Appendix. 539 



first segment, which, in most of the other species, are 

 completely absent ; but this does not justify their 

 separation, since the head spines of Heliconius, Eiuides, 

 and Colccnis Dido, which are of a considerable length, 

 are shorter than those of the next segment in Colcenis 

 Julia, and Dione Vanilla, and in Dione Juno they 

 dwindle down to two minute points, this last species also 

 bearing a short pair on the first segment. The larva of 

 Dione Juno is thus as closely related to that of Acrcea 

 Tltalia as it is to that of its congener Dione Vanilla. 



" If it were desired to form two distinct larval groups 

 this could not be effected on the basis of their differ- 

 ences in form, but could only be based on their food- 

 plants. The larvae of Heliconius, Eueides, Colcenis, and 

 Dione live on species of Maracujd (Passiflora) ; those of 

 Acrcea Thalia and A/alia on Compositae (Mikania and 

 Veronia). These larval groups would agree with those 

 founded on the form-relationships of the imagines, but 

 unlike the imaginal groups, which can be formed into 

 families, they would scarcely possess a generic value. 



" If we arrange the single species of caterpillars accord- 

 ing to their resemblances, this arrangement docs not 

 agree with that based on the resemblances of the 

 imagines, even if we disregard the different values of the 

 groups. The result is somewhat as follows : 



IMAGINES. 



(Nymphalideoui butterflies with tufts on 



(FmmilieO MARACUjX-GROUP. 



(Genera.) Heliconius. Eueides. Colarnis. Dione. Acraca. 



| , * s . * , | 



(Species.) Eturate. Isabella. Dido. Julia. Vanilla. Juno. Tkalia. 



I I I I I I 



LARV^i. 



* * 



