22 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



Prothoracic segment. — With ii large subspherical tubercle on each side bearing uuuierous 

 radiating hairs (Lasiocampida' of first stage) or pencils of hairs (Parorgyia); two autlers {H. 

 (/uttivittn, hiinidatd ;iiid H. ohlujiia). 



Second thoracic segment. — Two high slender spines. First stage of Aiti-sotd .•ienatoyia, A. stigma, 

 and Dryocampa rubicunda. 



Third thoraeir segment. — Two spinulose pappose flaps, Empretia stimuha. 



First, second, and third thoracic segments. — Each with a pair of high spines, Ciiheronia regal is 

 and Fades imperialis. 



Second and third thoracic segments. — Each with a pair of long horns, Sphingicampa hieolor. 



First and third thoracic segments. — In Stage I of the European Aglia tint (I'oulton). 



JPir.s/ al>dominal segment. — Movable tubercle in Schizura and Xyliuodes. 



Eighth alidominaJ segment. — The caudal horn of Sesia and most Sphingidie, Theosia, and 

 Endroniis, BomJ>yx mori, and other species — Sphingicampa, Fades, Citlieronia, and Aglia tan 

 (Stage I). 



So far as I am aware no one has suggested why these horns and high tubercles, and often 

 pencils of hairs, are restricted to these particular segments. As a partial explanation of the reason 

 it may be stated that the presence of tliese high tubercles, etc., is correlated with the absence of 

 abdominal legs on the segments bearing the former. It will also be noticed that in walking the 

 apodous segments of the caterpillar are more elevated and prominent than those to which the legs 

 are apjiended. They tend to bend or hump up, particularly the tii-st and the eighth abdominal, 

 the ninth segment being reduced to a minimum, and the tenth simply represented by the suranul 

 and paraiml plates, together with the last pair of legs. 



As is well known, the loopers or geometrid worms, while walking, elevate or bend up the part 

 of the body situated between the last tlioracic and lirst pair of abdominal legs, whicli are appended 

 to the seventh uromere. Now, in the larva of Nematocampa filamentaria, which bears two pairs of 

 remarkable filamental tubercles rolled \\\^ at the end, it is certainly very suggestive that these are 

 situated on top of the loop made by the caterpillar's body duiing progression, the lirst pair arising 

 from the second and the hinder pair from the fourth abdominal segment. 



It seems, therefore, that the humps or horns arise from the most ])rominent ])ortions of the, 

 body, at the p. tint where the body is most exposed to external stimuli; and the force of this is 

 especially seen in the conspicuous position of those tubercles which are voluntarily made to nod or 

 so move as to frighten away other creatures. Perhaps the tendency of these segments to loop or 

 hump up has had a relation of cause ami effect in inducing the h}i)ertropliy of the dermal tissues 

 entering into the composition of the tubercles or horns. 



Analogous i)ositions are in the vertebrates utilized, as in spiny, osseous fishes, or the sharks, 

 the horned Amphibia, or horned reptiles and horned mammals. The prominence of the foundation 

 ])arts, from which the tubercles arise, may lead to a determination of the blood toward such places, 

 and thus in well fed or overfed (possibly underfed) individuals induce a tendency to hyi)ertropliy, 

 which once set up in early generations led to the i)roduction of incipient humps which became more 

 developed as they proved useful and became preserveil in this or that form by natural selection. 

 On the other hand, the hypertrophy of certain piliferous warts would tend to cause an arrest of 

 development or a tendency to atrophy in the piliferous warts of adjoining segments. And in like 

 manner may the simple seta^ havci become liypertro])liied on account of their great utility as 

 deterrent organs, and become wonderfully modified in this and that direction in such aiid such 

 forms, until they became in recent geological times the common and normal inlieritance not only 

 of scattered species but of certain genera in scattered families, and even of entire families. 



It is to be observed, as one will see by referring to the sju'cial larval histories and the lecapit- 

 ulations which we have appended, that in the species of Schizura the evolution or hypertrophy of 

 the movable or nutant tul)ercles begins in the third stage at about the time when the young 

 caterpillars leave their common bn-thplace on the underside of the leaf and seek more conspicuous 

 feeding grounds on the outer edge or on the upper side of the leaf, where they are exposed to 

 the visits of ichneamons, or Tachina', or carnivorous Ilemiptera, or to the onset of open-mouthed 

 insectivorous birds. At the same time arise the bright colors, spots, and stripes, the very peculiar 

 V-shaped silver or yellowish white mark characteristic of the species of Schizura — these are per- 



