268 COMPARATIVE ANATOMY chap. 



as the latter grows at the expense of the yolk ami develops further, the sac becomes 

 smaller, so that when the embryo is hatched the size of the yolk-sac is only one third of 

 that of the young animal (Fig. 222 D). It must further be mentioned with regard to 

 the yolk sac that it is at no time in communication with the intestine. As the 

 embryo becomes constricted from the yolk the latter divides into two parts — an inner 

 part, lying inside the embryo, and an outer part, tilling the sac. These two parts 

 are connected by means of the stalk of tlie yolk sac, which jirojects downward from 

 the "head." The yolk within the embryo is divided into three unequal parts, the 

 largest of which fills the visceral dome ; another mass of considerable size fills the 

 "head," and these two masses are connected with a smaller portion lying in the 

 nuchal region. 



Loligo and Argonauta have a smaller yolk sac, round which the blastopore 

 grows at an earlier stage than in Sepia. The yolk sac of Argonauta is entirely 

 taken into the body before the latter has com]iletely closed ventrally. 



The quantity of nutritive yolk is still less in a Cephalopod {Omviatostrcphcs ?), 

 the spawn of which floats in the sea. Segmentation is in this case also partial and 

 discoidal, but the blastoderm almost completely encloses the yolk before any organ 

 develops on the germ, and no external yolk sac is formed. 



The results of the investigations hitherto made with regard to the germinal 

 layers, the development of the inner organs, and the inner differentiations of the 

 outwardly visible organs are so contradictory and in many cases so incomplete, 

 that no description of them is here attempted. Further investigation is much 

 needed. The development of the eye has already been described (p. 171), and that 

 of the hind-gut and ink-bag was illustrated (p. 197). 



Two important facts in the ontogeny of the Dibrancltia should be noted. (1) 

 In considering the arms as parts of the foot, it is important to notice that they 

 arise behind the rudiments of the head, and only secondarily come to lie round and 

 below the latter. The mouth, at quite a late stage, lies at the anterior end of the 

 circle of arms (Fig. 222 C). (2) The funnel consists of two separate lateral rudi- 

 ments, the free edges of which fuse secondarily. This point is important in connec- 

 tion with the separation of the two lobes of the funnel, which lasts throughout life 

 in Nautilus. For the view of the funnel as epipodium, (/. p. 116. 



The fact that the velum is wanting in the Cephalopod embryo must also be 

 noted. The absence of this organ is explained by the direct development of the 

 Cephalopoda within the egg capsule at the expense of a large quantity of nutritive 

 yolk. 



Investigations as to the development of the shell, and as to the nature of the 

 organ which has been called the shell gland, are much needed. 



XXIV. Phylogeny. 



No actual points of connection between the Molluscan phylum and anj' other 

 division of the animal kingdom have as yet been found ; the origin of the Mollusca 

 is therefore merely a matter of speculation. The present writer favours the view 

 that the Mollusca descended from animals like the Turhellaria, which had become 

 differentiated from the modern Platodcs by the acquisition of a hind-gut and a heart, 

 and the (at least partial) transformation of the genital cavity into a secondary 

 and primitively paired body cavity. There is a striking agreement in the nervous 

 system of the lower Molluscs {Chiton, Solenogastres, and in some respects the Dioto- 

 cardia) and that of the Platodes ; in both there is a ladder-like nervous system 

 with the principal trunks beset along their whole length with ganglion cells ; the 



