Methods in Germination Tests of Coniferous Tree Seeds 87 



The soil test is used, sand being the medium chosen, in 

 movable flats with inside dimensions of 4 inches deep by 12 inches 

 square for containers. When a new test is started in a flat, this 

 flat is filled level to the top with finely sifted sand just moist 

 enough to pack well, which is then pressed down with a board 

 just fitting inside of the flat. This board, which is y% inch thick, 

 has five strips nailed to it on the under side, each strip being 1 

 inch wide and J/s inch deep. These strips are spaced 1 inch 

 apart and the two outer strips are \y 2 inches from the edge of 

 the board. The board also has strips nailed on the top and pro- 

 jecting over the edge to prevent it from being pushed down more 

 than the required depth. After being pressed down with this 

 board the surface of the sand is just y% inch from the top of the 

 flat and five rows appear on the sand 1 inch wide, 1 inch apart, 

 the outer two rows \ l /> inches from sides of the flat. All the 

 rows are y% inch deep. Into each row is put a series of one 

 hundred seeds evenly distributed ; thus, each flat holds a com- 

 plete sample of five hundred seeds. Sand is then sifted over the 

 whole so as to more than fill the flat, the extra soil being 

 scraped off with a straight board. This results in a uniform sand 

 cover over the seeds to the depth of ^ inch, together with 

 practically uniform soil conditions in the seed bed. The slight 

 difference of compression of l /% inch between the rows and the 

 inter-row spaces appears negligible. 



The flat is then marked with the lot and sample designations 

 and the rows are given separate numbers. 



Next, the flat is placed on a bench in the greenhouse, where a 

 temperature range is kept within 25 degrees F., with a minimum 

 of 55 degrees F. and a maximum of 80 degrees F. The flat is 

 watered daily or often enough to keep the top of the sand moist. 



The date on which the germination test is commenced is 

 recorded. Daily observations are made and the date on which 

 germination commenced in each row is recorded. Then daily 

 counts are made of the number of seeds germinated in each row 

 and the final germination computed for each row at the end of 

 75 days, (since tests have shown that very few or no seeds 

 germinate after this period) by digging out and opening the 

 ungerminated seeds and adding the number of those with ap- 

 parently fertile embryos to the number already germinated. If 

 the final germination for all five rows does not vary more than 

 12 per cent, the sample is assumed to be representative. However, 

 if there is a variation of over 12 per cent, a new sample is 

 selected and the test repeated. If the resuks of the test show it 

 to be representative, the daily germinations of the five series 



