Disturbances are an important process structuring forests worldwide and have long been considered as significant drivers of dynamics and diversity [1
]. In tropical forests, disturbances such as hurricanes and tree-falls from lightning or wind events create a mosaic of forest patches of different microhabitats at varying stages of succession [4
] superimposed upon background topographical and soil variation. If niche partitioning permits coexistence along the axis of tolerance to disturbance, different species should be selected by such disturbances, and we would expect to find predictable suites of species associated with different disturbance regimes [7
]. This is indeed the case. Because of a trade-off between growth and survival, fast-growing pioneer species occur predictably in tree-fall gaps. Conversely, slow-growing shade-tolerant species survive well in closed canopy forest [8
]. Furthermore, resprouting of trees is high in areas affected by hurricanes [11
]. These large-scale disturbances are obvious in the forest and are important mechanisms of species coexistence and community structure [4
]. However, the contribution of gaps and their associated pioneer flora to species coexistence is likely to be limited, because full tree-fall gaps are relatively infrequent in lowland rain forest [13
] and most pioneer species occur at low abundance [12
]. Are there other kinds of disturbance in tropical forests affecting far wider areas of forest that influence species composition and community dynamics?
Several authors have suggested that smaller disturbances such as branch-falls have a greater and more widespread effect on forest dynamics and composition than might be apparent at first glance [13
]. In the absence of large canopy-opening disturbances, mortality in sub-canopy layers can increase light availability, free up resources and lead to faster sapling growth [16
]. Furthermore, these sub-canopy gaps occupy large proportions of forest [13
]. Sub-canopy openings occur from a variety of processes, including the death of sub-canopy trees, large branch falls, or in-filling of canopy gaps by trees in middle forest strata [14
]. In addition to freeing resources, these events may cause disturbances to younger life stages, such as branch-falls affecting the survival and growth of seedlings and saplings [18
]. Falling branches are heavy and can kill or physically damage small plants [22
]. If these branch-falls were prevalent enough of a selective force, species intolerant of this damage could be selected against in areas of high branch-fall. Peters et al. [15
] extend this argument to apply to arborescent palm leaves, which are heavy enough to physically damage anything underneath (> 15 kg dry mass and several metres in length [23
]). Given the ubiquity of large palms in western Amazonian forests (densities often exceeding 60 palms > 10 cm dbh ha-1
]), these authors suggested that falling leaves could select for species adapted to withstand this disturbance impact, such as those with storage organs or cotyledons that could easily resprout following damage. Indeed, Peters et al. [15
] found differences in sapling (0.5-2.5 m in height) communities in areas close to large arborescent palms compared to those with no influence of palm leaf fall. Although theirs was only a small study over 2.25 ha, it would seem that falling palm leaves could have a large effect on the species composition of forest communities.
However, falling branches and leaf fronds are not the only widespread small-scale disturbances in forests. Large frugivores such as tapirs and peccaries are known to cause heavy localised disturbance, dramatically increase mortality of seedlings and saplings [30
], and consume and disperse seeds [34
]. Peccaries also consume palm fruit [33
] and in a repeat study of Peters et al. [15
], Beck et al. [36
] showed that sapling communities were impacted by the presence of palms in forest containing peccaries, but not in forest where they had been hunted out. Beck et al. [36
] attributed the relationship between palms and saplings to peccaries rooting around palms and not to falling palm fronds. However, both studies [15
] suffered from using only static data, examining distributions of palms relative to other species without the ability to test how these distributions develop through differential growth or mortality.
Whatever the mechanism (falling fronds or rooting peccaries), the disturbance around palms should cause differences in seedling and sapling density, growth, mortality, and/or re-sprouting close to palms compared to areas further away. If falling fronds are the cause, there should be similar effects in the neighbourhoods of large, fruiting palms as well as large, but non-fruiting palms. Alternatively, if the disturbance is caused by peccaries, we would not expect an effect by non-fruiting palms. In addition, although palms are a common food source, many other species of large-fruited tree are also consumed by peccaries [33
]. Therefore, we would also expect an effect on the seedling and sapling communities around other large fruit-bearing tree species similarly eaten by peccaries. In this latter case, the effect of these small-scale disturbances may be even more widespread than Peters et al. [15
We should also consider that palms may have positive effects on some species. Evidence gathered by Connell et al. [13
] showed that smaller sub-canopy disturbances led to faster growth of understorey saplings, presumably because of increased light penetration or freed-up resources from the death of a sub-canopy individual. These sub-canopy gaps (without complete opening of the forest canopy) were very common in the forest (covering up to about 48% of forest). Falling palm fronds likely create such sub-canopy gaps and seedlings that survive palm or peccary damage may benefit from having (i) fewer competitors (seen in the lower local densities of seedlings and saplings), (ii) more light (because palms have a less dense canopy than other trees and because the falling fronds knock other branches down) and, (iii) more soil resources (for the same reasons). Ultimately, this increased availability of resources adjacent to palms could lead to higher growth or survival of saplings of some species.
In this study we use spatially-explicit data from long-term monitoring of seedlings, saplings in the neighbourhoods of large arborescent Arecaceae (mostly Iriartea deltoidea
) and reproductive Myristicaceae trees in a large forest dynamics plot in western Amazonia to examine the potential structuring effect of falling leaves and branches on tropical forest communities. We aim to test (i) whether large palms and other fruiting trees appear to impact seedling and sapling dynamics and distributions and (ii) whether any difference is due to the trees themselves or an interaction with large frugivores. We extend previous studies of these phenomena in two key areas. First, we use dynamic data to understand the processes that lead to the patterns observed in static data in previous studies. Second, we consider that small-scale disturbances and/or interactions with frugivores must also apply to other fruiting trees. In this case, we examined seedlings and saplings in the neighbourhood of fruiting adults in a common dioecious family at our site, the Myristicaceae, fruits of which are also often consumed by peccaries [33
]. The Myristicaceae have been studied here for 10 years, and detailed information is available on the sex of each reproductive tree. Comparing the sapling communities around male (non-fruiting) and female (fruiting) trees allows a rigorous comparison of fruit-eating by peccaries, and a counterpoint to understanding the impacts of disturbances associated with arborescent palms.