Effect of presence of the dingo (Canis lupus dingo) on risk-sensitive foraging of small mammals in forest ecosystems

Amanda Lu ‘13

Letnic Research Group, Hawkesbury Institute for the Environment, University of Western Sydney


Trophic regulation of mesopredators through top-order predators can have profound effects on ecosystem com­munity and diversity. In the absence of a top predator such as the dingo, an abundance of invasive mesopredators, such as cats and foxes, exerts strong selective pressures on native prey and can alter prey foraging behavior. When foraging in the presence of predators, prey must weigh predation risk against food gain. To examine the indirect impacts of dingo baiting on risk-sensitive foraging in forests, we measured differences in giving up densities of food resources (GUDs) and surveyed local populations of mesopredators and mammals. We hypothesized that in baited areas, mesopredators would be more abundant and prey would perceive greater predation risk. Forag­ing trays of peanuts were placed in baited and nonbaited study areas for four nights and the remaining peanuts measured as the GUD. A higher density of mesopredators (cats and foxes) and a lower density of small mam­mals were observed in baited versus nonbaited study sites. Consistent with foraging theory, rodents perceived significantly greater predation risk in baited areas than nonbaited areas. However, an abundance of medium and large mammals was not affected by the baiting regime. Ecosystem conservation management has strongly focused on the baiting of invasive predators in protected reserves. Our study suggests that the absence of a top predator positively affects mesopredator abundance and negatively affects small mammals in forests. For sus­tainable forest management, reconsideration of predator poison baiting regimes may be necessary to optimize ecosystem diversity and structure.



The presence of top predators in animal communities has important consequences for ecosystem structure and function. Predators can exert a top-down effect on herbivores and smaller predators through traditional trophic cascade theory. Suppression of herbivores can positively impact plant species and diversity while intraguild predation or competition can promote small mammal and bird diversity (Ripple and Beschta 2002). The keystone role of top-order predators amplifies their impact on prey populations and habitat diversity.

Predation can include both direct effects on populations through consumption and selective pressures on prey behavior. Top-order cascade effects can alter behavior to reduce prey activity or shift activity to safer habitats. Under optimal foraging theory, prey must weigh feeding rate against predation risk to determine whether to utilize a feeding patch (MacArthur and Pianka 1966). When food density is high and predation risk is low, prey benefit by continu­ing to forage in the harvest patch. However, when food density is low and predation risk is high, prey species must increase their vigilance; predation risk may outweigh any benefit from foraging. Testing foraging behavior of prey requires measuring habitat use, preferences, and the acceptance or rejection of patches (Brown et al. 1988).

While risk foraging preferences can be directly influenced by predation risk (Brown et al. 1992; Hughes and Ward 1993; Jacob and Brown 2000), the effect of top-order trophic cascades on foraging behavior has only been recently studied (Letnic and Dworjanyn 2011; Strauß et al. 2008). Under the mesopredator release hypothesis, removal of top predators has a direct positive effect on mesopredators and an indirect negative effect on prey diversity (Crooks and Soule 1999). Top predator interaction with mesopredators can include opportunistic intraguild predation on mesopredators or mesopredator avoidance of areas frequented by top predators due to perceived predation risk. Consequently, fol­lowing removal of top predators, changes in mesopredator behavior can include strong predation pressure on prey and can lead to local extinction (Johnson et al. 2007). Past studies on the mesopredator release hypothesis have focused on the ecological role of wolves in North America (Hebblewhite et al. 2005). In wolf-absent areas, elk populations were greater, so herbivory reduced vegetation complex­ity and structure, subsequently affecting songbird diversity.

In Australia, the dingo has been intensively studied as a key­stone species because of its predatory and competitive effects on invasive mesopredators (Glen et al. 2007). One of Australia’s most ecologically influential predators, the dingo arrived approximately 3500-4000 years ago with Asian seafarers (Corbett 1995). While originally brought as human companions, dingoes frequentlyescaped and quickly established wild populations. Soon after the introduction of the dingo, both native marsupial predators, the thylacine and Tasmanian devil, went extinct on the mainland, possibly due to outcompetition from the dingo. With the conver­sion of land to pastoral use, dingoes have been driven from much of southeastern Australia (Glen and Short 2000). Hundreds of introduced species have further contributed to the extinction or demise of native flora and fauna. Invasive mesopredators such as the European red fox (V. vulpes) and the house cat (F. catus) are widespread across the continent and have significant impacts on the biodiversity of small and medium-sized mammals.

Inverse relationships between dingo and fox activity have been found in southeastern Australia (Newsome et al. 1997). The survival of native marsupials has been closely correlated with high dingo density, supporting top predator maintenance of prey biodiversity (Johnson et al. 2007). In a parallel study on the impacts of dingoes in arid Australia, Letnic et al. (2009) found lower mesopredator abundance and herbivore activity in the presence of dingoes than in the absence of them, when compared across the New South Wales dingo fence. Within the last century, pastoral interests have led to heavy dingo poison baiting and thus the removal of a top predator.

Forest ecosystems in southeastern Australia hold a rich diversity of flora and fauna. Wild dingo populations persist in areas along the coast, while in other areas, ongoing baiting suppresses popula­tions. Previous regional studies have investigated dingoes and the competitive exclusion of foxes in New South Wales forests; however, human presence was attributed as a dominant factor in determining the distribution of foxes (Catling and Burt 1995). Evidence in sup­port of a negative relationship between fox and dingo abundance in the eastern forests suggests that the abundance of dingoes sets an upper limit on the abundance of foxes (Johnson and VanDerWal 2009). Extensive scat analysis of dingo and fox diets in the Sydney area has shown that foxes consume a greater range of prey than dingoes and that a greater proportion of their diet is comprised of small and medium sized mammals, birds, reptiles, and insects (DECC 2007). Intraguild predation and dietary competition have also been suggested to suppress foxes (Cupples et al. 2011).

Although risk-sensitive foraging in relation to dingo presence has been studied in arid Australia (Letnic and Dworjanyn 2011), little is known about the impacts of dingo presence on rodent foraging in forest ecosystems. Here, we study the indirect effects of dingo presence on mesopredator abundance and rodent foraging behavior in eastern New South Wales forests. We measured predator/prey abundance and perceived predation risk in baited and nonbaited areas. First, since dingo populations will be suppressed in baited areas, based on the mesopredator release hypothesis, a higher density of mesopredators is expected. Secondly, greater density of mesopredators will exert greater predation pressure on small mammals and influence the foraging behavior of prey. In accor­dance with optimal foraging theory, we predict higher perceived predation risk in baited areas.



Selection of poison baited and nonbaited study sites

Due to pressures from domestic livestock and urban settings, the presence of the dingo in NSW is now limited to the coastal ranges and the northwest corner of the state. In many areas, poison baiting is conducted to limit invasive preda­tor populations as part of conservation efforts. Two adjacent study areas were selected in the mid-North Coast to assess prey foraging behavior in the presence and absence of dingoes: Hat Head National Park (31o2’27’’S, 153o1’35’’E; baiting) and Limeburners Creek Natural Reserve/Goolawah National Park (31o18’8’’S, 152o53’49’’E; no baiting), which are forested habitats on the coast that form part of a system of eight coastal protected areas from the Manning River to Coffs Harbor. The parks were selected based on similar vegetation, habitat, landforms, and close proximity to minimize climactic factors (10 km apart). Hathead National Park contains 7,200 hectares of coastal land with vegetation communities of wet and dry heathland, littoral rainforest, eucalypt forest, and woodland (Hat Head NP Plan of Management 1998). Mean annual rainfall in the park is 1484.0 mm (1939-2011; Australian Bureau of Meterology). Intensive 1080 poison meat baiting in the past and currently in Hat Head has retarded dingo presence. Limeburners Creek NR was the first nature reserve declared on the north coast of New South Wales and covers an area of 9,123 hectares of wet and dry heathland, littoral rainforest, eucalypt forest and woodland, as well as fresh water and estuarine wetlands (Limeburners Creek NR Plan of Management 1998). Mean annual rainfall at Port Macquarie, 12.6 km away, is 1534.5 mm (1840-2010; Australian Bureau of Meterology). As part of the landcare initiative the Maria River Project, no dingo or fox baiting has taken place in Limeburners NR or adjoining properties since 2007. The park has a stable dingo population that is closely monitored through satellite tracking (NPWS 2011). Goolawah NP is an adjacent regional park of 534 hectares cre­ated in 2010 with similar vegetation and is managed by Limeburners Creek NR (no baiting regime).

Within each study area, eight sites were placed along roads at least 1 km apart (Figures 1 and 2). Baited sites were placed several kilometers apart from nonbaited sites.

Figure 1. Map of Hat Head National Park. Dots indicate locations of each site (Source: Google Maps).
Figure 2. Map of Limeburners Creek Nature Reserve and Goolawah National Park. Dots indicate locations of each site. Two sites (circled) were located on private land adjacent to Limeburners Creek NR with the permission of the owners Goolawah Co-op (Source: Google Maps).


Assessment of risk-sensitive foraging

Mammal foraging activity was assessed by measuring giving up densities (GUDs) through foraging trays in the eight sites for each study area. Under optimal patch use theory, resource depletion during foraging reduces the harvest rate and thus the benefit for the prey. Declining harvest rates therefore must be weighed against metabolic cost, predation risk, and missed oppor­tunity costs of foraging (Brown 1988). The giving up density (GUD) of food measures the point where predation risk outweighs harvest rate and the prey stops foraging. Assuming stable metabolic and opportunity costs between study areas, we predict that dingo regulation of mesopredator abundance should mediate predation risk from mesopredators. Since the density of food decreases as more food is harvested, the giving up density is a measure of the prey perception of foraging costs and predation risk. This method has previ­ously been established as a quantitative measure of perceived predation risk (Brown et al. 1997; Letnic et al. 2009).

Four artificial food patches were established at each of the sites using alu­minum trays (30 × 20 × 5 cm) provisioned with 20 peanut quarters each and randomly mixed with 1 L of sifted sand to create an even distribution and declining gain. The number of seeds used was replicated from previous work in arid Australia (Letnic and Dworjanyn 2011). To attract medium-sized mam­mals such as long-nosed bandicoots (Perameles nasuta), peanut quarters were coated with black truffle oil (Paull et al. 2010). Peanut trays were placed for four nights at each site. The first experiment was run at Hat Head for four nights (15-21 November 2011, heavy rains forced us to suspend experiments for two nights), and the second experiment was run at Limeburners Creek/Goolawah the following week (22-25 November 2011) for four nights. Both experiments were conducted under a waning moon. We also trialed mealworms as a bait attractant for medium-sized mammals in two food patches at each site for two nights of each experiment (Searle et al. 2008). Ten mealworms (larval stage of Tenebrio molitor) were placed in aluminum trays (22 × 16 × 5 cm) and mixed into 1 L of sand (Kovacs et al. 2011).

The area around each peanut or mealworm tray was swept so footprints could be observed. Each morning foraging trays were checked for signs of disturbance or tracks and the number of peanuts or mealworms remaining was counted as the giving up density (GUD). Foragers were identified by inspecting the tracks present on the sand or the chew marks on the nuts. Confirmation of the species that visited the trays was conducted by mounting an infrared game camera with movement sensors on one tray at each site (ScoutGuard). Cameras were set to 1 min video with 1 second interval from 6:00 PM to 7:10 AM the following day. Missing peanuts and dead or missing mealworms were replaced every morning and the surrounding sand swept.

Mean giving up densities were calculated as the mean number of peanut or mealworm pieces remaining per night per tray and analyzed using a general­ized linear model with study area (baited or nonbaited) as the variable factor through SPSS Statistics 18.

Small mammal trapping

To estimate abundance of small mammals in each study area, trapping grids were laid out at each of the eight sites. A total of 20 traps within each site were situated 10 m apart in a 50 m × 50 m square. All trapping grids were positioned approximately north-south at least 10 m away from the road. Fold­ing aluminum Elliott traps (33 × 10 × 10 cm) were baited for four consecutive nights with a mixture of rolled oats, peanut butter, honey, and truffle oil. Traps were checked from first light every morning. Captures were identified, weighed, and sexed. After first capture, individuals were marked on the tail with permanent marker to identify recaptures.

Predator activity

To assess presence of predators in the area, 16 sand plots at Hat Head and 16 sand plots at Limeburners/Goolawah were placed at 500 m intervals on the road. An area the width of the road × 1 m was swept each afternoon and the tracks observed and identified the following morning. Sand plots were maintained for four consecutive days at each study area (baited and nonbaited).

Faunal abundance

In order to consider the predatory impacts on prey, we compared abun­dances of prey populations in each study area through transect runs. Macropod and faunal abundance activity was measured through afternoon surveys. Two afternoon surveys in each study area were conducted between the hours of 5:00-8:00 PM on transects ranging from 9 m to 11.1 km along the road while driving at 15 km/h. Visibility along the road was 10 m to each side and an index of abundance was calculated as the mean number of animals sighted per kilometer surveyed. We conducted three spotlighting surveys for between 6.3 to 9.1 km in each study area from a four wheel drive traveling at 5 km/h. Visibility was 10 m on each side of the road. Abundance data was calculated as the mean number of animals sighted per a kilometer surveyed.

At each of the eight sites in both study areas, four scat surveys were under­taken on linear transects of 100 m length. The observer looked one meter to each side and all scats were tallied and identified. Indications of animal presence were also counted on each transect (goanna (Varanus) diggings, burrows, scratchings etc).

Medium-sized mammal camera trap

Activity of medium sized mammals was monitored using baited camera traps (ScoutGuard). A bait of honey, oats, peanut butter, and sardines was placed in a tube cylinder on the ground with openings. Mounted cameras were set to 15 second video recordings from 6:00 PM to 7:10 AM and collected after 4 nights for identification. Two camera traps were placed at each site.

Habitat complexity

To account for possible differences in vegetation between study areas, an assessment of habitat complexity at all eight sites in both study areas was con­ducted. Percentage canopy covered was measured by walking 200 m around the trapping grid in the site and recording canopy cover or sky every 5 meters. At every 10 m, the number of understory layers and maximum height within a 1 m radius was recorded. In order to measure the vegetation density, a horizontal coverboard method was used (Monamy and Fox 2000). A checkered 20 cm × 50 cm board of ten 10 cm × 10 cm squares was held 5 meters away from an observer at 0–20 cm; 20–50 cm; 50–100 cm; 100-150 cm; and 150-200 cm in height. Ground cover was also estimated by taking 50 steps within a 10 m × 10 m plot and recording green vegetation, leaf litter, log, or bare ground at each step. Four estimates of board and ground cover methods were made at each of the north, east, south, and west points of the trapping grid.



Foraging behavior

Takes by rodents from the peanut foraging trays were recorded at both the baited and nonbaited study areas (n=17, baited; n=23, nonbaited). Mean giving up densities were significantly lower in nonbaited study areas compared to baited study areas with p < 0.05 (Figure 3). Medium-sized mammal takes at peanut trays were not included in calculations due to variations in giving up density for medium versus small mammals. Camera footage confirmed rodent foraging at trays overnight.

Foraging from mealworm trays was not observed from any sites. Elevated precipitation during foraging nights and ant predation reduced survival of mealworms. Detection of mealworm escape or bird foraging was also difficult. However, the mean number of mealworms remaining was calculated as 6.27 ± 0.42 SE in baited and 7.52 ± 0.40 SE in nonbaited sites.

Four species were captured in total: R. fuscipes (bush rat), R.rattus (black rat), R. litroleus (swamp rat), A. stuartii (antechinus). Mean capture rates (captures per trapping night) were higher for all species in nonbaited study areas compared to baited study areas (Figure 4).

Figure 3. Mean peanut giving up densities (± SE) of rodents. The giving up density represents the number of peanuts remaining in 1 L of sand.
Figure 4. Mean abundance (± SE) of small mammals in captures per trapping night. Trapping nights were counted as the number of traps open per a night. R. fuscipes (bush rat), R.rattus (black rat), R. litroleus (swamp rat), A. stuartii (antechinus).

Predator sand plots

Dingoes were detected on plots at both the baited and nonbaited sites; however, the mesopredators cat (Felis catus) and fox (Vulpes vulpes) were detected only at the baited sites (Figure 5). Eastern grey kangaroos (Macropus giganteus) were more frequently detected in nonbaited sites while medium-sized marsupials (Perameles nasuta, Potorous tridactylus, and Pseudocheirus) had comparable densities at both study areas.

Figure 5. Mean abundance (± SE) on a predator sand plot per night over four nights. M. gigantus (Eastern grey kangaroo), P. nasuta, P. tridactylus, or Pseudocheirus (medium marsupials), S. scrofa (pig), F. catus (cat), V. vulpes (fox), C. lupus dingo (dingo), Varanus (goanna).

Medium-sized mammal camera traps

Abundance of fauna detected by the camera traps was calculated as the total number of individuals observed over a four-night period with 16 cameras in each study area (baited or nonbaited). Two spe­cies were observed at the baited site and three species were observed at the nonbaited site. Overall, abundance of fauna recorded was higher at nonbaited than baited sites (Figure 6).

Figure 6. Faunal abundance observed by infrared cameras. Total number of individuals recorded at 16 cameras over four nights.

Scat and trace count

Searches for scats and other traces yielded a greater diversity of animals per hectare in nonbaited than baited sites. Mean abun­dance was calculated by averaging the number of scats and traces per hectare. Abundance of medium-sized mammals: wallaby (M. rufogriseus or W. bicolor), bandicoot (P. nasuta), potoroo (P. tri­dactylus), echidna (T. aculeatus), sugar glider (P. breviceps), koala (P. cinereus), or wombat (V. ursinus) was higher in nonbaited than nonbaited sites. However, abundance of macropods was slightly higher in nonbaited areas. Dingo scats were found only in the non­baited study area.

Figure 7. Mean number (± SE) of scats or traces per a hectare in baited and nonbaited areas. M. gigantus (Eastern grey kangaroo), M. rufogriseus or W. bicolor (wallaby), P. nasuta (bandicoot), P. tridactylus (potoroo), T. aculeatus (echidna), P. breviceps (sugar glider), P. cinereus (koala), V. ursinus (wombat), S. scrofa (pig), C. lupus dingo (dingo).


Spotlighting transects were divided by the number of kilome­ters surveyed to calculate average abundance/km. Eight species were observed in total (6 in baited and 4 in nonbaited). Overall, abundance of medium-sized prey was similar in baited sites and nonbaited sites. Cats were observed in baited sites but not in the nonbaited sites.

Figure 8. Mean abundance (± SE) per km surveyed by spotlight. Three spotlights were conducted in the baited area and only two spotlights were con¬ducted in the nonbaited area due to inclement weather. M. rufogriseus (red-necked wallaby), P. tridactylus (potoroo), P. breviceps (sugar glider), P. norfol¬censis (squirrel glider), Trichosurus (brushtail possum), P. cinereus (koala), F. catus (cat).

Afternoon surveys

Abundances of fauna were calculated as the number per kilo­meter surveyed. The mean density of macropods sighted was much higher in baited than nonbaited sites.

Habitat assessment

Vegetation density calculated by the coverboard method was similar in both the baited and nonbaited study areas (Figure 9). Percent visibility at all heights except 150-200 cm was slightly higher in the baited sites. Ground cover composition was similar between both study areas and dominated by leaf litter followed by green vegetation, bare ground, and logs. However, the baited site was composed of more leaf litter (58.2 ± 3.4%) than the nonbaited site (48.6 ± 3.2%), while the proportion of green vegetation was greater in nonbaited than baited areas (Figure 11). While canopy cover was greater in unbaited sites (68.5 ± 6.4%) compared to (59 ± 3.2%) that in baited sites (Figure 12), understory complexity measured by mean understory layers and height was similar between study areas.


Figure 9. Mean abundance (± SE) per km surveyed in the afternoon. Only one survey was conducted at the nonbaited site due to inclement weather. M. gigantus (Eastern grey kangaroo), M. rufogriseus (red-necked wallaby), W. bicolor (swamp wallaby), L. europaeus (brown hare).
Figure 10. Mean percent visibility (± SE) through vegetation at varying height intervals above ground. Mean visibility at each study area was cal¬culated by averaging the proportion of squares visible out of 10 squares on a 50 cm x 20 cm checkered coverboard.Figure 11. Mean percent (± SE) ground cover composition in baited and unbaited study areas.


Figure 12. Mean percent canopy and understory complexity (± SE) in baited and unbaited study areas.


The results from this study support the mesopredator release hypothesis: in areas of baiting, increased mesopredator activity and subsequent suppression of small mammal abundance was observed. These findings are consistent with previous studies on the negative correlation found between dingoes and mesopredators (Kennedy et al., 2011; Letnic and Dworjanyn 2011). In accordance with the expectation that perceived predation risk is higher with greater predator density, giving up densities of rodents were higher in the baited study area. However, while small mammal abundance demonstrated the direct impacts of predation, medium-sized mam­mal and macropod abundance appeared to be unaffected by dingo presence.

Lower abundance and diversity of small mammals in baited areas was most likely indicative of increased mesopredator activ­ity. Presence of V. vulpes and F. catus on predator track plots was observed only in baited study areas. In studies where foxes have been removed, increases in small mammal populations have been observed (Kinnear et al. 2002; Dexter and Murray 2009). Further­more, evidence of native mammal and bird predation by F. catus and the subsequent decline of local populations has been well docu­mented (Dickman 1996). The exclusion of mesopredators by dingoes has been supported by studies of interactions between feral cats and dingoes, suggesting that distributions of feral cat populations in dense cover may reflect the use of cover as more protection from C. lupus dingo (Edwards et al. 2002).

The presence of C. lupus dingo tracks on sand plots in both study areas indicate that baiting regimes may not completely exclude din­goes from the area. Dingo tracking satellite data in the nonbaited sites (NPWS unpublished) indicate that while all dingoes tracked spent the majority of time within the nonbaited Limeburners Creek Nature Reserve, dingoes travel between the two parks. Therefore, our model may not strictly reflect dingo-present and dingo-absent ecosystems, instead reflecting greater dingo activity in the unbaited study area. However, dingoes that do travel through the baited Hat Head National Park frequently die from bait consumption. Accordingly, Limeburners Creek NR sustains a stable dingo popu­lation whereas Hat Head NP does not. While dingoes were only opportunistically sighted during the study, all sightings occurred in Limeburners Creek NR (unbaited; 3 adults, 2 juveniles). More extensive sand plot data is needed to clarify dingo abundance and activity in each study area.

Strong selection pressure by mesopredators on rodents predicts that anti-predator behaviors should develop in baited areas. Conse­quently, optimal foraging theory suggests that giving up densities should increase with increased predation risk (Brown 1988). The higher giving up densities observed in baited areas was supportive of our initial predictions. While higher GUDs and small mammal abundance data in baited areas were consistent with mesopredator presence, differences in GUDs may have been due to other factors. Density of rodents was higher for all species in the unbaited areas, which suggests the possibility of multiple encounters of prey per foraging patch. Consequently, giving up densities would reflect population density rather than the foraging choice of the last indi­vidual. Analyzing the number of individuals visiting a patch is difficult to determine from track data. However, none of the cameras placed observed more than one individual at a foraging tray. An additional alternative explanation that has been suggested for lower GUDs is that at higher rodent population densities, intraspecific competition for food resources results in a lower state of energy of rodents (proposed Brown et al. 1997, Searle et al. 2008, Letnic and Dworjanyn 2011). More extensive study of risk-sensitive foraging would be required to assess the validity of this concern. Similarly, risk-sensitive foraging based on microhabitat use was not examined. Small mammals may prefer foraging patches placed under cover or may forage more under a new moon rather than full moon (Brown et al. 1997).

Risk foraging was only evaluated in rodents and may differ in medium-sized mammals. The viability and accessibility of meal­worm foraging trays have been factors in the lack of foraging from patches. In future studies, a cover would need to be placed over the worms to prevent overheating or water filling the trays. Palatability by native fauna may be another deterrent, as cameras placed at the trays detected an echidna, but no mealworms were foraged.

Faunal abundance trends were inconsistent between the visual and scat/trace survey methods. While macropod abundance was higher in baited sites during afternoon and spotlighting surveys, dung transects and sand plots suggested greater macropod presence in nonbaited areas. Other studies of dingo-macropod interactions have found increases in kangaroo abundance at low dingo density due to predation (Corbett & Newsome 1987; Thomson 1992; Letnic et al. 2009). The similar large prey densities between baited and non­baited areas in this study may be reflective of macropod population influx from surrounding cleared land where dingoes are absent due to human exclusion. Macropods were visually observed on private land near Limeburners Creek NR during the study.

Overall, trends in the surveys of medium-sized mammals (bandi­coots, potoroos, echidnas, hares, koalas) demonstrated greater density in the nonbaited study area. These findings are consistent with predictions from the mesopredator release hypothesis. Previ­ous scat comparisons of fox and dingo diets elsewhere in Australia have shown resource partitioning. Small and medium-sized prey comprise a greater proportion of fox diets than dingo diets (DECC 2007; Mitchell and Banks 2005). However, glider and possum abun­dances were unaffected by dingo absence or presence. Arboreal species such as possums and gliders may be less influenced by mesopredation if foraging behaviors shift to minimize time spent on the ground.

Limitations of this study include the difficulties of comparing two spatially segregated habitats. Differences in vegetation between study areas could play a factor in the differences observed. Vegeta­tion density was similar in both study areas; however, canopy cover was greater in the nonbaited sites. A denser canopy could contribute to differences in forest ecosystem communities; nevertheless, under­story complexity and height were similar between sites. Previous study on the composition of ground dwelling mammals in NSW eucalypt forests suggests that understory complexity and the density of understory shrubs determines ecosystem composition (Catling and Burt 1995). Similarly, small mammal abundance is dependent on the regeneration of understory cover after a fire rather than on time (Monamy and Fox 2005). While differences in habitat between sites could have influenced the trends observed, understory com­plexity and height were comparable between study areas and thus less likely to play a determining role in foraging behavior.

Though the results of this study provide support for small mam­mal diversity and abundance under nonbaiting regimes, potential shortcomings include the short time period of the study and lack of replicates. Seasonality effects could play a role in predator-prey population fluctuations, which would not have been captured in our data. Furthermore, differences observed may be locally con­tained and influences of top order predators may not be consistent in other forest ecosystems. This preliminary study is the first part of a larger project that will examine several other forest ecosystems in New South Wales.



The influence of top predators on ecosystem communities through limitation on mesopredators and subsequent trophic cas­cade effects is gathering support as a consideration in ecosystem management. Invasive mesopredators threaten native wildlife and can result in local prey population extinctions. This study demon­strated the cascade effects of a top predator on foraging behavior of small mammals. With increased mesopredation, prey must com­pensate for increased predation risk by reducing foraging time. While baiting has been widely used in Australia to control invasive species, given that the presence of a top order predator can regulate mesopredation and thus promote small mammal diversity, nonbait­ing management of forest ecosystems may be of significant value.



I would like to thank Dr. Mike Letnic from the University of Western Sydney, my project advisor, for enabling me to conduct this study. Nick Colman, Ph.D. candidate, under whom I conducted fieldwork, was a wonderful mentor and coworker for my first expe­rience in the field. I was privileged to work as a part of his larger project in NSW forests.



Brillouin, L. 1956. Science and information theory. Academic Press, New York.

Brown JS. 1988. Patch Use as an Indicator of Habitat Preference, Predation Risk, and Competition Behavioral Ecology and Sociobiology , Vol. 22, No. 1 (1988), pp. 37-47

Brown, J., Kotler, B.P., Smith R.J., Wirtz, W.O. 1988. The Effects of Owl Predation on the Foraging Behavior of Heteromyid Rodents. Oecologia 76, 408-415.

Catling, P.C., Burt, R.J. 1995. Studies of the ground dwelling mammals of eucalypt forests in south-eastern NSW: the effect of habitat variables on distribution and abundance. Wild. Res, 22, 271-88.

Corbett L. K. 1995. The Dingo in Australia and Asia. UNSW Press, Sydney.

Corbett, L. K. & Newsome, A. E. 1987. The feeding ecology of the dingo. 3. Dietary relationships with widely fluctuating prey populations in arid Aus­tralia – an hypothesis of alternation of predation. Oecologia 74, 215–227.

Cupples, J.B., Crowther, M.S., Story, G., Letnic, M. 2011. Dietary overlap and prey selectivity among sympatric carnivores: could dingoes suppress foxes through competition for prey? Journal of Mammalogy 92, 590-600.

Department of Environment and Climate Change of NSW. 2007. Threatened and pest animals of Greater Southern Sydney. Sydney: Department of Envi­ronment and Climate Change.

Dexter N, Murray A. 2009. The impact of fox control on the relative abundance of forest mammals in East Gippsland, Victoria. Wild Res 36, 252-261.

Dickman, CR. 1996. Impact of exotic generalist predators on the native fauna of Australia. Wild. Bio 2, 185-195.

Edwards, G. P., Preu, N. D., Crealy, I. V. and Shakeshaft, B. J. 2002. Habitat selection by feral cats and dingoes in a semi-arid woodland environment in central Australia. Austral Ecology 27, 26–31.

Glen A. S. & Short J. 2000. The control of dingoes in New South Wales in the period 1883–1930 and its likely impact on their distribution and abundance. Aust. Zool. 31, 432–42.

Hebblewhite, Mark, Clifford A. White, Clifford G. Nietvelt, John A. McKenzie, Tomas E. Hurd, John M. Fryxell, Suzanne E. Bayley, and Paul C. Paquet. 2005. Human activity mediates a trophic cascade caused by wolves. Ecol­ogy 86, 2135–2144.

Hughes, JJ, Ward, D. 1993. Predation risk and distance to cover affect foraging behavior in Namib Desert gerbils. Animal Behavior 46, 1243-1245.

Jacob, J, Brown JS. Microhabitat Use, Giving-Up Densities and Temporal Activ­ity as Short- and Long-Term Anti- Predator Behaviors in Common Voles. Austral Ecology 32, 492-501.

Johnson, C. N. & VanDerWal, J. 2009. Evidence that dingoes limit abundance of a mesopredator in eastern Australian forests. Journal of Applied Ecology 46, 641–646.

Johnson, C.N., Isaac, J.L., Fisher, D.O. 2007. Rarity of a top predator triggers continent-wide collapse of mammal prey: dingoes and marsupials in Aus­tralia. Proceedings of the Royal Society of London B 274, 341–346.

Kennedy, M., Phillips, B., Legge, S., Murphy, S. & Faulkner, R. 2011. Do dingoes suppress the activity of cats in northern Australia? Austral Ecology. Online.

Kovacs, E.K., Crowther, M.S., Webb, J.K., Dickman, C.R. 2011. Populationa nd behavioural responses of native prey to alien predation. Oecologia. Online.

Letnic, M., Crowther, M. S. and Koch, F. 2009. Does a top-predator provide an endangered rodent with refuge from an invasive mesopredator?. Animal Conservation, 12, 302–312.

Letnic, M., Koch F. 2010. Are dingoes a trophic regulator in arid Australia? A comparison of mammal communities on either side of the dingo fence. Austral Ecology, 35, 167–175.

Letnic, M. and Dworjanyn, S. A. 2011. Does a top predator reduce the predatory impact of an invasive mesopredator on an endangered rodent?. Ecography, 34, 827–835.

Letnic, M., Greenville, A., Denny, E., Dickman, C. R., Tischler, M. K., Gordon, C. & Koch, F. 2011. Does a top predator suppress the abundance of an inva­sive mesopredator at a continental scale? Global Ecology and Biogeography 20, 343–353.

Macarthur, H., Pianka, A.L. 1966. On optimal use of a patchy environment. The American Naturalist 100, 603-609.

Mitchell, BD., Banks, PB. 2005. Do wild dogs exclude foxes? Evidence for com­petition from dietary and spatial overlaps. Austral Ecology 30, 581-591.

Monamy, V., Fox, BJ. 2000. Small mammal succession is determined by vegeta­tion density rather than time elapsed since disturbance. Austral Ecology 25, 580-587.

Newsome A. E., Pech R. P., Smyth R., Banks P. B. & Dickman C. R. 1997. Potential Impacts on Australian Native Fauna of Rabbit Calicivirus Disease. Environment Australia, Canberra.

Paull, David J., Claridge, Andrew W., and Barry, Simon C. (2011) There’s no accounting for taste: bait attractants and infrared digital cameras for detecting small to medium ground-dwelling mammals. Wildlife Research 38, 188–195.

Read, V.T., Malafant, KWJ, Myers, K. 1988. A comparison of grid and index-line trapping methods for small mammal surveys. Aust. Wildl. Res. 15, 673-87.

Ripple, W.J., Beschta, R.L. 2004. Wolves and the Ecology of Fear: Can Preda­tion Risk Structure Ecosystems? BioScience 54, 755-766.

Strauß A Solmsdorff KY Pech R Jacob J. 2008. Rats on the run: removal of alien terrestrial predators affects bush rat behavior. Behavioral Ecology and Sociobiology 62.

Thomson, P. C. (1992). The behavioural ecology of dingoes in north-western Australia. III. Hunting and feeding behaviour, and diet. Wildlife Research 19, 531–541.