Results
Short-term effects of ant occupants and herbivores on A. drepanolobium physiology- In our BACI experiment during the wet season, on Control trees we found large declines between 2017 and 2018 in leaf-level photosynthesis (24%) and transpiration (20%), which significantly differed from minor declines for photosynthesis (4%) and transpiration (7%) between wet seasons for Transition trees (site x year terms, photosynthesis: Fig. 1A; transpiration: Fig. 1C). Although leaf water potential range became more negative by 57% for Control trees and less negative by 3% for Transition trees from the 2017 to 2018 wet season, the site x year term in our GLMM was only marginally significant (Fig. 1E).
In our BACI experiment during the dry season, we observed a decline between 2017 and 2018 dry seasons in leaf-level photosynthesis that was significantly larger for Transition trees (54%) than for Control trees (25%) (site x year interaction term, Fig. 1B). Changes in leaf-level transpiration between dry seasons significantly differed for Transition trees (1% decline) and Control trees (35% increase) site x year interaction term, Fig. 1D). Despite these different trends in photosynthesis and transpiration, we found no significant variation in leaf water potential range due to any GLMM factors during the dry season (site x year interaction term, Fig. 1F).
Longer-term effects of ant occupants and herbivores on Acacia drepanolobium physiology - During wet seasons, we found significant differences in leaf- and canopy-level photosynthesis and transpiration as a function of invasion status, vertebrate herbivore exclusion from Invaded trees, and P. megacephala exclusion (full ANOVA results in Table S5). Trees in stands which had been invaded for ca . 5 years had 16% lower photosynthesis (Fig. 2A) and 17% lower transpiration (Fig. S1A) at the leaf-level than trees in uninvaded stands. Scaled up to the canopy level, Invaded trees had 68% lower photosynthesis (Fig. 2A), and 54% lower transpiration (Fig. S1A) than Uninvaded trees. The exclusion of vertebrate herbivores from Invaded trees was associated with a 19% increase in photosynthesis (Fig. 3A) and a non-significant 13% increase in transpiration (Fig. S2A) at the leaf-level, and was also associated with substantially higher canopy-level photosynthesis (365% higher; Fig. 3C) and transpiration (305% higher; Fig. 4A). Exclusion of P. megacephala ants did not significantly affect leaf-level photosynthesis (Fig. 3A) or transpiration (Fig. S2A), but their exclusion resulted in higher canopy-level photosynthesis (66% higher, Fig. 3C) and transpiration (52% higher; Fig. 4A). Despite these large differences between treatments in leaf and canopy transpiration, we only observed a significant but minor increase in leaf water potential range due toP. megacephala exclusion (14% less negative; Fig. 4C), and no significant differences due to invasion status or vertebrate herbivore exclusion from Invaded trees. Exclusion of C. mimosae and vertebrate herbivores from Uninvaded trees did not significantly impact leaf- or canopy-level photosynthesis or transpiration or leaf water potential range (Note S4).
During dry seasons, at the leaf-level we observed small increases in gas exchange and water potential range resulting from invasion status, herbivore exclusion from Invaded trees, and P. megacephalaexclusion, but no significant differences at the canopy level (Table S5). Invaded trees had significantly higher photosynthesis (13%; Fig. 2B) and transpiration (16%; Fig. S1B) at the leaf level, but we did not find significant differences in in scaled-up canopy-level estimates of photosynthesis (Fig. 2B) or transpiration (Fig. S1B). Invaded trees also had slightly less negative leaf water potential range (7% less negative; Fig. S1D) during the dry season. The exclusion of P. megacephala from Invaded trees resulted in a minor but significant decline (17%) in dry season leaf-level transpiration (Fig. S2B), butP. megacephala exclusion did not significantly affect leaf- or canopy-level photosynthesis (Fig. 3B & D), canopy-level transpiration (Fig. 4B), or leaf water potential range (Fig. 4D). Similarly, exclusion of vertebrate herbivores was associated with a minor but significant decline (13%) in leaf-level transpiration, but no significant change in leaf- or canopy-level photosynthesis (Fig 3B & D), canopy-level transpiration (Fig. 4B), or leaf water potential range (Fig. 4D). In contrast to trees from longer-term Invaded sites, Uninvaded trees did not exhibit different gas exchange or leaf water potential traits due to ant (C. mimosae ) exclusion, herbivore exclusion or their interaction (Note S4).