Mammal Community Composition Research Articles

Ecological Characteristics of Small Mammal Communities at a Superfund Site

Wildlife species can serve as biomonitors of environmental health and are prognostic of ecotoxicological consequences when contaminants are introduced into the environment. Small mammals, particularly rodents, comprise the majority of indicator species used in terrestrial biomonitoring studies; however, many biomonitoring studies address acute effects over relatively short periods. We still know little regarding effects of chronic exposure to contaminants on small mammals. The overall goal of this study was to determine ecological characteristics of small mammal communities inhabiting a heavy metal contaminated site, Tar Creek Superfund Site, compared to reference sites located in northeastern Oklahoma over a 2 y timeframe. Primary hazardous materials present at Tar Creek Superfund Site include lead, zinc and cadmium. Mark-recapture techniques were employed to test the hypothesis that structure and composition of small mammal communities inhabiting this contaminated site would be significantly altered compared to uncontaminated reference sites. Contaminated and reference sites were similar in vegetation compositional characteristics. Small mammal communities inhabiting Tar Creek Superfund Site had reduced species diversity, including richness and evenness, compared to reference sites. Furthermore, communities within Tar Creek Superfund Site were dominated by a single species, Peromyscus leucopus (white-footed mouse). Species composition was different between contaminated sites and reference sites as evidenced by detrended correspondence analysis, with contaminated sites being more similar to each other than to either reference site. No direct link between site contamination and disparities among most ecological characteristics could be established.

Influence of Woody Vegetation on Small Mammals in Tallgrass Prairie

We examined response of small mammals to woody invasion of tallgrass prairie in northeastern Kansas by sampling sites that ranged from 0 to 100% in woody cover (i.e., frequently burned prairie to unburned closed-canopy gallery forest; woody coverage mapped from satellite imagery). Abundance and biomass of small mammals initially increased with increasing woody vegetation, but then decreased to their lowest level at 100% woody cover. Richness was greatest (an average of seven species) where woody cover was ≤17% and decreased to one species where woody cover was 100%. Abundance of white-footed mice (Peromyscus leucopus) increased as woody cover increased from 0 to 62% cover, but decreased in forested sites (100% cover). Abundance of western harvest mice (Reithrodontomys megalotis) decreased as woody cover increased; they were not recorded in forested sites. Deer mice (P. maniculatus) were present in sites with no to low levels of woody vegetation, but were absent in sites with moderate to high levels of woody cover. The 11 species captured were recorded in sites that had little woody vegetation (≤17% cover), whereas only seven species were captured at sites where considerable woody vegetation (≥28% cover) occurred. Even white-footed mice and eastern woodrats, which are considered woodland forms, decreased in abundance or were absent from forested sites. Overall, the results demonstrate that even relatively small amounts of woody vegetation in prairie landscapes can alter abundance, biomass and species richness and composition of small mammal communities.

Role of habitat and landscape in structuring small mammal assemblages in hedgerow networks of contrasted farming landscapes in Brittany, France

In this study, we investigated the environmental factors driving small mammal (rodents and shrews) assemblages in permanent habitat patches in response to a gradient of agricultural intensification. Small mammals were sampled using a trapping standard method in the hedgerow networks of three contrasted landscapes differing by their level of land-use intensity and hedgerow network density (BOC1: slightly intensified; BOC2: moderately intensified and POL: highly intensified). We hypothesized that habitat and landscape characteristics have to be considered to understand the structure of local community. In that way, we carried out a multi-scale study using environmental variables ranging from local habitat (structure and composition of the hedgerows) to hedgerows neighbourhoods in a radius of 300 m (land cover and connectivity around hedges) and to landscape units (three sites). During 1 year, 24 hedgerows were sampled seven times, representing a total of 1,379 captures (86% of rodents and 14% of shrews) and eight species, dominated by the wood mouse (Apodemus sylvaticus) and the bank vole (Clethrionomys glareolus). Inter-site variability was significant and accounted for 18% of total variation in small mammal species abundances. But intra-site variability was also highlighted: species abundance profiles may differ greatly among hedgerows within a site. The more explanatory variables were identified at the different scales of the study: the landscape unit POL was shown to be an important factor in structuring the community, but the predominant factors explaining differences of abundances among hedgerows were about local habitat. In fact, the width of hedges and the tree species richness appeared to be significant and explaining the greatest part of the total variation of the small mammal community composition.

Response of terrestrial small mammals to varying amounts and patterns of green-tree retention in Pacific Northwest forests

To sustain native species in managed forests, landowners need silvicultural strategies that retain habitat elements often eliminated during traditional harvests such as clearcut logging. One alternative is green-tree or variable retention. We investigated the response of terrestrial small mammals to experimental harvests that retained large live trees in varying amounts (approximately 100, 75, 40, and 15% of original basal area) and patterns (aggregated versus dispersed) in mature coniferous forests of western Oregon and Washington. Treatments were applied in 36, 13-ha experimental units. We used pitfall traps to sample small mammals for 4 weeks each autumn during 2 years before and 2 years after treatments. We captured 21,351 individuals of 32 species. We analyzed effects of treatments on relative abundance of 12 species. As level of retention declined, we expected species associated with closed-canopy forests to decrease ( Sorex trowbridgii, Neurotrichus gibbsii, Peromyscus keeni, Myodes [ Clethrionomys] californicus, and M. gapperi); species associated with early successional habitats to increase ( S. vagrans, P. maniculatus, Microtus longicaudus, and Microtus oregoni); and habitat generalists to show little response ( S. monticolus, S. pacificus, and S. sonomae). As expected, M. californicus declined after harvest, and P. maniculatus and M. longicaudus increased. Sorex sonomae showed an unpredicted decrease. Other species did not show consistent changes. Responses of S. monticolus, S. sonomae, and M. gapperi varied among study areas. For M. gapperi, this variation was not explained by differences in habitat structure among areas. For all species, capture rates were similar in dispersed- and aggregated-retention units. Similarity in species composition between harvested sites and controls decreased with decreasing retention. Future sampling of these treatments is needed to assess long-term responses. Based on our initial results, green-tree retention strategies need to be sensitive to regional variation in environmental characteristics and small mammal community composition.

An Evaluation of the Contribution of Cultivated Allspice (Pimenta Dioca) to Vertebrate Biodiversity Conservation in Nicaragua

Tropical deforestation has emerged as one of the most important conservation challenges of our time, both because of the high species diversity and rates of endemism of tropical forests, and because of the rapid rate at which this process is proceeding. Recent studies indicate that areas of low-intensity agroforestry have similar levels of vertebrate diversity as some primary habitats, leading some researchers and conservationists to conclude that this type of commodity production could contribute to the conservation of biodiversity. We compared the composition of bird, mammal and herpetofaunal communities in primary forest, secondary forest, and pasture—and within the allspice productive systems that have replaced pasture. We found that mammal species richness was higher in primary forest than all other habitats; however for resident and migrant birds, amphibians and reptiles, species richness was similar between primary forest and the other habitats. Despite similarities in overall numbers of species, there were numerous species that were encountered only in primary habitats. We conclude that the cultivation of allspice in a mixed productive system can offset some of the losses to biodiversity; however it should be complemented by the establishment and maintenance of protected areas to accommodate populations of primary forest specialists that are unable to persist in altered habitats.

How does landscape use influence small mammal diversity, abundance and biomass in hedgerow networks of farming landscapes?

Over the last decades, profound changes in agricultural practices in the world have led to modifications of land-use as well as landscape structure and composition. Major changes resulted in enlargement of parcel size, increase of cultivated areas and drastic reduction of permanent elements such as woods, hedges or natural meadows. In this context we chose to investigate the composition and structure of small mammal communities in the hedgerow networks of three landscape units of Western France (Brittany) differing by their level of agricultural land-use intensity and hedgerow network density: a slightly intensified dense hedgerow network landscape unit (BOC1), a moderately intensified and fragmented hedgerow network landscape unit (BOC2) and a highly intensified landscape unit on an area reclaimed from the sea (POL). Characterization of small mammal communities was performed using live trapping on permanent habitats (eight hedges per landscape unit). In each of the 24 trapping units, a standardized method was used consisting of a baited 100-m trap-line. Diversity indices were used to compare the three communities. Species richness didn't vary across landscapes whereas Shannon's index of diversity underlined a clear difference between, on the one hand, the most intensified landscape unit (POL) which displayed the lowest diversity and, on the other hand, the two other less intensified units. The abundance of small mammals differed between the three sites: they were significantly more numerous in the hedges of the most intensified site than in hedges of the two other sites. Differences between species also appeared: for example, the Bank vole (Clethrionomys glareolus) was very characteristic of POL, whereas the Pygmy shrew (Sorex minutus) was much more associated with BOC sites. Within hedges, like for abundance, small mammal biomass was the highest in the most intensified site (POL > BOC2 = BOC1). On the contrary, at the landscape scale, biomass was the lowest in POL (BOC1 > BOC2 > POL) because of the fragmentation of the hedgerow network.

Estimating the relative abundance and temporal activity patterns of wood mice (Apodemus sylvaticus) by remote photography in mediterranean post-fire habitats

We investigated whether remote photography can be considered a valid method to assess the abundance and the temporal activity patterns of the wood mouse (Apodemus sylvaticus) in two Mediterranean post-fire habitats. We compared the results with those obtained by Sherman live-trapping, a conventional procedure to obtain information on the composition of small mammal communities. We observed a highly significant increase of abundance of wood mice in shrublands (early post-fire stage) in front of pinewoods (unburned during the last 50 years) with both capture methods (G = 104.83; p < 0.0001; d.f. = 1). 81 out of 95 photographic contacts of wood mouse were recorded in shrublands (85.3%) and the remainder 14 contacts in pinewoods (14.7%), and 78 out of 90 captures with Sherman traps were obtained in shrublands (86.7%), and the remainder 12 in pinewoods (13.3%). Remote photography showed the same proportions of captures than live-trapping in the two habitats studied (G = 0.07; p = 0.79; d.f. = 1). Temporal activity patterns of wood mice were biphasic during winter in shrublands, with lower values of activity near sunset and sunrise, and maximum four hours after sunset anf five hours before sunrise. We conclude that remote photography, despite pseudoreplication (impossibility to diferentiate individuals in the photographs), can be considered a valid method to estimate wood mice abundance and activity patterns in the post-fire habitats studied.

The impacts of forest clearance on lizard, small mammal and bird communities in the arid spiny forest, southern Madagascar

Madagascar is a global biodiversity hotspot threatened by forest loss, degradation and fragmentation, all of which are detrimental to the future survival of forest-dwelling organisms. For conservation purposes it is essential to determine how species respond to habitat disturbance, specifically deforestation. In this study we investigated the impacts of deforestation on three vertebrate communities, lizards, small mammals and birds, in an area of spiny forest subjected to anthropogenic forest clearance. Spiny forest has high levels of endemism, but conservation in this unique ecosystem is hindered by the lack of research. We undertook standardised trapping, time-constrained and timed species searches to assess species richness, species abundance and community composition of lizards, small mammals and birds in six areas of ‘forest’ and six ‘cleared’ areas. From surveys and opportunistic sightings we recorded a total of 70 species of birds, 14 species of mammals and 38 species of reptiles and amphibians. We found forest clearing to have a negative effect on species richness and community structure of all groups and identified loss of canopy cover as a driving factor behind this. However, the response and sensitivity to clearing varied between groups and species. Lizards (50%) and small mammals (40%) had the greatest decline in species richness in response to clearing as compared to birds (26%), although birds showed the greatest shift in community structure. The community in cleared areas contained more generalist and introduced species that have wider geographic ranges and habitat preferences, than those unique to the spiny forest. We found the first suite of species to suffer from forest clearance were those of high conservation priority due to their restricted geographic range. Our findings are discussed in relation to future spiny forest conservation and management.

MAMMALS, EDGE EFFECTS, AND THE LOSS OF TROPICAL FOREST DIVERSITY

Relative to the surrounding mainland forests, a subset of tree species dom- inates wind-exposed, forested islands in Gatin Lake, Panama. We explored how tree di- versity in these fragments has been affected by (1) impoverishment of the mammal com- munity and (2) changes in abiotic conditions following island formation ca. 90 years ago. To test effects of changes in the mammal community, we assessed seed and seedling survival for nine tree species in five forests: small islands with no mammals; small islands with spiny rats but no larger mammals; medium islands (intermediate mammal community); Barro Colorado Island (intermittently present puma and jaguar); and mainland forests (intact mammal community). To test effects of abiotic stress, we chose experimental sites at wind- exposed, wind-protected, and interior forest sites. We predicted that fragments with less diverse mammal communities would be characterized by (1) fewer seeds dispersed and cached, (2) lower long-term seed survival, and (3) higher rates of seedling herbivory by mammals. Where alteration of the environment has caused greater exposure to dry-season winds, we predicted that (4) germination and seedling establishment and (5) dry-season seedling survival would be low. Further, we expected that (6) dry-season seedling survival would increase if soil moisture levels were raised, but that (7) wet-season seedling survival is independent of wind exposure. In the larger forests, seed and seedling survival were low, and differences in mammal community composition had little effect. Small islands that supported only rats had the same low seed survival as larger forests. Elimination of all mammals resulted in much higher seed survival, and slightly higher seedling survival. Germination and seedling sur- vival were lower at wind-exposed than at protected sites, but irrigation had no effect on dry-season seedling survival. It appears that the dominant species on Gatdin Lake fragments have passed through a series of filters: seeds must first avoid being eaten by spiny rats, survive the first dry season, and then avoid seedling herbivory by rats. Biotic and abiotic factors determine seedling survival in tropical forests; synergistic changes in these factors can thus lead to dramatic losses of diversity.

THREE METHODS FOR ASSESSING RICHNESS AND COMPOSITION OF SMALL MAMMAL COMMUNITIES

Analysis of pellets of Tyto alba and scats of Genetta genetta and Shennan livetrapping were compared to assess richness and composition of small mammal communities in a Mediterranean area (NE Spain). Owl pellets provided 17 small mammal species (17,232 individuals), genet scats 14 species (2,145 individuals), and livetrapping 9 species (1,488 individuals). Owl pellets oversampled insectivores and grassland rodents and undersampled tree-dwelling and woodland rodents. Genet scats and livetrapping overs amp led woodland rodents and undersampled insectivores and grassland rodents. After controlling for sample size and elevation differences between methods by means of analysis of covariance (ANCOV A) and rarefaction, owl pellets contained higher richness for small samples(<50 individuals), and scats contained higher richness for large samples (> 100 individuals), both having higher richness than livetrapping regardless of sample size. We concluded that both indirect methods provided complementary infonnation of small mammal communities, detecting the 19 small mammal species known to be present in the study area.

The bioclimatic model: a method of palaeoclimatic qualitative inference based on mammal associations

ABSTRACTAim The bioclimatic model is a new method for palaeoclimatic reconstruction built on the assumption of a significant correlation between climate and mammal community composition. The goal of this approach is to infer past climatic conditions using mammal fossil associations as source data.Location The study used mammal faunas from all over the world to develop the bioclimatic model. As an example of the potential of the model, we have applied it to Quaternary faunas from Eurasia.Methods The proposed model was constructed by applying multivariate discriminant analysis to modern mammal faunas and climates from throughout the world. The model was validated with a different set of modern faunas than those used in the discriminant analysis, including some from transitional zones between different climates (ecotones). To test the reliability of the method in the Pleistocene, the results have been compared to those obtained with data from other disciplines, such as palaeobotany.Results The results obtained in the validation of the model show that more than 90% of the localities have been classified correctly. Comparisons of results in the late Pleistocene‐Holocene of Barová between a palaeobotanical study and the bioclimatic analysis show the latter to be highly accurate. The results for early Pleistocene faunas show somewhat drier and more open climatic conditions for Europe than the present day, with larger areas of steppe environments.Main Conclusions The bioclimatic model could be used to infer climatic conditions from mammal faunas. The results presented in this work provide a preliminary example of the potential that bioclimatic analysis has as a tool for palaeoclimatic inference. Finally, this method offers the opportunity to standardize data coming from vertebrate palaeontology for use in the construction and evaluation of climatic models.

Small mammals in young forests: implications for management for sustainability

Small mammals have been proposed as indicators of sustainability in forests in the Pacific Northwest and elsewhere. Mammal community composition and species abundances purportedly result from interactions among species, forest-floor characteristics, large coarse woody debris, understory vegetation, and overstory composition. Coarse woody debris is thought to be particularly important because of its diverse ecological functions; covers from 10 to 15% have been recommended based on retrospective studies of forests and small mammals. Unfortunately, ecological correlations are not necessarily indicative of causal relationships and magnitudes depend on composition of finite, usually non-random, cross-sectional samples. Retrospective studies must be replicated to confirm relationships. We conducted a large-scale, cross-sectional survey of 30- to 70-year-old coniferous forests in western Washington to determine if previously reported relationships would hold with an unrelated, larger sample. Coarse woody debris cover was 8.3±0.6% ( x ̄ ± S.E. , n=8 blocks of forest, range 4–13%). Understory cover was too low (18±8% for shrubs) to allow examining interactions between understory and coarse woody debris. Overstory composition covaried with coarse woody debris. One or two of four statistically extracted habitat factors (overstory composition, herbaceous cover, abundance of Acer circinatum, and abundance of Acer macrophyllum) accounted for 18–70% of variance in abundance of 11 mammal species. Our results support hypotheses that: (1) biocomplexity resulting from interactions of decadence, understory development, and overstory composition provides pre-interactive niche diversification with predictable, diverse, small-mammal communities; (2) these communities incorporate numerous species and multiple trophic pathways, and thus, their integrity measures resiliency and sustainability.

Pleistocene and. Holocene Records of Antilocapra Americana: A Review of the FAUNMAP Data

FAUNMAP is an electronic database documenting the late Quaternary (Pleistocene and Holocene) distribution of mammal species in the United States, developed at the Illinois State Museum with support from the National Science Foundation. The primary purpose of the database is to investigate evolution of mammalian communities, although individual species distributions are readily examined. With a Geographic Information System (GIS), changes in the distributions of individual species and their effects upon mammal community composition can be documented for the late Quaternary. As of 1994, it included data from 2919 sites in the contiguous 48 states covering the last 40,000 years. The database is highly incomplete and does not represent the entire locality distribution of species, primarily because only a select few cultural resource management reports were included. The FAUN MAP database was queried for Antilocapra americana, the sole living representative of a once-extensive family of pecoran artiodactyls. GIS maps were generated showing the distribution of Antilocapra americana from the Wisconsinan through the Holocene and the modern extant range. These maps reveal Antilocapra americana has been consistently present throughout what early twentieth century mammalogists consider the species’ historic range, with only an occasional locality outside those boundaries. These latter localities can be correlated with short-term shifts in the distribution of the short-grass prairie eastern border or slight changes in the western limits of the species range.

Historical examination of the status of large mammals in Aragon, Spain

Summary. - During the last century, human activities have promoted large changes in habitats, leading to changes in the abundance and composition of mammal communities. Some species have been favoured intentionally (specially those with high economic importance), or unintentionally (Le. opportunistic species), introduced (as some livestock species), or persecuted (Le. species that affect livestock or game). In this paper, we show and explain the changes in the composition of the large mammal community in Aragon, a northeastern Spanish area where there have been extinctions (lynx, Lynx sp.); introductions (moufflon, Ovis ammon; fallow deer, Dama dama); reintroductions (red deer, Cervus elaphus); and particularly natural expansions (wild boar, Sus scrota ; roe deer, Capreolus capreolus ; red deer; Pyrenean chamois, Rupicapra pyrenaica pyrenaica). Some taxa are currently almost extinct (brown bear, Ursus arctos; Pyrenean ibex, Capra pyrenaica pyrenaica) and others are recovering after getting close to extinction (wolf, Canis lupus). Three interconnected causes may explain the clear increment of ungulates in our study area: (1) the abandonment of rural activities, (2) the subsequent decrease of livestock, and (3) the increase of forest habitats. None of those species are currently below 150 % of their range in the 19th century. The setting up of hunting reservations and the design of a hunting legislation have also contributed to the observed changes. Traditionally, large predators have not been favoured by human activities, while compensatory protection measures carried out have been scarce or came too late. While some of these species are probably extinct in the area, the remaining occupy currently less than one fourth of their former range and need urgent conservation measures for to recover.

Small mammal abundance and community composition in Prince Edward Island National Park

Small mammal abundance and community composition in Prince Edward Island National Park

Small mammal communities and fire in the Brazilian Cerrado

AbstractThe Cerrado biome is a complex of open forests, gallery forests and savanna vegetation that harbours about 60 species of small mammals. Despite the frequent occurrence of fire in this habitat, few studies have analysed its effects on the small mammal communities in the Cerrado. The purpose of this study was to compare small mammal communities of two Cerrado areas with different fire frequencies (2–3 years or more than 18 years since the last fire). Species diversity, species composition, and relative abundance of the most common species of these communities were analysed in both areas before and after fire. One trapping grid was established in each of the two areas. These areas were sampled monthly or bi‐monthly during 1 year before and 1 year after prescribed fires. Both areas had communities with similar species composition, but the protected site had more species, before and after fire (seven and eight, respectively), than the frequently burned site (five and six species, respectively). Both sites showed significant post‐fire changes for a short time period (3 months), but only the area that had been protected against fire for more time showed significant post‐fire changes in relative abundance of the most common species (the murid rodents Bolomys lasiurus, Calomys callosus, C. tener, Oryzomys subflavus, and Thalpomys cerradensis) for the 1 year following fire. Overall diversity tended to be greater in the protected site both before and after fire. Species composition and diversity indices of the small mammal communities were relatively unaffected by the fires. Thalpomys cerradensis, however, appeared to use only areas that burn frequently and C. callosus showed an immediate post‐fire increase in population size in both areas, contrasting with a simultaneous decrease in number of B. lasiurus.

Small mammal communities and fire in the Brazilian Cerrado

The Cerrado biome is a complex of open forests, gallery forests and savanna vegetation that harbours about 60 species of small mammals. Despite the frequent occurrence of fire in this habitat, few studies have analysed its effects on the small mammal communities in the Cerrado. The purpose of this study was to compare small mammal communities of two Cerrado areas with different fire frequencies (2–3 years or more than 18 years since the last fire). Species diversity, species composition, and relative abundance of the most common species of these communities were analysed in both areas before and after fire. One trapping grid was established in each of the two areas. These areas were sampled monthly or bi-monthly during 1 year before and 1 year after prescribed fires. Both areas had communities with similar species composition, but the protected site had more species, before and after fire (seven and eight, respectively), than the frequently burned site (five and six species, respectively). Both sites showed significant post-fire changes for a short time period (3 months), but only the area that had been protected against fire for more time showed significant post-fire changes in relative abundance of the most common species (the murid rodents Bolomys lasiurus, Calomys callosus, C. tener, Oryzomys subflavus, and Thalpomys cerradensis) for the 1 year following fire. Overall diversity tended to be greater in the protected site both before and after fire. Species composition and diversity indices of the small mammal communities were relatively unaffected by the fires. Thalpomys cerradensis, however, appeared to use only areas that burn frequently and C. callosus showed an immediate post-fire increase in population size in both areas, contrasting with a simultaneous decrease in number of B. lasiurus.

Deer Mouse Predation on the Biological Control Agent, Urophora spp., Introduced to Control Spotted Knapweed

Field observations made in 1993 suggested that rodents were preying on spotted knapweed (Centaurea maculosa) seedheads, possibly targeting the gall fly larvae (Urophora SPP.) which overwinter within them. I conducted a brief study to determine the cause of seedhead predation and quantify gall fly predation. Stomachs were examined from 19 deer mice (Pero- myscus maniculatus) captured in the fall of 1993 and winter of 1997. All individuals had preyed upon gall fly larvae. The mean number of gall fly larvae found in 10 deer mouse stomachs in the winter of 1997 was 212.8. The minimum number of larvae consumed by these 10 animals for 1 night of foraging was 2686. Availability of a concentrated protein that is a readily accessible and abundant resource during winter may elevate deer mouse populations in knapweed-in- fested habitats. Increases in densities of deer mice due to gall fly presence could bring about shifts in composition of small mammal communities.

The structure of small mammal communities in some alpine habitats

Abstract We studied the composition of several small mammal communities living in different mountain and forest habitats of the central eastern Italian Alps. The small mammals were then grouped together, by cluster analysis, according to similarities in species and density. From the 22 stations investigated, five groups emerged, each one having also distinct environmental characteristics. We observed that spruce forest communities are grouped separately from those of mixed forests (larch and Swiss stone pine). We must stress the considerable difference existing between the small mammal communities living in different kinds of coniferous forests. The larch and Swiss stone pine forest seem to be able to support a greater density of small mammals, which includes in particular the bank vole ( Clethrionomys glareolus ).

Does Mammal Community Composition Control Recruitment in Neotropical Forests? Evidence from Panama

Patterns of seed predation, germination, and seedling herbivory were investigated in Panamanian forests. We hypothesized that seed and seedling survival would vary with differences in mammal community composition. We tested this hypothesis at five sites in mainland forests adjacent to Gatun Lake, full terrestrial mammalian granivore/herbivore communities with top predators; at five sites on Barro Colorado Island (BCI), also a full mammalian granivore/herbivore community but without the two largest cats; at one site each on five medium-sized islands, with rats, agouti, rabbit, and paca present; and on five small islands that support rats only. Experiments were replicated for Dipteryx panamensis, Gustavia superba, and Virola nobilis, all of which have large seeds. To assess seed removal, seeds were placed in wire exclosure cages and nearby outside the cages. There was no difference in removal rates between forest types, with almost all unprotected seeds removed at all sites. To assess post-removal seed fate, seeds of Gustavia and Virola were attached to threads and placed on the forest floor. All threaded seeds were victims of predation on small islands, whereas 34, 43, and 77% of threaded seeds were dispersed and buried on BCI, medium islands, and the mainland, respectively. To assess seedling herbivory, half of the wire exclosure cages were removed after germination, and seedling survival was assessed after 13–14 mo. Protection from mammals increased seedling survivorship by more than sixfold on the smallest islands, by threefold on the medium islands, by twofold on the mainland, and by less than twofold on BCI. The absence of the two largest cats and the exclusion of poachers from BCI was associated with lower seedling herbivory and higher seed predation than observed on the mainland. In contrast, extreme mammal defaunation on the small and medium islands had large and consistent effects on seedling recruitment, including increased seed predation and increased seedling herbivory relative to sites with more intact mammal communities.