166 PROCEEDINGS OF THE ACADEMY OF [Jan., 



preferred to this method of experiment, yet in the absence of such 

 observation much may be learned by comparing the behavior of the 

 same individual bird with different species of insects. 



The indirect evidence that Ph. philenor acts as a model and pos- 

 sesses the qualities of a model seems to me extremely strong. On 

 this hjT)othesis many facts receive their interpretation; without it 

 they are unexplained and meaningless. Philenor is one of the 

 " Aristolochia swallowtails," a section which is abundantly repre- 

 sented in tropical America and in the Orientaf Region, but, with the 

 exception of Ph. anterior in Madagascar, absent from the Ethiopian 

 Region. The mimicry we observe in North America is not only 

 repeated in both Regions where these swallowtails are abundant, 

 but repeated in a more convincing manner, because the patterns 

 are often far more elaborate, and because an " Aristolochia swallow- 

 tail" may break up into numerous geographical races with distinctly 

 difterent patterns which are mimicked in each locality by correspond- 

 ing races of the " Fluted swallo^vtails " and, in the Neotropical Region, 

 of the "Kite swallowtails." A good example is the Oriental Ph. 

 aristolochice with its subspecies mimicked by the females of Pap. 

 poJytes. Furthermore, there is in this case experimental evidence 

 that aristolochice is distasteful, and its slower, more flaunting flight 

 has often been remarked upon. In the Oriental Region species of 

 Pharmacophagus are also sometimes mimicked by day-flying moths, 

 and, in the Neotropical Region, not only by these, but by "Kite 

 swallowtails" (Cosmodesmus) and Pierines. Throughout the whole 

 range, as in North America, the mimicking "Fluted swallowtails" 

 are as a rule females, while on the other hand the "Kite swallowtails " 

 are mimetic in both sexes (33) . Just as the other much-mimicked 

 groups — the Danaince, Ithomiince, Heliconince, and Acrceince — are 

 themselves specially subject to mimicry — the genera or sections of 

 the same subfamily superficially resembling each other and also 

 resembling those of the other subfamilies — so is it in both respects 

 with the South American " Aristolochia swall(n\i;ails. " In every 

 way these butterflies behave like the great distasteful groups supply- 

 ing the best known models for mimicry. If we had no experimental 

 or other evidence that the Danaince are unpalatable, the indirect 

 evidence is strong enough to warrant at any rate a provisional accept- 

 ance of the hypothesis that they possess some peculiar means of 

 defence which renders them specially advantageous as models. For 

 wherever they are indigenous in the Old World they are mimicked 

 by butterflies of other groups, and even in North America, where 



