Permeability measurements were conducted with the flow-pump method on sand, sandy silt and silty clay specimens in a conventional triaxial system by introducing and withdrawing water at known constant flow rates into the base of a specimen with a flow-pump, and by monitoring the head difference induced across the length of the specimen with a sensitive differential pressure transducer. The results show that the previously reported advantages of the flow-pump method, compared with conventional constant head and falling head methods, were realized for permeability measurements in conventional triaxial equipment. These advantages are that direct flow rate measurements are avoided together with the associated errors that arise from the effects of contaminants on capillary menisci and the long periods of time involved in flow rate measurements; permeability measurements can be obtained much more rapidly and at substantially smaller gradients; errors from the small intercept in the otherwise linear flow rate versus hydraulic gradient relationship, and also from seepage-induced permeability changes, can easily be recognized and avoided or minimized. The results further show that the initial transient response of a specimen that precedes the steady state condition needed for a permeability measurement can require a substantial period of time. In this study the response times varied from a fraction of a minute for a sand specimen to more than 200 minutes for a silty clay specimen. Errors in permeability measurements from this transient response can easily be avoided with the flow-pump method but not with constant head and falling head methods.