This paper describes a method for estimating lung density, air volume and changes in fluid content from a non-invasive measurement of the electrical resistivity of the lungs. Resistivity in Ω m was found by fitting measured electrical impedance tomography (EIT) data to a finite difference model of the thorax. Lung density was determined by comparing the resistivity of the lungs, measured at a relatively high frequency, with values predicted from a published model of lung structure. Lung air volume can then be calculated if total lung weight is also known. Temporal changes in lung fluid content will produce proportional changes in lung density. The method was implemented on EIT data, collected using eight electrodes placed in a single plane around the thorax, from 46 adult male subjects and 36 adult female subjects. Mean lung densities (±SD) of 246 ± 67 and 239 ± 64 kg m(-3), respectively, were obtained. In seven adult male subjects estimates of 1.68 ± 0.30, 3.42 ± 0.49 and 4.40 ± 0.53 l in residual volume, functional residual capacity and vital capacity, respectively, were obtained. Sources of error are discussed. It is concluded that absolute differences in lung density of about 30% and changes over time of less than 30% should be detected using the current technology in normal subjects. These changes would result from approximately 300 ml increase in lung fluid. The method proposed could be used for non-invasive monitoring of total lung air and fluid content in normal subjects but needs to be assessed in patients with lung disease.