Feb. 2,1924 
Tissue Fluids in Cotton 
309 
DISCUSSION OF RESULTS 
The data presented in the foregoing sections lead to definite con¬ 
clusions as far as matters of fact concerning the physicochemical prop¬ 
erties investigated are concerned. The consistency of results for the 
two years, the close concordance of the results for the various subseries 
in the comprehensive experiment of 1921, and the large size of the 
differences when considered in comparison with their probable errors, 
leave no possible doubt as to the existence of differences between the 
tissue fluids of the Egyptian and Upland cotton, and between the ^F 1 
hybrid and the two parental types, with respect to the physicochemical 
properties investigated. 
This represents a distinct advance in our working knowledge of the 
characteristics of a highly important economic plant. The biologist 
would, however, like to see some explanation of the demonstrated dif¬ 
ferences in terms of more fundamental biological phenomena. He would 
at least like to be sure that every possible biological source of error is 
eliminated. 
We may, therefore, review again certain possible explanations of the 
differences which have been demonstrated. 
review of biological sources of error 
The probable errors of random sampling have been so reduced by the 
number of data accumulated that the conclusions are not open to criticism 
on the grounds of the results being due to chance. 
Differences between the plant forms due to differences in distribution 
over a field characterized by diversity in soil conditions have been fully 
eliminated, unless we assume that the forms are remarkably different in 
the depth of root penetration, and that the soil is sufficiently different at 
the various levels to influence the physicochemical properties in a different 
manner. 
At present we have no adequate information on this point. It is a 
question to which we hope to give some attention later. 
A possible source of error—and one which seems largely unavoidable 
during the stages of development preceding maturity—is to be found in 
the differences in the growth habit of the two groups (Egyptian and 
Upland) of cotton. This renders the collecting of wholly comparable 
leaves somewhat difficult, but the greatest care was exercised in this 
work. Ultimately it may be possible to refer the differences here demon¬ 
strated to some more fundamental physiological or morphological phe¬ 
nomenon involving stage of development, but at present we are unable 
to do so. 
differences between the first and second series of determi¬ 
nations 
We have now to consider certain outstanding features of these series of 
constants. 
Conspicuous among these is the fact that all the constants differ be¬ 
tween the first and the second series. This fact may be established by the 
comparison of the actual differences in the eight fundamental tables of con¬ 
stants (Tables X to XIII and XV to X VIII) with their probable errors. For 
convenience of reference and to facilitate the comparison of the various 
