274 
Journal of Agricultural Research 
Vol. XXVII, No. 5 
nearly as possible the same time. Probably not more than five min¬ 
utes elapsed in any case between the sampling of any two members of 
a group. Generally the time was much less than this. Probably no 
appreciable source of error was introduced in this way. 
In the case of the loss of any sample through breakage, contamination, 
or through any discernible evidence of an erroneous reading, new samples 
of both members of the duplet or of all three members of the triplet 
were taken before the completion of the whole series of determinations. 
Rows i and 5 were planted to self-fertilized Pima and self-fertilized 
Meade. Rows 3 and 7 were planted to Pima and Meade grown from 
bulk seed. Thus it is possible to consider in a preliminary way whether 
there are differences between the sap properties of these two classes 
(those grown from self-fertilized and those grown from open-fertilized 
seed) of plants. Again, rows 1, 3, 5, and 7 were planted to Pima 
Egyptian and to the Upland variety Acala. Thus it is possible to com¬ 
pare the two varieties of Upland cotton in a wholly preliminary way. 
If the samples had been taken systematically across the field, it is 
evident that changing conditions from day to day might be a source of 
large differences between the several rows. The collections were, there¬ 
fore, made from duplets or triplets selected as nearly as possible at random 
from different portions of the field. 
In order to avoid in so far as possible differences due to age, collec¬ 
tions were made primarily of matured leaves from the main stem or 
the largest branches of the plants. Great care was taken to select 
perfect leaves. 
A possible source of variation in the constants is personal equation in 
the making of the collections. The taking of the samples presented 
considerabie difficulty because of the limited time in which it was neces¬ 
sarily carried out in order to have environmental conditions as nearly 
comparable as possible. Three workers, therefore, took part in collecting 
the material. 
In the first series of collections the samples were taken by two work¬ 
ers, and notes were taken by the third. Had one of the two workers 
taken either Egyptian, Upland, or hybrid samples continuously, it seems 
quite possible that differences in personal equation might have influenced 
the results. But the two collectors generally alternated in the variety 
collected. Thus, any systematic influence of personal equation was 
avoided. 
It is evident, however, that if there be a real personal equation on the 
part of the different collectors, this will tend, in the system of collection 
adopted for the first series, to obscure the difference between the forms. 
To avoid this possibility, a modification of the plan was adopted in 
the second series of collections. One worker, who had enjoyed longer 
experience in botanical work than the others, collected all the leaves 
used for samples. Thus the possibility of an influence of personal equa¬ 
tion in the collection of the samples would seem to have been practically, 
if not entirely, eliminated in the second series. 
ANALYTICAL METHODS 
The samples of tissue were placed at once in rubber-stoppered glass 
tubes and in a very short time were in a cooling mixture for freezing an¬ 
tecedent to the extraction of the sap, in accordance with the findings of 
