PHYSIOLOGY OF SPONGES. 323 



excrete in a manner possibly protective for the colony * and are charged with 

 its reproduction. 



At this stage we may conceive the secretion of an attractive odour to have 

 become permanently associated with their nitrogenous excretion, with the 

 advantage of more frequent reinvigoration of the stock by conjugation with 

 locomotor units from other colonies. 



Development led to differentiation of purely reproductive cells from those 

 charged with excretion, the reproductive cells being placed in greater safety 

 by a covering of the excretory cells. The latter would still attract the 

 locomotor units, which, when engulfed, they would pass on for nuclear union 

 with the gonocytcs. This absorption from the exterior and transmission to 

 the interior would be simplified (as the reverse process has probablv been in 

 nephridial tubes) to perforation (cf. 1892, p. 182). Misdirected perforation, 

 placing the excretory surface of the sponge in communication with the 

 flagellar surface instead of with an oocyte, would sporadically occur, and 

 when so occurring would lead to a sudden increase in the hydraulic efficiency 

 of the flagella as a food-catching apparatus, supplies being thus brought to 

 centrally placed choanocytes which otherwise could only be reached by water 

 which had passed over their neighbours. Such misdirected perforation would 

 therefore be perpetuated by the increased prosperity of those colonies in 

 which it happened, their form being now in essence that of the rudimentary 

 Olynthus ; from this upwards " all transition from more to less primitive 

 canal systems exhibits an increase in the ratio " that determines the 

 mechanical efficiency of the canal system (1888 ; 1895, p. 29). 



The ceroids are fragmented archasocytes (see Minchin 1897, p. 500) which 

 (ante, p. 300) break out into the surrounding water and effect conjugation 

 with the archaeocytes of other colonies. 



My excuse for printing such a fanciful story is the extreme difficulty of 

 forming any conjecture as to possible value to the organism of the intermediate 

 stages in the evolution of a hydraulic system supplied by intracellular pores 

 (cf. MacBride, 1918, p. 52). If the pores were multicellular, we might 

 conceive them as originating as grooves or furrows in the flagellate surface, 

 with the advantage in every stage of bringing better supply to the more 

 central flagella. Such an evolution would fit in with many embryological 

 observations, and there is so much evidence that sponges are polyphyletic 

 that it is not impossible that both developments have occurred j\ But 

 in both Calcinea and Calcaronea we know that the pores are intracellular. 



* The nitrogenous excretion of the arehreocytes may be beneficial to the colony by being 

 either repulsive to foes or attractive to friends. But, in view of the known defecation by 

 collared cells, these probably get rid of their own nitrogenous poisons, and I do not now 

 believe that in this sense either the primitive archrcocytes or the specialised porocytes excrete 

 for the whole colony. Loisel's criticism on this point (1898, p. 203) is quite just. 



t In the lXmatiidie and the Ilexactinellidu there is no evidence of the presence of pore- 

 cells, aud it is doubtful if in either of these groups there is any sexual reproduction. 



