SPOROPHYLLS AND STAMENS AND PISTILS 345 
and averaging 
ah N32 raed tyty oe (ns, ep Ny) ¥Ox0y 
Ny + Me : 
[Zps] 
According to Harris’s result [z,s] is always positive. The 
formula shows that this must be the case when 
NOx” apy tt? + (ns aes Ny) O20y >o 
or when 
n,(o22 + forty) > Np(e,? + roze,). 
If o, and gy are equal this can happen only when 7, > mp, 
or in other words when the stamens are more numerous than 
the pistils. 
From the original data from which Harris has drawn the 
material for this study (see Harris for references) it is evident 
that the standard deviations for pistils and stamens are nearly 
equal and that the stamens are more numerous than the pistils. 
For the average values for Europe given by Pearson the mean 
number of pistils is 19.432 with a standard deviation of + 4.8508; 
the mean number for stamens 26.498 with a standard deviation 
of + 4.2562. The coefficient of correlation is + 0.5584. Sub- 
stituting these values we find the results obtained by Harris 
necessarily follow from the existing numerical relations. 
From this consideration it seems evident that the results 
obtained by Harris do not add anything to the observations on the 
numbers of pistils and stamens and their variabilities. Just so 
the higher coefficients of variation given by Harris in Table I of 
his paper result from the fact that the same value (standard 
deviation) is divided in the case of the pistils by a lower value 
(mean for pistils) than in the case of the stamens (mean for 
stamens). : 
It might be said that the total number of sporophylls and the 
per cent of stamens and pistils vary independently, in which 
case the expression (7 + s)(p + v) would represent the number 
of pistils. Disregarding the value sv, since it is small in com- 
parison to the other values, the expression becomes p(n + s) + nv 
and the same relations will hold as in the case discussed above. 
New YorK BOTANICAL GARDEN 
