188 I. IJIMA : HEXACTINELLIDA. I. 



To repeat, to me it seems certain that the embryo in a 

 very early stage of its development consists of a small assemblage 

 of uniform-looking cells, which differ in no distinguishable 

 feature from the archœocytes. If the resulting body were some- 

 thing comparable to a bud or a gemmula, I would probably have 

 felt no hesitation in concluding that the cells were really 

 archgeocvtes, and that we had here to do with a case of asexual 

 reproduction. But, free-swimming larvye, essentially similar to 

 those developed from ova in other sponges, being at issue, the 

 question whether true ova are not somehow complicated in the 

 cell-mass whence the larva arises, seems to claim to be brought 

 on the tapis, all the more, since our knowledge of the Hexacti- 

 nellidan ovum is far from being satisfactory. 



H. V. Wilson ('94) has made endeavors to show that in 

 certain Monaxonid species {Esperella fihrexilis, Tedania brucei), 

 larvae similar in all fundamental respects to those which develop 

 from eggs in other species, arise, exactly like gemmules, in an 

 asexual way. The ' gemmule larva,' as distinguished in genetical 

 relation from the ' egg larva,' develops from a simple cellular 

 mass, which originates by the multiplication and coming together 

 of certain ' mesoderm ' cells. I conceive the mode of origin and 

 growth of the archœocyte-congeries in the Hexactinellida to be 

 just the same, and it seems to me not impossible that in the 

 Hexactinellid larvie which I have seen, we have simply a new 

 case of the * gemmule larva ' or bud embryo. 



Serious doubts have been thrown on the accuracy of Wil- 

 son's observations concerning the ' gemmule larva ' by Maas 

 ('96) and MiNCHiN ('97). Especially the former writer lias 

 suggested that the ' gemmule larva ' may have arisen from an 

 egg just like any ordinary larva and that Wilson's idea of its 



