THE MENDELIAN PHENOMENA 17 



hybrid generation consisted of 1,064 plants, of which 

 787 were tall, and 277 were dwarf, a very close 

 approximation to a ratio of 3 to 1. 



Ten seeds from each plant of a hundred only 

 of these 787 tails were sown. Twenty-eight of 

 these produced tails ; seventy-two produced both 

 tails and dwarfs. This result departs consider- 

 ably from the expected one. Let us examine 

 it more closely. Of the three tails in the second 

 hybrid generation, two should produce tails and 

 dwarfs in the proportion of 3 to 1 ; one should 

 produce tails only. This is one way of expressing 

 the average result. Another way is to say that 

 two-thirds of the tails in the second hybrid generation 

 produce tails and dwarfs in the ratio of 3 to 1 ; and 

 one-third tails only. Of a hundred tails in the second 

 hybrid generation we should, therefore, expect sixty- 

 six or sixty-seven to produce tails and dwarfs in the 

 ratio of 3 to 1, and thirty-four or thirty-three to pro- 

 duce only tails, if the general statement I have made as 

 to these ratios is true. It will, therefore, be seen that 

 amongst the hundred tails of the second hybrid 

 generation which Mendel tested, the number of pure 

 tails was too low (twenty-eight as against thirty-four 

 or thirty- three), and the number of hybrid tails too 

 high (seventy-two as against sixty-six or sixty-seven). 

 This brings us up against a very important question, 

 namely, how close a fit between the so-called Men- 

 delian expectation and the actual result must we 

 expect ? The general rule is that the greater the 

 numbers of individuals dealt with, in the experiment, 



