Tumour hypoxia is known to be associated with aggressiveness and poor response to treatment, which has stimulated the development of several methods able to detect hypoxic tumours. To date, only one method, the oxygen microelectrode, has been used to provide pretreatment measures of tumour oxygenation that correlate with local control and disease-free survival. In an effort to validate new methods, comparisons have been made between the Eppendorf oxygen microelectrode, the comet assay, and hypoxia marker binding in tumours of patients undergoing curative treatment or palliative radiotherapy. These comparisons suggest that tumours with median oxygen tensions below 10 mmHg have relatively high hypoxic fractions as measured by the comet assay (> 0.20). The fraction of cells that binds pimonidazole, detected in cells obtained by fine-needle aspiration biopsy, correlates well with the hypoxic fraction measured using the comet assay. However, in general, hypoxic fractions measured by the comet assay and pimonidazole binding correlate only poorly with Eppendorf measurements performed for the same tumour. Factors that might be responsible for these differences, and problems associated with measuring the 'relevant' hypoxic population are discussed.