STUDIES OF THE COLLEMBOLAN EYE. 
BY J. E. GUTHRIE. 
Plate XVIII. 
In any primitive group of animals, there is an unusual interest 
attached to any organ which shows a simple structure, as having a pos- 
sible bearing upon the history of the organ in a more highly specialized 
condition, as found in higher, closely related groups. Sometimes organs 
of seemingly simple structure are very puzzling from the fact that we 
are at a loss to determine whether their condition is primitive, or is due 
to degradation or partial atrophy. Embryological studies are often of 
value in determining the case, but not always. In entomology we have 
surprisingly few embryological studies which are specific enough to guide 
us in such determinations. When a group of insects varies widely how- 
ever, in relation to any specified structure, a comparative study of its 
adult condition in the different members of the group may be of value. 
In the minute, wingless insects of the order (or sub-order) Collem- 
bola, often included, along with the true Thysanura, under the ordinal 
name of Thysanura as constituting a lowly branch upon the insectan 
phylogenic tree, we find no compound eyes, nor do we generally find any 
isolated ocelli. The ocelli are bunched or gathered together in the post- 
antennal regions on the right and left eye-spots. 
It is not impossible, I think, that each eye-spot may represent all that 
remains of an ancient compound eye, the few ocelli being modified sur- 
vivors of some of the component ommatidia. Embryological research 
must be looked to for the determination of this question if it ever is 
determined. If this should prove to be true, we must regard the com- 
pound eye as having undergone a certain very definite amount of decad- 
ence at a period remote enough to antedate the general branching off of 
species. A modern Thysanuran, MacMlis, shows many characters that 
must have been likewise possessed by the ancestor of the Collembola. 
MacMlis, moreover, has well-developed compound eyes; so that it seems 
not unlikely that the Collembolan eye has descended from that type. 
The present paper, however, has to do rather with the modifications 
of the eye from the Collembolan type, than with its possible pre-Collem- 
bolan history. 
The eyespots are situated dorso-laterally, caudad to the bases of the 
antennae, as may be seen in Figures 1 to 4. The first two show a typical 
horizontal-headed species; the following two a species with the head 
vertical. Each eyespot may contain eight or fewer ocelli, but never more 
than that number. There are abundant indications that when fewer 
than eight ocelli are present, the condition denotes atrophy of one or 
more of the oceller units, for eight seems to be the- typical number of 
ocelli for the Collembolan eye. Some whole genera, it is true, show a 
( 239 ) 
