RESULTS OF SEED TESTS 



Made July 1, 1919, to July 1, 1920 



During the past year 563 samples of seeds have been analyzed 

 for purity and germination. The total number of samples col- 

 lected by agents employed and directed by the Commissioner of 

 x\griculture was 772. Of this number, 390 were not guaranteed 

 for purity and germination and consequently have not b^en an- 

 alyzed or reported in this bulletin. For their own personal use 

 farmers and other individuals have sent in for analysis 185 sam- 

 ples. 



In presenting herewith the results of findings relative to the 

 official samples use is made of the customary allowances for vari- 

 ations in tests. A difference of 5 per centum is satisfactory in 

 germination and for purity the following "tolerance" formula is 

 used. In this formula "T" stands for the variation allowed and 

 "P" for the percentage of purity stated. T= "^ io"~ ^ • ^or 

 example, a sample of seed guaranteed 90% pure and 90% germin- 

 able would be considered satisfactory if our findings showed a 

 purity of 88.0 and a germination of 85.5. 



We realize that many factors enter into the processes of seed 

 testing which may cause variation in results obtained. Under 

 the most careful and painstaking methods of sampling slight differ- 

 ences may occur; analytical errors may be made in the tedious job 

 of separating and weighing the different lots of seeds, and small 

 errors are always easy to overlook ; moisture and temperature varia- 

 tions may often affect the viability of samples enclosed in small 

 envelopes while waiting to be tested; finally, molds and tempera- 

 ture fluctuations in the germinator, after the most rigorous precau- 

 tions have been observed, may result in very slight differences. 

 For these reasons we believe the analyst should recognize his limi- 

 tations and be more than willing to allow some variations, and to 

 attach little signiiicance to occasional small discrepancies. The 

 goal to be attained is practical, not analytical, results. It is only 

 when the slight variations are consistently repeated or extreme 

 differences are noted that the dealer's good intentions should be 

 seriously challenged. 



Whenever the first test on a sample of seed could not b( 



