Abstract. Optimisation of gas production from shale gas reservoirs depends critically upon a good understanding of the porosity and pore microstructure of the shale. Conventionally surface area measurements or mercury porosimetry have been used to measure the porosity in gas shales. However, these conventional methods have limited accuracy and only provide a bulk measurement for the samples. More recently, scanning electron micrography (SEM) and Focussed Ion Beam SEM (FIB-SEM) techniques have been applied in an attempt to address these limitations. Unfortunately, these two methods destroy the samples. In this research three-dimensional x-ray micro tomography (XRMT) imaging techniques were used to capture the structure of three samples and also compared to data from mercury porisimetry. The resulting data have been segmented in order to recognize individual pores down to a resolution of about 1 µm. Distributions of pore volume, pore size, pore aspect ratio, surface area to pore volume ratios and pore orientations were calculated from the XRMT data. It was found that the porosity obtained from XRMT measurements is smaller than that obtained using mercury porisimetry, the reason for which might be displacement of kerogen by the high pressures generated in the mercury technique, but is unlikely to be due to both techniques not being able to measure pores smaller that about 900 nm. Pore volume and size distributions showed all of the shales tested in this work to be multimodal with similar major modal values for volume and pore size. The pores also have a range of pore aspect ratios and surface area to pore volumes, including values indicating the presence of significant oblate spheroidal pores where the major axis is up to 330 times bigger than the minor axis. This has implications both for the connectedness of pores and the resultant gas permeability and the effectiveness of gas desorption processes into the gas shale's pores. These high aspect ratio pores were oriented both in dip and azimuth in preferential directions making it likely that the shale gas itself has significant anisotropy both for permeability and in its mechanical properties. Permeabilities calculated from the XRMT distribution data matched very well with permeabilities obtained by scaling considerations and typical values for similar gas shales.