Knowledge on the quantity and dynamics of rhizodeposition under ecologically realistic conditions may elucidate various aspects of soil organic matter dynamics. Data from a field experiment with C-14 pulse-labelling of spring wheat at different development stages, were used to estimate rhizosphere carbon fluxes. Not only the flux of C to the roots was assessed but also the fluxes of organic and inorganic release of root-derived material. C fluxes were calculated from curves fitted to data on shoot and root biomass and to data on C-14 distribution at different development stages. The C-14 distribution curves were extrapolated from the first labelling date (elongation stage) down to crop emergence and from the last labelling date (dough ripening stage) up to crop harvest, using different extrapolation procedures. The results show that while the maximum shoot growth rate occurred around ear emergence, the flux of C to the roots had a maximum around tillering. Over the entire growing season, shoot growth amounted to 5730 kg C ha-1 and 2310 +/- 90 kg C ha-1 was translocated belowground. Of this 920 +/- 150 kg C ha-1 was lost in root respiration and 500 +/- 120 kg C ha-1 was released as young photosynthate rhizodeposits, which are defined as organic materials released from the roots within 19 days after assimilation. Root growth amounted to 940 +/- 40 kg C ha-1, of which, however, 370 +/- 40 kg C ha-1 was lost again through root decay. Root turnover during the growing season, defined as root decay divided by root growth, was therefore 37-42%. Most of the organic input to soil (56-64%) occurred through rhizodeposition, while 36-44% was comprised in root biomass at crop harvest. The model used for the calculation of the carbon fluxes is discussed.