THE DEVELOPMENT. OF THE NERVOUS SYSTEM. 487 
which is consistent with the developmental history and final 
condition of the Arthropod brain, and renders a great number 
of facts intelligible. I advance it with some diffidence, as I am 
aware of the difficulties which beset any investigation of the 
development of the nerve-centres, and the ease with which 
the significance of a layer of cells may be misinterpreted, but 
I present it as a working hypothesis which may serve as a 
guide to future investigators. The evidence in its favour 
depends on but few preparations, and is the result of the 
study of a single insect; the difficulties of obtaining satis- 
factory evidence is very great owing to the extreme rarity of 
good preparations in the earlier stages of the development of 
the larva in the proper planes for the investigation of the 
question, as suitable sections are by no means easy to prepare. 
In examining the nymphoid state of the embryo, I discovered 
Fic, 64,.—A sagittal section of the neuroblast of a newly-hatched larva. d,. head 
discs ; 7, ventricular cavity of the hemisphere ; g, central stroma of the neuro- 
blast. 
a condition, which is figured in Pl. XV., Fig. 1, in which the 
preoral ganglia are apparently vesicular, and subsequently I 
found the observation borne out by the condition of the nerve- 
centres in the newly-hatched larva (Fig. 64). From these 
observations, I am strongly urged to the view that the remark- 
able mantle layer in the adult larva is the roof of a cavity, the 
remains of which form the ventricular cavities of the brain of 
the imago. 
It is true that the early vesicular stage, which appears to be 
the probable interpretation of my sections, has not been hitherto 
suspected in the Arthropod brain, and although Kowalevski [97] 
discovered a central cavity in the brain of Sagitta, he observed 
