Comment on “Was atmospheric CO2 capped at 1000 ppm over the past 300 million years?” by McElwain J. C. et al. [Palaeogeogr. Palaeoclimatol. Palaeoecol. 441 (2016) 653-658]

Franks et al. (2014) developed a model for estimating the concentration of atmospheric CO2 (ca) that can be applied to most stomata-bearing plant fossils. In a recent paper, McElwain et al. (2016a) proposed changes to two of the key inputs: mesophyll conductance to CO2 (gm) and CO2 assimilation rate at a known ca (A0). These proposed changes lead to increases in the model-estimates of ca. Here we show a lack of support for these proposed changes. First, the downward revision of gm is the result of a mathematical error by McElwain et al. (2016a) when describing the relationship between CO2 assimilation rate (An) and gm. Once corrected, values for gm are very similar to the values recommended by Franks et al. (2014). Second, the proposed ~ 2-fold upward revision of A0 is not supported by data from extant analogs or by hydraulic constraints from fossils. Moreover, the modelled estimates of An from Franks and Beerling (2009) are in fact most consistent with the modelled An in Franks et al. (2014) using their recommended A0 values, not those proposed by McElwain et al. (2016a). These results provide further support for the strategy of model implementation outlined in Franks et al. (2014).