371 



desirable. While still convinced that the character lending 

 itself best to the formation of primary groups and doing 

 least violence to natural affinities is the relation of the 

 labrum to the clypeus, I have found that it is undesirable 

 to retain the second of the three primary groups for which 

 I employed it in my former series of papers, and which was 

 characterized as having the relation to each other of the 

 labrum and clypeus intermediate between their relation in 

 the other two groups. The first essential in the artificial 

 grouping of an extensive series of species is to form the pri- 

 mary aggregates in such fashion that there can be no doubt 

 at all in distributing the species among them. In my former 

 classification I divided the species into those having the lab- 

 rum entirely below the level of the clypeus, those having the 

 summit of the labrum on a level with the clypeus, and 

 those having the clypeus below the level of the summit of 

 the labrum. Working through the long series of undescribod 

 species now in my hands, I have found that the transition 

 of the position of the summit of the labrum is very gradual, 

 so that in not a few species that would have to be placed 

 in the intermediate group the relation of the labrum to the 

 clypeus would differ very little indeed from its relation in 

 species that would fall into one or other of the other two 

 groups. I therefore propose to abolish the second primary 

 group altogether, and to distribute its species between what 

 I formerly called the first and the third groups, expressing, 

 however, the characters that distinguish those groups (now 

 the first and second) in altered terms, for the comprehension 

 of which a little explanation is necessary. In a very large 

 majority of the Heteronyces (including the whole of the 

 group formerly called the third) the relation of the labrum 

 to the clypeus is such that if the head be looked at somewhat 

 obliquely from behind, the middle part of the free outline of 

 the head is the labrum, and that that part is a convex curve 

 distinct from the curve of the clypeus on either side of it, 

 so that the free outline of the head appears trilobed, or at 

 least trisinuate. There is the greatest possible specific diver- 

 sity in the nature of this outline in other respects — the middle 

 lobe or arch (?.f., the labrum) varying from being much 

 more than a third of the whole free outline, to being little 

 more than a narrow inequality in the curve of the free out- 

 line (in some instances conspicuous only from certain points 

 of view), and from being but little projected forward (so 

 that the summits of the clypeal curves reach forward con- 

 siderably more than it does) to being a conspicuous proboscis- 

 like lamina protruding from the front of the head; but in 

 all the species which I group together by this character there 



