Onderstaande tekst is niet 100% betrouwbaar

is shown. By comparing the results obtained for the individual groups separately with the mean value we get the values of R given in Table 3.

Table 3.

Numb.

Limits of deel. Source A ^ g^ars P

-f 90° to -f 520 Br 2Ó9°.4 297 2 7 + 0.2

+ 520 to —20° Br 2Ö9°.8 2115 12.2 +0.6

— 20° to — 310 Br 273°.s 139 0.9 + 4 3

— 20° to —40° N 2Ó9°.6 114 0.8 +0.4

— 20° to — 350 Tay 2Ö20.8 198 0.6 — 6.4

— 40° to — 52° N 266°.7 38 0.3 —2.5

— 40° to — 520 Gi 269°.0 490 3.7 — 0.2

— 520 to —90° Auw. 265°'! 225 2.0 —4.1

It will be remarked that in the zone — 40° to — 520 the residuals both for Newcomb and for Gill are negative, whereas in the preceding table the combination of the two gives a positive value of R. The cause lies of course in the different distribution of the weights over the three-hour-groups. In the present table every star has got a weight i, in the preceding solution opposite three-hour-groups got equal weights in this zone. No importance need be attached therelore to the series of negative values in the second half of the column of R. If we consider the residuals as wholly accidental we may derive a mean error from them. Denoting the mean error ') belonging to the unit of weight by f0, we get:

(19 ) fo = ± 3°-4

and as a preliminary result for A and its mean error, from all groups together

(20 ) A = 2Ó9°.2 ± o°.7.

This value has been derived from 3517 stars if a star of Taylor is counted for Therefore if we denote by f100 the mean error in the result derived from a group of 100 stars distributed over the whole of the sky we get:

(21 ) ^100 — 4

Comparing this value with those for the mean errors in A, determined in A. N. n°. 3722 from different groups of stars and taking into account the numbers of stars from which they have been derived , the result must be considered satisfactory.

') In future we will always work with mtan errors.

Sluiten