352 SEMI-CENTENNiAL OF TORREY BOTANICAL CLUB 
vealed by such treatment of data. It is to be determined only by 
observation and by a refinement of methods of collecting data. 
Pearson's first studies ('o2b) pertaining to the heredity of the 
number of stigmatic bands in capsules of the Shirley poppy are of 
special interest, for here data were collected from all capsules. 
These data were statistically treated by three methods: 
I. The correlation of all offspring capsules with parental mean 
capsule, the various progenies grown in each locality being thrown 
together in a single correlation table. 
2. The comparison of the average variability of an array of 
offspring of a single parent plant with the variability of the off- 
spring population. Here the means for individual offspring were 
determined. 
3. A mathematical consideration of homotypic relationship 
in correcting the parental correlation determined by the first 
method. 
According to the first method the parental correlations for the 
different crops, as a whole, range from 0.3230 to 0.1220. The 
highest correlation of 0.3230 was obtained in the ''most starve- 
ling crop” which had few capsules per plant and the low correla- 
tion of 0.1220 was obtained in the crop that was most luxuriant 
in growth. Here is definite evidence that the greater vigor of 
growth affects individual variability by increasing very much the 
partial variability. Оп this account the method of collecting data 
and the statistical treatment give lower parental correlation when 
there is increased vigor. 
In one crop of 907 plants the means were determined separately 
for each plant and these were correlated with the mean of the 
parents. The value was 0.1561 as compared with 0.1864 obtained 
for the same crop by the first method. Here Pearson attributes 
the low parental correlation to ‘‘differentiation’’ and reports that 
the flowers that come out ‘‘early in the season have fewer bands 
than those which come later” and that “ће number of capsules to 
the individual plant, and the dates at which it produces them, tend 
to obscure the influence of pure heredity, and make the stigmata, 
however easy to count and deal with a by no means ideal char- 
acter to study heredity проп” (p. 72). 
The low values obtained for parental correlation were, however, 
