ATTACIDES. 273 



(i) Citheronia and Hemileuca (with two such nervures) fall together, 



(2) Automeris, Aglia, Saturnia and Attacus (with one). By a com- 

 bination of the neurational characters exhibited by the fore- and 

 hindwings, one gets : 



Group 1 (generalised) — (a) Citheronia, Hemileuca, with two anal nervures. 

 (b) Automeris, with only one, but with IV 2 of forewing stationary. 



Group 2 (specialised) — Aglia, Saturnia, Attacus, with one anal nervure. 

 [Aglia is to be considered the lowest, as nervure IV 2 is less intimately related to 

 IV^ than in Satumia.~] 



This Dyar states agrees with the classification on larval characters 

 f supra), and asserts that the neurational characters broadly interpreted 

 confirm the larval ones. Grote, however, maintains (Ent. Rec, x., 

 pp. 145 — 146) his views, and considers Aglia to be a specialised 

 Automerid, and Hemileuca a generalised Saturnian, explaining the 

 similarity of nervure VIII in Hemileuca and Citheronia as a matter of 

 convergence, and asserts that all the other imaginal characters 

 contradict the bringing of Hemileuca and Citheronia into one group. 

 He further states that the female antennae in Aglia are Citheronian 

 and Automerid in character, whilst those of Hemileuca are of the 

 Saturnian type. [See also Grote's paper on the subject (Proc. South 

 Lond. Ent. Society, 1897, pp. 82 — -85)]. Grote further states (in litt.) 

 that the subfamilies indicated by him could, at least in part, be 

 regarded as families, when his classification would read as follows : 



Fam. I : Attacid^e. 



Subfam. I : Attacinae. 



Subfam. 2 : Saturniinae . 

 Fam. II : Hemileucid^. 

 Fam. Ill : Automerid^e. 



Subfam. I : Agliinae. 



Subfam. 2 : Citheroniinae. 



A complete study of all the early stages of the Attacid (sens, lat.) 

 species can alone eliminate all the disturbing elements arising from con- 

 vergence, and give us a satisfactory classification based on true 

 ancestral characters. As a contribution to this end, Packard discusses 

 ( Bo??ibycine Moths of America, pp. 40 — 43) the salient ontogenetic 

 features of several Attacid species, separating what are most probably 

 congenital from later adaptational characters, and thus distinguishing 

 between ancestral characters denoting relationships and superficial 

 specialisations developed to meet individual needs. Packard has 

 selected the larvae of Platysamia (ceo-opia), Callosamia (promethea), 

 Telea (polyphenols) and Actias (hen a) as typical genera for treatment. 

 After noticing the great contrasts between the first and later larval 

 instars, both in armature and coloration, he gives the following 

 summary of what he considers the salient ontogenetic features : 



Platysamia (cecropia). — a. Congenital characters: (1) The setae in stage 

 1 blunt, slightly bulbous and glandular. (2) The tubercles all of the same size. 



(3) Body in stage 1 dark, almost blackish-green, head jet-black, tubercles yellowish- 

 green. (4) The homologue of the caudal horn shows plainly its double origin. (5) 

 The differences between the colours of the larva in the first and last stages very 

 marked, b. Evolution of later adaptational characters-. (1) The thoracic dorsal 

 tubercles in stage 2 and onwards are larger than the abdominal ones. (2) Five rows 

 of indistinct spots along the body in stage 2 (not so distinct as in S. cynthia, the body 

 being still dusky- green), not originating from lines. At the end of stage 2 the larva 

 is more like that of S. cynthia of the same age, the body being more yellow and the 

 black spots more distinct. The spots disappear at the end of stage 4. (3) The 

 thoracic dorsal tubercles deep orange, their homologues on the abdominal segments 

 amber-yellow, (4) The tubercles at the end of stage 2 and in stage 3 spotted with 



