MUSEUM OF COMPARATIVE ZOOLOGY. 81 



It might be diflBcult to determine whether in this group we have to do 

 with dichotomy, did not the tips of the margins at times reveal the fact 

 that there is no division of the ancestral series, but that a new one is 

 added at the side of an ancestral one (Plate VIII. Fig, 69), where of 

 the marginal individuals 4 is clearly median (ancestral) and 3 is lateral, 

 13 median and 12 lateral, etc. (see page 49). 



The members of the group of Cyclostomata seem to be closely related, 

 and the method of budding is so similar throughout the group that it 

 seems fair to interpret the more compact Tubuliporidie from the Crisiadse. 

 In Crista, as we have seen, individuals are placed in rows, from which at 

 intervals lateral rows are given off to the right or left. One may say 

 that typically these are given off from each individual to both the right 

 and the left, although in some cases, as in Figure 65% lateral branches 

 are typically given off alternately to the right and left, and are often 

 aborted. Perhaps the most general formula of all for Cyclostomes should 

 be that of two lateral branches from each individual, one or both of 

 which may remain undeveloped. Such a formula I believe to be also 

 the typical one for Bugula and its allies, and for the FlustHna and 

 Escharina. It would be written thus : — 



* 

 • b * 



* ♦ (*) • 

 »a* »oaa» ♦ » 



* . » 



(13) -?^ D (*) C (*) B (*) A 



* « ♦ 



*a* *oaa* » * 



* * (*) * 



. b • 



* 



in which the parenthesized asterisks indicate the presence of regenera- 

 tive tissue. This is identical with (12) and similar to (1). 



Braem ('90, pp. 130-133) has already called attention to the differ- 

 ence between Phylactolsemata and Gymnolaemata in the orientation of the 



VOL. XXII. — NO. 1. 6 



