journal article
LitStream Collection
Igapó seed patches: a potentially key resource for terrestrial vertebrates in a seasonally flooded forest of central Amazonia
Antunes, Ana Carolina;Baccaro, Fabrício;Caetano Andrade, Victor Lery;Ramos, José Ferreira;Da Silva Moreira, Roberto;Barnett, Adrian A;
doi: 10.1093/biolinnean/blz101pmid: N/A
Abstract In Amazonian igapó forests (seasonally flooded forests on blackwater river margins), the end of the annual flood pulse results in the formation of extensive mat-like seed patches. The seeds in these patches then germinate, forming a dense, highly heterogeneous, assemblage. Animal–plant interactions in these areas, as well as the influence that the patches have on the occurrence of herbivorous vertebrates, remain almost completely unstudied. Using camera traps in areas with and without seed/seedling patches, we tested the relationship between these seed accumulation sites and the presence of bird and mammal species. At the micro-scale (between treatments), vertebrate occurrence was not related to patch presence. At the larger scale (local), distance from adjacent upland (terra firme) forest and seed patch size were correlated with vertebrate distribution. The widespread occurrence of terrestrially active birds and mammals throughout igapó forests, not just where food resource densities were high, seems to be a compromise strategy between exploring the area to select the most favourable food items, and minimizing the risk of being predated when spending extended time foraging at the concentrated food sources represented by the seed patches. Our results underline the potential importance of igapó forests as a key habitat for a variety of terrestrial terra firme taxa, as well as emphasize the dynamic nature of this forest type, and should encourage further studies of this habitat and resource availability system. bird, Igapó forest, interaction, mammal, neotropical forest INTRODUCTION The ideal free distribution (IFD) concept developed by Fretwell & Lucas (1970) predicts that there will be more consumers in higher-quality patches than in lower-quality patches. Given that possible knowledge of resource availability is related to resource patch size, the predictions of the IFD model are, consequently, scale-dependent (Ives et al., 1993; Cosner, 2005). For example, at large spatial scales, competitors can be positively associated because a common resource is aggregated (Giller & Doube, 1994; Inouye, 2005; Schellhorn & Andow, 2005), whereas, at smaller scales, they may use behaviour or microhabitat selection to avoid direct competition (Byers, 1989; Ziv et al., 1993; Albrecht & Gotelli, 2001). Where a resource is not predictable (predator–prey interactions), models also consider that the movement of animals and the presence of refuges are important variables in determining individual distributions (Turchin, 1998; Mason & Fortin, 2017). For predators, prey movement is the main operational factor in any foraging–predation trade-off (Mitchell & Lima, 2002). Thus, predictions regarding predators and prey distributions depend on the level of predation risk, as well as the volume and type of resources available (Hammond et al., 2007). In the Amazon basin, blackwater floodplains along the Rio Negro and its tributaries cover an area of ~119 000 km2 (Melack & Hess, 2010). Annual floods with a very low input of dissolved and suspended nutrients are responsible for the low level of nutrients in these habitats (hereafter igapós) (Junk et al., 2015). This annual flooding may last for more than 200 days, and strongly influences the composition and zonation of plant assemblages along the flood-level gradient (Ferreira et al., 2010; Barnett et al., 2015; Junk et al., 2015). Trees in these forests have developed a suite of anatomical, physiological, morphological and phenological adaptations that allow their survival in this stressful environment, including high seed mass to compensate for the lack of nutrients, seedlings with rapid shoot elongation (which increases the chances of non-submergence for tall seedlings), high frequency of xeromorphic leaves, and seed dispersion modes highly dependent on the annual flood pulse (Parolin et al., 2004; Hawes & Peres, 2014; Parolin et al., 2016). Water- and fish-dispersed species show morphological adaptations that facilitate dispersion via increased buoyancy, such as spongy tissues or large air-filled spaces (Parolin et al, 2010). In igapó forests, ichthyochory and hydrochory are common modes of dispersion (Kubitzki & Ziburski, 1994; Correa et al., 2007). However, the flood pulse also influences resource availability for fruit- and seed-eating vertebrate species. This availability is markedly seasonal, with two peaks occurring annually. The first occurs at highest water levels, when most igapó trees synchronously produce fruits (Parolin et al., 2013), and the second occurs at the lowest water level, when many seeds that had been floating in the water lie exposed on the forest floor (Haugaasen & Peres, 2005, 2007). As the water level drops, floating seeds accumulate around obstacles such as floating logs and fallen trees, or in areas with low water current velocity (Barnett et al., 2012), often forming floating rafts of seeds several to many square metres in extent (Supporting Information, Fig. S1A). When flood waters recede, these accumulations are deposited intact on the forest floor, and many of the seeds within them subsequently germinate (Barnett et al., 2012). The two periods of resource availability in igapó forests coincide with periods of reduced fruit availability in the adjacent upland (terra firme) forest (Haugaasen & Peres, 2005). As a result of this fruiting asynchrony, arboreal and volant animals (during the high-water season), and terrestrial animals (during the unflooded period), move annually into igapó forests to take advantage of these pulses of food resource availability (Bodmer, 1990; Renton, 2002; Haugaasen & Peres, 2007). To date, little information is available on the distribution of terrestrial vertebrates in Amazonian flooded forests in general (Alvarenga et al., 2018; Costa et al., 2018), and such data are even rarer for igapó forests. Seed patches size may vary from 0.4 to 3.8 m2, with densities of up to 594 seedlings per m2, which contrasts strongly with the rest of the understorey vegetation, where seeds and seedlings are generally sparse (Supporting Information, Fig. S1B) (Barnett et al., 2012). Although the process of seed patch formation remains largely unstudied, the patches are potentially important in explaining patterns of seasonal occupation of unflooded igapó forest by mammal and bird species. Seed patch formation, distribution and composition are hard to predict, but all appear to be related to the phenology of igapó plants, as well as their form of dispersion and the interaction between topography and water dynamics as flood levels recede. The scale of the landscape in which this phenomenon is observed is also relevant. Certainly, patch initiation depends largely on the landscape along the river, such as small altitudinal variations in the topography of the terrain, as well as the presence of local impedimenta such as floating fallen trunks and large palm fronds. As a result, the size, abundance and position of such patches varies from year to year, influencing the predictability of this resource for local frugivores. Although the potential ecological importance of seed and seedling patches in the regeneration and long-term maintenance of igapó tree populations has been considered em passim (Ferreira et al., 2010; Piedade et al., 2010), no studies have looked at how seed patches affect the animals that inhabit the region. Here, we investigated how seed patches influence the local spatial patterning of herbivorous vertebrates in igapó forests. mainour main objective was to evaluate whether areas of seed and seedling accumulation are positively related to increased presence of mammal and bird species in unflooded igapó forest. Specifically, we aimed to: (1) compare vertebrate visitation frequency at seed patches and areas of forest floor lacking such patches (control areas), (2) compare fruit and seed consumption frequencies at patches and control areas, and (3) evaluate whether (i) patch size, (ii) distance from a patch to the edge of the nearest terra firme forest, and (iii) predator occurrence influenced vertebrate visitation frequency at such seed patches. This provides a baseline for future studies of the ecological dynamics of such seed patches: consumption and seed dispersal by animals (local-scale ecological processes), as well as the importance of floodplain forests (large-scale ecological processes) in furnishing key resources for terra firme-based vertebrates. MATERIAL AND METHODS Study area The study was conducted in Jaú National Park (JNP) (01°54′–01°57′S, 61°27′–61°28′W) (Fig. 1), located between the cities of Barcelos and Novo Airão, on the southern bank of the Rio Negro, some 220 km north-west and upstream of Manaus, the capital city of Amazonas State, Brazil. Average annual temperature is 26.7 °C, and annual average rainfall is ~2400 mm. The rainy season occurs from January to May and the lowest rainfall season is between June and October (Borges et al., 2004). The hydrological regime has a monomodal flood pulse with mean amplitude of 8.17 m (peld-maua.inpa.gov.br). Flood level maxima occur during the second half of June, and minima in early November [PELD MAUA website (peld-maua.inpa.gov.br); values calculated from the Agência Nacional de Águas (ANA) hydrological station data at Moura, 70 km downstream from the Park]. JNP includes some 22 720 km2 of igapó, terra firme and white sand forests, with 12% of the area occupied by igapó forest, seasonally flooded by the Rio Jaú (Fundação Vitória Amazônica, 1988). The study was conducted between September and November 2015, during the low water period. Figure 1. View largeDownload slide Map of Jaú National Park in Central Amazonia showing the locations of 36 1-ha plots in igapó forest (grey area) and three fire events (diamond) that occurred near the sampling area. We used the mask from Hess et al. (2015) to obtain the boundaries between flooded and non-flooded areas. Records of fires during fieldwork in the study area are from: http://www.inpe.br/queimadas/bdqueimadas (18 July 2018). Figure 1. View largeDownload slide Map of Jaú National Park in Central Amazonia showing the locations of 36 1-ha plots in igapó forest (grey area) and three fire events (diamond) that occurred near the sampling area. We used the mask from Hess et al. (2015) to obtain the boundaries between flooded and non-flooded areas. Records of fires during fieldwork in the study area are from: http://www.inpe.br/queimadas/bdqueimadas (18 July 2018). Sample design To study the interactions between vertebrates and seed patches, we established 36 (50 × 20-m) plots on the floor of unflooded igapó forest. These were always parallel to the river channel and at different flood levels. Within each plot we: (1) recorded the size of all seed patches; and (2) randomly selected one seed patch per plot (through a random draw, so that patches of a range of sizes were sampled), and positioned a camera trap within it (Reconyx). To test whether animals were selectively visiting those areas of forest floor with seed patches more often than random, an equal number of camera traps were positioned to provide photographic coverage of control sites (areas where no seed patches were present inside the study plots). Control camera placement was identical to that at seed patches, at the same flood level, a determination based on highwater marks on tree trunks, but in places where seed accumulation was not observed. Camera traps were set to take five photos per event, without pause. Sequential photos from the same species within 15 min were considered a single record. We considered this time interval sufficient to record the frequency of animal visitation at patches with abundant resources, because the focus of the study was not the abundance of animals in the area, but the frequency of visitation to seed patches. Cameras were installed in a sequence that began downstream and moved upstream as the study progressed. Fieldwork was divided into two campaigns: during the first (30 days during September and October) we sampled two sets of 12 plots (50 × 20 m) and, in the second (15 days during November), a further set of 12 plots. Site positions tracked the progressive drop of the Jaú river water level, and so moved ever closer to the middle of the river bed. However, no study plots were inundated during the study period. In each sampling, cameras remained in position, operating continuously, for 15 days. Between-plot distance was at least 1.5 km, with plots established along different elevations from the river level, so that, overall, the entire range of the flood-duration gradient was covered. Using 1 × 1-m sub-plots, we sampled the density and identified all seeds/seedlings in each seed patch and in all control areas where a camera trap was installed. We also counted those seeds/seedlings that were consumed (bite marks) within these sub-plots. Botanical material was collected in the field and seedlings and seeds were identified at the INPA Herbarium, Manaus. The density of patches per plot was calculated by summing the area of all patches within the plot and dividing by the total plot area (1000 m2). We used the mask from Hess et al. (2015) that mapped wetland extent, vegetation cover and inundation state from the entire lowland Amazon at 100-m resolution using mosaics from the Japanese Earth Resources Satellite (JERS-1), and which are free for use and available on the ORNL DAAC website at https://doi.org/10.3334/ORNLDAAC/1284. We used these mosaics to locate boundaries between flooded and non-flooded areas. To measure sampling point distances from the centre of each plot to the edge of the closest terra firme forest, linear measurements of distances in metres were made manually using the MEASUREMENT (Quantum GIS, v.2.10) tool. Data analysis We used a paired Wilcoxon test to determine whether vertebrate occurrence was higher at seed patches than at comparable areas of forest floor lacking such patches (control areas). The sample unit of this analysis was the total number of individual herbivorous vertebrates (potential direct consumers from the seed/seedling patches) recorded per plot over 15 days at patches and in control areas. Exploratory tests were performed to analyse possible relationships between the occurrence of specific vertebrate groups and the presence of patches. Several clusters were tested (frugivores, prey, rodents, large and medium, small, herbivorous, carnivorous), but none showed a significant relationship with patches at the micro scale. We therefore focused our analyses on the herbivorous vertebrate assemblage visiting the area of study. To compare seed and seedling consumption by terrestrial vertebrates within and outside patches, we also used a paired Wilcoxon test, comparing observed percentages of consumed plants (both seeds and seedlings present inside 1 × 1-m sub-plots) in patches and control plots. Spatial dependence between neighbouring plots (seed patches and control areas) was accounted for, as we used a statistical paired test. As preliminary analysis showed that seed patch size and patch density per plot were correlated (Pearson’s r = 0.78), we used only seed patch size in subsequent models. We used a generalized linear mixed model (GLMM) (Bolker et al., 2009) to test the effects of seed patch size, distance between patch and nearest terra firme forest edge, and occurrence of mammalian carnivores as possible influences on herbivore visitation frequency. For the last of these, we defined ‘occurrence’ as the total number of herbivorous vertebrates or carnivorous mammals recorded at seed patches (our dependent variable). At the beginning of the first campaign (low water season), water level dropped rapidly during the next few days (up to 40 cm per day – our pers. obs.), depositing seed patches on the forest floor. In contrast, during the second campaign, the river was at its lowest level, and many of the seeds present in the patches were already germinating. Therefore, flood depth was strongly correlated with the field campaign. Accordingly, we applied a Poisson error distribution to account for count data, and included field campaign as a random effect in the model to control for temporal and spatial dependence in the data. To test the model fit, we compared Akaike’s information criterion (AIC) of the full GLMM with the respective AIC value for the null model (intercept and random effect only). Full GLMMs were selected when delta AIC (the difference between the full and null models) was >2 (Zuur et al., 2009). To measure the importance of the random variables in our results, we calculated the marginal and conditional R2 for the GLMM. Marginal R2 provides the variation explained only by the fixed variables, while the conditional R2 gives the variation explained by fixed and random effects in the model (Nakagawa & Schielzeth, 2013). We also displayed the selected models graphically using the visreg package (Breheny & Burchett, 2013). The partial graphs show expected values of the dependent variable and expected values of the target independent variable if all other independent variables in the analysis are maintained at their median values. GLMMs were created with the lme4 package (Bates et al., 2015) in R (R Core Team, 2016). RESULTS We sampled a total area of 3.6 ha of unflooded igapó forest floor. Each 1000-m2 plot had on average 2.4 seed patches (± SD = 1.18). Within the total area sampled we recorded 89 seed patches. Seeds and seedlings from 61 identified species were recorded from the seed patches and control areas (Table 1). A further 11 morphospecies remain unidentified. There were higher numbers of seeds in patches than in control areas (W = 1194.5, P < 0.01), demonstrating that greater seed deposition does occur in some igapó locations than in others (Supporting Information, Fig. S2). The mean number of plant species in the patches was 5.58 (± SD = 2.82), while in control areas on average we found 2.80 species (± SD = 1.5). Patch size varied widely, the smallest found being 1.34 m2 and the largest around 715 m2. Their distribution did not follow a discernable pattern, and the distance between them was variable (2–40 m). Also, we observed no difference between plant species composition for seed patches and control areas per camera trap station (PERMANOVA, R = 0.01, P = 0.268) (Fig. S3). Table 1. Plant species and morphotypes present in seed patches and control areas in igapó forest during the dry season, Jaú National Park (occurrence at SP = seed patches and CA = control areas) Family Species/ morphotypes Occurrence Arecaceae cf. Geonoma sp.1 SP, CA Iriartella setigera SP Leopoldinia pulchra SP, CA Mauritiella armata CA Oenocarpus mapora SP, CA Bignoniaceae Handroanthus barbatus SP Chrysobalanaceae Couepia paraensis SP, CA cf. Hirtella sp.2 SP Licania sp.1, 2, 3, 4, 5, 6, cf. 7 SP, CA Clusiaceae Calophyllum brasiliense SP, CA Garcinia sp.1 SP Combretaceae Terminalia sp.1, 2 SP, CA Combretum sp.1, 2, 3, cf. 4 SP Connaraceae Connarus sp.1 SP Rourea sp.1 SP Ebenaceae Diospyros sp.1 SP Euphorbiaceae Alchornea sp.1 SP, CA Amaioua sp.1, 2 SP Hevea spruceana CA Mabea nitida SP, CA Mabea sp.1, 2 SP Elaeocarpaceae Sloanea sp.1 CA Fabaceae Acosmium sp.1 SP cf. Cynometra bauhiniifolia SP, CA Cynometra sp.1, cf. 2 SP Dipteryx odorata SP Dalbergia sp. 1 SP, CA Hydrochorea sp.1, 2 SP, CA Macrolobium acaciifolium CA Macrolobium sp. 1 SP, CA Ormosia sp.1, 2 SP Swartzia polyphylla SP, CA Swartzia sp.1, 2, 3 SP, CA Tachigali sp. 1, 2, 3 SP Gnetaceae Gnetum sp.1 CA Humiriaceae Humiriastrum sp.1, 2 SP Lecythidaceae Eschweilera tenuifolia SP, CA Lauraceae Ocotea cymbarum SP, CA Ocotea sp.1, 2 SP, CA Malphigiaceae Byrsonima sp.1 SP Menispermaceae Abuta sp.1 CA Mimosoideae Abarema sp.1 CA Myrtaceae Eugenia sp.1, 2 SP, CA Myrciaria sp.1 SP, CA Passifloraceae Passiflora sp.1 SP Polygalaceae Securidaca sp.1 CA Symmeria paniculata CA Rubiaceae Coussarea sp.1 SP Sapindaceae Allophylus sp.1 SP Sapotaceae cf. Elaeoluma sp.1, 2 SP Morphotype Morphotypes 1, 2, 3, 4 SP, CA Family Species/ morphotypes Occurrence Arecaceae cf. Geonoma sp.1 SP, CA Iriartella setigera SP Leopoldinia pulchra SP, CA Mauritiella armata CA Oenocarpus mapora SP, CA Bignoniaceae Handroanthus barbatus SP Chrysobalanaceae Couepia paraensis SP, CA cf. Hirtella sp.2 SP Licania sp.1, 2, 3, 4, 5, 6, cf. 7 SP, CA Clusiaceae Calophyllum brasiliense SP, CA Garcinia sp.1 SP Combretaceae Terminalia sp.1, 2 SP, CA Combretum sp.1, 2, 3, cf. 4 SP Connaraceae Connarus sp.1 SP Rourea sp.1 SP Ebenaceae Diospyros sp.1 SP Euphorbiaceae Alchornea sp.1 SP, CA Amaioua sp.1, 2 SP Hevea spruceana CA Mabea nitida SP, CA Mabea sp.1, 2 SP Elaeocarpaceae Sloanea sp.1 CA Fabaceae Acosmium sp.1 SP cf. Cynometra bauhiniifolia SP, CA Cynometra sp.1, cf. 2 SP Dipteryx odorata SP Dalbergia sp. 1 SP, CA Hydrochorea sp.1, 2 SP, CA Macrolobium acaciifolium CA Macrolobium sp. 1 SP, CA Ormosia sp.1, 2 SP Swartzia polyphylla SP, CA Swartzia sp.1, 2, 3 SP, CA Tachigali sp. 1, 2, 3 SP Gnetaceae Gnetum sp.1 CA Humiriaceae Humiriastrum sp.1, 2 SP Lecythidaceae Eschweilera tenuifolia SP, CA Lauraceae Ocotea cymbarum SP, CA Ocotea sp.1, 2 SP, CA Malphigiaceae Byrsonima sp.1 SP Menispermaceae Abuta sp.1 CA Mimosoideae Abarema sp.1 CA Myrtaceae Eugenia sp.1, 2 SP, CA Myrciaria sp.1 SP, CA Passifloraceae Passiflora sp.1 SP Polygalaceae Securidaca sp.1 CA Symmeria paniculata CA Rubiaceae Coussarea sp.1 SP Sapindaceae Allophylus sp.1 SP Sapotaceae cf. Elaeoluma sp.1, 2 SP Morphotype Morphotypes 1, 2, 3, 4 SP, CA View Large Table 1. Plant species and morphotypes present in seed patches and control areas in igapó forest during the dry season, Jaú National Park (occurrence at SP = seed patches and CA = control areas) Family Species/ morphotypes Occurrence Arecaceae cf. Geonoma sp.1 SP, CA Iriartella setigera SP Leopoldinia pulchra SP, CA Mauritiella armata CA Oenocarpus mapora SP, CA Bignoniaceae Handroanthus barbatus SP Chrysobalanaceae Couepia paraensis SP, CA cf. Hirtella sp.2 SP Licania sp.1, 2, 3, 4, 5, 6, cf. 7 SP, CA Clusiaceae Calophyllum brasiliense SP, CA Garcinia sp.1 SP Combretaceae Terminalia sp.1, 2 SP, CA Combretum sp.1, 2, 3, cf. 4 SP Connaraceae Connarus sp.1 SP Rourea sp.1 SP Ebenaceae Diospyros sp.1 SP Euphorbiaceae Alchornea sp.1 SP, CA Amaioua sp.1, 2 SP Hevea spruceana CA Mabea nitida SP, CA Mabea sp.1, 2 SP Elaeocarpaceae Sloanea sp.1 CA Fabaceae Acosmium sp.1 SP cf. Cynometra bauhiniifolia SP, CA Cynometra sp.1, cf. 2 SP Dipteryx odorata SP Dalbergia sp. 1 SP, CA Hydrochorea sp.1, 2 SP, CA Macrolobium acaciifolium CA Macrolobium sp. 1 SP, CA Ormosia sp.1, 2 SP Swartzia polyphylla SP, CA Swartzia sp.1, 2, 3 SP, CA Tachigali sp. 1, 2, 3 SP Gnetaceae Gnetum sp.1 CA Humiriaceae Humiriastrum sp.1, 2 SP Lecythidaceae Eschweilera tenuifolia SP, CA Lauraceae Ocotea cymbarum SP, CA Ocotea sp.1, 2 SP, CA Malphigiaceae Byrsonima sp.1 SP Menispermaceae Abuta sp.1 CA Mimosoideae Abarema sp.1 CA Myrtaceae Eugenia sp.1, 2 SP, CA Myrciaria sp.1 SP, CA Passifloraceae Passiflora sp.1 SP Polygalaceae Securidaca sp.1 CA Symmeria paniculata CA Rubiaceae Coussarea sp.1 SP Sapindaceae Allophylus sp.1 SP Sapotaceae cf. Elaeoluma sp.1, 2 SP Morphotype Morphotypes 1, 2, 3, 4 SP, CA Family Species/ morphotypes Occurrence Arecaceae cf. Geonoma sp.1 SP, CA Iriartella setigera SP Leopoldinia pulchra SP, CA Mauritiella armata CA Oenocarpus mapora SP, CA Bignoniaceae Handroanthus barbatus SP Chrysobalanaceae Couepia paraensis SP, CA cf. Hirtella sp.2 SP Licania sp.1, 2, 3, 4, 5, 6, cf. 7 SP, CA Clusiaceae Calophyllum brasiliense SP, CA Garcinia sp.1 SP Combretaceae Terminalia sp.1, 2 SP, CA Combretum sp.1, 2, 3, cf. 4 SP Connaraceae Connarus sp.1 SP Rourea sp.1 SP Ebenaceae Diospyros sp.1 SP Euphorbiaceae Alchornea sp.1 SP, CA Amaioua sp.1, 2 SP Hevea spruceana CA Mabea nitida SP, CA Mabea sp.1, 2 SP Elaeocarpaceae Sloanea sp.1 CA Fabaceae Acosmium sp.1 SP cf. Cynometra bauhiniifolia SP, CA Cynometra sp.1, cf. 2 SP Dipteryx odorata SP Dalbergia sp. 1 SP, CA Hydrochorea sp.1, 2 SP, CA Macrolobium acaciifolium CA Macrolobium sp. 1 SP, CA Ormosia sp.1, 2 SP Swartzia polyphylla SP, CA Swartzia sp.1, 2, 3 SP, CA Tachigali sp. 1, 2, 3 SP Gnetaceae Gnetum sp.1 CA Humiriaceae Humiriastrum sp.1, 2 SP Lecythidaceae Eschweilera tenuifolia SP, CA Lauraceae Ocotea cymbarum SP, CA Ocotea sp.1, 2 SP, CA Malphigiaceae Byrsonima sp.1 SP Menispermaceae Abuta sp.1 CA Mimosoideae Abarema sp.1 CA Myrtaceae Eugenia sp.1, 2 SP, CA Myrciaria sp.1 SP, CA Passifloraceae Passiflora sp.1 SP Polygalaceae Securidaca sp.1 CA Symmeria paniculata CA Rubiaceae Coussarea sp.1 SP Sapindaceae Allophylus sp.1 SP Sapotaceae cf. Elaeoluma sp.1, 2 SP Morphotype Morphotypes 1, 2, 3, 4 SP, CA View Large Total sampling effort was 1080 camera-trap*days. We obtained 155 photographic records of: eight rodent species (three of which remain unidentified), plus 11 other mammal and nine bird species (see Table 2 for a complete species list). However, herbivorous vertebrate visitation frequency was similar between seed patches and control areas (V = 162.5, P = 0.729) (Fig. 2). The relative number of seeds and seedlings consumed was also similar between seed patches and control areas (V = 102.5, P = 0.4723) (Fig. 3). Table 2. List of vertebrate species recorded by camera traps in igapó forest during the dry season at Jaú National Park Group/Species Common name Seed patches Control areas Total Mammals Carnivora Leopardus pardalis Ocelot 9 2 11 Leopardus wiedii Margay 3 5 8 cf. Leopardus wiedii 0 1 1 Panthera onca Jaguar 1 3 4 Pteronura brasiliensis Giant otter 2 0 2 Total carnivores 15 11 26 Rodentia Cuniculus paca Lowland paca 3 7 10 Dasyprocta leporina Red-rumped agouti 1 0 1 Hydrochoerus hydrochaeris Capybara 0 1 1 Isothrix cf. bistriata Yellow-crowned brush-tailed rat 0 1 1 cf. Makalata sp. 5 4 9 cf. Oecomys sp. 2 1 3 Proechimys sp. Spiny rat 4 2 6 cf. Proechimys sp. 2 3 5 Didelphimorphia Didelphis marsupialis Common ppossum 7 8 15 Philander sp. Four-eyed ppossum 9 11 20 Perissodactyla Tapirus terrestris Lowland tapir 0 1 1 Artiodactyla Pecari tajacu Collared peccary 3 1 4 Mazama guazoubira Gray brocket deer 0 1 1 Mazama nemorivaga Brown brocket deer 3 2 5 Primates Cebus albifrons White-fronted capuchin 3 4 7 Birds Anseriformes Cairina moschata Muscovy duck 2 2 2 Struthioniformes Crypturellus undulatus Undulated tinamou 1 4 5 Crypturellus sp. 1 1 2 Columbiformes Geotrygon montana Ruddy quail-dove 2 0 2 Leptotila rufaxilla Grey-fronted dove 4 5 9 Leptotila sp. 1 1 2 Galliformes Pauxi tuberosa Razor-billed curassow 2 6 8 Gruiformes Psophia leucoptera White-winged trumpeter 1 0 1 Pelecaniformes Tigrisoma lineatum Rufescent tiger-heron 1 1 2 Struthioniformes Tinamus guttatus White-throated tinamou 1 0 1 Tinamus cf guttatus 1 0 1 Tinamus major Great tinamou 3 0 3 Tinamus cf major 0 1 1 Tinamus sp. 0 1 1 Total non-carnivores 62 69 129 Group/Species Common name Seed patches Control areas Total Mammals Carnivora Leopardus pardalis Ocelot 9 2 11 Leopardus wiedii Margay 3 5 8 cf. Leopardus wiedii 0 1 1 Panthera onca Jaguar 1 3 4 Pteronura brasiliensis Giant otter 2 0 2 Total carnivores 15 11 26 Rodentia Cuniculus paca Lowland paca 3 7 10 Dasyprocta leporina Red-rumped agouti 1 0 1 Hydrochoerus hydrochaeris Capybara 0 1 1 Isothrix cf. bistriata Yellow-crowned brush-tailed rat 0 1 1 cf. Makalata sp. 5 4 9 cf. Oecomys sp. 2 1 3 Proechimys sp. Spiny rat 4 2 6 cf. Proechimys sp. 2 3 5 Didelphimorphia Didelphis marsupialis Common ppossum 7 8 15 Philander sp. Four-eyed ppossum 9 11 20 Perissodactyla Tapirus terrestris Lowland tapir 0 1 1 Artiodactyla Pecari tajacu Collared peccary 3 1 4 Mazama guazoubira Gray brocket deer 0 1 1 Mazama nemorivaga Brown brocket deer 3 2 5 Primates Cebus albifrons White-fronted capuchin 3 4 7 Birds Anseriformes Cairina moschata Muscovy duck 2 2 2 Struthioniformes Crypturellus undulatus Undulated tinamou 1 4 5 Crypturellus sp. 1 1 2 Columbiformes Geotrygon montana Ruddy quail-dove 2 0 2 Leptotila rufaxilla Grey-fronted dove 4 5 9 Leptotila sp. 1 1 2 Galliformes Pauxi tuberosa Razor-billed curassow 2 6 8 Gruiformes Psophia leucoptera White-winged trumpeter 1 0 1 Pelecaniformes Tigrisoma lineatum Rufescent tiger-heron 1 1 2 Struthioniformes Tinamus guttatus White-throated tinamou 1 0 1 Tinamus cf guttatus 1 0 1 Tinamus major Great tinamou 3 0 3 Tinamus cf major 0 1 1 Tinamus sp. 0 1 1 Total non-carnivores 62 69 129 View Large Table 2. List of vertebrate species recorded by camera traps in igapó forest during the dry season at Jaú National Park Group/Species Common name Seed patches Control areas Total Mammals Carnivora Leopardus pardalis Ocelot 9 2 11 Leopardus wiedii Margay 3 5 8 cf. Leopardus wiedii 0 1 1 Panthera onca Jaguar 1 3 4 Pteronura brasiliensis Giant otter 2 0 2 Total carnivores 15 11 26 Rodentia Cuniculus paca Lowland paca 3 7 10 Dasyprocta leporina Red-rumped agouti 1 0 1 Hydrochoerus hydrochaeris Capybara 0 1 1 Isothrix cf. bistriata Yellow-crowned brush-tailed rat 0 1 1 cf. Makalata sp. 5 4 9 cf. Oecomys sp. 2 1 3 Proechimys sp. Spiny rat 4 2 6 cf. Proechimys sp. 2 3 5 Didelphimorphia Didelphis marsupialis Common ppossum 7 8 15 Philander sp. Four-eyed ppossum 9 11 20 Perissodactyla Tapirus terrestris Lowland tapir 0 1 1 Artiodactyla Pecari tajacu Collared peccary 3 1 4 Mazama guazoubira Gray brocket deer 0 1 1 Mazama nemorivaga Brown brocket deer 3 2 5 Primates Cebus albifrons White-fronted capuchin 3 4 7 Birds Anseriformes Cairina moschata Muscovy duck 2 2 2 Struthioniformes Crypturellus undulatus Undulated tinamou 1 4 5 Crypturellus sp. 1 1 2 Columbiformes Geotrygon montana Ruddy quail-dove 2 0 2 Leptotila rufaxilla Grey-fronted dove 4 5 9 Leptotila sp. 1 1 2 Galliformes Pauxi tuberosa Razor-billed curassow 2 6 8 Gruiformes Psophia leucoptera White-winged trumpeter 1 0 1 Pelecaniformes Tigrisoma lineatum Rufescent tiger-heron 1 1 2 Struthioniformes Tinamus guttatus White-throated tinamou 1 0 1 Tinamus cf guttatus 1 0 1 Tinamus major Great tinamou 3 0 3 Tinamus cf major 0 1 1 Tinamus sp. 0 1 1 Total non-carnivores 62 69 129 Group/Species Common name Seed patches Control areas Total Mammals Carnivora Leopardus pardalis Ocelot 9 2 11 Leopardus wiedii Margay 3 5 8 cf. Leopardus wiedii 0 1 1 Panthera onca Jaguar 1 3 4 Pteronura brasiliensis Giant otter 2 0 2 Total carnivores 15 11 26 Rodentia Cuniculus paca Lowland paca 3 7 10 Dasyprocta leporina Red-rumped agouti 1 0 1 Hydrochoerus hydrochaeris Capybara 0 1 1 Isothrix cf. bistriata Yellow-crowned brush-tailed rat 0 1 1 cf. Makalata sp. 5 4 9 cf. Oecomys sp. 2 1 3 Proechimys sp. Spiny rat 4 2 6 cf. Proechimys sp. 2 3 5 Didelphimorphia Didelphis marsupialis Common ppossum 7 8 15 Philander sp. Four-eyed ppossum 9 11 20 Perissodactyla Tapirus terrestris Lowland tapir 0 1 1 Artiodactyla Pecari tajacu Collared peccary 3 1 4 Mazama guazoubira Gray brocket deer 0 1 1 Mazama nemorivaga Brown brocket deer 3 2 5 Primates Cebus albifrons White-fronted capuchin 3 4 7 Birds Anseriformes Cairina moschata Muscovy duck 2 2 2 Struthioniformes Crypturellus undulatus Undulated tinamou 1 4 5 Crypturellus sp. 1 1 2 Columbiformes Geotrygon montana Ruddy quail-dove 2 0 2 Leptotila rufaxilla Grey-fronted dove 4 5 9 Leptotila sp. 1 1 2 Galliformes Pauxi tuberosa Razor-billed curassow 2 6 8 Gruiformes Psophia leucoptera White-winged trumpeter 1 0 1 Pelecaniformes Tigrisoma lineatum Rufescent tiger-heron 1 1 2 Struthioniformes Tinamus guttatus White-throated tinamou 1 0 1 Tinamus cf guttatus 1 0 1 Tinamus major Great tinamou 3 0 3 Tinamus cf major 0 1 1 Tinamus sp. 0 1 1 Total non-carnivores 62 69 129 View Large Figure 2. View largeDownload slide Number of herbivorous vertebrate occurrences, compared for seed patches and control areas (shown in paired format by the dashed lines), in igapó forest. Figure 2. View largeDownload slide Number of herbivorous vertebrate occurrences, compared for seed patches and control areas (shown in paired format by the dashed lines), in igapó forest. Figure 3. View largeDownload slide Percentage of seeds/seedlings consumed by vertebrates in areas with and without seed patches (shown in paired format by the dashed lines). No significant difference was observed (P > 0.05). Figure 3. View largeDownload slide Percentage of seeds/seedlings consumed by vertebrates in areas with and without seed patches (shown in paired format by the dashed lines). No significant difference was observed (P > 0.05). The full GLMM had a better fit (AIC = 191.66) than the null model (AIC = 203.86). Seed patch size explained most of the changes in vertebrate assemblage in GLMM analyses that considered all 36 plots. Seed patch size (P < 0.001) (Fig. 4A) and distance to terra firme forest (P < 0.05) (Fig. 4B) were positively and negatively related, respectively, to vertebrate assemblage occurrence (Fig. 4). In contrast, the presence of predators failed to explain the occurrence of herbivorous vertebrates across the dry igapó forest (Fig. 4C). Thus, igapó seasonality (included in our model as a random effect) seems to be more important than the fixed factors. The full model (conditional R2) explained around 38% of the variation in vertebrate occurrence while the fixed effects (marginal R2) explained 9%. Figure 4. View largeDownload slide Relationship between vertebrate number and patch size (A), distance to terra firme forest (B) and number of predators (C). Regression lines are shown only for significant relationships (see text). The grey area represents the 95% confidence interval. Figure 4. View largeDownload slide Relationship between vertebrate number and patch size (A), distance to terra firme forest (B) and number of predators (C). Regression lines are shown only for significant relationships (see text). The grey area represents the 95% confidence interval. DISCUSSION At the local scale, patches with high resource density were not more attractive to potential consumers of seeds and seedlings. In addition, we found that the percentage of seed and seedling consumption did not differ within and outside the seed patches (Fig. 3). Although there was no difference in occurrence of the studied vertebrate assemblage between seed patches and control areas, a large number of vertebrate species and individuals are present in igapó forest when it is unflooded. For many vertebrate assemblages, the igapó forests provide, during the non-flooded season, not only an abundant source of food resources, but also water (because many terra firme forest streams are dry at this time: Bodmer, 1990; Haugaasen & Peres, 2007). In addition to the photographic records, other indications, such as rodent burrows, faeces, bitten seeds and paw-prints, also show this environment to be occupied or visited frequently by a variety of mammal species (Antunes et al., 2017; A. C. Antunes et al., unpubl. data). In comparison with studies carried out in várzea forests (white-water floodplains of the Amazon River and its tributaries) using a higher sampling effort, the total number of vertebrate records in this study was lower, although species richness was higher (Alvarenga et al., 2018; Costa et al., 2018). The number of records was probably affected by several fire events that happened in the JNP in 2015, some of them very close to the sampling areas of this study, and observed during the fieldwork (Fig. 1). Igapó forests have a low resilience to fire, and after a second event are unable to recover (Flores et al., 2017). This only reinforces the potential ecological importance of these seed and seedling accumulations, the animal–plant interactions that are associated with them (including seed consumption and dispersion by animals), and which species (plant and animal) are involved, all of which are currently unknown. Haugaasen & Peres (2007) demonstrated that the increased availability of fruits and seeds deposited on the ground in igapó forests, and new vegetative growth when water levels recede, coincide with the migration of fruit- and seed-eating vertebrates from terra firme forest. Bodmer (1990) found floodplain forests in Amazonian Peru also to be seasonally important for ungulates, and Barnett et al. (2012) reported seed patch use by otherwise arboreal igapó-dwelling primates that moved to the ground to exploit them. Our results contradict the central hypothesis proposed in this study and so raise the following question: why, given the presence of patches of abundant resources, are animals not distributed as predicted by the IFD? Ungulate herbivores and small mammals are well known to have good spatial memories with regard to aggregations of food resources (Howery & Bailey, 1999). Also, rodents detect, and prefer to forage in, larger and denser seed patches (Cabin et al., 2000; Baraibar, 2011); consequently, resource distribution can strongly influence the spatial use of an area by such animals (Fagan et al., 2013). However, because of how igapó seed patches form, their size and spatial distribution on the igapó forest floor is likely to vary significantly between years, decreasing the relevance of any long-term spatial memory of resource locations that an animal might deploy. Consequently, fruit- and seed-eaters visiting unflooded igapó forests are probably exploring the area extensively in search of resources that are unpredictable in size, composition and density. This characteristic violates one of the principles of the IFD, namely that individuals are aware of the resource profitabilities of each patch. In addition, as animals are unlikely to be territorial under such circumstances, the progressive depletion of patch value over time (a key element of resource use foraging: see Krebs et al., 1974; Kotler & Brown, 1990) is also likely to be highly variable in nature and extent. There may, in addition, be others factors in play. In this study, camera-traps recorded several carnivores (n = 26; 16% of records). The large numbers of paw-prints also attested to the frequent occurrence of such animals in the study habitat (A. C. Antunes, unpubl. data). The species present included those known to prey on small granivores (e.g. rats and mice), such as margay and ocelot, and those, such as the jaguar, that prey on larger species (e.g. paca, peccary and deer). In comparison to upland forests, an open landscape such as the unflooded igapó forest (de Almeida et al., 2016: and see Supporting Information, Fig. S4) offers greater visibility and consequently has an influence on the risk of predation by these predators. For the golden-backed uacari monkey (Cacajao ouakary), another igapó mammal that uses these seed patches, landscape composition is one of the key factors influencing its foraging behaviour: this primate avoids seed patches close to dense ground-based vegetation (Barnett et al., 2012). In addition, not all fruits and seeds available in patches are palatable, as plants may invest less in anti-herbivory defences when nutrients are plentiful (Coley, 1985). In the generally nutrient-poor study area, plants are known to invest heavily in anti-herbivory defences such as tannins and other phenolic compounds, which in turn are considered to reduce the carrying capacity for a variety of organisms, both directly (vertebrate and invertebrate herbivores) and indirectly (their predators and parasites) (Freeland & Janzen, 1974; Janzen, 1974; Marples, 2018). Also, many fruits and seeds are of a size compatible with the consumer (Antunes et al., 2017), so that energy reward in each patch is unpredictable. Coupled with the random distribution of most seeds and seedlings (which consequently require extensive travel by herbivores to find them), this could influence the ways in which potential prey species are consuming and moving between the resources available at seed patches and control areas (Mitchell & Lima, 2002; Schmitz et al., 2004). In prey species known to have well-developed spatial memory capacity (e.g. scatter-hoarding rodents: Smulders et al., 2010; Hirsch et al., 2013), it is known that they will randomize movements to decrease encounters with predators (Mitchell & Lima, 2002). Although patch formation is unpredictable at the small scale (between treatments), it is likely that at larger scales (e.g. along the river) there are more suitable sites for floating patches to be deposited, depending on topography and other landscape features, which makes the deposition of patches somewhat more predictable for the vertebrates that are present in these areas, especially those with larger home-ranges. Despite frugivorous vertebrate occurrences not being related to seed patches locally, they are positively related to seed patches sizes at larger scales. Other studies have also shown that species responses are scale-dependent (Garmendia, 2013; Quesnelle, 2014; Alvarenga et al., 2018). For várzea forests, mammal assemblage composition varies according to scale, which reinforces the importance of considering scale in ecological landscape studies (Alvarenga et al., 2018). In our study, the larger the size of patches, the greater the number of occurrences of vertebrates associated with them. According to the IFD, animals tend to forage in the highest quality patches, and therefore larger patches would be more attractive. Even though patch compositions are unpredictable, larger patches are more likely to have a greater energy reward for animals that feed on them than smaller patches. Distance to terra firme forests was an additional explanatory factor in frugivorous vertebrate occurrence at larges scales within JNP igapó forests: the closer to the terra firme forest, the greater the number of recorded occurrences. During the driest months of the year, water availability is known to affect the distribution of terra firme-dwelling animals very intensely (Paredes et al., 2017), so that animals move through igapó forests in search of water. Because all plots were established near the Rio Jaú main channel, the distance from study plots to terra firme forest was linked to the distance from the main remaining free-standing water source (Supporting Information, Fig. S5), once most small streams dry up (R. S. Moreira, pers. comm.). Accordingly, patches closer to terra firme forest represent the smallest distances to be travelled from the edge of the terra firme to the igapó forest during the search for resources and water, and consequently this equates to less time of exposure and lower risk of predation. The occurrence of this frugivore assemblage in the forest is independent of the presence of predators, probably due to prey avoidance of open or exposed areas, while seeking rapid access to water and valuable food resources in the larger patches. In várzea forests, some predators (such as the margay) use the habitat homogeneously, with no precise association with their prey, probably due to the large variety of prey that occur in these forests (Alvarenga et al., 2018). A similar process seems to occur in the igapó forest during the dry season, a period in which there is large-scale lateral movement of animals from terra firme to adjacent igapó forest in search of resources and water. The absence of a correlation between predator and prey occurrences on this scale may also be linked to refuge availability, because the igapó forest floor is very open compared with other tropical forest habitats, and the absence of cover is known to explain a lack of correlation in the spatial distribution of predators and prey (Sih, 1984, Formanowicz & Bobka, 1989). CONCLUSION Information on vertebrates in igapó forests remains very limited. River flood-pulse dynamics directly influence the distribution and composition of plant assemblages and, consequently, those of animals. Differences in water levels between field campaigns explained most of the variation in vertebrate occurrence we found. The lateral movement of frugivorous vertebrates between terra firme and igapó forests reinforces the importance of these floodplains as a source of food and water during the dry season (Haugaasen & Peres, 2007; Costa et al., 2018). At the large scale, the formation of seed and seedling aggregates on the forest floor appears to be important to the herbivorous guild that feed on them during this period, but this also highlights other little-known ecological aspects, such as seedling establishment and recruitment in the area, dispersal/predation of seeds by small mammals, and infestation of these seeds by insects, among others. A better understanding of these processes is needed for igapó forests, which must be understood as functionally linked mosaics of ecologically integrated habitats, which include upland forests and white-sand scrub-forest (campina). This is important for vertebrate conservation in Amazonia, because such a dual-habitat greatly enhances the vulnerability of species that use this seasonal resource utilization strategy (Flores et al., 2017). Consequently, this study contributes to understanding the dynamic nature of frugivorous animals on the igapó forest floor, as well as paving the way for future studies. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s website. Fig. S1. A, seed patch accumulating and floating during the flooded season in igapó forest. The arrow shows the seed patch limit. B, high density of deposited seedlings (arrow), contrasting with the rest of the forest floor during the unflooded season. Fig. S2. There are significantly higher numbers of seeds and seedlings in patches than in control areas (P < 0.01), confirming the formation of seed patches in contrast to control areas (shown in paired form by dashed lines). Fig. S3. Multidimensional scaling ordination of plant species composition at seed patches and control areas. Fig. S4. Image of igapó forest within the Jaú National Park during the dry season. The open nature of the understorey allows high visibility within the forest. Fig. S5. During the dry season, it is common for streams to dry-up, so that the main channel of the Jaú River, accessible only by crossing igapó forest, becomes the main source of water for many vertebrates. The figure shows a stream in the year that the sampling was performed (the entire open area is water-filled in the high-water season). ACKNOWLEDGMENTS This study was funded by The Rufford Foundation (grant number 16299-1 to A.C.A.) and The Percy Sladen Memorial Fund (a grant to A.A.B.). We are grateful to Marcelo Paustein Moreira and Angélica Resende for their help with the geoprocessing data. We thank the Chico Mendes Institute for Biodiversity Conservation (ICMBio) for their kind support and permission to conduct our research in Jaú National Park. Logistical support was also provided by the MAUA Working Group PELD-MAUA and INPA/Max Planck Project in Manaus. We would like to thank CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) for financial aid to A.C.Antunes. F. Baccaro is supported by CNPq #309600/2017-0 productivity grant. This is contribution number 16 from the Amazonian Mammal Research Group, INPA, Manaus, and number 28 from the Igapó Study Project. We also thank three anonymous reviewers for their helpful comments. REFERENCES Albrecht M , Gotelli NJ . 2001 . Spatial and temporal niche partitioning in grassland ants . Oecologia 126 : 134 – 141 . Google Scholar Crossref Search ADS PubMed WorldCat de Almeida DRA , Nelson BW , Schietti J , Gorgens EB , Resende AF , Stark SC , Valbuena R . 2016 . Contrasting fire damage and fire susceptibility between seasonally flooded forest and upland forest in the Central Amazon using portable profiling LiDAR . Remote Sensing of Environment 184 : 153 – 160 . Google Scholar Crossref Search ADS WorldCat Alvarenga GC , Ramalho EE , Baccaro FB , da Rocha DG , Ferreira-Ferreira J , Bobrowiec PED . 2018 . Spatial patterns of medium and large size mammal assemblages in várzea and terra firme forests, Central Amazonia, Brazil . PLoS One 13 : e0198120 . Google Scholar Crossref Search ADS PubMed WorldCat Antunes AC , Baccaro F , Barnett AA . 2017 . What bite marks can tell us: use of on-fruit tooth impressions to study seed consumer identity and consumption patterns within a rodent assemblage . Mammalian Biology-Zeitschrift für Säugetierkunde 82 : 74 – 79 . Google Scholar Crossref Search ADS WorldCat Baraibar B , Daedlow D , De Mol F , Gerowitt B . 2011 . Density dependence of weed seed predation by invertebrates and vertebrates in winter wheat . Weed Research 52 : 79 – 87 . Google Scholar Crossref Search ADS WorldCat Barnett AA , Almeida T , Spironello WR , Sousa Silva W , MacLarnon A , Ross C . 2012 . Terrestrial foraging by Cacajao melanocephalus ouakary (Primates) in Amazonian Brazil: is choice of seed patch size and position related to predation risk? Folia Primatologica 83 : 126 – 139 . Google Scholar Crossref Search ADS WorldCat Barnett AA , Silva WS , Shaw PJA , Ramsay RM . 2015 . Inundation duration and vertical vegetation stratification: a preliminary description of the vegetation and structuring factors in Borokotóh (hummock igapó), an overlooked, high-diversity, Amazonian habitat . Nordic Journal of Botany 33 : 601 – 614 . Google Scholar Crossref Search ADS WorldCat Bates D , Maechler M , Bolker B , Walker S . 2015 . lme4: Linear mixed-effects models using Eigen and S4. R package version 1 . 1 – 7 . 2014 . http://CRAN.R-project.org/package=lme4. Bodmer RE . 1990 . Responses of ungulates to seasonal inundations in the Amazon floodplain . Journal of Tropical Ecology 6 : 191 – 201 . Google Scholar Crossref Search ADS WorldCat Bolker BM , Brooks ME , Clark CJ , Geange SW , Poulsen JR , Stevens MHH , White JSS . 2009 . Generalized linear mixed models: a practical guide for ecology and evolution . Trends in Ecology and Evolution 24 : 127 – 135 . Google Scholar Crossref Search ADS PubMed WorldCat Borges SH , Duringan CC , Pinheiro MR , Camargo JLC , Murchie A . 2004 . Planejando o estudo da biodiversidade na Amazônia brasileira: uma experiência no Parque Nacional do Jaú. In: Borges SH , Iwanaga S , Durigan CC , Pinheiro MR , eds. Janelas para a Biodiversidade no Parque Nacional do Jaú. Uma estratégia para o estudo da biodiversidade na Amazônia . Manaus : Fundação Vitória Amazônica , 3 – 18 . Google Preview WorldCat COPAC Breheny P , Burchett W . 2013 . Visualization of regression models using visreg . The R Journal 9 : 56 – 71 . Google Scholar Crossref Search ADS WorldCat Byers JA . 1989 . Behavioural mechanisms involved in reducing competition in bark beetles . Ecography 12 : 466 – 476 . Google Scholar Crossref Search ADS WorldCat Cabin RJ , Marshall DL , Mitchell RJ . 2000 . The demographic role of soil seed banks. II. Investigations of the fate of experimental seeds of the desert mustard Lesquerella fendleri . Journal of Ecology 88 : 293 – 302 . Google Scholar Crossref Search ADS WorldCat Coley PD , Bryant JP , Chapin FS . 1985 . Resource availability and plant antiherbivore defense . Science 230 : 895 – 899 . Google Scholar Crossref Search ADS PubMed WorldCat Correa SB , Winemiller KO , Lopez-Fernandez H , Galetti M . 2007 . Evolutionary perspectives on seed consumption and dispersal by fishes . AIBS Bulletin 57 : 748 – 756 . WorldCat Cosner C . 2005 . A dynamic model for the ideal-free distribution as a partial differential equation . Theoretical Population Biology 67 : 101 – 108 . Google Scholar Crossref Search ADS PubMed WorldCat Costa HC , Peres CA , Abrahams MI . 2018 . Seasonal dynamics of terrestrial vertebrate abundance between Amazonian flooded and unflooded forests . PeerJ Preprints 6 : e26960v1 . WorldCat Fagan WF , Lewis MA , Auger-Méthé M , Avgar T , Benhamou S , Breed G , LaDage L , Schlägel UE , Tang WW , Papastamatiou YP , Forester J , Mueller T . 2013 . Spatial memory and animal movement . Ecology Letters 16 : 1316 – 1329 . Google Scholar Crossref Search ADS PubMed WorldCat Ferreira CS , Piedade MTF , de Oliveira Wittmann A , Franco AC . 2010 . Plant reproduction in the Central Amazonian floodplains: challenges and adaptations . AoB Plants 2010 : plq009 . Google Scholar Crossref Search ADS PubMed WorldCat Flores BM , Holmgren M , Xu C , van Nes EH , Jakovac CC , Mesquita RC , Scheffer M . 2017 . Floodplains as an Achilles’ heel of Amazonian forest resilience . Proceedings of the National Academy of Sciences USA 114 : 4442 – 4446 . Google Scholar Crossref Search ADS WorldCat Formanowicz DR Jr , Bobka MS . 1989 . Predation risk and microhabitat preference: an experimental study of the behavioral responses of prey and predator . American Midland Naturalist 121 : 379 – 386 . Google Scholar Crossref Search ADS WorldCat Freeland WJ , Janzen DH . 1974 . Strategies in herbivory by mammals: the role of plant secondary compounds . The American Naturalist 108 : 269 – 289 . Google Scholar Crossref Search ADS WorldCat Fretwell SD , Lucas HL Jr . 1970 . On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development . Acta Biotheoretica 19 : 16 – 32 . Google Scholar Crossref Search ADS WorldCat Garmendia A , Arroyo-Rodríguez V , Estrada A , Naranjo EJ , Stoner KE . 2013 . Landscape and patch attributes impacting medium-and large-sized terrestrial mammals in a fragmented rain forest . Journal of Tropical Ecology 29 : 331 – 344 . Google Scholar Crossref Search ADS WorldCat Giller PS , Doube BM . 1994 . Spatial and temporal co-occurrence of competitors in Southern African dung beetle communities . Journal of Animal Ecology 63 : 629 – 643 . Google Scholar Crossref Search ADS WorldCat Hammond JI , Luttbeg B , Sih A . 2007 . Predator and prey space use: dragonflies and tadpoles in an interactive game . Ecology 88 : 1525 – 1535 . Google Scholar Crossref Search ADS PubMed WorldCat Haugaasen T , Peres CA . 2005 . Tree phenology in adjacent Amazonian flooded and unflooded forests . Biotropica 37 : 620 – 630 . Google Scholar Crossref Search ADS WorldCat Haugaasen T , Peres CA . 2007 . Vertebrate responses to fruit production in Amazonian flooded and unflooded forests . Biodiversity and Conservation 16 : 4165 – 4190 . Google Scholar Crossref Search ADS WorldCat Hawes JE , Peres CA . 2014 . Fruit–frugivore interactions in Amazonian seasonally flooded and unflooded forests . Journal of Tropical Ecology 30 : 381 – 399 . Google Scholar Crossref Search ADS WorldCat Hess LL , Melack JM , Affonso AG , Barbosa CCF , Gastil-Buhl M , Novo EMLM . 2015 . LBA-ECO LC-07 wetland extent, vegetation, and inundation: Lowland Amazon Basin . Oak Ridge : ORNL DAAC . Available at: https://doi.org/ 10.3334/ORNLDAAC/1284. Google Preview WorldCat COPAC Hirsch BT , Kays R , Jansen PA . 2013 . Evidence for cache surveillance by a scatter-hoarding rodent . Animal Behaviour 85 : 1511 – 1516 . Google Scholar Crossref Search ADS WorldCat Howery LD , Bailey DW , Laca EA . 1999 . Impact of spatial memory on habitat use . Grazing Behavior of Livestock and Wildife 70 : 91 – 100 . WorldCat Inouye BD . 2005 . Scaling up from local competition to regional coexistence across two scales of spatial heterogeneity: insect larvae in the fruits of Apeiba membranacea . Oecologia 145 : 188 – 196 . Google Scholar Crossref Search ADS PubMed WorldCat Ives AR , Kareiva P , Perry R . 1993 . Response of a predator to variation in prey density at three hierarchical scales lady beetles feeding on aphids . Ecology 74 : 1929 – 1938 . Google Scholar Crossref Search ADS WorldCat Janzen DH . 1974 . Tropical blackwater rivers, animals, and mast fruiting by the Dipterocarpaceae . Biotropica 6 : 69 – 103 . Google Scholar Crossref Search ADS WorldCat Junk WJ , Wittmann F , Schöngart J , Piedade MT . 2015 . A classification of the major habitats of Amazonian black-water river floodplains and a comparison with their white-water counterparts . Wetlands Ecology and Management 23 : 677 – 693 . Google Scholar Crossref Search ADS WorldCat Kotler BP , Brown JS . 1990 . Rates of seed harvest by two species of gerbilline rodents . Journal of Mammalogy 71 : 591 – 596 . Google Scholar Crossref Search ADS WorldCat Krebs J , Ryan J , Charnov E . 1974 . Hunting by expectation or optimal foraging: a study of patch use by chickadees . Animal Behaviour 22 : 953 – 964 . Google Scholar Crossref Search ADS WorldCat Kubitzki K , & Ziburski A . 1994 . Seed dispersal in flood plain forests of Amazonia . Biotropica 26 : 30 – 43 . Google Scholar Crossref Search ADS WorldCat Mason TH , Fortin D . 2017 . Functional responses in animal movement explain spatial heterogeneity in animal–habitat relationships . Journal of Animal Ecology 86 : 960 – 971 . Google Scholar Crossref Search ADS PubMed WorldCat Marples NM , Speed MP , Thomas RJ . 2018 . An individual-based profitability spectrum for understanding interactions between predators and their prey . Biological Journal of the Linnean Society 125 : 1 – 13 . Google Scholar Crossref Search ADS WorldCat Melack JM , Hess LL . 2010 . Remote sensing of the distribution and extent of wetlands in the Amazon basin. In: Junk WJ , Piedade MT , Wittmann F , Schöngart J , Parolin P , eds. Amazonian floodplain forests: ecophysiology, biodiversity and sustainable management . London : Springer Science & Business Media , 43 – 59 . Google Preview WorldCat COPAC Mitchell WA , Lima SL . 2002 . Predator–prey shell games: large‐scale movement and its implications for decision‐making by prey . Oikos 99 : 249 – 259 . Google Scholar Crossref Search ADS WorldCat Nakagawa S , Schielzeth H . 2013 . A general and simple method for obtaining R2 from generalized linear mixed‐effects models . Methods in Ecology and Evolution 4 : 133 – 142 . Google Scholar Crossref Search ADS WorldCat Paredes OSL , Norris D , de Oliveira TG , Michalski F . 2017 . Water availability not fruitfall modulates the dry season distribution of frugivorous terrestrial vertebrates in a lowland Amazon forest . PLoS One 12 : e0174049 . Google Scholar Crossref Search ADS PubMed WorldCat Parolin P , De Simone O , Haase K , Waldhoff D , Rottenberger S , Kuhn U , Kesselmeier J , Kleiss B , Schmidt W , Pledade MTF , Junk WJ . 2004 . Central Amazonian floodplain forests: tree adaptations in a pulsing system . The Botanical Review 70 : 357 – 380 . Google Scholar Crossref Search ADS WorldCat Parolin P , Ferreira LV , Piedade MTF , da Cunha CN , Wittmann F , Arias ME . 2016 . Flood tolerant trees in seasonally inundated lowland tropical floodplains. In: Goldstein G , Santiago L , eds. Tropical tree physiology . Cham : Springer , 127 – 147 . Google Preview WorldCat COPAC Parolin P , Waldhoff D , Piedade MT . 2010 . Fruit and seed chemistry, biomass and dispersal. In: Junk WJ , Piedade MT , Wittmann F , Schöngart J , Parolin P , eds. Amazonian floodplain forests . Dordrecht : Springer , 243 – 258 . Google Preview WorldCat COPAC Parolin P , Wittmann F , Ferreira LV . 2013 . Fruit and seed dispersal in Amazonian floodplain trees–a review . Ecotropica 19 : 15 – 32 . WorldCat PELD MAUA (Programa de Pesquisas Ecológicas de Longa duração) .Sítio Igapó - Parque Nacional do Jaú (PARNA Jaú). Website visited on 14 February 2018. Available at: https://peld-maua.inpa.gov.br/crbst_5.html. Piedade MTF , Ferreira CDS , Franco AC . 2010 . Estrategias reproductivas de la vegetación y sus respuestas al pulso de la inundación en las zonas inundables de la Amazonía Central . Revista Ecosistemas 19 : 52 – 56 . WorldCat Quesnelle PE , Lindsay KE , Fahrig L . 2014 . Low reproductive rate predicts species sensitivity to habitat loss: a meta-analysis of wetland vertebrates . PLoS One 9 : e90926 . Google Scholar Crossref Search ADS PubMed WorldCat R Core Team . 2016 . R: a language and environment for statistical computing . Vienna : R Foundation for Statistical Computing . Available at: https://www.R-project.org/. Google Preview WorldCat COPAC Renton K . 2002 . Seasonal variation in occurrence of macaws along a rainforest river . Journal of Field Ornithology 73 : 15 – 19 . Google Scholar Crossref Search ADS WorldCat Schellhorn NA , Andow DA . 2005 . Response of coccinellids to their aphid prey at different spatial scales . Population Ecology 47 : 71 – 76 . Google Scholar Crossref Search ADS WorldCat Schmitz OJ , Krivan V , Ovadia O . 2004 . Trophic cascades: the primacy of trait-mediated indirect interactions . Ecology Letters 7 : 153 – 163 . Google Scholar Crossref Search ADS WorldCat Sih A . 1984 . The behavioral response race between predator and prey . The American Naturalist 123 : 143 – 150 . Google Scholar Crossref Search ADS WorldCat Smulders TV , Gould KL , Leaver LA . 2010 . Using ecology to guide the study of cognitive and neural mechanisms of different aspects of spatial memory in food-hoarding animals . Philosophical Transactions of the Royal Society of London B: Biological Sciences 365 : 883 – 900 . Google Scholar Crossref Search ADS WorldCat Turchin P . 1998 . Quantitative analysis of movement: measuring and modeling population redistribution in animals and plants . Sunderland : Sinauer Associates . Google Preview WorldCat COPAC Ziv Y , Abramsky Z , Kotler BP , Subach A . 1993 . Interference competition and temporal and habitat partitioning in two gerbil species . Oikos 66 : 237 – 246 . Google Scholar Crossref Search ADS WorldCat Zuur AF , Ieno EN , Walker NJ , Saveliev AA , Smith GM. 2009 . Mixed effects modelling for nested data. In: Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM, eds. Mixed effects models and extensions in ecology with R. Statistics for Biology and Health . New York : Springer , 101 – 142 . Google Preview WorldCat COPAC © 2019 The Linnean Society of London, Biological Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)