INTRODUCTION 13 



the regions, and hence the difficulty in determining how 

 many of the border-forms, which have obviously intruded 

 from the neighbouring regions, should be counted. 



(2) The uncertainty as to the limits of the genera. 

 This uncertainty has been greatly increased of late years by 

 the action of some zoologists in proposing a multitude of 

 unnecessary generic terms. 



When these two factors have been settled and the lists 

 constructed, a further difficulty is met with, and this is one 

 which depends very much on the individual fancy of the 

 author, namely, as to the percentage of peculiarity which 

 should be required to constitute a Region. 



Taking the first question in dispute, we find that Mr. 

 Allen, in his paper already quoted (2), gives a tabulated 

 list of the genera of his North Temperate realm, dividing 

 them into North American and Eur-asiatic ( = Palsearctic) 

 forms, and putting the individual genera into three cate- 

 gories, namely, those circuru -polar, or common to the Nearc- 

 tic and Paloearctic Regions (numbering thirty-two); those 

 peculiar to each Region (i.e. twenty-nine to the Nearctic and 

 forty-one to the Pala?arctic) ; and, finally, those which range 

 further south into the Neotropical Region on the one hand, 

 and into the Oriental and Ethiopian Regions on the other. 



Working from these tables we find that 38 per cent, of 

 the Nearctic genera and 42 per cent, of the Pakearctic 

 genera are confined to their respective Regions, while 42 

 per cent, in the case of the Nearctic and 34 per cent, in the 

 case of the Pakearctic are common to the two regions. 

 These last percentages include, however, several quite wide- 

 spread genera which can hardly be called circum-polar — such 

 as Sri a, 'us, Kid ii rojifrriix, Lcpus, Lutra, Canis, and Fells. 



These figures show that there is, as has indeed never 



