We report an experiment designed to identify the effect of elevated CO2 on species of phytoplankton in a simple laboratory system. Major taxa of phytoplankton differ in their ability to take up CO2, which might lead to predictable changes in the growth rate of species and thereby shifts in the composition of phytoplankton communities in response to rising CO2. Six species of phytoplankton belonging to three major taxa (cyanobacteria, diatoms and chlorophytes) were cultured in atmospheres whose CO2 concentration was gradually increased from ambient levels to 1000 parts per million over about 100 generations and then maintained for a further 200 generations at elevated CO2. The experimental design allowed us to trace a predictive sequence, from physiological features to the growth response of species to elevated CO2 in pure culture, from the growth response in pure culture to competitive ability in pairwise mixtures and from pairwise competitive ability to shifts in the relative abundance of species in the full community of all six species. CO2 altered the dynamics of growth in a fashion consistent with known differences among major taxa in their ability to take up and use CO2. This pure-culture response was partly successful in predicting the outcome of competition in pairwise mixtures, especially the enhanced competitive ability of chlorophytes relative to cyanobacteria, although generally statistical support was weak. The competitive response in pairwise mixtures was a good predictor of changes in competitive ability in the full community. Hence, there is a potential for forging a logical chain of inferences for predicting how phytoplankton communities will respond to elevated CO2. Clearly further extensive experiments will be required to validate this approach in the greater complexity found in diverse communities and environments of natural systems.