286 BRITISH LEPIDOPTERA. 



Saturniinae and Hemileucinae* '. Both are agreed that the Attacinae 

 and Saturniinae are closely allied, and when we examine these 

 carefully we find that the larvae of these agree in having " no single 

 tubercle on the 9th abdominal segment, in having a tubercular anal 

 plate, and no primitive first stage," whilst the larvse of the Agliinae 

 and Hemileucinae (as well as the Citheroniinae and Automerinae\) are 

 said to have a primitive first stage (see Dyar, Ent. Rec, x., p. 37) J. It 

 is quite clear, on larval characters, that the Attacidae (limited to 

 Attacinae and Saturniinae) is the most specialised family, whilst the 

 other subfamilies, that have a primitive first larval stage, belong 

 to more generalised families. Of the two subfamilies thus retained, the 

 Attacinae would appear to be the more specialised, for, as we have be- 

 fore noticed, Packard remarks (Proc. Am. Acad. Arts and Set'., 1893, p. 

 58) that, "as regards the larval tubercles, the species of Saturnia are 

 on the same plane with the embryo, just before exclusion, of the 

 more highly specialised forms of the group Attacinae" and adds 

 further that " an interesting series of parallelisms may be observed 

 in comparing the early and later stages of the larvae of the family, 

 e.g., while the late embryos of the Attacinae are, perhaps, paralleled 

 by the fully-grown larva of Saturnia, the fully-grown larva of the 

 most, or one of the most, generalised Attacinae, Platysamia, is on 

 the same plane of specialisation as the larva of Callosamia in its 

 third stage." He also considers that the great size of the Attacinae 

 (particularly Attacus atlas) appears to be a sign of recent specialisa- 

 tion, and the small size of Saturnia, apart from its other features, 

 suggests that it is a generalised form not departing greatly from 

 the normal size of the superfamily Bombyces. Dyar considers 

 that, of the more generalised families, the Agliidae is most closely 

 allied to the Attacidae (Saturniidae), resembling it in that the larvae 

 have " no single dorsal tubercle on the 9th segment, but have the anal 

 plate tubercular," and differing from it in that there is a primitive 

 first larval stage J, and that nervure IV 2 in the forewing of the imago 

 is less intimately related to IV , than in Saturnia. 



It will be gathered from our remarks that we are inclined to limit 

 the Attacidae to the subfamilies Attacinae and Saturniinae, excluding 

 Agliidae as of family rank. Meyrick appears to base his diagnosis 

 {Handbook, &c, p. 313) of the superfamily largely on these two 

 subfamilies, for he observes under his genus Saturnia : " Characters 

 those of the family." These family characters read as follows : 



Imago : Head densely rough-haired. Ocelli absent. Eyes glabrous. Tongue 

 rudimentary. Antennae (under -3) in S strongly bipectinated to apex. Labial palpi 

 very short, hairy. Thorax densely hairy above and beneath. Abdomen hairy. 

 Femora and tibiae densely hairy, posterior tibiae without middle spurs. Forewings : 

 lb furcate, 5 and 6 closely approximated from upper angle. 7 absent, 8 and 9 out 



* We are quite unable, on larval characters, to accept Grote's statement 

 (Ent. Rec, x., p. 146) that Hemileiica is a generalised J»aturnian in the sense of 

 including it as a subfamily of our Attacidae. The presence of two anal nervures in 

 the hindwing suggests at least a separate family, even if it be not, as suggested by 

 Dyar, a subfamily of the Cither oniidae. (bote's own opinion now (see anted, 

 p. 273) is that the Hemileucids should be treated as a distinct family. 



f It may be well to note here that Bodine places (Antennae of Lepidoptera, p. 43) 

 the subfamily Autonici inae, in which the antenna: of the ? have a single pair of 

 pectinations to each joint in the Saturniidae, with the Attacinae. which latter have, 

 in both sexes, two pairs of pectinations to a segment. He does not know the 

 antennal structure of the Saturniinae {sens, strict.). 



\ We do not agree that Aglia has a primitive first stage (see, footnote, anted, p. 272). 



