MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. ' 23 



As regards coloration, .1. nihicunda appears to be the most primitive or phylogenetieally 

 oldest species of the genus. When hatched (stage I) there are no stripes, the larva being of a 

 pale yellowish green, the traces of the longitudinal stripes appearing toward the end of the stage 

 and the larva, having fed, Ijecoming greener in color. The green color in this species alone per- 

 sists throughout larval life, the alternate stripes being lighter or darker green, accoi'ding to the 

 amount of green pigment deposited in the hypodermis. It is to be observed that the dark green 

 stripes are no more accentuated dorsally than on the sides. The green hue is a protective one, 

 most larva; being green, and hence this species needs less spiny structures than the others of the 

 genus. Here it might be observed, what is well known to everj- entomologist, that the under- 

 side of all lepidopterous larva" is usually paler than the upper, due to the lack of pigment. This, 

 as in all animals in which the underside of the body is lighter than the upper, is due to the 

 absence of direct light vays and of the resulting stimulus to the deposition of pigment, the under- 

 parts being in the shade. This is the case also in fishes, reptiles, birds, and mammals. 



Yet in some caterpillars there may be a dark ventral median line or stripe, and in ApaiflochK 

 torrefacta (Ft. I of this Monograph, Fl. IX. fig. tV) there are two or three conspicuous black 

 patches on the underside of the body, yet the larva may occasionally assume a posture revealing 

 these marks. 



It is noteworthy that in A. ruhicimda all the green stripes, so far as we have observed, 

 appear at once along the entire length of the body, not apparently first originating at the end of 

 the body. But further observations are needed as to this point. 



Origin of the lateral {! iifraspiracula r) red hand. — Nearly all lepiilopterous larv;e have along 

 the sides of the abdominal-segments 1-8 a lateral ridge, or salient, somewhat fieshy. convex, lon- 

 gitudinal fold, more or less regular, situated beneath the spiracles or spiracular line. This ridge 

 or fold, which is continued behind by the edge of the suranal plate, is apt in many caterpillars to 

 be pigmented with red or yellow, these tints appearing on the upper or more exposed surface, 

 while the lower side is pale or whitish, the shading being more or less gradual. 



In the second stage of A. ruljicanda this lateral ridge begins to be colored red, forming an 

 obscure line, which originates on the eighth and ninth abdominal segments and passes forward to 

 the first thoracic segment, though fainter on the anterior half of the l)ody. This reddish line may 

 have been inherited from some extinct form, because in the third and subsequent stages it disap- 

 pears from the front and middle of the body and becomes restricted to abdominal segments 8 and 

 9. On these segments it forms a very conspicuous mark, being bright crimson or vermilion. 

 The body is also a little flattened on these segments, so that the spots show from, above. I have 

 not speciallv noticed the larva while resting on a maple leaf, but am disposed to think that 

 these two vermilion or reddish patches harmonize with the reddish petiole of the leaf of this 

 tree, and are thus protective. It is also noticeable that this line, originating at the tail end of 

 the body, is another example of the origin of certain new kinds of markings at tlje posterior end 

 of the bod}'. 



In the majority of the species, namely, in three out of the five whose transformations are 

 known, the longitudinal l)ands are either pink or grayish, and this style of marking thus seems 

 to have been the one most advantageous to the species. The lower lines in ^1. rlrginiensis and 

 stigma, corresponding to the reddish line in ^4. rnhicuiida. is pink, and of the same hue as the 

 subdorsal line. The tint appears in stage II. and remains in all the subsequent stages, extending 

 along the whole length of the l)ody. They do not, so far as we have been able to observe, origi- 

 nate at the end of the body. In stage II there are no signs of them. 



.1. senatoria is the most conspicuoush' marked larva of the genus. Unlike ^-1. ruhicunda 

 and the other species, it remains gregarious throughout larval life, clustering in great numbers 

 on the terminal twigs of the oak. after the leaves have l)een devoured, so as to be very conspic- 

 uous, yet the}- appear to be avoided by insectivorous birds. Whether they are poisonous or not 

 we do not know, but the bright, deep ocherous stripes alternating with the black ones are prob- 

 ably warning colors. 



What in A. rid/ica/ida are dark green stripes become black in A. simatoria. while the deep 

 ocher or scotch snvitf ones correspond to the pale green stripes of ,1. ruhicunda. All the spines 



