Highlights
-
• Microhabitat sharing between host taxa correlates with more similar assemblages among host-generalist nematodes.
-
• Host-restricted nematodes (notably Syphacia spp.) contribute an independent coevolutionary signal to between-host dissimilarity.
-
• Opportunistic host behaviour leads to more dissimilar nematode assemblages in rodents.
-
• Host factors have a greater influence on nematode assemblages than the environment.
Introduction
Parasites are omnipresent, and their diversity is characterised by unique life history traits such as host and environmental association and a specific mode of transmission (Lee Reference Lee and Lee2002; Morand et al Reference Morand, Bouamer, Hugot, Morand, Krasnov and Poulin2006). This has led to several studies investigating how parasite communities are assembled and which factors drive spatial variation in these communities (e.g. Bush Reference Bush, Esch, Bush and Aho1990; Froeschke et al Reference Froeschke, Harf, Sommer and Matthee2010, Reference Froeschke, van der Mescht, McGeoch and Matthee2013; Krasnov et al Reference Krasnov, Shenbrot, Mouillot, Khokhlova and Poulin2005; Spickett et al Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017b; Timi and Poulin Reference Timi and Poulin2003). Parasite communities comprise parasite populations that share a spatiotemporal unit (Bush et al Reference Bush, Lafferty, Lotz and Shostak1997) and can be expressed as parasite species within a host individual (infracommunity) or host population (component community) (Bush and Holmes Reference Bush and Holmes1986; Holmes and Price Reference Holmes, Price, Kikkawa and Anderson1986). Since infracommunity structure is linked to the lifespan of the host, it is more prone to stochastic events compared to component communities that are formed over longer evolutionary time scales (Bush et al Reference Bush, Lafferty, Lotz and Shostak1997; Kennedy Reference Kennedy, Esch, Bush and Aho1990; Poulin Reference Poulin2007; Timi and Poulin Reference Timi and Poulin2003). In addition, most parasite life cycles include on-host and off-host stages, and consequently, both host and environmental factors can drive spatial variation in parasite communities (Fournié et al Reference Fournié, Goodman, Cruz, Cedeño, Vélez, Patiño, Millins, Gibbons, Fox and Cunningham2015; Warburton et al Reference Warburton, Budischak, Jolles and Ezenwa2023).
This study sets out to further elucidate the factors that drive spatial variation in endoparasitic nematode communities. The latter can broadly be separated into monoxenous (direct life cycle) and heteroxenous (indirect life cycle) taxa (Durette-Desset and Digiani Reference Durette-Desset and Digiani2005; Jrijer et al Reference Jrijer, Bordes, Morand and Neifar2015; Lee Reference Lee and Lee2002; Poulin Reference Poulin2007). Since parasitic nematodes are reliant on their hosts for their development, reproduction, and dispersal (Lee Reference Lee and Lee2002), it is expected that their community composition can be influenced by host characteristics such as diet, sociality, and geographic range (Arneberg Reference Arneberg2002; Bordes et al Reference Bordes, Blumstein and Morand2007; Leung and Koprivnikar Reference Leung and Koprivnikar2019). In addition, sympatric and phylogenetically closely related host species often harbour more similar nematode communities compared to allopatric and more unrelated host species (Poulin Reference Poulin2010; Titcomb et al Reference Titcomb, Pansu, Hutchinson, Tombak, Hansen, Baker, Kartzinel, Young and Pringle2022). Ectothermic nematode communities, however, can also be influenced by the external environment that the host species occupies (Froeschke and Matthee Reference Froeschke and Matthee2014; Froeschke et al Reference Froeschke, Harf, Sommer and Matthee2010; O’Connor et al Reference O’Connor, Walkden-Brown and Kahn2006) and, in particular, climate, soil type, and vegetation structure (Brouat et al Reference Brouat, Kane, Diouf, Bâ, Sall-Dramé and Duplantier2007; Lee Reference Lee and Lee2002; Solórzano-García et al Reference Solórzano-García, White and Shedden2023). In addition, for heteroxenous species, environmental conditions can also influence the survival and distribution of intermediate hosts, which are often arthropods (D’Bastiani et al Reference D’Bastiani, Campião, Boeger and Araújo2020). Irrespective of what the determining factors are, the presence of a nematode species in the definitive host at a specific locality implies that the environment at that locality is favourable for the survival of the specific nematode species (Morand et al Reference Morand, Bouamer, Hugot, Morand, Krasnov and Poulin2006).
To date, most previous studies investigating spatial variation in nematode assemblages have focused on fish, birds, and domestic animals (e.g. Bush Reference Bush, Esch, Bush and Aho1990; Carney and Dick Reference Carney and Dick2000; Karvonen and Valtonen Reference Karvonen and Valtonen2009; Karvonen et al Reference Karvonen, Cheng and Valtonen2005; Poulin Reference Poulin2010; Timi and Poulin Reference Timi and Poulin2003) and most studies dealing with rodents are restricted to one sampling locality (Klimpel et al Reference Klimpel, Förster and Schmahl2007; Ondríková et al Reference Ondríková, Miklisová, Ribas and Stanko2010), one host species (Montgomery and Montgomery Reference Montgomery and Montgomery1989), or have described parasite diversity and broad community composition and lacked life-history-based host comparisons (Chaisiri et al Reference Chaisiri, Chou, Siew, Morand and Ribas2017). In southern Africa, knowledge on the spatial variation of nematode assemblages associated with rodents is limited to a single study that compared two allopatric congeneric host species (Rhabdomys pumilio and Rhabdomys dilectus) (Spickett et al Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017b). We propose that to properly address the relative contribution of host traits and environmental factors on the spatial variation in helminth community compositions, multiple sympatric rodent species across different localities should be studied.
In South Africa, two Rhabdomys species, Micaelamys namaquensis, and Otomys unisulcatus (all family Muridae) occur sympatrically in the Western Cape Province (WCP) (Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015; Skinner and Chimimba Reference Skinner and Chimimba2005). In this region, the four-striped grass rat, Rhabdomys, is represented by R. pumilio in the Fynbos and Succulent Karoo biomes and R. intermedius in the Nama-Karoo biome (du Toit et al Reference du Toit, Jansen van Vuuren, Matthee and Matthee2012; Ganem et al Reference Ganem, Dufour, Avenant, Caminade, Eiseb, Tougard and Pillay2020; Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015). These two species are cryptic and, based on the currently available data, they are both diurnal, live in family groups and make simple nests below shrubs (Christian Reference Christian1977; Curtis and Perrin Reference Curtis and Perrin1979; Schradin et al Reference Schradin, Lindholm, Johannesen, Schoepf, Yuen, König and Pillay2012) with a diet that comprises seeds, fruits, plant material, and the seasonal consumption of arthropods (Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015; Perrin and Curtis Reference Perrin and Curtis1980). They are habitat generalists that are frequently recorded in low-lying areas (Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015; Schradin and Pillay Reference Schradin and Pillay2005). The Namaqua rock mouse, M. namaquensis, is a geographically widespread nocturnal species which nests in small family groups (Perrin Reference Perrin1981; Perrin and Curtis Reference Perrin and Curtis1980; Skinner and Chimimba Reference Skinner and Chimimba2005). Similar to Rhabdomys, M. namaquensis is omnivorous and feeds on herbage, seeds, and insects (Kerley Reference Kerley1992a), but they have a habitat preference for rocky outcrops (Macfayden Reference Macfayden2014; Skinner and Chimimba Reference Skinner and Chimimba2005). The Karoo bush rat, O. unisulcatus, is a diurnal solitary animal that may also live in small family groups occupying stick nests constructed of plant material (Brown and Willan Reference Brown and Willan1991; Makuya et al Reference Makuya, Pillay and Schradin2024; Vermeulen and Nel Reference Vermeulen and Nel1988) and, in contrast to the previous species, is regarded as a strict herbivore consuming mostly succulents such as Lycium horridum (Brown and Willan Reference Brown and Willan1991; du Plessis et al Reference du Plessis, Kerley and Winter1991). Similar to Rhabdomys, they also prefer low-lying areas where the two taxa frequently co-occur (Brown and Willan Reference Brown and Willan1991; Kerley Reference Kerley1992b; Vermeulen and Nel Reference Vermeulen and Nel1988). Of the three rodent taxa, O. unisulcatus is the largest (adult male average head and body length = 148 mm) compared to Rhabdomys and M. namaquensis (adult male average head and body length = 103 mm and 110 mm, respectively) (Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015; Skinner and Chimimba Reference Skinner and Chimimba2005). It is hypothesised that differences in diet, temporal activity patterns, as well as spatial and habitat niche use would lead to differences in the intra- and intertaxon nematode species compositions.
By studying three geographically sympatric and confamilial rodent host taxa with different life histories and microhabitat preferences, we aim to understand the mechanisms that shape the spatial variation in nematode assemblages in these rodents. It is predicted that hosts with more similar microhabitat-use (Rhabdomys and O. unisulcatus) will harbour more similar nematode communities than those associated with the rocky outcrop species M. namaquensis. Since more frequent contact exists between Rhabdomys and O. unisulcatus, it is predicted that they will have higher within- than between-locality similarity in nematode composition. Moreover, a more diverse range of local habitats occupied by the rodents can lead to nematode communities within taxa across localities being less similar than within localities (Dybing et al Reference Dybing, Fleming and Adams2013; D’Bastiani et al Reference D’Bastiani, Campião, Boeger and Araújo2020; Froeschke et al Reference Froeschke, Harf, Sommer and Matthee2010) since they will have differential contact with intermediate hosts or infective stages of the nematodes (Dybing et al Reference Dybing, Fleming and Adams2013; D’Bastiani et al Reference D’Bastiani, Campião, Boeger and Araújo2020; Froeschke et al Reference Froeschke, Harf, Sommer and Matthee2010). Finally, it is predicted that in the present study dissimilarity in nematode component communities will not necessarily increase with geographic distance (Poulin Reference Poulin2003) since it can be obscured by host factors, such as vagility (Poulin et al Reference Poulin, Krasnov, Mouillot and Thieltges2011), specialised habitat preferences and diets (Leung and Koprivnikar 2019) and the scale at which the study is done (Brouat and Duplantier Reference Brouat and Duplantier2007; Spickett et al Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017b; van der Mescht et al Reference van der Mescht, Warburton, Khokhlova, Stanko, Vinarski, Korallo-Vinarskaya and Krasnov2018).
Materials and methods
Study area, host collection, and parasite recovery
Rodents representing three life histories were trapped at 12 WCP localities during the spring and summer months of 2023 to 2025 (Supplementary Table S1). The sampling design (trap type, number of traps, bait mixture, and trap period per locality) was standardised across localities. Sherman live traps were baited with a mixture of oats and peanut butter, and each locality was trapped once for three to four days. Captured rodents, representing the desired morphospecies, were euthanised with isoflurane (Cape Nature Permit: CNN44-87-27373, Ethical Clearance: ACU-2022-17062) and tongue tissue was collected from each rodent for molecular confirmation where needed (data not shown). Animals were stored at −20 °C until processed in the laboratory.
Nematode removal
Each rodent was dissected to remove the gastrointestinal tract (GIT), which was also separated into the stomach, caecum, small intestine, colon, and rectum, and preserved in 70% ethanol for later processing. The various segments of the GIT (apart from the rectum) were placed in a Petri dish filled with water, cut open, after which the contents were mixed with the water. The water containing the contents from the GIT was systematically examined for the presence of nematodes using a stereomicroscope (Leica MZ75). The wall of each organ was also examined under a stereomicroscope (Leica MZ75) for the additional presence of nematodes. The rectum was first emptied by carefully pressing out the faecal pellets with untethered tissue forceps before the membrane was cut open and similarly microscopically examined for the presence of nematodes. Faecal pellets were disaggregated in water and also examined for the presence of nematodes. The recovered nematodes were initially identified and sorted by morphospecies and placed in 70% ethanol until further processing.
Nematode identification
A subset of each nematode morphospecies was slide-mounted using lactophenol. Microscopic identification was conducted using a Leica DM1000 light microscope fitted with a Zeiss Axiocam ERc 5s camera. Mounted specimens were initially identified to genus level and, where possible, to species level by referring to relevant keys and descriptions of various authors (Anderson et al Reference Anderson, Chabaud and Willmott2009; Bain et al Reference Bain, Mutafchiev, Junker and Schmidt-Rhaesa2014; Durette-Desset Reference Durette-Desset1979; Durette-Desset and Digiani Reference Durette-Desset and Digiani2005; Gibbons and Khalil Reference Gibbons and Khalil1982; Hoberg et al Reference Hoberg, Lloyd and Omar1995; Inglis et al Reference Inglis, Harris and Lewis1990; Junker et al Reference Junker, Lane, McRee, Foggin, van Dyk and Mutafchiev2013; Ortlepp Reference Ortlepp1937; Ortlepp Reference Ortlepp1939; Quentin Reference Quentin1970; Singh Reference Singh1948; Skrjabin et al Reference Skrjabin, Shikhobalova, Lagodovskaya and Skrjabin1960; Smales et al Reference Smales, Harris and Behnke2009). Specimens that were identified to genus level were allocated to morphospecies based on morphological comparisons (visually) and by conducting basic morphometric measurements on a subset of the samples. Final species identifications correspond to Schlemmer et al (Reference Schlemmer, Junker, Matthee and Matthee2026).
Data analyses
All analyses were conducted in the R statistical environment (R Core Team 2015). The mean abundance of each nematode species (total abundance of each nematode species divided by the total number of hosts) and the prevalence (percentage of hosts infected) were calculated following Bush et al (Reference Bush, Lafferty, Lotz and Shostak1997). To prevent sampling bias, only 6 of the 12 localities with more than 14 rodent individuals per taxon and where all 3 rodent taxa were trapped were included in the analyses. Nematode taxon richness was compared across localities for each rodent host species using sample-based rarefaction (species accumulation curves) using the function ‘specaccum’ of the vegan package (Oksanen et al Reference Oksanen, Simpson, Blanchet, Kindt, Legendre, Minchin, O’Hara, Solymos, Stevens, Szoecs and Wagner2015; method = ‘random’, 999 permutations), treating each host individual as a sampling unit and each nematode taxon as a ‘species’ so that richness could be compared fairly despite unequal numbers of hosts examined per locality (see Table 1). To prevent sampling artefacts, rare nematode species which had a total abundance of fewer than five individuals across the sampling range were removed. Thereafter, data entries were log (x +1) transformed and standardised (number of individuals of a given nematode species in each host individual (infracommunity) or mean abundance of a given nematode species at each locality (component community)) within a sample (a host individual, a host taxon, or a locality) by total abundance so that each entry was expressed as relative rather than absolute counts. The similarity in nematode species composition between the three sympatric rodent taxa and among localities within rodent taxa was compared by performing analyses of similarities (ANOSIM) using the vegan package in R (Clarke and Green Reference Clarke and Green1988; Clarke and Warwick Reference Clarke and Warwick2001; Oksanen et al Reference Oksanen, Simpson, Blanchet, Kindt, Legendre, Minchin, O’Hara, Solymos, Stevens, Szoecs and Wagner2015). ANOSIM operates directly on a dissimilarity matrix, so matrices of Bray–Curtis dissimilarities in nematode species composition between samples (host taxon individuals, host taxon populations or pair-localities) were first constructed (Bray and Curtis Reference Bray and Curtis1957). The null hypothesis for ANOSIM is that samples from different groups are, on average, as similar to each other as samples within the same group. The ANOSIM statistic, R, ranges from 0 to 1, with R = 1 indicating dissimilarity among groups, while R ≈ 0 suggests equal similarity within and among groups. R intervals are commonly interpreted in ecology as: <0, no structure; 0–0.1, little to no separation; 0.1–0.25, weak separation; 0.25–0.5, moderate separation; 0.5–0.75, strong separation; and 0.75–1.0, very strong separation. Each ANOSIM was performed based on 999 permutations.
Locality information and rodent sample sizes for Rhabdomys, Micaelamys namaquensis, and Otomys unisulcatus trapped at six localities in the Western Cape Province, South Africa, 2023–2025
Table 1. Long description
The table contains six columns: Sampling date, Locality code, Closest town, Rhabdomys, Micaelamys namaquensis, and Otomys unisulcatus.
* Row 1: 24 to 27 Oct. 2023, M f, Ceres. Rhabdomys: 20 (13 female: 7 male). Micaelamys namaquensis: 23 (10 female: 13 male). Otomys unisulcatus: 20 (17 female: 3 male).
* Row 2: 6 to 10 Nov. 2023, O b, Montagu. Rhabdomys: 21 (13 female: 8 male). Micaelamys namaquensis: 19 (6 female: 13 male). Otomys unisulcatus: 20 (9 female: 11 male).
* Row 3: 10 to 14 Dec. 2023, W b, Matjiesfontein. Rhabdomys: 20 (8 female: 12 male). Micaelamys namaquensis: 21 (10 female: 11 male). Otomys unisulcatus: 21 (15 female: 6 male).
* Row 4: 7 to 11 Oct. 2024, L v, Vanrhynsdorp. Rhabdomys: 22 (11 female: 11 male). Micaelamys namaquensis: 16 (8 female: 8 male). Otomys unisulcatus: 23 (12 female: 11 male).
* Row 5: 18 to 22 Oct. 2024, M w, Oudtshoorn. Rhabdomys: 20 (12 female: 8 male). Micaelamys namaquensis: 14 (9 female: 5 male). Otomys unisulcatus: 14 (10 female: 4 male).
* Row 6: 27 to 29 Jan. 2025, T w, Touwsrivier. Rhabdomys: 20 (10 female: 10 male). Micaelamys namaquensis: 20 (13 female: 7 male). Otomys unisulcatus: 20 (13 female: 7 male).
* Total: Rhabdomys: 123 (67 female: 56 male). Micaelamys namaquensis: 113 (56 female: 57 male). Otomys unisulcatus: 118 (76 female: 42 male).
Nematode species that primarily accounted for differences between groups were identified. The groups included: nematode infracommunities of the three sympatric host taxa, nematode component communities of the host populations at a given locality, and infracommunities for each separate host taxon between locality pairs. This was carried out by decomposition of the Bray-Curtis similarity into the contribution of each species through the routine Similarity Percentages (SIMPER) analysis in R using the ‘simper’ function in the vegan package.
The influence of geographic distance on similarities in nematode component communities within each host taxon was evaluated using multiple regression on distance matrices (MRM). MRM models the relationship between a multivariate response distance matrix, in this case Bray–Curtis dissimilarity matrix of component communities, and an explanatory matrix of pairwise geographic distances. Euclidean distances between localities were calculated with the stats package implemented in R. Model significance and regression coefficients were assessed through permutations of the response matrix while keeping explanatory matrices fixed. Rows and corresponding columns of the response matrices were permuted together, and the model’s coefficient of determination, along with the regression coefficients, were calculated to generate a null distribution (Legendre and Legendre Reference Legendre and Legendre1998; Lichstein Reference Lichstein2007). All probabilities were derived from 10,000 permutations, and the analyses were conducted with the ecodist package in R (Goslee and Urban Reference Goslee and Urban2007).
Results
A total of 576 rodent individuals representing 4 species were captured at 12 localities as part of a larger study (see Schlemmer et al Reference Schlemmer, Junker, Matthee and Matthee2026 and Supplementary Table S1). For the 6 localities included herein, a total of 9299 nematode individuals were recorded from 354 rodents (Rhabdomys n = 123, M. namaquensis n = 113 and O. unisulcatus n = 118) (Table 1) (see Supplementary Table S2 for nematode diversity across 12 localities). Rarefaction curves reached or closely approached an asymptote for all six locality–host species combinations within the range of individuals sampled, indicating that sampling effort was generally sufficient to capture the majority of nematode taxon richness present (Supplementary Figure S1). Within the 6 analysed localities, 11 nematode species were recovered across all host taxa (Table 2). Syphacia was the only genus representing two morphologically and molecularly dissimilar undescribed species. Syphacia sp. 1 was confined to Rhabdomys only, while Syphacia sp. 2 was restricted to M. namaquensis (see Schlemmer et al Reference Schlemmer, Junker, Matthee and Matthee2026). Hereafter, only data from the six localities included in the current study are reported. The most prevalent nematode species was Nematodirus sp. (23.16%), while the most abundant nematode species was Syphacia sp. 1 (16.97 ± 5.72) (Table 2). Neoheligmonella capensis was the only nematode that was shared between the three host taxa (Table 3). Rhabdomys harboured seven species, of which five represented monoxenous species, and the remaining two represented heteroxenous species (Table 3). Syphacia sp. 1 was the most abundant (48.85 ± 16.13) and prevalent (31.71%) in Rhabdomys and was recovered from five out of the six trapping localities (Tables 2 and 3). Micaelamys namaquensis harboured six species, of which three were monoxenous and the other three heteroxenous species (Table 3). Syphacia sp. 2 was the most abundant (8.30 ± 3.05) and prevalent (43.36%) species associated with M. namaquensis and occurred at all six localities (Tables 2 and 3). Lastly, O. unisulcatus harboured five species, of which only one was a heteroxenous species (Table 3). Nematodirus sp. was the most abundant (10.53 ± 2.72) and prevalent (38.98%) species in O. unisulcatus (Table 3).
Nematode species (n = 11) recovered from Rhabdomys, Micaelamys namaquensis, and Otomys unisulcatus trapped at six localities in the Western Cape Province, South Africa, 2023–2025. Data presented include prevalence (%P), mean abundance ± standard error (MA ± SE), and number of localities where presence was recorded (LOC)
Table 2. Long description
The table consists of four columns: Nematode species, percent P (prevalence), M A plus or minus S E (mean abundance plus or minus standard error), and L O C (number of localities).
Monoxenous species:
* Nematodirus species: 23.16 percent P, 4.18 plus or minus 0.95 M A, 5 L O C.
* Syphacia species 2: 13.84 percent P, 2.65 plus or minus 0.99 M A, 6 L O C.
* Neoheligmonella capensis: 12.43 percent P, 1.45 plus or minus 0.46 M A, 6 L O C.
* Syphacia species 1: 11.02 percent P, 16.97 plus or minus 5.72 M A, 5 L O C.
* Aspiculuris shikoloueta: 6.50 percent P, 0.73 plus or minus 0.26 M A, 5 L O C.
* Heligmonina spira: 0.28 percent P, 0.01 plus or minus 0.01 M A, 1 L O C.
* Trichuris species: 0.28 percent P, 0.01 plus or minus 0.01 M A, 1 L O C.
* Strongyloides species: 0.28 percent P, 0.003 plus or minus 0.003 M A, 1 L O C.
Heteroxenous species:
* Protospirura muricola: 2.26 percent P, 0.16 plus or minus 0.09 M A, 3 L O C.
* Abbreviata species: 2.26 percent P, 0.08 plus or minus 0.03 M A, 3 L O C.
* Mastophorus muris: 0.85 percent P, 0.03 plus or minus 0.02 M A, 2 L O C.
Nematode species recovered from respective host taxa, used in current analysis, in the Western Cape Province, South Africa, 2023–2025. Data presented include mean abundance ± standard error and prevalence (%) in brackets
Table 3. Long description
The table is organized into four columns: Nematode species, Rhabdomys (n = 123), Micaelamys namaquensis (n = 113), and Otomys unisulcatus (n = 118). Data is presented as mean abundance plus or minus standard error, with prevalence percentage in parentheses.
Monoxenous species:
* Aspiculuris shikoloueta: Rhabdomys 1.70 plus or minus 0.72 (12.20 percent); M. namaquensis 0; O. unisulcatus 0.41 plus or minus 0.20 (6.78 percent).
* Heligmonina spira: Rhabdomys 0.04 plus or minus 0.04 (0.81 percent); others 0.
* Nematodirus sp.: Rhabdomys 1.93 plus or minus 0.46 (29.27 percent); M. namaquensis 0; O. unisulcatus 10.53 plus or minus 2.72 (38.98 percent).
* Neoheligmonella capensis: Rhabdomys 2.15 plus or minus 0.92 (21.14 percent); M. namaquensis 0.14 plus or minus 0.07 (4.42 percent); O. unisulcatus 1.97 plus or minus 0.97 (11.02 percent).
* Strongyloides sp.: M. namaquensis 0.01 plus or minus 0.01 (0.88 percent); others 0.
* Syphacia sp. 1: Rhabdomys 48.85 plus or minus 16.13 (31.71 percent); others 0.
* Syphacia sp. 2: M. namaquensis 8.30 plus or minus 3.05 (43.36 percent); others 0.
* Trichuris sp.: O. unisulcatus 0.02 plus or minus 0.02 (0.85 percent); others 0.
Heteroxenous species:
* Abbreviata sp.: Rhabdomys 0.05 plus or minus 0.05 (0.81 percent); M. namaquensis 0.19 plus or minus 0.09 (6.19 percent); O. unisulcatus 0.
* Mastophorus muris: Rhabdomys 0; M. namaquensis 0.02 plus or minus 0.01 (1.77 percent); O. unisulcatus 0.06 plus or minus 0.06 (0.85 percent).
* Protospirura muricola: Rhabdomys 0.07 plus or minus 0.04 (2.44 percent); M. namaquensis 0.43 plus or minus 0.28 (4.42 percent); O. unisulcatus 0.
Analysis of similarity of infracommunities between host taxa
The ANOSIM results depicting the species composition of nematode infracommunities indicated a weak but significant overlap between Rhabdomys and O. unisulcatus (R = 0.18, P < 0.001). There was more overlap when comparing nematode species composition of infracommunities of Rhabdomys with O. unisulcatus than within Rhabdomys from different localities, whereas overlap was similar in O. unisulcatus (see below). SIMPER showed that most of the dissimilarity between Rhabdomys and O. unisulcatus was driven by 4 nematode species (Nematodirus sp., Syphacia sp. 1, Neoheligmonella capensis, and Aspiculuris shikoloueta), which together contributed 73.77% to dissimilarity (Table 4).
Nematode species that contributed most to the dissimilarity of nematode infracommunities between host taxa (MA = mean abundance after log (x + 1) transformation)
Table 4. Long description
The table is divided into three comparative sections based on host taxa.
Section 1: Comparison between Rhabdomys and O. unisulcatus.
- Nematodirus species: M A in Rhabdomys 25.16, M A in O. unisulcatus 74.48, Contribution 30.29 percent.
- Syphacia species 1: M A in Rhabdomys 34.84, M A in O. unisulcatus 0, Contribution 17.42 percent.
- Neoheligmonella capensis: M A in Rhabdomys 18.41, M A in O. unisulcatus 17.85, Contribution 14.54 percent.
- Aspiculuris shikoloueta: M A in Rhabdomys 18.57, M A in O. unisulcatus 7.18, Contribution 11.52 percent.
Section 2: Comparison between Rhabdomys and M. namaquensis.
- Syphacia species 2: M A in Rhabdomys 0, M A in M. namaquensis 79.68, Contribution 39.8 percent.
- Syphacia species 1: M A in Rhabdomys 34.84, M A in M. namaquensis 0, Contribution 17.4 percent.
- Nematodirus species: M A in Rhabdomys 25.16, M A in M. namaquensis 0, Contribution 12.6 percent.
- Neoheligmonella capensis: M A in Rhabdomys 18.41, M A in M. namaquensis 2.59, Contribution 9.8 percent.
- Aspiculuris shikoloueta: M A in Rhabdomys 18.57, M A in M. namaquensis 0, Contribution 9.3 percent.
Section 3: Comparison between O. unisulcatus and M. namaquensis.
- Syphacia species 2: M A in O. unisulcatus 0, M A in M. namaquensis 79.68, Contribution 39.84 percent.
- Nematodirus species: M A in O. unisulcatus 74.48, M A in M. namaquensis 0, Contribution 37.24 percent.
The ANOSIM analyses between the nematode infracommunities of Rhabdomys and M. namaquensis demonstrated a strong and significant separation (R = 0.56, P < 0.001). There was thus similar overlap in nematode species composition of infracommunities for Rhabdomys when comparing between host taxa (Rhabdomys vs M. namaquensis) and within host taxa from different localities (see below). In contrast, there was less overlap when comparing the composition of infracommunities of M. namaquensis with Rhabdomys than within M. namaquensis across localities (see below). SIMPER showed that most of the dissimilarity between Rhabdomys and M. namaquensis was driven by five nematode species (Syphacia sp. 1, Syphacia sp. 2, Nematodirus sp., N. capensis, and A. shikoloueta) that jointly contributed 88.90% to the dissimilarity (Table 4).
Lastly, the ANOSIM between O. unisulcatus and M. namaquensis indicated a very strong and significant separation (R = 0.75, P < 0.001) between rather than within their nematode infracommunities. Therefore, there was less overlap when comparing infracommunity species composition between these host taxa than within each host taxon across localities for O. unisulcatus and M. namaquensis, respectively (see below). Furthermore, SIMPER showed that most of the dissimilarity between O. unisulcatus and M. namaquensis was driven by 2 nematode species (Syphacia sp. 2 and Nematodirus sp.) that jointly contributed 77.08% to their dissimilarity (Table 4).
Analysis of similarity of component communities between host taxa
The ANOSIM of species composition of nematode component communities indicated a weak but significant separation between Rhabdomys and O. unisulcatus, which was more than within hosts (R = 0.23, P < 0.05). SIMPER analysis between the component communities of these rodents showed that most of the dissimilarity was driven by 4 nematode species (Nematodirus sp., Syphacia sp. 1, N. capensis, and A. shikoloueta) that jointly contributed 67.08% to dissimilarity (Table 5). A very strong and significant separation was observed between the component communities of Rhabdomys and M. namaquensis, which was more than within hosts (R = 0.83, P < 0.01), and this was mainly attributed to 2 nematode species (Syphacia sp. 1 and Syphacia sp. 2) with a joint contribution of 59.76% (Table 5). Lastly, the results from the ANOSIM of the nematode component communities between O. unisulcatus and M. namaquensis indicated a very strong and significant separation, which was more than within hosts (R = 0.96, P < 0.001). The SIMPER analysis showed that 3 nematode species (Syphacia sp. 2, Nematodirus sp., and N. capensis) contributed 73.77% to the dissimilarity between the component communities of O. unisulcatus and M. namaquensis (Table 5).
Nematode species that contributed most to the dissimilarity of nematode component communities between host taxa (MA = mean abundance after log (x + 1) transformation)
Table 5. Long description
The table is divided into three comparative sections based on host taxa.
Section 1: Comparison between Rhabdomys and O dot unisulcatus.
* Nematodirus species: M A in Rhabdomys is 18.77, M A in O dot unisulcatus is 53.08, contributing 20.51 percent to dissimilarity.
* Syphacia species 1: M A in Rhabdomys is 34.49, M A in O dot unisulcatus is 0, contributing 17.25 percent to dissimilarity.
* Neoheligmonella capensis: M A in Rhabdomys is 20.54, M A in O dot unisulcatus is 26.41, contributing 14.77 percent to dissimilarity.
* Aspiculuris shikoloueta: M A in Rhabdomys is 18.26, M A in O dot unisulcatus is 19.24, contributing 14.55 percent to dissimilarity.
Section 2: Comparison between Rhabdomys and M dot namaquensis.
* Syphacia species 2: M A in Rhabdomys is 0, M A in M dot namaquensis is 85.01, contributing 42.51 percent to dissimilarity.
* Syphacia species 1: M A in Rhabdomys is 34.49, M A in M dot namaquensis is 0, contributing 17.25 percent to dissimilarity.
Section 3: Comparison between O dot unisulcatus and M dot namaquensis.
* Syphacia species 2: M A in O dot unisulcatus is 0, M A in M dot namaquensis is 85.01, contributing 42.51 percent to dissimilarity.
* Nematodirus species: M A in O dot unisulcatus is 53.08, M A in M dot namaquensis is 0, contributing 26.54 percent to dissimilarity.
* Neoheligmonella capensis: M A in O dot unisulcatus is 26.41, M A in M dot namaquensis is 2.04, contributing 13.21 percent to dissimilarity.
Analysis of similarity of infracommunities within host taxa
The ANOSIM analysis of nematode species composition associated with Rhabdomys was more dissimilar among individuals from different localities than among individuals from the same locality (R = 0.58, P < 0.001). Infracommunities were therefore most dissimilar within Rhabdomys when compared to other host species (see above). The SIMPER analysis showed that 4 nematode species (A. shikoloueta, Syphacia sp. 1, N. capensis, and Nematodirus sp.) jointly contributed a mean cumulative contribution of 85.27% to the dissimilarity (Table 6). Micaelamys namaquensis individuals revealed little to no and non-significant separation among individuals from different localities compared with those from the same locality (R = 0.04, P > 0.05). In contrast, O. unisulcatus individuals exhibited weak but significant separation among individuals from different localities compared with those from the same locality (R = 0.15, P < 0.01). Infracommunities were therefore more similar within O. unisulcatus individuals than in comparison with other host taxa. The dissimilarity was caused mainly by the contribution of 3 nematode species (Nematodirus sp., N. capensis, and A. shikoloueta), which contributed 81.20% to dissimilarity (Table 5).
Nematode species that contributed most (maximum contribution > 20%) to the pairwise between-locality dissimilarity in species composition of infracommunities within host taxa (N = number of locality pairs where species contributed to dissimilarity)
Table 6. Long description
The table consists of four columns: Nematode species, Mean contribution (%), Maximum contribution (%), and N (number of locality pairs). It is divided into two main sections based on host taxa.
Section 1: Rhabdomys
* Aspiculuris shikoloueta: Mean contribution 29.63, Maximum 50, N 9.
* Syphacia sp. 1: Mean contribution 22.63, Maximum 35.83, N 15.
* Neoheligmonella capensis: Mean contribution 15.97, Maximum 31.82, N 15.
* Nematodirus sp.: Mean contribution 17.04, Maximum 34.84, N 15.
Section 2: O. unisulcatus
* Nematodirus sp.: Mean contribution 40.26, Maximum 50, N 15.
* Neoheligmonella capensis: Mean contribution 19.73, Maximum 25, N 12.
* Aspiculuris shikoloueta: Mean contribution 21.21, Maximum 25, N 15.
MRM analysis (distance decay) indicated that dissimilarity in nematode species composition among localities was not significantly correlated with geographic distance between localities across all 3 host taxa (r2 = 0.13, F = 2.44, coefficient = 0.04 for Rhabdomys; r2 = 0.09, F = 0.15, coefficient = 0.01 for M. namaquensis; r2 = 0.25, F = 1.54, coefficient = 0.06 for O. unisulcatus; P > 0.05 for each host taxon, respectively) (Supplementary Figure S2). In other words, localities that were close or far from one another have similar nematode assemblages.
Discussion
Statistical analyses of the nematode assemblages sampled across multiple sympatric rodent species from 6 different localities suggest that dissimilarity in nematode species composition between host taxa is primarily influenced by differences in microhabitat use mediated through life-history traits of host taxa. However, part of the between-host separation is driven by host-restricted nematode species (i.e. Syphacia spp.). In contrast, environmental influences appear to act indirectly through shaping host behaviour and habitat use rather than directly affecting nematode community composition.
Host effects on nematode community composition
Rodent taxa (O. unisulcatus and Rhabdomys) that share microhabitats and are sometimes sampled from the same nesting site share more similar nematode communities compared to the rodent taxon that occurs at the same locality but uses different microhabitats (M. namaquensis). This pattern was repeatedly observed at all 6 localities. Rhabdomys are considered an opportunistic species with a wide habitat tolerance, but they are most frequently encountered in low-lying areas (du Toit et al Reference du Toit, Pillay, Ganem, Relton, Child, Roxburgh, Do Linh San, Raimondo and Davies-Mostert2016; Froeschke and Matthee Reference Froeschke and Matthee2014; Ganem et al Reference Ganem, Dufour, Avenant, Caminade, Eiseb, Tougard and Pillay2020). This translates to more frequent interaction and resource sharing with O. unisulcatus individuals that prefer patches with shrubby vegetation in low-lying areas (Brown and Willan Reference Brown and Willan1991; Macfayden Reference Macfayden2014). Further, a behavioural study done in the Succulent Karoo found that O. unisulcatus and R. pumilio individuals share runways and compete for nesting sites, which further supports microhabitat sharing between these taxa (Schradin Reference Schradin2005). Additional support for the importance of microhabitat sharing is evident in the fact that nematode communities in the rock-dwelling M. namaquensis were the most distinct from the aforementioned taxa, despite the fact that both M. namaquensis and the 2 Rhabdomys species are omnivores (Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015). Since nematode species differ in habitat preference and climatic tolerance levels, which will influence their distribution within the landscape (Kafle et al Reference Kafle, Peacock, Grond, Orsel and Kutz2018; Spickett et al Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017a), it is argued that microhabitat sharing by the host species exposes them to the same set of nematode species and can thus also facilitate nematode exchange (Obanda et al Reference Obanda, Maingi, Muchemi, Ng’ang’a, Angelone and Archie2019; Titcomb et al Reference Titcomb, Pansu, Hutchinson, Tombak, Hansen, Baker, Kartzinel, Young and Pringle2022).
Nematode species that contributed most to the dissimilarity in community composition among the 3 genera were A. shikoloueta, N. capensis, Nematodirus sp., Syphacia sp. 1, and Syphacia sp. 2. The differences in the values of the SIMPER analyses were attributed to differences in the abundances of the nematode species or the absence of some nematode species in the species composition associated with the respective host taxa or localities. Some species were shared between host taxa, e.g. A. shikoloueta was found in Rhabdomys and O. unisulcatus (also see Spickett et al Reference Spickett, Junker, Froeschke, Haukisalmi and Matthee2020), but absent in M. namaquensis. In contrast, species within the genus Syphacia have been recorded as host-specific (Behnke et al Reference Behnke, Stewart, Smales, Cooper, Lowe, Kinsella, Bajer, Dwużnik-Szarek, Herman, Fenn, Catalano, Diagne and Webster2022). Members of the latter genus have a more restricted mode of transmission that includes retro-infection and the presence of eggs on the fur of the host; therefore, grooming serves as another transmission method for Syphacia spp. (Carlberg and Lang Reference Carlberg and Lang2004; Lee Reference Lee and Lee2002; Taffs Reference Taffs1976). The host specificity of Syphacia could explain the high contribution of Syphacia spp. to dissimilarity across all comparisons between host taxa. The host specificity of Syphacia spp. may itself have been shaped by long-term microhabitat-based host partitioning, so that host-specific parasites and host-generalist parasites reflect different temporal scales of the same ecological process. Since Syphacia sp. 1 and Syphacia sp. 2 are strictly host-restricted and therefore cannot be exchanged between host taxa regardless of microhabitat overlap, their disproportionate contribution to between-host dissimilarity represents a coevolutionary rather than an ecological signal.
Both N. capensis and Nematodirus sp. infect their host through the ingestion of their free-living larvae (Audebert et al Reference Audebert, Cassone, Kerboeuf and Durette-Desset2004; N’Zobadila and Durette-Desset Reference N’Zobadila and Durette-Desset1992). In the present study, vegetation was more often denser in the low-lying areas compared to the rocky outcrops (personal observation). Denser vegetation reduces soil temperature and could protect a water film on the soil surface from evaporation (Brouat et al Reference Brouat, Kane, Diouf, Bâ, Sall-Dramé and Duplantier2007; Howe and Smith Reference Howe, Smith, Gentry, Fuhrmann and Zuberer2021; Lee Reference Lee and Lee2002). This water film can result in higher mobility of free-living larvae of Nematodirus sp. and N. capensis and, coupled with the lower soil temperature, reduce desiccation risks (Anderson Reference Anderson2000; Lee Reference Lee and Lee2002). Given that Rhabdomys and O. unisulcatus occupy the low-lying areas, this may explain the higher recordings of Nematodirus sp. and N. capensis associated with them. In support of the latter, the abundance of N. capensis was the lowest in M. namaquensis, and Nematodirus sp. was completely absent from M. namaquensis. In addition, an unknown Syphacia species, as well as N. capensis, were part of the nematode species that contributed most to the dissimilarity of the communities within Rhabdomys individuals across localities in the current study and that of Spickett et al (Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017b).
Environmental effects on nematode community structure
Nematode infracommunities of the 3 murid rodent taxa showed variable differences among localities, with moderate dissimilarity in species communities observed for Rhabdomys species, while low to no dissimilarity was observed for O. unisulcatus and M. namaquensis. In the absence of measured environmental data, we speculate that the generalist nature of Rhabdomys (including their high adaptability to anthropogenic habitats) would assist in explaining the highest dissimilarity within the respective taxa across localities (Froeschke et al Reference Froeschke, Harf, Sommer and Matthee2010). Specifically, Rhabdomys inhabit a higher diversity of microhabitats, which could be complementary to the survival thresholds of the infective stages of more nematode species. Furthermore, upon sampling the rodents, it was noted that at certain localities, Rhabdomys was trapped in habitats with no disturbance, while at other localities, they were trapped in croplands or near/inside buildings. This could have led to differential local influences on the nematode composition associated with Rhabdomys among localities. Our finding corroborates the results of Froeschke and Matthee (Reference Froeschke and Matthee2014), who documented that nematode assemblages associated with R. pumilio are influenced by the disturbance of their habitats.
In contrast to the pattern observed for Rhabdomys, O. unisulcatus and M. namaquensis are habitat specialists mainly confined to undisturbed areas (Monadjem et al Reference Monadjem, Taylor, Denys and Cotterill2015; Skinner and Chimimba Reference Skinner and Chimimba2005). As mentioned above, O. unisulcatus prefers areas with dense natural vegetation cover and M. namaquensis prefers rocky outcrops. The restriction of these 2 genera to undisturbed and, compared with Rhabdomys, more uniform microhabitats could explain why they do not show strong dissimilarities across the 6 localities. The latter supports the notion that the habitat of the host is a strong determinant of nematode communities associated with rodents (Froeschke and Matthee Reference Froeschke and Matthee2014; Palmeirim et al Reference Palmeirim, Bordes, Chaisiri, Siribat, Ribas and Morand2014).
Habitat preference was previously suggested for nematode species and particularly for Aspiculuris, which shows a preference for more xeric habitats (Spickett et al Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017a, Reference Spickett, Junker, Froeschke, Haukisalmi and Matthee2020). Our study supports this suggestion in that A. shikoloueta was most abundant in the more xeric Succulent Karoo (data provided in Schlemmer et al Reference Schlemmer, Junker, Matthee and Matthee2026). In this specific instance, the higher prevalence in xeric conditions might be due to an increase in coprophagy by the host to combat resource-poor environments and conserve water (Spickett et al Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017a). The increase in coprophagy could therefore have led to an increase in the ingestion of these nematode eggs. An increase in coprophagy at xeric localities could have further contributed to the dissimilarity among infracommunities of Rhabdomys across localities.
The effect of geographic distance on community composition of nematodes
In the present study, no support was found for distance decay of similarity across host taxa (Poulin Reference Poulin2003; Poulin et al Reference Poulin, Krasnov, Mouillot and Thieltges2011). The absence of a significant correlation between community similarity and the geographic distance among these localities indicates that geographic distance was not a significant predictor of dissimilarity in this dataset. While this may suggest a role for host life-history differences, additional factors such as the limited spatial extent of sampling, environmental similarity across sites, or limited statistical resolution could also explain this pattern. The study by Spickett et al (Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017b) noted that the heterogeneity of climatic variables across localities could influence the outcome of the MRM analysis. Further, Spickett et al (Reference Spickett, Junker, Krasnov, Haukisalmi and Matthee2017b) conducted their study over a greater geographic area that included multiple and more dissimilar biomes in South Africa. The smaller geographic scale and more similar biomes included in the present study may have resulted in more homogeneous environmental conditions across localities, which may have led to fewer differences in nematode component communities (Palmeirim et al Reference Palmeirim, Bordes, Chaisiri, Siribat, Ribas and Morand2014).
In conclusion, our study highlights the fact that microhabitat sharing is an important determinant of similarity in nematode species composition found in sympatric host taxa in the same family. Environmental influences appear to be manifested primarily through host life-history characteristics rather than acting directly on infracommunity composition. The absence of distance decay in similarity is possibly confounded by the small geographic scale employed herein. Future work should include direct environmental covariates, such as precipitation and Normalised Difference Vegetation Index (NDVI), to formally partition host and environmental effects, expand sampling extent to larger geographic scales with more heterogeneous biomes, and investigate whether host-specific nematodes show patterns of codivergence with their hosts. Although efforts were made to ensure that each genus-level taxon represented a single morphospecies, it should be noted that genus-level identifications may encompass multiple distinct species, potentially masking differences in parasite assemblages. Therefore, future studies should aim to obtain species-level identifications to minimize potential biases.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0022149X26102077.
Acknowledgements
Private landowners and property managers are thanked for allowing us to trap on their properties. We thank Dr Kerstin Junker for her assistance in identifying the nematodes. We are grateful to various field assistants: Aidan Fitzgerald, Alyssa Little, Inge Raubenheimer, Jessica Kipling, Lisa Gouws, and Lola Singo.
Financial support
The research was funded by the National Research Foundation (NRF) in South Africa (Grant reference numbers: GUN 129276 and FBIS22100761503). Any opinion, finding, conclusion, or recommendation expressed in this material is that of the authors and the NRF does not accept any liability in this regard.
Competing interests
None.
Ethical standard
The study was approved by the Animal Care and Use Ethics Committee of Stellenbosch University (reference numbers: ACU-2022-17062, ACU-2024-30885). A Section 20 permit was obtained from the Department of Agriculture, Land Reform and Rural Development (no. 12/11/1/7/5(b)(6091 MG)), as well as collection permits from CapeNature (the Western Cape provincial nature conservation authority) CN44-87-27373 and CN44-87-33283).
Open access
