INTRODUCTION. xv 



Since, then, all attempts at a linear arrangement must be artificial, and 

 the classification of families into orders is impossible in the present state 

 of our knowledge, I have arranged the families in the old but, to a large 

 extent, artificial sequence adopted by Cuvier, which has at least the prac 

 tical value that it is well known, and thus obviates to a large extent the 

 trouble of reference to an index. I have been careful to point out under 

 each family whether, in the opinion of the best informed naturalists, it is 

 nearly connected or not with the families near which it is placed. 



There is no department of ornithology which has received more atten- 

 tion of late years than that of 



Geographical Distribution ; 



and there is no subject more intimately connected with the discrimination 

 of species and with the whole question of classification. The zoological 

 divisions into which naturalists divide the world are not the same as those 

 in common use among geographers. So far as these boundaries are deep- 

 sea, they may be said to be practically the same. Where there are no 

 changes of climate to make it imperative upon birds to migrate it is 

 remarkable how seldom they use their powers of flight to wander far from 

 home. Even the narrow channel of deep sea between Borneo and Celebes 

 marks an important boundary in the geographical distribution of birds, 

 whilst the shallow Mediterranean is of little significance. 



The land-boundaries of the zoological regions are climatic. The world 

 is divided into six or seven zoological regions. 



The Palsearctic Region contains the whole of the Old World north of the 

 desert of Sahara in Africa and north of the Himalayas in Asia. 



The Nearctic Region contains the New World north of the tropics, i. e. 

 north of Mexico. 



The Ethiopian Region consists of Africa south of the great desert, and 

 Southern Arabia. 



The Oriental Region consists of Asia south of the Himalayas, and the 

 islands of the Malay archipelago as far east as Borneo. 



Tlie Australian Region consists of the rest of the islands of the Pacific. 



The Circumpolar Region has the north pole for its centre, and extends 

 to the Arctic circle in the Old World, and somewhat further south in the 

 New ; but many naturalists do not recognize this region as zoologically 

 distinct. 



Perhaps the most interesting fact connected witli these divisions is that 

 in the tropical regions most birds vary much less than they do in the arctic 

 regions. If we eliminate the arctic genera, which are comparatively recent 

 emigi'ants, we shall find that the tropical species are generally well defined ; 

 they are obviously ancient residents who have well nigh exhausted the 

 variations required to adapt them to their surroundings, which must have 



