66 
G. H. Parker 
somewhat like that in Branchipus^ the optic ganglia and retina in 
-crustaceans have grown in complexity band in band tili a condition 
like that in Astacus was reached. This mutuai growth is quite 
natural when one reflects on the intimate physiological dependence of 
the retina and the ganglia. 
Aside from the question of the relation of ommatidia to one 
-another is that concerning their origin. This topic has already been 
discussed (cf. Parker, 91, pag. 118), and I recur to it only to add 
some fresh evidence. This evidence bears directly upon the opinion 
expressed by Watase (90) that each ommatidium is an elongated 
ectodermic Involution whose cavity has been obliterated by the ap- 
proximation of its walls, the cells of which, however, still retain 
their power of secreting chitinous substance. The secretion of the 
deeper cells forms the rhabdome, that of cells nearer the surface 
gives rise to the cone, and the superficial cells produce the facet; 
.in one sense, then, these three structures, rhabdome, cone, and 
.facet, are homologous. I bave already advanced arguments against 
this theory (cf. Parker, 91, pag. 128) : the arrangement of the cells 
in different ommatidia does not favor this idea; the small number of 
cells in the more primitive forms of ommatidia makes Involution 
mechanically difficult ; there are no transitional forms between those 
simple ocelli that are produced by Involution, and ommatidia; and, 
finally, there is absolutely no embryological evidence that omma- 
tidia are formed by Involution. To these objections I must now 
add the further one that the composition of the rhabdome as por- 
trayed in the preceding pages does not favor the idea that this 
structure is in any wise a secreted body, but rather that it is a 
living product of the retinular cell in much the same sense that 
muscle substance is the product of a muscle cell. These objections 
seem to me to afford sufficient ground for abandoning the theory 
that ommatidia bave arisen by Involution. They offer no difficulty, 
however, to the idea already expressed (Parker, 91, pag. 130) that 
the simplest type of ommatidium has been derived from a 
Cluster of cells in a continuous unfolded epithelium and that by a 
process of ceir division and differentiation the simpler type of 
ommatidium gave rise to the more complex. 
