E. Pennisi (“Forest giants are the trees most at risk,” News, 6 September, p. 962) interprets presentations of three studies as suggesting that “for trees, size is not strength, and forest giants are disproportionately vulnerable.” However, this conclusion is not well supported. The observation that lightning is a major cause of large-tree mortality on Barro Colorado Island (BCI) is best interpreted in context: when all sources of mortality are considered, small, not large BCI trees are most vulnerable to mortality (1). We have no a priori reason to assume that relative size vulnerabilities must reverse if mortality rates increase in the future. Rather than reflecting universally high drought vulnerability of large trees, the remotely-sensed observation of increasing mortality with tree height in California’s Sierra Nevada (2) likely reflects changing species dominance with height. During the drought, sizes of trees suffering greatest mortality varied widely among species, a consequence of idiosyncratic host-tree selection by different bark beetle taxa (3). Pines were the only common species with mortality that increased with size, and pines also increased in relative dominance with canopy height (4, 5). More broadly, increasing tree mortality across western USA has affected trees of all sizes (6), and a recent multi-continent compilation showed no consistent size vulnerability to drought (3). Because large trees typically are both the most highly valued and the most visible in aerial mortality surveys, the literature used to define the most damaging invasive forest pests (7) may be biased toward pests that kill large trees. Additionally, greater proportional mortality increase of large trees in the presence of invasive pests does not necessarily reflect greater absolute mortality rates relative to small trees (3, 8). Finally, Pennisi overlooked contrasting examples. For example, fire is increasing in many forests globally, and often disproportionately kills small trees (9). While large trees are certainly vulnerable to many ongoing environmental changes (10), they are not consistently the most vulnerable.