GENERAL CONSIDERATIONS ON THE MOLLUSCA. 323 



the shell-gland on the dorsal surface of the larva, and thus gives the 

 larva the special character of the Mollusc without at first affecting its 

 general appearance (Figs. 14, 15, 18, pp. 28,31 and 36 ; Figs. 51 and 

 56, pp. 1 25 and 135). Somewhat later, but also at a very early stage, 

 the foot appears on the ventral side of the larva. The very early rise 

 of this organ which may in a few cases be found even before the 

 Trochophore form fully develops, must be regarded as a shifting back 

 to an early period of embryonic development of this feature which was 

 only a recent acquisition. It is all the more easy to admit this, as the 

 shell-gland is found to arise exceptionally early in those forms, the 

 development of which show marked specialisation, as, for instance, in 

 the Unionidae (Fig. 22, p. 50) and in the Cephalopoda (Figs. 116 

 A, p. 255 and 131 D, p. 280). The shifting back of the shell to the 

 earliest possible embryonic period can easily be explained by its im- 

 portance as a protection to the larva, a fact which may be observed 

 in every Lamellibranch or Gastropod larva whether young or old ; 

 the slightest disturbance causes the animal rapidly to retreat into its 

 shell and thus to sink to the bottom of the water. 



The larvae of the Amphineura have no true shell-gland, a peculiarity 

 that would increase the resemblance between them and the Annelida 

 if their organisation were better understood. It must at present be 

 confessed that the larvae of the higher Molluscs are far more like those 

 of the Annelida than are the larvae of these more primitive forms, in 

 which we should expect a closer resemblance. The shell-plates of 

 Chiton, further, arise in the same region as the typical shell of the 

 higher forms (Fig. 5, p. 9). 



The manner in which the shell appears in the embryo favours the 

 view that it is derived phylogenetically from a cuticular dorsal cover- 

 ing, within and beneath which calcareous concretions were deposited. 

 The shell-plates of Chiton have, indeed, with some probability, been 

 traced back to the transformed spines of this animal (p. 12), but on 

 the whole it seems more likely that the Chiton shell, which consists 

 of a number of plates, arose in consequence of a secondary distribution 

 (determined by the manner of life of the animal) of an originally 

 continuous dorsal carapace.* From this flattened, bowl-shaped shell 



*The shell has also been regarded as a partly internal dermal skeleton in 

 consequence of its condition in Chiton, where it is traversed by strands of 

 connective tissue (Fig. 8, p. 12), and the retractors of the body inserted into 

 it have been thought to play an important part in its development (Thiele, 

 No. 20). This last factor is in any case of importance in connection with the 

 various modifications of the shell (where the latter is already present). 



