Carbon dioxide (CO₂) efflux from soils is one of the most important fluxes of the global carbon (C) cycle. CO₂ efflux measurements from soil are commonly used to investigate short-term soil organic matter turnover. Carbon recently assimilated by plants and released into the soil by exudation is very important in this short-term turnover, since it is a readily available C source for microorganisms, which becomes decomposed to CO₂ within a few hours to days. Root respiration represents a further important contribution to the total CO₂ efflux from soil, which must be considered separately from soil organic matter turnover. Various techniques based on ¹³C labelling, ¹⁴C labelling, and ¹³C natural abundance are used to separate these CO₂ flows and to quantify plant mediated C input into the soil as well as root-derived CO₂ fluxes. The separation techniques based on ¹³C natural abundance assume absence of ¹³C fractionation by root respiration and exudation, but this assumption was not thoroughly proven.

Coupling ¹³C natural abundance and ¹⁴C pulse labelling enabled us to investigate dependence of ¹³C fractionation on assimilate partitioning between maize shoots, roots, exudates, and root-respired CO₂. The amount of recently assimilated C in these four pools was controlled by three levels of nutrient supply: full nutrient supply (NS), ten times diluted nutrient supply (DNS), and deionised water (DW). Every container with one maize plant was sealed between root and shoot. Thus, root respiration could be examined in a closed system: air was pumped through the nutrient solution, CO₂ from root respiration was trapped in 1 M NaOH solution, and the resulting CO₂-free air was again pumped through the nutrient solution. The experiment consisted of three cycles started on days 14, 19, and 24. Each cycle included: 1) supply of the plants with full NS for recovery from DNS or DW for one day before labelling (CO₂ and exudates were collected during this period before second and third labelling), 2) labelling of shoots in a ¹⁴CO₂ atmosphere for 1.5 hours, and 3) trapping of CO₂ in NaOH and of exudates released into NS, DNS, or DW for four days. Total C, ¹⁴C activity, and d¹³C were determined in NaOH, in exudates, and in roots and shoots.

Increasing amounts of recently assimilated C in the roots (from 8 to 10% of recovered ¹⁴C in NS and DNS treatments) led to a 0.3‰ ¹³C enrichment with decreasing nutrient supply. A further increase of C allocation in the roots (from 10 to 13% of recovered ¹⁴C in DNS and DW treatments) resulted in an additional enrichment of the roots by 0.3‰. d¹³C of CO₂ evolved by root respiration was similar to that of the roots in DNS and DW treatments. However, if the amount of recently assimilated C in root respiration was reduced, the respired CO₂ in the NS treatment became 0.7‰ depleted in ¹³C compared to roots. Increasing amounts of recently assimilated C in the CO₂ from NS via DNS to DW treatments resulted in a 1.6‰ d¹³C increase of root respired CO₂. Thus, for both pools, i.e. roots and root respiration, increasing amounts of recently assimilated C in the pool with decreasing nutrient supply led to a d¹³C increase. In DW and DNS plants there was no ¹³C fractionation between roots and exudates. However, high-nutrient supply decreased the amount of recently assimilated C in exudates compared to the other two treatments and led to a 5.3‰ ¹³C enrichment in exudates compared to roots.

We conclude that ¹³C discrimination between plant pools and within processes such as exudation and root respiration cannot be accepted as fixed but strongly depends on the amount of C in the respective pool and on partitioning of recently assimilated C between plant pools.