ON THE SOURCES OF APPARENT POLYMORPHISM 
IN PLANTS, ETC. 
(EDITORIAL.) 
In the last number of Biometrika a note was inserted in the Miscellanea* 
warning biometricians against laying too great stress on " apparent modes " 
obtained by the mere inspection of frequency polygons. It was pointed out 
that the significance of such modes could only be tested by an application of the 
mathematical theory of random sampling. Now a great deal of argument in 
favour of the dimorphic and polymorphic character of plants has been based solely 
on emphasising irregularities in the seriation, which have no importance when the 
deviations due to random sampling are properly allowed for. Nor has the applica- 
tion of this very rough form of graphical analysis tended to error in botanical 
investigations only. We have also in mind the bold resolutions by crani- 
ologists of small groups of human skulls into distinct local races, because they 
formed peaked frequency distributions. On application of the theory of random 
sampling such peaks have on more than one occasion been found to be of no 
significance, and further they have been seen to actually disappear when a much 
larger number of skulls were available for measurement. 
Again, within the limits of the same homogeneous i"ace by measuring groups of 
individuals at different stages of seasonal or secular growth, or subject to different 
conditions of environment, we may easily obtain significant bi-modal or even multi- 
modal frequency distributions, which have no relation whatever to the existence of 
dimorphic forms or to " petites especes." For example, the offspring of Shirley 
poppies all grown from Hampden stock in six different parts of this country 
showed in the number of stigmatic bands on the capsules modes significantly 
different from each other and from tlie parent stock. There is no doubt that, had 
the countings been mixed together without attention being paid to local environ- 
ment, we should have spoken of the flower of this poppy as polymorphic in 
character, or suggested the existence of " petites especes." A still further emphasis 
of the irregularity of the seriation would have arisen, if the capsules had not been 
* Vol. I. p. 260. 
