38 BIOLOGICAL LECTURES. 



years ago that in Nereis each of the cells into which d* divides 

 buds forth several small cells (Fig. 5, b) t which do not enter 

 into the mesoblast-bands, though I did not correctly determine 

 their fate. More recently Conklin was able to show that a 

 similar process occurs in Crepidula (Fig. 5, a), and that the 

 cells thus formed are ento blast-eel Is which enter into the forma- 

 tion of the archenteron. On reexamining the matter in Nereis 

 I found the clearest evidence that the same was true here. In 

 both these cases, therefore, the posterior cell of the fourth 

 quartet is of mixed character, and divides into two mesento- 

 blasts, each of which first gives rise to a number of entoblast 

 cells (two in Crepidula, four or five in Nereis], the residue con- 

 stituting the mesoblast. In both these forms, therefore, the 

 ectoblast (and in Crepidula the ectomesoblast) are first com- 

 pletely segregated, and the archenteron which remains gives 

 rise to the mesoblastic pole-cells. The latter are, therefore, of 

 entoblastic rather than ectoblastic origin, and may be designated 

 as the entomesoblast. 



Further examination of these phenomena brings out some 

 highly interesting facts which seem to constitute a striking 

 case of ancestral reminiscence in cleavage. Several years ago 

 I found in two genera of annelids, Aricia and Spio, that the 

 small entoblast-cells of Nereis and Crepidula (i.e., those budded 

 forth from the two mesoblasts derived from the division of d^ 

 or M) are represented by a single pair of quite rudimentary 

 cells, scarcely larger than polar bodies (Fig. 5, e), which ap- 

 parently take no part in the building of the archenteron, and 

 can only be explained as vestiges or reminiscences of such a 

 process as occurs in Crepidula or Nereis. Later researches 

 have revealed the presence of these vestigial entoblasts in 

 several other forms, and have shown further that they are con- 

 nected by several intermediate steps with the larger functional 

 cells found in Crepidula. Thus in Amphitrite (Mead) and 

 Planorbis (Holmes) they are quite vestigial, agreeing essen- 

 tially in size and origin with those of Aricia. In Unio (Lillie) 

 they are considerably larger (Fig. 5, d), in Clymenella (Fig. 5, c) 

 they are as large as the mesoblastic moiety (Mead) ; while in 

 Crepidula (Fig. 5, a) their bulk surpasses that of the meso- 



