94 BULLETIN OF THE 
bered that at this time the ventral area is composed of a thin layer of 
ectodermic cells ; these cells offer the least resistance to the movement of 
the yolk mass, which, therefore, takes a ventrad direction. 
During the period of reversion Agelena presents an interesting resem- 
blance to certain conditions in the development of Oniscus. But accord- 
ing to Bobretzky (74, Fig. 15, hd.) the proctodeum in Oniscus arises 
some time before the formation of a tail-fold (1. ¢., Fig. 17, 77.), which, 
moreover, is never conspicuously indicated. In Agelena the tail-lobe is 
very prominent, and the fold which results in its formation appears long 
before the proctodzeal invagination. The appearance of this tail-fold, as 
seen in sagittal sections, is so like that presented by the first stage in the 
formation of the proctodeeum in Oniscus (I. ¢., Fig. 15) as to suggest the 
possibility that the infolding in the latter case is really a tail-fold and 
not the proctodeeum, in which event hd. of Fig. 15 would correspond to 
rf. of Fig. 16, and the proctodzeum in the latter figure would be a new 
invagination. The principal objection to this view, aside from the 
author’s reputation for accurate observation, lies in the closeness of the 
stages of Figs. 15 and 17, which would not seem to allow time for such 
radical changes. Another and perhaps sufficient objection is, that the 
invagination in question (Ad. Fig. 15) is lined with columnar epithelium 
like the proctodeum of the succeeding stage, and that the slight tail-fold 
is lined with flat cells. But whatever may be the truth with regard to 
Oniscus, I am certain that in Agelena the tip of what I have called the 
tail-lobe becomes the morphological end of the body, and that the proc- 
todeeum pierces the tip of this lobe after the reversion of the embryo 1s 
nearly completed, and the tail-lobe has become much shortened. 
One fundamental difference supposed to exist between the eyes of 
Arthropods and those of Vertebrates, relates to the direction in which the 
light traverses the retinal elements. In the vertebrate eye the light 
passes through the cells from their deep to their outer (genetically con- 
sidered) ends. In the arthropod eye the light was supposed to have the 
reverse direction; but that this difference does not exist in the eyes of 
Agelena is rendered apparent from its manner of development already 
described. If the proliferation of cells which precedes the invagination 
led directly to the formation of the eyes, the light would then traverse 
the percipient elements from their outer to their deep ends; and it is 
probable that an ancestral eye of this kind prevailed. In the process of 
invagination, however, this thickened portion— from which are formed 
the retinal elements — is completely inverted, and as a consequence the 
