BY R. J. TILLYARD. 681 



disposal was only excelled by that of Wingia, and I was able 

 to study the wings at all stages of pupal life. 



In order to save space in the figures, and also to show a 

 stage different from any of those exhibited in the case of Wingia, 

 I have drawn, in Text-fig. 99, a, both the t racheation and vena- 

 tion of the wings of a ten-days-old pupa. The pale bands, 

 which foreshadow the imaginal venation proper, are indicated by 

 the double dotted lines, each pair of which encloses a trachea, 

 represented by a continuous black line. Alongside this figure, 

 in Text-fig. 99, b, is shown the basal portion of the forewing in 

 the freshly-turned pupa, somewhat more enlarged. From this 

 we can see how well separated the bases of the main tracheae 

 are at first, and how closely they tend to approximate as pupal 

 life progresses. An exception is the group of three anal 

 tracheae, which arise, as usual, very close together in a single 

 bunch . 



The following important points should be noted: — 



(1) In the forewing, R 2+3 an ^ R4+5 eome °ff widely separ- 

 ated from the main stem of R . This is a characteristic of the 

 pupal forewing of all true Butterflies (with a few exceptions 

 in very highly specialised eases), and marks Euschemon de- 

 finitely as belonging to the Rhopalocera. 



(2) Although R 2+3 is much shorter than R 4+5 , it forks nearer 

 to the base than does the latter. This agrees with Xgleutes, 

 Wingia, etc., but differs from the more archaic condition seen 

 in the Hepialidae. 



(3) In the hindwing, Sc and Rj fuse to form a single vein; 

 but trachea Se is dominant, as in Xyleutesj whereas, in Carpo- 

 capsa, Wingia, and Doratifera, R'i is dominant. 



- (4) The development of a very broad humeral area in the 

 hindwing has thrown the frenulum back right to the extreme 

 base of the wing (Text-fig. 100), carrying with it the remnant 

 of the humeral veinlet, which appears as a pale band in the early 

 pupa, but does not become fully chitinised in the imago. Pro- 

 bably correlated with this is the shortening of the separate basal 

 portions of Sc and R a in the imago to form a small loop. This 

 loop is characteristic of the Besperiidae; and. in many genera, 

 the veihlet hm is to be seen projecting from it basad, much more 

 strongl'y than in Euschemon. In other Butterflies, this loop is 

 reduced to a minute vestige, or completely closed up. 



(5) In the early pupa, M g forms as a very distinct pale band 



