Introduction
In European apple orchards, the rosy apple aphid Dysaphis plantaginea (Passerini), the green apple aphid Aphis pomi (de Geer) and the woolly apple aphid Eriosoma lanigerum (Hausmann) are considered to be the three main aphid pests of economic importance (Marc & Canard, Reference Marc and Canard1997). In particular, D. plantaginea can be the most detrimental, causing yield reductions from 30% (Deberardinis et al., Reference Deberardinis, Baronio and Baumgartner1994; Blommers et al., Reference Blommers, Helsen and Vaal2004) to 80% (Qubbaj et al., Reference Qubbaj, Reineke and Zebitz2005). Moreover, several other aphid species attack the apple trees, such as Aphis spiraecola (Pagenstecher), Rhopalosiphum insertum (Walker), Myzus varians (Davidson) and Myzus ornatus (Laing). Only A. pomi has the apple tree Malus domestica (Borkh) (Rosales: Rosaceae) as a unique host, but E. lanigerum and A. spiraecola are also permanently present on apple trees because they are anholocyclic in Europe (no sexual reproduction). The other species include the apple tree in their life cycle as a primary host in spring. Their alate migrants lay eggs in autumn, close to apple shoots, and start to hatch in late March. The migration of D. plantaginea winged adults to their secondary host, the plantain herb Plantago spp. (Lamiales: Plantaginaceae), begins in May and lasts until the end of June (Bonnemaison, Reference Bonnemaison1959; Carroll & Hoyt, Reference Carroll and Hoyt1984). Due to the exponential growth rate of these aphids, and their high damage risk, insecticide control is usually initiated at very low population levels. For example, insecticide treatment is recommended when a single D. plantaginea fundatrix is observed on a sample of 100 shoots (Minarro et al., Reference Minarro, Hemptinne and Dapena2005).
The need to control these aphids by non-chemical methods has stimulated research into the potential of natural enemies to suppress these pests in apple orchards (Carroll & Hoyt, Reference Carroll and Hoyt1984; Mueller et al., Reference Mueller, Blommers and Mols1988; Brown, Reference Brown2004; Minarro et al., Reference Minarro, Hemptinne and Dapena2005; Fréchette et al., Reference Fréchette, Cormier, Chouinard, Vanoosthuyse and Lucas2008; Dib et al., Reference Dib, Simon, Sauphanor and Capowiez2010). Two phases of control, involving different groups of predators, can be distinguished. Early in the season, before the aphid colonies build-up, biological control of aphids requires resident and actively foraging predators that are efficient at finding aphids even at very low density. These predators can have low predation rates because of the large impact that removal of small numbers of individuals can have on subsequent prey densities (Murdoch et al., Reference Murdoch, Chesson and Chesson1985; Chang & Kareiva, Reference Chang, Kareiva, Hawkins and Cornell1999). At that time, the main natural enemies are usually generalist predators because they are able to remain present in absence of a target pest, sustained on alternative prey (Symondson et al., Reference Symondson, Sunderland and Greenstone2002). After colonies have established and the exponential growth phase has begun, biological control of aphids requires more specialised natural enemies that have a high predation rate and/or growth rate, and are actively searching for colonies (Murdoch et al., Reference Murdoch, Chesson and Chesson1985).
Among generalist predators of aphids, spiders possess characteristics that make them well-adapted for control of aphids early in the year. They are present on the canopy at the right time (Pekar & Kocourek, Reference Pekar and Kocourek2004; Simon et al., Reference Simon, Bouvier, Debras and Sauphanor2009), and some species actively hunt in winter (Marc & Canard, Reference Marc and Canard1997; Miliczky et al., Reference Miliczky, Horton and Calkins2008; Korenko & Pekar, Reference Korenko and Pekar2010). In many instances, spiders kill more prey than they consume (Riechert & Lockley, Reference Riechert and Lockley1984; Greenstone, Reference Greenstone1999). Wyss et al. (Reference Wyss, Niggli and Nentwig1995), looking at aphid remains in spider webs, found that web-building spiders contribute to the biological control of the D. plantaginea alate migrants in autumn. However, other studies, looking at communities of aphid predators (Brown, Reference Brown2004; Minarro et al., Reference Minarro, Hemptinne and Dapena2005; Dib et al., Reference Dib, Simon, Sauphanor and Capowiez2010), concluded that spiders should not be considered as biocontrol agents of aphids because of their low abundance and predation rates compared to other groups, such as earwigs, ladybirds, syrphid and cecidomyiid flies, lacewings and true bugs. As these studies were looking at predation on well-established aphid colonies, their conclusions neglected the potential impact of spiders during the early phase of colony establishment.
Here, we hypothesised that spiders are significant predators that contribute to the early-season control of aphid pests in apple orchards. Two supplementary hypotheses were tested in support of this primary question: (i) early-season spiders eat aphid fundatrices, (ii) later appearance of colonies is negatively correlated with spider abundance early in the season. Other more specialised natural enemies (e.g. parasitoids) arrive when colonies are already established and growing exponentially, but any delay to the onset of this growth phase will improve the ability of these specialists to limit aphid population growth (Murdoch et al., Reference Murdoch, Chesson and Chesson1985; Chiverton, Reference Chiverton1986). To address this question of early season predation, the numbers of aphids and canopy-dwelling spiders were followed from mid-March until the end of April in three organic orchards. Predation on aphids by spiders was studied by gut content analyses with diagnostic polymerase chain reaction (PCR) using aphid-specific primers. Gut content analysis of predation by spiders on aphids has been widely used in cereal crops (see references in Harwood & Obrycki, Reference Harwood and Obrycki2005; Kuusk et al., Reference Kuusk, Cassel-Lundhagen, Kvarnheden and Ekbom2008). This has been shown to be an effective method for analysing predation by fluid-feeders in the field without disrupting the system under study (Greenstone, Reference Greenstone1999; Symondson, Reference Symondson2002; King et al., Reference King, Read, Traugott and Symondson2008).
Materials and methods
Evaluation of the aphid specific primers
The aphid primers (Aphid F/Aphid R), originally designed by Chen et al. (Reference Chen, Giles, Payton and Greenstone2000) for detection of predation on cereal aphids, were tested for their ability to amplify DNA from the apple aphids D. plantaginea, A. pomi, A. spiraecola and E. lanigerum and the stonefruit aphids Myzus cerasi (Fabricius), Myzus persicae (Sulzer) and M. varians all collected locally. PCR was carried out in 10 μl volume reactions using 1 μl DNA template and the PCR protocol was adapted from Chen et al. (Reference Chen, Giles, Payton and Greenstone2000). PCR reactions contained 10 mM Tris-HCl, pH 9.0, 1.5 mM MgCl2, 1.0 μM of each primer, 50 mM KCl, 0.1 mM of each dNTP, 0.05 U μl−1 of Taq DNA polymerase (Promega) and were performed in a PTC-200 thermocycler (MJ Research, Watertown, MA, USA). After an initial denaturing step of 3 min at 94 °C, a cycle of 94 °C for 30 s, 52 °C for 30 s and 72 °C for 1 min, was repeated 34 times, and then a last extension period at 72 °C for 2 min was performed. PCR products were separated by electrophoresis in a 1.5% agarose gel with TBE buffer and visualised by ethidium bromide staining and photographed under UV light.
Spider collection and aphid monitoring
The three organic apple orchards were situated within a 5-km2 area near Avignon in south-eastern France. In early March 2010, shoots, on which at least one aphid egg was observed, were marked with coloured plastic strips. For statistical independence, each marked shoot was on a different tree. Respectively, 17, 14 and 30 shoots were marked in orchards 1, 2 and 3. To collect spiders within all canopy habitats, two cardboard traps (10-cm wide bands, secured with elastic bands) that mimic tree bark were installed on each sampled tree, one wrapped around the branch next to the marked shoot and the other around the main trunk (1 m above the ground). This trapping method is considered the best technique to sample spiders that use tree bark as a shelter (Mizell & Schiffhauer, Reference Mizell and Schiffhauer1987), as well as web-building spiders when the foliage is absent (Marc & Canard, Reference Marc and Canard1997). The shoot was considered the sampling unit for aphid numbers and the two cardboard traps the sampling unit for the spider community. The marked shoots were observed weekly for fundatrices and the presence of colonies on six occasions in early spring (22 and 31 March; 7, 14, 22 and 27 April). On 6 May, trees corresponding to the marked shoots were carefully examined, and the number of colonies and aphids per colonies were assessed. As D. plantaginea is easy to recognize by its grey colouration, we distinguished the D. plantaginea colonies from the other green aphid colonies (A. pomi, R. insertum, A. spiraecola, M. varians, M. ornatus) which cannot be separated to species in the field. Spiders were collected at the same time as the aphid monitoring except on the last date (27 April) for all orchards and on 22 March in orchard 2. In order to avoid local population depletion in orchard 3, where trees were young, spiders were retrieved in trees marked even (15 trees) one week and in trees marked odd (15 trees) the following week. At each sampling time, the cardboard traps were unwrapped from the tree and allowed to fall into a 4 cm × 25 cm tray. Predators were shaken out and the same trap secured again in the tree. In addition to spiders, small numbers of overwintering tenebrionids, ladybirds and earwigs were found, but as they are inactive at this time, they were not sampled or tested for predation on aphids. Spiders were collected individually into Eppendorf tubes then stored in a cool box (4 °C) for a maximum of 3 h and frozen at –20 °C in the laboratory. They were identified before DNA extraction. Identification was to genus level only for species for which juveniles cannot be determined to species. The DNA was extracted with the Qiagen DNeasy® Blood & Tissue extraction kit, resuspended in 200 μl of manufacturer's elution buffer and stored at −20 °C. Finally, extraction success was verified by amplification of the spider DNA with the universal arthropod primers LCO1490 and HCO2198 (Folmer et al., Reference Folmer, Black, Hoeh, Lutz and Vrijenhoek1994), and any possibility of cross-over contamination within batches of samples was excluded by the use of negative controls (water) (King et al., Reference King, Read, Traugott and Symondson2008).
Data analysis
Factors influencing aphid predation by spiders
The effects of orchard, spider identity (species or genus level), canopy habitat (branch or trunk), aphid population (number of aphids per shoot) and date were tested against the probability of detecting aphid DNA in the spider guts, using a general linear model with binomial family errors and logit link function (R Development Core Team, 2011).
Effect of spiders on the emergence of aphid colonies
The binary variable “presence or absence of aphid colonies on the shoot on 22 April” was set up from the aphid records. The abundances of spiders collected from 22 March to 22 April in the corresponding tree were summed. And so, the number of aphid-positive spiders was calculated as the sum of all positive spiders for aphid predation collected from 22 March to 22 April in the corresponding tree. The effects of orchard, spider abundance and number of aphid-positive spiders on the presence/absence of aphid colonies were tested using a generalised linear model with binomial family errors and logit link function.
Results
Amplification of orchard aphid DNA
The general aphid primers only failed to amplify the DNA from E. lanigerum, which was not observed in the orchards.
Aphid dynamics
The mean initial number of eggs per shoot was 1.3 ± 0.1 (mean ± SE), 1.0 ± 0.0 and 1.3 ± 0.2 in orchards 1, 2 and 3, respectively, and was not different between orchards (Kruskal-Wallis χ2 = 3.639, df = 2, P = 0.162). The first fundatrices were recorded on 22 March and the first D. plantaginea colony was observed on 22 April, two weeks after the first appearance of other aphid colonies (7 April). Dysaphis plantaginea colonies represented 23.5% of the total number of colonies on 27 April (ten colonies) and 71% of the total number of colonies on 6 May (27 colonies). At the end of shoot monitoring (27 April), 23.6% of the marked shoots (i.e. for which at least one aphid egg was initially observed) were infested by an aphid colony in orchard 1, 13.3% in orchard 3 and none in orchard 2. On 27 April, the aphid population was at the beginning of its exponential growth phase (fig. 1).
Numbers of spiders (bars, Y1-axis), aphid colonies (black line, Y1-axis) and aphids (dashed line, Y2-axis) in three apple orchards (a) orchard 1, (b) orchard 2 and (c) orchard 3. Spiders detected positive by the aphid-specific diagnostic PCR are represented by the shading on the bars.
Spider communities and other predators
In total, 149 spiders were collected in the three orchards (72, 20 and 57 in orchards 1, 2 and 3, respectively) (table 1). The mean numbers of spiders collected per tree and sampling occasion were, respectively, 0.85 ± 0.09, 0.34 ± 0.09 and 0.71 ± 0.22 (mean ± SE) in orchards 1, 2 and 3 and were significantly different between orchards (Kruskal-Wallis χ2 = 6.379, df = 2, P = 0.041). The spider communities differed between orchards. Orchard 1 and orchard 3 displayed species-rich communities; orchard 3 was dominated by Philodromus spp.; and orchard 2, particularly, lacked web-building and sit-and-wait spiders (table 1).
Composition of the spider assemblage and their predation on aphids detected by the aphid-specific diagnostic PCR in three orchards.
Predation by spiders on aphids
Spiders were found positive for aphid consumption from 22 March, which was two weeks before the first observation of colonies (fig. 1). In consequence, from 22 March to 7 April, it is likely that spiders attacked colony fundatrices. The probability of detecting aphid DNA in the spider guts increased with the number of aphids per shoot (F = 7.52, df = 1, P = 0.007) and depended on the date (F = 4.00, df = 4, P = 0.004) but was not explained by spider identity (F = 1.37, df = 20, P = 0.15), orchard (F = 2.40, df = 2, P = 0.09) and canopy habitat, i.e. branch or trunk, (F = 0.16, df = 1, P = 0.69). Temperature thresholds for most other potential natural enemies of aphids make it unlikely that these were active at this time (none were observed).
Effect of spiders on the presence of colonies
The presence of aphid colonies on 22 April was related to spider abundance (F = 6.035, df = 1, P = 0.017), number of aphid-positive spiders (F = 10.023, df = 1, P = 0.003) and orchard (F = 3.167, df = 2, P = 0.049) (fig. 1). The relationship was negative (parameter estimation = –1.134) for spider abundance only, showing that shoots from trees with more spiders were less likely to harbour an aphid colony. The presence of aphid colonies was positively related (parameter estimation = 2.321) to the number of aphid-positive spiders, suggesting density dependent predation on aphids.
Discussion
The impact of the early-season predation by spiders on aphid populations was studied by analysing early-season aphid population dynamics, spider abundances and predation rates on aphids. Gut content analyses revealed predation events on aphid fundatrices. Spider identity, and their location within the canopy, had no significant effect on predation rates, although numbers were low, and therefore this result should be treated with caution. Aphid abundance had a positive effect on predation rates. Moreover, the presence of an aphid colony on a shoot was negatively correlated with early-season spider abundance in the corresponding tree.
Previous studies hypothesized that spiders could play an important role in the control of pests during winter (Marc & Canard, Reference Marc and Canard1997; Marc et al., Reference Marc, Canard and Ysnel1999; Korenko & Pekar, Reference Korenko and Pekar2010) because some spiders are able to remain active in winter and hunt for prey at low temperatures (Korenko et al., Reference Korenko, Pekar and Honek2010). Indeed, our results clearly showed that spiders prey upon colony fundatrices. They also suggested that spiders were mobile within the canopy; as, within this limited sample of spiders, those collected on the trunk were equally positive for predation on aphids as spiders collected on the shoots where the eggs/fundatrices were present. Moreover, other potential predators of aphids were unlikely to have been active during March and early April, and activity was not observed. The temperature thresholds for development of D. plantaginea and A. pomi are 4.5 °C and 5.9 °C, respectively (Graf et al., Reference Graf, Baumgärtner and Delucchi1985), which means that predators must be active around this temperature threshold to prey upon fundatrices. Korenko et al. (Reference Korenko, Pekar and Honek2010) showed that Philodromus spp. and Anyphaenea accuentata (Walckenaer), which were present in their study orchard, had lower temperature thresholds of predatory activity than the developmental thresholds of D. plantaginea and A. pomi (1.2 °C and 3.5 °C, respectively). In consequence, these winter-active spiders are above their temperature threshold of activity when the first D. plantaginea and A. pomi eggs hatch. By contrast, temperature thresholds for other predators on apple trees are higher than those for the aphids, with 6 °C for the earwig Forficula auricularia (Helsen et al., Reference Helsen, Vaal and Blommers1998) and the hoverfly Episyrphus balteatus (Hart et al., Reference Hart, Bale and Fenlon1997), 9.0 °C for the ladybird Adalia bipunctata (Obrycki & Tauber, Reference Obrycki and Tauber1981) and 10.5 °C for the cecidomyiid fly Aphidoletes aphidimyza (Morse & Croft, Reference Morse and Croft1987). These predators were indeed reported to prey upon aphids later in the season when the first colonies are already established (Dib et al., Reference Dib, Simon, Sauphanor and Capowiez2010). Thus, temporal niche partitioning (Finke & Snyder, Reference Finke and Snyder2010) for aphids as a prey resource results from this difference in phenology among predators, reducing predator interference and limiting temporal refugia for pests (Symondson et al., Reference Symondson, Sunderland and Greenstone2002). Nevertheless, spiders are known to be intraguild predators (Hodge, Reference Hodge1999; Wise, Reference Wise2006), and their impact on other natural enemies should be studied before assuming niche complementarity between predators.
There are indicators suggesting that spiders may have been efficient at controlling aphid numbers in the present study. First, the time lag between the presence of the first fundatrices and their predation by predators is a critical factor for the appearance of colonies and their subsequent damage (Wyss et al., Reference Wyss, Villiger, Hemptinne and Muller-Scharer1999; Brown, Reference Brown2011). In particular, Brown (Reference Brown2011) concluded that limitation of D. plantaginea colony development is more efficient when predation occurs within the first week of colony establishment. Here, spiders tested positive for predation on aphids as soon as the first eggs hatched and were, thus, likely to have had a considerable impact on aphid population dynamics, for example delaying the aphid exponential growth phase (Birkhofer et al., Reference Birkhofer, Gavish-Regev, Endlweber, Lubin, von Berg, Wise and Scheu2008). Second, control of colony development is improved by high initial predator abundance (Wyss et al., Reference Wyss, Villiger, Hemptinne and Muller-Scharer1999). Here, the presence of a colony was indeed negatively correlated to the initial spider abundance on a tree. Finally, although prey preferences (Agustí et al., Reference Agustí, Shayler, Harwood, Vaughan, Sunderland and Symondson2003) and functional specialisation (Marc & Canard, Reference Marc and Canard1997) are common in spiders, this may not be the case in winter when prey are scarce (Marc et al., Reference Marc, Canard and Ysnel1999). Spider identity was not significant in explaining predation on aphids, but more work would be needed with larger sample sizes to confirm this. Results so far indicate that no specific taxa would need to be particularly encouraged for early-season control of aphids.
To conclude, our results support the hypothesis that spiders are important for the early-season control of aphids. Furthermore, the impact of spiders on aphid populations delays the exponential growth phase and, therefore, increases the effectiveness of other predators (and parasitoids) that are active later in the season. The high potential of spiders as intraguild predators, and positive and negative effects of spiders within a predator assemblage (Hodge, Reference Hodge1999), needs further study if their role in aphid suppression within conservation biological control programmes is to be fully understood.
Acknowledgements
This work was funded by the Centre Technique Interprofessionnel des Fruits et Légumes. Catherine Boreau de Roincé's PhD was granted by a CIFRE from the Association Nationale de la Recherche et de la Technologie. We are thankful to Laetitia Durlin and Jean-Michel Ricard for their help with field work, and to the farmers, Jean-François Toubon and Jean-Charles Bouvier for coordination.
