Do techniques based on super(15)N enrichment and super(15)N natural abundance give consistent estimates of the symbiotic dependence of N sub(2)-fixing plants?

The primary aim of this review is to determine if methods based on super(15)N enrichment (E) and super(15)N natural abundance (NA) give consistent estimates of the proportional dependence of N sub(2)-fixing species on biological N sub(2) fixation (P sub(atm)), and secondly to attempt to explain any inconsistencies that may be found. Published estimates of the symbiotic dependence of N sub(2)-fixing plants based on E and NA techniques applied in the same experiment were compared across scales from glasshouse pots to field plots to landscapes in agricultural and forest ecosystems, which included grain legumes, pasture and forage legumes, and woody perennials. A meta-analysis of the published data was based on correlation coefficients, box-plots and confidence intervals of means. In some studies, estimates were reference plant dependent for both E and NA techniques, indicating temporal and/or spatial variations in the natural and artificial distribution of super(15)N, which can sometimes result in erroneous negative estimates of symbiotic dependence. While significant correlations were obtained between E and NA estimates of P sub(atm) for each of the three groups of N sub(2)-fixing species, the probability that the methods provided estimates of P sub(atm) within -5 to +5 % of each other was 0.29 or was 0.54 within -10 to +10 % of each other. We have identified a number of interacting factors that may contribute to the inconsistent agreement between estimates of P sub(atm) by E and NA techniques, which underlines the need for a re-examination of the fundamental assumptions on which each method is based, and whether those assumptions are valid in any given situation.