MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 127 



Walker .states that there is in the 9 a slender zigzag reddish band near the outer border. 

 Expanse of wings, 52 lines. 



I had supposed that this was probably only a eliniatic variety of E. !iii]>eri<il!s. but on an 

 examination of the larva and pupa preserved in the Cambridge Museum, and kindly loaned me 

 by Mr. S. Henshaw, I find that there are diti'erenees in both larva and pupa, which appear to be 

 most probably constant. , If so, then E. cacieio) is probably the stem-form inhabiting Brazil, the 

 probable center of origin of the genus, and our Central and North American E. imper'tallx is 

 probably a later climatic species, its specific characters having been assumed as the result of its 

 exposure to a cooler, more northern climate. 



Zffym.— (PL XVII. tigs. 2, 2 A, 2 B. Copied from Burmeister.) 



An alcoholic specimen differs- from a blown example of E. imjjerudis from New York in 

 its much larger size. The head is dark, as was probably the body, as in Burmeister's tioure. 

 The dark head is yellowish brown on the sides. Prothoracic shield and armature just as in 

 E. imj^erialis^ the vestigia] spines low, not nearly so high as broad. The most marked difference 

 is seen in the dorsal and subdorsal second and third thoracic horns and in the caudal horn. 

 They are in E. cacicus very long and slender, and the spinules are fine and slender, acute, not 

 coarse, thick, and short as in E. imperialis. The horns are also slenderer at base and not sub- 

 conical as in E. imjjerialis, and are considerably longer in proportion. (See PI. XLVIII, fig. 4.) 



Another decided difference is the equality in size of the horns of the dorsal and subdorsal 

 series (///) in E. cacicus: in this species those of the subdorsal series are slightly thicker, but 

 very slightly shorter, than the dorsal ones, while in E. imperlalts they are about a third as long 

 and scai'cely higher than broad; they bear about 12 spinules. which are not short, broad, and 

 crowded together, as in E. iinperialis, t)eing on tlie conti'ary slender and delicate, very acute. 

 The caudal horn (tig. 4, c) is also long(>r and higher, less curved, thick at the base, and with 

 scattered, slender spinules. 



The bodv is very hairy, densely so. the hairs being pale horn-colored and rather longer 

 than in the most hairy E. hnperkdis. The suranal plate armed with granulations, which are 

 crowded as in E. unperialis, but coarser and rounded. Anal leg with somewhat crowded 

 granulations, which are larger and more rounded than in E. iinperialis. 



Burmeister states that the spines are rouye carmine, but those of our E. iinperialis are 

 always dark brown or yellow. The ground color of the body varies according to Burmeister in 

 being of a clear green or of an obscure graj'ish brown. We infer, therefore, that dichromatism 

 occurs in the South .\merican larva as well as in its representative in North America. 



The stigmata of E. cacicus are carmine red with a white ring (ceinfurt). The abdominal 

 legs are marked with a large lacquered red spot above the planta. 



Burmeister, stating that Boi.sduval gives to the North American type, which is smaller and 

 clearer in color, the name of imperialis, and to the South American that of A. cacicus, adds that 

 the individuals from Buenos Ayres resemble the North American form in size and color, with 

 some apparent modifications in ornamentation. Ijringing it nearer the Brazilian type. "This 

 form is known under the name of Basilona (or Crenadia) opaca.^^ (Atlas, p. -t6.) 



Pupa. — Ijcngth, 05 mm.; thickness, 20 mm. (PI. LVI, ffgs. 6, 6ff). It differs from tliat of 

 JE. iinpfrialis in the longer maxilla^ and legs. The labrum is more distinct; the head in front 

 less rugose, ))ut elsewhere, namely on the wings and body, the surface is more spinose (compare 

 figs. 5 and 6 and 5(/ with 6a). The hinder edges of the posterior abdominal segments are much 

 more spinose, the teeth or spines on the hind edge of the sixth abdominal segment being longer 

 than thick, and acute. The cremaster is smooth on the under side, whereas in the S of E. imperialis 

 there are longitudinal ridges; on the upper side it does not specially differ from the other species 

 named. The armature on abdominal segments 8-10 is much as in E. ii/ip>eriali.s. 



Food plants. — In Brazil, according to Burmeister, it lives on different plants; he found it at 

 Rio de Janeiro on Mangifera indica, at Buenos Ayres on Enjthrina crista gcdli. 



Geographical distr Unit ion. — Rio Janeiro (Agassiz, Museum of Comp. Zoology, Cambridge, 

 Mass.): Rio Janeiro (British Museum): Brazil (Beske in Boisduval). 

 Vol. 9—05 9 



