306 FROM NEBULA TO NEBULA 



of variables are abnormals, and (2) that abnormal stars 

 are invariably subnormals. By this I mean, that after 

 the collapse of a star, whether by spontaneous eruption 

 or by puncture, it for a time drops to the temperature of 

 stars of lesser mass, and, secondly, that in this collapsed 

 condition it is bound to exhibit changes in its light 

 especially fitful changes. For illustration, compare the 

 similar behavior of a coal or wood fire. Let it, however, 

 not be concluded that only recently collapsed stars are 

 subject to these variations, for our own sun is now in a 

 normal stage for a star of his size, yet his sun-spot cycle 

 is certainly a true variation. Apropos of this subject, 

 let me quote another passage from Doctor Campbell 

 (Stellar Motions, p. 283) : 



A study of the periods of variable stars brings out most 

 curious relations. The periods clearly have preferences for cer- 

 tain lengths. There are a large number whose variations from 

 maximum to minimum and back to maximum are completed in 

 approximately one day, and many whose periods are half a day 

 or less. As the period lengthens from one day, the number of 

 variables decreases rapidly until we reach the few of eleven-day 

 period. There are relatively very few variables with periods be- 

 tween ii days and no days. Variable-star nature seems to abhor 

 this interval. Beginning with no days, the number of variables 

 increases rapidly, with increase of period, up to a maximum at 

 345 days; and the number of variables then decreases rapidly 

 until we reach periods in length approximately 450 days. There 

 are few with periods longer than 450 days, and our information 

 concerning them or their periods is extremely meagre. It is of 

 interest to note that Omicron Ceti, with average period 331 days, 

 is but one of a great number which make up the maximum near 

 345 days. 



Long-period variable stars differ from short-period variables 

 in important particulars. The former vary in brightness through 

 a wide range, usually from three to eight visual magnitudes. 

 Short periods, on the contrary, have small ranges of brightness, 

 varying through 0.2 of a magnitude or less up to a maximum of 

 1^2 magnitudes, with few exceptions. The long-period variables 

 are all reddish in color, apparently indicating that the atmos- 

 pheres of these stars are dense, absorbing the violet rays and 

 transmitting the waves of longer length, or that we are dealing 

 with low-temperature radiations. Short-period variables, on the 

 contrary, are all yellow or white in color. Chandler has found 

 that there exists a relation between the length of period and red- 



