The metabolic theory of ecology (MTE) predicts the ubiquity of the of 3/4 scaling exponent relating metabolic rate (MR) to body mass, as well as cell-size invariance coupled with body-size dependence of cellular MR in quickly dividing cells. An alternative prediction is that MR scales interspecifically with a coefficient that is between 2/3 and 1, depending on the cell size and cell MR, which is mostly driven by the cell surface-to-volume ratio. We tested (1) the contribution of cell size to interspecific differences in MR and (2) whether the cell size-MR relationship is mediated by genome size (GS), which usually correlates positively with cell size. We tested (1) and (2) using erythrocyte area as a proxy for cell size in 14 eyelid geckos, which belong to a monophyletic group exhibiting large body-size variation. The scaling of standard MR (SMR) was significantly lower than 3/4, whereas mass-specific SMR correlated with erythrocyte area in both phylogenetically adjusted and conventional analyses. This points to cell-size variation as the factor governing metabolic rate scaling, which questions predictions of the MTE. However, the nonsignificance of the correlation between mass-specific SMR and GS undermines the strength of the relation between GS and cell size, at least in these species.