Preconstruction treatment or the selection and design of a foundation system both rely on accurate estimates of the potential heave of the supporting expansive soil. Swell percentage of expansive soils is commonly obtained from oedometer swell tests under fully lateral restraint conditions. In situ heave, however, is not a one-dimensional volume change but more likely a three-dimensional process. Consequently, surface heave predicted from results of oedometer tests is usually conservative and largely differs from heave actually observed in the field. To account for the discrepancy between oedometer and in situ boundary conditions, some researchers have suggested a lateral restraint factor to be applied to heave predictions evaluated based on parameters obtained from oedometer swell tests. This paper presents a laboratory testing procedure by which the vertical and the volumetric swells of compacted expansive soil samples are concurrently measured using a hydraulic triaxial stress path apparatus. A series of triaxial swell tests is conducted and the results are compared with the volume change measured in specimens tested under one-dimensional loading conditions in the oedometer. The percentages of vertical swell obtained from the triaxial swell tests are found to be considerably lower than the corresponding values measured in the oedometer tests. For the triaxial swell tests, the ratio of swell in the vertical direction to the volumetric swell, SR, is evaluated. The test results indicate that SR increases with increasing applied confining pressure, and for a particular swell test, the value of SR is not constant but rather changing with elapsed swelling time.