MEMOIRS OF THE NATIONAL x\.CADEMY OF SCIENCES. 121 



divided into two oval lanceolate black plates, the small ends poiutiug toward the head. Otherwise 

 the body is marked as in the above-desmibed specimen, except that there are black spots at the 

 base of the middle abdominal lejis. Tlie hairs are not represented as so clavate as in the other 

 specimen. It is possible that the latter is iu the second stage, but if so, the suranal jilate would 

 not i)robably be so large and entire. 



Tliird stiujc. — Length, 7 mm. (probably not of the normal length, owing to confinement), July 

 30. About ready to molt, as the prothoracic segment is somewhat swollen. The black prothoracic 

 plate still persists, aud the hairs arising from it are about twice as long as those elsewhere, but 

 the black suranal plate has disappeared; the anal legs are still slight, and the body beyond tlie 

 sixth abdonunal segment is upraised. The reddish color has deepened, and the yellowish lines 

 are more distinct, while the spiracular line, inclosing the distinct black spiracles, is pale lilac- the 

 middle abdominal legs do not appear to be spotted. 



Fourth stage. — Length, 10 mm., August 13 (evidently underfed and unnaturally small). The 

 head is large, as wide as the body in front; the cervical shield still persists, as do the clavate 

 hairs. The eiilor has now changed to a dark reddish brown, above and beneath, with longitudinal 

 gray stripes seen from above aud four seen sideways; the additional stripe is the infraspiracular- 

 one, while the spiracular one has moved up, the spiracles being situated between them. 



The following is a description of another larva of this stage received from Mr. Angus,. 

 August 2.~>: 



Length, lo mm. Head shining black, as wide as the body. 



The body of the usual cylindrical shape, rather slender, dark pitchy reddish brown all over. 

 Prothoracic shield transversely oblong, not so square at the corners as in D. perxpicua. There are 

 four dull whitish rather obscure lines on each side, which are of nearlj- the same width and of 

 exactly the same color; they are somewhat irregular on the edges, being somewhat broken and 

 of the same distance apart. The lowest or infraspiracular line is a little wider and more distinct 

 than the others, aud extends along the lateral ridge. The body beneath is of the same color as 

 above. The suranal plate is black, rounded ; the anal legs are black at the tips. The middle 

 abdominal legs are stained black above the plautw, and the thoracic legs are black. The hairs 

 are long and white; those on first thoracic segment, and eighth and ninth abdominal, longer thau 

 those elsewhere; those on the prothoracic segment stand up and curl over the head, and two or 

 three of them are as long as the three thoracic segments put together. The spiracles are black. 



Fifth stage.— heugth, 28 mm., August 29. Very different from the fourth stage, the color 

 being still darker, while only two grayish lines are seen from above, and two lines when the larva 

 is seen from the side. Tlie two dorsal and the supraspiracular lines have disappeared. The body 

 is now clothed with numerous soft fine gray hairs, many of which are nearly as long as the body. 

 The anal legs are still smaller than in the preceding stage. 



Recapitulation. — 1. In this species the larvai of the first four stages apparently have clavate 

 glandular hairs, an unusually late persistence. 



2. Tlie body is reddish in the three first stages, but becomes dark in the fourth, while in D. 

 mini-stm the body is reddish in the fourth, being less iirecocious tlian iu this species. 



3. The loss of two of the longitudinal stripes iu stage V is noteworthy, and the habits of the- 

 larva should be noted by the future observer to learu the probable cause of such a change ; also why 

 in B. miiiistra, and perhaps in other si)ecies, there is such a decided change iu the general color 

 and stripes in the last as compared with the penultimate stage. 



4. The black suranal plate seems in Stage I to be entire, and to divide in two at the end of the- 

 stage, not being present in the third stage. It is to be ho])e:l that those who may hereafter rear 

 the S[)ecies of Datana will preserve specimens of the earlier stages in alcohol for future study. 



Habits. — Eggs, August; larviE, August and September; adults. May aud August; localities,. 

 Kansas, Missouri, Indiana, New York, Maine, and District of Columbia; food plants, walnut, 

 hickory, larkspur, thorn. (Riley MS. notes.) 



Fooil plaufs. — The larvas prefer black walnut, but feed on hickory, butternut, etc. (Angus),-: 

 walnut (Pilate); '-Live together iu large companies on walnut {Juglans), hickory {Garya), beech,. 

 {Fiif/us), and also on oak {(^uercus), but very rarely" (Beutenmiiller); willow, honey locusf, thorn,, 

 and apple (Hiley). In Kansas, Juijlans nigra (Popenoe). 



