We investigated the utility of combining micro-scale computed tomography (micro-CT), image analysis and geostatistics to quantify pore geometry at spatial scales ranging from 4.4 [mu m to 2 mm. To facilitate this, we investigated soil taken from an old permanent upland pasture known to support very high levels of microbial diversity and that had not been cultivated for many centuries. Aggregates (< 3 mm in diameter) from three treatments (control, sewage sludge amended and biocide treated) derived from the site were imaged using synchrotron-based computed microtomography. Image analysis was used to determine aggregate porosity and pore shape parameters, and semivariance analysis was used to measure the spatial correlation of pore space within the three land treatments. For all plots high porosities were observed (c. 30%) at scales below 3 mm. High variations of porosity were also observed, ranging from 22% to 47%, but no significant differences among treatments were found. No significant differences among treatments were found in the distribution of pores within aggregates, as revealed by semivariance analysis, or in pore shape parameters. No treatment effects were observed. However, the work presented here shows that the combination of approaches adopted has great potential for quantifying the soil microbial physical habitat. Future work, should investigate the use of these novel techniques in more controlled soil ecosystems to provide an exciting new way of understanding the soil-microbe interactions at appropriate scales. (c) 2005 Elsevier B.V. All rights reserved