De Vries : Age and Area 



167 



independent of special organizations or 

 so-called adaptations. 



In following this discussion, many 

 readers will object that there are too 

 many single instances which do not 

 seem to be in accordance with this new 

 law. In order to reply to this objection, 

 Willis has introduced his principle of 

 averages, or what we might call his 

 statistical method. A general law can 

 only be extracted from the study of 

 individual cases, if these are combined 

 in sufficient num])ers to smooth the 

 effects of their special features. Age 

 and area should not be used as a com- 

 parison of single cases, but for small 

 groups of allied forms, such as the 

 species of a genus, the genera of a 

 tribe or the tribes of a family. Expe- 

 rience has proved that averages for 

 groups of about ten such types suffice 

 to eliminate the effects of most of 

 the deviations. Whenever possible, 

 groups of fifteen to twenty species 

 should be preferred. In doing so the 

 method may be extended to biological 

 or ecological groups of plants as well as 

 to systematical ones. 



When one considers such groups and 

 compares them with other related sec- 

 tions, the effects of age will show clear- 

 ly, because the other factors in disper- 

 sal will either be pulling the same way 

 upon all, or will cancel one another by 

 pulling in different directions. Age, 

 however, is always pulling alike upon 

 all species. The greater the number of 

 allied forms taken, and the greater the 

 length of time considered, the more 

 clearly will the effects of age show. 



In applying this statistical method 

 to the flora of Ceylon, one has to com- 

 pare two main lines of geographical 

 facts. One of them relates to the 

 probable age of the species, the other 

 to their present distribution on the 

 island. Three main divisions are pro- 

 posed, one containing the endemic spe- 

 cies, the second those confined to Cey- 

 lon and Peninsular India, and the third 

 the forms of wider distribution. The 

 distribution on the island itself is de- 



rived from the data given in Flora of 

 Ceylon, by Trimen and Hooker. These 

 authors divide all species into six 

 classes, viz.: i, very common; 2, com- 

 mon ; 3, rather common ; 4, rather rare ; 

 5, rare ; and 6, very rare. The number 

 of species belonging to each of these 

 classes is then multiplied with the fac- 

 tor indicating their degree of rarity, 

 and this is done for each of the three 

 groups of dift'erent dispersal. From 

 these figures averages may be calcu- 

 lated, which give the degree of rarity 

 for any group under consideration. In 

 this way Willis finds. 



No. of Rar- 

 Species ity. 



Mean rarity of all species... 2,809 3-5 



Species of wide distribution.. 1,508 3.0 



Of Ceylon and Peninsular India 492 3.5 



Species endemic to Ceylon... 809 4.3 

 Species of all the zt, endemic 



genera 52 4.5 



Species of Doona (endemic). 11 4.6 

 Species of Stemonoporus 



(endemic) 15 5.4 



From this table we see that the spe- 

 cies of widest distribution are the 

 commonest, that those of Ceylon and 

 Peninsular India have just the mean 

 degree of rarity, whereas the endemics 

 are relatively rare. The species of the 

 endemic genera, and especially of the 

 only two genera which are rich in 

 endemic species, are the rarest of all 

 on the island. Further tables, which 

 cannot be reproduced here, show that 

 the different families of this flora be- 

 have in the same manner, showing that 

 no one family has any particular ad- 

 vantage over another. The same rule 

 prevails for the genera and even for 

 ecological groups of species. Every- 

 where the order of rarity within the 

 island is parallel to the degree of the 

 distribution outside of it. 



Confirmation of this law was then 

 obtained by working out the flora of 

 New Zealand, of the Hawaiian archi- 

 pelago and of other oceanic islands 

 after the same method. The orchids 

 of Jamaica, Callitris, a coniferous spe- 

 cies of Australia, the ferns of New 



