Feb. a, 1924 
Tissue Fluids in Cotton 
3 i 5 
A casual glance at Table XXI will show that the Egyptian series on 
the one hand and the Upland and the hybrid series on the other differ 
materially with respect to the correlations between the first and second 
series of determinations. All of the coefficients for A, k , and k/A in 
the Pima series are positive and in general may be considered significant 
in comparison with their probable errors. The comparable coefficients 
for the Upland series and for the hybrid plants are characterized by low 
values of . the correlation coefficients. These may be either positive or 
negative in sign. They are of the same general order of magnitude as 
their probable errors. No one of the 21 coefficients can be considered 
certainly significant in comparison with its probable error. 
The coefficients for hydrogen-ion concentration are low throughout, 
and can not be considered statistically significant in comparison with 
their probable errors. 
Taking the averages 25 of the constants for the parental groups and for 
the hybrid groups, we note that all of the coefficients for the Upland 
plants are numerically less than 0.11. One of them is positive while 
three are negative in sign. The averages for A, k, and k/A for the 
hybrid plants are less than 0.07. The average value of the two coeffi¬ 
cients for hydrogen-ion concentration in the hybrid plants is — 0.10. 
Contrasted with these, we find in the Pima Egyptian series the follow¬ 
ing averages: + 0.418 for A, r= +0.259 for k , r= +0.299 for k/A, 
and +0.086 for P H . Clearly these correlations for the Egyptian 
series are far higher than those for the Upland series. 
We have considered various possible explanations of this difference in 
correlation, but since actual data are wanting for the adequate substan¬ 
tiation or refutation of any of the suggested possibilities, it seems wisest 
to omit all discussions of suggested explanations until further experi¬ 
mental evidence, now being collected, can be analyzed. 
That the behavior of the two types of plants is quite different is further 
suggested by a consideration of the cross correlations between the con¬ 
stants of the plants of the same duplet or triplet in the first and second 
series, as set forth in Table XXII. The coefficients in this table were 
determined primarily as a means of obtaining the probable errors of the 
differences between the first and second series of differences between the 
Egyptian and the Upland types. The few words of discussion are 
purely incidental. The problem will, we hope, receive more detailed 
treatment on the basis of further data now being collected. 
The coefficients presented show the relationship between the first 
determination on the Egyptian plants of a duplet or triplet and the 
second determination on the Upland or hybrid plants of the same duplet 
or triplet; or, conversely, the constants show the correlation between the 
first determination on the Upland or hybrid plants of a duplet or triplet 
and the second determination on the Pima plants; or, finally, the coeffi¬ 
cients show the relationship between the first determination on the 
Upland plants of a triplet and the second determination on the hybrid 
plants, or the first determination on the hybrid plants and the second 
determination on the Upland plants. 
The first two columns of Table XXII show the plants which furnished 
the materials for the first and second determinations, respectively. 
For convenience of comparison these are arranged in pairs. The first 
25 The averages of the correlations have been computed for the Egyptian and the Upland plants by 
weighting the constants for rows i and 5 and for 3 and 7 by two and that for rows 2, 4 and 6 by three, and 
taking 7 as N in the determination of the average for the whole experiment. 
