268 BRITISH LEPIDOPTERA. 



the structure of the subspiracularwart, by consolidation of iv and v, would 

 make the Saturniids ( = Attacids) one of the specialised superfamilies in 

 the same line of descent as the generalised Lacosomids. His action, 

 however, in placing (loc. cit., pp. 53 — 54) the latter in his heterogeneous 

 Cossina, and his statement that they cannot be properly classed with the 

 Satumiina on larval characters, though the difference is not funda- 

 mental, leaves us in great doubt as to his real opinion. He adds : " In 

 the. Satumiina the unpaired dorsal tubercle is not an invariable character, 

 in one section it is absent on the 9th abdominal, and in one genus 

 (Anisota) on the 8th, so that it is not difficult to imagine the Lacosomidae 

 to represent the most generalised condition of the Satumiina in which 

 the consolidation of tubercle i has not taken place on either segment." 

 With regard to the larval specialisation of Attacids and Sphingids, 

 Dyar offers further remarks (" Class. Lep. Larvae," Ann. N. York 

 Acad. Set., viii., pp. 200 — 201), but they are too general to be of 

 much service here. Nothing, however, that this author has written, 

 lays itself out for criticism more than does his attachment of the 

 Notodonts and Lymantriids to the Lachneid branch of his phylogenetic 

 tree ( Proc. Bost. Soc. Nat. Hist., xxvii., p. 146, and Journal N. Y. 

 Ent. Soc, iv., pi. iii). Here he combines the Notodonts, Eupterotids, 

 Lymantriids, Lemoniids and Lachneids in one evolutionary stem. 

 The synopsis which he here gives of the characters for the separation 

 of the Attacids and Sphingids appears, however, to be sound 

 enough. This reads : 



Subprimary setae absent or all greatly obscured after the first moult by 

 secondary characters : 



1. Tubercles iv and v united, all the setae borne on 

 prolonged tubercles subject to various modifications. 

 Usually an unpaired tubercle on the 8th or 9th abdominal 



segment . . . . . . . . . . Saturniides 



(Attacides; . 



2. Tubercles iv and v remote, v higher than iv, 

 obscured after first moult. An unpaired process bearing 



tubercles i on the 8th abdominal segment . . Sphingides. 



Packard gives {Proc. Amer. Phil. Soc, xxxi., pp. 139 — 192) an inter- 

 esting account of the Citheroniidae {Ceratocampidae), supposed by some 

 lepidopterists to be the connecting group between the Attacids and 

 Sphingids. He considers the group to be well circumscribed, and the 

 most generalised forms to be D ryocampa and Anisota, whilst Sphingicampa 

 may be regarded as transitional, connecting Dryocdmpa and Anisota with 

 Eacles and atheroma. His note that the larvae of Aglia (tan) and 

 Packs {imperialis) in their third stadia resemble each other very 

 closely is interesting, still more so, that Aglia, at its last ecdysis, 

 passes (so far as the larval appearance is concerned) from one family 

 to another, and that "the ontogenetic development of this larva 

 epitomises that of two families, whereas that of most Bombyces is 

 simply usually only an epitome of that of a subdivision of a family, 

 or of a small group of genera." Packard, after further discussion, 

 suggests that Aglia should be regarded as the type of a distinct subfamily 

 of Ceratocampidae (= Citheroniidae), and that the latter family might thus 

 be divided into the two subfamilies, Ceratocampinae and Agliinae ; whilst 

 from their larval and imaginal characters, and in their mode of 

 spinning a cocoon, he is disposed to consider the Hemileucidae as a family 

 closely allied to, though distinct from, the Citheroniidae ( Ceratocampidae). 

 Later he makes (Bombycine Moths of America, p. 39) Aglia a connecting 



