CH. XVI] ENDEMISM AND DISTRIBUTION: GENERA 175 



ever phenomena are shown by space should also be more or less 

 paralleled by those shown by size. We are thus led on to the 

 investigation of the sizes of genera, to see whether they may not 

 show some definite relationships to one another, such as might 

 be expressed by the aid of curves. 



Very little investigation is required to show that this is indeed 

 the case. If we take 16/1 to mean 16 genera of one species each, 

 3/2 to mean 3 genera of two species each, and so on, then examine 

 the endemic flora of all the islands of the world, and pick out 

 those genera that are actually endemic to the islands, one finds 

 that all the islands show the same type of arrangement, as may 

 be seen in the following list of examples : 



Table showing the numhers of Endemic Genera 

 of different Sizes upon a number of Islands 



Azores, Canaries, Madeira 16/1, 3/2, 1/4 



Borneo .... 59/1,8/2,2/4,1/5 



Ceylon .... 19/1,2/2,1/3,1/5,1/11,1/15 



Cuba .... 58/1,9/2,2/3,2/4,1/6 



Hawaiian Islands . . 14/1,6/2,7/3,4/4,3/5,2/6,2/7,1/9, 



1/11,2/12, 1/14, 1/17, 1/28 

 Japan .... 54/1,9/2,1/3,2/4,1/8 

 Java .... 57/1,2/2 



Madagascar . . . 191/1,37/2,10/3,7/4,9/5,2/6,2/7, 



1/8, 5/10, 1/12, 1/18, 1/20 

 New Caledonia . . 73/1, 27/2, 6/3, 4/4, 4/5, 4/6, 2 7, 



1/9, 3/10, 2/12, 1/15 

 New Zealand (proper) . 22/1, 2/2, 3/3, 2 4, 1/5, 1/9 

 Socotra .... 17/1, 1/2, 1/3 



On such large islands as Madagascar, where there are many 

 endemics, the same phenomenon is shown even by single families. 

 Thus the Madagascar Compositae show 11/1, 2/2, 1/3, 1/5, and 

 1/10, the Rubiaceae 14/1, 3/2, 1/4. 



Every island in the world that possesses any endemic genera 



at the aetual fact, which is all we have to go upon, that A occupies a large 

 and B a small area. My way of looking at the same fact is to suppose that 

 A is older than B. This is really a much more simple explanation, especially 

 when we remember that the areas occupied by the different species in a 

 genus, or the different genera in a family, usually increase fairly regularly 

 from very small to large. If one have areas represented by 1,2,3, 4, 5, 6, 

 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, it seems an unnecessarily 

 oblique way of looking at the facts to say that 1, 2, 3, 4. and 5 must be 

 regarded as dying out, while 16 to 20 are to be looked upon as successful 

 and expanding species, and no two authors can agree about whetiier the 

 intermediate species 6 to 15 are one thing or the other. It is far more 

 simple to regard all as still in process of expansion, but that some, by reason 

 of greater age and perhaps other advantages, have grown larger than others. 



