Oerminal Streak o/Mysis. 191 



plates, too, lie at first close behind, but in later stages close in 

 front of the ectodermal primitive cells. At last all the primi- 

 tive cells split up into smaller cells. 



When tlie organs (nervous system and appendages) begin 

 to develop from the germinal streak the regular arrangement 

 of the cells in rows gradually disappears, since they commence 

 to divide in different planes. I am entirely unable to state 

 positively how many of the original longitudinal rows enter 

 into the formation of the ventral nerve-chain ; it seems to me 

 most probable that it is only the median row and the one lying 

 next to it on each side which take part therein. Another 

 process, however, is very distinctly recognizable in the forma- 

 tion of the ventral chain ; for the ectoderm-cells which are 

 destined for the ])roduction of the ganglion-cells become 

 developed as primitive cells, which, by means of budding, 

 give rise to rows of smaller cells towards the interior in a 

 manner precisely similar to that wdiicli Wheeler * has 

 described in the case of Insects. Yet in Mysis the ^' neuro- 

 blasts " (as Wheeler terms them) are not covered by the 

 epidermis, but actually represent the most superficial layer of 

 cells of the region in question, and, so far as I am able to 

 observe, persist throughout as epidermis-cells, while in the 

 case described by V/heeler they are situated within the 

 epidermis. 



A torsion of the germinal streak, such as I recently 

 described as occurring in Gammarus f, does not take place in 

 Mysis. 



As is evident from what has been stated above, besides 

 many points showing great agreement with the conditions 

 which are found in Gammarus, several noticeable deviations 

 from what is seen in the last-named form also occur. In this 

 connexion the existence of the lai'ger primitive cells at the 

 posterior end of the germinal streak of Mysis is especially 

 worthy of mention, since these are wanting in Gammarus. 

 On the whole Mysis is a more convenient and more easily 

 manipulated object for the study of the processes which are 

 here alluded to. The above results were derived partly from 

 the study of series of sections, but chiefly from the examina- 

 tion of transparent surfaccipreparations. J. Nusbaum \, the 

 most recent monographer of the development of Mysis, has 



* Wheeler, "Neuroblasts in the Arthropod Embryo," Journal of Mor- 

 phology, vol. iv., 18'J1, p. 337 et seq. 



t Zool. Anzeiger, 1892, no. 39G. 



X J. Nusbaum, " L'embryologie de Mysis chamceleo (Thompson)," Arch, 

 de Zool. exp. et g^n., s€r. 2, t. v., 1887, p. 123 et seq. 



