METAMORPHOSIS IN LEPIDOPTERA. IS 



died. Those that moulted or did so partially exhibited peculiar charac- 

 teristics, such as — the loss of the hooks of the prolegs, a structural 

 change in the true legs, an alteration of the maxillae, the development 

 of a short haustellum, and the two normal palpal processes much 

 lengthened. The antennas were the organs most altered, the basal 

 joint large and bladder-like, whilst the remainder was from 1mm. -3mm. 

 or more in length, in one considerably longer, folded together, as the 

 pupal antenna is, within the larval head. Before moulting it became 

 brown and transversely ribbed, just like the pupal antenna, but had a 

 soft white terminal joint with a terminal bristle. . . . The eyes 

 were more or less altered. In nearly all, the six eye-spots could be 

 observed, but the two upper were usually smaller or disappearing. 

 They had amongst them yellow, raised, clearly defined patches, whilst 

 in several of the larvae, a crescentic slightly raised mark lay between 

 the eye-spots and the base of the antenna. This was clearly the 

 crescentic mark of the glazed eye, both from its form and its position ; 

 the latter not at first recognisable, until it was remembered that the 

 face becomes bent forwards in the pupa and the antenna thrown back- 

 wards. . . . The larva?, though possessing normal jaws, did not 

 attempt to eat . . . and exhibited to a great extent in the characters 

 indicated, some development towards the pupal stage. Chapman con- 

 siders that the condition of the glazed eye shows clearly that the 

 crescent is an appendage to the eye, and not a part of the eye itself, 

 since it is distinctly separate from the larval eyes just as in the pupa, 

 and it is outside the area which often shows indication of the hexagons 

 of the imaginal eye, and beneath which that commences its develop- 

 ment. The observer suggests whether the diminished nutrition was, 

 in this experiment, not commenced before a certain amount of change 

 towards the pupal stage had occurred, and that this could not be 

 undone. He thinks that had the diminution of nutrition been begun 

 earlier (i.e., before these changes had sufficiently advanced to be of 

 importance) it is quite possible that the further larval instar would 

 have been of an entirely larval character. 



The normal number of moults in the Acronyctids (taken as a 

 whole) is five, although a number of species do, occasionally, reach the 

 last (sixth) instar in four moults, by omitting the fifth, and when, as 

 in Pharetra rumicis, that instar has a special distinctive marking, or 

 coloration, or arrangement of hairs, these larvae never exhibit that 

 particular phase. In P. rumicis this is by no means uncommon, most 

 broods presenting some examples of it ; it has also been noted in 

 Pharetra menyanthidis, P. auricoma, Acronicta leporina, and Apatela 

 aceris, and no doubt it occurs, if more rarely, in all the other British 

 species. This variation has no relation to sex, and Chapman found that 

 it did not represent an attempt of the larva to press forward so as to 

 become double-brooded, although P. rumicis (next to P. auricoma) most 

 frequently produces autumnal specimens that probably represent a 

 double brood, nor did he observe any difference in size in the full 

 grown larva; or imagines. It seems to be a spontaneous variation, the 

 meaning and use of which have yet to be discovered. Jocheaera alni 

 alone, of the British species, has four moults as the normal number, 

 but, in rare instances, it moults five times like the others, and when it 

 does so, the larva, in the extra penultimate instar, differs from any of 

 those preceding it, and shows a transition between the juvenile and 



