44 



DISCONTINUOUS VARIATION. 



tiation into two species in respect of four out of the 

 seven characters measured, and so is obviously, in the 

 authors' opinion, to be regarded as composed of two 

 more or less confluent species. 



This " precise criterion of species " suggested by 

 Davenport and Blankinship has much to recommend it, 

 but probably it would generally be considered that an 

 index of isolation of only 50 per cent, is too small a dif- 

 ference to merit specific distinction. Perhaps it would 



70 



40 



20 



4 6 8 10 12mm. 

 Mid-Stem Diameter 



FIG. 12. Distribution of Mid-Stem Diameters in TypTia. 



be better, therefore, to increase it to 90 or 95 per cent. 

 In any case, it must, I think, be admitted that a slight 

 degree of confluency, or overlapping of the curves of 

 variation, ought not to compel one to assume the exist- 

 ence of only a single species, though this is the view 

 which has been generally held in the past. On the 

 other hand, the fact of a group of organisms showing 

 absolutely discontinuous variation in respect of some 

 apparently unimportant characteristic ought not to 



