18 BULLETIN 1056, IT. S. DEPARTMENT OF AGRICULTURE 
long rainy season which averaged only 55 per cent germination and 
which ran as low as. 29 per cent. 10 Also tests of seed from cotton 
harvested in the early fall months showed an average of 90 per cent 
germination, but that from mature cotton harvested in January 
averaging only 43 per cent and from frosted cotton harvested during 
the same month only 14 per cent germination. 11 
MAKING TESTS 
One or more representative samples of each lot of seed should be 
taken at the time of sacking and marked with the lot number, for 
use in making germination tests. The sample should contain at least 
1 ounce of seeds, about 300, in order that duplicate tests, using 100 
seeds in each, may be made, and a part of the sample retained for 
check. If there is any probability that a lot of seed has become dam- 
aged from any cause prior to harvesting and ginning, such seed 
should be tested as soon as ginned or delinted and held in bulk pend- 
ing the results of the germination test before being recleaned and 
sacked. With the possible exception of cases where there may be a 
shortage of high germinable seed of a particular variety, there is no 
reason for retaining damaged seed for planting purposes. If such 
seed is sold for planting purposes a second test should be made at the 
time it is sacked, and sales should be made on the basis of this latter 
test. 
The equipment necessary in making reliable germination tests is not 
expensive. Commercial germinators are on the market at around $50. 
Tests also may be made in sand or soil or by the use of blotting paper 
or canton flannel placed between ordinary dinner plates and kept 
warm and moist. Samples also may be sent to the seed laboratory, 
Bureau of Plant Industry, Department of Agriculture, Washington, 
D. C, or to the branch seed laboratory, College Station, Tex., for 
test without cost. 
The method of making the test will depend upon the nature of 
one's relation to the planting cottonseed business and the quantity 
of seed sold annually. The main point is that a reliable test should 
be made. The method of testing by cutting the seed with a knife 
and examining and tasting the kernel is inaccurate, if not fallacious, 
can not be relied upon, and should not be used or advocated. 
It is not proposed to fix a minimum or a standard percentage of 
germination at this writing, but rather to show that germination di- 
rectly affects the commercial as well as agricultural value of plant- 
ing cottonseed. The comparative commercial value of different 
grades of seed is governed largely by their agricultural value. A 
dead seed has no value for planting purposes. For this reason it is 
essential that accurate records of germination tests be kept, so that 
the information may be passed on or made available to the farmer or 
other purchaser. These records should include the lot number of the 
seed from which the sample was taken, the date on which the test 
was made, and the percentage of germination as shown by the test. 
The duration of the test and any general remarks having a bearing 
on the viability of the seed or the vigor and healthfulness of the pros- 
pective young plants may be included. 
10 TESTING COTTONSEED FOR GERMINATION*. Ark. Agl\ Exp. Sta„ Cir. 22. 
u COTTONSEED FOR PLANTING purposes. Ark. Agr. Exp. Sta., Cir. 37. 
