Figure Legends (also included below figures in this draft form)
Figure 1. Leaf photosynthetic rate (Amax-leaf ), leaf transpiration rate (Eleaf) and leaf water potential range of Acacia drepanolobium trees in Transition sites (that were invaded by P. megacephala in December 2017 after initial wet and dry season surveys) and in paired Control sites (that were protected byC. mimosae throughout the study). Effect tests are reported in each panel, with significant interactions terms indicating an effect of invasion distinct from interannual change. In wet seasons, Amax-leaf and Eleaf in Control trees were significantly lower and leaf water potential range was higher in 2018 than in 2017, yet these leaf physiological variables for Transition trees did not differ from 2017 (immediately before invasion) to 2018 (ca . 6 months after invasion) (A ,C ,E ). In dry seasons, trees in Transition sites had lowerAmax-leaf in 2018 (ca . 9 months after invasion) than in 2017 and this difference was significantly larger than the interannual decline for Control trees (B ). Also, in dry seasons, Eleaf increased between 2017 and 2018 for Control trees but was consistent for Transition trees (D ), and leaf water potential range was consistent for all trees (F ).
Figure 2. Differences in leaf- (Amax-leaf ; i.e., per-unit-leaf-area) and canopy-level (Amax-canopy ; i.e., canopy photosynthetic capacity) photosynthesis (means ­± SEM) ofP. megacephala - vs. C. mimosae- occupied Acacia drepanolobium adults in wet and dry seasons at long-term Invaded and Uninvaded sites. (A) Trees occupied by P. megacephala workers have significantly lower Amax-leaf andAmax-canopy than do uninvaded trees in wet conditions; (B) invaded trees have higherAmax-leaf during dry conditions, butAmax-canopy did not significantly differ for invaded and uninvaded trees during the dry season. Results of pairwise comparisons are indicated as significant (* P < 0.05, ** P < 0.001, *** P < 0.0001) or not significant (NS).
Figure 3. Differences in photosynthesis (means ± SEM) ofP. megacephala -occupied Acacia drepanolobium adults in a 2x2 full-factorial experiment (presence/absence of ants, large herbivores) conducted at 3 long-term Invaded sites in wet and dry seasons. Photosynthetic indices are estimated at the leaf- (Amax-leaf ) and canopy-level (Amax-canopy ). Results of effect tests are reported in panels. In wet seasons, (A ) vertebrate herbivory causes decline in Amax-leafand (C ) vertebrate herbivory and invasive ant presence both cause declines in Amax-canopy ; whileAmax-leaf andAmax-canopy are lower during dry seasons, vertebrate herbivores and invasive ant presence does not affect either response variable in dry conditions (B and D ).
Figure 4. Differences in canopy transpiration capacity (Ecanopy ) and leaf water potential range (means ± SEM) for Acacia drepanolobium adults in a 2x2 full-factorial experiment (presence/absence of ants, large herbivores) conducted at 3 long-term Invaded sites in wet and dry seasons. Significant results of effect tests are reported in subfigures (factors at left, P values at right; refer to main text for full output): in wet seasons, (A ) vertebrate herbivory and P. megacephalapresence both cause declines in Ecanopyand (C ) invasive ant presence causes a small but significant reduction in leaf water potential range. In the dry season, (B ) vertebrate herbivory and P. megacephala presence both cause declines in Ecanopy but do not affect leaf water potential range (D ). The diurnal change in leaf water potential (C and D ) is likely driven by a combination of increased transpirational water loss (A andB ) and hydraulic resistance.