Feb. 2,1924 
Tissue Fluids in Cotton 
281 
heavy rain. This left the fine silt in such a state that collections could 
not be taken up for the second series until August 19. The second 
series was completed on August 27. 
In the first series collections were made from the single Egyptian and 
the single Upland plant adjoining the hybrid plant or adjoining each 
other. This was done in order to obtain the samples from plants sub¬ 
jected to as nearly identical environmental conditions as possible. 
The method of collection followed in the first series has the disadvan¬ 
tage that the materials upon which the physical constants are to be 
based are drawn from single individuals, and are therefore subject to the 
variations characteristic of individual plants. 
In taking up the second series of collections, it was evident that the 
two plants of Egyptian and Upland cotton from which the first samples 
had been taken were so depleted of the more mature leaves that it would 
be desirable to extend the range of individuals somewhat, and to include 
the adjoining plants in the taking of the samples representing the two 
parent forms. This could not be done in the case of the hybrid, which 
was represented by but a single individual in each triplet, but there was 
little difficulty in obtaining adequately large samples from the F* hybrid 
plants, which were larger than the others. 
When the second series of determinations was taken up, it seemed de¬ 
sirable to make all the readings in duplicate. Time was available for 
this extra safeguard because of the standardization of every detail of the 
laboratory routine which had been worked out in the first series of deter¬ 
minations. The conductivities were reread, frequently with a different 
resistance; the freezing-point determinations were repeated on the same 
sample of fluid, thawed after the first freezing; and the hydrogen-ion 
determinations were made in duplicate, or repetition readings were made 
with the same electrodes. The average of these duplicate determina¬ 
tions have been used in the final calculations of the statistical constants. 
Except for a very few cases in which there was an obvious slip of the pen 
in recording the reading, the results of the first and second determinations 
check remarkably well. 
Limitation of space precludes the publication of the individual deter¬ 
minations, nearly 3,000 in number, upon which our conclusions are based. 
It is necessary, therefore, to summarize the determinations by treating 
them statistically. 
This may be done in two ways: First, the determinations may be 
arranged in an orderly manner and the frequency distribution of the 
magnitudes considered; second, statistical constants may be determined 
for the various subseries, and conclusions based on the comparison of 
these statistical constants, with due regard to their probable errors. 
We have first of all to consider the groupings of the materials for 
seriation and for the determination of statistical constants. While the 
constants are available for the duplets or triplets of the individual rows, 
it does not seem expedient for present purposes to seriate the constants 
by rows, or to determine the statistical constants for such a detailed 
grouping of the materials. We have therefore combined the rows 
according to variety and origin of seed. 11 Thus rows 1 and 5, rows 3 and 
7, and rows 2, 4, and 6 have been combined. 
11 In working with the original records for another purpose we have been careful to note that the con¬ 
clusions drawn for the combined materials of the individual varieties are in general substantiated by the 
results for the individual rows. Because of the influence of the number of determinations on the probable 
errors of random sampling, the constants for the individual rows will show greater irregularity than those 
based^On the combined records Of two or more rows. 
