Forest allocation of net primary productivity (NPP) to reproduction is poorly quantified globally, despite its critical role in forest regeneration and a well-supported trade-off with allocation to growth. Here, we present the first global synthesis of a biometric proxy for forest reproductive allocation (RA) across environmental and stand age gradients. We find that ecosystem-scale RA increases ~ 60% from boreal to tropical forests. Nonlinear relationships with climate metrics are important, explaining nearly 14% of variation, but are not the sole predictors of RA. The influence of forest age is comparable to climate in terms of effect size, and metrics of soil fertility also showed small but significant relationships with RA. These results provide strong evidence that RA at the ecosystem scale is mediated by climate, forest age, and soil conditions, and is not a globally fixed fraction of positive NPP, as assumed by most vegetation and ecosystem models.

Frugivory interactions are essential ecological processes for the regeneration of tropical forests, ensuring ecosystem resilience following natural disturbances. However, little is known about how frugivory interactions are shaped by anthropogenic disturbances, especially in Amazonia - one of the richest ecosystems of the planet. Here, we investigate how selective logging and forest fires impact both arboreal and terrestrial frugivory interactions in Amazonian forests. We focus on four forest disturbance classes: Undisturbed, Logged, Logged-and-3y-burned (burned three years before our sampling), and Logged-and-17y-burned (burned 17 years before sampling). In total, we sampled 4,670 frugivory interactions at the community level, in a sampling effort of 31,484 hours. Undisturbed forests sustained a significantly higher number of species and interactions only when compared to logged-and-17y-burned forests. Selective logging and forest fires did not significantly alter the structural properties of the frugivory networks, which were highly modular, moderately specialised, poorly connected and non-nested. Regarding community composition, we detected high β-diversity of plant and frugivore species, and frugivory interactions, both among and within forest classes, mainly driven by spatial turnover. Logged-and-17y-burned forests hosted the most unique interaction composition compared to undisturbed forests. Our study provides empirical evidence that disturbance-mediated drift in frugivore interactions may occur many years after the initial disturbance events.