Pedogenic carbonate can serve a useful tool in paleoenvironmental reconstructions and dating soils in arid regions. However, its applicability as a paleoecological and chronological proxy is complicated by poor understanding of the processes of diagenetic alteration of carbonate in soils and paleosols. Although it is generally accepted that carbonate re-crystallization is most intensive in upper soil horizons and decreases with depth, its rates are difficult to quantify. To estimate the rate of carbonate re-crystallization, we used the isotopic exchange between primary carbonates of loess and 14C respired from the rhizosphere of wheat that was artificially labeled in a 14CO2 atmosphere under controlled conditions. Although similar approaches were frequently used in various rhizosphere and CO2 flux studies, to our knowledge this is the first time that such an approach has been used for carbonates. Young plants of spring wheat were grown in closed pots (7 cm in height) containing loess with 27% CaCO3. The hole around the plant was hermetically sealed by a silicon paste. All plants were labeled simultaneously in a labeling chamber having an air fan for intensive internal air circulation. A label consisting of 210 kBq of 14C as Na214CO3 solution was put in a vial connected with tubings to the labeling chamber. 5 M H2SO4 was added to the Na214CO3 solution to release 14CO2 into the chamber. Assimilation took place within 2 hours after the 14CO2 pulse had been applied. Thereafter, the plants were grown under normal atmosphere conditions. One to four labellings were applied. During 5 days after each labeling, the CO2 evolved from root-loess compartment (representing root-derived CO2) was trapped in 50 ml of a 1.0 M NaOH solution. One week after each labeling, the plant roots were removed from loess, and organic substances were washed out. 14C incorporation in CaCO3 was estimated by addition of H3PO4 to loess, trapping of evolved CO2 in NaOH and measuring its 14C activity. The CaCO3 re-crystallization rates were calculated according to 14C specific activity of root-derived CO2 trapped from the pot and 14C activity recovered in CaCO3 of loess. An ascending number of 14CO2 pulses (1 … 4) showed a linear increase of rhizosphere 14C recovered in the CaCO3 of loess. Based on this connection, the initial re-crystallization rates of loess carbonate were calculated by linear regression: for used loess containing 27% CaCO3, the initial rate of carbonate re-crystallization was 0.000029 d-1. Subsequently, using linear and exponential approaches with different lengths of growing season, we extrapolated the observed CaCO3 re-crystallization on longer time periods. The calculations show that at least 100 years, but probably between 400 and 2000 years, are necessary for full (99%) re-crystallization of the CaCO3 of loess. We suggest a general equation for calculating the remaining not re-crystallized CaCO3 depending on time (t): %CaCO3(t) = 100  exp(-t  0.00078  Growing-Season-Length / 365 / initial-CaCO3-percentage). It should be emphasized that this relationship is valid for the uppermost 7 cm of a loess soil at an initial stage of pedogenesis. To test the model, we compared it with radiocarbon dates for carbonate in the uppermost 8 cm of soil profiles from regions with different lengths of the growing season. The comparison indicated a good agreement between these 14C dates and our calculations. The findings in this work suggest that there is a mean residence time for carbonate materials in soil which may represent a limiting factor for chronological resolution of proxy records from pedogenic carbonate.