MUSEUM OF COMPARATIVE ZOOLOGY. 51 



There is something of a tendency for lateral hranches to be given off 

 in the same generation from closely related branches. Thus (Fig. 65") 

 from the primary individual, o, of the stock, two individuals, a median 

 and a lateral one, arise. Each gives rise in its first generation to two 

 individuals, a median and a lateral. Of these four individuals each 

 [^ives rise at the end of three generations of median buds to two buds, 

 a median and a lateral. Comparing 2 and 10, first descendants of the 

 two branches arising from the second individual of 8, we find that each 

 gives rise to lateral branches fi-om their first individual and from their 

 fourth. Comparing 14 and 19, first descendants of the two branches 

 arising from the first individual of 8, we find each giving rise to lateral 

 branches from their first individuals. The law breaks down, however, 

 when an attempt is made to carry it to extremes. 



The fourth rule is not always so pronounced in Crisia eburnea as 

 elsewhere, although lateral budding seems to be slightly more frequent 

 at the mai'gin. 



The extreme marginal branches usually attain far fewer generations 

 than the more intermediate ones ; thus, in Figure Go", branch 20 ends 

 in the 7th generation and branch 13 in the 7th also, while the more 

 intermediate branches 15 and 18 attain 12 and 14 generations respect- 

 ively. So, too, while the outer branches 6 and 1 contain respectively 

 10 and 11 generations, the inner branches reach 12 and 14. 



It is very noticeable that the outer branches give rise to more indi- 

 viduals than the intermediate ones. Figure 65" will serve to illustrate 

 this also. Here the outer branch 4, the intermediate 8, and the outer 

 15 possess, together with the branches arising from them, 33, 28, and 40 

 individuals respectively. Farmer ('91, p. 168) finds this true for his 

 Crisia ramosa, for he says, " It is frequently remarked that the longest 

 and most branched parts of the colony are lateral branches, and not 

 parts of the main stems." 



There is, in the long run, a deci'ement in the rate of increase of indi- 

 viduals in successively older generations, yet it is not so regular a one 

 as that which we found to exist in Bugula. Thus, in the seven gen- 

 erations which even the shortest branches shown in Figure 65* had 

 attained, the average increase of the number of individuals in the 

 second, third, and fourth generations over the number in the preceding 

 is 67% ; in the fifth, sixth, and seventh, 44%. The generations beyond 

 the seventh are not complete ; they would have contained more indi- 

 viduals at a later period, when the branches which have now attained 

 only seven generations had grown. Thus the number of individuals 



