ATTACIDES. 265 



Superfamily VIII : ATTACIDES*. 



Possibly no group of lepidoptera of similar extent has received 

 more attention than, or has been so generally studied as, this, yet the 

 authorities are at present entirely at variance as to the limits of 

 the families and their relationship to each other, whilst the origin 

 of the superfamily and its relationship to other superfamilies, 

 although generally agreed upon, are far from satisfactorily determined 

 when one considers matters of detail. Great differences exist be- 

 tween the main families into which the Attacids are divided in 

 the egg, larval, pupal and imaginal stages, these differences being 

 largely due to specialisation in various directions, and are particularly 

 marked in the larval and imaginal stages. In the larvae one can 

 readily mistake secondary developments for primary structural 

 characteristics, and thus be easily misled as to real relationships. 

 Packard observes (Proc. Amer. Acad. Arts and Sciences, 1893, p. 

 58) that the tubercles of the fullgrown larva of Satumia (pavonia, 

 pyri, &c.) are on the same plane of development as are those of 

 the embryo, just before exclusion from the egg, of the more highly 

 specialised Attacinae ( Callosamia, &c), and that the fullgrown larva 

 of one of the most generalised of the Attacinae ( Platysamia, &c.) is only 

 on the same plane of specialisation as the larva of Callosamia 

 in its third instar. Dyar, referring to the same fact, says 

 {Trans. N.Y. Acad. Sci., 1894, pp. 54 — 55) that, "in a section 

 [Satumiidae (—Altacidae)] of this group, the primitive first larval stage 

 is wanting, the larvae hatching in an advanced degree of 

 specialisation. In the more generalised forms, the tubercles have 

 only a single seta, the base usually prolonged into a stiff (often 

 branched) chitinous rod, absent in the case of tubercles ii, and but a 

 single rod bears iv and v. Dorsally, on the 8th abdominal segment, 

 the two tubercles i are usually consolidated, and sometimes also 

 on the 9th abdominal segment ; or, the tubercles may be enlarged 

 and bear a crown of hairs, or become developed into a bunch of 

 spines." From an examination of the newly-hatched Attacid larva, 

 Dyar thinks that the single process or wart below the spiracle 

 (mentioned above as " a single rod ") is derived from tubercles iv 

 and v consolidated and not from v alone. Wailly's notes 

 (Ent., xxix., p. 354) are not only suggestive as showing other 

 remarkable larval specialisations, but hint distinctly that many 

 so-called African Attacids may not belong to this family (Attacidae) 

 at all. He describes the larva of a species allied to Satumia 

 suruka as "a false Geometer, beautifully black, ornamented 

 on its segments with thorny projections, yellow on the last eight 

 segments and pink on the first. The body is covered with spots of 

 the same colour as the tubercles. The spiracles are black, bordered 

 with yellow ; the claspers of a fine shining black." It is difficult to 

 know how an Attacid larva [sens, strict.) can be "a false Geometer," 

 but Wailly goes on to state that other South African species, placed by 

 systematists in the genera Antheraea, Gynanisa and Bunaea, have very 

 different habits from typical Attacids. Thus the larvae of the three 



* The oldest plural name, applicable to the group, appears to be Linne's 

 Attaci, which was given in 1767, and it, therefore, seems to us necessary to call 

 the superfamily ATTACIDES. We understand that Grote concurs with this although 

 he has in his recent writings used the superfamily name Saturniides [Die 

 Saturniiden, p. 24). 



