4 THE FAUNA OF SCOTLAND. 



But the converse of this, namely, that the possession of the 

 fewest number of joints marks the most recent origin, does not 

 necessarily follow. For, as has just been remarked, it is precisely 

 in those groups which now have the greatest and fewest number 

 of joints that sexual selection has played the greatest rule in 

 causing tlie greatest difference between the sexes ; and there can, 

 I think, be no hesitation in affirming, that the peculiar antennae 

 of Hylotoma have been produced by that agency, since everything 

 tends to show that the original ancestor of the family must have 

 had multe-articulate antennae, while in • the other Hymenoptera, 

 as well as among insects generally, it is among the lower tribes 

 that secondary sexual characters are most marked. 



So far as the division into tribes and sub-tribes goes, my 

 arrangement does not differ much from that of Thomson, whose 

 work (except in so far that nothing or next to nothing is said 

 about the larvae), is one of the best and most suggestive that 

 has yet appeared on the family. We differ in so much that 

 I form Lo2)hyriis and its allies into a distinct tribe, instead of 

 regarding them as only a sub-tribe of the Tcnthredina. The 

 Nematides again are so sharply cut off from the rest of the family, 

 not only by the structure of the perfect insects, but also by the 

 larvae having 20 legs, that I follow Prof. Zaddach in forming them 

 into a tribe. They are naturally united to the Tenthredina by the 

 strong affinities Dineura has with Ilojjiocampa. The last mentioned 

 genus too has gall-making species, while in the alar neuration it 

 is almost identical with Dineura. AVith the sub-tribes again I 

 form the Phyllotoma section into a sub-tribe (as was done by 

 Newman under the name of Druidae, without his having a very 

 clear idea of the group), which their peculiar habits and structure 

 appear quite to warrant. Strongijlogaster, Poecilosoma, Taxonus, and 

 Emijhjtus, I form also into a sub-tribe. These genera appear 

 to me to have been most unnaturally separated in our books, 

 notwithstanding that they are very closely related in many 

 respects, while the larvae are very similar, and all agree in 

 pupating in stems (either of their food plants or others), without 

 spinning any cocoons. 



To this union it may be objected that the four genera exhibit con- 

 siderable diversity in their alar neuration. It is true that EinpJujtus 

 has only three marginal cellules, while Poecilosoma has one, and 

 StrongyJogaster two medial cellules in the posterior wing. But 



