74 K. kishi.n'ouyï:. 



brain (fig. 34 ex. gr.). The invagination is in the margin of the brain 

 where none of the peculiar cell-groups above referred to is f«nind. 

 After the appearance of this invagination there is found another jiaired 

 e]Ml.)lastic invagination, along the antei'ior internal corner of the l)rain 

 (figs. 13, 30, 38. ////. ///•.). This new invagination is also on the 

 maro-in of the brain, where the peculiar cell-grou])s are not found. I 

 believe that the external invagination is homologous with the invagi- 

 nation of the lateral vesicle of the spider and the internal one to tlie 

 semicircular groove. These cephalic invaginations of Limulus, how- 

 ever, are very shallow and disappear before the separation of the 

 nervous svstem from the epiblast takes place. They are already not 

 found in tlie eml)rvo in the stage of fig. 14. Ihit tlie two cell-masses 

 which were thus invaginated are found separate in the deep part (figs. 

 37, 51). The part of the brain formed l)y the external groove liecomes 

 the oj^tic ganglii^n of the lateral eyes as in the case (^f the spider. The 

 optic ganglion of tlie niedian eyes is produced from the cpildastic 

 thickening forming the united mouth of the invagination for tlie 

 median eves (fig. 69, g. uird. r. see infra). Tiie thickening is always 

 between, and touches on each side, the two eephalic ganglia (figs. 14, 

 37, 69). 



At first the anteriin- and lateral borders of the lirain were directly 

 on the anterior and lateral borders of the ventral ])late ; but as the 

 oTowth of the brain la o-s l)ehind that of the ^-entral plate, it becomes 

 gradually se])arated from the margin of the ventral plate. 



The brain whicli was almost circular in its outline, as its height 

 was almost equal to its breadth (figs. 12, 34), becomes afterwards 

 semicircular (fig. 13), later still, sic-kle-shaped (fig. 14), and shortly 

 before hatching almost str;iight (fig. '59). 'I'his is caused pfobal)ly by 

 the ])eculiar de\'elopment (if segments. All the segments seem to 

 develop most vigorously ai the line inidway between tlie ventral aiid 



