216 STELLAR PHOTOGRAPHY. 



the negative residuals, 23, 13, 5, 2, and 1 have the values —0.1, —0.2, —0.3, —0.4, 

 and —0.5. An examination of the entire number of 163 estimates showed that the 

 frequency •with which the various tenths of a magnitude were employed varied 

 greatly; the figure 5 occurred 61 times; 0, 47 times; 8, 33 times; 3, 15 times; 2, 

 7 times ; 4, 5 times ; 7, 4 times ; 6, once ; 1 and 9, not at all. A test of the rela- 

 tive sensitiveness of the plates is afforded from the average value of the residuals 

 relating to the images contained on each. The means for Plates 209, 248, 327, and 

 361, are +0.046, —0.040, +0.031, and —0.009. Accordingly, the stars on Plate 209 

 appear on the average about one twentieth of a magnitude fainter than on the 

 average of the four plates. These corrections are so small that it is hardly neces- 

 sary to apply them. 



Measures were made of the positions of the images on Plate 327, to determine 

 the degree of precision to be expected from the application of photography to 

 transit instruments. The shortest exposures do not seem to have been much less 

 tlnn a second, owing to the time required to cover and uncover the glass. The 

 images formed by the shorter exposures were minute circular dots, which for the 

 fainter stars did not exceed 0.008 cm. in diameter. They were not perfectly sym- 

 metrical, probably owing to the diffraction Avhen the lens was nearly covered. All 

 the stars were, however, subjected to the same conditions. About a dozen of the 

 brighter stars formed images even with the shortest exposures. The intervals 

 between these images were measured by attaching the photograph to a micrometer- 

 screw having a pitch of one twenty-fourth of an inch, by which it could be moved 

 across the field of a microscope. The magnifying power was only ten diameters, 

 and could doubtless have been increased with advantao;e. Eight settings were made 

 on each star, seven on the centre of the dots formed by the exposures of two seconds, 

 of one second, and of half a second, and one on the end of the last trail. Nine stars 

 were measured in this way, and the first of them was taken a second time to see 

 if the instrument had moved. The intervals were next found by taking the first 

 differences of these readings. The means of the ten values of these seven intervals, 

 expressed in I'evolutions of the micrometer-screw, were 0.380, 0.230, 0.139, 0.353, 

 0.145, 0.142, and 0.497. Each revolution is equivalent to 13'.87. The residuals 

 found by subtracting these mean values from the observed intervals are given in 

 Table V., expressed in seconds of time. As in Table IV., the first column serves to 

 designate the star by its number in the Catalogue of Wolf. 



