Experiments on nitrogen fixation by nodulated lucerne

Abstract

Field experiments at five sites (two on clay loam, two on a sandy soil and one on chalky loam) compared dry-matter yields and nitrogen contents of lucerne and ryegrass. These tested Rhizobium inoculation and fertilizer treatments in an experimental design proposed by Vincent and Nutman (IBP 69(66), amended 57/68). The crop and inoculation treatments (using large inocula of effective or ineffective strains) were set out in strips across each site to minimise contamination, and the fertilizer treatments (nitrogen, limeetc.) distributed at random within strips. This proved satisfactory and allowed valid statistical analysis. The massive inoculation with the ineffective strain suppressed the few naturally occurring effectiveR. meliloti at some of the sites, and provided a basis of comparison for assessing the amounts of nitrogen fixed by the native strains and by the introduced effective strains. However, it did so only during the first year because the introduced ineffective strains were overgrown by effective strains in the second year. Hence, ryegrass without nitrogen fertilizer was a better measure of the amounts of nitrogen obtainable from the soil. Estimates of fixation ranged from 0 to more than 300 kg N ha−1 at the different sites, treatment comparisons indicating the main factors limiting fixation. Parallel experiments were done using a soil core technique described by Vincent. These gave very similar results to the field trials in their first years.