174 ENDEMISM AND DISTRIBUTION: GENERA [pt. ii 



spermum in Atlantic North America and north-east Asia), about 

 19 in the next class, and 24 or more in the lowest class. 



This type of distribution corresponds to that of the species of 

 Doona, Gymnema, Cissampelos, etc., described in the preceding 

 chapter (p. 157). But the Cijrtandra type (p. 159) can also be 

 matched, e.g. by the family Monimiaceae (37), in which there 

 are 22 genera with small areas (the largest being New Guinea and 

 Celebes) and 49 species in all (average 2-2 species per genus), 

 5 genera with areas of moderate size (and 22 species, average 4-4), 

 and 5 with areas of large size (and 196 species, average 39-2). 

 These larger areas overlap one another to some extent in some 

 cases, but there is no single genus covering, or nearly covering, 

 the range of the family. 



All these groups of genera, it will be seen, give indications, even 

 when considered singly, that the areas they occupy go with their 

 number of species, and if taken in groups, the applicability of 

 Size and Space is clearly obvious. 



So far, in dealing both with endemic (and other) species, and 

 with endemic (and other) genera, we have been considering only 

 the areas occupied by them, and we have seen that these are 

 graduated from many very small areas through a good many of 

 a size somewhat larger up to a tail of a very few that occupy the 

 largest areas. Plotted graphically, as in fig. on p. 162, the 

 numbers always form a hollow curve. 



But now, if age be the chief determinant of spread ^ as would 

 appear to be the case from all the figures that have been given, 

 and from the success of the many predictions based upon it that 

 have been made; and if Size and Space be equally valid, then it 

 would seem that the sizes of the genera {i.e. their numbers of 

 species) in any group of endemics should also be arranged in a 

 hollow curve. If Age, Size, and Space (or Area) go together, then, 

 as age is the only active^ factor of the three, it is clear that what- 



1 As already pointed out, age of itself effects nothing, but the fact that 

 dispersal goes so largely with age shows that the various factors that are 

 operative produce an average or resultant effect, so that in twice the time, 

 twice the dispersal will occur, unless barriers (physical or ecological) inter- 

 fere. The essential difference between this view and the older one is that 

 under Age and Area all species (with few exceptions) are looked upon as 

 enlarging their area, instead of a few doing so, and many contracting theirs. 



Manv people take the pojtular view, which is based, it must be remem- 

 bered, upon an assumed efficacy of natural selection for which as yet there 

 is little proof, that species with "small areas of distribution owe the fact that 

 those areas are small to the competition of other more successful types. 

 But there is little evidence for such a belief. It is simply a way of looking 



