72 Cooperatice Investigatlons on Plauts 



iiKÜvidual Standard deviatioiis as fainilies takeu, as well as the raean and Standard 

 dcviatioi) of the population of 907 plants. The niean of the plant means was 

 1375 Stigmata, and their Standard deviation 1'370. The average variability 

 (or S.D.) of plants tVom the same parentage was 1-272, and the weighted niean 

 Square deviation — i.e. \fS{na-)!N, where a is the standani deviation of a fraternity 

 of n, and N the total nuniber of plants — was 1287. If /• be the pareutal corre- 

 lation 1 Vi — r', S being the total plant Variation, should be that of an array 

 of offspring. Hence equating this to r287 we find ?• = '3427. Had we equated it 

 to 1'272 we should liave found r = '.3716. In either case this, cur second niethod, 

 gives a value much closer to that ''.i^QO of the third niethod, than the valiie "2595 

 fouiul for the slope of the Kiifield crop rcgre-ssimi line. 



Fiiially to illiistrate the second niethod niore completcly a correlation table 

 has actnally bcen formed for the nieans of parent and otfspring plants in the case 

 of the Enfield I crop. The value of the parental nican plant is 12(39 of the 

 offspriiig plant, 13"75, of their respective Standard deviations H99 and 1-370; 

 tlie correlation is ■1561 and the regression coefficient of offspring on parent plant 

 "1784. Nor is this result really to be much wondered at, f)r not only are the 

 Mmeans of the stigmatic bands much infiuenced by the number <if capsules on the 

 ; plant, but the flowers that come out early in the season have fewer bands than 

 those which come later. Thns taking the O.xford record for first flowers on 

 322 plants, we divided them iuto 161 carlier flowers of each family and the 161 

 later flowers, and found a correlation between the number of stigmatic bands 

 of the first fiower and its lateness in Coming ovit of 2078. The number of 

 capsrdes to the individual plant, and the dates at which it produces them, tend 

 to obscure the iiifluence of pure heredity, and make the Stigmata however easy 

 to coimt and dcal witli a by no nieans ideal character to study heredity upon. 



To avoid the ditficult}' due to differentiation in capsules, a correlation table was 

 formed for 327 first flowers and the means of the parent plants in the Oxford 

 crop. We found parental mean 1274, Standard deviation r240; Stigmata of first 

 flowers of offspring 1102, Standard deviation r828; correlation '2438, and re- 

 gression of offspring first capsule on parental mean = •3-594. There can be small 

 doubt that, if we had compared the Stigmata on the first fiower of the parent with 

 those on the first flower of the offspring, we should find heredity in poppies as intense 

 as in f ircarm or cophalic inde.\ for the case of man. It is true that the Enfield I 

 and Kidderniinster crops show a reniarkably low parental correlation, but both were 

 large crops and it is highly probable that the environment of the strips devoted to 

 special fainilies was to some e.xtent differeutiated, and this nuiy well have weakened 

 the apparent strength of heredity*. 



Fiually if we attempt to obtain the parental correlation by our third method, 

 we have the foUowing table : 



* Mr John Notcutt assures me that for the Eiddorminstcr crop thcrc was no diScreutiation in 

 environment apparont for either light, sheltcr or soil. 



