Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Action: Exclude livestock from semi-natural habitat (including woodland) Terrestrial Mammal Conservation

Key messages

Read our guidance on Key messages before continuing

  • Nine studies evaluated the effects of excluding livestock from semi-natural habitat on mammals. Six studies were in the USA, two were in Spain and one was in Australia.



  • Abundance (9 studies): Four out of eight studies (including four site comparisons and four controlled studies), in the USA and Spain, found that excluding grazing livestock led to higher abundances of mule deer, small mammals and, when combined with provision of water, of European rabbits. One study found higher densities of some but not all small mammals species when livestock were excluded and the other three studies found that grazing exclusion did not lead to higher abundances of black-tailed hares, California ground squirrel burrows or of five small mammal species. A site comparison study in Australia found more small mammals where cattle were excluded compared to high intensity cattle-grazing but not compared to medium or low cattle-grazing intensities.


Supporting evidence from individual studies


A controlled study in 1982–1984 on a shrubland site in California, USA (Roberts & Tiller 1985) found that inside a cattle-exclusion fence, there were more mule deer Odocoileus hemionus than there were outside it. This result was not tested for statistical significance. Over six sampling events, 192 faecal pellet clumps were counted inside the enclosure compared to 138 outside it. In June 1982, a prescribed burn was carried out across 4 ha of land. A 0.25-ha enclosure (cattle proof but not deer proof) was established on the burned area. Relative deer presence inside and outside the enclosure was assessed by counting pellet-groups in September 1982, February, August, and November 1983 and March and July 1984. Counts were made along 18 transects (5 m long) inside the enclosure and 18 outside the enclosure.


A replicated, site comparison study in 1990–1992 in a desert in south-central California, USA (Brooks 1995) found that excluding livestock led to more small mammal species, and higher densities of some small mammal species, compared to sheep-grazed areas. More species of small nocturnal rodents were found in ungrazed (3.7 species/sample) than in grazed areas (2.5 species/sample), and diversity was higher in ungrazed areas in all three years (data reported as diversity indices). The densities of three of five species were higher in ungrazed than in grazed plots (long-tailed pocket mouse Chaetodipus formosus: 26 vs 6 animals/ha; Merriam’s kangaroo rat Dipodomys merriami: 31 vs 13; southern grasshopper mouse Onychomys torridus: 3 vs 0 respectively). The densities of the other two species did not differ significantly between grazed and ungrazed plots (little pocket mouse Perognathus longimembris: 29 vs 30 animals/ha; deer mouse Peromyscus maniculatus: 1 vs 0). Two pairs of 65-ha plots were established in 1990 with one plot inside an area fenced since 1978–1979 and one outside, in an area grazed by sheep (grazing intensity not stated). Over five periods of four to six nights, in May 1990–March 1992, mammals were caught in 64 Sherman traps/plot, 10 m apart.


A replicated, site comparison study in 1994 in a desert site in California, USA (Brooks 1999) found that in areas where livestock were excluded, there were fewer black-tailed hares Lepus californicus, compared to in sheep-grazed unfenced areas that were also driven over by off-road vehicles. Fewer black-tailed hares were found in fenced plots (0–1.5 hares/survey; 11 droppings/m2) compared to in unfenced plots (1–4 hares/survey; 22–31 droppings/m2). Two 2.25-ha plots that were fenced in 1980 were compared to two plots that were grazed by sheep (and driven over by off-road vehicles). Sites were matched for environmental variables. Hare numbers were estimated in May and July 1994 by counting the number of hares seen on four 1.25-km-long transects and the number of droppings in sixty 40 × 50-cm sampling units in each plot.


A replicated, site comparison study in 1998–1999 of a riparian grassland area in Pennsylvania, USA (Giuliano & Homyack 2004) found that stream margins, fenced to exclude grazing livestock, had a higher species richness and abundance of small mammals than did unfenced margins. There were more species in fenced stream margins (4.4 species/site) than in unfenced margins (2.6 species/site). More small mammals overall were caught in fenced (21.2/site) than in unfenced (9.7/site) margins. Three species were sufficiently abundant to analyse individually. There were more individuals in fenced than unfenced margins for meadow voles Microtus pennsylvanicus (fenced: 8.0; unfenced: 5.3 individuals) and meadow jumping mouse Zapus hudsonius (fenced: 9.1; unfenced: 3.5 individuals). No significant difference was found for short-tailed shrew Blarina brevicauda (fenced: 3.8; unfenced: 2.4 individuals). Nine 100-m-long riparian margins, fenced one to two years previously, were compared with nine 100-m-long unfenced (cattle-grazed) riparian margins. Three types of small-mammal trap were operated continually throughout April–July in 1998–1999.


A replicated, randomized, controlled study in 1991–1994 in grassland and savanna in California, USA (Fehmi et al. 2005) found that excluding grazing livestock did not increase the number of California ground squirrel Spermophilus beecheyii burrows. Changes in the number of active ground squirrel burrows, relative to pre-experiment numbers, did not differ between ungrazed and grazed plots (60–100% vs 40–100% of pre-experiment numbers). The spatial distribution of active burrow entrances did not differ between ungrazed and grazed plots (2.6–3.4 vs 2.2–4.1 m between nearest burrows). Three sites, each with four plots, were studied. Half of plots were in grassland, and half were in savanna. Half had cattle-exclusion fencing and half were cattle-grazed from spring to summer. Three ground squirrel colonies were mapped in each plot in autumn 1991 (pre-experiment). Fencing was erected late in 1991 and burrows were further mapped in autumns of 1992–1994.


A replicated, site comparison study in 1999–2001 of a grassland area in Castilla y Lyón, Spain (Torre et al. 2007) found more small mammals in plots from which cattle were excluded, compared to grazed plots. More individual small mammals were caught in grazing exclusion plots (0–16 individuals/plot) than in grazed plots (0–3 individuals/plot). Three species of mammal were found; white-toothed shrew Crocidura russula (61.6% of captures), common vole Microtus arvalis (31.9%), and wood mouse Apodemus sylvaticus (6.5%). Six grazing exclusion plots (2–10 ha) were established in reforestation areas in grasslands grazed by 2–10 cattle/ha. These areas were reforested in 1990, but few planted trees survived. Eight live traps were placed in each of 22 trapping plots (11 inside and 11 outside cattle exclosures). Traps were operated for three consecutive nights during September–October 1999 and 2000 and in June 2000 and 2001.


A controlled study in 2005–2007 in open forest and scrubland at a site in Córdoba province, Spain (Catalán et al. 2008) found more European rabbits Oryctolagus cuniculus in a plot that was fenced to exclude large herbivores and with artificial warrens and water provided, than in an unmanaged area. Interventions were all carried out in the same plot, so their relative effects could not be separated. Average rabbit pellet counts were higher in the plot where the interventions were deployed (first year: 0.33 pellets/m2/day; second year: 1.08 pellets/m2/day) than in the unmanaged plot (first year: 0.02 pellets/m2/day; second year: 0.03 pellets/m2/day). A 2-ha plot was fenced to exclude large herbivores in March 2005. Rabbits and predators could pass through the fence. Five artificial warrens were installed and water was provided at one place. No management was carried out in an otherwise similar plot. Rabbit density was determined by monthly counts of pellets, from March 2005 to March 2007, in 0.5-m2 circles every 100 m along a 1-km transect in each plot.


A site comparison study in 1993–2007 on a shrubland site in South Australia, Australia (Read & Cunningham 2010) found that excluding cattle increased abundances of small mammals compared to high intensity cattle grazing but not to medium or low grazing intensities. The average number of small mammals/sample at ungrazed points (3.6 individuals) was higher than with intensive cattle grazing (1.7 individuals) but not higher than the numbers with medium- (5.0) or low-intensity cattle grazing (7.7). Species richness followed a similar pattern (ungrazed: 1.7 species; intensive grazing: 1.2 species; medium grazing: 1.7, low intensity grazing: 2.2 species). Livestock were fenced out from an approximately 9 × 9-km area in 1986. Small mammals were sampled using pitfall traps for a 10-day period in either December or January 1993–1996 and again in 2007. Five points were sampled inside the enclosure (ungrazed) with 13 outside (grazed). Cattle grazing intensity was determined by dung counts. Low intensity grazing was <12 dung/ha, medium grazing was 12–100 dung/ha and intensive grazing was >120 dung/ha.


A replicated, controlled study in 1998–2006 in sagebrush shrubland previously affected by wildfire in California, USA (Germano et al. 2012) found that excluding livestock did not alter the abundance of five small mammal species. Over eight years, abundance of San Joaquin antelope squirrel Ammospermophilus nelson did not differ significantly between areas where livestock were excluded (4–38 animals/plot) and grazed areas (2–29 animals/plot). The same pattern was true for short nosed kangaroo rat Dipodomys nitratoides nitratoides (1–55 vs 3–58 animals/plot), Heermann’s kangaroo rat Dipdomys heermanni (0–4 vs 0–22), giant kangaroo rat Dipodomys ingens (0–4 vs 0–3), and San Joaquin pocket mouse Perognathus inornatus inornatus (1–10 vs 1–17). Four 2.6-km2 areas were grazed by cattle and four 25-ha areas were fenced to exclude livestock. To estimate antelope squirrel abundance, 64 traps, baited with oats, at 40-m intervals, were established in each plot. To estimate abundance of other small mammals, 144 traps, baited with bird seed, were established in each plot at 10-m intervals. Traps were set for six consecutive days and nights in July–September 1998–2006.

Referenced papers

Please cite as:

Littlewood, N.A., Rocha, R., Smith, R.K., Martin, P.A., Lockhart, S.L., Schoonover, R.F., Wilman, E., Bladon, A.J., Sainsbury, K.A., Pimm S. and Sutherland, W.J. (2020) Terrestrial Mammal Conservation: Global Evidence for the Effects of Interventions for terrestrial mammals excluding bats and primates. Synopses of Conservation Evidence Series. University of Cambridge, Cambridge, UK.