434 
ASTRONOMY: SHAPLEY AND SHAPLEY 
color. That the star with the shortest known perod (3.2 hours) should 
show an abnormal color variation suggests that for some unknown 
reason it may be a real exception. To determine whether or not this 
is the case, observations were undertaken at Mount Wilson with the 
60-inch reflector. Four series of plates were obtained (in all more than 
300 exposures); each of the first two covers a complete period, and as 
Seed 27 plates were alternated with isochromatic plates (used with a 
yellow color filter), each series yields for the study of color phenomena 
complete, simultaneously determined photographic and photovisual Hght 
curves. 
The photovisual curves thus obtained show no great deviation from 
Kron's mean visual curve. The ranges are 0.72 mag. and 0.85 mag., 
as compared with Kron's mean of 0.76 mag. The curves for the two 
nights differ in certain particulars from each other, and as both are smooth 
and well determined, the observed differences are probably real. This 
behef is supported by an investigation of the visual observations pub- 
lished by Kron, which show a much greater diversity in the shape of 
the nightly curves than can possibly be accounted for by errors of 
observation. There are, in general, two types of maximum; one is 
narrow and pointed, the other is broad, round-topped, and somewhat 
lower. The accompanying figure gives examples taken from the photo- 
metric observations by Guthnick at Berlin. 
The Mount Wilson photographic curves differ from each other in the 
same way as the photovisual; for the same maximum the photographic 
and photovisual curves are very similar in form but differ from those 
of the other night. 
A comparison of the Mount Wilson curves yields the following results : 
The photographic maximum occurs simultaneously with the photovis- 
ual within the errors of observation. The magnitudes at minimum 
phase agree for the two nights, being photovisually 12.17 and 12.17, 
and photographically 12.50 and 12.49, respectively. This gives a color 
index at minimum of +0.32, corresponding to spectral type A8 if the 
usual relation between color index and spectrum is adopted for this 
particular case. In the magnitudes at maximum, however, there is no 
such agreement. Photovisually they are 11.45 and 11.32, photographi- 
cally 11.53 and 11.74, so that the color index at maximum varies from 
+ 0.08 to +0.42. 
It is thus seen that so far as our data are decisive the photographic 
range is not constant. On one night it is 0.97 mag., on the other 0.75 
mag. The smoothness and definiteness of the curves makes it im- 
possible to attribute this large difference to observational errors entirely. 
