THE DEVELOPMENT OF THE COMPOUND EVE. 541 
characteristic of the Ametabola. In many insects, as the 
noctuid and, crepuscularian Moths (Lepidoptera heterocera), the 
retina is segregate—that is, groups of ommatea have distinct 
nerves, just as in the larval Neuroptera (Libellula). In the 
Muscide and the imago of the Libellulida the retina is always 
continuous, and a single thick decussating nerve supplies the 
whole eye. In the Gnats and Tipulide (Diptera nematocera) 
the retin are always, so far as I know, segregate, and the 
compound eyes come into functional activity in the active 
nymphs in the aquatic species. The several stages in the 
development of the continuous retina of the Dragon-flies 
(Libellula) are represented in Fig. 70. 
Historical and Critical—Recent writers on the development 
of the compound eye in Arthropods have almost entirely 
devoted their attention to an attempt to form a plan or 
scheme which shall serve as a morphological type to bring 
into accord the various kinds of compound eye, which they 
believe exist in Arthropods. 
The statements made are chiefly supported by little evidence 
and much argument ; some, as Carriére [232] and Parker [250], 
maintain that the whole structure originates in a simple 
thickening of the epidermal layers, and there is no doubt, as 
far as the dioptron is concerned, this statement is partly 
correct. Others, as Bobretzki [253], Reichenbach [255], 
Kingsley [256], and Patten [239], state that it is formed as an 
invagination of the epidermic layer. That the eye is situated 
in many Arthropods, both Crustacea and Insecta, either per- 
manently or temporarily in an invagination of the integument 
is indubitable—compare for example the eye in Apus (Bernard 
[257]), and the invaginated eye-disc of the dipterous larva—so 
that, in some Arthropods at least, the truth is apparently on 
the side of the invagination theory. All these observers, how- 
ever, think that the eye is developed from the whole invagina- 
tion, which becomes, according to them, flattened out, so that 
three layers are subsequently formed, and they are all at 
variance as to what becomes of these layers. According to 
Reichenbach, the cavity of the vesicle disappears at an early 
