THE PHYLOGENY OF THE LEPIDOPTEROUS PUPA. 93 



ciently rigid to transmit to the base of the cocoon the thrust necessary 

 to open it, whilst too few would not give a good ladder for the 

 abdominal segments to climb by their vermicular action. 



It is almost universal that the pupa- skin is used as a ladder in this 

 way, still, there are exceptions. In the Cerurids, for instance, the 

 cocoon closely embraces the pupa, and the necessary thrust is taken by 

 the moth from the base of the cocoon ; in some Noctuids the pupa- 

 case collapses and the moth makes its escape by the vigorous action of 

 the hind legs. These, and other instances, are, however, all secondary 

 derivatives of further evolution. Similarly, the two other conditions 

 suggested as necessary to the establishment of the obtect pupa are not 

 found to be essential on further evolution. The cocoon, again, may 

 become dense, and have to be softened by special fluids, or cut by 

 special appliances. 



Amongst the pupa?-obtecta? there is here and there a tendency 

 shown to further solidification, but this always affects both the 5th 

 and 6th segments equally, and in certain Arctiids we have actually, or 

 very nearly, a solid pupa, as in Spilosoma. 



In one instance, that of the Nolidac, we find a pupa with only one 

 free segment. In this pupa, there are arrangements of the appendages 

 that prove that it is not derived from any ordinary obtect pupa ; the 

 egg also proves that this family is not related, not only to any families 

 with which it is usually associated, but even, at all directly, to any 

 Macro-lepidopterous family. 



Amongst the Micro-Lepidoptera many families have obtect pupa?, 

 and a few have completely solid pupa?. In some cases, as in some 

 species at present accepted as Gelechiids, this stage may have been 

 reached from an obtect form, but, in others, it seems likely that the 

 solid form has been attained more directly, and, as it were, by a short 

 cut. This seems likely to be the case in the Lyonctiidae (Cemiostoma, 

 &c). Whether others, such as Enicostoma, Bedellia, &c, will prove to 

 be more nearly related to each other, and to the Gelechiids referred to, 

 remains to be seen. Other Gelechiids may, or may not, be really 

 related to these, but they present a restriction of movement that may 

 be a step towards fixation. In these, only antero-posterior movement 

 is permitted. These Gelechiids are clearly close to the Depressariids 

 that are similarly limited. 



So far, we have treated pupa? as if they always had a cocoon of 

 some sort. This is almost correct with regard to the Lepidoptera- 

 Phalama?, but the butterflies form so large a class that they are more 

 than an exception. Originating from some incomplete pupal form, 

 they must, directly on reaching the obtect condition (in the Hespe- 

 riids), have abandoned cocoons and started on a fresh course of 

 evolution thus opened up to them. They were no longer affected by 

 the question of how to escape from a cocoon, and thus had no need to 

 preserve two free segments, and so, on different occasions, in each 

 family, and possibly for different reasons, these were successively lost, 

 and thus we find, in the higher members of each family, that the pupa 

 is solid and without movement. Our knowledge of butterfly pupae is 

 sufficiently full to enable one to he sure that the solid pupa? of Parnas- 

 sius, of Anthocaris, of several Lycaenids, of Danaids, and of some 

 Satyrids, are all of independent origin, and it is not improbable that 

 five is by no means the full number of such separate origins of solid 



