Does labelling frequency affect N rhizodeposition assessment using the cotton-wick method?

The aim of the present study was to test and improve the reliability of the 15N cotton-wick method for measuring soil N derived from plant rhizodeposition, a critical value for assessing belowground nitrogen input in field-grown legumes. The effects of the concentration of the 15N labelling solution and the feeding frequency on assessment of nitrogen rhizodeposition were studied in two greenhouse experiments using the field pea (Pisum sativum L.). Neither the method nor the feeding frequency altered plant biomass and N partitioning, and the method appeared well adapted for assessing the belowground contribution of field-grown legumes to the soil N pool. However, nitrogen rhizodeposition assessment was strongly influenced by the feeding frequency and the concentration of labelling solution. At pod-filling and maturity, despite similar root 15N enrichment, the fraction of plants' belowground nitrogen allocated to rhizodeposition in both Frisson pea and the non-nodulating isoline P2 was 20 to more than 50% higher when plants were labelled continuously than when they were labelled using fortnightly pulses. Our results suggest that when 15N root enrichment was high, nitrogen rhizodeposition was overestimated only for plants that were 15N-fed by fortnightly pulses, and not in plants 15N-fed continuously. This phenomenon was especially observed for plants that rely on symbiotic N2 fixation for N acquisition, and it may be linked to the concentration of the labelling solution. In conclusion, the assessment of nitrogen rhizodeposition was more reliable when plants were labelled continuously with a dilute solution of 15N urea.