- Top of page
As in many parts of the world, spiders in South Africa are a widespread, highly diverse arthropod group (Foord et al., 2011; Dippenaar-Schoeman et al., 2012). Baseline studies, focused mostly on the north-eastern provinces, have produced much information on spider diversity and distribution in natural and agricultural systems (Foord et al., 2011; Dippenaar-Schoeman et al., 2013). Yet, surveys in the Western Cape Province, which comprises most of the Cape Floristic Region (CFR), a global biodiversity hotspot (Myers et al., 2000), have been limited. Recent surveys in the CFR suggest that the region's spider diversity may be comparable to that of the most species-rich biomes (Haddad & Dippenaar-Schoeman, 2009) and the expansion of survey efforts in the Western Cape has been recommended (Haddad & Dippenaar-Schoeman, 2009; Foord et al., 2011).
The CFR is an important agricultural area. A third of the CFR has been transformed by agriculture (Rouget et al., 2003), with vineyards being one of the dominant crops, comprising 95% of South Africa's vineyards. Future transformation of natural habitat for vineyard cultivation and its associated impacts on biodiversity is a key threat to CFR ecosystems (Fairbanks et al., 2004). Fortunately, conservation initiatives and production schemes that promote sustainable production and protection of natural habitat on CFR wine farms (CAPE, 2011; Conservation at Work, 2013; IPW, 2013; WWF, 2013) provide opportunities for biodiversity conservation in the vineyard landscape. There is considerable scope for research that will inform these practices to help make the landscape more habitable for native species.
Spiders often dominate predator assemblages in South African agroecosystems (Dippenaar-Schoeman et al., 2013), including vineyards, where, in terms of abundance, they can comprise up to 64% of the ground-living predator assemblage (Gaigher, 2008; Gaigher & Samways, 2010). Because they are such a consistent component of the natural enemy assemblage, there is increasing interest in the biological control potential of spiders in South African crops (Dippenaar-Schoeman et al., 2013). Evidence from other crops suggests that they can reduce densities of pests such as citrus psylla (Van den Berg et al., 1992), cotton bollworm, and red spider mite (Dippenaar-Schoeman et al., 1999). They may be important as part of the natural enemy complex within sustainable production systems. Investigation into their ecology in vineyards is, therefore, interesting both from a conservation and agricultural point of view.
We focus here on the diversity and distribution of spiders in the CFR vineyard landscape. We aim to contribute to the known spider distribution data in the Western Cape and to examine the effects of local land-use on spider diversity, focusing specifically on the effect of farming intensity and the contribution of natural habitat fragments. Based on these results, their potential for pest suppression in CFR vineyards is discussed.
- Top of page
A total of 1586 spider individuals were collected representing 45 species from 16 families (Table 3). Three families contributed 80% of the total spider abundance, i.e. Gnaphosidae (39.5%), Lycosidae (23.5%), and Amaurobiidae (16.3%). Family diversity differed between the three habitat types, with natural sites having a higher family diversity than both vineyard types (Fig. 1). Natural habitat supported 16 families, and the overall assemblage was dominated by Gnaphosidae (33.2%), Amaurobiidae (22.4%), Lycosidae (12.7%), and Linyphiidae (10.8%). Nine families were recorded in organic vineyards and Gnaphosidae (44.3%) and Lycosidae (31.5%) dominated the assemblage. Conventional vineyards supported 11 families and were also dominated by Gnaphosidae (41.9%) and Lycosidae (27.4%) (Fig. 1). Families that are common and common-to-rare in South Africa were well-supported by all habitats, but some rare families such as Ammoxenidae and Palpimanidae were absent from vineyards (Table 3).
Table 3. Total abundance of spider species recorded pooled for the three habitat types
|Family|| ||Genus and species||Natural habitat||Organic vineyards||Conventional vineyards|
|Amaurobiidae||(R, WE)||Chresiona sp.||131||65||63|
|Ammoxenidae||(R, WA)||Ammoxenus sp.||33||0||0|
|Clubionidae*||(C, WA)||Clubiona sp.||1||1||0|
|Genus & species undet.||2||0||0|
| Clubiona umbilensis ||0||1||0|
|Cyrtaucheniidae||(R, WA)||Ancylotrypa sp. 1||0||0||3|
|Ancylotrypa sp. 2||1||0||0|
|Gnaphosidae||(C, WA)||Camillina sp.||8||9||4|
|Genus & species undet.||13||18||11|
|Zelotes sp. 3||6||0||2|
|Genus & species undet.||2||15||0|
| Trachyzelotes jaxartensis ||89||79||51|
| Drassodes solitarius ||51||93||89|
| Zelotes fuligineus ||9||8||4|
| Setaphis subtilis ||4||0||3|
| Zelotes oneili ||4||26||12|
| Pterotrichia varia ||7||2||3|
| Upognampa parvipalpa ||1||4||0|
|Idiopidae||(CR, WA)||Ctenolophus sp.||1||0||0|
|Linyphiidae||(C, WE)||Genus & species undet.||18||23||20|
|Genus & species undet.||21||5||1|
|Genus & species undet.||23||1||2|
|Genus & species undet.||1||2||2|
|Liocranidae||(R, WA)||Rhaeboctesis sp.||1||0||1|
|Lycosidae*||(C, WA)||Geolycosa sp.||11||5||5|
| Pardosa crassipalpis ||2||66||7|
|Pardosa sp. 2||29||49||64|
|Palpimanidae||(R, WA)|| Diaphorocellus biplagiatus ||29||0||0|
|Philodromidae*||(CR, WA)||Thanatus sp.||1||4||2|
|Pisauridae||(CR, WE)||Rothus sp.||1||0||0|
|Salticidae*||(C, WA)||Pellenes sp. 2||10||14||22|
|Pellenes sp. 3||8||2||0|
|Pellenes sp. 1||1||0||1|
|Scytodidae||(C, WA)|| Scytodes testudo ||3||2||2|
|Theridiidae**||(C, WE)|| Steatoda capensis ||1||0||1|
|Steatoda sp. 2||2||0||0|
|Thomisidae*||(C, WA)|| Xysticus subjugalis ||8||2||1|
|Total abundance per habitat type||585||574||427|
Figure 1. Rank abundance plot of spider families in (a) natural habitat, (b) organic vineyards, and (c) conventional vineyards, pooled for the three localities.
Download figure to PowerPoint
Sites differed in spider abundance (χ2 = 228.72, P < 0.0001; Fig. 2a). In Localities 1 and 3, natural habitat had a significantly higher abundance than organic (ZLocality1 = 4.20, P < 0.0001; ZLocality3 = 2.82, P < 0.005) and conventional vineyards (ZLocality1 = 3.86, P < 0.0001; ZLocality3 = 7.64, P < 0.0001). In Locality 2, the organic vineyard had the highest abundance, which was significantly higher than the natural habitat (ZLocality2 = 6.91, P < 0.0001), and conventional vineyard (ZLocality2 = 3.85, P < 0.0001).
Figure 2. Mean spider (a) abundance and (b) species richness per habitat type within the three localities (±SE). Means with letters in common are not significantly different at P < 0.05.
Download figure to PowerPoint
Sites also differed in spider species richness (χ2 = 46.84, P < 0.0001; Fig. 2b). Natural habitat had the highest species richness in two of the localities, being significantly higher than the organic (ZLocality1 = 2.18, P < 0.05) and conventional vineyards (ZLocality1 = 3.34, P < 0.001) in Locality 1, and higher than the conventional vineyard in Locality 3 (ZLocality1 = 3.38, P < 0.001). Organic vineyards had higher species richness than conventional vineyards in all localities, but these differences were non-significant.
In terms of overall assemblage structure, there was considerable overlap between sites (RANOSIM = 0.44, P < 0.001). Nevertheless, nMDS revealed that natural habitat assemblages were still distinct from all vineyard assemblages (Fig. 3). Vineyards grouped according to locality rather than management class (Fig. 3). A total of 22 species were shared between all habitats, whereas natural sites had nine unique species, and organic and conventional vineyards each had one unique species.
Figure 3. nMDS ordination of Bray–Curtis similarities from square-root transformed species abundances per site, showing the grouping of sites according to habitat type (natural habitat and vineyards) and locality.
Download figure to PowerPoint
Key discriminating species between assemblages of different habitat types were abundant species that varied in their relative abundances between habitats (Table 4). Chresiona sp. and Trachyzelotes jaxartensis contributed to differences between natural habitats and vineyards, being most abundant in natural habitats, and also more abundant in organic than conventional vineyards. Hogna sp. and Pardosa sp., both Lycosidae, a frequently encountered family in agro-ecosystems (Dippenaar-Schoeman & Jocqué, 1997), as well as Drassodes solitaries, distinguished vineyards from natural sites and were more abundant in organic than conventional vineyards. An unidentified Linyphiid (sp. 1) differed between vineyard types, being more abundant in organic vineyards (Table 4).
Table 4. Results from SIMPER analyses showing mean relative abundances of key discriminating species and their contributions to dissimilarities between natural habitats, organic vineyards, and integrated vineyards
|Average dissimilarity: 65.14%||Mean abundance||Dis/SD||% Con||Cum % Dis|
|Natural habitat||Organic vineyards|
| Trachyzelotes jaxartensis ||2.97||2.63||1.21||6.64||22.99|
| Drassodes solitarius ||1.70||3.10||1.11||7.31||30.30|
|Pardosa sp. 2||0.97||1.63||1.09||6.01||36.31|
|Average dissimilarity: 68.72%||Natural habitat||Conventional vineyards|| || || |
| Trachyzelotes jaxartensis ||2.97||1.70||1.09||7.37||20.13|
| Drassodes solitarius ||1.70||2.97||1.08||7.75||27.88|
|Average dissimilarity: 62.65%||Organic vineyards||Conventional vineyards|| || || |
| Drassodes solitarius ||3.10||2.97||1.15||7.86||16.07|
| Trachyzelotes jaxartensis ||2.63||1.70||1.13||8.40||24.47|
|Linyphiidae sp. 1a||0.77||0.67||1.01||5.82||40.27|
Sites differed in terms of their environmental variables (Table 2). Natural sites generally had the highest plant species richness, plant height, indigenous plant cover, and percentage natural habitat in a 500 m radius. On average, organic vineyards had the highest weed cover and leaf litter depth and dry weight, whereas conventional vineyards had the highest intensity indices for insecticide, fungicide, herbicide, and fertiliser (Table 2). Five of the original 13 environmental variables had a significant effect on spider abundance and species richness: percentage weed cover, plant species richness, insecticide intensity, tillage intensity, and distance to nearest natural habitat (Table 5).
Table 5. Results of a generalised linear model with Poisson distribution and log-link function, indicating the influence of various biotic and abiotic variables on spider abundance and species richness
|Variables||Spider abundance||Spider species richness|
|% Weed cover||1||105.18|| <0.0001 ||1||18.66|| <0.0001 |
|Plant species richness||1||48.43|| <0.0001 ||1||6.01|| 0.01 |
|Leaf litter depth||1||0.00||0.96||1||0.42||0.52|
|Insecticide intensity||1||5.40|| 0.02 ||1||4.13|| 0.04 |
|Tillage intensity||1||62.25|| <0.0001 ||1||12.66|| <0.001 |
|% Natural habitat in 500 m radius||1||0.12||0.73||1||0.05||0.82|
|Distance to nearest natural habitat||1||72.18|| <0.0001 ||1||10.71|| <0.001 |
Season had a significant effect on spider abundance (χ2 = 23.34, P < 0.0001; Fig. 4a) and species richness (χ2 = 9.79, P < 0.005; Fig. 4b). Spider abundance and species richness increased from May (autumn) to October (spring) at all sites. Significant differences between seasons were seen more frequently in the vineyard sites than natural sites (Fig. 4a and b).
Figure 4. Mean spider (a) abundance and (b) species richness per site in autumn and spring (±SE). Site abbreviations: N1–N3 = natural habitats, localities 1–3, O1–O3 = organic vineyards, localities 1–3, C1–C3 = conventional vineyards, localities 1–3. * = autumn and spring samples significantly different at P < 0.05.
Download figure to PowerPoint