ASEXUAL REPRODUCTION IN SAGARTIA 181 
(pp. 176, 177), that the plane of fission is more or less nearly per- 
pendicular to the directive plane. That directive endocoels are 
avoided, gives additional emphasis to the tendency to divide in 
complete non-directive endocoels. 
The inequality of the parts resulting from a division is strik- 
ing. The relative size of the parts may be roughly measured 
by the number of complete mesenteries. Of fourteen divisions 
of diglyphic specimens into two parts each, no case of equal 
distribution of the complete mesenteries between the two parts 
is found, whereas cases of extreme inequality do occur. The 
other specimens show in general a similar lack of equality of 
parts. One might explain the inequality in the products of 
division of a regular hexameric individual on the ground that the 
tendency to divide in complete endocoels and the tendency to 
divide in a plane perpendicular to the major axis of the mouth 
overcome any tendency to divide into equal parts. But this does 
not hold for the unequal division of a regular octameric specimen 
such as no. 9 (table 3). In this specimen the tendencies for the 
fission plane to pass through complete endocoels and to pass per- 
pendicular to the major mouth axis would both be satisfied by a 
division which should produce equal parts. Yet this division 
was very unequal. In a number of other cases the inequality of 
parts is much greater than is demanded by the tendencies to 
divide in complete endocoels and along the favored plane. It 
should be noticed that, in every case where a recent regeneration 
zone is indicated in the old part (tables 3 to 6), the piece con- 
taining this regenerating part possesses the greater number of 
complete mesenteries. It is possible that there is a tendency 
toward an equality in which different regions have values 
dependent upon age. The tendencies of the planes of fission so 
far as analyzed are summarized on pages 226, 227. It remains 
to be pointed out that, even among specimens having the same 
number of siphonoglyphs and complete mesenteries, the im- 
mediate products of division are quite diverse in form. The 
same processes, dealing with specimens already varying as 
shown in tables 3 to 7, lay a foundation for the most extreme 
numbers of mesenteries and siphonoglyphs found in the fully 
