Freshwater carbonates (tufas) develop today from the Arctic to the tropics, many being localized about springs and upper water courses. Some Quaternary tufas, especially in the Mediterranean region, extend over tens of square kilometres and exceed 30 m in thickness. Radiometric dating of Holocene deposits shows that many have accumulated at an average rate of 1 mm year−1. However, local precipitation may be much faster and some Holocene deposits may even have outpaced their tropical marine carbonate counterparts. Recently, the study of active sites has attempted to quantify the precipitation mechanisms which lead to tufa deposition. However, field observation and sampling procedures suffer from the inherent disadvantages of uncontrolled fluctuations in environmental conditions during the study programme. These disadvantages compromise any interpretations, particularly where controls on spar versus micrite precipitation are concerned. Many of these problems have been overcome in the current study by the construction and operation of laboratory mesocosm flumes which simulate the natural conditions (e.g. pH, flow rate, ambient temperature and daylight) in which freshwater carbonate (tufa) is deposited. Three mesocosms were supplied with natural river water from tufa precipitating streams and two mesocosms were supplied with UV-treated (sterile) river water from the same source. One of the untreated flume mesocosms was linked with a calcium reactor, which replaced calcium ions removed during the precipitation process in order to maintain tufa growth over extended experimental runs. Low-magnesium calcite precipitates (both rhombic sparite grown from long-crystallite dendrites and short-crystallite dendrite triad precursors) and micrite peloids (grown from spherulitic precursors) were precipitated in intimate association with biofilm (extracellular polymeric substances) within the four mesocosms supplied with natural river water. Virtually, no tufa-like precipitate was obtained from the flumes supplied with UV-treated river water. A second extended run flume experiment was also carried out for comparison purposes using a calcium hydroxide solution in deionized water. Collectively, these experiments provide convincing evidence confirming that the presence of a microbial biofilm strongly influences the precipitation of carbonates in riverine freshwater settings. In particular, experimental results show that micro-peloidal micrite and short-crystallite calcite dendrites are only produced in the presence of microbial extracellular polymeric substances.