58 



the imagines show that each species as a whole has continued 

 its progress. Before leaving the larva I must mention that 

 if we admit of a separate kind of evolution for each stage of 

 lepidopterous life we must also admit of an evolution for 

 each stage of the larval life, as the habits and environment 

 may differ in the different stages. We can see this does 

 occur if we look closely enough. It is, however, more 

 marked in some species than in others ; for instance, Papilio 

 machaon, Dicraimra vinula, and Phisia moneta are familiar 

 examples of larvae differing in different instars. Bucculatrix 

 ulmella is at first a leaf-miner, and later the larva with an 

 altered appearance feeds exposed. But perhaps the best 

 instance of all is that of Phyllocnistis siiffusella, which in three 

 of its instars is so unlike that the uninitiated would take the 

 larva, in these three stages for three quite distinct organisms. 

 We have not quite done with the larva yet, for though I 

 believe it undergoes some change, it is still a larva while 

 spinning the cocoon. This change is quite plain in Phylloc- 

 nistis, as when the larva is full fed it undergoes an ecdysis 

 and is completely changed in appearance. It does not feed 

 at all in this instar, and its sole function is to spin its cocoon. 



The lepidopterous cocoon may be a tough structure, like 

 that of Lasiocampa quercus, or open network, like that of 

 Pkitella inaciilipcnnis, or simply a boss of silk, like that of 

 Vanessa io. In every case its form has been, no doubt, 

 brought about by evolution, and must be taken into account 

 when treating of the life-history of any species. 



As regards pupal development I can say but little, as I 

 have only made very few observations on that stage of 

 lepidopterous life. So long as the pupa is enclosed, whether 

 it be in an earthen cell, a dense silken or an open network 

 cocoon, it seems to retain, more or less, what may be called 

 the normal pupal shape, such as that of the Noctuid, Tineid, 

 or even those oiHesperia nialvce and Parnassius apollo; though, 

 of course, all of these differ vastly in many particulars. 

 When, however, the pupa dispenses with a cocoon and 

 becomes exposed, it develops various protuberances and 

 loses the normal shape, as we see in most butterflies, in the 

 geometrid genus Ephyra, and in the genera Elachista and 

 Bedellia among the Tineina. It seems that these various 

 protuberances and points which we see on the pupa of the 

 Pierids, Nymphalids, especially on that of Limenitis sibylla 

 and others, are special developments of the pupa. I have 

 observed that in Pieris brassicce these prominences are not 

 fully developed until some minutes after the larval skin has 



