Introduction

Shelterbelts, hedgerows and windbreaks are landmarks of human settlements and agricultural landscapes throughout the world (Baudry et al. 2000). The widening interest in these non-crop habitats arises from their ecological, cultural and economical value in farm forestry. Non-crop habitats act as refuges for species that cannot survive elsewhere in farmlands (Stephens et al. 2006). Many arthropods, both beneficial and harmful, use these habitats (Dix and Leatherman 1988; Majer et al. 2000) for food, oviposition and shelter (Pasek 1988; Thomas et al. 1991, 1992; Pollard and Holland 2006) and as overwintering sites (Dennis et al. 1994; Thomas and Marshall 1999). Within shelterbelts, arthropods could serve as prey for birds (Gámez-Virués et al. 2007), contribute to crop pollination (Pontin et al. 2006), and aid natural pest suppression (Tsitsilas et al. 2006). Although few studies have explored how the botanical composition of woody non-crop habitats such as shelterbelts can affect their arthropod community (e.g., Pollard and Holland 2006; Tsitsilas et al. 2006), extensive work on the effects of other non-crop habitats such as field margins (Olson and Wäckers 2007) suggest enhanced impact of natural enemies within crops (Marino and Landis 1996; Steffan-Dewenter et al. 2001; Bianchi et al. 2006, Fiedler et al. 2008). Key examples of this are the suppression of pest mites in pastures adjacent to shelterbelts by predators, which appeared to be maintained by the groundcover vegetation within shelterbelts (Tsitsilas et al. 2006); and the colonization of vineyards adjacent to a corridor of flowering plant species by generalist predators (Nicholls et al. 2001). In addition, several European landscape-scale studies highlight non-crop habitats as relevant landscape components for a better conservation of biodiversity (e.g., Kruess and Tscharntke 1994; 2000; Thies et al. 2003; Bianchi et al. 2006; Billeter et al. 2008).

Although provision of nectar does not always lead to enhanced parasitism of pests (Heimpel and Jervis 2005), scope to increase the value of shelterbelts for natural enemies by manipulating groundcover vegetation is possible from the success of this approach in several crop systems. For example, an increase in the abundance of the leafroller parasitoid Dolichogenidea tasmanica (Hymenoptera: Braconidae) occurred when groundcover plants flowered in a New Zealand apple orchard (Irvin et al. 2000) and vineyard (Berndt et al. 2002), and an elevated rate of parasitism of Epiphyas postvittana (Lepidoptera: Tortricidae) eggs by Trichogramma carverae (Hymenoptera: Trichogrammatidae) when groundcover plants flowered in an Australian vineyard (Begum et al. 2006). Laboratory experiments indicate how effects of flowering groundcovers measured under field conditions may operate (see Jervis et al. 1996). Longevity of Diadegma insulare (Hymenoptera: Ichneumonidae) and Cotesia marginiventris (Hymenoptera: Braconidae) (Johanowicz and Mitchell 2000) was greater when these parasitoids had access to L. maritima flowers compared with control (water only), longevity of Microctonus hyperodae (Hymenoptera: Braconidae) increased when flowers of Fagopyrum esculentum (Polygonaceae) and Coriandrum sativum (Apiaceae) were available (Vattala et al. 2006), and longevity and fecundity of T. carverae was greater in the presence of L. maritima than other treatments (Begum et al. 2006).

Despite the high level of interest in conservation biological control (DeBach 1964; Ehler 1998) over the last 15 years, little work has focused on habitat manipulation methods to suppress arthropods infesting linear plantations of trees in agroecosystems, ‘farm shelterbelts’, which provide benefits such as reduction of wind erosion and salinity risk, preservation of soil moisture, and protection for crops, pastures and livestock (Baudry et al. 2000). Anoplognathus, Heteronychus arator and Heteronyx elongatus (Coleoptera: Scarabaeidae), Phylacteophaga froggatti (Hymenoptera: Pergidae), Mnesampela privata (Lepidoptera: Geometridae) and Phaulacridium vittatum (Orthoptera: Acrididae) are few of the most common and damaging herbivores of eucalypts in farms and plantations (CSIRO 1996; Loch and Floyd 2001). Although other herbivores occur usually in low numbers their damage could affect commercial farm forestry by reducing the success of establishment of eucalypts (Loch and Floyd 2001). For example, Ardozyga stratifera (Meyrick) (Lepidoptera: Gelechiidae), a leaf-tier, causes malformation of leave shoots by joining them together (Common 1990; Loch and Floyd 2001).

The present study is the first empirical assessment to test amenability of Australian shelterbelt wasps to manipulation via plant community. This study aimed to measure the effects of incorporating Lobularia maritima (L.) Desv. (Brassicaceae) as flowering groundcover in a native-tree shelterbelt on the (i) diversity of parasitoid and phytophagous wasps and (ii) parasitism of A. stratifera larvae that were naturally infesting Eucalyptus blakelyi Maiden (Myrtaceae) trees. L. maritima was selected since it has been found a successful herbaceous resource in similar manipulation experiments that aimed at conservation biological control (e.g., Ambrosino et al. 2006; Begum et al. 2006; Irvin et al. 2006).

Materials and methods

A randomized block experimental design was established in a 3-year-old shelterbelt (20 × 230 m) located in the Charles Sturt University farmland in Orange city, central-west NSW (33°15′S; 149°07′E; 849 m asl). The temperature during the study was 23.9 ± 1.4°C maximum and 8.8 ± 1.5°C minimum. The shelterbelt had five rows of native trees (4 m inter-row spacing) including the following taxa: E. blakelyi (48 trees), E. macrorhyncha F. Muell. ex Benth. (48 trees), E. melliodora A. Cunn. ex Schauer (56 trees), E. pauciflora Sieber ex Spreng. (48 trees), E. viminalis Labill. (48 trees), Acacia dealbata Link (12 trees), A. implexa Benth. (12 trees), A. vestita Ker Gawl. (12 trees) and Casuarina cunninghamiana Miq. (36 trees). Shrubs of Callistemon sieberi DC. (18 trees) and Leptospermum myrtifolium Sieber ex DC. (18 trees) were also present.

Twenty-four E. blakelyi trees (mean height 157.9 ± 65.4 cm) were selected based on their uniform levels (10–20% of leaves attacked, visual estimations) of natural infestation by A. stratifera larvae, which was the only herbivorous arthropod established on the trees. The 24 E. blakelyi trees were used to investigate the diversity of wasps and parasitism of A. stratifera larvae in a randomized block design. Each block comprised two trees, for one of the trees, a groundcover was established by manually removing the existing sward of Phalaris aquatica L. (Poaceae) and planting glasshouse-grown L. maritima (cv. Carpet of Snow) plants to cover 1 m2 of the area around each nominated tree (hereafter, ‘treatment’). The other tree in each block retained the original groundcover of P. aquatica (hereafter, ‘control’). L. maritima plants were replaced, whenever necessary, to provide consistent groundcover of flowering plants.

Sampling of Hymenoptera commenced in September 2005, two weeks after establishing the L. maritima groundcover, and was continued till March 2006 (spring–summer). Wasps were collected fortnightly using a vacuum sampler (Weed Eater®, Model GB1 30v, Poulan Co. Shreveport, Louisiana, USA) by placing the sampler tube to enclose the branches. E. blakelyi foliage and the groundcover within 50 cm radius of the tree trunk were sampled for 30 s per plot. Vacuum samples from each plot were stored in separate zip-lock bags at −18°C for observations later.

All wasps were identified to the family level and assigned to morphospecies. Genus and species were determined, wherever possible, using the limited databases because the Australian arthropod fauna is one of the most poorly known (Austin et al. 2004; Raven and Yeates 2007). Determinations of taxa were made using Stevens et al. (2007), Austin et al. (2005), Naumann (1991), and by consulting appropriate professionals (vide acknowledgement). All morphospecies were assigned to functional groups: ‘parasitoids’ or ‘phytophagous’ (including gall inducers) following Stevens et al. (2007), Austin et al. (2005), La Salle (2004), Naumann (1991) and Gauld (1984). Numbers of morphospecies and individuals of each taxon, and each functional group, were used in calculating the Shannon diversity index, H′ (Magurran 1988) in each treatment, and a t-test (significance P < 0.05) was performed to measure statistical differences (Magurran 1988). Mean abundance of morphospecies was calculated by pooling all sampling dates together as some of the parasitoids and phytophages occurred in very low numbers. Only the mean abundance of those ‘common’ species of wasps (represented by more than five individuals and with an arithmetic mean greater than 1, after pooling all samples) was analysed with one-way ANOVA (GenStat® 10) after square-root transformation (√x + 0.5).

Parasitism of Ardozyga stratifera larvae

Ardozyga stratifera larvae form a shelter in which to live and feed by tying Eucalyptus leaves together. Thirty of these shelters were collected from each of the 24 trees of E. blakelyi in October 2005 and March 2006. A. stratifera larvae were removed from the shelters and individually kept in plastic containers (250 ml) with lids that allowed ventilation and under laboratory conditions (25 ± 3°C, 13 h light period). Wet cotton balls were added to retain moisture within the containers. Containers were checked twice a week to record emergence of parasitoids and fresh field-collected E. blakelyi leaves, without any arthropod infestation, were supplemented periodically. Parasitoids within the containers were counted and stored in 70% ethanol. Treatment effects on the abundance of parasitized and non-parasitized A. stratifera larvae were analysed with a χ2 test (GenStat® 10).

Results

In total, 783 wasps, including both parasitoids and phytophages, were assigned to 50 species and 20 families. Of these wasps, 478 were parasitoids (44 species; 19 families) obtained from the treatment (H′ = 2.78, evenness = 0.734) and 91 were parasitoids (20 species; 12 families) obtained from the control (H′ = 2.43, evenness = 0.810) (Table 1). Shannon diversity index for parasitoids differed significantly between treatment and control (t 1,142 = 2.84, P < 0.01) (Table 1). Braconidae (F 1,11 = 36.2, P < 0.001), Diapriidae (F 1,11 = 7.86, P = 0.017), Encyrtidae (F 1,11 = 42.22, P < 0.001), Eulophidae (F 1,11 = 10.27, P = 0.008), Pteromalidae (F 1,11 = 41.37, P < 0.001) and Scelionidae (F 1,11 = 11.06, P = 0.007) showed the greatest species richness and abundance in the treatment (Table 1). Seven parasitoid taxa (mentioned below) were adequately abundant to be considered ‘common’. Cotesia sp. (Braconidae) (F 1,11 = 22.65, P < 0.001), Pteromalus sp. (Pteromalidae) (F 1,11 = 42.98, P < 0.001), Anagyrus sp. (Encyrtidae) (F 1,11 = 33.44, P < 0.001), Entedoninae sp. (F 1,11 = 14.81, P = 0.003) and Eulophidae sp. 1 (Eulophidae) (F 1,11 = 8.43, P = 0.014) were significantly more abundant in the treatment than in the control; and Cirrospilus sp. (F 1,11 = 3.99, P = 0.071) and Eulophidae sp. 2 (Eulophidae) (F 1,11 = 0.50, P = 0.495) did not show any statistical difference (Fig. 1).

Table 1 Diversity of parasitoid and phytophagous wasps collected from the Lobularia maritima (Lm) treatment and control
Full size table
Fig. 1
figure 1

Mean (±SE) abundance of the common parasitoid taxa: C, Cotesia sp.; P, Pteromalus sp.; A, Anagyrus sp.; E, Entedoninae sp., Ci, Cirrospilus sp., Eu1, Eulophidae sp. 1; and Eu2, Eulophidae sp. 2; and the common phytophagous taxa M, Megastigmus sp.; and O, Ophelimus sp.; in the Lobularia maritima (Lm) treatment and control. Asterisks denote differences between treatments (one-way ANOVA ** P < 0.01, *** P < 0.001)

Full size image

Phytophagous wasps were considerably less diverse than parasitoids: 34 individuals of Ophelimus sp. (Eulophidae) and 21 individuals of Megastigmus sp. (Torymidae) were obtained from the treatment (H′ = 0.41, evenness = 0.598), whereas 134 individuals of Ophelimus sp., 24 individuals of Megastigmus sp. and a singleton of Pergidae were obtained from the control (H′ = 0.28, evenness = 0.258) (Table 1). Phytophagous wasps were more abundant in the control, but the numerical difference between treatment and control was not statistically significant. Megastigmus sp. and Ophelimus sp. were adequately abundant to be considered ‘common’, although their abundance did not differ between treatment and control (Fig. 1).

Parasitism of Ardozyga stratifera larvae

Fifty-five and 12 A. stratifera larvae were obtained from the treatment and the control, respectively, in October 2005. Of these, 11 larvae from the treatment and one from the control were parasitized by Diaulomorpha sp. 1 (Eulophidae). Despite the numerical difference in parasitism of A. stratifera larvae, treatment effect was not statistically significant (χ 21  = 0.91, P = 0.304). In April 2006, 55 and 31 A. stratifera larvae were obtained from the treatment and the control, respectively. Of these, 17 larvae from the treatment were parasitized by Sierola sp. (Bethylidae), Diaulomorpha sp. 2 (Eulophidae) and one undetermined parasitic wasp; and three A. stratifera larvae from the control were parasitized by Glyptapanteles sp. (Braconidae) and Neostomatoceras sp. (Chalcididae). Of these parasitoids from A. stratifera larvae, only Sierola sp. (n = 5 individuals) were represented in the vacuum sample catches. Parasitism of A. stratifera larvae was significantly greater in the treatment than in the control (χ 21  = 5.01, P = 0.025) in April 2006.

Discussion

This study shows that incorporation of L. maritima groundcover around individual E. blakelyi trees had positive effects on the diversity of parasitic wasps, indicating their amenability to manipulation via the plant community.

The positive effect of floral resources on the activity and density of several parasitoids has been demonstrated extensively in laboratory (e.g., Baggen and Gurr 1998; Jacob and Evans 2004; Short and Steinbauer 2004) and field conditions (e.g., Siemann et al. 1998; Bostanian et al. 2004; Lavandero et al. 2005). Most of these investigations have established that the nectar of herbaceous plants (e.g., L. maritima) added value to the agro-ecosystems (e.g., orchards, Bugg and Waddington 1994) through enhanced longevity of parasitoids (Johanowicz and Mitchell 2000; Begum et al. 2006), although greater longevity was not always achieved when tested under laboratory conditions (Sivinski et al. 2006). The positive effects of the floral resources on diversity of parasitoids are consistent with, but not necessarily indicative of, nectar feeding behaviour of the wasps (Heimpel and Jervis 2005). Although floral resources are critical for some parasitoids, other factors, such as moderated microclimate, shelter and supporting alternative host or prey organisms, will need to be factored here as having key effects on the efficacy of parasitoid performance.

Parasitism of A. stratifera larvae was numerically higher in those collected from the treatment than from the control, but was statistically significant only in April 2006. Three of the five parasitoid taxa reared from A. stratifera larvae belonged to Eulophidae (two) and Braconidae (one), which were the families, among six, that showed greater abundance and species richness in the treatment. These results indicate that flowering-plant groundcover plays a role in either attracting or retaining parasitoids (also see Stephens et al. 1998), and possibly affecting longevity and fecundity by making available nectar. A. stratifera larvae were numerically more in the treatment on both sampling occasions (October 2005; April 2006), thus an aggregative response by the parasitoids to A. stratifera density is also possible and could account for differences in parasitism.

The present work establishes a positive relationship between wasps and flowering groundcover at small scale (1 m2), by adding value to the performance of shelterbelts. Effects of floral resources on parasitoids and predators have been investigated at the farm (Fitzgerald and Solomon 2004; Woodcock et al. 2005) and landscape scales (Thies et al. 2003; Bianchi et al. 2005). Presence of L. maritima influenced the abundance and the activity of parasitoids at a small spatial scale trialled in this field-based experiment. Parasitoids such as Hyposter sp. (Ichneumonidae), Macrocentrus sp., Trichogramma sp. (Trichogrammatidae) (Freeman-Long et al. 1998) and Lysiphebus testaceipes (Braconidae) (Fernandes et al. 1997) have shown movements up to 76, 30, 6, and 60 m distances, respectively, into the crop areas after visiting a floral source. Diadegma semiclausum (Hymenoptera: Ichneumonidae) reached distances of over 80 m in four days after feeding on a plot with flowers, but parasitism rates on Plutella xylostella (Lepidoptera: Plutellidae) were lower in a plot without flowers separated by a distance of 60 m (Lavandero et al. 2005). These findings establish that the distances, which the wasps travel and the scale at which they parasitize hosts are not always related (Lavandero et al. 2005). Moreover, effects of groundcover vegetation on parasitoids at small and large spatial scales are not always consistent. For example, the addition of L. maritima to apple orchards infested with the leafroller E. postvittana did not influence parasitism by Dolichogenidea spp. (Hymenoptera: Braconidae) but the characteristics of the surrounding landscape affected the rates of parasitism (Bell et al. 2006). Therefore, evaluation of potential food plants for parasitoids and their amenability to manipulation in shelterbelts and similar non-crop habitats will need to consider the scale effects and also use larger and more widely spaced plots than the present study.

These findings are in broad agreement with those reported by Tsitsilas et al. (2006) who have investigated the impact of arthropod predators in shelterbelts and adjacent cropland on the pest arthropods of pastures in Victoria Australia. The present study provides evidence that the enhanced abundance of parasitoids impacts on the pest arthropods of shelterbelts, which are attracting greater attention as farm forestry. However, the effects of shelterbelt botanical composition, including not only groundcover vegetation, on diversity and activity of wasps and other arthropods needs to be investigated further, particularly in relation to identifying plant species that provide benefit to natural enemies more selectively and to ‘scaling-up’. The present work using the food plant L. maritima as a flowering groundcover for the conservation of the existing biological control in a native vegetated shelterbelt indicates that parasitoids are amenable to manipulation via addition of flowering groundcover.