Aim Tropical elevation gradients are natural laboratories to assess how changing climate can influence tropical forests. However, there is a need for theory and integrated data collection to scale from traits to ecosystems. We assess predictions of a novel trait‐based scaling theory, including whether observed shifts in forest traits across a broad tropical temperature gradient are consistent with local phenotypic optima and adaptive compensation for temperature. Location An elevation gradient spanning 3,300 m and consisting of thousands of tropical tree trait measures taken from 16 1‐ha tropical forest plots in southern Perú, where gross and net primary productivity (GPP and NPP) were measured. Time period April to November 2013. Major taxa studied Plants; tropical trees. Methods We developed theory to scale from traits to communities and ecosystems and tested several predictions. We assessed the covariation between climate, traits, biomass and GPP and NPP. We measured multiple traits linked to variation in tree growth and assessed their frequency distributions within and across the elevation gradient. We paired these trait measures across individuals within 16 forests with simultaneous measures of ecosystem net and gross primary productivity. Results Consistent with theory, variation in forest NPP and GPP primarily scaled with forest biomass, but the secondary effect of temperature on productivity was much less than expected. This weak temperature dependence appears to reflect directional shifts in several mean community traits that underlie tree growth with decreases in site temperature. Main conclusions The observed shift in traits of trees that dominate in more cold environments is consistent with an ‘adaptive/acclimatory’ compensation for the kinetic effects of temperature on leaf photosynthesis and tree growth. Forest trait distributions across the gradient showed overly peaked and skewed distributions, consistent with the importance of local filtering of optimal growth traits and recent shifts in species composition and dominance attributable to warming from climate change. Trait‐based scaling theory provides a basis to predict how shifts in climate have and will influence the trait composition and ecosystem functioning of tropical forests.

The network of minor veins of angiosperm leaves may include loops (reticulation). Variation in network architecture has been hypothesized to have hydraulic and also structural and defensive functions. We measured venation network trait space in eight dimensions for 136 biomass‐dominant angiosperm tree species along a 3,300 m elevation gradient in southeastern Peru. We then examined the relative importance of multiple ecological and evolutionary predictors of reticulation. Variation in minor venation network reticulation was constrained to three axes. These axes described reconnecting vs. branching veins, elongated vs. compact areoles compact vs. elongated and low vs. high‐density veins. Variation in the first two axes was predicted by traits related to mechanical strength and secondary compounds, and in the third axis by site temperature. Synthesis. Defensive and structural factors primarily explain variation in multiple axes of reticulation, with a smaller role for climate‐linked factors. These results suggest that venation network reticulation may be determined more by species interactions than by hydraulic functions.

This study used high-precision radiocarbon bomb-pulse dating of selected wood rings to provide an independent validation of the tree growth periodicity of Pseudolmedia rigida (Klotzsch & H. Karst) Cuatrec. from the Moraceae family, collected in the Madidi National Park in Bolivia. 14C content was measured by accelerator mass spectrometry (AMS) in 10 samples from a single tree covering over 70 yr from 1939 to 2011. These preliminary calendar dates were determined by dendrochronological techniques and were also used to select the samples for 14C AMS. In order to validate these preliminary dates using the established Southern Hemisphere (SH) 14C atmospheric concentration data set, the targeted rings were selected to be formed during periods before and after the 14C bomb spike nuclear tests (i.e. 1950s–1960s). The excellent agreement of the dendrochronological dates and the 14C signatures in tree rings associated with the same dates provided by the bomb-pulse 14C atmospheric values for the SH (SHCal zone 1–2) confirms the annual periodicity of the observed growth layers, and thus the high potential of this species for tree-ring analysis. The lack of discrepancies between both data sets also suggests that there are no significant latitudinal differences between the 14C SHCal zone 1–2 curve and the 14C values obtained from the selected tree rings at this geographic location (14°33′S, 68°49′W) in South America. The annual resolution of P. rigida tree rings opens the possibility of broader applications of dendrochronological analysis for ecological and paleoclimatic studies in the Bolivian tropics, as well as the possibility of using wood samples from some tree species from this region to improve the quality of the bomb-pulse 14C SHCal curve at this latitude.

Abstract—The taxonomically difficult genus Polylepis (Rosaceae) contains several problematic species complexes, including one around P. sericea. We here reevaluate species boundaries in this complex based on morphological characters and climatic niches. We recognize five species within the P. sericea complex, based primarily on leaf characters: P. albicans (northwestern Peru), P. argentea (central Peru) described here, P. canoi (southeastern Peru to central Bolivia), P. ochreata (Ecuador to southernmost Colombia), and P. sericea (Venezuela to central Colombia). We provide full synonymy, diagnostic notes, lists of selected specimens examined, and a key to the species of the complex.