488 Mr. A. T. Masterman on the Nutritive and 



mately becomes of it In tliis part of the sponge is very difficult 

 to make out. 



The transformed choanocyte either completes the course of 

 intracellular digestion itself, or it is devoured by mesodermal 

 phagocytes together with its contained nutritive particles. 

 The choanocyte, however, at this stage so closely resembles 

 the mesodermal cells that its ultimate fate cannot be followed 

 in sections. 



We have certain cases described which lead one to su|)pose 

 that the cells may be devoured by phagocytes. Thus 

 Dendy (4) figures a phagocytic cell in close contact with a 

 choanocyte, and I have observed a few appearances in these 

 sections which appear to be phenomena of a like nature. 

 Delage (3) found that in the larva of ISpongiUa the " ciliated " 

 cells are engulfed in the indifferent amoeboid cells, and some 

 at least are digested ; tiie others, on the other hand, emerge 

 later and form the endodermal pinnacocytes and choanocytes. 

 These and other instances show that it is not an unknown 

 phenomenon for the choanocyte to be eaten by the phago- 

 cytes. 



At the same time there is not sufficient evidence in this 

 sponge for such an assumption, and perhaps the real process 

 is that intracellular digestion is conducted by the incurrent 

 choanocyte, but that if this laden cell is unable, through an 

 excess of nutrient material or through a low metabolic activity, 

 to conduct the ])rocesscs of digestion itself, it falls a prey to 

 the mesodermal cells, just as do degenerating tissues in 

 the phenomenon of phagocytosis occurring in the develop- 

 ment and life-history of many forms. 



The radial chambers appear always to be lined almost 

 uninterruptedly with choanocytes in all later stages of diges- 

 tion, and we may therefore conclude from this that the vacated 

 places of the immigrating choanocytes are tilled up either by 

 re-transformed choanocytes or by mesodermal cells which 

 assume the collared phase. One or both of these processes 

 must obviously take place in order that the sponge mav not 

 be reduced to an amoeboid mass without radial chambers. 

 This result has actually been obtained in several cases cited 

 below, when tlie sponges had been overfed. 



We thus find that : — 



1. The process of ingestion is conducted almost entire/// by 



the choanocytes. 



2. These choanocytes travel to the interior of the colony, 



and in doing so assume the shape of mesodermal 

 phagocytes. 



3. The process of digestion is conducted in the so-called 



