130 
ASTRONOMY: ADAMS AND JOY 
selected from among those classified by the Harvard observers, and these 
spectra have been used for comparison purposes as well as that of the sun. 
The spectra have been compared by the aid of a Hartmann spectro-comparator. 
A summary of the results is given in brief form in the accompanying table. 
It became clear from the very first comparisons that the hydrogen lines are 
abnormal in the Cepheid spectra, being in some cases several times as strong 
as in the sun, although in other respects the spectra do not differ greatly. We 
have accordingly adopted the plan used in previous investigations of making 
two determinations of spectral type. The first is based upon the hydrogen, 
lines alone; the second upon the more general features of the spectra, in 
particular, the intensities of the arc lines of the various elements, and other 
characteristics which seem to be primarily a function of general spectral type. 
MAXIMUM OF 
LIGHT 
MINIMUM 
Hydrogen 
General 
spectrum 
Hydrogen 
General 
spectrum 
F 0 
F 9 
F 8 
F 9 
F 3 
F 7 
F« 
F 8 
SZTauri 
F 0 
F 8 
F 6 
F 8 
F 0 
Go 
F 7 
Gi 
F 0 
F 7 
F 8 
Go 
Ap 
Go 
F 6 
G 0 
F 2 
F 9 
F s 
Go 
F 3 (near max) 
F 9 
F 8 
Gi 
F 2 (near max) 
F* 
F 9 
Go 
Mean 
Fx 
F 9 
F, 
G 0 
Three conclusions may be drawn from these results: First, that the hydro- 
gen lines are abnormally strong in all of these stars, the difference translated 
into spectral type amounting to eight divisions of the Harvard scale at maxi- 
mum of light: Second, that at minimum this difference is reduced, amounting 
in the average to but three divisions: Third, that there is little or no difference 
in the general spectrum at maximum and minimum so far as the criteria here 
used are concerned. Certain differences in other respects will be referred to 
presently. The general conclusion may therefore be drawn that the variation 
in spectral type among the Cepheid variables is mainly a variation in the 
intensity of the hydrogen lines. 
The determination of spectral type of several Cepheid variables by Shapley 5 
and of 5 Cephei by Adams and Shapley 6 , were based wholly on the hydrogen 
lines, the anomalous behavior of these lines in many stars not being fully 
recognized when the earlier work was done. A comparison of the present 
spectral determinations from the hydrogen lines with those of Shapley for 
seven stars common to the two lists shows a close degree of accordance in the 
amount of variation at maximum and minimum of light. The values are seven 
