Optimising sampling methods for small mammal communities in Neotropical rainforests

相似文献   

Effects of alien pine plantations on small mammal community structure in a southern African biodiversity hotspot

Commercial plantations and alien tree invasions often have substantial negative impacts on local biodiversity. The effect of plantations on faunal communities in the fire‐adapted fynbos vegetation of the Cape Floristic Region biodiversity hotspot is not yet well quantified. We studied small mammal community structure in alien Pinus radiata plantations and adjacent fynbos regenerating after clear‐felling of plantations on the Cape Peninsula, South Africa. Small mammal sampling over 1,800 trap‐nights resulted in 480 captures of 345 individuals (excluding recaptures) representing six species. Significantly more species, individuals (12 X) and biomasses (29 X) of small mammals occurred on recovering fynbos sites compared to plantations. This was commensurate with a higher diversity of plant growth forms, vegetation densities and live vegetation biomass. Only one small mammal species, the pygmy mouse (Mus minutoides), was consistently trapped within plantations. Fynbos sites were dominated by three small mammal species that are ecological generalists and early successional pioneer species, rendering the recovering fynbos slightly depauperate in terms of species richness and evenness relative to other studies done in pristine fynbos. We make three recommendations for forestry that would facilitate the restoration of more diverse natural plant communities and progressively more diverse and dynamic small mammal assemblages in a key biodiversity hotspot.     

Effects of roads on small mammal diversity and abundance in the northern Serengeti,Tanzania

The high biodiversity of small mammal species in the Serengeti ecosystem provides this ecosystem with important conservation value. However, whether the extensive development of roads has negative impacts on the small mammal population has not been tested. Small mammal population diversity and abundance were examined in this study using live trapping at sites close to (experimental) and away (control) from the main gravel road during the short rainy seasons in November and December 2011 and 2012. A total of 138 individuals from three orders representing six families and fourteen species were collected over 4,860 trap nights. There were no significant differences in the species richness, diversity or abundance of small mammals between the control and experimental sites (p > 0.05), suggesting that the current gravel road does not have a significant impact on the small mammal population. These findings were ascribed to the availability of favourable habitats at both distances as a result of little road usage due to poor conditions. Should the road be improved, the control of anthropogenic activities in the area should be given high priority. Continuous monitoring of the small mammal populations in the area is recommended.     

Indications of higher diversity and abundance of small rodents in human‐influenced Sudanian savannah than in the Niokolo Koba National Park (Senegal)

Expansion of human activities in the Sudanian savannah is changing natural habitats together with their animal communities. The abundance and diversity of small mammal communities were investigated in human‐influenced savannahs in close proximity to Niokolo Koba National Park (NKNP) in south‐eastern Senegal and compared with those from NKNP. Two localities were sampled in natural and two in human‐influenced areas during two dry and two rainy seasons. Total sampling effort was 5400 trap‐nights. Ninety‐eight individuals of eleven species were captured outside NKNP, while only 28 individuals of five rodent species were caught inside the park (all of which were found also outside NKNP). The most abundant species outside NKNP was Mastomys erythroleucus, while inside NKNP it was Praomys daltoni. Both relative abundance and diversity were significantly higher in human‐influenced landscape than in natural savannahs of NKNP. The difference in species richness may be linked with lower population densities of certain species in NKNP and the effect of traditional agriculture, which may support the presence of species typical for deforested landscape. Furthermore, food availability, and predator abundances are discussed as possible reasons for the higher rodent abundances outside NKNP.     

Life on high: the diversity of small mammals at high altitude in South Africa

The Great Escarpment is the major mountain system in South Africa, yet very few biological surveys have been conducted outside of the well-known Drakensberg section. This is surprising given the important role that mountains play in local and global biodiversity patterns. In this study, small mammal diversity and community composition were estimated at three high altitude (>1,700?m) sites within the Sneeuberg Mountain Complex (SMC) of the Great Escarpment, South Africa from June 2009 to May 2010. The influences of selected environmental variables on diversity were also tested. Of 423 live-captures, a total of 292 unique individuals of 12 small mammal species (one shrew, one elephant shrew and 10 rodents) were identified during 5,280 trap nights. No single environmental variable could account for the variation observed in diversity measurements but vegetation height appeared to be the most important factor to influence the number of individuals captured. It is hypothesised that the high species richness and diversity of small mammals observed in the SMC compared to other parts of the Great Escarpment is due to the SMC being located in a transition zone of the Grassland and Nama-Karoo biomes. Our results suggest that the SMC could be important in conserving small mammal species from western and eastern assemblages across South Africa.     

Small mammal trapping in tropical montane forests of the Upper Nilgiris, southern India: An evaluation of capture-recapture models in estimating abundance

Capture-mark-recapture was used to study small mammal populations in tropical montane forests in southern India. Eleven plots in six montane forest patches were sampled from February-October, 1994. Six species were captured, including four rodents and two shrews. PROGRAM CAPTURE was used to derive estimates of density of the most abundant species in the study area,Rattus rattus Linnaeus. The coefficient of variation of the density estimate was used as an index of precision. The coefficient of variation decreased exponentially with increasing capture probability and with an increase in trapping duration. The coefficient of variation and the capture probability were not correlated with estimates of density. The density estimate increased with trapping duration, as did trap mortality. The latter may have been due to the trend of increased mortality with recaptures of the same individual, which in turn may have been due to weight loss over consecutive captures. Estimates of density derived using four estimators were different for 2, 3, 4 and 5 days of trapping. The coefficient of variation was highest for the generalized removal estimate and lowest for Darroch’s estimate. The models and estimators could not be applied to more than one species, and for this species, only in select habitats in a few seasons. Therefore, models of density estimation developed for temperate areas may not be suitable for tropical habitats due to low densities of small mammals in these habitats.     

Influence of habitat characteristics on small mammals in a Mexican high-altitude grassland

Small mammals in a high-altitude grassland area close to Mexico City were studied. Populations of 10 species were censused using live traps in 48 sample quadrats. Within each quadrat, vegetation characterization, including complete floristic listings, cover values for species and layers and values of habitat modification, were assessed. Habitats were described according to plant communities identified using ordination and classification methods. Nine different plant communities were obtained. Densities and abundance of all small mammal species were calculated for each of the habitats classified. Peromyscus alsloni was the most abundant species in all habitats, reaching maximum densities of 55 ha⁻¹ in pine forest with dense ground and herb layer. Peromyscus melanotis also occurred in all habitats but at lower densities (maximum 29 ha⁻¹). Reithrodontomys megalotis was found in all habitats except in tall dense grassland. Densities for this species were generally low (1-9 ha⁻¹) but reached 19 ha⁻¹ in short dense grassland. All other species were largely absent from 4–8 habitats and showed very low densities (0.75–4 ha⁻¹). The densities of the more abundant species were largely correlated with more open habitats and higher indices of habitat modification. Lower altitude grassland habitats have a greater abundance of small mammals and a higher species richness than the medium and higher altitude, physiognomically more complex habitats. Species richness was highest in tall pine-alder forest with a species-rich, dense herb layer and lowest in pine forest with dense ground and herb layers. Species richness was positively correlated with overall small mammal density.     

Interspecies interference and monitoring duration affect detection rates in chew cards

Pest monitoring methods should provide unbiased accurate estimates of pest densities and locations, while also minimizing time‐in‐field and costs. Recent pest mammal monitoring studies have found that chew cards are more effective than conventional mammal monitoring methods, but little experimental work has been done to determine optimal experimental duration or quantify the risks of saturation by one species biasing detections of other species. Here, we used chew cards in three sites within Awarua wetland (Southland, New Zealand) to investigate the optimal amount of time required to detect targeted pest species (rats, possums and mice), and to examine the potential of rats and possums to bias detection rates of other species. We found depressed detections of possums and rats where a contraspecific had been detected on a card, which is consistent with previous studies of a similar duration on interspecies interference. This experiment is the first to analyse the rates at which species detections accrue over the course of a survey, and we found rat detections lagged behind possums for the first four nights. We modelled the effect of survey duration and relative rat abundance on the likelihood of further possum detections. Duration and rat abundance interacted, meaning there are trade‐offs to be considered with regard to duration: shorter durations may avoid the risk of saturation in areas of high pest density, but risk not sampling sparse or neophobic populations. Our data suggest that chew cards remain one of the most sensitive pest monitoring tools for rats and possums, compared to conventional methods such as tracking tunnels and wax tags. In areas of moderate pest densities, we suggest that a duration of five nights is optimal for detecting pests. However, in areas of high pest density the sensitivity of chew cards may render them unsuitable because of saturation and interspecies interference effects.     

Mammal density and patterns of ectoparasite species richness and abundance

Patterns of species richness, prevalence and abundance of ectoparasites have rarely been investigated at both the levels of populations and species of hosts. Here, we investigated the effects in changes in small mammal density on species richness, abundance and prevalence of ectoparasitic fleas. The comparative analyses were conducted for different small mammal species and among several populations during a long-term survey. We tested the hypothesis that an increase in host density should be linked with an increase in parasite species richness both among host species and among populations within host species, as predicted by epidemiological models. We also used host species density data from literature. We found that host density has a major influence on the species richness of ectoparasite communities of small mammals among host populations. We found no relationship between data of host density from the literature and parasite species richness. In contrast with epidemiological hypotheses, we found no relationships between abundance, or prevalence, and host density, either among host species or among host populations. Moreover, a decrease in abundance of fleas in relation with an increase in host density was observed for two mammal species (Apodemus agrarius and A. flavicollis). The decrease or the lack of increase in flea abundance in relation with an increase in host density suggests anti-parasitic behavioural activities such as grooming.     

Ecological correlates of mammal β‐diversity in Amazonian land‐bridge islands: from small‐ to large‐bodied species

Aim

Mega hydroelectric dams have become one of the main drivers of biodiversity loss in the lowland tropics. In these reservoirs, vertebrate studies have focused on local (α) diversity measures, whereas between‐site (β) diversity remains poorly assessed despite its pivotal importance in understanding how species diversity is structured and maintained. Here, we unravel the patterns and ecological correlates of mammal β‐diversity, including both small (SM) and midsized to large mammal species (LM) across 23 islands and two continuous forest sites within a mega hydroelectric reservoir.

Location

Balbina Hydroelectric Dam, Central Brazilian Amazonia.

Methods

Small mammals were sampled using live and pitfall traps (48,350 trap‐nights), and larger mammals using camera traps (8,160 trap‐nights). β‐diversity was examined for each group using multiplicative diversity decomposition of Hill numbers, which considers the importance of rare, common and dominant species, and tested to what extent those were related to a set of environmental characteristics measured at different spatial scales.

Results

β‐diversity for both mammal groups was higher when considering species presence–absence. When considering species abundance, β‐diversity was significantly higher for SM than for LM assemblages. Habitat variables, such as differences in tree species richness and percentage of old‐growth trees, were strong correlates of β‐diversity for both SMs and LMs. Conversely, β‐diversity was weakly related to patch and landscape characteristics, except for LMs, for which β‐diversity was correlated with differences in island sizes.

Main conclusions

The lower β‐diversity of LMs between smaller islands suggests subtractive homogenization of this group. Although island size plays a major role in structuring mammal α‐diversity in several land‐bridge islands, local vegetation characteristics were additional key factors determining β‐diversity for both mammal groups. Maintaining the integrity of vegetation characteristics and preventing the formation of a large set of small islands within reservoirs should be considered in long‐term management plans in both existing and planned hydropower development in lowland tropical forests.

     

Land Use and Small Mammal Predation Effects on Shortgrass Prairie Birds

ABSTRACT Grassland birds endemic to the central shortgrass prairie ecoregion of the United States have experienced steep and widespread declines over the last 3 decades, and factors influencing reproductive success have been implicated. Nest predation is the major cause of nest failure in passerines, and nesting success for some shortgrass prairie birds is exceptionally low. The 3 primary land uses in the central shortgrass prairie ecoregion are native shortgrass prairie rangeland (62%), irrigated and nonirrigated cropland (29%), and Conservation Reserve Program (CRP, 8%). Because shortgrass–cropland edges and CRP may alter the community of small mammal predators of grassland bird nests, I sampled multiple sites on and near the Pawnee National Grasslands in northeast Colorado, USA, to evaluate 1) whether small mammal species richness and densities were greater in CRP fields and shortgrass prairie–cropland edges compared to shortgrass prairie habitats, and 2) whether daily survival probabilities of ground-nesting grassland bird nests were negatively correlated with densities of small mammals. Small mammal species richness and densities, estimated using trapping webs, were generally greater along edges and on CRP sites compared to shortgrass sites. Vegetation did not differ among edges and shortgrass sites but did differ among CRP and shortgrass sites. Daily survival probabilities of artificial nests at edge and CRP sites and natural nests at edge sites did not differ from shortgrass sites, and for natural nests small mammal densities did not affect nest survival. However, estimated daily survival probability of artificial nests was inversely proportional to thirteen-lined ground squirrel (Spermophilus tridecemlineatus) densities. In conclusion, these data suggest that although land-use patterns on the shortgrass prairie area in my study have substantial effects on the small mammal community, insufficient data existed to determine whether land-use patterns or small mammal density were affecting grassland bird nest survival. These findings will be useful to managers for predicting the effects of land-use changes in the shortgrass prairie on small mammal communities and avian nest success.     

Performance of ground-based and aerial survey methods for monitoring wildlife assemblages in a conservation area of northern Tanzania

Validating and improving field-sampling techniques for estimating wildlife community composition and population size is essential for wildlife management and conservation. We conducted ground distance sampling surveys along line transects and block counts from a small aircraft in Manyara Ranch in Northern Tanzania and contrasted estimates of species richness and species-specific densities from both sampling techniques. We used regression analyses (logistic regression and generalized linear mixed models) and model selection to investigate whether a species’ body size, group size, body color, as well as vegetation cover explained the variation in species presence/absence and relative density differences in aerial vs. ground-based sampling. Ground surveys detected significantly more species than aerial surveys. However, aerial surveys detected three species that were missed by ground surveys (African lions, African buffalo, and spotted hyena). Model selection suggested that species with smaller body mass and small group sizes were more likely to be missed in aerial surveys. Densities estimated from the aerial surveys were generally but non-significantly lower than the densities estimated from the ground surveys, with the exception of density estimates for African elephants which were slightly higher from aerial surveys. Density differences between the two methods were greater for species with small group size, light body color, and in areas with denser vegetation cover; these variables explained 75% of the variation in density differences between the two survey methods. Albeit being similar in operational costs in our relatively small study area, ground surveys yielded (1) more complete information with respect to wildlife community composition and (2) density estimates were mostly higher and (3) more precise and (4) appear more feasible to be implemented in community-based conservation schemes.     

Community‐based wildlife management area supports similar mammal species richness and densities compared to a national park

Community‐based conservation models have been widely implemented across Africa to improve wildlife conservation and livelihoods of rural communities. In Tanzania, communities can set aside land and formally register it as Wildlife Management Area (WMA), which allows them to generate revenue via consumptive or nonconsumptive utilization of wildlife. The key, yet often untested, assumption of this model is that economic benefits accrued from wildlife motivate sustainable management of wildlife. To test the ecological effectiveness (here defined as persistence of wildlife populations) of Burunge Wildlife Management Area (BWMA), we employed a participatory monitoring approach involving WMA personnel. At intermittent intervals between 2011 and 2018, we estimated mammal species richness and population densities of ten mammal species (African elephant, giraffe, buffalo, zebra, wildebeest, waterbuck, warthog, impala, Kirk's dik‐dik, and vervet monkey) along line transects. We compared mammal species accumulation curves and density estimates with those of time‐matched road transect surveys conducted in adjacent Tarangire National Park (TNP). Mammal species richness estimates were similar in both areas, yet observed species richness per transect was greater in TNP compared to BWMA. Species‐specific density estimates of time‐matched surveys were mostly not significantly different between BWMA and TNP, but elephants occasionally reached greater densities in TNP compared to BWMA. In BWMA, elephant, wildebeest, and impala populations showed significant increases from 2011 to 2018. These results suggest that community‐based conservation models can support mammal communities and densities that are similar to national park baselines. In light of the ecological success of this case study, we emphasize the need for continued efforts to ensure that the BWMA is effective. This will require adaptive management to counteract potential negative repercussions of wildlife populations on peoples' livelihoods. This study can be used as a model to evaluate the effectiveness of wildlife management areas across Tanzania.     

Random versus Game Trail-Based Camera Trap Placement Strategy for Monitoring Terrestrial Mammal Communities

Camera trap surveys exclusively targeting features of the landscape that increase the probability of photographing one or several focal species are commonly used to draw inferences on the richness, composition and structure of entire mammal communities. However, these studies ignore expected biases in species detection arising from sampling only a limited set of potential habitat features. In this study, we test the influence of camera trap placement strategy on community-level inferences by carrying out two spatially and temporally concurrent surveys of medium to large terrestrial mammal species within Tanzania’s Ruaha National Park, employing either strictly game trail-based or strictly random camera placements. We compared the richness, composition and structure of the two observed communities, and evaluated what makes a species significantly more likely to be caught at trail placements. Observed communities differed marginally in their richness and composition, although differences were more noticeable during the wet season and for low levels of sampling effort. Lognormal models provided the best fit to rank abundance distributions describing the structure of all observed communities, regardless of survey type or season. Despite this, carnivore species were more likely to be detected at trail placements relative to random ones during the dry season, as were larger bodied species during the wet season. Our findings suggest that, given adequate sampling effort (> 1400 camera trap nights), placement strategy is unlikely to affect inferences made at the community level. However, surveys should consider more carefully their choice of placement strategy when targeting specific taxonomic or trophic groups.     

Elevational gradients of small mammal diversity on the northern slopes of Mt. Qilian, China

Aim Small mammal species richness and relative abundance vary along elevational gradients, but there are different patterns that exist. This study reports the patterns of distribution and abundance of small mammals along the broader elevational gradient of Mt. Qilian range. Location The study was conducted in the Mt. Qilian range, north‐western China, from June to August 2001. Methods Removal trapping was conducted using a standardized technique at 7 sites ranging between 1600 and 3900 m elevation within three transects. Correlation, regression and graphical analyses were used to evaluate the diversity patterns along this elevational gradient. Results In total, 586 individuals representing 18 nonvolant small mammal species were collected during 20 160 trap nights. Species composition was different among the three transects with 6 (33%) of the species found only within one transect. Elevational distribution and relative abundance of small rodents showed substantial spatial variation, with only 2 species showing nonsignificant capture frequencies across elevations. Despite these variations, some general patterns of elevational distribution emerged: humped‐shape relationships between species diversity and elevation were noted in all three transects with diversity peaks at middle elevations. In addition, relative abundance was negatively correlated with elevation. Conclusions Results indicate that maximum richness and diversity of nonvolant small mammals occurred at mid‐elevations where several types of plants reached their maximum diversity and primary productivity, and where rainfall and humidity reached a maximum. It is demonstrated that the mid‐elevation bulge is a general feature of at least a large portion of the biota on the Mt. Qilian range.     

Mammal population densities at a global scale are higher in human‐modified areas

Global landscapes are changing due to human activities with consequences for both biodiversity and ecosystems. For single species, terrestrial mammal population densities have shown mixed responses to human pressure, with both increasing and decreasing densities reported in the literature. How the impacts of human activities on mammal populations translates into altered global density patterns remains unclear. Here we aim to disentangle the effect of human impacts on large‐scale patterns of mammal population densities using a global dataset of 6729 population density estimates for 468 mammal species (representing 59% and 44% of mammalian orders and families). We fitted a mixed effect model to explain the variation in density based on a 1‐degree resolution as a function of the human footprint index (HFI), a global proxy of direct and indirect human disturbances, while accounting for body mass, trophic level and primary productivity (normalized vegetation index; NDVI). We found a significant positive relationship between population density and HFI, where population densities were higher in areas with a higher HFI (e.g. agricultural or suburban areas – no populations were located in very high HFI urban areas) compared to areas with a low HFI (e.g. wilderness areas). We also tested the effect of the individual components of the HFI and still found a consistent positive effect. The relationships remained positive even across populations of the same species, although variability among species was high. Our results indicate shifts in mammal population densities in human modified landscapes, which is due to the combined effect of species filtering, increased resources and a possible reduction in competition and predation. Our study provides further evidence that macroecological patterns are being altered by human activities, where some species will benefit from these activities, while others will be negatively impacted or even extirpated.     

Untangling the roles of fire,grazing and rainfall on small mammal communities in grassland ecosystems

In grassland systems across the globe, ecologists have been attempting to understand the complex role of fire, grazing and rainfall in creating habitat heterogeneity and the consequences of anthropogenic control of these factors on ecosystem integrity and functioning. Using a South African grassland ecosystem as a model, we investigated the impact of fire and grazing pressure on small mammal communities during three differing periods of a rainfall cycle. Over 2 years, 15,203 trap nights revealed 1598 captures of 11 species (nine rodents, one macroscelid and one insectivore). Results highlighted the importance of the interplay between factors and showed that the role of fire, grazing and rainfall in determining small mammal abundance was species-dependant. While no two species were affected by the same environmental variables, grass cover or height was important to 56% of species. Considered independently, high rainfall had a positive influence on small mammal abundance and diversity, although the lag period in population response was species-specific. High grazing negatively affected overall abundance, but specifically in Mastomys coucha; fire alone had little immediate impact on small mammal diversity. Six months after the fire, vegetation cover had recovered to similar levels as unburned areas, although small mammal diversity and richness were higher in burned areas than unburned areas. Grazing levels influenced the rate of vegetation recovery. In conclusion, low-level grazing and burning can help to maintain small mammal biodiversity, if conducted under appropriate rainfall levels. A too high grazing pressure, combined with fire, and/or fire conducted under drought conditions can have a negative impact on small mammal biodiversity. To maintain small mammal diversity in grassland ecosystems, the combined effects of the previous year’s rainfall and existing population level as well as the inhibition of vegetation recovery via grazing pressure need to be taken into consideration before fire management is applied. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Studying Large Mammals With Imperfect Detection: Status and Habitat Preferences of Wild Cattle and Large Carnivores in Eastern Cambodia

Studying large mammal species in tropical forests is a conservation challenge with species’ behavior and ecology often increasing the probability of non‐detection during surveys. Consequently, knowledge of the distribution, status, and natural history of many large mammal species in Southeast Asia is limited. I developed occupancy models from camera‐trapping data, thereby accounting for imperfect detection at sampling sites, to clarify the status and habitat requirements of four globally threatened or near threatened large mammals (banteng Bos javanicus, gaur Bos gaurus, dhole Cuon alpinus, and leopard Panthera pardus) in Mondulkiri Protected Forest, eastern Cambodia. Camera traps were operational for >3500 trap nights with 202 photographic encounters of the four study species. Model averaged occupancy estimates were between 5 percent (leopard) and 140 percent (gaur) higher than naive estimates (i.e., proportion of camera‐trap sites species recorded from) thus highlighting the importance of accounting for detectability during conservation surveys. I recommend the use of an occupancy framework when using camera‐trap data to study the status, ecology, and habitat preferences of poorly known and elusive species. The results highlight the importance of mixed deciduous and semi‐evergreen forest for wild cattle in eastern Cambodia and I emphasize that these habitats must be considered in conservation planning across the Lower Mekong Dry Forest Ecoregion.     

Plant and small mammal richness correlate positively in a biodiversity hotspot

Hotspots of biodiversity are important areas in facilitating an understanding of species richness and its maintenance. Herbivores can increase plant richness by reducing dominant plant species thus providing space for subdominant species. As small mammals are abundant in the Succulent Karoo and therefore might affect plant richness by means of herbivory, we tested if this mechanism might exist in the Succulent Karoo in southern Africa, a biodiversity hotspot due to its extraordinary plant richness. At ten ecologically different study sites we measured plant and small mammal richness and diversity and determined 11 abiotic factors including soil composition, altitude and rainfall. We found positive correlations between plant richness and the number of small mammal species. A general linear model revealed that the number of small mammal species was more important than abiotic factors in explaining variation in plant richness. To test whether small mammals might directly influence plant richness, we studied the influence of the bush-Karoo rat Otomys unisulcatus, a central place forager, on the plant community. The immediate surroundings of occupied O. unisulcatus nests showed significantly higher plant richness than control areas. We conclude that small mammals can have a positive effect on plant richness in the Succulent Karoo. While experimental data are needed to support these correlative results, the results of our study indicate that areas of high small mammal richness should be included in conservation programs of the Succulent Karoo.     

Habitat disturbance and small mammal richness and diversity in an Atlantic rainforest area in southeastern Brazil.

Since disturbance is an important ecological factor affecting species diversity in natural environments, the increasing human occupation rate in Brazilian Atlantic rainforest, which supports about 50% of Brazil's human population, has resulted in intense habitat degradation and fragmentation. Within this rainforest, animal and plant species have been lost at a high rate, and biological and diversity is presently vulnerable. Various animals community studies along a gradient of environmental disturbances have shown that the highest species diversities occur in habitats with intermediate levels of disturbance frequency and intensity. In the present study, which was carried out in the Atlantic forest of Ilha Grande (23 degrees 11' S and 44 degrees 12' W), an island located on the southern coast of Rio de Janeiro State, southeastern Brazil, we estimated species richness, diversity, and abundance of small mammals in three forest areas having different disturbance levels. This was done, in each of these areas and in an anthropic area that they surround, by establishing trails 200 m long, in which points were marked at 20 m intervals. The work involved a total effort of 4800 trap/nights. We also measured some habitat variables at each site in order to evaluated their disturbance levels. Our data showed that the two most conserved forests had the lower species richness and small mammal diversity, while in the anthropic area wild species were absent. The forest with an intermediary level of disturbance showed higher values for species richness and diversity, with the anthropic area presenting the highest disturbance level.