Catch crops can effectively decrease nitrate leaching in arable cropping systems but their long-term impacts on nitrogen mineralization are not well known. This study quantified the effects of continuous catch crops on net N mineralization, crop N uptake, crop N use efficiency and N leaching in three long-term (13-17 years) field experiments in northern France. Mustard was grown every year at Boigneville, radish every year at Thibie and ryegrass every 2 years at Kerlavic. The mean N content of catch crop residues at these sites was 33, 36 and 35 kg ha(-1) yr(-1) and their mean C:N ratio was 13, 17 and 28, respectively. Net mineralization was calculated with a mass balance of soil mineral N using measured inputs and outputs. Catch crop establishment enhanced annual mineralization by on average 26, 18 and 9 kg N ha(-1) yr(-1) respectively during the 13-17 year period. The difference in mineralization rate between catch crop and control treatments (extra mineralization) was positive from the first year at Boigneville, whereas it was negative or nil during the first 3-5 years at Thibie and Kerlavic. At the latter sites, the extra mineralization rate increased significantly over time at a rate of 2.0 and 2.6 kg N ha(-1) yr(-2). At the end of the experiment, cumulative extra mineralization represented 72%, 60% and 23% of the total N added by catch crop residues at Boigneville, Thibie and Kerlavic, respectively. Repeated catch crops significantly increased N uptake and N use efficiency by main crops at Kerlavic and Thibie, but not at Boigneville. The efficiency of catch crops in reducing N leaching persisted over the years at all sites. A model simulating C and N dynamics during catch crop decomposition was able to reproduce the pattern of extra N mineralization kinetics with the various catch crop species, but underestimated the range of variation between sites. Better predictions were obtained when C or N inputs due to catch crops were increased by 10-57%, suggesting that the actual inputs could be markedly greater than those measured in catch crop residues. According to the model, the C:N ratio of catch crop residues largely explained differences in mineralization due to different catch crop species.