008 W. J. SOLLAS. 



come to form a part. Thus the gastrula^ would appear to 

 result from a remarkable rearrangement of the cells of the 

 morula^ the blastula stage being rapidly slurred over in order 

 to economise space. 



The commencing gastrula is seldom so normal in character 

 as represented in fig. 19, and never retains the simple form 

 there exhibited for long ; for as the advancing hypoblast grows 

 inwards to meet the epiblast, both become deeply folded (figs. 

 24 and 26), so that by the time both layers have become 

 approximated to form a single bilamellar wall the embryo has 

 assumed the form of a complicately folded sac (figs. 27, 28, 32). 

 Hence not only the blastula but the gastrula stage is slurred 

 over, for the embryo does not wait till a gastrula form is com- 

 pleted before commencing to fold; but the folds arise pari 

 passu with the growth of the gastrula, which thus as a definite 

 stage in the development can scarcely be said to occur at all. 

 The embryo which so results may be regarded as a young 

 sponge, complete in most essential characters, save as regards 

 the presence of pores, which are not in my sections to be 

 observed. The outer sinuses of the folds may be regarded as 

 rudimentary incurrent canals, and the inner sinuses as rudi- 

 mentary flagellated chambers, with their ducts as yet undif- 

 ferentiated, as they permanently remain in Halisarca Du- 

 jadini. The rest of the archenteric cavity probably gives rise 

 to the excurreut canal system. Difficult as the history of the 

 young sponge is, we have yet, however, by no means exhausted 

 its complications. 



Frequently, indeed generally, the young gastrula exhibits a 

 considerable number of infoldings of approximately equal 

 strength (fig. 18), so that at this stage, in the absence of any 

 diff'erentiation of the component cells (and none exists), it is 

 impossible to determine which of the folds will give rise to the 

 hypoblast and which will remain as epiblast. Since, however, 

 a folded sac with a bilamellar wall is invariably the final term 

 of the series of changes observed, one must conclude, in order 

 to explain its formation, that one of the several commencing 

 folds grows faster than the rest, and at length meeting them 



