PEGANTHA SMAEAGDINA. 93 



Development of the parasitic generation of Pegantha smaragdina. 



Plates 22-26. 



Up to the present time the only accounts of the development of a sec- 

 ondary, parasitic generation of medusae within the parent which have 

 appeared are those by Metschnikoff f86 a ) and Stchelkanowzew (: 06) for 

 Cunina proboscidea Metsch., and these two, though agreeing as to the main 

 features, differ from each other in many details. An important difference 

 between Cnnina and Pegantha is the region in which development takes 

 place ; for in the former, according to both authors above cited, it is re- 

 stricted to the mesogloea and endoderm of the subumbrella in the region of 

 the gastric pockets and peronial canals, while in Pegantha I have found 

 it only in the gelatinous substance of the exumbrella, in the region of, and 

 close to the margin of the gastric cavity. It occurs only in the smaller of 

 our two specimens ; i. e. the one in which no gonads are to be seen. 



Although I have not been able to trace the earliest stages, or to find any- 

 thing suggesting fertilization, there is no reason to doubt that the develop- 

 ment is that of a fertilized ovum, as Stchelkanowzew (: 06) has shown to be 

 the case in Cunina proboscidea; for although no discrete gonads are present in 

 the specimen, yet it is possible that the ova in this species as well as in C. 

 proboscidea develop, not in localized masses, but irregularly over the oral 

 surface of the gastric wall, where their small size and resemblance to ordi- 

 nary ectoderm cells would render them difficult to recognize. Amoeboid cells 

 such as Metschnikoff describes are very common in various regions of the 

 jelly of the present species, and I can strongly support the accuracy of his 

 observations that they divide by mitosis (PI. 22, fig. 8), and may unite in 

 pairs (PI. 22 figs. 10, 11). I have found no evidence, however, that such pairs are 

 the two resultants of the previous division of a single parent cell, or that the 

 union is anything more than an accidental one. Stchelkanowzew has sug- 

 gested (: 06, p. 459) that these amoeboid cells (which he did not observe) 

 are in reality the spermatozoa. The facts that they differ enormously 

 in size (cf. PI. 22, fig. / with figs. 2-k), and especially that they show mitotic 

 division, suggest that this explanation is not correct. 



The earliest stage undoubtedly belonging to the developmental series is 

 represented in PI. 22, fig. 12. It contains two nuclei, and study of several 

 specimens has shown beyond question that it consists of two cells, one en- 

 closing the other. The inner cell, the future larva, is spherical, its nucleus 



