68 SECTIONAL ADDRESSES. 



balistic habit of supplementing their own synthetic products by engulfing 

 the spores of their neighbours or their fragmentary mortal remains. Some 

 carried cannibalism so far as to give up altogether the colouring matter 

 which in sunlight made carbohydrates out of their environment, living 

 instead on carbohydrates and proteins formed by productive contem- 

 poraries. 



These seceders formed many specialised groups, of which the choano- 

 flagellates, like many of the much higher ciliates, adopted the policy of 

 fixing themselves to the ground by their blunt ends, so that the flagellum 

 made a current carrying the desired spores, which passed over the sticky 

 body. For this function they changed the tractellar action of the flagellum 

 (still preserved in sponge larvae) into a pulsellar action 20 — a change 

 which may be effected merely by lengthening the flagellum. Also, by 

 means of the well-known transparent collar, they masked the part of 

 the flagellum nearest the cell, where the motion is purely lateral and 

 only drives the food away from the adhesive surface without assisting 

 the current. 



Where two choanoflagellates, side by side, drove their currents in the 

 same direction, each reinforced the other and each cell obtained more food 

 than when working alone. 30 Consequently choanoflagellates began to 

 form colonies of various types. One, the Proterospongia of Saville Kent, 

 has been hailed by all as marking a stage towards the evolution of sponges ; 

 but from this flat gelatinous crust it is very difficult to imagine the 

 advantageous steps which led to an Olynthus supplied by intracellular 

 pores ; I would conceive the ancestor as a saucer-shaped Proterospongia 

 colony fixed by a narrow base. I propose the name of Porifera vera for 

 the sponges descended from this form, and, like it, having collar-cells 

 based on interstitial jelly that contains other cells and is sheathed by 

 other cells from the water. Such sponges, calcareous, horny, and tetracti- 

 nellid, are contrasted with the Porifera nuda, the Hexactinellida, in which 

 the collar-cells and most other cells are connected with each other by 

 protoplasmic threads, but lie otherwise naked in the water. Hexacti- 

 nellida must be descended independently from some colonial choano- 

 flagellate like a Codonosiga, whose branches developed the faculty of 

 anastomosis ; and they are of a lower grade altogether than the true sponges. 

 Living in the permanent currents of abyssal depths, they have not found 

 necessary the hydraulic mechanism for a powerful outflow, which is found 

 in every other group of sponges 31 ; their hydraulic organisation is limited 

 to a separation of the effluent from the afferent channels, but they have 

 developed the secretion of cubic opal, on the skeleton crystals of which the 

 network of their naked cells is extended across the current. They re- 

 produce by ciliate gemmules : Ijima finds no gametes, but we are at 

 liberty to imagine that the free-swimming gemmule may possibly produce 

 them, as in Volvox and other flagellates. 



Among the monaxon members of Dendy's Tetraxonida many sponges, 

 of the general appearance of Porifera vera, have spicules which show, 

 by a symmetry about three planes at right angles to each other, that they 



29 Geoffrev Lapage, 1925 : Q.J. M.S. Ixix, pp. 477, 465. 



30 Q.J.M.S. 1923, lx\ai, p. 298. 

 31 Jbid. p. 312. 



