The effective time response of a tensiometer embedded in soil of low hydraulic conductivity (e.g. clay) is greater than that measured in free water. Theoretical analysis, based on simplifying mathematical and physical assumptions, is developed to estimate this influence of the soil, and shows that for a typical rapid response tensiometer (e.g. a pressure-transducer type), a time constant of 0·5 s in a loam can lengthen to one of 10 min in a very heavy clay. Experiments indicate that the theory probably underestimates the true value. To increase the response time in soils of low conductivity, it is necessary to increase the gauge sensitivity (i.e. the change in pressure per unit change in volume of fluid entering the gauge), rather than change the cup conductance, which has negligible effect. In practice, the available choice of design parameters is limited, and the analysis is of more practical use for assessing the potential of a given system for a given purpose, and for interpreting the significance of observations.