119 



What may be the purpose of the bristles, and upon what conditions 

 their diversity of form depends, I am not able to state, because the life- 

 history of these caterpillars is not sufficiently known, in fact upon this 

 point no observations have as yet been made. The shovel-formed 

 bristles are undoubtedly an inheritance from the ancestors of the Pa- 

 pilio caterpillars, and we find them in the larvae of quite different fami- 

 lies, as, for example, in Danais Archippus, Colias Philodice and Satyrus 

 Eurytus, all of which I have examined. If they have any meaning at 

 all in the Papilionidce it is only in the first stage, for they lose their 

 shovel-formed enlargement after the following moults, become propor- 

 tionately shorter, and finally quite obsolete. In brief, we conclude 

 that in the early stages we must regard them as rudimentary organs, 

 which, like the warts on which they stand, have lost their meaning for 

 the insect, and are replaced by other elements. 



In P. Philcnor, which constitutes in its development an aberrant 

 form, the bristles become rudimentary at a still earlier period, and in 

 the first stage we find on each wart for the most part only one long 

 bristle corresponding to the typical form. As the warts are developed 

 into the prominent horns already mentioned, the bristles are soon sup- 

 pressed. 



Let us now try whether my observations upon the few species of 

 Nymphalidce which have been at my disposal may be utilized in the 

 same manner as I have done with the Papilionidce. 



Here also we see that the early stages resemble each other more or 

 less completely; that in all the species examined of the genus Meliicca, 

 for example, these stages are quite in accordance. The young cater- 

 pillar is covered with hair, that is to say, provided with long bristles 

 which are arranged in the usual longitudinal rows. In contrast to the 

 Papilionidce, the brisdes stand singly on inconspicuous elevations of 

 the skin. But after the first moult a state of affairs is introduced such 

 as we find in the first stage of the Papilionidce-, namely, tall, conical 

 warts appear, beset with numerous brisdes, so that the caterpillar ap- 

 pears much more hairy than in the first stage. These warts continue 

 to increase with every new moult, so that they lose nothing in circum- 

 ference in proportion to the size of the caterpillar. The course of 

 development, therefore, is exactly in inverse ratio to that seen in the 

 Papilionidce. In that family the warts were present in the first stages 

 only as a heritage from a supposed prototype, and from that point 

 continually diminished. In the Nymphalidce they are a newly-acquired 

 character which does not appear until the second stage, and which in 

 the following stages is maintained or still further perfected. 



In the Papilionidce the coloring and markings take the precedence, 

 and suppress the warts and bristles; in the family we are now consider 

 ing the first named elements play a more subordinate part, and in their 

 place the hairy covering acquires an important meaning for the species 



