13 



Epichnopterygids, since, though they may possess no scales on the 

 dorsa of the pectinations (a few species still retain a few), neither do 

 they possess any sense-hairs, the dorsa being bare and smooth 

 except perchance for an odd sensory bristle, stiffer anil shorter 

 than the sensory hairs, which in nearly all the Psychids, above 

 and including Fumeidas, are very long. 



The Apteronidae are a branch of the Psychini, but have very spe- 

 cialised antennae ; the pectinations are very short, the sense-hairs 

 short and hidden away on the ventral aspect of the antenna, 

 beneath a great development of the dorsal scaling, here consisting 

 of hair-scales, which, projecting at the site of each pectination, give 

 the antenna the appearance of being serrated. 



It is curious that this differentiation of the highest Psychids 

 into two groups by antennal structure, should coincide with a 

 differentiation by the presence or absence of long anterior tibial 

 spurs. The Acanthopsychini (CEkcticus, La?isdowni, Metura, 

 Accuitlwpsyche, etc.) possess the spurs and lose the scaling of the 

 pectinations ; they also have special elaboration of the venation 

 of the opposed wing margins (generally). The Psychini that 

 retain the antennal scales lose the tibial spurs, and as a rule 

 have less elaboration of neuration. 



The cochlidid antenna with complete scale covering might be 

 treated in an off-hand manner as parallel with that of Nepticula, 

 etc., as an immediate result of the plasticity of antennas of the 

 lowest families. This would, I think, be an erroneous method. 

 No doubt it is a result of the plasticity of the antenna in the 

 most generalised groups, for the Cochlidids, curiously and elabo- 

 rately specialised as they are in some respects, are essentially a very 

 primitive form. 



The reasons for rejecting the ready method are that the forms 

 having the completely scaled antenna are the higher and not the 

 lower in the family, and that the scaling is abundant and irregular, 

 and not in the two regular rows we find in so many forms. 



This type of scaling is in so many places associated with pec- 

 tination, that, if it seems reasonable to do so, one looks for associa- 

 tion with pectination here. 



Now the Cochlididae are especially a family with pectinate an- 

 tennas, and are immediately or collaterally descended from the 

 Megalopygids— a family with antennae most fully pectinated. The 

 megalopygid antenna is also a most fully scaled antenna, and one 

 might say of it, and in some degree also of the pectinated antenna 

 of the Cochlidids, that, if it were not pectinated, it would certainly 

 be completely scaled. In Lagoa crispata the pectinations are not 

 scaled internally (nor are they in Cochlidids), but they are not only 

 scaled down their free outer borders, but are expanded at their tips, 

 where they have a very dense mat of scales, such as I do not think 

 I have seen in any other family, but reminding one very much of the 

 thickened scaled end of the pectinations in Incurvaria vuiscalella. 



