254 CARNEGIE INSTITUTION OF WASHINGTON. 



The adaptation of the 10-inch photographic triplet for use with an 

 objective prism has rendered possible a systematic investigation by Mr. 

 Shapley of all the brighter Cepheids. The spectrum of every Cepheid 

 variable investigated, whatever its period or apparent magnitude, 

 undergoes a periodic variation in spectral type synchronously with the 

 change in light, similar to that observed for RS Bootis. At and near 

 the time of maximum light the spectrum reaches its bluest stage; at the 

 minimum it is reddest. On the average the change is nearly a spectrum 

 interval, and appears to include successive stages which are almost if 

 not actually identical with known types of spectra. This apparently 

 general characteristic of Cepheid variables supports the hypothesis that 

 the immediate cause of their light variation is to be found in periodic 

 disturbances in their atmospheres. Further, the phenomenon of con- 

 spicuous and rapid variation of spectral type probably has considerable 

 significance in the more general problems of the physical evolution of 

 stars. In the furtherance of this investigation more than 300 spectro- 

 grams of Cepheids have been made. For 18 stars the spectra were found 

 to be distinctly variable, and in no case has the spectrum of a Cepheid 

 remained constant when its light changes. 



For a further study of the differences in the spectra at maximum 

 and minimum hght, high-dispersion spectrograms of 5 Cephei were 

 made by Mr. Adams and Mr. 8hapley with the 60-inch telescope. In 

 addition to the change in spectral class, these yielded other evidence 

 that the temperature of the gases constituting the star's absorbing 

 envelope is higher at maximum than at minimum. At maximum the 

 high-temperature lines, such as those of Fe, Ti, Sr, and Cr, are very 

 strong and the low-temperature lines weak; while at minimum the 

 reverse is the case. The radial velocity was determined both from the 

 Fe lines showing large pressure displacements in laboratory experiments 

 and from those showing small displacements. The difference is such 

 as to indicate that the pressure at the radiating surface of the star is 

 less than one atmosphere, and both in sign and amount is in good agree- 

 ment with the pressure shifts shown by the solar lines. 



NEBULAK SPECTROSCOPY. 



Photographs of the spectra of several spiral nebula? have been made 

 with the small spectrograph at the primary focus of the 60-inch 

 reflector. The most important of these is an 80-hour exposure upon 

 N. G. C. 4594, with the slit along the axis of the spindle. A long sUt 

 was used and, with the aid of a silvered glass plate ruled at regular inter- 

 vals, comparison spectra were photographed at intervals of 1 nun. 

 on the plate. The spectrum represents approximately the central half 

 of the nebula and is of type G5. Measurements were made on the 

 successive spectral strips to a distance of about 2' on either side of the 

 nucleus. The results may be summarized as follows : 



