8 BULLETIN 644, IT. S. DEPARTMENT OF AGRICULTURE. 
On a recent visit to southern Georgia the writer was asked to com- 
pare two varieties of cotton growing in the neighborhood, for the 
benefit of the local cotton farmers. Ten 5-locked bolls of each of the 
two varieties were secured and examined, with the results shown in 
Table IXL Variety B was the local favorite, since it had the higher 
percentage of lint. In this section of Georgia, although the smaller 
farmers pick their own cotton, they had not appreciated the disad- 
vantages of the variet}'- they were growing, even from the standpoint 
of the labor of picking. 
INCREASING THE LINT PERCENTAGE DOES NOT ALTER THE COST 
OF PRODUCTION IF THE LINT INDEX REMAINS CONSTANT. 
In the case referred to, in which the same lint index was traced 
through all percentages of lint from 25 to 42, the labor of harvesting 
the crop and the efficiency of the laborers themselves are the same in 
each case. For, since the lint index, which determines the number of 
bolls to the pound of fiber, is constant, the number of seeds and of 
bolls producing a pound of fiber also remains constant. 
There is another relation of the lint index which has not been 
worked out as yet, but which may be suggested here as a possibility. 
Both Tables I and II show that the higher lint indexes are asso- 
ciated with the heavier seeds. Heavy seeds have a relatively larger 
percentage of kernels to hulls than smaller or lighter seeds, and the 
oil content may be found to be associated also with heavier seeds. 
IMPROVED METHODS FOR OBTAINING LINT PERCENTAGES. 
The usual method of obtaining the percentage of lint in cotton 
varieties is to weigh a random specimen of the seed cotton and gin 
it ; then reweigh the seed, calculate the percentage of seed, and set 
down the difference as the percentage of lint. Few workers in cot- 
ton selections weigh the lint after ginning and calculate the per- 
centage of lint directly, These operations occasion considerable 
labor and care in making the various records, and, of course, the 
more numerous the calculations and entries in the records the. greater 
the liability to error. Owing to these and other considerations it 
has been found advantageous to begin with a standard sample of seed 
cotton of 100 grams in weight, a method which avoids the necessity 
for recording the original weight of the specimen. After this stand- 
ard sample has been ginned the seeds are weighed. Each gram of 
seed then represents 1 per cent of the original seed cotton. The 
difference between the weight of the seed and 100 grams is the weight 
of the lint removed in ginning and is also the percentage of lint. 
Thus, this method avoids the necessity for recording the net weight 
of the seed and of calculating the percentage of lint. 
Since this procedure for finding the percentage of lint has been in 
operation, a balance has been placed on the market equipped for the 
