278 
Journal of Agricultural Research 
Vol. XXVII, No. 5 
A second series of samples was taken on August 18. Sample 6 re¬ 
presents Acala Upland cotton growing between two rows of Pima Egyptian 
(sample 5). On this date it was feasible to secure samples of Pima 
Egyptian (7), Sea Island (8), Acala Upland (9), and Kekchi Upland (10) 
in close proximity in the same border. Pima Egyptian is compared 
with all the other types, and Sea Island is compared with the two Upland 
varieties. A comparison of Pima Egyptian (sample 11) with Gossypium 
herbaceum var. Wa-Gale (sample 12) from the Transcaucasian region 
was also made on this date. 
On August 22 a third series of comparisons was made with the follow¬ 
ing samples from the yield test plots of Pima Egyptian, Lone Star Upland, 
and Durango Upland. In the comparisons of Pima Egyptian and Lone 
Star Upland cotton, samples 13-14 were taken near the south end of the 
border, samples 15-16 near the middle, and samples 17-18 near the 
north end. In the same test plot and on the same date a comparison 
of Pima with Durango was also made on the basis of samples near the 
south end (19-20), near the middle (21-22), and near the north end 
(23-24) of the border. In another border we took Pima Egyptian 
and Holdon Upland for comparison with the Fj hybrid between the two 
species. (Samples 25-26.) For constants for the hybrid see sample 34, 
p. 299. In the same border we took Pima Egyptian (27) and Acala 
Upland(28) for comparison with an Fj hybrid. (See No. 35, elsewhere.) 
The constants for the individual samples are compared in Table I. 
Comparing the values of freezing-point lowering and calculated osmotic 
pressure in Pima Egyptian cotton with the 13 determinations 9 on the 
several varieties of Upland cotton, we note that without exception the 
osmotic concentration is higher in the Egyptian cotton than in the 
Upland types. 
Calculating the statistical constants for freezing-point lowering, A, in 
Pima and Upland, the data presented in Table II were obtained, the 
mean being the average value of the constant, S. D. the standard devia¬ 
tion, C. V. the coefficient of variation, and r, the coefficient of correlation 
between the two variables denoted by the subscript. 
Table II .—Statistical constants for freezing-point lowering in Pima and Upland cottons 
in IQ20 
Pima, P. 
Upland, U. 
Mean.. 
1.3OO ± 0.022 
o.ii8±o.oi6 
9.100 
I.I 44 ± 0.022 
0.120 + 0.016 
10.456 
S. D. 
C. V. 
rpu = +0.7439 ±0.0836 
In determining the probable error of the difference we must remember 
that because of field and meteorological heterogeneity (21) the values 
of osmotic pressure in associated plants of Pima and Upland types may 
be correlated. As a matter of fact this correlation is found to be r = o. 744. 
This shows that the difference in salt content of the soil or differences in 
9 In one case (No. 7) a single sample of Egyptian cotton is compared with two varieties closely associated 
in the same plot of Upland cotton. In calculating the statistical constants for the whole series of each type, 
this one sample of Egyptian cotton is therefore used twice. 
