INTRODUCTION. 



ate one of the most cosmopolitan groups of birds, and undoubtedly 

 the most cosmopolitan of Passerine birds. From these two cha- 

 racters, their high development and the cosmopolitan character of 

 their distribution, we may infer that they are an example of the 

 most numerous, most recent, and most nearly allied groups of 

 birds. In fact they form one enormous genus. There has been no 

 time or opportunity for intermediate forms to die out. Nature 

 has not yet split them up into genera ; in fact she has as yet only 

 very imperfectly segregated them into species. Under these cir- 

 cumstances their subdivision into generic groups is to some extent 

 arbitrary. I have endeavoured to choose my types of genera as 

 carefully as possible, and to define their limits (;'. e. to draw a line 

 where nature has di-awn none) with as much judgment as possible. 

 The keys to the genera are, I admit, very artificial, and will require 

 modification when new species are discovered ; but I have endea- 

 voured to characterize the genera at length, and in most cases I 

 trust that the instinct of the ornithological student will enable him 

 to tell a Chat, for example, from a Eedstart, without being able to 

 define the characters of each genus. 



It is also due to the student that some explanation should be 

 given of the way in which conspecies *, or forms between which 

 the difierence is only subspeeific, are treated in this volume. In 

 the previous volumes of the ' Catalogue of Birds ' subspeeific forms 

 are recognized. This is an immense stride upon the hard and tight 

 system of previous English ornithologists who refuse to acknow- 

 ledge the imperfectly segregated species which undoubtedly exist 

 in nature, because forsooth their binomial system of nomenclature 

 docs not easily lend itself to their discrimination. The American 

 system, clumsy as it is, has undoubtedly the advantage of being 

 far in advance of the old mode. We must all admit that Corone 

 macrorhyncha, var. levaiUanti is most objectionable as a name for 

 an extreme form, and one, moreover, -which leaves the intermediate 

 forms nameless. It is very important that no mistake should be 

 made as to what constitutes a conspecies. Two forms may be very 

 closely allied ; but if the difference between them, however small it 

 may be, is constant, and is not attributable to age, sex, or season, 



* I am indebted tu Professor Solilegel for tbis word. 



