94 Mr. W. Marshall 07i the 



and enclosed between other tissue-elements j it is therefore 

 the upper extremity, in which the clear respiratorj plasma has 

 collected, that is compelled to free itself from obstructive 

 surroundings ; the cell therefore loses its purely prismatic form 

 and acquires a process which is accessible to the water on all 

 sides. But this alone does not suffice, even if it widens 

 upwards ; it is compelled to enlarge its surface still more, 

 and this is effected by its quitting its simple cylindrical 

 or conical form and becoming converted into a funnel. This 

 does not imply, as previous observers have sometimes stated, 

 that the flagellate cells could not take in nourishment, but, 

 as we see from the other endodermic cells, for that purpose 

 they need not acquire a collar. Thus in ray view there exists 

 a very essential difference between the functions of the collar 

 in the flagellate cells of the Sponges and in the Choanofla- 

 gellata, of which latter Blitschli remarks * : — " There is 

 unanimity among observers that the collar, at least in the 

 Cryptomonadina, i? an organ connected with the reception of 

 food." 



In this way I come to the conclusion that the Flagellata 

 and the flagellate cells of the Sponges absolutely stand in no 

 phylogenetic connexion, but that the two peculiarities, which 

 agree so remarkably, are due rather to adaptations sui generis. 

 The flagellate cells of the endoderm of the young sponge, 

 probably even in one which has originated from an amphi- 

 blastula, ai e not at once to be identified with the flagellate 

 cells of the larva ; first of all the flagella disappear, and then 

 (after the cell has become flattened as an endodermic cell, and 

 then again extended with a fresh accumulation of the clear 

 respiratory plasma at the free pole), so soon as the flow of 

 water becomes possible, they again make their appearance 

 together with the collar ; in the sponge-larvee in which the 

 endoderm is formed by division of the coenoblastema — and 

 these are probably the majority — there can be no question at 

 all of any such connexion. 



While I am perfectly in agreement with Schulzein denying 

 any relationship between the Sponges and Choanoflagellata, 

 our views with regard to the degree of relationship between 

 the Sponges and Cnidaria are, as already remarked, very 

 divergent, and I will now endeavour to support and establish 

 my opinion in opposition even to such serious objections as 

 Schulze puts forward. 



As we have already seen, Schulze, while placing the onto- 

 genetic processes in Sponges in the first line, as justly required 



* Bronn's ' Klassen und Ordnungen,' Neue Bearb. Bd. i. p. 885. 



