748 ON THE VALIDITY OF SOME FORMS OF MIMICRY. 



accustoiixed to breed on a hedge of Madras thorn in an isolated 

 povsition in my garden ; the i-esults of such counts showed 73 wet 

 forms, 1 intermediate, and 19 dry, and it was not vintil January 

 28th that the wet forms were entirely replaced. 



The butterfly takes exactly a fortnight to pass from larva to 

 imago, and thus we have approximatel}^ three or four broods of 

 wet season forms produced under diy season conditions. We 

 must assume that as the butterfly is seasonally dimorphic it stands 

 in need of protection, yet so far as I covikl ascertain the wet forms 

 suffered no diminution though exposed to what were, or should 

 have been, adverse circumstances. 



The butterflies rested during the night and in cloudy weather 

 on the under surfaces of the leaves of Vincta sp. ?, a small flowering 

 shrub with pink flowers and small oval green or frequently yellow 

 faded leaves. It often collected gi-egariously, two or even three 

 being on the same leaf and perhaps ten on the same plant. The 

 position was an admirable one for pi-otection from the wet, and 

 also from small pi-edatory foes which seldom look upwards when 

 hunting for prey. The appearance of the plant is the same 

 throughout the year, and the butterfly derived no advantage from 

 its change from one seasonal foi-m to the othei-. 



The following experiment makes me still further doubtful of 

 these effects being due to natural selection. 



If we take the pupa of a somewhat similarly coloured butterfly 

 which is not seasonally dimorphic, such as Papilio demodocus or 

 Fapilio demoleus, and expose it to a hot dry temperature, we can 

 produce an insect with much of the yellow on the under surface 

 replaced by red. I am almost persuaded that these rusty red 

 spots ai-e a vestige of a charactei- at one time common to certain 

 Pierines and Papilionines which is more or less reproduced by 

 heat and dryness, if of sufficient duration and intensity under 

 natural conditions in the Pierines, but in the Papilionines only 

 under artificial stimulation of a like but exaggerated character. 



Conclusions. 



1 . It has been shown that in Bourbon and Mauritius there are 

 no butterfly- eating birds or reptiles ; so that the cases of mimicry 

 occurring there cannot be due to their influence. 



2. In Ceylon it has been conclusively shown that the butterfly- 

 eating reptiles are impai'tial feedei'S. 



3. That a trained obsei'ver can distinguish the majority of 

 these mimetic butterflies at a distance of about twenty or thirty 

 feet and frequently at the same number of yards ; and this being 

 so, it is cei'tain that a bii'd which has to depend for its existence 

 on its powers of observation, could after a few failures be able to 

 discriminate them at the same and probably at a considerable 

 greater distance. 



4. That Drongos feed largely upon Euplojas, and this being so, 



