406 MARGARET ROBINSON. 



and a small median strand of cells whicli is clearly an invagi- 

 nation. 



Nusbaum (1887) describes a median double cord of nervous 

 tissue in tlie brain of a My sis embryo, and in his account of 

 the development of Oniscus (1886) he states that the brain 

 in the optic region consists of four lobes, the two outer ones 

 only being the optic lobes. This, indeed, is the state of things 

 in this stage. The median cord here more closely resembles 

 that in My sis than it does the tiny invagination figured by 

 Reichenbach (1886). 



In neither of his papers does Nusbaum say how this median 

 cord originates. That it does arise from the band of cells 

 which connects the optic ganglia at their first appearance is, 

 I think, certain, and from its appearance in figs. 31 and 32 

 it seems probable that it arises by invagination. 



The central mass is itself double, and the optic ganglia are 

 slightly bilobed, so that we may say that each ganglion here 

 consists of two parts — a small inner one whicli is single, and 

 a large outer one which is bilobed (fig. 31). But it seems to 

 me that the real state of things is better expressed by saying 

 that the protocerebrum in this region consists of a pair of 

 large lateral bilobed optic ganglia, and a central mass of 

 nervous tissue which is double. 



Tracing the nervous system farther back we iind that the 

 central mass increases in size, and the two lobes of each optic 

 ganglion gradually fuse into one (figs. 32 — 34) . In a section 

 just anterior to the first antenna (fig. 35) one can see a still 

 further separation of the central mass from what is here 

 the rest of the deuterocerebrum, while in the region of the 

 first pair of antenna) the median mass is partly separated 

 from the ganglia by a layer of ectoderm cells (fig. 36,/.). 

 These colls have fusiform nuclei which look as though they 

 might belong to connective-tissue cells, but the mesoderm at 

 this stage is liot sufiiciently advanced for that to be the 

 case. 



On tracing the median mass backwards one finds that it 

 first becomes single (fig. 37), and finally, in the region of the 



