VARIATION 21 



Seven heads was the commonest result, eight and 

 six were less common, and so on, until we reach the 

 classes 0, 1, 12, and 14, which do not occur at all. Our 

 curve in this case is much more irregular than in the 

 previous one, the reason being that in this case we are 

 dealing with a very much smaller number of results, 

 viz. 150, instead of over 8000. 



Both of these curves approach, as will be seen by 

 comparison with the ideal type shown in Fig. 5, which 

 is known as the normal curve of variability. 



We should get an extremely close approximation to 

 this ideal type if, say, we were to toss a lot of a hundred 

 coins ten thousand times, 

 and plot the result. It is the 

 curve which expresses " pro- 

 bability " in a result which 

 depends on "sheer chance " 

 or on a multitude of small 

 and independent causes, and 

 variability in many char- 

 acters of living organisms is ^ 



Of this type. ^ 6< _ The NormaJ variability 



The appearance of a giant Curve, 



may be compared to a toss- 

 ing of ten coins in which all turn up heads. The extreme 

 stature is the result of the many independent factors 

 influencing stature all happening to act in one direc- 

 tion, just as the extreme result in tossing is due to 

 the coins having all happened to fall on the same 

 side. 



Occasionally, however, variability follows different 

 rules. If, for instance, we were to count the number 

 of " petals " on a series of marsh-marigold flowers, we 

 should find that the commonest number was five, and 

 that there were frequently six or seven, and occasion- 

 ally eight, but never less than five. The curve repre- 

 senting such variability would therefore be of a quite 

 one-sided type, as is shown in Fig. 6. 



Such extreme cases are comparatively rare, but a 

 certain departure from symmetry, or skewness as it is 



