28 o 
Journal of Agricultural Research 
Vol. XXVII, No. 5 
the Upland cottons. The differences in the ratios are relatively small. 
The average of the eight differences which indicate higher ratios of k to A 
in the Egyptian cotton is 0.00123, whereas the average of the five cases 
which indicate higher ratios of k to A in the Upland cotton is -0.00167. 
The general average for the 13 differences is 0.00012 ±0.00034. Here the 
probable error has been determined by the formula given above, the 
statistical analysis having given the constants presented in Table V. 
Table V .—Statistical constants for the ratio of electrical conductivity to freezing point 
lowering in Pima and Upland cottons in IQ 20 
Pima, P, 
Upland, U. 
Mean. 
0.02779 ±0.00105 
0.00560 ± 0.00074 
20.17 
0.02767 ±0.00104 
°- 00 555± 0 - 00 °73 
20.04 
S. D. 
C. V. 
rpu — +0.9479 ±0.0189 
Turning back to a comparison of the Egyptian with cottons other 
than the Upland, we note from samples 7 and 8 that the osmotic pressures 
and electrical conductivities are higher in Pima than in Sea Island cot- 
ton but that the ratio of electrical conductivity to freezing-point lowering 
is lower in Pima. Samples 11 and 12 show the same relationship to 
hold in a comparison of Pima Egyptian with the Transcaucasian Gos- 
sypium herbaceum var. Wa-Gale. 
While these two cases taken alone are altogether inadequate as bases 
for conclusion concerning differences between Pima Egyptian cotton and 
other species of the genus Gossypium, it is to be noted that the differ¬ 
ences are of precisely the same kind as are found in the comparison of 
the Egyptian and Upland types. 
The results, therefore, strengthen the conclusion to be drawn from 
a comparison of Egyptian and Upland cottons, and suggest the interest 
of a more extensive comparative study of the different cotton species. 
INVESTIGATIONS IN 1921 
The organization of the experimental details of the work carried out 
in 1921 has been fully described above. It remains merely to state that 
two complete series of determinations were made and to note certain 
differences in the conditions of these series. 
The sampling of the culture was begun on August 6, and the first 
series of samples—80 triplets and 60 duplets, 360 samples in all, less a 
few for which plants were not available—was completed on August 16. 
At the time the first series was taken the soil moisture was ample 
but the surface layer was not wet. The plants at the beginning were in 
a rapidly growing condition, with but few bolls developing. Their 
growth was, however, being checked by the inadequacy of soil moisture 
when the collections were completed. 10 
The plants were given an ample irrigation immediately after the 
last samples were taken. After this irrigation there was an unusually 
n Irrigation was delayed somewhat beyond the time at which water might advantageously have been 
applied in order that the first series of determinations might be completed in the period between two 
irrigations. 
