elements, therefore, which by the side of these have lost their meaning-, 

 or would have a disturbing effect on the development of the new char- 

 acter, are gradually compelled to yield. The particular kind of mark- 

 ings will also have a part to perform; thus, on a uniformly brown or 

 g-reen color, such as the nearly related Papilio Turnus and Troilus 

 display, the protuberances project very prominently, and on the other 

 hand they disappear almost entirely after the seconcl moult. The warts 

 and bristles would exert a still more disturbing influence on the pecu- 

 liar markings of Papilio Ajax with their numerous black rings; hence 

 we may account for the remarkably sudden change from the second to 

 the third stage. All at once the markings make their appearance, and 

 at the same time the previously strongly developed warts and bristles 

 vanish. I have already mentioned that I possess a preparation in 

 which the second skin is just becoming detached, and under it the 

 characteristic form and color of the third stage are plainly visible. 



With such markings as, for example, Papilio Asterias, Brevicauda, 

 Machaou (and Alexanor) display, the warts will on the contrary be 

 much less disturbing, as they scarcely project on the dark, oblique 

 bands. Accordingly we find that they do not entirely disappear until 

 the last moult. 



We may then probably infer that the larvae of the Papilionidce de- 

 scend from forms which, with dull coloring and inconspicuous mark- 

 ings, were covered by strong, bristle-bearing warts. All the caterpil- 

 lars examined by me in the first and even in the second stage, conform 

 to this hypothetical prototype. 



The larvae of the PapilionidcE now existing show for the most part 

 brilliant colors and conspicuous markings (rings, bands, ocelli, etc.), 

 and the original forms were forced to yield to these elements. Many 

 intermediate steps have bridged over the great gap between that pro- 

 totype and the larvae of our own period, and these steps are perpetuated 

 to us in the various forms of the successive stages of development. 

 Every caterpillar repeats during its moults, in a more or less perfect 

 manner, the history of its descent. 



I have heretofore attempted to prove that the kind of markings 

 sometimes shows the necessity of a very sudden disappearance of the 

 rudimentary organs. Were this not the case we could very easily de- 

 termine from the rudiments of the warts the age of Papilio larvae, that 

 is to say, the remoteness of each species from its ancestral form. Those 

 caterpillars, namely, which kept the warts longest would be the oldest, 

 standing nearest to the prototype; those which lose the warts in the 

 early stages would be the most recent, furthest removed from the pro- 

 totype. In the first category, therefore, we should include, for ex- 

 ample, Papilio Machaon, Asterias and Brevicatida; and this seems not 

 unreasonable, because in them the dark color remains to the third 

 stage without decided markings. Brevicauda would appear still older 



