164 ENDEMISM AND DISTRIBUTION: SPECIES [pt. ii 



is one species covering the whole range of dispersal, the classifica- 

 tion can be carried into greater detail, but even in such cases as 

 Cyrtandra, where there is not such a species, the phenomenon 

 can be quite clearly seen. It is evidently perfectly general, and 

 we shall see many further examples of it in the next chapter, 

 and go on to consider its general bearings in later chapters. 



When one goes on to examine into the genera and families to 

 which endemic species chiefly belong, one discovers that in most 

 countries the bulk of the endemic species do not belong to the 

 endemic genera. Even in a region of such marked endemism as 

 the Hawaiian Islands (37), for example, where there are several 

 very large endemic genera, only 225 out of 581 endemic species 

 belong to the endemic genera, or 38 per cent. In New Zealand 

 less than 5 per cent, do so, and in Brazil perhaps 10 per cent. 



The numerous endemic species that do not belong to endemic 

 genera are found on examination to belong, not, as one might 

 perhaps expect, to small and broken genera, Avhich we have been 

 accustomed to consider moribund, but in greater proportion to 

 the larger and more important genera. The average number of 

 species in a genus, taking the whole world, is about 12-7, and in 

 the Hawaiian Islands, taking the first hundred genera in the 

 flora (37), we find that of the 47 that contain endemics, but are 

 themselves widely dispersed, 36 are above the average size in 

 the world, and have 102 local endemics, while 11 are below, and 

 have 22 endemics. Of these 11 belong to Lipochaeta, which only 

 occurs outside these islands as a single species in the Galapagos. 

 The average size of the whole 47 genera (in the world) is 97 

 species, or eight times the average. Of these genera 8 are cosmo- 

 politan in their dispersal, 11 are tropical and subtropical, 8 are 

 tropical, these three categories including 57 per cent, of the total 

 (cf. Chapter xii, Size and Space). A further 9, bringing the total 

 to 76 per cent., occur in both Old and New AVorlds. 



If we turn to New Zealand, and take the first 100 genera (37), 

 of those with endemics 43 are above the average in the world, 

 and only 14 below, while the average world-size of one of these 

 57 is 73 species, or six times the world-average. The same thing 

 shows Avherever I have tested it. For example, if one take the 

 first 100 genera in Vol. iv of the Indian Flora (37), most of them 

 as it happens being Asclepiads, which are imusually small genera, 

 one finds 52 non-endemic genera, of which 38 are above, and 

 14 below, the average world-size. The remaining 48 are largely 

 endemic genera, for India, like all large areas, has a greater pro- 



