THE EXTERNAL STRUCTURE OF THE LEPIDOPTEROUS LARVA. 47 



lost on the posterior one ; the sub-stigmatal tubercles are retained 

 and approximated, the anterior one of the four on the base of the 

 leg seems to have been moved up, forming tubercle vi, which is 

 thus anterior ( = pre-spiracular). This explanation accounts for the 

 possible formation of the pre-spiracular tubercle as such, for it will be 

 observed that, whereas tubercle v of Dyar is the typical sub-spiracular 

 tubercle of the more specialised families, tubercles iv and vi, typically 

 originating below the spiracle, according to Dyar, become respectively 

 the post-spiracular and pre-spiracular in special instances. In all 

 the other families of the Lepidoptera, Dyar states that the middle 

 tubercle of the three on the middle sub-segment is lost, but the upper 

 on the posterior sub -segment is retained ; the two (iv and v) below 

 the spiracle are also retained, as in the Psychids, but they are either 

 approximated (sometimes even united to form a compound sub- 

 spiracular tubercle, as is Margarodia), or separated so as to form two 

 distinct tubercles, viz., the sub-spiracular and post-spiracular, whilst 

 of the four tubercles at the base of the leg, the posterior one (not the 

 anterior one, as is the case in the Psychids) is moved up to form 

 tubercle vi. 



The tendency for tubercles iv and v to coalesce so as to form a 

 compound sub-spiracular tubercle, appears to be characteristic of the 

 larvas which comprise, in its broad lines, Comstock's Microfrenatje or 

 Generalised Frenat;e, whilst the tendency for tubercles iv and v to 

 separate and form post-spiracular and sub-spiracular tubercles, re- 

 spectively, appears to be characteristic of his Specialised Frenatje. 



Dyar notes, and if it held good it would be very curious, that " it is a 

 striking fact that we do not find a series of intergrading forms between 

 the single-haired tubercle and the many-haired wart, though both may 

 occur in different genera of the same family," and he considers that 

 this is explicable on the principle of discontinuous variation, which is 

 insisted upon by Bateson. He says that in the lower (more generalised) 

 families we have the simple and primitive form of tubercle ; in the more 

 specialised families we find a modification, which consists in the tubercles 

 becoming enlarged and many-haired. In these compound tubercles each 

 hair arises from its own minute tubercle, and the whole are borne upon 

 an enlarged base or wart. Modification then takes place in the higher 

 groups, by a reduction in the number of tubercles, the reduction taking 

 place : — (a) By coalescence, (b) By unequal development and final 

 obliteration of particular ones. (This is discussed later in chapter.) 



We have seen that in some of the more specialised larvae there is a 

 general tendency to the reduction of tubercles, so that some may 

 entirely disappear. In some cases, however, the bases of the tubercles 

 are developed into long fleshy processes, carrying aborted setae, as in 

 the case of certain larvae of the Nymphalids, Papilionids, etc. In other 

 cases, the setae remain as glandular hairs, in some instances secreting an 

 urticating (? odorous) fluid, or the hairs themselves become highly 

 specialised, and greatly increased in number, forming brushes, tufts, 

 plumes, etc., as in the larvae of Acronyctids, Liparids, Arctiids, etc. 



One of the most striking modifications of the tubercles is seen in 

 the caudal horn of the Sphingides. This is an unpaired dorsal process 

 on the 8th abdominal segment. A figure of the larva of Deilephila 

 enphorbiae (Weismann, Studies in the Theory of Descent, PI. v., fig. 38) 

 in its first skin, shows that the two setae of tubercle i are borne on the 



