270 BRITISH LEPIDOPTERA. 



A. Pectinations ventral, single pair to each segment . . Bombycidae. 



AA. Pectinations dorsal , 



B. Single pair to a segment . . . . . . . . Hemileucidae. 



BB. Two pairs to a segment 



C. Distal portion of clavola filiform .. .. CiTHERONHDyE. 



CC. Clavola pectinate throug-hout . . . . Saturniid/e. 



He notes (loc. at.) : " The Hemileucidae exhibit a line of develop- 

 ment distinct from that of the Bombycidae on the one hand, in having the 

 pectinations dorsal instead of ventral, and from the Cither onii'dae and 

 Saturniidae on the other, in having a single pair of pectinations to a 

 segment. Thus it appears that the Hemileucids belonged to the 

 branch which produced the Citheroniids and Saturniids after it had 

 separated from the branch giving rise to the Bombycidae. The 

 Citheroniidae and the Saturniidae have followed the same line of 

 development in that they both have two pairs of dorsal pectinations to 

 a segment. The first family has not progressed so far as the second ; 

 its members do not have their antennae pectinate throughout. The 

 antennae of these families show a high degree of development also in 

 the arrangement of the hairs of the third type. The spaces between 

 the pectinations are nearly filled with long, interlacing hairs, which 

 are regularly arranged in a distinct line of from two to three rows 

 extending continuously from the apex of one pectination along the 

 dorso-lateral surface of the segment to the apex of the other pectination 

 of the same side.* Still another feature showing remarkable develop- 

 ment of sense-organs is exhibited in the higher forms, especially in 

 Tropaea and Samia of the Saturniidae. This is the multiplication of 

 cones. While in nearly all other moths cones are limited to one or 

 at most two to a segment, we here find them literally heaped up on 

 the ventro-distal edge of the segments of the distal portion of the 

 antennae ; and on the pectinations also there are often several either 

 at the apex or along the pectinations at various intervals. Fig. 7 

 shows this condition in the antenna of Tropaea lunaV A number of 

 African forms and some others, that appear to be very close to 

 typical Attacids, differ, nevertheless, in having antennae with 

 single pectinations. Eudaenwnia seems exceedingly close to Actias, 

 yet it has but a single pair of pectinations to a segment, and there 

 are many similar parallels. It follows that a second pair of pectina- 

 tions can be lost per sa/tum, and frequently is so lost, and that some 

 conditions, especially African ones, promote this tendency, or, taken in 

 conjunction with remarks already made ( anted, pp. 265 — 267) about 

 the forms of larvae and methods of pupation in African species, it is 

 by no means impossible that an Attacid stirps, beginning as low down 

 as the Hemileucidae, and, failing to develop the second pair of 

 pectinations, did, nevertheless, undergo an evolution, very parallel, 

 not generally merely, but in considerable detail, with the ordinary 

 Attacids. Such an occasion would not be out of accordance with 

 what is found in many other portions of the biological field ^Chap- 

 man). 



As a more satisfactory subdivision, based on larval characters 

 alone, Dyar gives the following tabulation : 



A single dorsal tubercle on Qth abdominal segment 



A pair of tubercles on anal plate . . . . . . OthkroniiDjE. 



No tubercles on anal plate .. .. .. .. Hemileucids. 



*Bodine's fig. 54 represents the arrangement in Dryocampa rubicunda. 



