VARIATION 



375 



involve differences in the number of constituent parts, such as 

 perianth-segments, leaflets in a compound leaf, etc. 



It is easiest to study variation if some character is selected 

 which is capable of exact measurement, as, for instance, the 

 length of the leaf in the Privet, the length of the Runner Bean 

 seed, or the number of 

 ray-florets in the Daisy. 

 Taking the first of these 

 cases as an example, it 

 will be found that if, say, 

 a thousand leaves are 

 measured, the difference 

 between the length of the 

 shortest and the longest 

 is quite considerable, and 

 that the majority of the 

 leaves are of an average 

 size. If all the thousand 

 leaves be classified ac- 

 cording to their lengths 

 into separate groups, 

 differing b)? increments of 

 one millimetre, and the 

 number of leaves in each 

 group counted, those 

 containing the smallest 

 number will be found to 

 be situated at the two 

 extremes. Between these 

 points the number of 

 leaves of each particular 

 length will be found to 

 increase with considerable regularity, as the size of its individuals 

 approaches that exhibited by the majority. This most frequent 

 size is termed the mode, and often corresponds very closely to 

 the arithmetical mean of all the measurements. 



By plotting a curve, in which the ordinates represent the 

 number of individuals in each group, and the abscissa: tha re- 

 spective lengths, the variation can be graphically represented 



Fig. 2i8. — Three equal-aged individuals of 

 the Marsh Cudweed (Gnaphalium 

 iiliginostim) from the same locality, 

 showing individual variation. All 

 three are in fruit. (Natural size.) 



