Capacitance probes are increasingly being used to monitor volumetric water content (VWC) in field conditions and are provided with in-built factory calibrations so they can be deployed at a field site without the requirement for local calibration. These calibrations may not always have acceptable accuracy and therefore to improve the accuracy of such calibrations soil-specific laboratory or field calibrations are required. In some cases, manufacturers suggest calibration is undertaken on soil in which the structure has been removed (through sieving or grinding), whilst in other cases manufacturers suggest structure may be retained. The objectives of this investigation were to (i) demonstrate the differences in laboratory calibration of the sensors using both structured and unstructured soils, (ii) compare moisture contents at a range of suctions with those predicted from soil moisture release curves for their texture classes (iii) compare the magnitude of errors for field measurements of soil moisture based on the original factory calibrations and the laboratory-based calibrations using structured soil. <br><br> Grinding and sieving clay soils to < 2 mm and then repacking the clay to bulk densities similar to in-situ field bulk densities was found not to represent the same field conditions for accurate VWC. When adding > 50 % water to the ground and sieved soil samples, dielectric values to VWC > 50 % were observed to be significantly lower than using undisturbed soil cores taken from the field and therefore undisturbed soil cores were considered to be better to calibrate capacitance probes. Generic factory calibrations for most soil sensors have a range of measurement from 0 to 50 %, which is not appropriate for the studied clay-rich soil, where ponding can occur during persistent rain events, which are common in temperate regions.