CLASSIFICATION OF LEPIDOPTERA. 107 



diagrammatic table that supplements it, gives us but little practical 

 help in our knowledge of the broad lines of evolution along which 

 the Lepidoptera have travelled. Unfortunately we are not able 

 to follow the table, even in its broadest lines, for reasons that are 

 self-evident, <?.//., the derivation of Pyralids from Pterophorids, and 

 the latter from Alucitids, which Chapman has shown f to be impossible ; 

 the derivation of Lasiocampids from Lithosiids, and Notodonts from 

 Lasiocampids, which the eggs show to be equally impossible, and so on. 



It is quite evident that the evolution of the many specialised super- 

 families has taken place from the generalised, and that the former are 

 the most recent evolutionary products of certain stems of which the 

 generalised are the older offshoots. What is needed, then, is some 

 character (or characters) that will not slice off horizontally, as it were, 

 all the branches of the genealogical tree, leaving (1) the upper super- 

 families, composed of the DETECTS or SPECIALISED-FRENAT^E, and (2) the 

 lower, comprising the INCOMPLETE or GENERALISED-FRENA, but one 

 which will give us clues as to the development of the branches themselves 

 vertically, and separate into their own particular branch the specialised 

 and generalised superfamilies belonging thereto. In this way alone can 

 we get a true conception of the genealogical relationship of the various 

 families to each other. 



It might have been supposed that Dyar's studies of the larval tubercles 

 would have led him to have constructed a tree satisfying the necessary 

 conditions, but it has not done so. One of his latest J (if not the latest) 

 pronouncements on the subject, satisfactory as it is in many ways, 

 leaves us much as we were. It works out as follows : 



I. Tubercles iv and v approximate or consolidated. 



1. Tubercles i and ii remote MICROLEPIDOPTERA. 



2. Tubercles i and ii consolidated ANTHROCEKINA. 



3. Tubercles i and ii remote, ii disappearing at 



the first moult BOMBYCINA. 



II, Tubercles iv and v remote. 



1. Tubercle iv behind the spiracle, v below it .. NOCTUINA 



2. Tubercle iv below, v in front of the spiracle SPHINGINA. 



3. Tubercles iv and v in line, except in some 

 Nymphalidae, where secondary armour is 

 developed KHOPALOCEKA. 



From this Dyar gets the following groups : 



I. The MICEOLEPIDOPTERA, including the Psychidae, Cossidae, Pyralidae, 



Tortricidae, Sesiidae, Tineidae and Lacosomidae. 

 II. The ANTHEOCEKINA, including the Pteroplioridae, Anthroceridae, Pyro- 



morphidae, Megalopyyidae and Eucleidae. 

 III. The BOMBYCINA, including the Citheroniidae, Hemileucidae, Saturniidae, 



and Bombycidae. 

 IV. The NOCTUINA, including the Notodontidae, Thyatiridae, Geometridae, 



Drepanidae, Agaristidae, Noctuidae, Cymbidae, Lithosiidae, Pericopidae, 



Arctiidae, Euchromiidae, Lymantriidae, and perhaps also the Thyri- 



didae, Dio2>tidae and Lasiocampidae. 

 V. The SPHINGINA, including the Sphingidae. 

 VI. The EHOPALOCEBA, including the families usually associated under this 



term. 



* American Naturalist, 1895, p. 803. 



t Trans. Ent. Soc. Land., 1896, pp. 129 et seq. and Entom. Record, vol. vii., 

 pp., 268 et seq. 



t " Relationship of Pyralidffi and Pterophoridaa from the larvae," Ent. Neivs t 

 Feb., 1895. 



