MEM0IK8 OF THE NATIONAL ACADEMY OF SCIENCES. 13 



species the trapezoid iiiiidc by tubercles / and // is present on abdominal segments 1 to S. In ^1. 

 stigjna (PI. LI) in stage III // have disappeared on segments 1 to 7, and in this stage and later 

 stages in all the species these tubercles (//) seem to have become atrophied, but retained on the 

 eighth abdominal segment. In A. virgiiriensis of the last stage tubercles // on this segment are 

 nearlj' obsolete and white, like the neighboring secondary tubercles in the specimens examined, 

 but this may be an individual variation, as in all the tubercles there is some diti'erence in 

 development. In a full-grown larva of ^1. stignui the tubercles il on this segment are larger 

 than in A. virginienslx and black or chitinized and pigmented. In A. rvhicunda of the last 

 stao-e the tubercles il are fairly distinct, but not pigmented, while in ^1. senatoria they are larger 

 than in the other species and pigmented dark reddish-brown like the other dorsal and lateral 

 spines. Why this pair of tubercles are retained on this segment and disappear on those in 

 front is not easy to explain, unless it be that they are protected by the overhanging caudal 

 horn, for they are longest and best developed in Citheronia. which has the largest caudal horn 

 of any of the subfamily. 



The lower tubercles, supra- and infra-spiracular, are in the same line as the spiracles ///. 

 being directly above the spiracle and in a line with tul)ercle /, as in the family generally and 

 the Saturniida?, as already observed by Dyar. 



Absence of tubercle i on the ninth aMominal segment. — In referring to the unpaired dorsal 

 tubercle on the ninth abdominal segment Doctor Dyar" states that it corresponds to //. "as 

 tubercle / is not present on the ninth segment."' In the Ceratocampina? there is but a single pair 

 of dorsal tubercles on this segment, fused in the later stages; since they are situated near the 

 hinder edge of the segment, I take it they are tubercles //, but otherwise I do not now see how 

 they can be distinguished from tubercle /. 



The history of the compound or fused horn or tubercle on this segment will be seen b}- 

 reference to Pis. XLV and XLVI of Adelocephala hicolw. 



In Anisota the double nature of this tubercle is plainly seen in the first stage of all the 

 species. In A. virgin iensis (PI. L, //) the two setiferous tubercles // are seen to arise from a com- 

 mon swollen base; in stage III the tubercle is slightly but distinctly forked, and the paired nature 

 of the spine is seen in the full-fed larva in the pair of bristles arising from the end of the tuber- 

 cles (PI. LII, figs. 11), Ic) not showing any terminal division. In ^1. stigma (PI. LII, tig. 3) the 

 separate swollen bases of the seta; in stage I are seen to ai'ise from a common large tubercle; 

 in stage II (fig. 3a) the sette are shorter but separate at base; while after the second molt the 

 tubercle becomes solid, pigmented, as in A. virginiensis. In A. senatoria, stage I (PI. LIII, tig. 

 1), there are, as in all the other species, no traces of tubercle i and the double tubercle ii arises 

 as in the other species, but in stage II (tig. la, ii), besides the primitive seta?, which are wide 

 apart at base, there have arisen two secondary minute seta (showing that this species is younger 

 than virginiensis and stigma); in stage III (tig;. Ii, ii) the tubercle is much less advanced than 

 in the other two species mentioned, since the tubercular base of each seta is still separate from 

 its fellow, the two tubercles arising from a common base; from an oblique point of view the 

 tubercle is seen to be more deeply forked than in the other two species mentioned (tig. IS, // a). 

 In A. rithicunda, stage I (tig. 2, ii), the tubercle is as in ^1. senatoria. In stage III the tubercle 

 seen vertically and from behind or obliquely (fig. 2h, a) to be deeply forked, and bearing at least 

 8 or 10 fine secondary setiferous tubercles, which suggest that this species is the most specialized 

 and therefore recent .species of any of the others, this view being corroborated by the other 

 features of its armature and coloration. In the final stage the tubercle is slightly forked. 



In Sj'ssphinx, Eacles. and Citheronia, tubercle ^■ appears to be absent in stage I and through- 

 out larval life. The fused tubercles ii in Adelocephala hicojor is in the last stage much reduced, 

 is simple, conical, not pigmented, but whitish, though showing the scars of two small seta?. 



In Syssphinx molina, judging from Burmeister's figure, it probably is much as in Adeloceph- 

 ala Mcolor, reduced and scarcely distinguishable except by position from the secondary tubercles. 

 It appears to be a little more prominent in A. icardi, and still more so in A. urgyracantha. 



aLoc. cit., p. 232. 



