THE CANADIAN ENTOMOLOGIST. 377 



An examination of the figures presented in this table reveals the fact 

 that 6,037 occur in only a single one of the fourteen regions, while the 

 other 2,504 are of more general distribution. Of the 6,037 endemic 

 genera 3,713, or 61.4 per cent, of the endemic forms are found only in 

 the Tropical American, Indian, Australian or extra-tropical South 

 American regions. Expressing the relationship somewhat differently, we 

 find that the 3,713 found exclusively in regions 3, 4, 6 or 7 are 43.4 per 

 cent, of the total 8,541 genera of flowering plants. In the Indian region 

 occur exclusively 13. i per cent, of all genera of flowering plants, in the 

 tropical American region 23 per cent., in the Australian region 5.1 per 

 cent., and in the extra-tropical South American region 2 per cent. 

 Altogether about 70 per cent, of all genera of flowering plants occur 

 exclusively in one or the other of the fourteen regions recognized — an 

 average of 5 per cent, per region. The average for the four regions, 3, 4, 

 6 and 7, which particularly interest us in the present problem, is 10.8, 

 while for the other ten regions it is 2.6 per cent. Turning now to the 

 total number of genera of plants occurring in these several regions, we 

 find that in the Indian region occur 28.9 per cent, of all genera, in the 

 Tropical American region 36.6 per cent., in the Australian region 13.5 per 

 cent., and in the extra-tropical South American region 7.9 per cent. The 

 average for these four regions is 21.7 per cent., while for the other ten 

 regions it is 9.8 per cent. 



It is clear that in the regions which have been indicated as those in 

 which apically dehiscent genera are most abundant, the whole flora is also 

 much richer, both as a whole and in endemic forms. This renders neces- 

 sary the greatest care in comparison. 



The Solanum-Cassia type seems to be the one best adapted for a 

 first examination of the hypothesis of specialized distribution; in numbers 

 it is next to the largest type, it is the one of the widest geographical range, 

 and it is made up of representatives from several widely-separated families, 

 so that its uniformity of structure and localized distribution cannot so 

 readily be accounted for on the supposition of community of descent. 

 The distribution of the genera is shown in table B. 



