I I 2 BIOLOGICAL LECTURES. 



spending respectively with those of the gastrula, then both 

 atrial chamber and ganglion arise from entoblast. If, on the 

 other hand, we consider both these layers as ectoblast (on 

 account of their origin) then the alimentary canal of the bud 

 arises from that layer. Should it hereafter be shown that the 

 atrial chamber of Botryllus arises in the egg-development from 

 the entoblast, the difficulty regarding the ganglion would remain 

 as great as ever. If with Hjort we say that the layers of the 

 bud are not germ-layers "in the ordinary sense," we do no 

 more than restate the problem. 



Analogous contradictions have recently been observed in the 

 asexual reproduction and regeneration of worms. In these 

 cases, as a rule, the new parts are derived from the corre- 

 sponding parts of the old body and in accordance with the 

 demands of the germ-layer theory. In the regeneration of the 

 tail of Luinbriculus, for example, Miss Randolph has shown ^ 

 that the new tail is derived from a mass of cells which have 

 nearly the same arrangement of germ-layers as in the egg- 

 embryo and are derived from the corresponding layers of the 

 old body. The later researches of v. Wagner and Schmidt 

 seem, however, to show that in this same animal both procto- 

 daeum and stomodaeum are regenerated from the entoblast 

 instead of arising from the ectoblast, as in the egg-embryo. 

 In the Rhabdocoelous genera Microstoma and Stenostoma the 

 new pharynx is regenerated (in the process of asexual division) 

 from the mesoderm (parenchyma) as shown by the concurrent 

 researches of v. Wagner and Ott, while in the egg-embryo 

 (of the related genus Mesostoma) it is invaginated from the 

 ectoblast. 



It is plain that in such cases of asexual development as these 

 the developmental test of homology breaks down more or less 

 completely ; for parts that are undoubtedly homologous (gan- 

 glion of the egg-embryo and bud-embryo of Botryllus, etc.) 

 differ totally in mode of origin even with respect to the germ- 

 layers. It may be urged that in regeneration and agamo- 

 genesis, development is condensed and abbreviated so as no 

 longer to repeat the phyletic development, and this is no 



^ Journ. Morph., VII, 1892. 



