68 



XI. IMPURITIES IN SAMPLES OF GARDEN 



SEEDS. 



Over one hundred packages of seeds have been carefully exam- 

 ined for impurities, and in ninety separate instances the results 

 have been tabulated and compared. This examination consisted 

 in counting every seed in the sample, counting the impurities, 

 weighing the seeds and the impurities, and determining, so far as 

 possible, the character of the impurities. The percentages of im- 

 purities, both by number and weight, have been calculated. From 

 these analyses it is easy to draw conclusions as to the probable ex- 

 tent of adulteration or impurity in garden seeds. No evidence 

 of adulteration was found, and weed seeds were few and unimpor- 

 tant. In some cases the sample had not been properly cleaned, 

 but in general the more important seeds were very free from im- 

 purities. The impurities were very largely immature and imper- 

 fect seeds. The average of impurities, by number, was found to 

 be 2.76 per cent., and by weight, 1.38 percent. The investiga- 

 tion appears to indicate that there is no necessity for seed-control 

 stations in this country, for the purpose of preventing dishonesty 

 and carelessness in the sale of garden seeds. The detailed results 

 will soon appear in Agriculhiral Science. 



GENERAL SUMMARY. 



I. The results of a seed-test depend very largely upon the 

 known conditions under which the test is made : 



1. Variations in temperature may cause variations in rapidity 

 of sprouting. 



2. An essentially constant temperature of about 74° gives quick- 

 er results than an ordinarily variable temperature of a similar 

 mean. 



3. It is probable that any constant temperature gives quicker 

 results than a variable temperature of which the mean is the same 

 as the constant temperature. 



