538 ADAM SEDGWICK. 



and cells for acquiring food, i. e. the differentiation supposed 

 by the invaginate gastrula hypothesis. For it is clearly neces- 

 sary on this view that the nutritive cells should be in direct 

 contact with the external medium, and this they are not on 

 Metschnikoff^s view. 



Then again, what justification do we find in the animal 

 kingdom for the hollow blastula hypothesis ? With the pos- 

 sible exception of Volvox, I know of no form at all approach- 

 ing a hollow blastula in structure. 



These difficulties are avoided by the hypothesis that the 

 primitive form was solid, which also suits the facts quite as 

 well. (1) It is no longer necessary to suppose the migration 

 inwards of cells. (2) There are a considerable number of 

 instances in the animal kingdom of lowly organised solid 

 forms which give a certain amount of justification to the 

 hypothesis of a solid ancestor. There is Trichoplax (No. 29), 

 the Orthonectidee, the whole of the Protozoa, the acoelous 

 Turbellaria, and finally the Sponges, or at least some of 

 them under certain conditions. The latter case is of particular 

 interest, and deserves a little attention here. 



Metschnikoff (No. 22, p. 372) states that Halisarca, when 

 overfed with carmine, loses its canals and becomes a mass of 

 amoeboid cells containing swallowed food, and surrounded by 

 a common envelope of ectoderm. The same fact has been 

 observed by Lieberkiihn in Spougilla (No. 20), in winter. 

 From these observations, and others by Hseckel and Carter 

 (quoted by Metschnikoff, No. 22, p. 361), it appears that under 

 certain nutritive conditions, the flagellated endoderm cells of 

 sponges may lose their flagella and become amoeboid, and the 

 whole sponge revert to the condition of the larva of Aply siua 

 (Shulze, No. 30, PI. 24, fig. 30) of a protoplasmic network 

 with nuclei at the nodes, and a cortical layer of ectoderm. 



Metschnikoff (No. 22) has further observed that in many 

 cases the food of the sponge passes directly into the paren- 

 chyma and is not found in the collared endoderm cells ; while 

 in other cases it is found both in the cells of the endoderm 

 and parenchyma. These facts, as Metschnikoff points out, seem 



