MEIMOIRS OF THE NATIONAL ACADEMY OP SCIENCES. 1 1 



II.— HINTS ON THK EVOLUTION OF THE BRISTLES, SPINES, AND TUBERCLES OF NOTODONTIAN 



AND OTHER CATERPILLARS.' 



It i.s not improbable, that, as a rule, all caterpillars at first lived on grasses, lierbaceous ami 

 low-growing- plants generally, and that gradnally they began to climb trees, as the latter became 

 <levelopcd, and in time became adapted to an arboreal station. As is well known, no decidnou.s 

 trees or flowering jjlants appeared in such numbers as to form genuine forests before the Cretaceous 

 lieriod, and about that time in geological history began to ap])ear the kinds of insects which visit 

 flowers and trees that blossom. 



The species of the great lepidopterous family Noctuiibe, of which we have in the L'uited States 

 alone over a thousand species, are, as a rule, low feeders. Certain species of jNIaniestra and of 

 Agrotis, ordinarily feeding on grasses ami low herl)s, will however, especially early in the spring, 

 ascend trees and shrubs of diti'erent kinds and temporarily feed upon the buds; and in summer 

 a species of Jlaniestra will ascend currant bushes in the night and cut oft" the young, fresh shoots. 



In the grouj) of forms represented by Catocala, Tl(>m()ptera,and Pheocyma we have true 

 tree inhabiting cateri)illars, and, like the Notodontians and dendricolous (leometrids, their bodies 

 difter lemarkablj'from those of the low feeders, being variously spotted and mottled with shadc^s 

 of brown and ash, to assimilate them to the color of the bark of the tree they rest upon, and are, 

 besides, provided with dorsal and lateral humps and warts, to further assnnihue them, in outline 

 as well as in color, to the knots and leaf-scales on the smaller branches and on the twigs among 

 which they feed. And then there is the small groui) of Noctuobombyces, represented by species 

 of Apatela, I'latycerura, Raphia, Charadra, and their allies, which closely ''mimic'" the hairj', 

 penciled, or spiny arboreal Bombyces.- It should, however, be observed that this is scarcely a 

 case of mimicry, but rather of adaptation; the presence of hairs, pencils, spines, and bi'istles being 

 apparently due to the caterpillars having changed their environment from herbs to trees, and being 

 subjected to the same conditions as the Bombyces themselves.' 



In the exclusively low feeding caterpillars of certain groups of butterflies the body is usually 

 smooth and adorned with lines and spots, while the general feeders and many arboreal forms are 

 often variously sjiined and tnbcrculated, yet many s])ined caterpillars of butterflies feed on low 

 herbs. ^ The Sphingid;e in pait feed on low plants and in part on trees, and they do not, except as 

 regards the caudal horn, exemplify our thesis. 



'This section Is reprinted with some alterations from an article in the Proceedings of the Boston Societj' of 

 Natural Historj-, sxiv, 1890, pp. 482-515, 556-559. 



-Of 34 si>ecies of North America Noctno-bombyces, whose transformations are known, all except 1 feed upon 

 trees. (See Edwards's catalogue.) 



^It is hardly necessary for us to express our entire disagreement with the view of ilr. A. G. Butler, that these 

 Noctuid:c are really Xotodoutiaus, or in any way allied to them. It seems to us that the characters which he uses.to 

 remove them from the Noctuidie are superficial and adaptive. Nearly twenty-five years ago I satisfied myself, after 

 an examination of the denuded head and wings, that the Noctuo-bombyces were true Noctuida\ and did not depart 

 essentially from the typical genera. 



■•While many, though not all, butterfly larv;p, as shown by Seudder and W. H. Edwanls, have spine-like "land- 

 ular hairs in the first stage, which may in some cases persist into one or two later stages, tbe body in many species, 

 especially in those which are not general feeders, but select low-growing, herbaceous plants, becomes smooth and 

 ornamented with stripes or spots. However, as a rule, butterfly larva- can not be divided, as the Bombyces, etc., 

 into high and low feeders; yet from Scudders "Classified list of food plants of American butterflies" (Psyche, 1889) the 

 following facts and couclusious may bo stated: 



Hesperidd'. — Out of 45 species enumerated, all l)ut 6 feed on herbs and especially on grasses, and those which 

 feed on tall shrubs or trees, such as EpargijrcKs lilijnis and 5 species of Thanaus, stand at the head of the group, which, 

 as every liody knows, is the lowest family of butterflies and nearest related to tlio moths. 



PnpHionkUv. — Of the 6 species enumerated, 3 fceil on trees as well as shrubs and herbs; 1 of these, however 

 (P. cresphnnles), feeds on trees alone. None of this family are hairy or spiued when mature, except /'. pltilcnoi; with 

 its peculiar flexible, spike-like growths. 



Pieriiiw. — Of 10 species, all feed on herbs, rarely on low shrubs, and none are armed with hairs, bristles, or 

 spines. The other two groups (Lijcanida- and Xi/mphalidtv) are general feeders, occurring indiflerently on herbs, vines, 

 and trees, except the striking ca.se of the 8 Satyrina>, which feed exclusively on grasses and herbs (E. portlatidia, 

 however, sometimes frefjuentiug the Celtis). The very spiny Argyunis larvre feed on Viola. It should also be noted 



