EVOLUTION OF THE STELLAR SYSTEMS. LUBY 
we suppose either mass to contract still further, it is evident that the rotation will begin 
to exceed the orbital motion ; and the tides raised in either mass by the attraction of the 
other will lag, and tidal friction will henceforth play just the part we have already 
described. 
Starting from a different point of view, Darwin was already at work on essentially 
the same problem when Poincaré’s paper appeared, and he held his results back for 
nearly a year longer, hoping to make application of the principle Poincaré had 
announced. In this second method of treatment two masses of homogeneous fluid were 
brought so close together that the tidal distortions of their figures caused them to coalesce 
into one mass; set in motion asa rigid system, the problem was to find the resulting 
figure of equilibrium. It turned out to be a dumb-bell with equal or unequal bulbs 
according to the relations of the primitive masses. Thus we see it proved from two 
The Apoid of Poincaré, showing how a rotating mass of 
fluid separates into two unequal parts. 
independent points of view that a division such as I assumed in 1888 can theoretically 
take place ; and among actual nebule of space such division seems to be a general law. 
During the years of 1896 and 1897, I have examined a number of such objects in the 
southern hemisphere, and find them substantially as drawn by Herschel many years ago. 
Burnham and Barnard had previously assured me that the interpretation of the figures 
of double nebulse based on the drawings of Herschel was in accord with the phenomena 
of nature, but the studies more recently made with the great Lowell telescope supple- 
ments their large experience in a very happy manner, and may be said to remove the 
last doubt that could attach to the division of nebule by the process of fission. 
Before concluding these remarks it ought to be pointed out that in space we have 
to deal with masses which are not homogeneous, nor are the nebule by any means 
incompressible ; yet many considerations lead us to believe that in most cases the density of 
