ON THE DEVELOPMENT OF THE POND-SNAIL. 387 



canal proceeds, which diverticula penetrate the unorganised 

 yelk, and, filling up the position once occupied by it, become 

 the two lobes of the Cephalopod liver. This process is 

 probably a general one throughout the animal kingdom, with 

 variation in non-essentials. 



Muscular layers and muscles. — I have above spoken of 

 fusiform cells arranged as a layer on the inner surface of the 

 body-wall, and as surrounding the alimentary canal and 

 bilobed mass of pellucid cells. These represent the meso- 

 blastic elements of the embryo. I have not been able to 

 find in the early stages of Lymn<2us a layer or group of 

 undifferentiated embryonic cells lying definitely between the 

 gastrula's body-wall and stomach ; such a layer would be a 

 mesoblast. It is possible that there are some such loosely placed 

 cells during a particlar phase of the development, just as 

 there are in Pisidium, but they are derived either from the 

 enclosing ectoderm or the invaginated endoderm in the first 

 instance. The appearances are strongly in favour of the 

 fusiform cells which lie in apposition to the epidermic cells 

 of the body-wall, being derived from ectoderm or epiblast cells. 

 The most noticeable groups develop at the circumference of 

 the shell-gland (PI. XVII, fig. 8 mii). The processes which 

 pass from the gastrula-endoderm-cells to the body-wall appear 

 (fig. 19), eventually to become muscular, but whether they 

 should then be attributed to the latter or the former is 

 doubtful. The term Triploblastic is applicable to Lymnceus 

 and to other m.olluscs in which there is no definitely consti- 

 tuted layer intermediate between the gastrula's ectoderm 

 and endoderm, since in it and them, as in all the groups 

 filiated to Vermes the musculature has not only relations to 

 the outer w^orld and to the gastric space, but to a third inter- 

 posed space — the hsemolymph cavity — in which a vascular 

 system and blood-lymph spaces develop. 



It is clear (and in saying this I am qualifying, though not 

 recalling, what I have stated in my essay in 'Ann. and Mag. 

 Nat. History,' June, 1873) that a mesoblast or third inter- 

 mediate layer must either be derived from epiblast or hypo- 

 blast (either or both), and so cannot be spoken of as of co- 

 ordinate value with those two layers; or, on the other hand, 

 it must be a separate entity originating simultaneously with 

 the epiblast and hypoblast from the egg-cell or its segmenta- 

 tion products. The latter case is one which certainly has 

 not been usually contemplated in the use of the term " me- 

 soblast" or " middle layer," and there is very small warrant 

 for assuming it as expressing the historic or phylogenetic 

 mode of origin of the layer in question. It is possible that 



VOL. XIV. NEW SER. C C 



