DEVELOPMENT OF THE GAPE SPECIES OF PEEIPATUS. 529 



to imply that the endoderm cells are not mainly nutritive, but 

 that their main function is to cause the currents through the 

 sponge body, and that the food brought by these currents 

 passes into the parenchyma, through the walls of the passages^ 

 to be digested by the so-called mesoderm cells. 



The collared cells are thus inconstant, and appear to be 

 merely parenchyma cells specially modified under certain 

 conditions and capable of passing back into their original form 

 when the need for them has passed away. When they vanish 

 the canal system also goes and the sponge becomes solid so far 

 as the latter is concerned. Inasmuch as the parenchyma cells, 

 and probably also the cells of the ectoderm, are all connected 

 by their processes (except in the cases in which they break 

 away and become amoeboid), it is clear that the sponge in this 

 condition and in the case of Shulze's larva already referred to, 

 is a syncytium, and but little more than a multinucleated 

 Protozoon. It differs from such a Protozoon simply in the 

 greater development of the vacuoles (spaces between the 

 cells) of the central portions, and in the presence of a distinct 

 cortical layer of nuclei. 



In some cases this assumption by a sponge of the Protozoon 

 form is much more marked than in the above cases. I refer 

 to the case of the well-known form Haliphysema, described 

 by Haeckel (No. 10) as a sponge with an axial ciliated 

 chamber traversing it, and by Lankester (No. 18) as a mul- 

 tinucleated Rhizopod. It is difficult to believe that either 

 of these distinguished naturalists can have made the mistake 

 implied by these contradictory observations ; and the only way 

 of reconciling the latter that I know of, is to be found in the 

 above view, viz. that Haliphysema, like Spongilla and 

 Halisarca possesses under certain conditions, the power of 

 becoming solid : that in certain conditions in which it was found 

 by Heeckel, it approximates to the sponge, while under other 

 conditions in which it has been found by Lankester and 

 Saville-Kent it loses its sponge-like structure and comes to 

 resemble a multinucleated Protozoon. There are certain 

 points in Lankester's description of the soft parts which favour 



