The experiments described are part of a broader investigation programme launched by Seiler in the 1980s (Seiler, Maloszewski and Behrens 1989; Seiler and Behrens 1992; Seiler 1999) in the Southern Franconian Alb, a karstified Upper Jurassic (Malm) carbonate formation in Southern Germany (north of Ingolstadt, with the main investigation area lying between Danube and the river Altmuehl), whose character, determined by two consecutive sedimentation cycles, is dominated by two contrasting facies types: the bedded limestones with a fissure/solution channel porosity amounting to 2% but with syngenetic porosities under 2%, and the dolomitised reefs with matrix porosities reaching 5-10%. The present investigation is specifically concerned with the reasons behind the lower denitrification activity in the bedded facies, as compared to the reef facies. With variable, but generally low DOC supply, and average `bedded'/`reef' concentration ratios differing from species to species: chloride — 1.5, nitrate — 2.2, sulphate — 1.2, sodium — 1.6, potassium — 1.6, atrazine — 1.9 (values reproduced after Seiler and Hartmann 1997), it is clear that the porosity contrast alone cannot account for these values, and that several (possibly coupled) chemical transformations are involved alongside with transport in this karst system. A number of papers have already been devoted to biodenitrification in the Southern Franconian Alb (Hausner, Lawrence, Wolfaardt et al. 2000; Miiller, Assmus, Hartmann et al. 2000; Naumann, Vomberg, Hartmann et al. 2001 etc.) and the issue of (organic) carbon availability has been strongly emphasized, but there has been no attempt, so far, to

look at the experimental findings through the eyes of a kinetic model for nutrient consumption and biomass growth. Here, a qualitative interpretation for the observed differences in nitrate levels is first sought in terms of carbon availability and/or limiting nutrient residence times. Second, one would like to obtain quantitative estimates for the relevant denitrification parameters as a basis for predicting future contamination levels reliably (as diffuse nitrate input also evolves with time under the influence of changing policies). Despite some ambiguity in parameter estimation, the model helps identify the factors determining the fate of nitrate in different facies types of the Southern Franconian formation. The question of what is actually used as an electron donor in biodenitrification, additionally or alternatively to DOC, has yet to be answered by different methods.