Maech 23, 1906.] 



SCIENCE. 



443 



Attention is called to the importance of 

 this eclipse in advancing our knowledge of 

 the corona. Of eclipses in the immediate 

 future, although a total one occurs every 

 year from 1907 to 1912, that of 1907 is 

 likely to be the most important because the 

 paths of its successors are for the most 

 part inaccessible. The best region for the 

 1907 eclipse is available by means of rail- 

 ways recently constructed in Russian terri- 

 tory. Also, about 600 miles northwest of 

 Peking, the eclipse track is rather difficult 

 of access in Mongolia. The writer has 

 computed, among other things, the positions 

 of ten possible stations, exact local times 

 of the four phases, durations of totality, 

 position angle of first contact, and sun's 

 altitude and azimuth at middle of eclipse. 

 (To be published in The American Journal 

 of Science.) 



Some Variable Star Problems: S. I. Bailey. 



About ten years ago more than five hun- 

 dred variables were found in globular 

 clusters; and further discoveries of this 

 kind are possible with instruments of large 

 size among the faintest stars in the clusters 

 already examined, and in clusters which 

 have not yet been searched. 



The variables in the clusters Omega Cen- 

 tauri, Messier 3 and Messier 5 have now 

 been studied. There are in these clusters 

 348 variables ; for 273 of these periods have 

 been determined, averaging about half a 

 day. Changes in brightness are very 

 rapid, and in many cases the regularity 

 of the periods is so great that these stars 

 would serve as good celestial time-keepers. 



We have observed the clusters from five 

 to twelve years, obtaining from five to ten 

 thousand returns of maxima, so that 

 periods are obtained which are correct 

 within a few tenths of a second. For 

 seventy-five of the variables periods have 

 not been determined. Of these, sixty-five 

 are too difficult, on account of nearness of 



other stars, smallness of the range of varia- 

 tion, or both. The remaining ten variables 

 present peculiarities not yet understood. 



While our observations for most of these 

 variables are satisfied by a simple, uniform 

 period, there are a number of cases in 

 which it has been necessary to introduce 

 a second term, implying a secular variation 

 in the period. Several have been found 

 which are best explained by the supposi- 

 tion that they are doubles, both components 

 varying with the same period but with 

 alternate maxima. 



The presence of large numbers of vari- 

 ables in some clusters, and few, or none, in 

 others, calls for an explanation. It may 

 be that the axes of revolution of all mem- 

 bers of such stellar systems are parallel. 

 Variability, if associated with revolution, 

 might then depend on the direction from 

 which the system is seen. Stars which 

 vary for us may be unvarying when seen 

 from some other point in space. Again, 

 variability doubtless marks an epoch in the 

 development of stars. This epoch may be 

 in the past for some, present for others, 

 and future for still others. 



The typical form of light-curve for vari- 

 ables in clusters is difficult to explain on 

 the theory of axial revolution. But Dr. 

 Curtis has shown recently that a relation 

 exists between motion in the line of sight 

 and variations in brightness in the ease of 

 W Sagittarii. Much may be learned by an 

 extension of this research to other variable 

 stars. 



Maxima and Minima of Variable Stars of 

 Long Period: Annie J. Cannon. 

 The original form of the bibliography of 

 variable stars undertaken several years ago 

 at Harvard College Observatory and for 

 which all the material has been collected, 

 was recently modified owing to duplication 

 of much of the work by the Gesellschaft 

 Committee in their proposed complete 



