644 



SCIENCE 



[N. S. Vol. XXXVII. No. 956 



star does not belong to the group. The proper 

 motions of most of these stars have been measured 

 recently by Van Maanen and found to be ex- 

 tremely small and their parallaxes are below the 

 limits of error of measurement according to Kap- 

 teyn's results. 



A comparison of three photographs of the spec- 

 trum of Nova Geminorum II. taken in August, 

 September and November leads to the following 

 conclusions: (1) The principal nebular bands are 

 slightly more intense on the last photograph, while 

 the hydrogen and helium lines remain very nearly 

 constant. (2) A very marked change occurs in 

 the line X 4687 of the principal series of hydrogen. 

 On the August photograph it is very faiut, while 

 in November it is fully half as strong as the 

 intensely bright band X 4640. This line showed 

 evidence of rapid variation on earlier photographs 

 as well, gaining greatly in intensity between May 

 5 and May 10. Its behavior should prove of great 

 value as bearing on the physical condition of the 

 star. (3) Two other bands, one at X 4522 and 

 the other at X 4605 are considerably stronger on 

 the last photograph. (4) The widths of the bright 

 bands have remained remarkably constant through- 

 out the history of the Nova. Measures on the 

 hydrogen, helium and nebular bands show no ap- 

 ■ preciable change from the photographs of April 

 and May. The positions of the centers of the 

 bands also remain as on earlier photographs, being 

 displaced from one to two Angstroms toward 

 longer wave-lengths. (5) All of the more prom- 

 inent bright bands except those at X 4522, X 4605 

 and X 4640 have broad faint absorption bands 

 nearly symmetrically placed upon them. In sev- 

 eral cases the dark bands contain one or more 

 narrow absorption lines. A remarkable case of 

 this sort is the line at X 4337.5 which has been 

 measured upon all of the photographs taken since 

 March. The bright bands are terminated on either 

 side by bright maxima, the violet member of which 

 is the strongest on the plates of August and No- 

 vember. (6) In addition to the bright bands the 

 spectrum of the star almost certainly shows an 

 extremely faint continuous spectrum probably 

 crossed by dark lines. The M line of calcium is 

 seen as a dark line on the November photograph 

 and yields a value of the radial velocity of about 

 + 5 km. 



From these considerations it is evident that the 

 spectrum of this Nova and probably of other 

 NovEe as well is by no means so simple in its later 

 history as has sometimes been supposed. The 



great width of the emission bands, the presence of 

 well-defined selective absorption within them, the 

 persistence of the displacement of their centers 

 toward longer wave-lengths, as well as the marked 

 variation in intensity of some of the important 

 bands such as X 4687, all go to show that the 

 physical conditions present are the most complex, 

 and must differ greatly from such as produce an 

 ordinary nebular spectrum. 



Sate of LigM Changes in Various Celestial Ob- 

 jects: S. I. Bailey. 



A discussion of the variable stars in Messier 3 

 shows some examples of extraordinary rates of 

 increase in light. This globular cluster is a faint 

 hazy star of about the sixth magnitude to the 

 naked eye. Its marvelous character is not sug- 

 gested even in a small telescope. At Arequipa 

 with an exposure of 100 minutes with the 13-inch 

 Boyden Refractor about 1,200 stars were shown. 

 These plates showed stars somewhat fainter than 

 the 16th magnitude. Among these were found 137 

 variables among 900 stars actually examined for 

 variability, or one in seven. The total number of 

 stars in the cluster is very great. On a plate made 

 by Ritchey on Mt. Wilson with the 60-inch re- 

 flector giving an exposure of four hours not less 

 than 30,000 stars are seen, if we include the cen- 

 tral mass where an exact count is impossible. 

 Among the variables found in this cluster the 

 maximum rate of increase in four cases appears 

 to be more than six magnitudes an hour. The 

 mean maximum rate of increase of all the vari- 

 ables is about two and a half magnitudes an hour. 

 It is doubtful whether any other celestial object 

 has so great a known rate of variation as six 

 magnitudes an hour, although it seems probable 

 that this rate may be exceeded in the case of 

 NovEe. Except for Novas great rates of change 

 appear to be rare, x Cygni has a range of nearly 

 ten magnitudes, but this enormous change takes 

 place during so long a tune that the rate per hour 

 is only a fraction of a magnitude. Eros has per- 

 haps the shortest period of a known object. From 

 one maximum to the following is only about two 

 and a half hours. The range may be a magnitude 

 or more at times, and the light curve closely 

 resembles a sine curve. Its maximum rate of 

 change is probably never more than two or three 

 magnitudes per hour. Some of the Algol variables 

 change very rapidly. TJ Cephei and W Delphini 

 are good examples. The rate per hour, however, 

 of any known Algol star does not exceed two or 

 three magnitudes per hour. 



