42 BULLETIN 1105. U. S. DEPARTMENT OE AGRICULTURE. 



tion. Another year's record may give decisive results; but it is 

 anticipated that no thoroughly consistent relationship will be es- 

 tablished, for the reason that conditions on the cultivated, as well as 

 on the natural, plots are not uniform. 



The effect of bringing the clay subsoil into the surface strata has 

 been mentioned in the discussion of germination. Greenhouse tests 

 have shown much slower growth in compact clay than in sandy and 

 loamy soils. Field tests show a similar relation with respect to 

 survival. We should expect the effects of cultivation to vary with 

 the character of the soil. Where the loamy surface layer extends 

 below the depth reached with the mattock, no injurious effects should 

 result ; but where the more clayey soil comes within 8 inches of the 

 surface, or where the mattock has gone below the usual depth, the 

 effect might be unfavorable. 



Opposed to the influence of clay content stands the favorable 

 effect of eliminating root competition. Undoubtedly the removal 

 of herbaceous plants on a plot of 5 by 10 feet would tend to increase 

 soil moisture in the lower strata reached by seedling roots in their 

 second year. But some vegetation has come back on all of the plots 

 (PI. VI), and in many cases the luxuriance of these plants makes 

 up for the deficiency in numbers. Moreover, nearly all the plots 

 are within reach of tree roots, which would tend to utilize any sur- 

 plus of soil moisture. On the whole, therefore, conditions of soil mois- 

 ture are probably little, if any, better on the cultivated plots than 

 on those bearing the original vegetation. 



Winter protection is another factor. Other things being equal, 

 the loss from winterkilling and ground heaving is greatest on bare 

 ground. As previously stated, the cultivated plots are no longer 

 bare, but the proportion of bare surface is greater than on the nat- 

 ural plots. Although no figures on late fall survival are available 

 for 1919, the favorable moisture conditions in the fall of 1919 and 

 early spring of 1920 make it safe to attribute the mortality between 

 September, 1919. and May or early June, 1920, almost entirely to 

 winterkilling and heaving. This conclusion is further strengthened 

 by the fact that the winter was open, with a prevalence of rain instead 

 of the usual deep snows. During the entire period there was no 

 question of deficient soil moisture, except in so far as it may havle 

 been rendered unavailable by freezing. The winter of 1919-20 fav- 

 ored both winterkilling and ground heaving. Theoretically, a some- 

 what higher rate of mortality was to have been expected on the cul- 

 tivated than on the natural plots. Average figures, however, show 

 practically no difference. Surprisingly little loss from any cause 

 occurred during the winter of 1920-21. 



