CLEAVAGE AND FORMATION OF THE GERM-LAYERS. 27 



with the ectoderm (Fig. 13 C). It is not known whether the 

 blastopore closes from behind forward, so that its relations to 

 the mouth and anus are still uncertain. [In Unio (LiLLiE) the 

 blastopore is said to close by the forward growth of its posterior 

 lip, the ventral plate.] 



In the Unionidae also there is an invagination-gastrula, but the 

 archenteron is here still smaller than in Cyclat (GoETTE, Fig. 23 

 A-C, e, p. 51). 



The presence of an invagination-gastrula in the Unionidae was first ob- 

 served by RABL in 1876 (No. 43), and SCHIEBHOLZ in 1878 (Nos. 47-49), and 

 the gastrula was said to have a specially large archenteron (Fig. 22, p. 50), but it 

 is impossible to reconcile either the position or the large size of this arch- 

 enteric invagination with the later development of the embryo, especially as 

 the alimentary canal is at first very inconspicuous. According to GOETTE'S 

 recent description (No. 15), this deep depression represents the shell-gland 

 which is here specially strongly developed, the archenteron, on the contrary, 

 being reduced (Fig. 23 A-C, sd and e). The subject will be discussed more 

 in detail in connection with the further development of the Unionidae (p. 49). 

 [LiLLiE (No. III.) has shown definitely that the large invagination observed 

 by RABL and SCHIERHOLZ was the temporarily inturned shell-gland, the true 

 archenteron being very small.] 



Between the extreme cases of epibolic and embolic gastrulation, 

 such as are offered by Gydas on the one hand and Teredo on the 

 other, Ostrea forms to a certain extent a transition. In the Euro- 

 pean as well as in the American Oyster, the micromeres have been 

 observed to grow round the macromeres, of which there are only one 

 or two present at this stage (Fig. 14 ^4). HORST and BROOKS agree 

 in denying the presence up to this stage of a cleavage-cavity, but 

 such a cavity arises as soon as the macromeres increase in number. 



As the micromeres even during epibole projected slightly beyond 

 the macromeres at the vegetative pole, a depression arises in this 

 region. When the macromeres now divide, a stage arises, with an 

 almost triangular blastopore, which cannot be distinguished from an 

 invagination-gastrula (Fig. 14 j3). During these processes, important 

 alterations have taken place in the form of the embryo ; an invagina- 

 tion closely resembling the archenteron in form, the so-called shell- 

 gland, has appeared (Fig. 145 and 0, sd). To this and the further 

 transformation of the embryo (C-E) we shall return later. 



Conditions similar to those in the Oyster are found also in the Lamelli- 

 branchs examined by LOVEN (Modiolaria and Cardium), in which the abund- 

 ance of yolk determines the early circumcrescence of the entoderm-elements, 



