369 



the ribs themselves coming out as broad, hght bands against 

 this darker background. The whole arrangement shows a 

 striking similarity to the wing-sheath pattern of E. cardamines, 

 the onlv difference being the absence in the latter of the series 

 of light intra-marginal spots. 



According to my opinion, the primitive nature of this inter- 

 pupal wing-colour pattern is proved b\' the following character- 

 istics : 



1. It is the same on upper and underside, and on fore- and 

 hind wings. 



2. It follows strictly the arrangement of the wing-nervures. 



3. It is very simple in its constitution. 



4. It contains those elements of the definite pattern which 

 are common to the different members of the Vanessa family. 



5. It shows a great and fundamental agreement with the 

 pattern of the pupal forewing sheath. 



The sudden change into the final or imaginai pattern must 

 be regarded as a phenomenon of extreme abbreviation of this 

 part of the phylogenetic development, perfectly comparable 

 to the unforeshadowed change of the full-grown caterpillar 

 into the imago-like chrysalis. In the ancestry of our Rhopa- 

 locera a period must have existed when butterfly-like insects 

 possessed semitransparent wings, showing a colour-pattern 

 similar to that of the pupal wing-sheath of to-day or to that of 

 the primitive intra-pupal wing. Whether these ancestral wings 

 were already covered with scales, or at least with hairs, or only 

 M'ith a more or less sculptured cuticula, and whether the colouring 

 matter had already penetrated into the chitinous structures or 

 was still lingering in the hypodermic cells, must be left undecided. 



2. Other VancssidcB (fig. i). 



Besides T'. nrticœ, I in\-estigated the wing development in 

 the pupie of V. io and Pvramcis cardai, and found that in these 

 three l'aacssidœ an identical primitive pattern arises in one 

 and the same manner, especially witli regard to the series of 

 light intra-marginal spots. The division of the forewing in a 

 lighter and a darker area, however, is much less marked in 

 V. io, and hardly if at all visible in /'. cardiii. 

 •i7 



