io6 Records oj the Indian Miisenni. [Vol. XXII, 



tion of A at its junction with CUj might l^e brought about as a 

 purely imaginal character. Also, the pigment bands of the larval 

 wing are so wide that, even if some slight angulation is really present 

 there, it might easily be overlooked. As regards the double row 

 of cells below C;(.j in the hind-wing, it may be noted that a closely 

 similar set occurs in the forewing of Needham's specimen (text- 

 fig. 5) ; so that we may regard this character as being probably a 

 variable one in the different wings of separate individuals of E. 

 supersies. Even if we grant the presence of these differences, they 

 are not of greater moment than such as we should expect to find 

 in two species of the same genus, from such widely different locali- 

 ties as Japan and the Himalayas. 



From the above evidence, it may, I think, be legitimately 

 concluded that the larva belongs to the genus Rpiophlchia. 



The Suborder Anisozygoptera. 



We have still to answer the question as to whether the 

 erection of the Suborder Anisozygoptera by Handhrsch was 

 justifiable ; and, in particular, whether the discovery of this larva 

 adds to or detracts from HandUrsch's opinion. 



It seems clear that we must now answer this question in 

 Handlirsch's favour. For a more complete blending of Anisop- 

 terous with Zygopterous characters, within one single larval type, 

 could scarcely have been hoped for, even by the most ardent 

 supporter of Handlirsch's view. The larval evidence is so strong 

 that, taking it in conjunction with the imaginal characters already 

 known, I have no hesitation now in accepting Handlirsch's Sub- 

 order Anisozygoptera ; and, consequently, a family Epiophlehiidae 

 must also be recognised. Also, as it is clear for many reasons that 

 the fossil type Heterophlebia. from the English Lower Lias, is a close 

 ally of Epiophlebia , and the same is true of the genus Triassoles- 

 tes, from the Upper Trias of Ipswich, Queensland, it follows that 

 this Suborder is the oldest of the three at present existing, so far 

 as our evidence goes. The fact that we have, in Heterophlebia, a 

 type in process of changing from an Anisozygopteron to a true 

 Anisopteron, by formation of a true triangle and supertriangle in 

 the hind- wing, seems to indicate clearly enough that the Anisoptera 

 are descended from the Anisozygoptera ; and, indeed, true Anisop- 

 tera do not appear in the fossil record before the Upper Lias. 

 There does not seem to be as definite evidence that the Zygoptera 

 are descended from the Anisozygoptera ; for there are certainly 

 some venational characters, notably the absence of an oblique vein, 

 in which the Calopterygidae still remain more generalised than 

 Epiophlebia and its allies. Our decision in this case would probab- 

 ly rest upon what forms amongst the earliest known fossil 

 Odonata we were prepared to accept as Anisozygopterous, on the 

 evidence of their wing-venation, and what forms we considered 

 Zygopterous. Probably the earliest true Odonate type combined 

 the more generalised characters of the Calopterygidae on the one 

 hand with those of the Epiophlehiidae on the other. Such a type 



