March 1/, 1905.] 



SCIENCE. 



Alo 



superposed. In the spectrum of 27 Canis 

 MajoriSj bright hydrogen was present in 

 March, 1890, April, 1895 and October, 

 1897. Numerous photographs of both 

 these spectra on other dates showed all the 

 lines to be wholly dark. 



It is evident that a large field of investi- 

 gation lies open to the spectroseopist among 

 these bright-line stars. 



Variable Stars in Large Nebulous Regions: 



Henrietta S. Leavitt. 



Since last March a special study of the 

 distribution of groups of variable stars has 

 been in progress at the Harvard College 

 Observatory. As one result of this investi- 

 gation, four hundred and fourteen new 

 variables have been discovered and an- 

 nounced. Seventy-three of these are in 

 Orion, one hundred and fifty-two in the 

 Large Magellanic Cloud, fifty-seven in the 

 Small Magellanic Cloud, one hundred and 

 five in Scorpius, ten in Carina and seven- 

 teen in Sagittarius. The results of this 

 study up to the present time may be sum- 

 marized as follows : 



First, as regards distribution, it has be- 

 come evident that gtoups of variable stars 

 are strongly localized. Of the ninety-nine 

 confirmed variables at present known in the 

 constellation of Orion, south of the equator, 

 eighty-nine are within the limits of Bond's 

 map of the region surrounding the Nebula 

 of Orion, and of these all but four are 

 found in less than half this area. The en- 

 tire region thus finally limited is nebulous. 

 The large number of variables discovered 

 in the two Magellanic Clouds is in marked 

 conti'ast with the small nvmiber found in 

 the surrounding regions. The neighbor- 

 hood of the Trifid Nebula in Sagittarius is 

 noticeably poor in variable stars, and so 

 also is the neighborhood of the nebula 

 about 77 (Jaririw. Yet these are two of the 

 most densely crowded regions of the Milky 

 Way. In Scorpius, after subtracting 



thirty-three variables which were found in 

 the cluster Messier 4, there are still left 

 more than four times as many variables as 

 were found in an area in Sagittarius ap- 

 proximately equal in extent and far richer 

 in stars. 



Secondly, a certain order of brightness 

 appears, on the whole, to prevail among the 

 variable stars of each group, those in the 

 central condensation of the Large JMagel- 

 lanic Cloud being the faintest, and those 

 in Scorpius the brightest. 



Finally, it is probable that different 

 types of variability prevail in different re- 

 gions. In both of the Magellanic Clouds, 

 a large proportion of the variables appear 

 to have very short periods, while in 

 Scorpius the reverse may prove to be the 

 case. Many of the variables in the Nebula 

 of Orion remain faint during the greater 

 part of the time, but occasionally show a 

 striking increase of brightness. Whether 

 these flashes of brilliancy occur regularly 

 is not yet known. 



The researches here described supple- 

 ment the remarkable discovery by Professor 

 Bailey, of large numbers of variable stars 

 in clusters. They are similar to these car- 

 ried on by Professor Wolf, of Heidelberg, 

 who has amnounced lists of new variables 

 in Orion, Aquila and Vulpecula. Evi- 

 dently a further study of the distribution 

 of groups of variable stars will be intensely 

 interesting in its bearing iipon the problems 

 of stellar evolution. 



PUt netarij Spectrograms : Percival Lowell. 



These spectrograms were made by Mr. 

 V. M. Slipher and the lantern slides of 

 them by Mr. C. 0. Lampland, both of the 

 Low(41 Observatory staff, and were pre- 

 sented by the director, Professor Lowell. 



1. Solar spectrum, photographed Novem- 

 ber 30, 1903—59 dark lines can easily be 

 counted between G and Hy. 



2. Spectrum of £ Prgasi, plio*ographed 



