314 
The Evolution of Binary Systems [oh. xi 
greater than this. This gives a rough classification according to the actual 
dimensions of the system. The last column gives the theoretical distribution 
s 2 /Si to be expected for 45 systems for which IJl 1 has the uniform value (H)9. 
TABLE XXV. Separations in Triple Systems (Russell). 
Observed ratio of 
separations (s 2 /»i) 
Number of systems 
Class I 
Class II 
Theory 
(f 2 /h = 0-09) 
>0-40 
1 
5 
1 
0-40 to 0-30 
2 
— 
2 
0-30 to 0-20 
3 
— 
3 
0’20 to 0"15 
5 
— 
41 
0-15 to 0*10 
8 
3 
9| 
0-10 to 0-05 
m 
2 
16 
0-05 to 0-025 
14$ 
2 
8 
<0-025 
13 
7 
1 
The table shews that the systems in Class II do not conform at all to the 
theoretical law of distribution. Down to a separation ratio of about 005, the 
systems in Class I conform closely to the distribution to be expected in 
systems with an actual separation ratio 0'09 oriented at random in space. 
Below an apparent separation ratio of 0’05 there are too many systems to be 
accounted for in this way, but the excess could be attributed to a further 
group of systems having a separation ratio of less than 0‘09—somewhere 
about 0 04. 
Apart from these precise figures, it is clear that, so far as Class I is con 
cerned, the law of distribution of s 2 /sj is just such as might statistically be 
expected for a number of systems in which Z 2 /^i had values ranging from 
about 0'09 downwards. This distribution fully conforms to the theoretical 
requirements for systems generated by successive processes of fission. 
The systems in Class II fall into two sharply defined groups. A group of 
14 for which s 2 /s 1 is less than 015 may very possibly be interpreted as systems 
with a separation ratio ranging from about 0 06 downwards arranged at 
random in space, and so may possibly have been formed by fission; but a 
group of 5 with a separation ratio greater than 040, cannot possibly be so ex 
plained. Russell, following an earlier suggestion of Moulton’s, supposes that 
these may have evolved out of separate condensations in the nebula from 
which the stars were originally formed. 
283. Russell interprets his investigation as shewing that the majority of 
binary systems whose separation is less than 1000 years’ proper motion, have 
been evolved by fission. The investigation shews that they may have been 
so evolved, but not that they must; it shews that a group of systems which