• competition;
  • nestedness;
  • null models;
  • Rodentia;
  • Soricomorpha;
  • species richness


  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Identifying nonrandom species composition patterns predicted by assembly rules has been a central theme in community ecology. Few studies have investigated the prevalence of multiple drivers on species composition patterns in small mammal assemblages in the Old World. This study investigated seasonal changes in rodent and shrew diversity in eleven savannah vegetation types in South Africa. We tested whether species composition patterns are nonrandom with respect to predictions from Diamond's assembly rules, niche limitation hypothesis and nestedness hypothesis. Species richness estimators indicated that inventories for the rodents (80%) and shrews (100%) were relatively complete. Rodent (n = 11 species) diversity and shrew (n = 5 species) diversity were highest in summer and lowest in autumn. Rodent richness was highest in the Terminalia sericea bushveld and woodlands and lowest in the Drypetes arguta sand forest, whilst shrew richness was highest in the T. sericea bushveld and woodlands and lowest in the Acacia nilotica/Dichrostachys cinerea open shrub savannah. We found no support for the predictions of competition and nestedness hypotheses and suggest that this was probably due to the high seasonal and annual variability in rodent and shrew diversity.


Le fait d'identifier des schémas de composition des espèces prédits par des règles en matière d'assemblages est un thème central de l'écologie communautaire. Peu d'études ont étudié la prévalence de multiples facteurs sur les schémas de composition d'espèces dans les assemblages de petits mammifères de l'Ancien Monde. Cette étude s'est intéressée aux changements saisonniers de la diversité des rongeurs et des musaraignes dans 11 types de végétation de savane en Afrique du Sud. Nous avons testé si les schémas de composition d'espèces sont non aléatoires par rapport aux prédictions des règles d'assemblage de Diamond, l'hypothèse de limitation des niches et l'hypothèse d'« imbrication » (nestedness). Les estimateurs de la richesse en espèces indiquaient que les inventaires des rongeurs (80%) et des musaraignes (100%) étaient relativement complets. La diversité des rongeurs (n = 11 espèces) et des musaraignes (n = 5 espèces) était la plus élevée en été et la plus faible en automne. La richesse en rongeurs était la plus haute dans la brousse à Terminalia sericea et la forêt, et la plus faible dans la forêt sableuse à Drypetes arguta, alors que la richesse en musaraignes était la plus élevée dans la brousse à T. sericea et la forêt, et la plus faible dans la savane arbustive ouverte à Acacia-nilotica/Dichrostachys cinerea. Nous n'avons rien trouvé qui vienne étayer les prédictions des hypothèses de compétition et d'« imbrication », et nous suggérons que c'est probablement dû à la forte variabilité saisonnière et annuelle de la diversité des rongeurs et des musaraignes.


  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The question whether communities are influenced predominantly by biotic interactions such as competition, abiotic drivers such as fragmentation, or chance events has been a central theme in community ecology for at least a 100 years (Gotelli & Graves, 1996). Since Diamond's (1975) study of the coexisting bird species of the Bismarck Archipelago, the identification of nonrandom species composition patterns predicted by assembly rules has been at the centre of intense theoretical and statistical scrutiny (Gotelli & Graves, 1996). Powerful techniques such as null model analyses have revealed that animal assemblages comprising fewer co-occurring species than expected by chance, in line with Diamond's first assembly rule, are common (Gotelli & McCabe, 2002). There is also evidence for nonrandom patterns predicted by other competition hypotheses including Fox's favoured states model (Fox & Brown, 1993) and niche limitation hypothesis (Wilson, 1987).

In contrast to competition, the nestedness hypothesis (Patterson & Atmar, 1986) invokes abiotic mechanisms such as differential colonization or extinction rates of species, or distance and area effects, to explain nested species composition patterns where species at species-poor sites represent subsets of species at species-rich sites (Wright et al., 1998). Nestedness patterns have been described in insular assemblages (Patterson & Atmar, 1986) and in fragmented habitats (Boecklen, 1997; Fischer & Lindenmayer, 2005).

Given that biotic filters such as competition should have a strong influence on the community structure of animals such as bats that have life histories characterized by low fecundity, low predation risk, long life expectancy and stable populations (Schoeman & Jacobs, 2008, 2011), abiotic processes rather than competition should influence the species composition of similarly sized mammals such as rodents (order Rodentia) and shrews (order Soricomorpha) that have life histories characterized by fluctuating populations, high reproductive rates and short life expectancy. In support, there is evidence that rodent population numbers fluctuate seasonally and are positively correlated with temperature and rainfall (Muteka, Chimimba & Bebbett, 2006; Yarnell et al., 2007). Further, significant nested patterns have been detected in rodent assemblages from North American and Asian deserts (Patterson & Brown, 1991; Kelt et al., 1999), from Egypt (Abu Baker & Patterson, 2011), and in Finnish shrew assemblages (Patterson, 1990). However, nonrandom co-occurrence patterns consistent with competition hypotheses have been found in New World rodent assemblages (Brown & Kurzius, 1987; Kelt, Taper & Meserve, 1995; Kelt et al., 1999) and in Egypt (Abu Baker & Patterson, 2011), as well as in shrew assemblages in Australian and North American forests (Fox & Kirkland, 1992; McCay et al., 2004). However, in many studies, co-occurrence patterns of assemblages across large geographical scales comprising heterogeneous environmental conditions (e.g. vegetation types, topography, climate) are compared with predictions from either competition or nestedness hypotheses. Focussing on one process only and integrating heterogeneous sites in co-occurrence analyses might lead to false conclusions about species assembly because the effects of competition and nestedness cannot be disentangled (Gotelli & Graves, 1996).

In this study, we investigated the seasonal diversity of rodent and shrew assemblages in eleven different savannah vegetation types in South Africa. We used a battery of null model analyses to test predictions from Diamond's (1975) assembly rules, the niche limitation hypothesis (Wilson, 1987) and the nestedness hypothesis (Patterson & Atmar, 1986). Rodent and shrew assemblages were analysed separately because detection of nonrandom species composition patterns consistent with theory is more likely among ecologically and phylogenetically similar species (Patterson & Brown, 1991).

Materials and methods

  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study area

Study sites were situated in eleven different vegetation types (Fig. 1, Table 1) in Phinda Private Game Reserve (PPGR; 27°40′S–27°55′S; 31°12′E–32°26′E) that were classified by van Rooyen & Morgan (2007) based on dominant tree and shrub layers using the TURBOVEG and MEGATAB software (Hennekens & Schaminée, 2001). The minimum distance between study sites was 320 m, and the maximum distance was 21.7 km, which is probably outside the range of small mammal dispersal distances and home ranges (e.g. Monadjem et al., 2011); hence, species turnover among sites was unlikely. PPGR covers approximately 21,402 ha and is situated 30 km from the eastern coast of Maputaland, with the southern tip of the Lebombo mountains bordering the reserve on the south-west. Altitude ranges from 50 m in the north-east to 340 m in the south-west. The region experiences a hot and humid subtropical climate.


Figure 1. Map of Phinda Private Game Reserve showing the locations of the eleven study sites in different savannah vegetation types (after van Rooyen & Morgan, 2007). See Table 1 for abbreviation of vegetation types

Download figure to PowerPoint

Table 1. Vegetation types of Phinda Private Game Reserve (after van Rooyen & Morgan, 2007) where rodents and shrews were surveyed from July 2009 to May 2010
Vegetation typeAbbreviationDominant treesDominant shrubsSoil type
Acacia karroo shrub A. karroo A. karroo, Euclea divinorum, Acacia nilotica, Berchemia zeyheri, Ziziphus mucronata Dichrostachys cinerea, Gymnosporia senegalensis, Gymnosporia buxifolia Coarse sandy clay to clayey
Acacia luederitzii/E. divinorum dense thickets and woodlands A. luederitzii/E. divinorum A. luederitzii, Berchemia zeyheri, Ziziphus mucronata, Dombeya rotundifolia, Spirostachys africana, Balanites maughamii, Euphorbia cooperi, Schotia brachypetala, Galpinia transvaalica E. divinorum, Cissus rotundifolius, Rhoicissus tridentata, Phyllanthus reticulatus, Dombeya rotundifolia Poorly drained, clayey
A. nilotica open shrub savannah A. nilotica A. nilotica D. cinerea, Gymnosporia senegalensis, E. divinorum Dark clayey
A. nilotica/D. cinerea open shrub savannah A. nilotica/D. cinerea A. nilotica, D. cinerea, Ziziphus mucronata, Berchemia zeyheri, S. africana, Sclerocarya birrea E. divinorum, Gymnosporia senegalensis, Rhus quenzii, Coddia rudis, Euclea racemosa Fine sandy loam
A. nilotica/Hyphaene coriacea pan systems and woodclumps on termitaria A. nilotica/H. coriacea A. nilotica, H. coriacea, Ziziphus mucronata, Acacia burkei, Coddia rudis, D. cinerea, E. divinorum, Gymnosporia buxifolia, Gymnosporia senegalensis, Croton steenkampiana, Rhus guenzii, Euclea racemosa Coarse sandy loam to sandy clay loam
Combretum apiculatum open savannah and grasslands C. apiculatum C. apiculatum, Ziziphus mucronata, Acacia burkei, Pavetta edentula, Sclerocarya birrea, Ficus stuhlmannii, Ficus abutilifolia D. cinerea, E. divinorum, Rhus guenzii, Rhus gracillima, Gymnosporia senegalensis Fine sandy clay
Drypetes arguta sandforest D. arguta D. arguta, Newtonia hildebrandtii, Cleistanthus schlechteri, Wrightia natalensis, Strychnos henningsi Uvaria caffra, Salacia leptophylla, Toddaliopsis bremekampii, Cola greenwayii, Croton steenkampiana, Hyperacanthus amoenus Sandy loam
Floodplain grassland F. grassland NoneD. cinerea, Azima tetracantha, Coddia rudis, Rhus guenzii, Flueggia virosa,N/A
H. coriacea Palmveld H. coriacea H. coriacea NoneN/A
S. africana dense woodlands on floodplains and riverbanks S. africana Schotia brachypetala, Pappea capensis Ehretia rigida, Capparis tomentosa Poorly drained clayey
Terminalia sericea bushveld and woodlands T. sericea T. sericea, Combretum molle Brachylaena discolor, Strychnos madagascariensis, Strychnos spinosa, Acacia burkei, Ziziphus mucronata, Sclerocarya birrea Rhus guenzii, Dalbergia obovata, Grewia monticola, Schotia capitata, Coddia rudis, D. cinerea, Gymnosporia senegalensis Sandy

Rodent and shrew sampling

We sampled rodents and shrews during July 2009 (winter), November 2009 (spring), March 2010 (summer) and May 2010 (autumn). Rodents and shrews were trapped with pitfall traps and Supa-Kill MRT1 catch-alive rodent traps (Scientific Envirocare cc, Kempton Park). Each pitfall trap formation at a study site consisted of seven pitfall traps. Pitfall traps consisted of 20-l buckets that were buried in the ground with the rim of the bucket at ground level. The buckets were placed 4 m apart from rim to rim in a Y formation (Fig. 2). The arms of the Y formation were arranged at 120 degrees apart with a 0.4 m high drift fence made of plastic sheeting, anchored with metal droppers, at 1 m intervals connecting the pitfall traps. The pitfall traps were not baited and left open for seven consecutive nights. Sixteen Supa-Kill MRT1 catch-alive rodent traps were set on five consecutive nights per site. The traps were placed on the ground in a 4 × 4 trapping grid with 6 m between transects and between traps (Fig. 2) Traps were checked every morning (6–8 h) and every afternoon (14.30–17.00 hours) when the catch-alive traps were re-baited with a mixture of peanut butter, oats and sunflower oil (Hughes, Ward & Perrin, 1994).


Figure 2. Pitfall trap formation and catch-alive trap grid used at each study site

Download figure to PowerPoint

We identified captured rodent and shrew species in the field based on the keys of Taylor (1998) and Skinner & Chimimba (2005). We released rodents that could be identified at the point of capture after clipping the fur on the dorsal side to avoid recording the same individual twice. Rodents that could not be identified to species level, and shrews were euthanized and deposited as voucher specimens in the Durban Natural Science Museum.

Statistical analysis

Richness and abundance (number of individuals) data were square root–transformed to meet the assumptions of normality and equal variances. We used two-way ANOVAs to determine the effects of season and vegetation types, as well as the interaction effect between season and vegetation type, on rodent and shrew species richness and abundance. Post hoc Tukey tests were performed on significant ANOVAs. We used Statistica version 6.0 (Statsoft, 2004) to perform statistical calculations.

To assess the completeness of our inventories, we used species richness estimators (Gotelli & Colwell, 2001). Species richness estimators extrapolate the expected number of species based on the sampling effort (Gotelli & Colwell, 2001). We used two species richness estimators; the Chao 1 (Chao, 1987) is sensitive to both the number of singletons and doubletons, whilst the Jackknife 1 (Burnham & Overton, 1978) is mostly sensitive to the number of singletons. These estimators have been shown to perform well, even in datasets with a limited number of samples (Walter & Morand, 1998).

To standardize comparisons of rodent and shrew species richness in the different vegetation types, we used sample-based rarefaction curves (Gotelli & Colwell, 2001). Sample-based rarefaction assumes random sampling from similar sized areas with randomly distributed individuals that are taxonomically similar and takes heterogeneity of the data into account (Gotelli & Colwell, 2001). To calculate species richness estimators and perform rarefaction, we used the software program EstimateS (version 8.2, Colwell, 2009).

To test for nonrandom species composition patterns, we used the co-occurrence module of Ecosim software (version 7.70, Gotelli & Entsminger, 2004). We used four indices to quantify species composition patterns. The C-score (Stone & Roberts, 1990) measures the average number of checkerboard units between all possible pairs of species and should be significantly larger than expected by chance in a competitively structured assemblage. The number of checkerboard species index quantifies the number of species pairs that never co-occur in any site and should also be significantly larger than expected by chance in a competitively structured assemblage. The number of species combinations (Pielou & Pielou, 1968) tests Diamond's (1975) first and second assembly rules: there should be significantly fewer unique species-pair combinations. The niche limitation hypothesis predicts that in a competitively structured assemblage, the variance of species richness, quantified by the V-ratio (Schluter, 1984), should be significantly smaller than expected by chance.

The Sim9 algorithm (Gotelli & Entsminger, 2004) was used to randomize the original matrix, that is, row and column totals were fixed. This algorithm has a good Type 1 error rate and is powerful in detecting nonrandom patterns even in noisy data sets (Gotelli & Entsminger, 2004).

Nestedness of rodent and shrew matrices was assessed in the four seasons using BINMATNEST (Rodríguez-Gironés & Santamaría, 2006). BINMATNEST reorders the presence/absence matrix and then calculates the degree of order or disorder in the maximally packed binary matrix as a measure in temperature, which ranges from 0 (absolute nestedness) to 100 (no nestedness). We quantified the degree of nestedness with null model three as it has a low risk of Type I error (Rodríguez-Gironés & Santamaría, 2006).


  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Rodent and shrew diversity

In total, eleven rodent species were captured, representing eight genera from one family, and five shrew species, representing two genera from one family (Table 2). The Chao 1 and Jackknife 1 species richness estimators indicated that our species inventories were ca. 80% complete for the rodents and 100% complete for the shrews.

Table 2. Abundance and species richness of rodents and shrews in eleven vegetation types of Phinda Private Game Reserve, surveyed in 2009 and 2010. See Table 1 for description of vegetation types
SpeciesnVegetation types
Acacia luederitzii/Euclea divinorum Acacia karroo Acacia nilotica A. nilotica/Dichrostachys cinerea A. nilotica/Hyphaene coriacea Combretum apiculatum Drypetes arguta Floodplain grassland H. coriacea Spirostachys africana Terminalia sericea
Aethomys ineptus 8 11    3111
Dendromus melanotis 13    7 1   8
Dendromus mystacalis 7   113 1 1 
Dendromus mesomelas 1          1
Dendromus cf. nyikae 1        1  
Lemniscomys rosalia 157  1 2    5
Mastomys natalensis 61 22113117  6 1
Mus minutoides 393416510 3 34
Steatomys pratensis 12 11311  3 2
Gerbilliscus leucogaster 8    5   3  
Saccostomus campestris 1   1       
Species richness 24466513537
Crocidura fuscomurina 1811 413 33 2
Crocidura hirta 1324   1 3 12
Crocidura silacea 7 3       13
Suncus infinitesimus 6 1  12 1  1
Suncus lixus 92    1 41 1
Species richness 34012404225

Temperature and precipitation varied among seasons (Winter: Max 22.5 ± 3.0°C, Min 9.7 ± 1.7°C, 0 mm; Spring: Max 24.3 ± 4.0°C, Min 15.0 ± 2.5°C, 80.8 ±9.5 mm; Summer: Max 28.3 ± 2.5°C, Min 17.9 ±1.5°C, 47.7 ± 5.0 mm; Autumn: Max 26.6 ± 3.1°C, Min 14.1 ± 2.1°C, 7.2 ± 3.7 mm; South Africa Weather Bureau). There were significant differences in rodent species richness and abundance among seasons (F3,304 = 6.09, P < 0.001; I3,304 = 5.65, P < 0.01, respectively), vegetation types (F10,297 = 4.60, P < 0.001; F10,297 = 4.96, P < 0.001, respectively) and the interaction between season and vegetation type (F30,278 = 2.01, P < 0.01; F30,278 = 1.81, P < 0.01, respectively). Post hoc Tukey tests revealed that rodent species richness and abundance were highest in summer and lowest in autumn. Sample-based rarefaction showed that rodent species richness was highest in the Terminalia sericea bushveld and woodlands and lowest in the Drypetes arguta sandforest (Fig. 3a).


Figure 3. Rarefaction curves for (a) rodents and (b) shrews captured during four seasons between 2009 and 2010 in different vegetation types in Phinda Private Game Reserve. Rodents: Acacia karroo (SD ± 0.88), Acacia luederitzii/Euclea divinorum (SD ± 0), Acacia nilotica (SD ± 1.54), A. nilotica/Dichrostachys cinerea (SD ± 1.59), A. nilotica/Hyphaene coriacea (SD ± 1.43), Combretum apiculatum (SD ± 0.66), Floodplain grassland (SD ± 0), H. coriacea (SD ± 1.18), Spirostachys africana (SD ± 0.85), Terminalia sericea (SD ± 1.64). Shrews: A. karroo (SD ± 1.14), A. luederitzii/E. divinorum (SD ± 0.47), A. nilotica/D. cinerea (SD ± 0), A. nilotica/H. coriacea (SD ± 0.81), C. apiculatum (SD ± 1.14), F. grassland (SD ± 0.66), H. coriacea (SD ± 0.62), S. africana (SD ± 0.81), T. sericea (SD ± 0.75). See Table 1 for description of vegetation types

Download figure to PowerPoint

Mus minutoides occurred in the greatest number of vegetation types, whilst Mastomys natalensis was the most frequently captured rodent species (Table 2). Abundance of M. minutoides was significantly different among seasons (F3,304 = 3.7997, P < 0.01) but not among vegetation types or the interaction between season and vegetation type (all P > 0.05). Abundance of M. natalensis differed significantly among the different vegetation types (F10,297 = 8.613, P < 0.001), but not among seasons or interaction between vegetation type and season (all P > 0.05).

There were significant differences in shrew species richness and abundance among seasons (F3,304 = 5.14, P < 0.001; F3,304 = 4.91, P < 0.01, respectively), vegetation types (F10,297 = 2.71, P < 0.01; F10,297 = 2.62, P < 0.01, respectively) and the interaction between season and vegetation type (F30,278 = 2.63, P < 0.0001; F30,278 = 2.75, P < 0.001, respectively). Post hoc Tukey tests revealed that shrew species richness was highest in summer and lowest in autumn. Shrew species richness was highest in the T. sericea bushveld and woodlands and lowest in the Acacia nilotica/Dichrostachys cinerea open shrub savannah (Fig. 3b).

Crocidura fuscomurina occurred in the greatest abundance and vegetation types (Table 2). Abundance of C. fuscomurina was not significantly different among seasons, vegetation types or the interaction between season and vegetation types (all P > 0.5).

Nonrandom co-occurrence patterns predicted by competition and nestedness hypotheses

We found no support for competition hypotheses. The observed C-score, number of species combinations and number of checkerboard species pairs were not significantly different from scores expected by chance (Table 3). Additionally, the variances in species richness between rodent and shrew assemblages were not significantly smaller than expected by chance (Table 3). Rodent and shrew assemblages were also not significantly nested (Table 3).

Table 3. Observed and expected C-score, number of species combinations (No. spp comb), number of checkerboard species pairs (No. check spp pairs), V-ratio indices and nestedness temperature (Tobs) of rodent and shrew assemblages in Phinda Private Game Reserve
MatrixDiamond's assembly rulesNiche limitationNestedness
C-scoreNo. spp combNo. check spp pairsV-ratioTemperature
ObsExpP valueObsExpP valueObsExpP valueObsExpP valueTobs (°)P value
All seasons3.943.930.4610.0010.930.0611.0013.040.221.601.601.0023.160.11
All seasons2.802.810.649.008.880.720.


  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We collected a total of sixteen rodent and shrew species, and species richness estimators indicated that the inventories were relatively complete. Dendromus cf. nyikae is a first record for KwaZulu-Natal; specimens have previously been sampled in the Limpopo Province, South Africa and the Inyanga district of eastern Zimbabwe (Skinner & Chimimba, 2005). Although C. fuscomurina is relatively rare in KwaZulu-Natal (Taylor, 1998), it was the most frequently caught shrew species during this survey. Similarly, we regularly caught Suncus lixus and Suncus infinitesimus that are rare in museum collections (P.J. Taylor, personal communication). However, our estimates of species richness may still be conservative, reflected in part by the small sample sizes for many species. Using a larger sampling effort, four additional rodent species, Steatomys krebsii, Thallomys paedulcus, Graphiurus murinus and Grammomys dolichurus, were captured at the neighbouring Mkuze Game Reserve (Taylor et al., 2007). Furthermore, our sampling techniques were probably ineffective for capturing arboreal species like Thallomys paedulcus, G. murinus and G. dolichurus that are also known to be trap-shy (A. Monadjem, personal communication).

In this study, rodent and shrew species richness and abundance varied seasonally, being highest during the wet season and lowest during the dry season. There is evidence that the seasonal variation in rainfall influences the onset and termination of the breeding season of small mammals (Leirs, Verhagen & Verheyen, 1994; Monadjem, 1998; Makundi, Massawe & Mulungu, 2006). During the rainy season, resource availability is at its highest (Lima et al., 2001), including vegetation cover, seed densities (Gutiérrez et al., 1993) and invertebrate abundance (Mortelliti & Boitani, 2009). On the other hand, many rodent assemblages in southern Africa exhibit the opposite pattern with highest numbers at the end of summer or autumn and lowest numbers in spring or early summer (Monadjem & Perrin, 2003; Avenant & Cavallini, 2007).

Species richness of rodents was highest in the T. sericea bushveld and woodlands and lowest in the D. arguta sandforest. This may be a reflection of the high plant species diversity of the T. sericea bushveld and woodlands (van Rooyen & Morgan, 2007) that, in turn, provide food and habitat resources to resident small mammals (Tews et al., 2004). Conversely, the D. arguta sandforest is characterized by poorly developed ground layer, with little or no grass and sandy soils (Mucina & Geldenhuys, 2006). Although there is typically a positive correlation between rodent species richness and vegetation complexity (Els & Kerley, 1996; Monadjem, 1997; van Deventer & Nel, 2006), the relationship is a complex one. It is dependent on how rodents perceive their habitat and may vary considerably between and within species (Tews et al., 2004). For example, Dendromus melanotis is normally associated with tall stands of grasses such as Hyparrhenia and Merxmuellera spp. and shrubs of the savannah biome (Skinner & Chimimba, 2005), yet we caught one specimen in the D. arguta sandforest. Generalist species that have wide habitat tolerances and broad diets, for example, M. minutoides and M. natalensis (Monadjem, 1997; van Deventer & Nel, 2006; Mulungu et al., 2011), were caught in most of PPGR's vegetation types.

Shrew species richness was also highest in the T. sericea bushveld and woodlands but lowest in the A. nilotica/D. cinerea open shrub savannah. Crocidura fuscomurina and Crocidura hirta were the most frequently captured shrew species at most of the sites, suggesting both species have a wide habitat tolerance. According to Skinner & Chimimba (2005) and Taylor (1998), C. hirta prefers habitats characterized by dense vegetation with deep litter and proximity to water, but there is a paucity of information on habitat requirements for C. fuscomurina.

We found no support for Diamond's (1975) assembly rules or the niche limitation (Wilson, 1987) hypothesis in rodent or shrew assemblages, even in the dry season when resources are probably more limiting than in the wet season. Our results are therefore consistent with the hypothesis that competition should have a minor influence, if any, on the community structure of rodents and shrews that have life histories characterized by high fecundity, high predation risk and short life expectancies (Bronson, 1985). Instead, abiotic processes such as rainfall (Leirs, Verhagen & Verheyen, 1994) and fire (Monadjem & Perrin, 2003) may drive assemblage patterns. On the other hand, morphological and behavioural characteristics of species not quantified in this study may facilitate resource partitioning and coexistence. For example, there is a significantly positive relationship between body size of shrews and the size of invertebrate prey (Pernetta, 1976), and rodents can partition habitat by vertical stratification of foraging activities (Maisonneuve & Rioux, 2001; Hannibal & Caceres, 2010).

Although nestedness is a common phenomenon in many ecological systems (Patterson, 1990; Patterson & Brown, 1991; Ulrich & Gotelli, 2007), we found no evidence of significantly nested patterns in the assemblages. Nested patterns were also not found in rodent assemblages in the Eastern Cape, South Africa (Kryštufek, Haberl & Baxter, 2008). Three mechanisms are necessary for the development of a nested community structure: a common biogeographical history, similar ecologically comparable environments and the hierarchical organization of niche relationships between species (Patterson & Brown, 1991). The rodent assemblages of the different vegetation types experienced various levels of disturbance at different times. For example, after proclamation in 1991, PPGR consisted of two separate areas, the northern section which included the T. sericea, Acacia nilotica and Hyphaene coriacea vegetation types and the southern section which included the Acacia nilotica/Euclea divinorum, Acacia karroo, Spirostachys africana, Floodplain grassland and Combretum apiculatum vegetation types. It is also unlikely that similar biotic and abiotic conditions characterize the vegetation types because they differ in dominant tree and shrub species, and soil types. Hence, species diversity of rodent and shrew assemblages varied both seasonally and yearly in the different vegetation types. Hierarchical organization of niche relationships requires graded differences in factors such as colonization abilities, extinction risk or overlapping resource requirements (Kelt et al., 1999), but our data do not test these hypotheses.

To conclude, rodents and shrews sustain many vertebrate predators in healthy ecosystems, and contribute significantly to the cycling of nitrogen and other nutrients in grasslands through the deposition of urine and faeces (Clark et al., 2005). Moreover, rodents act as keystone species in many ecosystems (Ernst & Brown, 2001) and are therefore useful indicator species in predicting the consequences of human land use or climate change (Cameron & Scheel, 2001). It is therefore important to accurately survey small mammal diversity and identify the processes that drive their community assembly. Species richness estimators indicated that our richness estimates were relatively complete, hence justifying our null model analyses on species composition patterns. One reason we found no support for predictions from competition and the nestedness hypotheses may be because rodent and shrew diversity varied between seasons and vegetation types. Alternatively, our classification of assemblages based on vegetation types may be at the wrong spatial resolution. More broad-scale assemblages defined on different vegetation classifications (e.g. Mucina, Rutherford & Powry, 2005) or alternative environmental features (e.g. soil or climate) may reveal nonrandom species composition patterns. Similarly, fine-scaled vegetation characteristics not quantified in this study, such as vertical grass and shrub structure, may be significantly correlated with rodent and shrew species composition (Layme, Lima & Ernst, 2004). Future work should combine abundance data with species composition indices (Ulrich & Gotelli, 2011) at different spatial scales and in opposition to processes such as predation, climate and microhabitat characteristics to better determine the extent of competition and nestedness on rodent and shrew species composition.


  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We are grateful to the management of PPGR and the Nyati antipoaching unit for their logistic support and all the volunteers for their assistance in the field. We thank four anonymous reviewers for their comments on a previous version of this manuscript. MCS is grateful to the University of KwaZulu-Natal for financial support.


  1. Top of page
  2. AbstractRésumé
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Abu Baker, M. & Patterson, B.D. (2011) Patterns in the local assembly of Egyptian rodent faunas: co-occurrence and nestedness. J. Arid Environ. 75, 1419.
  • Avenant, N.L. & Cavallini, P. (2007) Correlating rodent community structure with ecological integrity, Tussen-die-Riviere Nature Reserve, Free State province, South Africa. Integr. Zool. 2, 212219.
  • Boecklen, W.J. (1997) Nestedness, biogeographic theory, and the design of nature reserves. Oecologia 112, 123142.
  • Bronson, F.H. (1985) Mammalian reproduction: an ecological perspective. Biol. Reprod. 32, 126.
  • Brown, J.H. & Kurzius, M.A. (1987) Composition of desert rodent faunas: combinations of coexisting species. Ann. Zool. Fenn. 24, 227237.
  • Burnham, K.P. & Overton, W.S. (1978) Estimation of the size of a closed population when capture probabilities vary among animals. Biometrica 65, 623633.
  • Cameron, G.N. & Scheel, D. (2001) Getting warmer: effects of global climate change on distribution of rodents in Texas. J. Mammal. 82, 652680.
  • Chao, A. (1987) Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43, 783791.
  • Clark, J.E., Hellgren, E.C., Parsons, J.L., Jorgensen, E.E., Engele, D.M. & Leslie, D.M. Jr (2005) Nitrogen outputs from faecal and urine deposition of small mammals: implications for nitrogen cycling. Oecologia 144, 447455.
  • Colwell, R.K. (2009) EstimateS version 8.2: Statistical estimation of species richness and shared species from samples. (accessed on 25 October 2012).
  • van Deventer, M. & Nel, J.A.J. (2006) Habitat, food, and small mammal community structure in Namaqualand. Koedoe 49, 99109.
  • Diamond, J.M. (1975) Ecology and Evolution. Harvard University Press, Cambridge.
  • Els, L.M. & Kerley, G.I.H. (1996) Biotic and abiotic correlates of small mammal community structure in the Groendal Wilderness Area, Eastern Cape, South Africa. Koedoe 39, 121130.
  • Ernst, S.K.M. & Brown, J.H. (2001) Delayed compensation for missing keystone species by colonisation. Science 292, 101104.
  • Fischer, J. & Lindenmayer, D.B. (2005) Perfect nested or significantly nested – an important difference for conservation management. Oikos 109, 458494.
  • Fox, B.J. & Brown, J.H. (1993) Assembly rules for functional groups in North American desert rodent communities. Oikos 67, 358370.
  • Fox, B.J. & Kirkland, G.L. Jr (1992) An assembly rule for functional groups applied to North American soricid communities. J. Mammal. 73, 491503.
  • Gotelli, N.J. & Colwell, R.K. (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecol. Lett. 4, 379391.
  • Gotelli, N.J. & Entsminger, G.L. (2004) EcoSim: Null models software for ecology. Version 7.0. Acquired Intelligence Inc. & Kesey Bear. (accessed on 11 October 2011).
  • Gotelli, N.J. & Graves, G.R. (1996) Null Models in Ecology. Smithsonian Institution Press, Washington.
  • Gotelli, N.J. & McCabe, D.J. (2002) Species co-occurrence: a meta-analysis of J.M. Diamond's assembly rules model. Ecology 83, 20912096.
  • Gutiérrez, J.R., Meserve, P.L., Jaksic, F.M., Contreras, L.C., Herrera, S. & Vásquez, H. (1993) Structure and dynamics of vegetation in a Chilean arid thorn scrub community. Acta Oecol. 14, 271285.
  • Hannibal, W. & Caceres, N.C. (2010) Use of vertical space by small mammals in gallery forest and woodland savannah in south-western Brazil. Mammalia 74, 247255.
  • Hennekens, S.M. & Schaminée, J.H.J. (2001) TURBOVEG, a comprehensive data base management system for vegetation data. J. Veg. Sci. 12, 589591.
  • Hughes, J.J., Ward, D. & Perrin, M.R. (1994) Predation risk and competition affect habitat selection and activity of Namib Desert gerbils. Ecology 75, 13971405.
  • Kelt, D.A., Taper, M.L. & Meserve, P.L. (1995) Assessing the impact of competition on community assembly: a case study using small mammals. Ecology 76, 12831296.
  • Kelt, D.A., Rogovin, K., Shenbrot, G. & Brown, J.H. (1999) Patterns in the structure of Asian and North American desert small mammal communities. J. Biogeogr. 26, 825841.
  • Kryštufek, B., Haberl, W. & Baxter, R.M. (2008) Rodent assemblage in a habitat mosaic within the Valley Thicket vegetation of the Eastern Cape Province, South Africa. Afr. J. Ecol. 46, 8087.
  • Layme, V.A., Lima, A.P. & Ernst, W. (2004) Effects of fire, food availability and vegetation on the distribution of the rodent Bolomys lasiurus in an Amazonian Savanna. J. Trop. Ecol. 20, 183187.
  • Leirs, H., Verhagen, R. & Verheyen, W. (1994) The basis for reproductive seasonality in Mastomys rat (Rodentia: Muridae) in Tanzania. J. Trop. Ecol. 10, 5566.
  • Lima, M., Juillard, R., Stenseth, N.C. & Jaksic, F. (2001) Demographic dynamics of a neotropical small rodent (Phylotis darwini): feedback structure, predation and climatic factors. J. Anim. Ecol. 70, 761775.
  • Maisonneuve, C. & Rioux, S. (2001) Importance of riparian habitats for small mammal and herpetofaunal communities in agricultural landscapes of southern Québec. Agric. Ecosyst. Environ. 83, 165175.
  • Makundi, R.H., Massawe, A.W. & Mulungu, L.S. (2006) Breeding seasonality and population dynamics of three rodent species in the Magamba Forest Reserve, Western Usambara Mountains, north-east Tanzania. Afr. J. Ecol. 45, 1721.
  • McCay, T.S., Lovallo, J., Ford, W.M. & Menzel, M.A. (2004) Assembly rules for functional groups of North American shrews: effects of geographic range and habitat partitioning. Oikos 107, 141147.
  • Monadjem, A. (1997) Habitat preferences and biomass of small mammals in Swaziland. Afr. J. Ecol. 35, 6472.
  • Monadjem, A. (1998) Reproductive biology, age structure, and diet of Mastomys natalensis (Muridae: Rodentia) in a Swaziland grassland. Z. Saugetierkd. 63, 347356.
  • Monadjem, A. & Perrin, M. (2003) Population fluctuations and community structure of small mammals in a Swaziland grassland over a three-year period. Afr. Zool. 38, 127137.
  • Monadjem, A., Mahlaba, T.A., Dlamini, N., Eiseb, S.J., Belmain, S.R., Mulungu, L.S., Massawe, A.W., Makundi, R.H., Mohr, K. & Taylor, P.J. (2011) Impact of crop cycle movement patterns of pest rodent species between field and houses in Africa. Wildl. Res. 38, 603609.
  • Mortelliti, A. & Boitani, L. (2009) Distribution and coexistence of shrews in patchy landscapes: a field test of multiple hypotheses. Acta Oecol. 35, 797804.
  • Mucina, L. & Geldenhuys, C.J. (2006) Afrotemperate, subtropical and azonal forests. The Vegetation of South Africa, Lesotho and Swaziland. Vol. 19 (Eds L. Mucina and M.C. Rutherford). South African National Biodiversity Institute, Pretoria.
  • Mucina, L., Rutherford, M.C. & Powry, L.W. (2005) Vegetation Map of South Africa, Swaziland and Lesotho. SANBI, Pretoria.
  • Mulungu, L.S., Mahlaba, T.A., Massawe, A.W., Kennis, J., Crauwels, D., Eised, S., Mondjem, A., Makundi, R.H., Katakweba, A.A.S., Leirs, H. & Belmain, S.R. (2011) Dietary differences of the multimammate mouse, Mastomys natalensis (Smith, 1834), across different habitats and seasons in Tanzania and Swaziland. Wildl. Res. 38, 640646.
  • Muteka, S.P., Chimimba, C.T. & Bebbett, N.C. (2006) Reproductive seasonality in Aethomys namaquensis (Rodentia: Muridae) from Southern Africa. J. Mammal. 87, 6774.
  • Patterson, B.D. (1990) On the temporal development of nested subset patterns of species composition. Oikos 59, 330342.
  • Patterson, B.D. & Atmar, W. (1986) Nested subsets and the structure of insular mammalian faunas and archipelagos. Biol. J. Linn. Soc. 28, 6582.
  • Patterson, B.D. & Brown, J.H. (1991) Regionally nested patterns of species composition in granivorous rodent assemblages. J. Biogeogr. 18, 395402.
  • Pernetta, J.C. (1976) Diets of the shrews Sorex araneus L. and Sorex minutus L. in Wytham grassland. J. Anim. Ecol. 45, 899912.
  • Pielou, D.P. & Pielou, E.C. (1968) Association among species of infrequent occurrence: the insect and spider fauna of Polyporus betulinus (Bulliard) Fries. J. Theor. Biol. 21, 202216.
  • Rodríguez-Gironés, M.A. & Santamaría, L. (2006) A new algorithm to calculate the nestedness temperature of presence-absence matrices. J. Biogeogr. 33, 924935.
  • van Rooyen, N. & Morgan, S. (2007) Vegetation Types of Phinda Private Game Reserve. Ecotrust CC. report, Phinda Private Game Reserve.
  • Schluter, D. (1984) A variance test for detecting species associations, with some example applications. Ecology 65, 9981005.
  • Schoeman, M.C. & Jacobs, D.S. (2008) The relative influence of competition and prey defenses on the phenotypic structure of insectivorous bat ensembles in southern Africa. PLoS One 3, e3715.
  • Schoeman, M.C. & Jacobs, D.S. (2011) The relative influence of competition and prey defenses on the trophic structure of animalivorous bat ensembles. Oecologia 166, 493506.
  • Skinner, J.D. & Chimimba, C.T,. (2005) The Mammals of the Southern African Subregion. Cambridge University Press, Cape Town.
  • Stone, L. & Roberts, A. (1990) The checkerboard score and species distributions. Oecologia 85, 7479.
  • Taylor, P.J. (1998) The Smaller Mammals of KwaZulu-Natal. University of Natal Press, Pietermaritzburg.
  • Taylor, P.J., Rautenbach, A., Schoeman, M.C. & Combrink, X. (2007) A Winter Survey of the Smaller Mammals of the uMkhuze Section of the iSimangaliso Wetland Park, KwaZulu-Natal Province, South Africa. Operation Wallacea Ltd., Lincolnshire.
  • Tews, J., Brose, U., Grimm, V., Tielbörger, K., Wichmann, M.C., Schwager, M. & Jeltch, F. (2004) Animal species diversity driven by habitat heterogeneity/diversity: the importance of keystone structures. J. Biogeogr. 31, 7992.
  • Ulrich, W. & Gotelli, N.J. (2007) Null model analysis of species nestedness patterns. Ecology 88, 18241831.
  • Ulrich, W. & Gotelli, N.J. (2011) Null model analysis of species associations using abundance data. Ecology 9, 33843397.
  • Walter, B.A. & Morand, S. (1998) Comparative performance of species richness estimation methods. Parasitology 116, 395405.
  • Wilson, J.B. (1987) Methods for detecting non-randomness in species co-occurrences: a contribution. Oecologia 73, 579582.
  • Wright, D.H., Patterson, B.D., Mikkelson, G.M., Cutler, A. & Atmar, W. (1998) A comparative analysis of nested subset patterns of species composition. Oecologia 113, 120.
  • Yarnell, R.W., Scott, D.M., Chimimba, C.T. & Metcalfe, D.J. (2007) Untangling the roles of fire, grazing and rainfall on small mammal communities in grassland ecosystems. Oecologia 154, 387402.