| Autor: |
Cowling, D. W., Green, J. O., Green, S. M. |
| Zdroj: |
Journal of the British Grassland Society; Dec1964, Vol. 19 Issue 4, p419-424, 6p, 1 Chart |
| Abstrakt: |
Data from an experiment already reported (2) are used to examine a statistical method proposed by Walker et al. (8) for apportioning the nitrogen harvested in grass in a grass/ legume sward to legume, fertilizer and soil. Multiple regressions of the yield of N in the grass (Gn) upon N harvested in the associated legume (Cn) and N applied in fertilizer (Fn) were highly significant; but they were very variable, mainly because of the strong (negative) correlation between Cn and Fn. In some cases the coefficient for clover N was negative, in some cases the coefficient for fertilizer N exceeded unity; and the residual constant varied widely among 4 sets of plots on the same soil. The interpretation of such equations in physical terms is criticized on these grounds. The use of pure grass swards in conjunction with mixed swards gave estimates of the recovery of fertilizer N, and of the transfer of N from clover to grass, which agreed better with other published data than did the coefficients in a multiple regression. A simple regression of the estimate of transferred N in the grass upon the N harvested in the clover yielded a second constant. This constant, which varied considerably from one series of plots to another, is thought to indicate interactions between the various sources of N, which would be attributed to a direct effect of soil N if the multiple regressions were interpreted as Walker et al. suggest. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
