1 88 



REPORTS ON INVESTIGATIONS AND PROJECTS. 



the report of the preceding year have been traced to small unsymmetrical 

 temperature deformations of the mirror. The fluctuations in the average 

 value of the correction from plate to plate seem also to be due to this cause. 

 Coma, in the case of the reflecting telescope, is extremely sensitive to small 

 changes of figure, and a small variation in temperature is sufficient to pro- 

 duce an appreciable change in the size of images not on the axis. Owing to 

 the distribution of the light intensity within the image, an unsymmetrical def- 

 ormation of the mirror must produce corresponding irregularities in the 

 distance error when different directions from the axis are considered. 



The changes of figure involved are usually so slight that, under ordinary 

 conditions of seeing, it is quite impossible to study them in detail with the 

 knife-edge when used visually. On this account arrangements were made 

 whereby the actual state of the figure may readily be photographed, and the 

 result above stated has been derived from a comparison of the photographic 

 record of the figure with simultaneous determinations of the distance error. 



In explanation of the observed irregularities in the distance error it should 

 be stated that formerly it was not possible to cover the reflector with the 

 canopy when the Cassegrain spectrograph was in position. As the arrange- 

 ment of the program was such that the photometric observations immediately 

 followed those with the spectrograph, part of the photometric plates were 

 exposed when the mirror was in an abnormal condition. The instrument is 

 now covered with the canopy at the end of each night's work, and as a result 

 the difficulty has largely been eliminated. 



From the experience thus far gained it appears that with care in the mat- 

 ter of temperature control reliable photometric results may be obtained 

 through the use of an average value of the distance correction. 



The results for the magnitude scale of the Polar Sequence are based upon 

 15 plates and include 31 separate determinations of the scale from magni- 

 tude 10.5 downward to 17.6, the lower limit being shown on but a single 

 plate, however. Exposures with diaphragms of 32 and 14 inches and a wire- 

 gauze screen, in combination with exposures with full aperture, were used 

 for the derivation of the scale. The differences in scale for Mount Wilson 

 and Harvard are as follows : 



From magnitude 10.5 to 15.5 the deviations 

 are insignificant. The divergence for the fainter 

 stars requires further investigation. The mean re- 

 sults for the two diaphragms and the wire-gauze 

 screen are in satisfactory agreement throughout. 

 This was to have been anticipated, for, as pre- 

 viously reported, the diffraction effect in the case 

 of the diaphragms is negligible over a range of 

 5 magnitudes. The extension of the scale in the direction of both brighter 

 and fainter stars is now in progress. 



