76 BULLETIN 1105, U. S. DEPARTMENT OF AGRICULTURE. 



seed trees. Analysis of the data shows no such relationship when 

 we consider all the trees above 11 inches in diameter, but a surpris- 

 ingly close relationship when only the trees above 20 inches are 

 taken. Dividing the number of seedlings at the first count, in June, 

 1920, by the number of trees over 20 inches in diameter, the quotients 

 for the group-selection, scattered seed-tree, and shelterwood methods 

 are 1,040, 1,050, and 1,120, respectively. In other words, there were 

 roughly 1,100 seedlings for each tree. In view of the fact that 

 these counts were made 9 months after germination, and taking 

 into account the great difference in distribution of diameter classes 

 on the three areas, the constancy of this ratio must be regarded 

 as partly accidental, but that there is a definite relationship can 

 scarcely be doubted. The obvious conclusion to be drawn from 

 these observations is that the smaller trees, those between 11 and 

 approximately 20 inches in diameter, 16 are of relatively little im- 

 portance in reproduction. This deduction is substantiated in a broad 

 way by the findings under " Seed supply," which show that no less 

 than 5 blackjacks of the average run above 11 inches are required 

 to produce the same quantity of seed as a single free-standing yellow 

 pine of the type usually selected for seed production. On the 

 scattered seed-tree area, 75 per cent of the trees are over 20 inches 

 in diameter, whereas on the group-selection area tmly 30 per cent 

 fall in this class. It is not surprising, therefore, that the scattered 

 seed-tree area with a total of only 3.6 trees per acre should have 

 nearly three- fourths as many seedlings as the group-selection area 

 with its 14.2 trees per acre. Similarly, the shelterwood cutting, 

 which in June, 1920, had almost twice as many seedlings as the 

 group-selection cutting, evidently owed its superiority, not so much 

 to the 6 additional trees per acre in the 12 to 20 inch group, as 

 to the additional 3.2 trees per acre in the group above 20 inches, and 

 particularly to those above 25 inches. 



The real test of the three methods with respect to reproduction 

 comes in survival. Unfortunately the seedling counts on these plots 

 have not been made so frequently and regularly as on the plots nearer 

 the experiment station. Nevertheless, the records probably show the 

 trend of mortality with a fair degree of accuracy. From July 1, 

 1920, to June 15, 1921, the percentage of survival on the various cut- 

 ting divisions was as follows : Group selection 26, scattered seed tree 

 35, shelter wood 43. From June 15, 1921, to September 10 of the 

 same year, the percentages were 73, 97, and 88, respectively. The 

 group-selection area is lowest in both periods. The reasons for this 

 are not plain. More frequent examinations on plots near the experi- 



16 It has not been definitely determined where the lower diameter limit of effective 

 seed production lies. No doubt this varies with a number of factors. Sufficient infor- 

 mation is available, however, to warrant the adoption of 20 inches as a rough average 

 figure. 



