302 
Journal of Agricultural Research 
Vol. XXVII, No. s 
Since (i) the hybrid is not intermediate between the two parents with 
respect to the freezing point depression of its leaf tissue fluids but is 
characterized by a lower osmotic concentration than either of the parental 
forms, and since (2) the Egyptian parent is characterized by a higher 
osmotic concentration than the Upland parent, it is inevitable that the 
difference between the hybrid and the Egyptian parent should be far 
greater than that between the hybrid and the Upland parent. This is 
clearly evident in both the average value of the differences and in the 
ratio of these differences to their probable errors. The differences are, 
however, in all cases clearly significant in comparison with their probable 
errors. The ratios range from 5.8 to 29.1, and therefore leave no possible 
doubt as to the significance of the differences. 
The differences between the first and second series of Egyptian and 
Upland cottons have been discussed above (p. 290). The comparisons 
of the constants for the first and second collection of hybrids (column 5 
of Table XV) shows that in the heterozygous plants, as well as in the 
homozygous individuals of the parental strains, the second collection 
is characterized by a lower average osmotic concentration than the first. 
A remarkable fact in regard to these determinations is that in each of 
the six possible comparisons between the hybrid and the parental forms 
the hybrid shows greater decrease in osmotic concentration between the 
first and second collections than do either of the parent forms. 21 
A comparison of the differences between the hybrid and the parent 
forms in the first and second series, Hi-Ej, Hj-Uj, H 2 ~E 2 , and H 2 -U 2 by 
means of the differences between them (H 2 -E 2 )-(H 1 -E 1 ), (H 2 -U 2 )-(H 1 - 
UJ is possible on the basis of the values in column 8 of Table XV. These 
show that in every instance the difference is larger in the second series 
than in the first. 22 Thus the differentiation between the parent and the 
hybrid individuals tends to become greater with the advance of the 
season, just as the differentiation of the Egyptian and Upland forms 
has been shown to become larger. This rule holds, notwithstanding 
the fact that the actual values of the osmotic concentration are in all 
cases (in both hybrid and parental individuals) lower in the second 
series than in the first. 
Comparing the seasonal increases in the numerical magnitudes of the 
differences between parent and hybrid forms with their probable errors, 
calculated by the method indicated on page 290, we note (Diff/Edm. in 
column 8 of Table XV) that in general they may be considered statistically 
significant. 
The percentage differences (in the final column of Table XV) have 
been calculated with the constant for the parental form as a base. In 
the comparisons between hybrid and Egyptian cotton, it appears that 
the hybrid has an osmotic concentration from 4.8 to 7.4 per cent lower 
than that of the Egyptian parent in the first series of determinations. 
In the second series of determinations the hybrid is characterized by an 
osmotic concentration of about 12 per cent lower than that of the 
* l A possible explanation of this difference may be found in the fact that there was but a single plant of 
the hybrid in each instance. It is possible that because of the necessity for a somewhat closer picking 
of leaves in the case of the single hybrid plant, the average age of the leaves may have been somewhat less 
in the second collection than in the case of the controls. This is, however, merely a suggested possibility, 
A solution of the problem must await further investigation. 
n Since the differences between parent and hybrid forms are so taken that a negative sign indicates a 
lower value of freezing-point depression in the hybrid than in the parent, and since consistency in treat¬ 
ment throughout the paper demands the regarding of signs in taking the differences between the differences 
in the first and second series of determinations, a negative sign for the differences between the differences 
indicates that the differences between parent and hybrid are greater in the second series than in the first. 
