DISCONTINUOUS VARIATION. 



45 



compel one to regard it as composed of two distinct 

 species. It is, of course, often impossible to tell whether 

 any given characteristic is important or not, and hence 

 we must recognise that a really precise and univer- 



40-i 



i 6 8 10 12 14 1'6 18 20 22 UUU 



Leaf-Wid'tti 

 FIG. 13. Distribution of Leaf -Widths in Typha. 



sally applicable definition of a species is, and always 

 must be, unattainable. 



It is probable that variation series in the Vegetable 

 Kingdom often give double humped or multiple 

 humped curves, even if the material examined is as 

 homogeneous as it is possible to obtain it. Possibly, if 

 only individuals of the same stock were examined, 

 they would be found to give single humped curves, but 

 if material collected from different parts of the same 

 district, or even of the same field, is to be regarded as 

 composed of so many local races or sub-varieties, then 

 the determination of the variations in many plant 



