484 NATURAL HISTORY OF BIRDS. 



of this vast multitude of forms are confessedly anything but natural. Several single 

 characters have been employed, but in every case the result has been that some forms 

 were separated from their nearest allies, the relationship of which is so clear that it 

 cannot be disputed ; hence the systematist was compelled to take them into a group 

 from which the characters given excluded them and made them ' exceptions.' It has 

 repeatedly been attempted to establish sections higher than families, but without suc- 

 cess. "Wallace's proposition to employ the number of primaries as an easy means of 

 separating large groups has received considerable favor, but the fact that none of the 

 groups are without 'exceptions,' and that nine and ten primaries may occur within 



the same genus, makes the scheme valueless in 

 a natural classification. The fact is, really, that 

 the tenth (first) primary is not absolutely want- 

 ing in the ' nine-primaried ' Passeres, but its 

 size is so extremely reduced as to become quite 

 or nearly invisible in the old birds, the more so 

 since its position is also slightly changed, as it 

 is forced up on the upper surface of the wing. 

 Not so m the young, however. In a very young 



FI. 235. - primaries and first secondaries of specimen of the snow-flake (Plectrophemtx niva- 

 wing of fledgling piectrophennx enlarged one # g \ a t nine-primaried ' species, now before me, 



and one half times, from above ; a, alula ; c9p, 



covert of penultimate, and cWp, covert of last t ] ie outermost (tenth) primary is plainly visible 



primary ; gc, greater coverts ; p, primaries ; 9/>, J J 



penultimate primary ; Wp, ultimate (first) pri- j n Jt S natural position, and with the COlTCSpOnd- 

 uiary ; s, secondaries. 



ing great covert in its proper place, that is, in 



the interval between the ultimate and penultimate primaries, as shown in the accom- 

 panying cut (Fig. 235). From this will be easily perceived how perverse is the method 

 of counting the primaries from the edge of the wing, since, in nine-primaried birds, 

 the feather which is usually called the first in reality corresponds to the second of 

 the ten-primaried species. By counting from the secondaries, no difficulty is 

 experienced. 



Since Cuvier's days, however, the Passeres have most commonly been gathered 

 into groups according to the shape of the bill, and the section names ending in -rostres 

 are familiar to everybody. For convenience they have been employed even in the 

 latest systems, though confessedly only in default of something better. Believing 

 that their retention is a decided obstacle to a natural arrangement, we shall take no 

 further notice of them. It may be proper to enumerate them, however, since they 

 play a considerable role in the ornithological nomenclature. According to Sclater's 

 arrangement of the laminiplantar Passeres, these sections are as follows : 



Dentirostres, comprising thrushes, wrens, warblers, tits, Old World fly-catchers, pipits, etc. ; 



Latirostres, for the swallows; 



Curvirostres, creepers and nuthatches; 



Tenuirostres, sun-birds, honey-suckers, etc. ; 



Conirostres, finches, weaver-birds, tanagers; 



Cultrirostres, crows, pies, jays, starlings, grackles, birds-of-Paradise, etc. 



We do not claim, however, that the arrangement to be applied in the following is 

 quite natural in all its detail. The exact affinities of many forms are as yet very ob- 

 scure, or in dispute, while in many other instances we are still ignorant of the nature 

 of some important characters, whether we shall regard them as generalized or spe- 

 cialized features. But we must warn against any criticism charging unnaturalness on 



