786 



SCIENCE 



[N. S. Vol. XLIII. No. 1118 



In drawing such conclusions, however, even 

 from a thorough knowledge of a flora, one is 

 often exposed to lay too much stress on stri- 

 king but exceptional instances, whereas it is 

 only averages which may really be relied upon. 

 For this reason Willis has worked out a method, 

 which gives a large degree of accuracy and 

 thereby affords a firm and unattackable basis 

 for his deductions. Trimen divides all species 

 into six classes: Very Common, Common, 

 Eather Common, Rather Rare, Rare and Very 

 Rare, and his estimates are thoroughly reliable, 

 as is shown by the clearness and regularity of 

 the results derived from them. 



In order to compare two or more groups of 

 species Willis multiplies the number of spe- 

 cies in them, belonging to each of these classes 

 by a factor indicating the degree of rarity ac- 

 cording to the estimates of Trimen. These 

 factors are 1 for very common, 2 for common 

 and so on, up to 6 for very rare. In this way 

 averages may be calculated, which give the 

 relative degree of rarity for any group under 

 consideration. 



Next, the plants of Ceylon are divided into 

 three main groups, one containing the endemic 

 species, the second those confined to Ceylon 

 and Peninsular India, and the third the forms 

 of wider (although often not very much wider) 

 distribution. In this way Willis finds: 



No. of Species Rarity 



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



Species of wide distribution. 1,508 3, 



Of Ceylon and Peninsular 



India 492 3,5 



Species endemic to Ceylon. . . 809 4, 3 



Species of all 23 endemic 



genera 52 4, 5 



Species of Doona (endemic) . . 11 4, 6 



Species of Stemonoporus (en- 

 demic) 15 5, 4 



Thus the species of wide distribution are the 

 commonest, those of Ceylon and India have 

 just the mean degree of rarity, but the endem- 

 ics are relatively rare, the rarest of all being 

 the species of the endemic genera and espe- 

 cially those of the only two genera which are 

 rich in endemics. Results of the same kind. 



obtained by applying this method to different 

 manners of bringing the species of this island 

 into groups, are given in numerous tables, the 

 study of which will be of great importance to 

 all scientists interested in the subject. 



One of the main results is that the variation 

 in rarity between the different families or 

 groups of families of Ceylon-endemics is small, 

 and goes to show that no one family has any 

 particular advantage over another. In com- 

 paring the genera with one another the same 

 rule prevails, independent of the question 

 which genera are chosen and from which point 

 of view the comparison is made. 



The order of rarity: Ceylon, Ceylon and 

 Peninsular India, Wider Dispersal, holds 

 throughout all comparisons with extraordinary 

 regularity. It is obvious that some general 

 law must be underlying these phenomena. 



If the endemic species had originated by nat- 

 ural selection of infinitesimal steps and in 

 response to the local conditions, which are 

 obviously the only conditions that matter when 

 the species first appears, they must have been, 

 from this very origin, better adapted to these 

 conditions than their parent species. Accord- 

 ing to the theory of natural selection it would 

 follow that they must surpass their forerunners 

 in the struggle for life and soon spread to a 

 higher degree of commonness. But as the 

 table shows, the reverse is true. Yet they have 

 had ample time even for gaining a compara- 

 tively wide dispersal. Several recently intro- 

 duced species have spread to the stage of very 

 common, often in a few years. Tithonia 

 diversifolia, one of the Compositse, began to 

 spread about 1866 and in 1900 was all over 

 the island in damp enough spots. Mikania 

 scandens began to spread ten years ago and is 

 already common all around Peradeniya. Many 

 other instances could be given, since about 60 

 introduced species have become more or less 

 common in the island. 



Of the 809 endemics of Ceylon only 90 are 

 now common and only 19 very common in the 

 island, the remainder are mostly rare or very 

 rare. If they did not conquer their parents 

 and spread into larger areas, it is obvious that 

 they were not especially adapted to the eondi- 



