PEGANTHA SMARAGDINA. 95 



of the larva through the endoderm; at least it is absent in larvae which 

 have penetrated into the gastric cavity. 



The endoderm originates, as both Stchelkanowzew and Metschnikoff agree, 

 by secondary delamination, there being no trace of any invagination. Even 

 in such an early stage as that represented in PI. 23, fig. 7, the four or five 

 central cells are already sharply defined from the surrounding layer, and as 

 development progresses the distinction is made more and more evident both 

 by the differentiation of the endoderm cells, and by the development of a gelat- 

 inous layer between the two germ layers. The process in all respects agrees 

 with the descriptions of Metschnikoff, the resemblance between his figures 

 and corresponding stages in Pegantha being very striking. 



Later development '. — After entering the gastric cavity of the parent, the 

 larva grows rapidly in size and becomes spindle shaped. The ectoderm cells 

 multiply proportionately with the growth ; but the endoderm cells multiply 

 very slowly, but increase in bulk (PI. 23, figs. 7-11), and gradually assume a 

 chordate outline until they resemble the cells of the endodermic core of an 

 adult tentacle. Indeed larvae at this stage very much resemble small 

 detached tentacles. The earliest stage in which I have found a gastric 

 cavity is represented in PL 24, fig. 2. I have been unable to find stages 

 in the formation of the cavity. At a slightly more advanced stage the 

 mouth, already foreshadowed by the thinness of the wall in PI. 24, fig. 2, breaks 

 through. At about this same time the development of the primary tentacle 

 commences. Apparently the main portion of the larva goes to form the 

 tentacle, as is the case with the two primary tentacles in Solmundella 

 (Metschnikoff, '86 a ). A second tentacle is developed opposite the first; 

 other tentacles then appear irregularly, specimens having been observed 

 with three, four, and five. 



Budding. — Larvae with four or more tentacles often show stages in bud- 

 ding very similar to those described in detail by Metschnikoff ('7-1) for 

 Ctinina prolifera. He likewise seems to have observed one or two instances 

 of this mode of multiplication iti Gunina proboscidea ('8G n , taf. 12, figs, h, /■>'), 

 although Stchelkanowzew, not having found any himself, doubts their occur- 

 rence. In Pegantha smaragdina budding larvae are very numerous in the pen i 

 nial and ring canals of the parent-host (PI. 19, fig. 5). Budding takes place 

 at the aboral pole, and comprises both germ layers, as well as a portion of 

 the gastric cavity (PI. 24, figs. S, 5). Surface views of a number of stages 

 are represented on PI. 25. The bud may become constricted and Beparate 



