612 



SCIENCE 



[N. S. Vol. XXV. No. 642 



and the form of the light curve at that 

 point is not determined. 



C. P. D.— 63° 2,485 was at first supposed 

 to be of the Algol type, but has been found 

 instead to be a short-period variable of 

 unusual interest. As has already been an- 

 nounced in a recent circular, the light 

 curve resembles that of an Algol variable 

 with a minimum covering about half the 

 period, but the light appears to be chan- 

 ging continuously, though very slightly, 

 even when near maximum. As the range 

 is only four tenths of a magnitude, four 

 independent observations were made on 

 each plate. By taking means, the acci- 

 dental errors were reduced one half. The 

 mean magnitudes were then arranged in 

 the order of phase, and the mean phase 

 and magnitude was taken for each succes- 

 sive group of five plates. The average 

 deviation from the light curve of the points 

 thus obtained is only ± 0.02 magn., and 

 may be compared with the deviations 

 found with the best photometric measures. 



The large proportion of Algol variables 

 among those discovered in this region is 

 interesting. Of thirty-six variables an- 

 nounced last month, four have already 

 been shown to be of the Algol type, and 

 others, of which the observations have not 

 yet been discussed, are supposed to belong 

 to the same class. On the other hand, of 

 more than seventy variables in Scorpius 

 and Ophiuchus, announced two years ago, 

 few, if any, appear to vary in this manner. 

 The same is true of the Magellanic Clouds. 

 The present study of the distribution of 

 variable stars is in a very early stage and 

 much generalization from the results as yet 

 obtained is rash. Yet it has already be- 

 come evident that certain kinds of vari- 

 ables are apt to be found in groups. It 

 is desirable that the number of persons en- 

 gaged in this research should be increased, 

 so that the systematic survey of the 

 heavens may be completed with a reason- 



able degree of thoroughness during the 

 next few years. Evidently it is of the 

 highest importance to ascertain the types 

 to which new variables belong, even if their 

 number is too large to permit the computa- 

 tion of all the periods. 



A Peculiar Binary System: Ekic Doo- 



LITTLE. 



Latitude Terms of Long Period, from the 

 Flower Observations: C. L. Doolittle. 

 This series embraces a period of seven 

 years and nearly two months, and is prac- 

 tically homogeneous in all respects. As a 

 term having a period of about six years 

 has been supposed to be indicated by sim- 

 ilar series elsewhere, it seems desirable to 

 examine my results for evidence on this 

 point. 



Without making any assumption as to 

 the law of latitude variation, the intervals 

 from minimum to minimum were found 

 graphically. The maxima were not em- 

 ployed for this purpose, as they were less 

 clearly indicated. Six periods were found 

 as follows: 



1st 459 days 



2d 410 



3d 470 



4th 439 



5th 446 



6th 416 



Mean 440 days 



5 periods = 2,200 days 

 6 Julian years = 2,191.5 days 



The difference 8.5 days is of no importance 

 for present purposes. 



This close agreement seems to fall in line 

 with the possible existence of a term having 

 a period of about six years. 



The method of procedure was as follows : 

 For each interval from minimum to mini- 

 mum, terms of the form x sin 6 -{-y cos 

 were removed from the given values of the 

 latitude. A series of 103 equations was 

 then obtained of the following form: 



