Crystallization of calcium oxalate is studied mainly in the diluted healthy urine using scanning electron microscopy (SEM), and is compared with the crystallization in the diluted pathological urine. It suggests that the average sizes of calcium oxalate crystals are not in direct proportion to the concentrations of Ca2+ and Ox2- ions. Only in the concentration range of 0.60-0.90 mmol/L can larger size of CaOx crystals appear. When the concentrations of Ca2+ and Ox2- ions are 1.20, 0.80, 0.60, 0.30 and 0.15 mmol/L in the healthy urine, the average sizes of calcium oxalate crystallites are 9.5 X 6.5, 20.0 X 13.5 and 15.0 jj,m X 10.0 jj,m, respectively, for the former three samples after 6 d crystallization. No crystal appears even after 30 d crystallization for the samples of concentrations of 0.30 and 0.15 mmol/L due to their low supersaturations. The results theoretically explain why the probability of stone forming is clinically not in direct proportion to the concentrations of Ca2+ and Ox2- ions. Laser scattering technology also confirms this point. The reason why healthy human has no risk of urinary stone but stone-formers have is that there are more urinary macromolecules in healthy human urines than that in stone-forming urines. These macromolecules may control the transformation in CaOx crystal structure from monohydrate cal-cium oxalate (COM) to dihydrate calcium oxalate (COD). COD has a weaker affinity for renal tubule cell membranes than COM. No remarkable effect of the crystallization time is observed on the crystal morphology of CaOx. All the crystals are obtuse hexagon. However, the sizes and the number of CaOx crystals can be affected by the crystallization time. In the early stage of crystalli-zation (1-6 d), the sizes of CaOx crystals increase and the number of crystal particles changes little as increasing the crystallization time due to growth control. In the middle and late stages (6-30 d), the number of crystals increases markedly while the growth rate changes little due to the nucleation control.