92 BRITISH LEPIDOPTERA. 



wards became free. It does not follow, however, that a family with 

 three free segments is higher than one with four, as they may be 

 evolved in different stirpes ; all that is certain is that the one with 

 four free segments is not derived from the one with three ; the latter 

 being the higher, qua pupa, proves that if one is derived from the 

 other, it is the derivative, but it is not proved that there is any 

 derivation at all. 



It is an almost universal rule that the pupa-incompleta does not 

 completely emerge from the cocoon, but is retained, as it were, half- 

 emerged, usually by the elasticity of the cocoon seizing the softer 

 abdominal segments and by other devices ; assisted always, no doubt, by 

 the imago, finding the thorax and appendages free of the cocoon, at 

 once directing its efforts to quitting the pupa-case, and not to further 

 extrusion of the pupa from the cocoon. One of the means of checking 

 the pupa at the right degree of emergence is a cremastral cable extend- 

 ing from the bottom of the cocoon to the cremastral hook of the pupa, 

 and sufficiently loose to tighten at the right point. This is met with 

 in some Tortricids. It is of course important that the pupa-case 

 should be retained by the cocoon, or the moth would have difficulty in 

 getting rid of it. A pupa deprived of its cocoon often produces a 

 cripple for this reason. The cremastral cable is interesting as suggest- 

 ing one possible way by which the next great advance in pupal evolution 

 took place. This is the step from the pupa-incompleta to the pupa- 

 obtecta. 



The pupa-obtecta remains in the cocoon, the imago alone emerges. 

 It has solved the problem that we may picture as having been set before 

 it from the first, of how to emerge from the cocoon without imaginal 

 jaws. Unlike the Coleoptera and the Hymenoptera, it leaves the 

 cocoon at the same time as it casts the pupal skin, and performs its 

 expansion of wings, hardening of cuticle, etc., outside ; but, in respect 

 of leaving the pupa- skin in the cocoon, it has got back to their more 

 primitive type. 



The obtect stage was probably reached by several different lines of 

 advance from the pupaB-incomplette, and we may suppose it to have 

 begun in some measure accidentally, the imago coming out of the pupa 

 before emerging from the cocoon, possibly by the cremastral cable having 

 been too short, or from some other accident, the cocoon being 

 sufficiently loose or valvular to allow of the moth escaping. We have, 

 in Epermenia, preserved a transition stage. In this genus, the free 

 7th abdominal segment in the male pupa makes it belong to the 

 Incompletae, its non-emergence from the cocoon shows that it has 

 acquired the habit of the Obtectae. 



It is clear that the obtect habit might arise at any time from the 

 pupa-incompleta after it had made some progress. Three conditions 

 were necessary : (1) A cocoon flimsy or valvular enough for the 

 moth to break through. (2) A pupal head and thorax sufficiently 

 rigid to be held back by the opening in the cocoon (or a cremastral 

 cable) . (3) A pupal skin altogether rigid enough to form a base from 

 which the moth could force its way out. 



Once established, the obtect habit so enforced this last condition 

 that, with hardly an exception, the obtect pupa has only two free 

 segments, and, in this respect, is identical in both sexes. Too large a 

 number of free segments would obviously make the pupa-case insuffi- 



