198-200] 



Summary 



201 



and this relation may be represented by a curve such as RO described in the 

 direction of T^ decreasing. In the intermediate stage in which the gas does 

 not obey the gas laws, but is still highly compressible, the law relating E and 

 T t cannot be precisely specified. Clearly, however, the curve by which it is 

 represented in fig. 40 must depart asymptotically from PQ and approach 

 asymptotically to RO. Thus the actual sequence of changes in E and r l\ will 



Fig. 40. 



be represented by a curve such as PLMNO in fig. 40. The mass of gas starts 

 at a low temperature, increases to a maximum temperature and cools again. 

 Meanwhile the emission of energy will remain constant until approximately 

 the stage at which the maximum temperature is attained, after which it falls 

 steadily and rapidly to zero. 



200. It will at once be seen that this theoretical result describes exactly 

 Russell's theory of the order of stellar evolution, of which a brief account was 

 given in 13. Russell, while pointing out that his theory was in accordance 

 with theoretical principles, based the evidence for it mainly upon a diagram 

 of observed absolute magnitudes of stars*. In this diagram the spectral 

 class, giving a rough measure of the temperature, was taken as abscissa while 

 the ordinate measured the absolute magnitude. The stars in the redder 

 spectral classes (M, K, G) were found to fall into two detached, or nearly 

 detached, groups. In an upper group the absolute magnitude was approxi- 

 mately independent of spectral type ; in the lower group it varied rapidly 

 with spectral type, falling off towards the red end of the scale. These two 

 groups of stars form what Russell calls "giant" and "dwarf" stars respectively. 



Clearly Russell's diagram provides powerful confirmation of our theoretical 

 diagram shewn in fig. 40, the stars along the branch PL being giant stars, and 

 those along the branch ON being dwarfs. Further confirmation has recently 

 been afforded by the investigations of Adams and Joy f already referred to in 

 13. Here 500 stars are considered and these are again found to fall into 



* Nature, 93, p. 242 (May 7, 1914), and Popular Astron. 22 (1914), p. 11. 

 t Astrophyt. Journ. 46 (1917), p. 334. 



