136 SCLATER ON THE GENERAL DISTRIBUTION OF AVES. 



tion of the Australian type, and I should be mclined for the pre- 

 sent not to separate New Zealand and the Pacific Islands generally 

 from the Australian division. We should have, therefore, in the 

 Old World one temperate region and three tropical ; the eastern 

 palseotropical or Australian advancing rather farther to the south 

 than the others, the Indian or middle palseotropical being the most 

 northern of the three. 



In the New World we can simply divide the continent into 

 northern and southern divisions ; the northern, or Nearctic region, 

 extending down the centre of the table-land of Mexico, and show- 

 ing some indication of parallelism to the Palaearctic by the presence 

 of certain temperate types ; the Neotropical or southern (which 

 embraces the whole of the rest of this great continent) being 

 wholly free from any admixture of the sort, and in fact exhibiting, 

 in my opinion (with the exception possibly of New Guinea), by 

 far the richest and most peculiar Avi-fawna of the world's surface. 



Having thus pointed out what I consider to be the primary 

 divisions of the earth, — taking ornithology as our guide, I propose 

 to devote a few lines to each region separately, noticing its appa- 

 rent limits, its peculiarities, and most characteristic forms, and 

 attempting to give an approximate estimate of the comparative 

 abundance of ornithic species within its area. 



The subjoined plan will serve to give at one view an illustration 

 of my ideas as to the arrangement of these primary Avi-fawncd of 

 the earth's surface. It must, however, be recollected that the 

 calculations made as to the number of species to a square mile, can 

 be only looked upon as mere attempts at approximations. Even 

 in the whole general calculation, the presence of two variable ele- 

 ments — in the first place the number of square miles (about which 

 geographers still give the most conflicting statements), and in the 

 second place, the number of species of birds, concerning which 

 ornithologists are as yet by no means agreed, greatly increases the 

 uncertainty of the ratio deducible from them ; and in working out 

 the ratios in the respective regions, it is of course still more difii- 

 cult to attain to any great degree of accuracy. 



Taking however the whole number of square miles of dry land 

 at 45,000,000, and the number of species of birds at 7500, which 

 are both of them moderate estimates, we have on the average a 

 single species to each 6000 square miles. In the different regions 

 we shall attempt to show how far this ratio is departed from. 



The zoological kingdoms or primary divisions are of course natu- 

 rally separable into secondary divisions or provinces, but it would 



