The two-ponding depth(TPD)analysis procedure of single-ring infiltrometer data can yield invalid results,i.e.,negative values of the field-saturated soil hydraulic conductivity or the matric flux potential,denoting failure of the two-level run.The objective of this study was to test the performance of the TPD procedure in analyzing the single-ring infiltrometer data of diferent types of soils.A field investigation carried out in western Sicily,Italy,yielded higher failure rates(40%)in two clay loam soils than in a sandy loam soil(25%).A similar result,i.e.,fine-textured soils yielding higher failure rates than the coarse-textured one,was obtained using numerically simulated infiltration rates.Soil heterogeneity and reading errors were suggested to be factors determining invalid results in the field.With the numerical data,allowing a less generic definition of soil heterogeneity,invalid TPD results were occasionally obtained with the simultaneous occurrence of a high random variation(standard deviation≥0.5)and a well developed structural correlation for saturated hydraulic conductivity(correlation length≥20 cm).It was concluded that a larger number of replicated runs should be planned to characterize fine-textured soils,where the risk to obtain invalid results is relatively high.Large rings should be used since they appeared more appropriate than the small ones to capture and average soil heterogeneity.Numerical simulation appeared suitable for developing improved strategies of soil characterization for an area of interest,which should also take into account macropore effects. The two-ponding depth (TPD) analysis procedure of single-ring infiltrometer data can yield invalid results, ie, negative values ​​of the field-saturated soil hydraulic conductivity or the matric flux potential, denoting failure of the two-level run. Objective of this study was to test the performance of the TPD procedure in analyzing the single-ring infiltrometer data of diferent types of soils. A field investigation carried out in western Sicily, Italy, yielded higher failure rates (40%) in two clay loam soils than in a sandy loam soil (25%). A similar result, ie, fine-textured soils yielding higher failure rates than the coarse-textured one, was obtained using numerically simulated infiltration rates. Small heterogeneity and reading errors were suggested to be factors determining invalid results in the field .With the numerical data, allowing a less generic definition of soil heterogeneity, invalid TPD results were occasionally obtained with the simultaneous occurrence of a high random variation (st andard deviation≥0.5) and a well developed structural correlation for saturated hydraulic conductivity (correlation length ≥20 cm) .It was approved that a larger number of replicated runs should be planned to characterize fine-textured soils, where the risk to obtain invalid results is relatively high. Larger rings should be used since they have been more than than small ones to capture and average soil heterogeneity. Numerical simulation was suitable for developing improved strategies of soil characterization for an area of ​​interest, which should also take into account macroprocessor effects .