CRANGON VULGARIS. 117 



genera (Lina, Pj^rrhocoris, etc.) a niimber of amoeboicl 

 cells in the centre of the yolk which appear to be in the 

 process of division and to be connected together by a pro- 

 toplasmic network. Ayers ('84), studying GEcanthus, 

 finds in the yolk both amoeboid nuclei and amoeboid cells, 

 some of which migrate to the surface, and the cells, join- 

 ing each other by a fusion of the protoplasmic filaments, 

 form the blastoderm. Patten ('84) could not see the 

 nuclei arise into the blastema in the living egg of Phryg- 

 anids. In his earliest stages, the blastema, though not 

 divided into distinct cells, was nucleated, whilebelow this 

 were numerous amoeboid cells, distributed through the yolk 

 and connected by protoplasmic filaments. In later stages 

 these cells have almost entirely disappeared, while the 

 blastoderm has become much thicker and the nuclei more 

 numerous ; from which the conclusion is obvious tbat the 

 nuclei formerly seen in the yolk have migrated to the sur- 

 face. Korotneff ('84) does not recognize a blastema in 

 Gryllotalpa. He has the blastoderm arise by a migration 

 of amoeboid cells to the surface. 



In other eggs (e. g., Aphis, Metschnikoff, Witlaczil) 

 it has not been shown that the nuclei, before leaving the 

 centre of the egg, have each their own proper protoplasmic 

 envelope ; but in these cases there can not be the slightest 

 doubt that the segmentation proper takes place, at first, 

 not on the surface, but in the centre of the egg. Accord- 

 ing to Witlaczil ('84), the nuclei in Aphis do not reach 

 the surface until sixteen of them are formed. These ob- 

 servations, as well as those quoted before, show that the 

 view of Robin ('62) that the nuclei of the blastoderm 

 arise b}^ budding is as little justified by facts as that of 

 Weismann ('63) that they arise spontaneously. They 

 do, however, conclusively show that we do not have here 

 a "superficial" segmentation, but instead one which is 



