Karl Pearson and David Heron 
207 
correct method; but this will lead us to results absolutely opposed to those of 
Mr Yule's association and colligation. 
We ourselves, however, doubt the existence of this discrete unit; we have 
only come across it in theoretical Mendelian investigations and doubt whether 
the " unit character " which is absent or present has any existence in somatic 
classifications. Of this we should like to give two or three illustrations. 
The first illustration we take is from a paper* by Professor E. M. East, 
entitled "The Mendelian Notation as a Description of Physiological Facts." 
Professor East is a vigorous Mendelian making the very best defence he can 
of the Mendelian notation in the present parlous condition of Mendelian theory — 
which assumes the truth of the unity of the unit. The cross which Professor 
East cites is one between a "long" corolla and a "short" corolla race of Nicotiana. 
Speaking of other Mendelian investigations into size, Professor East writes of 
them : " No criticism could be made except that certain of the characters used 
varied considerably in the mother varieties and therefore were presumably not 
homozygous for all character factors*!". This criticism is apparently answered by 
a recent investigation of the writer's, as yet unpublished, where two species, 
Nicotiana forgetiana and Nicotiana alata grandiflora, were crossed. As seen by 
the table, the corolla length is very slightly variable in either species j, nor is it 
affected to any extent by environment, yet each species was absolutely reproduced 
by recombination in the F 2 generation." 
TABLE X. 
Frequency Distributions for Length of Corolla in a Cross between Nicotiana 
forgetiana (314) and N. alata grandiflora (321). 
Class 
Centres in 
Millimetres 
Designation 
Totals 
SO 
25 
SO 
35 
40 
45 
50 
55 
60 
65 
70 
75 
80 
85 
90 
314 
9 
133 
28 
170 
321 
1 
19 
50 
56 
32 
9 
167 
(314x321)^ 
3 
30 
58 
20 
111 
(314x321) F 2 
5 
27 
79 
136 
125 
132 
102 
105 
64 
30 
15 
6 
2 
828 
Now let us call short all " below 60 mm." and long all " above 60 mm." We 
have all the offspring " below 60 mm." Hence there is " dominance " of short 
corolla, and we may apply the magic formulae : 
(DD)x(RR) = 4>{DR), 
(DR) x (DR) = (DD) + 2 {DR) + {RR), 
* The American Naturalist, Vol. xlvi. p. 639. 
t Note the writer's interpretation of the results by the preconceived theory ! 
t The author does not tell us how many plants were grown in each generation. In the parent 
generation presumably only two. How many were considered in F x and F 2 ? The difference of 
variability would largely turn on this factor. 
