ABSTRACT We have developed a mathematical model of inert gas uptake and distribution based on the physiological characteristics of the human body. Using the inert gas levels calculated by the model we have calculated total gas content of each tissue type taking account of metabolic gas exchange. When the total gas pressure is greater than environmental pressure a bubble forms; the model calculates the change of volume of this separated gas phase. These calculations are repeated at frequent intervals throughout the decompression procedure. Thus we have predictions of separated gas phase volume for any tissue of the body, at any point of the path from arterial blood to venous blood and tissue oxygen levels. We have run this model to predict gas phase volume for a number of decompressions carried out with pigs and with humans in which we have used ultrasonic bubble detection. The agreement between predicted separated gas volume and bubble count is very good. The model enables us to examine the effect on bubble formation of changing major characteristics of the decompression profile. Thus we can begin to plan profiles with minimum bubble formation and compare predictions with bubble counts in pigs. Ultimately we should be able to design bubble-free decompressions,---but perhaps we are being too optimistic. INTRODUCTION The increased use of professional divers in commercial activities in the last two decades has social and medico-legal implications for the future. For the first time in history there will be a population of aging men who have spent a significant proportion of their working lives in an hyperbaric environment and have been subjected to the consequential decompression procedures. We do not yet know much about the long-term health consequences of diving but evidence is beginning to emerge that diving does create apparently irreversible changes and deterioration of normal function. As increasing numbers of professional divers enter the seventh and eighth decade of life there will almost certainly be a new population of men suffering long term industrial health changes. The diving industry has made remarkable advances in the reduction of decompression stickiness (DCS) so that now figures as low as 1 incident in 2000 dives are possible. However the evidence is clear that bubbles are a possible cause of some long-term damage and that bubbles form during many decompressions which are apparently clear of DCS. It is now 40 years since the concept of "silent" bubbles was first presented (Fulton 1951) and subsequent work using modern methods of detecting bubbles have confirmed the theory. Spencer and Clark (1972) reported the detection of pulmonary artery bubbles by precordial monitoring by Doppler in subjects free of DCS symptoms. More recently in a series of 74 divers studied during alr or gas switch dives from 40 metres 25 had bubbles in the pulmonary artery detected by Doppler though none had DCS.