3 °S 
Journal of Agricultural Research 
Vol. XXVII, No. 5 
In the determination of the percentage differences between the P H 
of the hybrid and that of the parent, we have considered that the proper 
base for the calculation of a relative difference is the deviation of the 
value of P H from that which may be taken as indicating neutrality. Thus 
in the calculation of these relative differences we have used the formula 
100 (. H-P ) 
' 6.860-P ”’ 
where H and P represent the average P H values for parent, P, and hybrid, 
H , respectively. This is in agreement with the practice in comparing 
the two parental types. 
The percentage differences calculated in this way range from 1.1 to 6.7. 
Turning to the comparison of the first and second sets of determinations 
on the hybrid plants, we note that in all cases the average value of P H 
is higher in the second series than it is in the first. Thus the tissue fluids 
in the second series of determinations are more nearly neutral than in 
the first series. These values range from 4.29 to 6.95 times as large 
as their probable errors. They may be considered unquestionably 
significant. 
Concerning this result we have no explanation to offer, Caution 
must be observed in interpreting it as a regular developmental change, 
since in an irrigated crop grown under the severe conditions of the South¬ 
western deserts many factors must vary from one series of determinations 
to another. An interpretation of the differences in acidity should be 
one of the purposes to be considered in the planning of any future investi¬ 
gations in this field. 
Turning to the correlations (Table XVIII, column 7), we note that the 
coefficients measuring the relationship between the hydrogen-ion con¬ 
centration of the hybrid and of the plants of the parental form growing 
beside it in the fields are in all cases positive in sign. This result sub¬ 
stantiates the findings in the comparison of the two types of cotton 
(Table XIII, p. 297), which shows clearly that acidity is to some extent 
determined by the varying conditions of the soil or of the atmosphere 
at the time at which the samples were taken. 
The second series of correlations is numerically larger than the first. 
This indicates either (a) that the plants have become more closely related 
to their substratum or ( b ) that their atmospheric environment differs 
more widely from time to time in the second series than in the first, or 
(c) that the determinations in the second series were made with a con¬ 
siderably higher order of precision than those of the first series. 
These results indicate clearly the importance of further investigations 
along these lines. 
Comparing the differences between the hybrid and the parent form 
in the first series with that in the second series (Table XVIII, column 8) 
we find that in the comparison with Egyptian cotton the differences are 
in all cases greater in the first than in the second series, but these differ¬ 
ences can not be considered significant in relation to their probable 
errors. In the comparison with Upland cotton the differences are numeri¬ 
cally greater in the first series for rows 3 and 7 and for rows 1, 3, 5, and 7, 
but numerically greater in the second series for rows 1 and 5. 
