MOUNT WILSON SOLAR OBSERVATORY. 219 



nitudes brighter. The mean of a few measures of its radial velocity 

 gives a value almost identical with that of the larger star, thus mak- 

 ing the physical connection of the two practically certain. The 

 radial velocity of the fainter star may be subject to a small variation, 

 although the fact that it is a very close visual binary complicates 

 observations of its velocity. 



CLASSIFICATION OF SPECTRA. 



The classification of the spectra has been continued regularly in 

 connection with the measurement, for the most part by Mr. Kohl- 

 schiitter. The results for about 300 stars of types A and B were 

 communicated to Professor Pickering and have been employed by 

 him in a comparison of the relative advantages for classification pur- 

 poses of spectra obtained with sht spectrographs and with the objec- 

 tive prism. The spectral classification of the stars with measured 

 parallaxes has also been thoroughly revised and has proved of great 

 value in an investigation of the luminosity curve of K-type stars now 

 being completed by Mr. Kohlschutter. In this connection he is 

 giving considerable attention to the possibility of finding in the 

 relative intensities of certain spectral lines criteria bearing on the 

 absolute luminosities of stars. The recognition of the important 

 applications of the registering micro-photometer devised by Professor 

 Koch to the determination of the intensities of spectral lines is lead- 

 ing to a study of additional lines to be used for classification purposes 

 on the basis of their absolute intensities. 



SPECTRA OF STARS IN THE HERCULES CLUSTER M 13. 



Two photographs of the spectrum of the Hercules cluster Messier 

 13 have been obtained by Mr. Pease with the new focal-plane spec- 

 trograph. The first plate had an exposure of 21 hours with the slit 

 set at a distance of 11" south of the center of the cluster. The 

 second had an exposure of 22 hours with the slit 18" north of the 

 center. The direction of the slit was east and west for both photo- 

 graphs and its width was 0.025 mm. for the first plate and 0.050 mm. 

 for the second. The guiding was as accurate as possible throughout 

 and as a result the spectra of the individual stars are separated clearly. 

 Of these it has been possible to classify 6 on the first photograph and 13 

 on the second, a total of 19 separate stars. The spectra are divided 

 as follows, taking as the unit five divisions of the Harvard scale : 



AO 2 



A5 5 



FO 2 



F5 8 



GO 2 



It is clear from these results that no satisfactory conclusions as 

 to the spectra of the stars in globular clusters can be drawn from 

 photographs of their integrated fight. 



