IDENTIFYING SIGNIFICANT ASSOCIATIONS 



Regression analysis was used to identify the soil 

 variables significantly associated with the release of 

 western redcedar. The dependent variable in the 

 regression models was the difference between the DDS 

 for released trees and the predicted nonreleased DDS 

 for the same tree (DDS "released" - DDS 

 "nonreleased"). Fourteen soil variables were tested for 

 association with the dependent variable. The predicted 

 diameter growth of each tree was included in each 

 regression model to eliminate variation not related to 

 the release response of the sample trees. Transformed 

 diameter increments for the periods 0-5, 5-10, and 

 10-15 years after treatment were used as dependent 

 variables in the regression models. 



Results 



Amounts of several soil nutrients and soil pH were 

 associated (P <0.05) with the growth difference between 

 released diameter growth and predicted nonreleased 

 diameter growth of western redcedar. Amounts of soil 

 nitrogen were significantly related to the difference 

 between released diameter growth and the predicted 

 nonreleased diameter growth of western redcedar as 

 denoted by the significant coefficients for the nitrogen 

 variables (table 6). The positive regression coefficients 

 for the nitrate and ammonium variables in the regres- 

 sion models indicated that trees growing on soils with 

 larger amounts of these forms of soil nitrogens had 

 greater release response than trees growing on soils 

 with lesser amounts of ammonium and nitrate. In 

 contrast, amounts of total soil nitrogen were negatively 

 associated with western redcedar diameter-growth 

 response to release. In other words, trees growing on 

 soils with the larger amounts of total nitrogen had 

 poorer diameter growth release response when 

 compared to trees growing on soils with lesser amounts 

 of total nitrogen. 



Amounts of soil sulfate and potassium showed 

 positive relationships with the release of western 

 redcedar from overhead and surrounding competition. 

 Both sulfate and potassium had positive regression 

 coefficients in the 0-5- and 10-15-year models (table 6). 

 The regression coefficients indicated that western 

 redcedar growing on soils with the larger amounts of 

 sulfate and potassium had greater diameter growth 

 response to release than western redcedar growing on 

 soils having lesser amounts of sulfate and potassium. 



Amounts of soil iron and copper, along with soil pH, 

 had negative relationships with the diameter increment 

 of released western redcedar. The regression coeffi- 

 cients for soil iron, and copper were significant only for 

 0-5- and 5-10-year periods, with the regression 

 coefficient for soil pH being significant for all growth 

 periods (table 6). These regression coefficients were 

 evidence that the response of western redcedar to 

 release was greater on soils with the lesser amounts of 

 copper, iron, and lower pH values, compared to tree 

 diameter growth on soils with the larger amounts of 

 copper, iron, and higher pH values. 



Significant interactions between soil characteristics 

 were detected using two variable regression models. For 

 the 0-5-year time period since release the significant 

 interactions identified were total nitrogen-iron and pH- 

 potassium (table 7). During the 5-10-year time period, 

 the significant interactions detected were total nitrogen- 

 copper, total nitrogen-iron, pH-iron, total nitrogen- 

 potassium, pH-sulfate, and total nitrogen-nitrate. Only 

 two significant interactions, total nitrogen-potassium 

 and pH-ammonium, were detected for the 10-15-year 

 period. The significant interactions were evidence that 

 for a given amount of soil constituent the response of 

 western redcedar to release was not proportional to 

 changes of the other constituent in the interaction. 



Discussion 



Several soil characteristics were significantly 

 (P£ 0.05) associated with the growth difference between 

 the diameter of released western redcedar and the 

 predicted diameter growth of nonreleased western 

 redcedar. Soil variables, including total nitrogen, 

 ammonium, nitrate, sulfate, copper, potassium, iron, 

 and pH had significant regression coefficients in the 

 release response models. 



Soil pH directly influences the uptake of soil nutrients 

 by plants. Soil copper, boron, zinc, manganese, 

 potassium, and iron are not readily absorbed by plants 

 at higher soil pH's (Alban 1958). The influence that soil 

 pH has on the availability of soil nutrients for plant 

 growth could explain why western redcedar growing on 

 high pH soils had the poorer response to release. Soil 

 pH could be limiting the uptake of soil nutrients, 

 thereby influencing the diameter increment of released 

 trees. Data for this study did show interactions of pH 

 and potassium having negative relationships with the 

 response of western redcedar to release. 



Nitrogen nutrition of forest trees is usually more 

 complex than is the nutrition of other major elements. 

 Nitrogen is most readily available for tree growth in the 

 nitrate and ammonium forms that often can be less 

 than 10 percent of the total nitrogen in a forest soil 

 (Jorgensen 1967). Much of the total nitrogen is 

 immobile or nearly so in organic matter and long 

 periods of time are involved in its breakdown 

 (Shoulders and McKee 1973). Switzer and others (1968) 

 showed that soils of a southern pine ecosystem con- 

 tained 1.700 lb (1 905 kg/ha) of total nitrogen per acre 

 but only 34 lb (38 kg/ha) per acre were available for 

 tree growth. Therefore it would be possible for western 

 redcedar to have a greater response to release on soils 

 with the greater amounts of available nitrogen, but also 

 a poor response to release on soils with the large 

 amounts of total nitrogen because it is possible that 

 only a small amount of the total soil nitrogen was 

 available for tree growth. 



Relationships of tree growth to soil nutrients can 

 easily be confounded by other tree and site variables. 

 Elevation and aspect can influence the availability of 

 soil nutrients (Leaf 1956). as can the tree genotype 

 (Steinback 1971). Likewise, tree rooting habit and the 

 ability of tree roots to penetrate forest soils can affect 



10 



