050 THE PANORrOID COMPLEX, iii., 



neura of to-day; and the reduction in the number of veins is 

 quite obvious also. It is, therefore, necessary to study, first 

 of all, those types in which the hindwings have been preserved. 

 We shall then be able to compare the forewings of those types 

 with the isolated forewings of Palaeontina and Palaeocossus, in 

 order to determine whether these two latter genera are related 

 to them, or whether, perhaps, they may not be regarded as true 

 Homoneura, not related to the rest. 



First of all, then, we must make a comparison of the four 

 hindwings shown, and note the points common to all in Hand- 

 lirsch's restorations. They all agree in one very remarkable 

 character, viz. that the media is reduced to a single straight vein. 

 As far as I can see, from the photographs given by Handlirsch, 

 there is no doubt on this point. Rs is two-branched in every 

 case; so that it is clear that its two branches may be named 

 R2+3 and R 4+5 respectively. R x is present in all; and in at 

 least one case a perfectly separate Se is also present (Lima- 

 codites) . Cu is three-branched in all cases; and one naturally 

 concludes that these three branches are the homologues of the 

 three branches existing in the Order to-day. (But it may be 

 noted that it would be possible to interpret the vein marked 1A 

 in Limacodites as Cu 2 in which case the vein marked Cu la must 

 be a branch of M captured by Cu 1 , as may be seen in many 

 highly specialised forms existing to-day. 



The hindwing of the Palaeontinidae, then, had the following 

 veins present-: — Sc (? sometimes absent), R 1} R 2+ 3, R44-5, M> 

 Cuj , Cui b , Cii2, 1A, 2A (? sometimes absent). This is a possible 

 total of ten veins at the most, out of the thirteen or fourteen 

 that should be present in a truly Homoneurous hindwing. Thus 

 the conclusion is arrived at that the Palaeontinidae whose hind- 

 wings are preserved are not Homoneura. 



They must therefore be either (a) Heteroneura derived from the 

 same stem as those of the present clay, or (5) Heteroneurous 

 forms derived from the original Homoneurous stem by an en- 

 tirely separate line of descent. 



Now, as we shall see in our examination of recent Hetero- 

 neurous types below, these latter have all been evolved along a 

 single line of descent, from what Turner has called the 

 Protocossid ancestral type (32) . The method of reduction, in 

 the evolution of that type, has been solely by loss of the 

 branches of Rs, and not by loss of the branches of M. Further, 



