Abstract
1 Grasses using the C4 photosynthetic pathway dominate today's savanna ecosystems and account for ~20% of terrestrial carbon fixation. However, this dominant status was reached only recently, during a period of C4 grassland expansion in the Late Miocene and Early Pliocene (4–8 Myr ago). Declining atmospheric CO2 has long been considered the key driver of this event, but new geological evidence casts doubt on the idea, forcing a reconsideration of the environmental cues for C4 plant success. 2 Here, I evaluate the current hypotheses and debate in this field, beginning with a discussion of the role of CO2 in the evolutionary origins, rather than expansion, of C4 grasses. Atmospheric CO2 starvation is a plausible selection agent for the C4 pathway, but a time gap of around 10 Myr remains between major decreases in CO2 during the Oligocene, and the earliest current evidence of C4 plants. 3 An emerging ecological perspective explains the Miocene expansion of C4 grasslands via changes in climatic seasonality and the occurrence of fire. However, the climatic drivers of this event are debated and may vary among geographical regions. 4 Uncertainty in these areas could be reduced significantly by new directions in ecological research, especially the discovery that grass species richness along rainfall gradients shows contrasting patterns in different C4 clades. By re-evaluating a published data set, I show that increasing seasonality of rainfall is linked to changes in the relative abundance of the major C4 grass clades Paniceae and Andropogoneae. I propose that the explicit inclusion of these ecological patterns would significantly strengthen climate change hypotheses of Miocene C4 grassland expansion. Critically, they allow a new series of testable predictions to be made about the fossil record. 5 Synthesis. This paper offers a novel framework for integrating modern ecological patterns into theories about the geological history of C4 plants.