708 
The Revised Draper Catalogue: ANNIE J. CANNON. 
Owing to a general desire among astronomers 
for the class of spectrum of many more stars than 
have yet been studied, work has been begun upon 
a new catalogue of stellar spectra, to be called the 
Revised Draper Catalogue. The whole sky will be 
covered by photographs taken with a prism placed 
in front of the object glass of the 8-inch Draper 
and Bache telescopes, the exposure being generally 
one hour. It is believed that all stars of the 
eighth magnitude will be included, as well as many 
fainter ones. All the classification will be made 
by the writer, using the notation described in the 
Harvard Annals, volume 56, pages 66 to 69. 
This work can be done rapidly, since previous 
study and classification of more than five thousand 
spectra taken with the various Harvard telescopes 
have made each division and subdivision a definite 
picture in the mind. Three assistants are working 
daily upon the laborious identifications, the reduc- 
tions and the clerical part of the catalogue. Seven 
thousand spectra have already been classified, and 
it is estimated that if we have equally good photo- 
graphs for the whole sky, the catalogue will con- 
tain one hundred thousand stars. It is proposed 
to print it in sections in the order of right ascen- 
sion, of which the portion from 0° to 6" will form 
the first volume. 
Notes on the Determination of the Elements of 
Algol Variables: H. N. RUSSELL. 
Further study of this problem on which a report 
was made to the society in August, 1910, shows 
that, when there is no constant period at minimum 
(i. e., when the eclipse is partial), it is possible to ~ 
represent the observed light curve within the error 
of ordinary observations by arbitrarily choosing 
any value within certain limits for the ratio of the 
radii of the two stars, and then determining the 
other elements in a suitable manner. 
The various sets of elements, however, give dif- 
ferent depths for the secondary minimum; and if 
this has been observed the problem becomes de- 
terminate, unless the primary and secondary min- 
ima are of nearly equal depth. In the latter case 
additional data (which can sometimes be supplied 
by spectrographic observation) are necessary if 
the elements are to be definitely determined. 
Tables have been prepared which facilitate the 
numerical solution of these problems. With slight 
modifications, these may also be used in the case 
of variables of the Beta Lyre type, in which the 
two stars are very close, and are distorted into 
prolate ellipsoids by their mutual attraction. 
SCIENCE 
[N.S. Vou. XXXV. No. 905 
The Eclipsing Variables W Crucis and W Urse 
Majoris: H. N. RUSSELL. 
Good light-curves of these stars have been de- 
termined, the first at Harvard by Miss Leavitt, the 
second at Potsdam by Muller and Kempf and by 
Baldwin. Both are of the B-Lyre type, and in 
each case the observations can be very satisfac- 
torily represented on the eclipse theory. 
W Crucis has a period of 198.5 days and a 
range from 8™.9 to 9™.5, with a secondary maxi- 
mum of 9™.2. The system consists of two stars, 
one twice as bright and two and a half times as 
large as the other. The ratio of the longer and 
shorter axis of the ellipsoidal stars is 7:6. The 
relative orbit has an eccentricity of 0.04, and at 
periastron the surfaces of the two stars are sepa- 
rated by a distance slightly exceeding the diameter 
of the smaller. At principal maximum the smaller 
star is totally eclipsed by the larger. Increase and 
decrease of light lasting 14 days and totality 15 
days. 
The density of the larger star can not exceed 
1/160 that of the smaller star, 1/10 that of the 
air under ordinary conditions. The spectrum is 
G pec. with bright lines. 
This system is evidently in a very early stage 
of development, and if comparable with the sun 
in mass and surface brightness, must be at a dis- 
tance of many thousand light-years. 
The writer is greatly indebted to Professor 
Pickering and Miss Leavitt for unpublished ob- 
servations of this star. 
W Urse Majoris varies from 7™.9 to 8™.5 in a 
period of 8 hours, within which there are two 
equal and equidistant maxima and minima. Its 
variations may be very satisfactorily represented 
on the assumption that the system consists of two 
stars, equal in size and brightness, with longer 
axes 4/3 of the shorter axes, revolving in a circular 
orbit, and separated by only 2/5 of their longer 
diameters. At maximum one star obscures half 
the disk of the other, the eclipse lasting one and a 
half hour. 
The light-curve may be equally well represented 
with any ratio of the two radii between unity and 
0.57—the surface brightness of the two stars being 
equal. The mean density of the system is 2.4 
times the sun. This star is also of Spectrum G, 
but must represent a very different stage of evolu- 
tion from the other. 
The Solar Rotation: J. S. PLASKETT. 
This paper gives an account of further work on 
the spectroscopic determination of the solar rota- 
