CH. XI] APPLICATION OF AGE AND AREA 109 



of both can be graphically represented by " hollow curves^," like 

 those in the fig. on p. 155 (and cf. clearer figure on p. 174), with 

 very many species occurring upon very small areas, the numbers 

 rapidly diminishing towards the areas of moderate size and then 

 more slowly to those of large size. 



In view of these and many other facts brought up, and of 

 which a simimary is given upon p. 159, it is no longer possible, 

 except in comparatively rare cases, to regard endemic species 

 either as relics or as special local adaptations ; though of course 

 if not adapted to the local conditions as they existed at the time 

 of their birth, they would be promptly killed out by natural 

 selection. The explanation offered by Age and Area, that species 

 of very small area of dispersal are in general young beginners, 

 and that area occupied increases with age, seems the only pos- 

 sible one for the great majority of species. Not only so, but age 

 proves to be by far the most important factor in the dispersal. 



In Chapter xvi Endemism and Distribution of Genera are 

 dealt with, and it is shown that the phenomena presented are 

 exactly parallel to those exhibited by species, and that the dis- 

 tribution of endemic genera is similarly a miniature of that pre- 

 sented by genera as a whole. The areas occupied by the genera 

 of a given family are arranged like those occupied by the species 

 of a given genus. There are very many upon comparatively 

 small areas, and many on the areas just a little larger, whilst 

 there are but few upon areas that are really large. As one would 

 expect from a consideration of the hyjDothesis of Size and Space, 

 one finds that the sizes of the genera themselves (in number of 

 their species) go mainly with the area occupied, so long as one 

 keeps to the allied forms of a single family. The bulk of the 

 genera of very small area are monotypic, or have but one species 

 each, Avhile the bulk of those of very large area have very many 

 species (average 59), those with intermediate size of area having 

 intermediate numbers of species. Plotting of the genera, whether 

 by size or by area, thus gives hollow curves. While the latter 

 represents their geographical distribution, the former obviously 

 represents their evolution. 



The same hollow curve type of distribution shows ilsclf if one 



^ The "hollow curve" arises when numbers are plotted as a graphic curve 

 which are large for the first two or three cases (e.g. in tlie fig. on p. 174 the 

 first three are -10, 15, and 8, or much more than half the total of 100), and 

 then taper away gradually in a tail (e.g. the remaining 37 are divided among 

 the groups of families from the 4th to the 29th). There is a large drop from 

 the first to the second, and from the second to the third or fourth. 



