16 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



known tho liirva is hiitched with thi.s pair of liorns alone, and it is signiticant that they are 

 situated on the most exjOTsed part of the body. 



In stage I the dorsal spines are throughout longest in A. I'irghiieiisiK (those of tho tliird 

 thoracic segment largest in this species), and in this respect this species is the most primitive or 

 nearest its supposed Adelocephala ancestor; the dorsal spines are longest in A. stiguia^ are shorter 

 in A. senatorin, and much the shortest in ^1. nihicunda. 



It is noteworthy that in tiie full-fed larva the dorsal horns of the second thoracic segment 

 are longest in A. riiyinensis, shorter in ^1. .fenatoria, and shortest in ^1. ruhicunda. 



The question arises as to the probable cause of the reduction in all the spines in this genus. 

 This may perhaps be answered by the fact that they do not need a defensive armature, owing to 

 the style of coloration, the l)ody })eing striped, like so many unarmed caterpillars, especially 

 Noctuidw, etc. Moreover, the eggs are laid in large patches of over 200, the young larv» live 

 gregariously on the under side of leaves of the oak, maple, etc., which bear no spines, while this 

 gregarious habit in all but A. rnhicunda is maintained throughout larval life, as they, especiallj' 

 A. senaforiu, are seen to be crowded together in clusters on the partly defoliated twigs of the 

 trees, somewhat after the manner of caterpillars of Datana. 



The 'iurnnal plate and its i/iod(p\'i(t/'>/is. — This is the reduced tenth abdominal segment, and 

 is flattened triangular, rather large, usually ending in a fork. 



In its perhaps most generalized form, that of Adeloceph<di( hlenJor. the suranal plate in stage I 

 is armed with three well-developed setiferous tubercles, the tubercles themselves conical ))ut 

 higher than thick at the base, and bearing a long slender seta. The plate is forked at the end, 

 the two terminal tubercles giving it a forked appearance; there is also a pair of setiferous 

 tubercles similar in size and shape to the others, each one situated a little way in from the edge 

 and behind the middle. There are no secondary setiferous tubercles. This description will also 

 apply to Syssphinx in its first stage. 



In the succeeding stage (II) and all the later stages the end of the plate is distinctly forked 

 and the surface of the plate is rough, with live secondary' setiferous tubercles, while the primary 

 ones are but little larger than the secondary ones and merely hold their own in size; on tlie base 

 of the two teiniinal tul)ercles forming the fork are several secondar_y setiferous tubercles. 



In the last stage the setaj are much reduced in length, nearly atroj)hied in many cases, and 

 the surface of the plate is rather evenly granulated. 



In Anlmta ruhictoidd (Pis. LII, LlII) the suranal plate is rather short, rounded at the end, 

 having lost the salient features of that of Adelocephala. The setiferous tul)ercles in stage I are 

 minute, though the bristles themselves are long, hair-like. The typical number is live on each 

 side, three marginal ones, one situated within, beyond the middle of the plate, and the pair of 

 terminal ones. After the tirst molt the pair of terminal tubercles are much larger than before, 

 giving the plate a forked appearance. In the last stage the primary tubercles are large, distinct, 

 and black-pigmented. 



In stage I of A. tndiieunda and .senator/a the plate is simplest; in ^4. riiyinu'iis/i< and stiff ma 

 it is more specialized and somewhat pigmented. In the second to the final stages, however, of 

 A. senatoria the plate is more stoutly spinose than in an\' other species, the lateral and two 

 terminal spines or tubercles being solid, conical, and pigmented l)rown black (tig. 2). 



In Eacles imperialix the armature of the suranal plate is very formidable in the two tirst 

 stages (Pis. XLVIII, tigs. 1, 2), especially before the tirst molt. The tirst tubercle, i. e., that 

 neai'estthe ninth segment, on each side is very large, four-headed, and as long as the entire plate. 

 In the second stage, the tubercle spines are reduced in size but retain the same structure, the 

 seta? remaining; but after the second molt the great basal spines become I'cduced to low, flattened, 

 broad, hubbly vestigial tubercles, with secondary tubercles. The plate is l)laik in the last stage, 

 with the edges yellow, while the granulations, or low stout conical tubercles, are coarse and 

 render the surface of the plate rough. 



In CitheroniareffaJist\\Q armature of the suranal plate is similar to that of Eacles, on the same 

 plan, but the tul)ercles and their spinules are nuich longer and slenderer, as seen in PI. LIV, tigs. 



