Dynamic dough density measurements were applied to monitor the rate of production of carbon dioxide gas during proving and its partitioning between the liquid phase and the bubbles in bread dough. The effects of yeast concentration, temperature, mixing speed, headspace pressure and sugar level were investigated. Increasing yeast level increased the rate of carbon dioxide production, as did increasing temperature up to 40°C, beyond which the production rate decreased. Mixing at low pressures resulted in fewer bubbles in the dough and a smaller interfacial area for mass transfer into bubbles; consequently the carbon dioxide concentration in the dough increased rapidly initially. This was followed by rapid growth of the few bubbles present, as a result of the high carbon dioxide concentration in the liquid phase, and sudden and rapid decrease of the dough density. Mixing at higher speeds increased the air content, but gave slightly slower rates of growth of the dough piece. The apparent drop in the rate of carbon dioxide production as proving proceeded was not caused by depletion of sugars, but rather by loss of gas from the dough piece. The results were compared with simulations of proving of bread doughs, giving good qualitative agreement; however, loss of gas from the dough pieces caused the rate of carbon dioxide production to be underestimated and resulted in deviations from the simulations.