Evaluation of the xylem ureide method for measuring N2 fixation in six tree legume species

The xylem ureide method, based on analysis of xylem sap for N solutes associated with N2 fixation (ureides) and soil-N use (NO3−, α amino-N), is used for estimating Pfix (proportion of plant N derived from N2 fixation) of herbaceous species of the leguminous tribes Phaseoleae and Desmodieae and may have application for tree legumes. We report experiments, involving Calliandra calothyrsus, Leucaena leucocephala, Gliricidia sepium, Sesania grandiflora, Desmodium rensonii and Codariocalyx gyroides, to: (i) determine for each species the composition of N solutes of xylem sap and to calibrate, where possible, the xylem ureide method using 15N isotope dilution as an independent measure of Pfix; and (ii) establish appropriate procedures for sampling xylem sap in field studies. In the first experiment, the tree legumes were grown for 55–77 weeks and sampled repeatedly (6–9 times) for vacuum-extracted xylem sap, shoot dry matter, total N and 15N. Asparagine was the predominant amino compound in xylem sap of all species. The relative abundance of ureide-N in xylem sap (ureide-N as a proportion of ureide-N + α amino-N + NO3−N) was compared with Pfix for five of the six tree legumes. Sap of S. grandiflora could not be analysed for ureides and NO3− using the standard colorimetric methods because of high levels of non-specific, background colour. For D. rensonii and C. gyroides, relative ureide-N ranged from less than 20% (Pfix of zero) to around 65 and 40%, respectively (Pfix of 100%). Functions describing the relationships between relative ureide-N (y) and Pfix (x) were: y = 18.3 + 0.446x (r = 0.91) for D. rensonii; and y = 8.49 + 0.279x (r = 0.92) for C. gyroides. With the other three species, relative ureide-N was not related to Pfix. We concluded that the xylem ureide technique is suitable for estimating Pfix for D. rensonii and C. gyroides only. In the second experiment, established trees of G. sepium, D. rensonii, C. calothyrsus and C. gyroides were used to assess sampling protcols. Results indicated that vacuum extraction of sap should proceed immediately the branch is cut, that plants need to be sampled around noon to eliminate diurnal effects and that branch-to-branch variation in relative ureide-N values within a single tree was minor.