222 SCIENCE PROGRESS. 



themselves in hemispherical groups, which become the 

 future chambers. The cellules intermediaires, which we 

 last saw in the meshes of the syncytial network, begin 

 to flatten and arrange themselves round the exhalant 

 lacunae to form their epithelium. Those of the inter- 

 mediaires not used up in this way become the connective 

 tissue of the adult. The amoeboid cells, after disgorging 

 their prey, become the wandering cells of the adult. 

 Occasionally, however, some of the ciliated cells become 

 actually digested when engulfed by the amoeboid cells. 



The above assertions of Delage with regard to the 

 process of metamorphosis are certainly such as zoologists 

 will find it difficult to credit without very strong proofs of 

 their correctness. The following reasons are advanced 

 by him in support of his statements : in Spongilla, where 

 this process of cannibalistic phagocytosis is most complete, 

 there are to be found in the interior of the sponge, at the 

 commencement of the second day, neither chambers nor 

 free cells in sufficient number to give rise to them. Apart 

 from the spicules there is nothing beneath the epidermis 

 except the groupes polynuclees and the cellules inter- 

 mediaires. At this period small cells begin to appear, at 

 first few, but soon innumerable, which arrange themselves, 

 to form the chambers, and in proportion as these cells 

 appear the granules in the groupes polynuclees disappear. 

 When the chambers are fully formed the groupes poly- 

 nuclees have disappeared, leaving in their place only the 

 amoeboid cells. " There is thus an evident correlation 

 between the newly appeared cells and the globules of the 

 groupes polynuclees." If these globules were vitelline 

 granules they should gradually diminish and disappear, but, 

 on the contrary, at the moment when the new cells appear 

 the globules increase in size and separate from the amoeboid 

 cells. 



Moreover, if the countless cells which form the chambers 

 do not come from these globules, they should be derived 

 from pre-existent elements, either from the amoeboid cells 

 or the intermediaires, which one would expect to find in 

 active division. But, as a matter of fact, neither the one 



