XXXIII] LITERARY WORK, ETC., 1 887-1905 211 



In the year 1894 I read a paper to the Cambridge 

 Natural Science Club on the question, " What are Zoological 

 Regions?" which was printed in Nature (April 26). But 

 my conclusion — that the six regions first defined by Dr. P. L. 

 Sclater are, for all practical purposes of the study of distribu- 

 tion, the most convenient and those which best illustrate the 

 actual facts of nature — was contested by my friend, Professor 

 Alfred Newton as regards the Nearctic and Palsearctic 

 regions, which he contended formed but one natural region. 

 I therefore thought it necessary to go into the subject in 

 more detail, and contributed a paper to NatiLral Science in 

 the following June, entitled " The Palsearctic and Nearctic 

 Regions compared as regards the Families and Genera of 

 their Mammalia and Birds." 



The first of these papers was for the purpose of showing 

 that, to be of any practical use to naturalists, zoological regions 

 must be so defined as to serve to elucidate the distribution 

 of all land animals. This will be evident if we consider the 

 results of the contrary view, that many classes, orders, and 

 even families, require a special set of regions to exhibit their 

 distribution with any approach to accuracy. Now as there are 

 some hundreds of these groups in the animal kingdom, we 

 should, perhaps, require fifty or a hundred sets of zoological 

 regions — each set differing in the number of regions and in 

 the boundaries of each, involving a different set of names in 

 each case. The result would be that each specialist would 

 have his own set of regions, with different names and 

 different boundaries ; and as no one could be familiar with 

 all these, the conclusions of each could be unintelligible and 

 useless to others. With one set of regions, on the other 

 hand, the distribution in every case can be described in terms 

 which would be intelligible to all ; and the comparison of the 

 distribution of groups differing in powers of dispersal and in 

 other ways, would often lead to an explanation of the differ- 

 ences of distribution, which is the whole aim and end of the 

 study, and which, so far as I can see, can be arrived at in no 

 other way. 



The second and more technical paper was for the purpose 



