44 Proceedings of the Royal Physical Society. 



bounded by ectoderm cells, one row thick in most places, 

 but continued here and there into the subjacent mesoderm 

 in the form of strings of cells several rows deep, and 

 occasionally forming short tubules or caeca. ' These may 

 branch or join one another, or may be united by strings 

 of ectoderm cells having no lumen (PL III., Fig. 6, ec. 

 cce). 



The mesoderm of the wall of the invaginated tube is thick. 

 In some places where it forms the projections into the lumen 

 (see PL I., Fig. 16) it is thicker than the mesoderm of the 

 non-retractile part of the body wall. The projections probably 

 correspond to the terminal lobes of Figs. 6 and 7, and the 

 external ridges of Fig. 5. The mesoderm of the invaginated 

 tube is a homogeneous matrix, with a few fusiform connective 

 tissue cells scattered through it. It usually has a considerable 

 number of lacunse, especially in its inner or ectodermal part, 

 and opposite to the insertion of the mesenteries (PL III., 

 Figs. 1 and 5), where the lacunse extend outwards towards 

 the endodermal surface. This arrangement of lacunse is just 

 what is to be expected from the fact that the invaginated tube 

 is merely the upper part of the body wall retracted (see PL 

 I., Fig. 13). The mesoderm also encloses the ectodermal 

 strings of cells or caeca, given off from the invaginated tube. 

 In its outermost layer, just inside the endoderm, the mesoderm 

 contains a delicate layer of longitudinal muscles (PL III., 

 Fig. 1) continuous with those of the mesenteries. Probably 

 these are retractors which aid the muscles of the mesenteries 

 in invaginating this part of the body wall. 



The endoderm which forms the outermost layer in the 

 wall of the invaginated tube is exactly like that lining the 

 non-retractile part of the body wall. As a rule it appears 

 to be thickest between the points of insertion of the mesen- 

 teries (PL III., Fig. 5, en). 



Tubules lined by endoderm cells and penetrating into the 

 subjacent mesoderm are found here also, but they are neither 

 so large nor so numerous as the similar tubules in the non- 

 retractile part of the body wall. Whether these endodermal 

 tubules ever communicate, either directly or by means of 

 the lacunae, with the ectodermal caeca given off from the 



