22 :\IEM()IKS OF THE NATIONAL ACADEMY OF SCIENCES. 



IV. PROTECTIVE ARMATURE BOTH IN SHAPE AND COLOR AND 

 DEFENSIVE MOVEMENTS. 



Keferrint^ to the agjjrossive or det'ciisiive nioveiiiont.s of the caterpillars of Adelocepliala. 

 Dr. Jewett remarks: " L'p to the third molt the lar\ie when alarmed move the protulierances 

 on the second and third segments as if to frighten away the intruder." 



In my Notes (p. 15M) I remark regardino- them in the second stage: '• In tiiis stage upon 

 touching or teasing the larva the thoracic spines spring out; at the same time the head, together 

 with the thoracic region, jerk violently, as if to beat oti' an intruder. Also, when two caterpillars 

 meet they evidently attack each other, l)utting and striking with their horns like two hostile goats, 

 deer, or cattle. It seemed evident, after repeated observations, that the great thoracic spines are 

 of real defensive use." Again, in referring to a larva in the fourth stage: "it would appear 

 probable that the formidable spines of the grown-up caterpillar save it not infrequently from 

 being swallowed by birds, though the horns are proliably of greater use in the earlier stages, 

 when they are much longer and much more movable, in frightening away ichneumons and 

 Tachina?. For example, even when 20 nmi. in length, a larva w'as seen when teased to spread 

 apart its great arm-like horns, while the full-fed ones did not notice such stimulus." 



It was also added that in the larva.' of the last stage, instead of being green as in the earlier 

 stages, they become roseate pale coral red, and "not so near in tint to the spines of the food 

 jilant as in the voung. When the caterpillar is at rest the\' are held close together in a recurved 

 jjosition, and in the grown-up larva when touched they are not moved or the body jerked in 

 response to such stimulus." 



"The horns in Sphingicampa are not held spread out as in C ngalis. l)ut those of each pair 

 are constantly held close to each other. The horns and the six silvery, opalescent, shining tuber- 

 cles probably ))ecome terrifying by the movements of the larva. The latter are turned on and 

 thi-ow their light out suddenly like flashes, and may thus have a deterrent etlect on their 

 enemies." (Pi'oc. Amer. Phil. Soc, XXXI, p. 157.) 



It is worthy of note that the larva' of most of the s])ecies of this genus feed on spiny plants, 

 one, ^1. li!c<ihi)\ on the spiny or honey locust. 



In Brazil, according to Peters, A. trardii lives on a spiny mimosa whose leaves on the under- 

 side arc violet lirown: A. i<i(h<(tuiiiJafa lives on a spiny climbing plant, ^4. hrci'is on a thorny 

 mimosa in the gardens of Kio Janeiro, and the caterpillars of an undescribed species on an 

 acacia. 



These caterpillars, with their remarkably long spines, including those figured by Peters and 

 by Burmeister, have the dorsal spines of the abdominal segments nearly or quite as long as those 

 on the thoracic segments. As these larva^ are quite large, and their spines in some cases from a 

 half to an entire centimeter in length, they would easily be mistaken by birds and lizards, as 

 well as monkeys and other mammals, for the smaller spines of the trees or shrubs on which they 

 feed, especially when not in motion, and it is a question whether they are not more useful as 

 mimicking the spines of their food plants than as simply defensive structures. Yet a l)ird, or 

 even a lizard, might hesiitate l)efore swallowing such spinose creatures. 



As to the origin of these great spines and their adaptation as protective structures it is not 

 easy to frame a hypothesis. To sujjpose that the}' arose by natural selection is scarcely an ade- 

 quate explanation, though this may account for their jMeservation after they have once been 

 developed. 



V. COLORATION IN THE LARV^ OF THE CERATOCAMPIN^E. 



We will bi'gin with the species of Anisota, in which the armature, owing to the reduction of 

 all but the pair of slender thoracic horns, is reduced, and becomes subordinate to the c()loration. 

 This is especially seen in ^1. siJKifAirldwhGxe. the alternating deep ocherous and black-brown liands 

 are so prominent, the effect of the black spines in most cases being lost, owing to the position 

 of those of th(> dorsal and subdorsal series in the dai'k l)r()wn or blackish stripes. 



