Each of >20 independent evolutions of C4 photosynthesis in grasses required reorganization of the Calvin-Benson-cycle (CB-cycle)within the leaf, along with coordination of C4 -cycle enzymes with the CB-cycle to maximize CO2 assimilation. Considering the vast amount of time over which C4 evolved, we hypothesized (i) trait divergences exist within and across lineages with both C4 and closely related C3 grasses, (ii) trends in traits after C4 evolution yield the optimization of C4 through time, and (iii) the presence/absence of trends in coordination between the CB-cycle and C4 -cycle provides information on the strength of selection. To address these hypotheses, we used a combination of optimality modelling, physiological measurements and phylogenetic-comparative-analysis. Photosynthesis was optimized after the evolution of C4 causing diversification in maximal assimilation, electron transport, Rubisco carboxylation, phosphoenolpyruvate carboxylase and chlorophyll within C4 lineages. Both theory and measurements indicated a higher light-reaction to CB-cycle ratio (Jatpmax /Vcmax ) in C4 than C3 . There were no evolutionary trends with photosynthetic coordination between the CB-cycle, light reactions and the C4 -cycle, suggesting strong initial selection for coordination. The coordination of CB-C4 -cycles (Vpmax /Vcmax ) was optimal for CO2 of 200 ppm, not to current conditions. Our model indicated that a higher than optimal Vpmax /Vcmax affects assimilation minimally, thus lessening recent selection to decrease Vpmax /Vcmax .