This article describes an updated computer model which attempts to simulate known renal reabsorption and secretion activity through the nephron (NEPHROSIM) and its possible relevance to the initiation of calcium-containing renal stones. The model shows that, under certain conditions of plasma composition, de novo nucleation of both calcium oxalate (CaOx) and calcium phosphate (CaP) can take place at the end of the descending limb of the Loop of Henle (DLH), particularly in untreated, recurrent idiopathic CaOx stone-formers (RSF). The model incorporates a number of hydrodynamic factors that may influence the subsequent growth of crystals nucleated at the end of the DLH as they progress down the renal tubules. These include the fact that (a) crystals of either CaOx or CaP nucleated at the end of the DLH and travelling close to the walls of the tubule travel at slower velocities than the fluid flowing at the central axis of the tubule, (b) the transit of CaOx crystals travelling close to the tubule walls may be delayed for up to at least 25 min, during which time the crystals may continue to grow if the relative supersaturation with respect to CaOx (RSS CaOx) is high enough and (c) such CaOx crystals may stop moving or even fall back in upward-draining collecting ducts (CD) owing to the Stokes gravitational effect. The model predicts, firstly, that for small, transient increases in plasma oxalate concentration, crystallisation only takes place in the CD and leads to the formation of small crystals which are comfortably passed in the urine and, secondly, that for slightly greater increases in the filtered load of oxalate, spontaneous and/or heterogeneous nucleation of CaOx may occur both at the end of the DLH and in the CD. This latter situation leads to the passage in the final urine of a mixture of large crystals of CaOx (arising from nucleation at the end of the DLH) and small crystals of CaOx (as a result of nucleation originating in the CD). As a result of the higher calcium and oxalate concentrations in the urine of RSF, these patients have an increased probability of initiating CaOx crystallisation in the DLH and so of going on to form the large crystals and aggregates found in their fresh urines, but not in the fresh urines from normal subjects (N). These predictions are supported by evidence from clinical studies on six RSF and six normal controls (NC) who were maintained for 4 days on a fixed basal diet. Their patterns of CaOx crystalluria were measured on the second day of the basal diet and after a small dose of sodium oxalate was given before breakfast on the fourth day of the study. The model also shows that the tubular fluid of RSF is more likely than that of N to reach the conditions necessary for de novo nucleation of CaP at the end of the DLH. This may occur following either a small increase in ultrafiltrable phosphate, as a result of ingestion of a high phosphate-containing meal, or a small decrease in the proximal tubular reabsorption of phosphate resulting, for example, from increased parathyroid activity. CaP crystals initiated at this point may heterogeneously nucleate the crystallisation of CaOx under the high metastable conditions of RSS CaOx which frequently exist in the urines of RSF. Under certain conditions, it is predicted that CaP crystals, initiated at the end of the DLH and travelling close to the tubular walls where their transit time is increased, might also be able to grow and agglomerate sufficiently to become trapped at some point in the CD and lead to the formation of Randall's Plugs in the Ducts of Bellini. Currently, work is under way to incorporate data on the growth and aggregation of crystals of CaP into NEPHROSIM to confirm the likelihood of this phenomenon occurring. The model shows that an increase in plasma calcium is unlikely to lead to spontaneous nucleation of either CaOx or CaP at the end of the DLH unless the concentration of plasma calcium reaches values usually associated with the cases of primary hyperparathyroidism. The most likely cause of spontaneous CaOx crystal formation at the end of the DLH is a small increase in plasma oxalate; the most likely cause of spontaneous CaP crystal formation at the end of the DLH is either an increase in plasma phosphate or a decrease in the fractional reabsorption of phosphate in the proximal tubule. The model predicts that the maximum volume of CaOx crystalluria that is likely to occur in a given urine is a function of both the RSS CaOx and the oxalate/calcium ratio in the final urine. These data explain why the volume of CaOx crystalluria is in the order UK normals < UK recurrent stone-formers < Saudi Arabian recurrent stone-formers which, in turn, probably accounts for the very high incidence of CaOx-containing stones found in Saudi Arabia compared with that in the UK.