Selective history is thought to constrain the extent and direction of future adaptation by limiting access to genotypes that are advantageous in a novel environment. Populations of Chlamydomonas previously selected at high CO2 were either backselected at ambient levels of CO2, or selected at levels of CO2 that last occurred during glaciation in the Pleistocene. There was no effect of selective history on adaptation to either level of CO2, and the high CO2 phenotypes were evolutionarily reversible such that fitness in ambient CO2 returned to values seen in controls. CO2 uptake affinity improved relative to the ancestor in both ambient and glacial CO2, although wild-type regulation of CO2 uptake, which deteriorated during previous selection at high CO2, was not restored by selection at lower levels of CO2. Trade-offs in both CO2 uptake affinity and growth were seen after selection at any given level of CO2. Adaptation to ambient and glacial-era levels of CO2 produced a range of phenotypes, suggesting that chance rather than selective history contributes to the divergence of replicate populations in this system.