ME.MOIKS OF THE NATIONAL ACADlvMY OF SCIENCES. 19 



coniiiioii foundiitiDii or \v:u-t-liko tuhcrclc. To quoto fi-oni my account: " On tlic cii^htli segment 

 i.s a .single central dorsal, black, oval, niodorately prominent wart, which is twice as large as the 

 largest on the ninth segment; it is transverse, bearing a bristle at each end, thus having plainly 

 originated from what was once two sejjurate warts." At the end of this stage, before the tirst 

 molt it becomes a double, large black tubercle, still ending in a i)air of sctiu; after molting 

 (stage 11) it is " now well developed, high, conical, and fleshy, slightly inclined backward, dark 

 at tip. and still liearing two bristles, though the dark chitinous spine is obsolete; the horn-like 

 tubercle is half as high as the segment is thick."' In the next stage it is nearl}- as long as the 

 eighth .segment is thick vei'tically. "The horn is slightly retractile in this stage, and the ba.se 

 is movable, being capaV)le of withdrawal and extension and is distinctly nutant, the apex some- 

 times hanging over baclcward." In the fourth stage the horn becomes larger, higher, and more 

 acute than before; "it is freely elevated or allowed to fall over backward, is soft and flexible, 

 but very slightly retractile, and bears a few scattered, tine bristles." In the tinal stage the horn 

 is high, stiff, not granulated, but somewhat annulated. The horn is more like that of a sphingid 

 than a ceratocampid. in not forming a solid spine, though this is not iuvariabl3' the case in the 

 Ceratocampida'. 



In the allied genus Dasylophia the larva is hatched with the two tubercles i on the eighth 

 abdominal .segment still separate, and thus represents an earlier stage in the ontogeny than in 

 Pheosia. In stage I this pair of tubercles each end in a l)ristle; in stage II and later stadia the 

 tubercles are fused to form a low, flattened tubercle, the seta^ being lost. 



In Ilyparpax aurora and the species of Schizura, whose segments are nearly all dorsally 

 humped, each side ending in a seta, the fleshy tubercles on the eighth abdominal segment are as 

 large or nearly as large as those on the rirst abdominal segment. In .V. L-pI'moultx the fusion of 

 what corresponds to the caudal horn is seen in PI. XXVI, flg. 1, tubercles il having apparently 

 become atrophied. The cause of the fusion of these tubercles, whether on the first or eighth 

 abdominal segment, is obscure, but proliably it is a mechanical one. 



In some respects the most specialized genus is Heterocampa, and in two of the species whose 

 ontogeny is known, i. e., H. ohJlqua and //. tjutfiiutta, certain segments are armed with a pair 

 of dor.sal chitinized .solid spines or horns. Those on the eighth abdominal segment are well 

 developed, being long, erect, and bearing a few secondary spines. (See Pis. XXX and XXXI of 

 my Monograph of the Notodontidw, also fig. 8.5.) Their bases are, however, wide apart and 

 there are no signs of a tendency to fusion. They disappear at the first molt, and are probably 

 either defensive structures inherited from some earlier form, or" have arisen with comparative 

 rapidity, a sport or mutation due to stimuli from without. The ninth abdominal segment in 

 in these two species is well developed, but smaller. 



These caudal horns do not appear in the first stage of 11. liinalatd, IL inanteo., IL untcohir. or 

 IL. astarte, which are thus proved to be more recentl}' evolved. In the tailed forms, such as 

 Heterocamjxt un/cohrr. ^lacrurocampa and Cerura, it is possible that the growth energv 

 expended in the productit)n of the armature has been transferred to a process of hypertrophy of 

 the anal legs. 



Reference should also be made to the species of Notodonta, Sj'mmerista, etc., which bear a 

 hump or broad, low tubercle on the eighth abdominal segment, apparently resulting from the 

 fusion of the primitive setiferous tubercle. 



The caudal luira of tlu- Botnhycida'. — In the majority of the species of this interesting group, 

 represented by Bomhy.v mori, there is a well developed caudal horn. 



In stage I the larva is similar to a young Malacosoma (Clisiocampa), having the fine bail's 

 arising in tufts from small warts, but already the two warts on the eiglith abdominal .segment 

 have united into one. 



On examining a series of larvw in the fourth stage the caudal horn is seen to be soft, fleshy, 

 thick at the base, and rather densely clothed with comparatively coarse seta^ it rather suddenly 

 contracts toward the end, which is somewhat acute, pointed, the ti]) dark chitinized, and bearing 

 no signs of its originally double origin; the slender end of the horn bears no hairs and is very 

 finely granulated. 



