Our project consisted of a standardized experiment which we replicated at 6 sites, in Germany, Japan, China, Papua New Guinea, and two study sites in Australia. The work at each of the study sites consisted of the manipulative experiment during which we prevent access of predators to saplings in forest understory and to branches in canopy. We worked with 42 focal plant species, preselected based on their traits, phylogeny and abundance. We marked 280 saplings in understory (35 individual per each of the 8 plant species) and assigned them to the 7 treatments (i.e. 5 individuals per treatment – CN1, BIRD, BAT; CN2, ALL, VER, ANT). Similarly, we select 96 large branches in canopy and assigned them to 4 treatments (CN2, ALL, VER, ANT). Exclosures of vertebrates (VER), birds (BIRD) and bats (BAT) were in form of large cage exclosures, which were placed at sites permanently or opened daily in the morning and in the afternoon to prevent access only of the respective predator. Exclosure of ants (ANT) were in form of a sticky insect barrier applied to the branches or trunk. ALL treatment then combined both treatments – cages and sticky barrier. Once we set the treatments, we marked permanently and count all leaves within the enclosed saplings or branches, photographed their herbivory damage, and collect all arthropods to start the experiment. After that, we managed the smaller bird & bat experiment (BIR, BAT, CN1 treatments) for a month, and surveyed the effect of the treatment after that. It means, we sampled newly established arthropod communities and re-scanned all leaves, keeping the leaves for chemical analyses. In the meantime, we kept surveying communities of predators we manipulated. Birds and bats are surveyed by point-count method and ants were collected from baits. After 5 months, we terminated the experiment the same way, as the short one-month experiment. About 6 months later, we replicate the whole experiment once again at the same site. Thus we obtained 2 full replicates for each of the 6 study sites. At each of the study site, we also surveyed predation rata by exposing dummy caterpillars. This work is a supportive to the main experiment. We found very interesting results contradicting all previous patterns found in forest understories, and we explained them by the abundance of predators, which is rarely measured in similar studies. Specifically, at different latitudes, we recorded between 95 and 845 insectivorous birds per 4 hours long survey. Considering the biomass of birds and their feeding specialization we obtained critically important data which we will be very important for global understanding of the importance of predators. We found also striking patterns in the abundance and richness of ants, for which we confirmed various feeding strategies by stable isotopes, supporting thus their weak effect on lower trophic levels.
