88 Forest Club Annual 



of seed in each sample are averaged, and the daily germination 

 curve plotted. From this, the practical germination per cent and 

 the final germination per cent for the sample is established. 



The advantage of testing each sample of 500 seeds in series 

 of 100 is self-evident. There is, of course, always a possibility 

 that even in a carefully selected sample of 500 seeds there may 

 be a number of abnormal seeds which will influence the result 

 markedly one way or the other. If the 500 seeds are tested as 

 one series, there is no possible way of detecting such an abnor- 

 mality, and the final result is likely to be misleading. But in 

 testing the 500 seeds as series of one hundred each, considerable 

 opportunity is offered to detect any such abnormality and, if 

 necessary, repeat the test. Even this is open to the criticism 

 that the abnormality may be evenly divided over the 5 series of 

 100 seeds each, but this is not very likely. At any rate, the 

 chance for error is greatly reduced even if it is not entirely 

 eradicated. 



NECESSITY OF STANDARDIZING GERMINATION TESTS. 



The investigator must not lose sight of the fact that, as yet, 

 methods of testing coniferous seeds are not definitely worked 

 out. Furthermore, seed testing is not an exact science for which 

 a hard and fast set of rules to be invariably followed can be 

 promulgated. The factors influencing any given seed test are 

 too varied to be brought under exact control. Then, too, special 

 methods must be evolved for special investigations, which still 

 further complicates the problem. 



However, it is highly desirable that methods of making 

 germination tests of coniferous tree seeds should be standardized 

 as rapidly as is compatible with our increasing knowledge of the 

 subject. This is absolutely essential if germination tests are to 

 be of any but local value and if they are to be of practical use to 

 the forest nurseries at large. For example, the relative value of 

 the seed means nothing now if one sample of Douglas fir col- 

 lected in the Cascade Mountains of Oregon gives a practical 

 germination of 67 per cent and a final germination of 75 per cent 

 while another collected in central Colorado gives a practical 

 germination of 50 per cent and a final germination of 60 per cent 

 unless the tests were made by the same establishment. Of course, 

 it is evident that for exact experiments the comparative tests 

 must be made at the same time, in the same place and under the 

 same conditions ; but the results obtained from a standard method 

 applied at all testing establishments would be sufficiently ac- 

 curate and comparable to prove of great value in nursery and 

 field sowing. 



