24 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



in .4. seJiatoria. except the horns, l)lend with the bhick tints, except tliose in the ocherousi 

 infnispiracular line. These black and ocht'r lines appear in stage IV. the black ones arriving in 

 stage II. In stage III the larva is still olixe green, with yellowish dorsal -and lateral lines. The 

 head in this species is persistently black from the tirst to the last stage. It is tinally to be 

 observed that in this genus color and ))i-ight stripes are of more importance than spines, whereas 

 in Adelocephala the armature is emphasized and the markings are a quite subordinate feature. 



Ci)l<irati<m in A(MocephaJ(i.—\x\ the fully grown caterpillars the body is grass or pea green, 

 with usually no gay markings, and the spines bristle with large stout sal)er-like spines, those of 

 the abdominal segments being in certain species nearly as large as those on the two hinder 

 thoracic segments. Hence they are admirably protected from attack, since the horns are both 

 defensive, and also harmonize both in shape and color with the spines of their food plants. 



Not so, however, with their ancestors, which, judging by the life history of A. hlroJor. espe- 

 cially its markings in stage I. were, liesides wearing long. knoVjbed, reddish horns, gaily and 

 conspicuously striped. A. hieolor on hatching is green, with three lateral white stripes and a 

 narrow thread-like red dorsal line ending on the red caudal horn. The head is black, while the 

 eight thoracic horns are all red. or reddish, including the suliglobular ends. As regards the 

 remarkable distinctness of these stripes, we infer that if we take into account the coloration alone 

 Adelocephala is a more primitive form than any other of its subfamily. Its ancestors mu.st have 

 been so marketl at larval maturity, and here we have a very clear example of the crowding back 

 to the tirst stage of the ontogeny of featui'cs characterizing the later and last one. and its life 

 history is a partial recapitulation of that of its ancestors. The horns in the earlier stages, when 

 the larv.T? may V)e huddled together on the underside of a leaf, ai'e of no use to them, and in the 

 tinal stages they tend more and more to be reduced and to lose their brilliant red coloring, though 

 in the mature ancestral caterpillars both the distinct lines, the threatening' forest of long banner- 

 like horns at the front and one at the end of the body must have rendered them inedible, or at 

 least unattractive to insecti\orous animals. 



The ancestral caterpillar must have lived exposed to the sunlight, as the green stripes are 

 due to the greater abundance of chlorophyll, the white to a deticiency of pigment, such lines 

 being the eii'ects of light and shade and very common among lepidopterous larva\ 



In the second stage the head is reddish, and the spines still red, the small ones alongf the abdo- 

 minal region yellowish. A new colorational feature is a narrow, thread-like red line comiecting 

 the spiracles, becoming yellow on tlie edge of the suranal plate. 



In stage III there are the same tints, but the lower side of the red lateral line is shaded 

 whitish yellow, while the line on the ninth segment and along the edge of the suranal plate is all 

 yellow. This line in front passes up to the base of the second pair of thoracic horns, so as to give 

 the appearance of a continuous red line, this being a concealment feature. 



In stage IV the red portion of this line becomes purplish above, probably due to the reflection 

 of the purplish parts of the leaves. Now all the spines have lost their red color and have become 

 j-ellowish, hence nmch less conspicuous than before, and in the last stage the body is plain green, 

 the shades given otf from the pale yellowish horns and silver dorsal spines, as well as the purplish 

 red and white of the lateral strij^e blending with the hues of the leaves and spines of the honey 

 locust. It should also be observed that the yellow edge of the suranal plate shows from aboxe. 



It may be concluded that all the colors and hues, as well as the spines of the beautiful larva 

 of A. hic(iloi\ naturally blend with or match the hues of the leaves and spines of its food plant. 



Certain of the Brazilian forms have apparently no red lines. In one we po.ssess (PI. Ill, tig. 1) 

 the body and spines are uniformly grass green. In A. aiyymccnf/ia, figured by Burmeister (our 

 PI. XLIX, tig. 1). the body is uniformly green, and there are no red lines or spots, while the spines 

 are yellowi.sh. 



In the Brazilian ^1. trard/ the body is g-recn. but the larger dorsal spines are very brilliantly 

 silvery, the smaller ones violet, and on the side of the body are three conspicuous, very large, 

 wide white bands, and the head and thoracic legs are also white. The larva> of ^4. !<ul>angulata 

 and h't'vis have a bluish head and lateral stripe. 



