274 



Forestry Quarterly 



three samples vary more than 5 per cent from their mean the 

 process is repeated. Usually a 5 per cent variation from the 

 mean weight is allowed for seeds making 1,000 or more to the 5 

 grams, and a 10 per cent allowance for larger seeds. The dry 

 weight, number per hectoliter, and other customary measures are 

 listed for agricultural seeds. 



The three samples of 200 seeds previously counted and weighed 

 may be used for the germination test, which is made in triplicate, 





,J ^P6^M!lF:^=jhZ^sm-^m^-== 



CL, C^/oss strc/ys, 

 C. 3i/r?ser7 /?c/r/7er. 



their mean result being used as the figure representing the true 

 germination per cent. However, should the result of one of the 

 tests vary more than 10 per cent from the mean of the three, 

 this figure must be disregarded, and if the results show more 

 than 6 per cent variation between the dififerent parallel tests the 

 whole test must be repeated. 



Practically all tests are made in the Jacobsen germination ap- 

 paratus, which consists of a square metal tank about five inches 

 high within which narrow strips of glass rest upon an inside rim 



