ps. 19, 1935, 665.
Wager, R. E. : A method of measuring fatigue of the eyes. J. Educ. Ps. 13, 1922, 561.
— : Fixation-accommodation rates as factors in reading. J. Educ. Ps. 15, 1924, 579.
Wiegand, C. F. : Ueber die Bedeutung der Gestaltqualitat für das Erkennen von Wörtern. Z. f. Ps. 48, 1918, 161.
Whipple, G. M. and Curtis, J. N. : Preliminary investigation of skimming in reading. J. Educ. Ps. 8, 1917, 333.
Zeitler, J. : Tachistoskopische Untersuchungen iiber das Lesen. Phil. Stud. 16, 380.
APPARATUS.
Clark, B.: A camera for simultaneous record of horizontal and vertical movements of both eyes. Am. J. Ps. 46, 1934.
Delabarre, E. B. : A method of recording eye-movements. Am. J. Ps. 9, 1898, 572.
Schackwitz, A.: Apparat zur Aufzeichnung von Augenbewegungen beim Zusammenhangenden Lesen (Nystagmograph). Z. f. Ps. 63, 442.
Travis, R. C. : The Dodge mirror-recorder for photographing eye-movements. J. Gen. Ps. 7, 1932, 311.
Weiss, A. P. : The focal variator. J. Exp. Ps. 2, 1917, 106.
LEGIBILITY OF PRINTING TYPES.
Ahrndt, A. : Etwas iiber den guten Werksatz. Klimsch’s Jahrbuch 1937, 16. Frankfurt am Main, Klimsch.
Akimoff, A. : La presse quotidienne genevoise du point de vue de Vhygiène d'ail. Genève, 1912.
Baird, J. W. : The legibility of a telephone directory. J. Appl. Ps. 1, 1917, 30.
Bentley, M. : Leading and Legibility. Psychol. Monogr. 30, 1921, Nr. 3.
Blackhurst, J. H. : Investigations in the hygiene of reading. Baltimore, Warwick & York, 1927.
Burtt, H. E. and Basch, C. : Legibility of Bodoni, Baskerville Roman and Cheltenham type faces. J. Appl. Ps. 7,. 1923, 288.
Cohn, H. L. and Rübenkamp, R. : Wie sollen Biicher und Zeitungen gedruckt werden ? Braunschweig, F. Vieweg & Söhne, 1903.
Committee on the best faces of type and modes of government printing, Report of. London, Stationery Office, 1922.
Committee of the British Association on the influence of schoolbooks on eye-sight. Report of. London, John Murray, 1916.
Dockeray, F. C. : Span of distinct Vision in reading and the legibility of letters. J. Educ. Ps. 1, 1910, 123.
Gilliland, A. R. : The effect on reading of changes in the size of type. El. Sch. J. 24, 1923, 138.
Greene, E. B. : The legibility of typewritten material. J. Appl. Ps. 17, 1933, 713.



//• Greene, E. B. : The relative legibility of Linotyped and typewritten material. J. Appl. Ps. 18, 1934.
v. Hovde, H. T. : The relative effects of size of type, leading and context. J. Appl. Ps. 13, 1929, 601 ; 14, 1930, 1.
Kirschmann, A. : Antiqua oder Fraktur. 3d ed. Leipzig, Deutscher Buchgewerbeverein.
Lobsien, M. : Ueber Lesbarkeit von Fraktur und Antiqua. Langensalza, H. Beyer & Söhne, 1918.
Lyon, O. C. : The telephone directory. Bell Telephone Quarterly, 3, 1924, 174. Mergenthaler Linotype Company : The legibility of type. Brooklyn.
Pergens, E. : La lisibilité des caractères d’impression. Ann. d’ocul. 132, 1904, 402. Pratt, C. C. : A note on the legibility of items in a bibliography. J. Appl. Ps. 8, 1924, 362.
Pyke, R. L. : Report on the legibility of print. London, Stationery Office, 1926. Roethlein, B. E. : The relative legibility of different faces of printing type. Am. J. Ps. 23, 1912, 1.
Schackwitz, A. : Die versuchsmdssige Lösung des Schriftstreites. Leipzig, K. F. Koehler, 1915.
Soennecken, F. : Fraktur oder Antiqua. Bonn, F. Soennecken, 1913.
Terry, P. W. : The reading problem in arithmetic. J. Educ. Ps. 12, 1921, 365. Tinker, M. A.: Reading reaction for mathematicaIformulae. J. Exp. Ps. 9, 1926, 444.
— : The relative legibility of the letters, the digits, and of certain mathematica!
signs. J. Gen. Ps. 1, 1928, 472.
— : How formulae are read. Am. J. Ps. 40, 1928, 476.
— : Numerals versus words for efficiency in reading. J. Appl. Ps. 12, 1928, 190.
— : The relative legibility of modern and old style numerals. J. Exp. Ps. 13, 1930, 453.
— : The influence of form of type on the perception of words. J. Appl. Ps. 16,
1932, 167.
— and Paterson, D. G. : Time-limit versus work-limit. Am. J. Ps. 42, 1930, 101.
— : Studies of typographical factors influencing speed of reading, I. The influence
of type form on speed of reading. J. Appl. Ps. 12, 1928, 359.
— : Studies etc. II. Size of type. Ibid. 13, 1929, 120.
— : Studies etc. III. Length of Line. Ibid. 13, 1929, 205.
— : Studies etc. IV. Effect of practice on the equivalence of test forms. Ibid.
14, 1930, 211.
— : Studies etc. V. Simultaneous variation of type size and line length. Ibid. 15,
1931, 72.
— : Studies etc. VIII. Space between lines or leading. Ibid. 16, 1932, 388.
— : Studies etc. IX. Reductions in size of newspaper print. Ibid. 16, 1932, 525.
— : Studies etc. X. Styles of type face. Ibid. 16, 1932, 605.
— : Studies etc. XI. Role of set in typographical studies. Ibid. 19, 1935, 647.
— : Studies etc. XIII. Methodological considerations. Ibid. 20, 1936, 132. Urban, F. M. : A remark on the legibility of printed types (on numerals). Am.
J. Ps. 23, 1912, 454.
Webster, H. A. and Tinker, M. A.: The influence of type face on the legibility of print. J. Appl. Ps. 19, 1935, 43.
INFLUENCE OF COLOUR AND SURFACE OF PAPER.
Committee on Glare of the Illuminating Engineering Society, Report of. Diffusing media III. Paper and inks. Tr. 111. Eng. Soc. 10, 1915, 379.
Holmes, G. : The relative legibility of black print and white print. J. Appl. Ps. 15, 1931, 248.
Preston, K., Schwankl, H. P., and Tinker, M. A. : The effect of variation in color of print and background on legibility. J. Gen. Ps. 6, 1932, 459.
Stanton, F. N. and Burtt, H. E. : The influence of surface and tint of paper on the speed of reading. J. Appl. Ps. 19, 1935, 683.
Sumner, F. C.: Influence of color on legibility of copy. J. Appl. Ps. 16, 1932, 231.
Tinker, M. A. and Paterson, D. G. : Studies etc. VI. Black type versus white type. J. Appl. Ps. 15, 1931, 241.
— : Studies etc. VII. Variations in color of print and background. Ibid. 15, 1931, 471.



Tinker, M. A. and Paterson, D. G. : Studies etc. XII. Printing surface. J. Appl. Ps. 20, 1936, 128.
Webster, H. A. and Tinker, M. A. : The influence of paper surface on the perceptibility of print. Ibid. 19, 1935, 145.
ATMOSPHERE-VALUE.
Berliner, A. : Atmospharenwert von Drucktypen. Z. f. angew. Ps. 17, 1920, 165. Da vis, R. C. and Smith, H. J. : Determinants of feeling tones in type faces. J. Appl. Ps. 17, 1933, 74.
Franken, R. B. : Tests to determine the best type for your advertising. Printer’s Ink Monthly, 9, 1924, 31.
Hevner, K. : Experimental studies of the affective value of colors and lines. J. Appl. Ps. 19, 1935, 385.
Klaqes, L. : Handschrift und Charakter. Leipzig, J. A. Barth, 1932.
— : Graphologischés Lesebuch, Ibid. 1933.
— : Grundlegung der Wissenschaft vom Ausdruck. Ibid. 1936.
— : Grundlagen der Charakterkunde. Ibid. 1928.
Poffenberger, A. T. and Franken, R. B. : Appropriateness of type faces. J. Appl. Ps. 7, 1923, 312.
Prandtl, A. : Experimente über den Einfluss von gefühlsbetonten Bewusstseinslagen auf Lesezeit und Betonung. Z. f. Ps. 60, 1911, 26.
Schiller, G. : An experimental study of the appropriateness of color and type in advertising. J. Appl. Ps. 19, 1935, 652.
Walton, W. E. and Morrison, B. M. : A preliminary study of the affective value of colored lights. J. Appl. Ps. 15, 1931, 294. Contains bibliography. Zeylmans van Emmichoven, F. W. : De werking der kleuren op het gevoel. Utrecht, W. de Haan, 1923.
ART AND SCIENCE OF ADVERTISING.
Periodicals : Art and Industry (Annual : Modern Pubücity) ; Arts et Métiers Graphiques (Annual : Publicité 19. .) ; Buch- und Werbekunst; Gebrauchsgraphik ; Printer's Ink; Die Reklame (till 1933); De Reclame; O ff. Org. Gen. v. Reclame (till 1938); Revue der Reclame. These are principal sources for commercial art.
Eliasbero, W. : Reklamewissenschaften. Brünn, Rudolf M. Rohrer, 1936. Farrar, G. P. : Typography of advertisements that pay. New York, Appleton. Hollingworth, H. L. : Advertising and Selling. New York, Appleton, 1913. Lucas, D. B. : The optimum length of advertising headlines. J. Appl. Ps. 19, 1935. Poffenberger, A. T. : Applied Psychology. lts principles and methods. New York, Appleton. 1927.
— : Psychology in Advertising. Chicago, A. W. Shaw, 1925.
Raffé, W. G. : Poster Design. London, Chapman & Hall, 1932.
Reiner, Imre : Typography in advertising. Art & Ind. Dec. 1937.
Rheinstrom,' C. : Psyching the ads. New York, Covici Friede, 1929.
Roels, F. : Psychologie der reclame. Amsterdam, H. J. W. Becht, 1938.
Scott, W. D. : The psychology of advertising. New York, Dodd, Mead ; London,
Pitman, 1932.
Sherbow : Effective type using for advertising. Sherbow, 1922.
— : Type chart for advertising, ibid. 1921.
— : Making type work. New York, Century, 1926.
Tolmer, A. : Mise-en-page. London, The Studio, 1931.
Zur Westen, W. von : Reklamekunst. Bielefeld-Leipzig, Velhagen & Klasing, 1903.
REPRODUCTIONS. THE ALPHABET.
The specimens of the foundries and of the manufacturers of composing machines remain the principal sources for the study of type designs.
Alphabete des Morgen- und Abendlandes. Berlin, Reichsdruckerei, 1924.
Brandi, K. : Unsere Schrift. Göttingen, Vandenhoeck & Rupprecht, 1911.



Carlyle, P. and Oring, G. : Letters and lettering. New York-London, McGrawHill, 1938.
Chappell, W. : The anatomy of lettering. New York-Toronto, Loring & Mussey. Degering, H. : Die Schrift. Berlin, Ernst Wasmuth, 1929.
Freshwater, G. J. and Bastien, A. : Printing types of the world. London, Pitman, 1931.
— : Dictionary of advertising and printing. Ibid. 1930.
Goudy, F. W. : The alphabet. Pelham NY, Bridgman.
— : The elements of lettering. Ibid.
Handbuch der Schriftarten. Leipzig, A. Seemann, 1926—’35, with supplts. Contains the entire German type production.
Hoffmann, H. : Schriftatlas. Stuttgart, Julius Hoffmann, 1930.
Horodisch, A. : Die Schrift im schonen Buch unsrer Zeit. Berliner Bibliophilenabend, 1931. Not for sale.
Lettervormen. Amsterdam, Ahrend, 1938.
McMurtrie, D. C. : Alphabets. Pelham NY, Bridgman.
— : Type Design. Ibid.
Taylor : The alphabet. London, 1883.
Tregurtha, C. Maxwell : Type and typefaces. London, Pitman.
Waite, H. E. : Alternative type faces. London, Techn. Publ. Cy., 1936.
Wetzig, E. : Ausgewahlte Druckschriften. Leipzig, Verein Leipziger Buchdr.Besitzer E. V., 1925.
MISCELLANEOUS WORKS ON TYPES AND TYPOGRAPHY.
Arts et les techniques graphiques, Les. No. spéc. 59, Arts et Mét. Graph. 1937.
Contains also an introduction to the Peignot type.
Audin, M. : Le livre. Son architecture, sa technique. Lyon, 1924.
Bauer, F. : Chronik der Schriftgiessereien in Deutschland und den deutschsprachigen Nachbarlandern. Offenbach am Main, 1928.
— : Normung der Buchdrucklettern. Leipzig, Deutscher Buchgewerbeverein. Beaujon, Paul : On the choice of type faces. Monot. Ree. 1933, 2. Biedermann, L. W. von : Die deutsche Schriftgiesserei. Eine gewerbliche Biblio-
graphie. Herausgeg. v. O. Jolles. Berlin, 1923.
Birrell and Garnett : Catalogue of (ƒ) Typefounders’ specimens; (II) Books printed in types of historie importance ; (III) Works on typefounding, printing and bibliography. London, 1928.
Bogeng, G. A. E. : Geschichte der Buchdruckerkunst. Hellerau, 1930. Claudin, A. : Histoire de Timprimerie en Trance. Paris, 1900—’04.
Cyliax, W. : Die Beziehungen zwischen Typographie und Graphik. Typ. Mitt. 1936, 4.
Faulmann, C. : Illustrirte Geschichte der Buchdruckerkunst. Wien, 1881—’82. Goudy, F. W. : The capitals of the Trajan column at Rome. New York, Oxford University Press, 1936.
Hessel, A. : Von der Schrift zum Druck. Z. d. Deutschen Vereins f. Buchw. 1923, 89. Hoffmeister, H. : Die Entstehung einer Schrift. Leipzig, Deutscher Buchgew. ver. 1927.
Johnson, A. F. : Periods of typography : The Italian sixteenth century. London, 1925.
— : French sixteenth century printing. London, 1928.
— : Type designs. London, Grafton, 1934.
Jones, H. : Type in action. London, Sidgwick & Jackson, 1938.
Kautzsch, R. : Wandlungen in der Schrift und in der Kunst. Frankfurt a. M., Gutenberg Gesellschaft, 1929.
Kittredge, W. A. : Getting people to read your book through goodtypography.
Printer’s Ink Monthly. 2, 1921, 23.
Koch, Paul : The making of printing types. The Dolphin 1, 1933, 24.
Legros, L. A. and Grant, J. C. : Typographical printing surfaces. London, Longmans, Green, 1916.
Lorck, C. B. : Handbuch der Geschichte der Buchdruckkunst. I (—1750), II (—1880). Leipzig, 1882—83.



McMurtrie, D. : Modern typography and lay-out. Chicago, Eyncourt Press. Morison, S. : First principles of typography. Cambridge Univ. Press, 1936.
— : Decorated types. The Fleuron 6, London, 1928.
— : Towards an ideal type. Ibid. 2, 1924.
— : Type designs of the past and present. London, The Fleuron, 1926.
— and Jackson, H. : A brief survey of printing history and practice. Ibid. 1923. Müller, Aug. : Lehrbuch der Buchdruckerkunst. Leipzig, J. J. Weber, 1926. Poortenaar, J. : Boekkunst en grafiek. Antwerpen, De Sikkel; Amsterdam,
Poortenaar, 1935.
Renner, P. : Typographie als Kunst. München, Georg Müller, 1922. Out of print.
— : Revolution der Buchschrift. Gutenberg Festschrift 1925, 279.
— : Die Zukunft unsrer Druckschrift. Typ. Mitt. 1926, 86.
— : Schriften der Zukunft. Ibid. 25, 1928, 189.
— : Type und Typographie. Arch. f. Buchgew. 65, 1928, 453.
— : Neue Ziele des Schriftschaffens. Graph. Betrieb, 1929, 187.
— : Formprobleme der Druckschrift. Imprimatur 17, 1930.
Sauberlich, O. : Buchgewerbliches Hilfsbuch. Leipzig, O. Brandstetter.
Tiemann, W. : Mehr Respekt vor dem Buchstaben ! B.D.G.-Blatter, Festnr. Aug.
1929.
Tschichold, J. : Die neue Typographie. Berlin, Bildungsverb. d. deutschen Buchdr. 1927. Out of print.
— : Eine Stunde Druckgestaltung. Stuttgart, F. Wedekind, 1930.
— : New life in print. Spec. Nr. Commercial Art, July 1930.
— : Schriftmischungen. Typ. Mitt. 1935, 2.
— : Typographische Entwurfstechnik. Stuttgart, F. Wedekind, 1932.
— : Typographische Gestaltung. Basel, Benno Schwabe, 1936.
Unger, A. W. : Die Herstellung von Büchern, lllustrationen, Akzidenzen u.s.w. Halle a S., 1923.
Updike, D. B. : Printing types, their history, forms and use. Cambridge, Mass.
Harvard Univ. Press, 1937. Oxford University Press.
Vinne, Th. L. de : Plain printing types. New York, Am. Printer, 1900. London, Raithby, Lawrence.
Warde, Mrs. B. L. : Plumed hats and modern typography. Comm. Art and Ind., Jan. 1932.
— : Size of print, in : Penrose's Annual 1938.
Wattenbach, W. : Das Schriftwesen im Mittelalter. Leipzig, 1896.
Weise, O. : Schrift und Buchwesen in alter und neuer Zeit. Leipzig, 1899. ^ilson, G. F. : Look at types with eyes that see. Comm. Art and Ind., Dec. 1934.
PERIODS ÓR GROUPS IN TYPOGRAPHY.
Important articles appeared in every volume of The Fleuron, The Dolphin and Ars Typographica.
Art of the book, The. Ed. by Ch. Holme, London, The Studio, 1914. Contains : W. D. Orcutt : The art of the book in America ; L. Deubner : The art of the book in Germany.
Bauer, Konrad F. : Bestiarium Typographicum. Z. f. Bücherfr. 21, 1932, 192 (on 18th. century display-types).
Buchkunst, Die neue. Studiën im In- und Ausland. Ed. by R. Kautzsch for Gesellschaft der Bibliophilen, Weimar, 1902. Contains an article by H. C. Marillier on contemporary English typ. and by Franz Blei on cont. American typ. Buchkunst 1. Leipzig, Staatl. Akad. f. graph. Künste u. Buchgew. 1931. Contains : S. J. A. Symons : Die Englischen Pressen der Gegenwart; S. H. de Roos : Ueber die neuzeitliche Buchkunst und Schriftgestaltung in den Niederlanden ; Maximilien Vox : Das schone Buch in Frankreich. Eine Uebersicht über zehn Jahre; Dr. Ernst Naményi : Ungarische Buchkunst; I. Tugenhold : Die Kunst des Buchgewerbes in der Sovjet- Union ; Arthur Novak : Die buchkiinstlerischen Bestrebungen in der Tschechoslovakei.
Cockerell, S. C. : A note by William Morris on his aims in founding the Kelmscott Press, etc. Hammersmith, 1908.



Grautoff, O. : Die Entwicklung der modernen Buchkunst in Deutschland. Leipzig, Hermann Seemann Nachf., 1902.
Gress, E. G. : Fashions in American typography, 1780—1930. New York, Harper, 1931.
Halliday Sparling, H. : The Kelmscott Press and William Morris, master-craftsman. London, 1924.
Halsbeke, Ch. L. van : L'art typographique dans les Pays Bas depuis 1892. Maastricht-Bruxelles, 1929.
Haupt, G. : Schrift und Handwerk. Offenbacher Sonderheft. Philobiblon 7, 1934.
Herrmann, G. : “Die Jugend” und ihr Künstlerkreis. Z. f. Bücherfr. 4, 1900, 2-3.
Hesse, R. : Le livre d'art du XlXe siècle a nos jours. Paris, 1937 (?). Deals with French typography.
Johnston, P. : Biblio-typographica. A survey of contemporary fine printing style. New York, 1930.
Krimpen, J. van : Typography in Holland. The Fleuron 7, 1930.
Livre d'art international, Le, No. spéc. 26. Arts et Métiers Graph. 1931. Contains articles by Bertrand Guégan on cont. typ. in all countries, and by Jean Bruller on modern French typ.
Modern book production. Ed. by C. G. Holme. London, The Studio, 1928.
Morison, S. : Modern Fine Printing. London, 1925.
— : Four centuries of fine printing, 1500—1914. London, Ernest Benn, 1924.
— : Meisterdrucke aus vier Jahrhunderten. Transl. by A. Simons. Berlin, Ernst
Wasmuth, 1934.
Nonesuch Century, The : An appraisal, a personal note and a bibliography of the first hundred books issued by the press, 1923—1934. London-New York, 1936.
Pollard, A. W. : Private presses and their influence on the art of printing. Ars Typographica 4.
Printing Number of The Times. London, 1912 and 1929.
Printing in the 20th century. London, The Times, 1930.
Reichner, H. : Die Druckerkunst in den Vereinigten Staaten von Nordamerika. Wien, H. Reichner, 1927.
Simon, O. and Rodenberg, J. : Printing of today. London-New York, 1928.
Thibaudeau, F. : La lettre d'imprimerie. Paris, 1921. On French types.
Tschichold, J. : Die gegenstandslose Malerei und ihre Beziehungen zur Typographie der Gegenwart. Typ. Mitt. 1935, 6.
ARTISTS.
P. Behrens, by G. Haupt, in : Die neue Buchkunst, Weimar, 1902.
Raffaello Bertieri, Eine Stunde mit —, by Pizzutto. Arch. f. Buchgew. 1934, 3.
W. A. Dwiggins, by Paul Standard, Penrose’s Annual 1937. London, Lund Humphries.
O. Eckmann, by H. Loubier, in : Die neue Buchkunst, Weimar, 1902.
Otto Eckmann’s kunstgewerbliche Entwicklung, by M. Osborn. Deutsche Kunst und Dek. 2, 1900, 313.
Eric Gill, by E. Ingham, Penrose’s Annual 1935, London, Lund Humphries.
—, by Mrs. B. L. Warde. Comm. Art and Industry, June 1933.
Otto Hupp, by F. Bauer, Klimsch’s Jahrbuch 1934. Frankfurt am Main.
Edward Johnston and English lettering, by Anna Simons. Monogr. künstlerischer Schrift II. Berlin-Leipzig, Heintze & Blanckertz.
Rudolf Koch, Designer of letters, by J. Rodenberg, Penrose’s Annual 1935.
Rudolf Koch und die Offenbacher Schreiberwerkstatt, by R. Conrad. Gebrauchsgraph. 4, 8.
—, by E. Engel, Klimsch’s Jahrbuch 1934.
Rudolph von Larisch, by F. H. Ehmcke. Leipzig, Staatl. Akad. f. Graph. K. u. Buchgew. 1926.
Anti-Larisch, by G. Kühl. Arch. f. Buchgew. 42, 1905.
William Morris, designer, by C. G. Crow. London, The Studio, 1934. Also for further references.



William Morris, his influence on American printing, by F. W. Goudy. Philobiblon 7, 1934, 4.
Anna Simons, by Dr. E. Hölscher. Monogr. künstlerischer Schrift I. Berlin-Leipzig, Heintze & Blanckertz.
Walter Tiemann, by J. Zeitler, in : Satz und Druck, Musterheft 1936. Berlin, Deutscher Drucker.
D. B. Updike, An account of the work of —, The Merrymount Press of Boston,
by Dr. G. P. Winship. Wien, H. Reichner.
—, The Merrymount Press, by P. Standard, Penrose’s Annual 1936.
E. R. Weiss zum 50en Geburtstag, 12 Oct. 1925, by H. Reichner. Leipzig, Insel-
verlag, 1926.
E. R. Weiss, Gesprache mit — über Buchkunst, by R. Schacht. Typ. Mitt. 23, 174. SPECIAL TYPES.
Beaujon, Paul (Mrs. B. L. Warde) : The book of verse. Monotype Recorder, 35, 2, 1936.
Italie typography, as demonstrated in some recent work on the Ludlow. Forew. by W. A. Kittredge. Chicago, 1927.
McMurtrie, D. C. : The Ludlow system of slug composition. Chicago, 1927. Meynell, F. : The typography of newspaper advertisements. London, Ernest Benn, 1929.
Morss, R. D. : The neglected schoolbook. Monot. Ree. 34, 2, 1935.
SchonfielD, H. J. : The new Hebrew typography. London, Dennis Archer. Schwarz, H. : Ueber die Schriften fiir den Zeitungssatz. Arch. f. Buchgew. Gebr. graph. 65, 217.
The changing newspaper. Monot. Ree. 35, 1, 1936.
The modern timetable. Monot. Ree. 34, 1, 1935.
GOTHIC TYPES (BLACK LETTER).
Bauer, F. : Die gotische Druckschrift. Klimsch’s Jahrbuch 1934, 10.
Bauer, Dr. Konrad F. : Leonhard Wagner, der Schöpfer der Fraktur. Z. f. Bücherfr. 25, 1936.
Black Letter, lts history and current use. Monot. Ree. 36, 1, 1937.
Crous, E. and Kirchner, J. : Die gotischen Schriftarten. Leipzig, Klinkhardt & Biermann, 1928.
Haupt, G. : Deutsche Schriften. Klimsch’s Jahrbuch 1934, 15.
Kautzsch, R. : Die Entstehung der Frakturschrift. Mainz, 1922.
Koch, R. : Die deutsche Schrift. Klimsch’s Jahrbuch 1934, 82.
HANDLETTERING.
Periodical : Die Zeitgemasse Schrift. Berlin-Leipzig, Heintze & Blanckertz. Hewitt, Graily : The pen and type design. Oxford Univ. Press, 1926. Johnston, E. : Writing and illuminating and lettering. London, Pitman.
— : Schreibschrift, Zierschrift und angewandte Schrift. Transl. by Anna Simons.
Leipzig, Klinkhardt & Biermann, 1928.
Koch, R. : Das Schriftschreiben. Typ. Mitt. 23, 1926, 82, 173.
Larisch, R. von : Unterricht in ornamentaler Schrift. Wien, Staatsdruckerei, 1926.
— : Ueber Zierschriften im Dienste der Kunst. München, 1899.
— : Beispiele künstlerischer Schrift. Wien, Anton Schroll, 1900.
Lettering of today. Ed. by C. G. Holme. London, The Studio, 1937. Contains : Dr. E. Hölscher : The principles of hand lettering; A. Fairbank : Calligraphy; Anna Simons : Lettering in book product ion; Percy J. Smith : Lettering in association with architecture ; R. Haughton James : Lettering in advertising.
Wade, Cecil : Modern lettering from A to Z. London, Pitman.







STELLINGEN.
I.
De leesbaarheidscoëfficiënt van een lettertype kan slechts bepaald worden voor een beperkte categorie van lezers, m.a.w. er bestaat ten aanzien van leesbaarheid geen gemiddelde lezer, noch een gemiddelde tekst of lettersoort.
II.
Het onderzoek van Schulte (Prakt. Ps. 2, 28) over de optimale letterspatie verwaarloost verschillen in grondvorm, dikteverdeeling en grootte van de letters ; de conclusies kunnen daarom niet gegeneraliseerd worden.
III.
Voor de vermoeidheidsstudie van het oog ten bate van het leesbaarheidsonderzoek kan het electromyogram en de registratie van den ooglidreflex, bij voortgezette leesperioden en constante opmerkzaamheid, met vrucht gebruikt worden.
M. Luckiesh en F. K. Moss, J. Appl. Ps. 22, 140.
IV.
Beschouwd als grafische neerslag van uitdrukkingsbewegingen kunnen lettervorm en mise-en-page, evenals het ornament, waardevol materiaal leveren voor de studie van stijlgeschiedenis.
V.
Het verdient aanbeveling om voor de bepaling van de relatieve inprentingskracht van annonces zich te bedienen van de methode van BARKLEY. k l barkley, J. Appl. Ps. 15, 390; 16, 74.
VI.
Dat Klages en zijn geestverwanten, in het voetspoor van Nietzsche, den Geest (Wil) meenen te moeten bestrijden, in plaats van hem (zonder waardeering) te aanvaarden als tegenpool van de Ziel, is historisch te verklaren als verzet tegen den tijdgeest.



VII.
De quantitatieve en qualitatieve verbetering van de arbeidscurve (incl. ongevallen) door inschakeling van geautoriseerde pauzes, is in de eerste plaats te danken aan verbetering in moreel (houding tegenover de bedrijfsleiding en tegenover den arbeid zelf).
Vgl. E. Mayo : Human problems of an industrial civilization ;
M. S. Viteles : Industrial psychology.
VIII.
Terecht wijst Baumgarten de onderscheiding van Klages tusschen echte en onechte karaktereigenschappen af; haar meening echter, dat een schijneigenschap iets zou voorwenden, is onjuist.
Fr. Baumgarten : Die Charaktereigenschaften.
IX.
De „mimische teekens” (Lersch) verkrijgen hun volledige beteekenis eerst in het licht van de geheele somatische expressie.
Ph. Lersch : Mimik und Physiognomik.
X.
Het is de uiteindelijke taak van de praktische karakterkunde om in den geest van C. G. Jung de aangeboren en verworven eigenschappen en bekwaamheden van elk individu tot een harmonisch geheel te ontwikkelen, zonder vervorming tot een specifiek mannelijk heldentype.
XI.
Het hooren en beoefenen van dansmuziek heeft beteekenis voor de ontwikkeling van de rijpere jeugd, aangezien het in staat stelt om le. innerlijke belevingen adequaat uit te drukken, en 2e. in de „helden der jazzwereld” eigen idealen te belichamen.
XII.
De parallel tusschen de muziek van J. S. Bach en de jazz, het eerst door Jean Wiener getrokken, is in staat het wezen van de uitdrukkingswijze van den eerstgenoemde te verhelderen.
XIII.
Het aantal lesuren in de klassieke talen mag onder geen beding nog meer beperkt worden, wanneer de gymnasiale opleiding haar groote vormende werking wil behouden.
Vgl. N. G. M. van Doornik ; De moderne gymnasiast tegenover zijn klassieke vorming. Groningen, 1938.