Ungulate diets reveal patterns of grassland evolution in North America

Equids from the late Miocene (Hemphillian North American Land Mammal Age) of Texas evolved C4 dominated diets earlier than their Nebraska counterparts, leading researchers to hypothesize a northward expansion of C4 grasses between 8 and 6 Ma. However, competing hypotheses from multiple paleoecological proxies support one of three hypotheses: that northward C4 expansion occurred at the expense of C3 woodlands, C3 grasslands, or was not occurring during the Hemphillian. We test these alternative hypotheses by comparing the ecosystems of Hemphillian Texas and Nebraska (6.3–7.5 Ma) using hoofed mammal dietary reconstruction (hypsodonty and digestive strategy) and by incorporating published tooth enamel isotopes, paleosol isotopes, and phytolith assemblage data. We also use hypsodonty, mesowear, and microwear in a detailed comparison of two localities that are roughly contemporary (Cambridge, Nebraska at 7 Ma and Coffee Ranch, Texas at 6.6 Ma) to rule out time averaging as an explanatory variable. Through analogy with modern African communities, we reconstruct Texas and Nebraska ecosystems as bushland and woodland, respectively. Using the proportions of hypsodont taxa we estimate mean annual precipitation values of 1217 mm/year for Nebraska and 1368 mm/year for Texas. Using rarefaction we also do not find differences in the richness of brachydont, hypsodont, and hindgut fermenting taxa. We therefore conclude that both ecosystems were largely similar in hoofed mammal faunal structure during the late Miocene. When compared to published paleosol and phytolith data, which show no differences in C3/C4 biomass among latitudes, our results allow us to reject hypotheses of northward C4 expansion during the late Miocene. We therefore suggest that the equid enamel isotopes may have been biased by two potential factors: the primary reliance on a single locality of appropriate age and possible behavioural lag in C4 feeding among Nebraska equids. These results suggest that the formation of a modern C3/C4 latitudinal gradient occurred later than is previously suggested by equid enamel carbon isotopes.

► We tested three hypotheses about the late Miocene expansion of C4 plants.
► We use hoofed mammal diets to reconstruct ecosystems in Texas and Nebraska.
► We compare to published paleosol and tooth enamel isotopes as well as phytoliths.
► We find no differences between Texas and Nebraska during the late Miocene.
► Formation of a North American C3/C4 latitudinal gradient occurred more recently.