A challenge for the development of terrestrial biosphere models (TBMs) and associated land surface components of Earth system models (ESMs) is improving representation of carbon (C) exchange between terrestrial plants and the atmosphere, and incorporating biological variation arising from diversity in plant functional types (PFTs) and climate (Sitch et al., 2008; Booth et al., 2012; Prentice & Cowling, 2013; Fisher et al., 2014). Accounting for patterns in leaf respiratory CO2 release in darkness (Rdark) in TBMs and ESMs is crucial (King et al., 2006; Huntingford et al., 2013; Wythers et al., 2013), as plant respiration - roughly half of which comes from leaves (Atkin et al., 2007) - releases c. 60 Pg Cyr-1 (Prentice et al., 2001; Canadell et al., 2007; IPCC, 2013). Fractional changes in leaf Rdark as a consequence of climate change can, therefore, have large impacts on simulated net C exchange and C storage for individual ecosystems (Piao et al., 2010) and, by influencing the CO2 concentration of the atmosphere, potentially feedback so as to alter the extent of future global warming (Cox et al., 2000; Huntingford et al., 2013). There is growing acceptance, however, that leaf Rdark is not adequately represented in TBMs and ESMs (Huntingford et al., 2013; Smith & Dukes, 2013), resulting in substantial uncertainty in future climate predictions (Leuzinger & Thomas, 2011); consequently, there is a need to improve representation of leaf Rdark in predictions of future vegetation-climate interactions for a range of possible fossil fuel-burning scenarios (Atkin et al., 2014). Achieving this requires an analysis of variation in leaf Rdark along global climate gradients and among taxa within ecosystems; and establishing whether relationships between leaf Rdark and associated leaf traits vary predictably among environments and PFTs (Wright et al., 2004, 2006; Reich et al., 2006; Atkin et al., 2008). PFTs enable a balance to be struck between the computational requirements of TBMs to minimize the number of plant groups and availability of sufficient data to fully characterize functional types, and the reality that plant species differ widely in trait values. Most TBMs contain at least five PFTs, with species being organized on the basis of canopy characteristics such as leaf size and life span, physiology, leaf mass-to-area ratio, canopy height and phenology (Fisher et al., 2014). Although classifications that are directly trait-based are emerging (Kattge et al., 2011), PFT classifications are still widely used in TBMs and land surface components of ESMs. As such, discerning the role of PFTs in modulating relationships between leaf Rdark and associated leaf traits will provide critical insights.