Agriculture, Ecosystems and Environment · 2018-12-10 · under sub-humid conditions in the sub-tropics. Weeds can contribute to plant diversity in croplands, ... It has been shown

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment

journal homepage: www.elsevier.com/locate/agee

Spider community shift in response to farming practices in a sub-humidagroecosystem of southern AfricaNilton Mashavakurea,⁎, Arnold B. Mashingaidzea, Robert Musundirea, Nhamo Nhamob,Edson Gandiwac, Christian Thierfelderd, Victor K. Muposhica School of Agricultural Sciences and Technology, Chinhoyi University of Technology, Private Bag 7724, Chinhoyi, Zimbabweb Institute of Research, Innovation and Technological Solutions, Zimbabwe Open University, 29 Samora Machel Avenue, P. O. Box MP1119, Harare, Zimbabwec School of Wildlife, Ecology and Conservation, Chinhoyi University of Technology, Private Bag 7724, Chinhoyi, Zimbabwed CIMMYT, P. O. Box MP 163, Mount Pleasant, Harare, Zimbabwe

A R T I C L E I N F O

Keywords:Biological pest controlFertilizationNo-tillagePredator relationshipsSpider diversityWeeding effects

A B S T R A C T

Most spiders are generalist predators and important biological control agents of various insect pests of agri-cultural crops. A study was conducted to determine the impact of cultural practices on the abundnace anddiversity of soil surface-dwelling spiders (Araneae). Two experiments were conducted at the Chinhoyi Universityof Technology experimental farm, Zimbabwe, over the 2013/2014 and 2014/2015 cropping seasons. The twoexperiments were conducted using a split-split-plot design arranged in randomized complete blocks using tillage,mulching, fertilizer and weeding management as factors, with spider diversity being a response variable. Tillageand mulching had strong effects on spider composition. In the first experiment that involved tillage system as themain plot factor, conventional tillage had a negative effect on ground dwelling taxa as evidenced by high ne-gative taxon weights of Lycosidae, Gnaphosidae and Salticidae. In the second experiment, mulching had strongpositive effects on ground dwelling spiders with the strongest being Lycosidae followed Gnaphosidae andThomisidae. The no-tillage option increased richness by 14.5% compared to conventional tillage. The effectivenumber of species was higher in the no-tillage option (exp^Hʹ = 2.2) than in conventional tillage (exp^Hʹ = 1.8).Our results suggest that no-tillage and retention of plant residue on the soil surface facilitate the abundance ofground and plant wandering spiders. More research is required to assess the specific benefits associated with thisincreased abundance, such as biological pest control.

1. Introduction

Spiders are beneficial predators that have been reported to prey on avariety of insect pests in agroecosystems including aphids, caterpillarsand beetles, forming an important component of biological pest control(Clough et al., 2005; Menalled et al., 2007). In agroecosystems, thedistribution pattern, abundance and diversity of spiders is usually ne-gatively affected by high external input farming systems such as me-chanical tillage, crop residue removal, fertilizer application and highweeding intensity (Butt and Sherawat, 2012). Agricultural managementactivities frequently cause structural degradation of habitats, followedby a concomitant loss in spider abundance and diversity. Farmingsystems that conserve biodiversity can play an important role in theenhancement of important ecological processes such as biological pestcontrol (Scherr and McNeely, 2008). Evidence shows that productionsystems such as no-tillage, reduced use of chemical fertilizers and

reduced weeding frequency support abundant and diverse communitiesof both predators and prey (Chaplin-Kramer and Kremen, 2012). Whenno-tillage is combined with retention of about 30% plant residue on thesoil surface, it can provide some benefits such as increased crop yieldsand biological activity (Thierfelder and Wall, 2010; Soane et al., 2012;Boscutti et al., 2015; Thierfelder et al., 2015). Ground surface dwellingspiders normally take refuge and sometimes overwinter in the soil andplant debris. Their survival, diversity and distribution are strongly af-fected by structural disturbances and distribution of resources (Chaplin-Kramer and Kremen, 2012). It follows that no-tillage, which providesstable microhabitats, preserves nesting sites and reduces the mortalityof arthropods, enhances increased spider abundance, diversity andother associated ecosystem services (Stinner and House, 1990). How-ever, little is known about the impacts of tillage on spider dynamicsunder sub-humid conditions in the sub-tropics.

Weeds can contribute to plant diversity in croplands, providing

https://doi.org/10.1016/j.agee.2018.11.020Received 7 August 2018; Received in revised form 14 November 2018; Accepted 24 November 2018

⁎ Corresponding author.E-mail address: [email protected] (N. Mashavakure).

Agriculture, Ecosystems and Environment 272 (2019) 237–245

0167-8809/ © 2018 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

T

suitable habitats for diverse populations of herbivorous arthropods(Haddad et al., 2009). It has been shown that weed control practicesdestroy food resources for herbivorous insects, resulting in reducedpopulation and diversity of available prey for spiders (Butt andSherawat, 2012). On the other hand, fertilizer application increasesplant biomass production, supporting high numbers of herbivores aswell as detritivores, thus enhancing predator abundance (Siemann,1998). Whilst plant biomass production increases under high fertilityconditions, plant diversity has been shown to decline due to thinning ofless competitive species (Suding et al., 2005). Evidence suggests thatoptimum combinations of tillage system, fertilizer application rate,mulching rate and weeding intensity can enhance spider conservationand biological pest control (Shennan, 2008).

Spiders occupy a high trophic level and are classified into threemain functional groups: web builders that trap their prey using silkwebs, ground - and plant wanderers which wander around on soil andplant surfaces in search of their prey (Stinner and House, 1990; Kromp,1999; Uetz, 1999). Ground wanderers, which require ground surfacesubstrates for shelter and ambushing prey, may respond more readily tochanges in ground surface architecture than plant wanderers. On theother hand, plant wanderers may be affected by plant density becausethey require standing plant biomass, especially during the humidseason when spider and prey activity is generally high. Moreover,seasonal changes in environmental conditions such as temperature,moisture and humidity influence the response of spider communitydiversity to habitat manipulation.

The intermediate hypothesis predicts that community diversity ishigh when the frequency and intensity of disturbance is at an inter-mediate scale (Connell, 1978). Studies from different parts of the world,mostly Europe, America and Australia, have shown the positive effectsof no-tillage on agroecosystem biodiversity (Whitehouse et al., 2009;Rendon et al., 2015). Although increased arthropod species richnessand diversity have been found to be beneficial for biological activity,the impact of agroecosystem complexity on arthropod diversity andnatural enemy populations is not fully understood (Kerzicnik et al.,2013). Reports of work in no-tillage systems, especially from the de-veloping world such as sub-Saharan Africa, are very scarce. The present

study was conducted to determine the response of spiders under dif-ferent management practices involving no-tillage systems in Zimbabwe.We tested the hypotheses that spider abundance and diversity: (1) de-crease with increasing intensity of mechanical soil disturbance, (2) in-crease with increasing amount of plant residue cover, (3) increase withincreasing amount of fertilizer application, and (4) decrease with in-creasing intensity of weeding.

2. Materials and methods

2.1. Site description

Two experiments (Experiment 1 and Experiment 2) were conductedat Chinhoyi University of Technology experimental farm, Zimbabwe,(17○20′S, 30○14′E). The site is situated in a sub-tropical environmentwith an altitude of 1140 m. The mean annual rainfall for the studyperiod is 850 mm whilst mean maximum temperature during summer is27 ○C and mean minimum temperature during the cold dry season is 7○C (Climatemps.org, 2017). The soils found in the study site are Cam-bisols (WRB, 1998). The site has three distinct seasons: (1) the rainyseason that occurs from mid-November to mid-April each year and isgenerally hot and humid (hot-humid season), (2) the cool, dry andsunny season which occurs between mid-April to mid-August (cool dryseason), and (3) a short dry season of intense heat which occurs be-tween mid-August and mid-November (hot dry season)(Climatestotravel.com, 2017).

2.2. Experimental design and treatments

The two experiments were designed as a split-split-plot in rando-mized complete blocks with three replications. The treatments used inthe two experiments are summarized in Table 1. Measurements wereconducted over two crop growing seasons, that is, 2013/2014 and2014/2015 seasons; approximately 13 months after the establishmentof a CA experiment in November 2012.

Table 1Summary of treatments used in the two experiments at Chinhoyi University of Technology experimental farm between January 2014 and August 2015.

Factor Levels Description

Experiment 1 Experiment 2

Tillage system Basin planting Planting basins were made using hoes, each basin measuring about15 × 15 × 15 cm

Planting basins were made using hoes, each basinmeasuring about 15 × 15 × 15 cm

Rip line seeding Rip lines were marked using a tractor-mounted ripper to a depth ofabout 15 cm

Not applicable

Conventional tillage (CT) Land preparation was done using a disc plough as primary tillagefollowed by secondary tillage using a disc harrow.

Land preparation was done using a disc plough asprimary tillage followed by secondary tillage usinga disc harrow.

Fertilizer regime No fertilizer (NF) No fertilizer was applied. Not applicablePrecision ConservationAgriculture rate (Low fertilizerrate, LF)

One handful of manure per planting station + 80 kg ha−1 compoundD fertilizer (8 N: 14 P2O5: 7 K2O) + 80 kg ha−1 ammonium nitrate(34.5 % N) [35.2 kg ha−1 N: 12.2 kg ha-1 P2O5: 6.6 kg ha-1 K2O].

Not applicable

Medium fertilizer rate (MF) 100 kg ha−1 compound D fertilizer (8 N: 14 P2O5: 7 K2O) + 100 kgha−1 ammonium nitrate (34.5 % N) [41.5 kg ha−1 N: 14 kgha−1P2O5: 7 kg ha−1K2O].

Not applicable

High fertilizer (HF) rate 200 kg ha−1 compound D fertilizer (8 N: 14 P2O5: 7 K2O) + 200 kgha−1 ammonium nitrate (34.5 % N) [83 kg ha−1 N: 28 kg ha−1P2O5:14 kg]

Not applicable

Weeding intensity Weeding twice Weeding at two and four WACE. Weeding at two and four WACE.Weeding three times Not applicable Weeding at two, four and six WACE.Weeding four times Weeding at two, four, six and eight WACE. Weeding at two, four, six and eight WACE.Clean weeding No weed growth was allowed throughout the growing season. No weed growth was allowed throughout the

growing season.Mulching rate No mulching Not applicable No mulch was applied to these plots.

Low mulching rate Not applicable 4 t ha−1 maize stover mulch were applied.Medium mulching rate Not applicable 8 t ha−1 maize stover mulch were applied.High mulching rate Not applicable 12 t ha−1 maize stover mulch were applied.

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

238

2.2.1. Experiment 1: effects of tillage, fertilizer application and weeding onspider abundance and diversity

In particular: (1) the main plot factor was tillage system comprisingbasin planting, rip-line seeding and conventional tillage (CT); (2) thesub-plot factor was fertilizer regime comprising the no-fertilizer option,micro dosing (low fertilizer rate), medium fertilizer rate and high fer-tilizer rate and (3) the sub-sub-plot factor was weeding regime (i.e.weeding twice, weeding four times and clean weeding). There were,therefore, 36 treatment combinations resulting in a total of 108 plots.Crop residue was removed after harvesting in the CT plots and retainedin the no-tillage plots.

2.2.2. Experiment 2: effect of tillage, mulching and weeding on spiderabundance and diversity

In particular: (1) the main plot factor was a tillage system com-prising basin planting and CT; (2) the sub-plot factor was mulching rate(no-mulching, low mulch, medium mulch and high mulch) and (3) thesub-sub-plot factor was weeding regime (weeding twice, weedingthrice, weeding four times and clean weeding). In this experiment,there were 32 treatment combinations resulting in a total of 96 plots.Crop residue was completely removed from all plots. Mulching treat-ments were applied immediately after planting the test crop.

2.3. Agronomic practices

Basins were made using hand-held hoes; each basin measuringabout 15 × 15 × 15 cm (Sims et al., 2012). Rip-lines were made to adepth of about 10–15 cm using a tractor-mounted ripper. Conventionaltillage was done using a disc plough to a depth of 25–30 cm as primarytillage followed by secondary tillage using a disc harrow. Primary til-lage was done in May soon after harvesting the test crop and secondarytillage was done in October of each year. An initial fertilizer (7% N,14% P2O5 and 7% K2O) was applied at planting, about 5 cm below and5 cm beside the crop seed. Top dressing fertiliser (ammonium nitrate,34.5% N) was applied 4–6 weeks after crop emergence (WACE), on thecondition that at least 40 mm of rainfall had been received. Weedingwas done using a hand-held hoe according to treatments (Table 1). Inthe first experiment, all the fertilizer was applied according to treat-ments (Table 1). In the second experiment initial fertilizer was appliedat a rate of 400 kg ha−1, being compound fertilizer (7% N, 14% P2O5,7% K2O) and top dressing fertilizer at 400 kg ha−1, that is ammoniumnitrate (34.5% N). In each experiment, treatments were applied on thesame plots every rainy season since the establishment of the tillagetrials in the 2012/2013 cropping season. In the 2013/2014 and 2014/2015 cropping seasons, a local medium maturity maize hybrid variety(ZAP61) was used as the test crop, planted at four seeds per plantingposition at a plant spacing of 0.9 m inter-row by 0.5 m intra-row. Thecrop was thinned to two plants per planting position at three WACE togive a plant population of 44,444 plants ha−1. Sub-sub plots were 7.2 mwide × 8 m long and separated by 1.8 m pathways. Measurements weretaken from the four central rows of the sub-sub plots after discarding0.6 m from either side of each row.

2.4. Treatment effects on spider abundance and diversity

To determine the effect of tillage, fertilizer, mulching and weedingon spider abundance and diversity in the agroecosystem, spiders weresampled using unbaited pitfall traps placed within each net plot (4central rows). Pitfall trapping is a method that is widely used in sam-pling of soil surface dwelling spiders (da Silva et al., 2008). The trapswere made of plastic jars measuring about 13 cm diameter and 1000cm3 in volume. The traps were half filled with a mixture of 20% alcoholto collect and preserve fauna samples. Two pitfall traps were randomlyset up within each net plot at least 2 m apart, and sampling was over

two crop growing seasons in the 2013/2014 and 2014/2015 agri-cultural calendar years. During each sampling period, traps were left inplace for seven days, and were emptied after every two days by filteringout arthropod specimens through a strainer. For each seven day sam-pling period, spider catches from the two traps in each plot were pooledtogether to form one sample. The traps were then rested for 14 days bycovering them with plastic sheets to avoid continuous trapping of ar-thropods during the resting period. A total of 216 pitfall traps in ex-periment 1 and 192 in experiment 2 were maintained throughout theperiod of study, giving a combined total of 408 pitfall traps for bothexperiments.

Arthropod specimens were placed in 70% alcohol for further proces-sing. Spiders were counted and identified to family level in the laboratoryaccording to (Dippenaar-Schoeman and Jocque, 1997). Three main func-tional groups were identified: ground wanderers, plant wanderers and webbuilders (Uetz, 1999). Ground wanderers and plant wanderers wanderaround on soil and plant surfaces in search of their prey whilst webbuilders trap their prey using silk webs (Whitmore et al., 2002).

2.5. Statistical analyses

A direct gradient analysis, Canonical Correspondence Analysis(CCA), was used to establish the relationship between managementpractices (tillage, fertilizer rate, mulching rate and weeding regime)and spider family composition for each experiment separately. Dataanalysis using CANOCO 5 (ter Braak and Smilauer, 2012) indicated thatthe spider data had gradient lengths of 3.6 SD (experiment 1) and 4.8(experiment 2) units which suited further analysis using CCA (Šmilauerand Lepš, 2014). CCA analyses also showed that tillage (experiment 1)and mulching (experiment 2) had the strongest effects on spider com-munity composition. Spider data was further examined using principleresponse curve (PRC) analysis in CANOCO 5 (ter Braak and Smilauer,2012). PRC is a multivariate technique which is suitable for the analysisof repeated measures which is designed to test and display the effects oftreatments and their changes over time (Whitehouse et al., 2005). In thePRC, treatment curves are presented relative to a standard. In thisstudy, the standard treatments were conventional tillage for experiment1 and no-mulching for experiment 2. The PRC is derived from partialredundancy analysis (RDA), with the Y-axis representing axis 1 of theRDA, whereas the X-axis represents the sampling period (Whitehouseet al., 2005, 2014). Permutation tests were conducted on the first ca-nonical axis of the RDA using the Monte Carlo method to test if the PRCexplained significant treatment variance. A second graph showingtaxon weights was plotted together with the PRC. The taxon weightrepresents the weight of each single taxon for the response given in thePRC diagram. A taxon with a high weight shows that its actual responseis more likely to follow the response that is shown in the PRC. A highpositive value of a taxon weight shows that the contribution of thetaxon to the accomplishment of the PRC is also high whilst a taxon witha high negative value shows a reversed image to the PRC (Moser et al.,2007). In this study, taxa whose weights occurred between –0.5 and 0.5had little effect on the PRC and were not considered for further statis-tical tests (Whitehouse et al., 2014).

Diversity indices, i.e., Shannon-Wiener entropy index, richness andevenness were computed using Paleontological Statistics (PAST)package version 3.14 (Shannon and Weaver, 1949; Hammer et al.,2001). The Shannon-Wiener entropy index was converted to true di-versity using the formula: exp^Hʹ, in order to estimate the effectivenumber of species (Jost, 2006).

Damaged spiders that could not be identified to the relevant taxo-nomic levels were excluded from the diversity index estimation.According to Shapiro-Wilk’s test for normality and Bartlett’s test forhomogeneity of variances, respectively, spider diversity data were foundto be normal and needed no transformation. However, abundance data

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

239

required transformation and were log(x+1.5)-transformed.The mixed model (Residual Maximum Likelihood, REML) repeated

measures procedure was used to determine the effect of tillage, fertilizer,mulching and weeding on spider diversity data and abundance of thosespider families whose taxon weights were above 0.5. Tillage, fertilizer,mulching and weeding treatments were used as fixed effects while sam-pling season (hot humid, cold dry or hot dry) was used as a random effect.In each experiment, the models included the three treatments and all theirinteractions as fixed effects. For the first experiment, the fixed model was:Constant + Block + Tillage + Fertilizer + Weeding + Tillage ×Fertilizer + Tillage × Weeding + Fertilizer × Weeding + Tillage ×Fertilizer × Weeding. For the second experiment, the fixed model was:Constant + Block + Tillage + Mulching + Weeding + Tillage ×Mulching + Tillage × Weeding + Mulching × Weeding + Tillage ×Mulching × Weeding. The random model was expressed as: Season.Analyses were done using GenStat Release for Windows Version 10 (VSN-International, 2011)

3. Results

3.1. Spider characterization and weather parameters

The 2013/2014 cropping season was slightly more humid than the2014/2015 season (Table 2). A total of 4425 spider (Order: Areneae)specimens representing three functional groups (ground wanderers,plant wanderers and web builders) from 13 families, were collectedbetween January 2014 and August 2015. Of all the spiders collectedduring the study, 88.3% were ground wanderers from four families(Corinidae, Ctenidae, Gnaphosidae and Lycosidae), 10.7% were plantwanderers from four families (Oxyopidae, Salticidae, Thomisidae andPhilodromidae), whilst web builders (Araneidae) comprised 1%(Table 2).

3.2. Treatment effects on spider abundance and diversity

Constrained ordination using CCA showed most spider families wereassociated with no-tillage (Fig. 1). The first Axis accounted for 45.44%variability (eigenvalue = 0.0157) whilst the second Axis accounted for24.15% variability (eigenvalue = 0.0095) of the cumulative explainedfitted variation. Tillage had a strong effect on species composition, asboth conventional tillage (Conv) and rip lines (Ripping) align with thefirst axis in the ordination diagram. No-fertilizer treatment, thoughshort, aligned in the illustration, with clean weeding giving indicationsthat the environmental variable Cle had a stronger but similar effect onspecies composition. High fertilizer (HF) had little effect on speciescomposition. The family Ctenidae with a low abundance of the sevenindividual spiders (Fig. 1) was strongly associated with rip line seedingand low fertilizer (LF) application.

In the second experiment, Axis 1 explained 45.02% (eigenvalue =0.0125) and Axis 2 explained 24.15% (eigenvalue = 0.0067) of thecumulative explained fitted variation, respectively (Fig. 2). CCAshowed that mulching had the strongest effect on community compo-sition; medium mulching (MM) and no-mulching (NM) align with thefirst axis whilst low mulching (LM) and high mulching (HM) align withthe second axis (Fig. 2). Weeding three times had a weaker but similar

Table 2Total number of spiders collected from each family and functional group in twoexperiments at Chinhoyi University of Technology experimental farm betweenJanuary 2014 and August 2015.

Spider family Functional groups of spiders

Experiment 1 Experiment 2

WB GW PW WB GW PW

Aranedae 24 ‾ ‾ 15 ‾Gnaphosidae ‾ 2182 ‾ ‾ 295 ‾Lycosidae ‾ 703 ‾ ‾ 725 ‾Oxyopidae ‾ ‾ 42 ‾ ‾ 30Saltisidae ‾ ‾ 183 ‾ ‾ 30Thomisidae ‾ ‾ 87 ‾ ‾ 30Coridae ‾ 1 ‾ ‾ ‾ ‾Philodromidae ‾ ‾ 14 ‾ ‾ ‾Tetragnathidae 1 ‾ ‾ ‾ ‾ ‾Ctenidae ‾ 7 ‾ ‾ ‾ ‾

Functional groups of the spiders are: Ground wanderer (GW), plant wanderer(PW) and web builder (WB).

Fig. 1. Canonical correspondence analysis (CCA) bi-plot ofspider family composition in relation to tillage, fertilizer andweeding treatments at Chinhoyi University of Technologyexperimental farm between January 2014 and August 2015.Notes: Lycosidae [Lyco], Salticidae [Salti], Gnaphosidae[Gnap], Thomisidae [Thom], Philodromidae [Phil],Oxyopidae [Oxyo], Araneidae [Aran] and Ctenidae [Cten].Basin planting [BAS], rip line seeding [RIP], conventionaltillage [CONV], no-fertilizer [NF], low fertilizer [LF], mediumfertilizer [MF], high fertilizer [HF], weeding twice [TWI],weeding three times [THR], weeding four times [FOU] andcleaning weeding [CLE].

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

240

effect on community composition than medium mulching. High ferti-lizer (HF) had little effect on community composition. The familySalticidae was associated with no mulching whilst Oxyopidae was as-sociated with high mulching.

The PRCs showed a significant effect of tillage system (F = 30.4,DF = 7, P = 0.002) and mulching rate (F = 11.4, DF = 7, P = 0.04) onspider community composition (Figs. 3 and 4). Throughout the studyperiod, the two no-tillage treatments tracked together and the highestdeviations in spider community composition were observed during thehot dry season of 2014, that is two years after commencement of theexperiment (Fig. 3). Two ground wanderers, Lycosidae and Gnapho-sidae, and a plant wanderer, Salticidae, had the highest, second highestand third highest negative taxon weights respectively, implying that thefamilies were more abundant in no-tillage than in conventional tillagetreatments. Their activity densities were correspondingly found to besignificantly (P < 0.05) higher under minimum than conventionaltillage in both the 2013/2014 and 2014/2015 cropping seasons(Table 3). Differences in spider community composition between mul-ched and no-mulch treatments were most evident during the hot humidand cold dry seasons of 2014 (Fig. 4). The highest deviation in spidercommunity composition was observed in the low mulch treatment.However, this trend was not consistent across the sampling seasons.Lycosidae had the highest positive taxon weight, followed by Gna-phosidae and a plant wanderer, Thomisidae, suggesting that these threetaxa had stronger effects of increasing spider abundance. In the 2013/2014 cropping season, abundance levels of Gnaphosidae and Thomi-sidae were significantly (P < 0.05) highest in plots treated with 4 tha−1 of mulch (Table 4). However, mulching had no effect on spiderabundance in the 2014/2015 cropping season (Table 4).

3.3. Treatment effects on spider diversity

In the first experiment, no-tillage significantly increased spider di-versity by 27.8% (F-value = 6.87, df = 2, P = 0.051). However, there

were no significant differences across the tillage treatments for spiderdiversity in the second experiment (F-value = 2.38, df = 1, P = 0.263)and evenness in both experiments (Experiment 1: F-value = 4.98,df = 2, P = 0.082; Experiment 2: F-value = 6.96, df = 1, P = 0.119).Moreover, the species richness of spiders was significantly differentacross the tillage treatments (experiment 1: F-value = 52.67, df = 2,P = 0.001; experiment 1: F-value = 18.75, df = 1, P = 0.049). Thethree diversity measures for the spiders across the fertilizer, mulchingrate and weeding rate treatments were not significantly (P > 0.05)different (Table 5).

4. Discussion

4.1. Spider characterization, abundance and diversity

Our study focused on the response of spider community compositionto tillage, mulching, fertilizer and weeding management. We found acomplex community of spiders comprising different families from dif-ferent feeding functional groups. Ground and plant wanderers werenumerically the most dominant, together contributing 99% of the totalspider catches. The observed trend is similar to the findings of Weeksand Thomas (2000) who reported spiders from these two functionalgroups to be most abundant (85%) relative to other spider groups. Itshould be noted that the trapping method used in our study results incollection of more ground than aerial arthropods, and this partly ex-plains the reduced abundance of aerial spiders such as web builders inour study sample (da Silva et al., 2008).

Our results also showed that spider faunae, particularly groundwanderers and plant wanderers, were associated with no-tillage.Significant differences were also observed between conventional tillageand the two no-tillage treatments. These results are consistent withprevious observations (Hatten et al., 2007; Mutema et al., 2013;Henneron et al., 2014) and further demonstrated that besides its en-hancement of total spider abundance, no-tillage increases abundance of

Fig. 2. Canonical correspondence analysis (CCA) bi-plot ofspider family composition. in relation to tillage, mulching andweeding treatments at Chinhoyi University of Technologyexperimental farm between January 2014 and August 2015.Lycosidae [Lyco], Salticidae [Salti], Gnaphosidae [Gnap],Thomisidae [Thom], Oxyopidae [Oxyo] and Araneidae[Aran].Notes: no-mulch [NM], low mulch [LM], medium mulch[MM] high mulch [HM], weeding twice [TWI], weeding threetimes [THR], weeding four times [FOU], clean weeding [CLE],basin seeding [BAS] and conventional tillage [CONV].

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

241

spider taxa such as Lycosidae, Gnaphosidae and Saltisidae. The ob-served effects of tillage on wandering spider groups (ground wanderersand plant wanderers) are most likely related to their impact on thephysical structure of the soil microhabitat. Conventional tillage breaksthe soil and buries crop residue, destroying suitable habitats for re-production, shelter and ambush points for wandering spider groups.Additionally, it has direct negative impacts on arthropods, includingcrushing and exposure to desiccation and predators. On the other hand,non-disturbance under no-tillage reduces spider mortality and henceenables proliferation and sustenance of a stable spider populationgrowth (Gavish-Regev et al., 2008; Benhadi-Marin et al., 2013). Some

species of Lycosidae such as Geolycosa fattier, G. missouriensis, G. rogersiand Hogna lenta construct burrows in the ground that serve as retreatsin the advent of predators and adverse weather conditions (Suter et al.,2011; Dippenaar-Schoeman et al., 2013). Tillage destroys these bur-rows yet the burrows are more persistent under no-tillage, leading toproliferation of spiders under no-tillage as opposed to conventionaltillage. Tillage also destroys the vegetative food resources of herbi-vorous arthropod prey of spiders. Previous studies showed that groundwanderers such as Lycosidae play an important density-dependent rolein controlling agricultural pests that include Helicoverpa spp., aphids(Aphididae), Trichoplusia spp. and Plutella spp. (Zhao et al., 1989;

Fig. 3. Principal response curves and species scores of spiderpitfall trap data sets for experiment 1 at Chinhoyi University ofTechnology experimental farm between January 2014 andAugust 2015.Notes: Lyco (Lycosidae), Gnap (Gnaphosidae), Thom(Thomisidae), Oxyo (Oxyopidae), Aran (Araneidae), Salti(Saltisidae), Phil (Philodromidae, Cten (Ctenidae); January –April 2014 (hot and humid season, HH-14), May – August2014 (cold and dry season, CD-14), September – November2014 (hot and dry season, HD-14), December 2014 – April2015 (hot and humid season, HH-15), May – August 2015(cold and dry season, CD-15), Conv (conventional tillage), Rip(rip line seeding) and Basin (basin planting).

Fig. 4. Principal response curves and species scores of spiderpitfall trap data sets for experiment 2 at Chinhoyi University ofTechnology experimental farm between January 2014 andAugust 2015.Notes: Lyco (Lycosidae), Gnap (Gnaphosidae), Thom(Thomisidae), Oxyo (Oxyopidae), Aran (Araneidae), Salti(Saltisidae);January – April 2014 (hot and humid season, HH-14), May –August 2014 (cold and dry season, CD-14), December 2014 –April 2015 (hot and humid season, HH-15), May – August2015 (cold and dry season, CD-15), control (no mulch ap-plied), high mulch (12 t ha−1 mulch applied), low mulch (4 tha−1 mulch applied), medium mulch (8 t ha−1 mulch ap-plied).

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

242

Suenaga and Hamamura, 2015; Rendon et al., 2016). Plant wandererssuch as Salticidae are polyphagous predators of insect pests includingHelicoverpa zea, Tetranychus spp., Spodoptera spp., leaf hoppers (Cica-dellidae) and weevils (Coleoptera) (Randall, 1982; Mansour et al.,1995; Dippenaar-Schoeman et al., 1999). Nyffeler (1999) showed thatinsects constitute about 75–90% of total prey of ground wanderers andplant wanderers combined.

In the second experiment, Lycosidae had the strongest effect of in-creasing abundance in mulched plots, followed by Gnaphosidae andThomisidae. These results concur with results from Whitmore et al.(2002), Dippenaar-Schoeman et al. (2013) and Hossain and Begum(2015). The effect of mulch could be tracked to its ability to providerefugia which act as ambush points for ground wandering spiders, aswell as for shelter and construction of silk sacs for trapping prey. Fur-thermore, it is possible that the mesic microhabitat created by mulchprovided suitable conditions for the ground and plant wanderers’ eco-logical requirements, concomitantly resulting in their increased sur-vival and reproductive capability. Our results show that in the morehumid year of the 2013/2014 cropping season, application of 4 t ha−1

of mulch fostered the highest spider abundance whereas during thedrier 2014/2015 cropping season, application of 12 t ha−1 of mulchhad the highest spider abundance. It is, therefore, possible that highermulching rates above 4 t ha−1 may dampen the effect of tillage in re-ducing spider abundance.

Fertilizer application and weeding regimes did not significantly

affect spider abundance and diversity in this study; contrary to thefindings of some studies done under temperate environments inPortugal, Germany and United States of America by Benhadi-Marinet al. (2013), Haddad et al. (2000) and Hertzog et al. (2016). There isno evidence from our study to support that fertilizer increased plantbiomass production, herbivorous arthropods and subsequently influ-encing spider community structure. It is most likely that the variancebetween our results and those cited above can be explained by thedifferences in climatic conditions of the study sites, particularly tropicalversus temperate regions.

Our results suggest that Lycosidae, Gnaphosidae, Salticidae andThomisidae are among the potential drivers of the ground spidercommunity response to agricultural management practices. Thesespider faunae are among some of the most numerous generalist

Table 3Tillage effects on activity density of spiders (Areneae) under a tillage ×fertilizer × weeding experiment at Chinhoyi University of Technology experi-mental farm, Zimbabwe between January 2014 and August 2015.

Tillage treatments F-statistic P-value SED

Basin Rip Conv

2013/2014 cropping seasonOverall spiders 11.1 b 10.2 b 7.1 a 19.83 < 0.001 0.027Gnaphosidae 6.6 b 6.7 b 4.5 a 5.29 0.006 0.030Lycosidae 2.5 b 2.0 b 1.0 a 15.33 < 0.001 0.027Salticidae 0.4 b 0.4 b 0.1 a 10.20 < 0.001 0.0152014/2015 cropping seasonOverall spiders 1.3 a 1.2 a 2.0 b 3.13 0.046 0.036Gnaphosidae 2.3 2.9 2.7 1.37 0.257 0.034Lycosidae 0.7 b 0.4 a 0.4 a 3.84 0.023 0.022Salticidae 0.5 b 0.6 c 0.3 a 3.58 0.030 0.023

Means followed by the same letter in the same row are not significantly dif-ferent (treatment means were separated using standard error of the differencesof means (SED) at P ≤ 0.05). Data were log(x+1.5) transformed values. Rip(rip line seeding), basin (basin planting), conv (conventional tillage), and SED(standard error of difference between means).

Table 4Spider activity density as affected by mulch application rate under a tillage × mulching × weeding experiment at Chinhoyi University of Technology experimentalfarm, Zimbabwe between January and August 2015.

Mulching treatments F-statistic P-value SED

No mulch 4 t ha−1 8 t ha−1 12 tha−1

2013/2014 cropping seasonOverall spiders 4.8 a 7.6 b 5.6 b 6.5 b 6.69 < 0.001 0.038Gnaphosidae 1.1 a 1.9 c 1.3 ab 1.5 bc 3.55 0.016 0.037Lycosidae 2.6 4.4 2.9 3.5 2.32 0.077 0.043Thomisidae 0.2 a 0.3 c 0.3 ab 0.2 bc 5.91 < 0.001 0.0222014/2015 cropping seasonOverall spiders 0.9 0.7 0.8 1.1 0.76 0.517 0.036Gnaphosidae 0.1 0.1 0.0 0.1 0.89 0.448 0.014Lycosidae 0.4 0.4 0.4 0.6 1.03 0.381 0.029Thomisidae 0.1 0.1 0.1 0.1 0.76 0.517 0.015

Means followed by the same letter in the same row are not significantly different (treatment means were separated using standard error of the differences of means(SED) at P ≤ 0.05). Figures in parentheses represent log(x+1.5) transform values. SED (standard error of difference between means).

Table 5Effect of tillage, fertilizer, mulching and weeding regime on mean spider taxarichness, diversity and evenness index (E) at Chinhoyi University of Technologyexperimental farm, Zimbabwe, between January 2014 and August 2015.

Treatment Richness index Exp^Hʹ Evenness index

Tillage system Exp 1 Exp 2 Exp 1 Exp 2 Exp 1 Exp 2Basins 3.194 b 3.042 b 2.16 b 2.37 0.711 0.789Rip-lines 3.444 c – 2.15 b – 0.650 –CT 2.611 a 2.729 a 1.83 a 2.25 0.735 0.844P-value 0.001 0.049 0.048 0.461 0.082 0.119SED 0.083 0.072 0.099 0.132 0.028 0.021Fertilizer rateNo fertilizer 3.037 2.07 0.724Low rate 3.370 2.09 0.646Medium rate 3.074 2.10 0.714High rate 2.852 1.93 0.711P-value 0.358 0.433 0.131SED 0.2837 0.114 0.035Weeding regimeClean weeding 3.000 2.792 2.09 2.27 0.694 0.809Weeding twice 3.028 2.833 2.00 2.24 0.708 0.838Weeding three times – 3.125 – 2.32 – 0.792Weeding four times 3.222 2.792 2.05 2.43 0.694 0.827P-value 0.544 0.499 0.753 0.772 0.088 0.454SED 0.218 0.254 0.011 0.196 0.032 0.030Mulching rateNo mulch 2.583 2.12 0.8414 tonnes ha−1 3.083 2.36 0.7748 tonnes ha−1 3.125 2.48 0.80612 tonnes ha−1 2.750 2.29 0.846P-value 0.387 0.604 0.089SED 0.354 0.271 0.029

Means with different superscripts in the same column are significantly different(treatment means were separated using standard error of the differences ofmeans (SED) at P ≤ 0.05). SED: standard error of difference between means.Exp^Hʹ: effective number of species.

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

243

invertebrate predators that prey on arthropods (Rahman et al., 2007).They survive, disperse and find their prey in field crop ecosystemsbetter than web builders (Plaza et al., 2011). The higher abundance ofthese spider families in no-tillage and mulching regimes suggests thepotential of these management practices in enhancing natural biolo-gical pest control efficacy in CA-based agroecosystems. The consistentlyhigh density of overall spider faunae under no-tillage and mulch ad-dition further suggests that these families can be a reliable biologicalcontrol agent under no-tillage.

4.2. Treatment effects on spider diversity

Even though our research plots were just three years old, our find-ings show that the trajectory of predator diversity in an agroecosystemcan be detected at such early stages of no-tillage and should thereforebe an important early indicator of improved soil quality under no-til-lage systems. Previous studies suggested that agroecosystem biodi-versity increases with the age of the no-tillage plots (Tabaglio et al.,2009). Increased spider richness due to no-tillage suggests that thispractice has the potential to increase diversity. Hypothetically, themigration of new families into the young no-tillage-based agroeco-system should compensate for any loss in taxa evenness caused by thedominance of species that already exist in the microhabitat. This haspotential for a net increase in the community diversity index.

5. Conclusion

The present study confirmed our first hypothesis and showed thatno-tillage increased abundance of some predatory ground wanderers(Lycosidae and Gnaphosidae) and plant wanderers (Thomisidae), whilstalso increasing spider diversity. Moreover, our results partly supportedthe second hypothesis and found increased abundances of Lycosidae,Gnaphosidae and Thomisidae in mulched plots. However, mulching hadno effect on spider diversity. Furthermore, in contrast to our third andfourth hypotheses, we found no effect of both fertilizer addition andweeding intensity on spider abundance and diversity. Our results form astrong basis for promoting integrated pest management under no-tillagecombined with crop residue retention. This is because predatory groundand plant wandering spiders such as Lycosidae, Gnaphosidae andThomisidae were increasingly enhanced as soil disturbance throughtillage was decreased and crop residue was retained. The spider familiesare important natural enemies of agricultural crops. We recommendthat no-tillage and mulch retention should be employed for enhance-ment of spiders for natural biological insect pest control in croppingsystems. Further studies are required to determine if increases in pre-datory ground wanderers and plant wanderers under no-tillage mayhelp to suppress the insect pest activity associated with no-tillage.

Acknowledgements

This research was jointly funded by Chinhoyi University ofTechnology (CUT) and the German Academic Exchange Program(DAAD). The Consultative Group on International Agricultural Research(CGIAR) program on MAIZE supported this study by providing time forC. Thierfelder to contribute to this study through its Flagship Project 1,that is Sustainable Intensification (www.maize.org). We are grateful tothe Natural History Museum of Zimbabwe’s Arachnology Departmentstaff, particularly Dr. Moira Fitzpatrick and Mr. Nyathi, for offering therelevant training in morphological identification of spiders. Dr. AnsieDippenaar-Schoeman donated color spider identification handbooksthat were very useful in the identification of spider specimens. We arealso grateful to Mr. Shadreck Nembaware who language-edited thisdocument. Our special gratitude goes to Messrs. Stewart Maganyani,Albert Chinyati and Lawrence Gweme as well as CUT field staff whoassisted with crop management, data collection and data collation.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in theonline version, at doi:https://doi.org/10.1016/j.agee.2018.11.020.

References

Benhadi-Marin, J., Pereira, J.A., Barrientos, J.-A., Bento, A., Santos, S.A.P., 2013. Araneaecommunities associated with the canopies of chestnut trees in the northeastern partof Portugal: the influence of soil management practices. Eur. J. Entomol. 110,501–508.

Boscutti, F., Sigura, M., Gambon, N., Lagazio, C., Krusi, B.O., Bon-fanti, P., 2015.Conservation tillage affects species composition but not species diversity: a com-parative study in northern Italy. Environ. Manage. 55, 443–452.

Butt, A., Sherawat, S.M., 2012. Effect of different agricultural practices on spiders andtheir prey populations in small wheat fields. Acta Agric. Scand. B 62, 374–382.

Chaplin-Kramer, R., Kremen, C., 2012. Pest control experiments show benefits of com-plexity at landscape and local scales. Ecol. Appl. 22, 1936–1948.

Climatemps.org, 2017. Chinhoyi Climate & Temperature.Climatestotravel.com, 2017. Climate in Zimbabwe: Temperature, Rainfall, Prevailing

Weather Conditions, When to Go, What to Pack.Clough, Y., Knuess, A., Kleijn, D., Tscharntke, T., 2005. Spider diversity in cereal fields:

comparing factors at local, landscape and regional scales. J. Biogeogr. 32,2007–2014.

Connell, J.H., 1978. Diversity in tropical rain forests and coral reefs. Science 199,1302–1310.

da Silva, P.M., Aguiar, C.A.S., Niemela, J., Sousa, J.P., Serrano, R.M., 2008. Diversitypatterns of ground beetles (Coleoptera: Carabidae) along a gradient of land-use dis-turbance. Agric. Ecosyst. Environ. 124, 270–274.

Dippenaar-Schoeman, A.S., Jocque, R., 1997. African Spiders: an Identification Mannual.Biosystematics Division, ARC - Plant Protection Research Institute, Pretoria.

Dippenaar-Schoeman, A.S., Van Den Berg, A.M., Van Den Berg, A., 1999. Spiders in SouthAfrican cotton fields: species diversity and abundance (Arachnida: Araneae). Afr.Plant Prot. 5, 93–103.

Dippenaar-Schoeman, A.S., Foord, S.H., Haddad, C.R., 2013. Spiders of the SavannaBiome. . University of Venda, Thohoyandou, Agricultural Research Council, Pretoria,South Africa.

Gavish-Regev, E., Lubin, Y., Coll, M., 2008. Migration patterns and functional groups ofspiders in a desert agroecosystem. Ecol. Entomol. 33, 202–212.

Haddad, N.M., Haarstad, J., Tilman, D., 2000. The effects of long-term nitrogen loadingon grassland insect communities. Oecologia 124, 73–84.

Haddad, N.M., Crutsinger, G.M., Gross, K.J., Haarstad, J., Knops, J.M.H., Tilman, D.,2009. Plant species loss decreases arthropod diversity and shifts trophic structure.Ecol. Lett. 12, 1029–1039.

Hammer, Ø., Harper, D.A.T., Paul, D.R., 2001. Past: Paleontological Statistics SoftwarePackage for Education and Data Analysis.

Hatten, T.D., Bosque-Pérez, N.A., Labonte, J.R., Guy, S.O., Eigenbrode, S.D., 2007. Effectsof tillage on the activity density and biological diversity of carabid beetles in springand winter crops. Environ. Entomol. 36, 356–368.

Henneron, L., Bernard, L., Hedde, M., Pelosi, C., Villenave, C., Chenu, C., Bertrand, M.,Girardin, C., Blanchart, E., 2014. Fourteen years of evidence for positive effects ofconservation agriculture and organic farming on soil life. Agron. Sustain. Dev. 35,169–181.

Hertzog, L.R., Meyer, S.T., Weisser, W.W., Ebeling, A., 2016. Experimental manipulationof grassland plant diversity induces complex shifts in aboveground arthropod di-versity. PLoS One 11, 1–16.

Hossain, M.M., Begum, M., 2015. Soil weed seed bank: importance and management forsustainable crop production- a review. J. Bangladesh Agric. Univ. 13, 221–228.

Jost, L., 2006. Entropy and diversity. Oikos 113.Kerzicnik, L.M., Peairs, F.B., Cushing, P.E., Draney, M.L., Merrill, S.C., 2013. Spider fauna

of semiarid Colorado agroecosystems: diversity, abundance and effects of crop in-tensification. Environ. Entomol. 42, 131–142.

Kromp, B., 1999. Carabid beetles in sustainable agriculture: a review on pest controlefficacy, cultivation impacts and enhancement. Agic., Ecosyst. Environ. 74, 187–228.

Mansour, F., Bernstein, E., Abo-Moch, F., 1995. The potential of spiders of different taxaand a predacious mite to feed on the carmine spider mite, a laboratory study.Phytoparasitica 23, 217–220.

Menalled, F.D., Smith, R.G., Dauer, J.T., Fox, T.B., 2007. Impact of agricultural man-agement on carabid communities and weed seed predation. Agric. Ecosyst. Environ.118, 49–54.

Moser, T., Rombke, J., Schallnass, H.-J., Van Gestel, C.A.M., 2007. The use of the mul-tivariate Principal Response Curve (PRC) for community level analysis: a case studyon the effects of carbendazim on enchytraeids in Terrestrial Model Ecosystems(TME). Ecotoxicology 16, 573–583.

Mutema, M., Mafongoya, P.L., Nyagumbo, I., Chikukura, L., 2013. Effects of crop residuesand reduced tillage on macrofauna abundance. J. Org. Syst. 8, 5–16.

Nyffeler, M., 1999. Prey selection of spiders in the field. J. Arachnol. 27, 317–324.Plaza, E.M., Kozak, M., Navarrete, L., Gonzalez-Andujar, J.L., 2011. Tillage system did not

affetc weed diversity in a 23-year experiment in Mediterranean dryland. Agric.Ecosyst. Environ. 140, 102–105.

Rahman, L., Chan, K.Y., Heenan, D.P., 2007. Impact of tillage, stubble management andcrop rotation on nematode populations in a long-term field experiment. Soil TillageRes. 41, 11–18.

Randall, J.B., 1982. Prey records of the Green Lynx spider, Peucetia viridans (Hentz)

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

244

(Araneae, Oxyopidae). J. Arachnol. 10, 19–22.Rendon, D., Whitehouse, M.E.A., Hulugalle, N.R., Taylor, P.W., 2015. Influence of crop

management and evironmental factors on wolf spider assemblages (Araneae:Lycosidae) in an Australian cotton cropping system. Environ. Entomol. 44, 174–185.

Rendon, D., Whitehouse, M.E.A., Taylor, P.W., 2016. Consumptive and non-consumptiveeffects of wolf spiders on cotton bollworms. Entologia Experimentalis et Applicata158, 170–183.

Scherr, S.J., McNeely, J.A., 2008. Biodiversity conservation and agricultural sustain-ability: towards a new paradigm of ‘ecoagriculture’ landscapes. Philos. Trans. Biol.Sci. 363, 477–494.

Shannon, C.E., Weaver, W., 1949. The Mathematical Theory of Communication.University of Illinois Press, Urbana.

Shennan, C., 2008. Biotic interactions, ecological knowledge and agriculture. Philos.Trans. Biol. Sci. 363, 717–739.

Siemann, E., 1998. Experimental tests of the effects of plant productivity and plant di-versity on grassland arthropod diversity. Ecology 79, 2057–2070.

Sims, B.G., Thierfelder, C., Kienzle, J., Friedrich, T., Kassam, A., 2012. Development ofthe conservation agriculture equipment industry in Sub-Saharan Africa. Appl. Eng.Agric. 28, 813–823.

Šmilauer, P., Lepš, J., 2014. Multivariate Analysis of Ecological Data Using CANOCO 5.Cambridge University Press.

Soane, B.D., Ball, B.C., Arvidsson, J., Basch, G., Moreno, F., Roger-Estrade, J., 2012. No-till in northern, western and south-western Europe: a review of problems and op-portunities for crop production and the environment. Soil Tillage Res. 118, 66–87.

Stinner, B.R., House, G.J., 1990. Arthropods and other invertebrates in conservation til-lage agriculture. Annu. Rev. Entomol. 35, 299–318.

Suding, K.N., Scott, L., Gough, C.L., Clark, C., Cleland, E.E., Gross, K.L., Milchunas, D.G.,Pennings, S., 2005. Functional- and abundance-based mechanisms explain diversityloss due to N fertilization. Ecology 102, 4387–4392.

Suenaga, H., Hamamura, T., 2015. Effects of manipulated density of the wolf spider,Pardosa astrigera (Araneae: Lycosidae), on pest populations and cabbage yield: a fieldenclosure experiment. Appl. Entomol. Zool. 50, 89–97.

Suter, R.B., Stratton, G.E., Miller, P.R., 2011. Mechanics and energetics of excavation by

burrowing wolf spiders, Geolycosa spp. J. Insect Sci. 11, 1–15.Tabaglio, V., Gavazzi, C., Menta, C., 2009. Physico-chemical indicators and micro-

arthropod communities as influenced by no-till, conventional tillage and nitrogenfertilisation after four years of continuous maize. Soil Tillage Res. 105, 135–142.

ter Braak, C.J.F., Smilauer, P., 2012. Canoco Reference Manual and User’s Guide:Software for Ordination, Version 5.0.

Thierfelder, C., Wall, P.C., 2010. Rotations in conservation agriculture systems of Zambia:effects on soil quality and water relations. Exp. Agric. 46, 309–325.

Thierfelder, C., Matemba-Mutasa, R., Rusinamhodzi, L., 2015. Yield response of maize(Zea mays L.) to conservation agriculture cropping system in Southern Africa. SoilTillage Res. 146, 230–242.

Uetz, G.W., 1999. Guild structure of spiders in major crops. J. Arachnol. 27, 270–280.VSN-International, 2011. GenStat for Window, 10th edition. VSN International, Hemel

Hempstead, UK.Weeks, R.D., Thomas, O.H., 2000. Habitat and season in structuring ground-dwelling

spider (Araneae) communities in a shortgrass steppe ecosystem. Community Ecosyst.Ecol. 29, 1164–1172.

Whitehouse, M.E.A., Wilson, L.J., Fitt, G.P., 2005. A comparison of arthropod commu-nities in transgenic Bt and conventional cotton in Australia. Environ. Entomol. 34,1224–1241.

Whitehouse, M.E.A., Hardwick, S., Scholz, B.C.G., Annells, A., Ward, A., Grundy, P.,Harden, S., 2009. Evidence of a latitudinal gradient in spider diversity in Australiancotton. Austral Ecol. 34, 10–23.

Whitehouse, M.E.A., Wilson, L.J., Davies, A.P., Cross, D., Goldsmith, P., Thompson, A.,Harden, S., Baker, G., 2014. Target and nontarget effects of novel “triple-stacked” Bt-transgenic cotton 1: canopy arthropod communities. Environ. Entomol. 43, 218–241.

Whitmore, C., Slotow, R., Crouch, T.E., Dippenaar-Schoeman, S.A., 2002. Diversity ofspiders (Aaneae) in a savanna reserve, Northern Province, South Africa. J. Arachnol.30, 344–356.

WRB, 1998. World Reference Base on Soils. FAO-ISRIC, Rome, Italy.Zhao, X.M., Qi, J.C., Yan, R.P., 1989. Preliminary report on biological characters of

Pardosa astrigera (Araneae: Lycosidae) and its use in the control of the cotton aphid.Nat. Enemies Insects 3, 110–115.

N. Mashavakure et al. Agriculture, Ecosystems and Environment 272 (2019) 237–245

245