2.2. Sampling
We sampled isoprenoid emissions from 30 trees across two permanent plots for each forest type: upland forest, white-sand forest and ancient river terrace forest. Plots were distributed at least one kilometer from each other along the ATTO access road (Fig. 1). Tree species were previously identified (with individual vouchers collected) and confirmed taxonomically. Two species were chosen from a preliminary selection based on (i) their abundance in the PELD MAUA tree species inventory (Andreae et al. 2015), (ii) their contribution to a wider distribution in central Amazonia (ter Steege et al . 2013; Fauset et al. 2015), and (iii) their occurrence in at least two of the forest types of the study experimental site. Protium hebetatum (Burseraceae) occurred across the three forest types: upland forest, white-sand forest and ancient river terrace forest.Eschweilera grandiflora (Lecythidaceae) was present in two forest types: upland forest and ancient river terrace forest. In addition, we selected the second most dominant species in stem abundance, biomass and productivity in central Amazonia - Eschweilera coriacea(Lecythidaceae) (ter Steege et al . 2013; Fauset et al.2015), but this species was only present in the plots of white-sand forest. For each species we selected and sampled five trees with similar diameters in each forest type where the species occured. Their height and position in the canopy varied with species life history. Leaf traits and isoprenoid emission were repeatedly sampled in May 2019 (wet season) and September 2019 (dry season). In addition, a smaller subset of samples were previously taken in December 2018 (dry-to-wet transition season). See table 1 for more details on the weather and climatic conditions for each campaign.