Mi\ A. R. Wallace on the Natural Airangement of Birds. 195 



perhaps, for the determination of affinities, of as much import- 

 ance as any which can be pointed out. Observations of this 

 nature have been applied by him to an arrangement of the 

 Passeres ; not as a perfect scheme, but as a starting-point to guide 

 future inquiries. One portion of this arrangement, with the 

 families included in which he is best acquainted, he now wishes 

 to submit to the judgment of ornithologists. 



The method illustrated at the commencement of this paper, of 

 marking off certain groups from the general mass, has been 

 satisfactory, because the portions so severed have been not only 

 capable of definition, but have contained only species which have 

 agreed in all essential points of their structure and oeconomy. 

 They have therefore met with general acceptance, and in all the 

 different systems of ornithologists these groups have scarcely 

 suffered any variation. But in attempting to carry out this 

 system in a farther division of the Passeres, no such satis- 

 factory and generally accepted results have been produced. No 

 syste matist has been satisfied with the arrangements of his 

 predecessors, and, after an endless variety of divisions and sub- 

 divisions, we are as far off fi-om any generally accepted system of 

 arrangement as ever. 



The reason of this wc conceive to be, that we have to deal 

 with a mass of species in which the series is so nearly complete, 

 that there are no more of those great chasms sej^arating consi- 

 derable portions from each other, and that the affinities are so 

 intricate and minutely varied, and so cut up as it were by minor 

 gaps between genera and families, that any attempt to form 

 great and well-marked subdivisions must fail, for the simple 

 reason that such are not marked out by nature. In such a case 

 an arrangement may be possible, but a classification may not be 

 so. We must therefore give up altogether the principle of divi- 

 sion, and employ that of agglutination or juxtaposition. We 

 shall best explain our meaning by pointing out the en'ors we 

 conceive to have been produced by the former method in most 

 modern classifications. 



The system of Cuvier, as modified by Vigors, Swainson, and 

 G. R. Gray, may be fairly taken as that most generally in use, at 

 least in this country. The tribes of the Conirostres, Deutirostres, 

 Fissirostres, Tcnuirostres, and Scansores, are said to be the na- 

 tural divisions of the Passeres, the main difference of opinion 

 being as to whether the Scansores should or should not form a sepa- 

 rate order, a question we believe to be of no importance what- 

 ever. These divisions being accepted, every bird is forced into 

 one of them, and the result has been the most incongruous and 

 unnatural combinations. For instance, in the Tcnuirostres are 

 combined the Humming Birds and the Sun Birds {Nectarinia), 

 families which in a natural arrangement would have, in our opi- 



13* 



