60 AGE AND AREA [pt. i 



frequency), then the number 1 attached to a species will mean 

 that it has the maximum, the number 5 the minimum, dispersal 

 in Ceylon. And we can find the average distribution of a group, 

 whether Ceylon endemics or widely distributed, by multiplying 

 the number under each head from 1 to 5 by the number of that 

 head, adding up all the marks thus obtained, and dividing by 

 the total number of species. Thus we obtain : 



Now the actual number of species under each of these heads 

 in the whole flora is (VC 285), C 670, RC 555, RR 429, R 415, 

 \B. 455. If we take the average rarity of the last five classes, 

 we find it to be just over 2-7. The average rarity of an endemic 

 we have seen to be 3-4, and of a wide 2-3, while the remaining 

 species, which are endemic to Ceylon and South India, show a 

 rarity of 2-7, the same as the whole flora. The difference of 1-1 

 in average rarity between wides and endemics represents over 

 a quarter of that between the most and the least widely dis- 

 tributed species (1 and 5, difference 4). In other words, the 

 most widely distributed species in Cej^lon, on the average, are 

 those that show a distribution abroad to a greater distance than 

 merely to Peninsular India; then follow those that reach the 

 peninsula, and the least widely distributed are those that are 

 found in Ceylon only. Taking the estimates of actual area given 

 above for the different classes, the differences actually found 

 indicate that an endemic has an average area about 40 miles 

 in diameter, a "wide" one of 80 miles, or four times as large. 



A cursory examination of other floras soon showed that their 

 species behaved in the same way, occupying areas of all sizes, 

 overlapping in the same manner, and with their endemics 

 occupying areas from many small up to few large, and the wides 

 the re\erse. At the same time, the figures for the Ceylon flora 

 indicated clearly that this graduation of areas, wides largest, 



