Prey abundance and intraguild predation between Adalia bipunctata ( Coleoptera : Coccinellidae ) and Macrolophus pygmaeus ( Hemiptera : Miridae )

Macrolophus pygmaeus Rambur (Hemiptera: Miridae) and Adalia bipunctata (L.) (Coleoptera: Coccinellidae) are two predatory insects commonly used as biological control agents. In order to determine the incidence with which both species attack and eat each other [Intraguild predation (IGP)], the direction and symmetry of the interaction between A. bipunctata and M. pygmaeus were characterized. In addition, whether the intensity of IGP between these two predators increased when the number of extraguild prey Acyrthosiphon pisum (Harris) (Hemiptera: Aphididae) decreased, was also determined. Unidirectional sensu stricto IGP between A. bipunctata and M. pygmaeus was recorded: when IGP occurred, A. bipunctata was always the IG predator that killed and ate M. pygmaeus, the IG prey. However, the intensity of IGP was a function of the abundance of the extraguild prey, A. pisum, since IGP increased when the number of extraguild prey decreased. These results are discussed in terms of theoretical models that predict stability and the outcome of using natural enemies to control pests.


INTRODUCTION
Intraguild predation (IGP) is a type of omnivory, which is defined as the "killing and eating species that use similar, often limiting, resources and that are potential competitors" (Polis et al., 1989;Polis & Holt, 1992).In IGP there are three organisms that interact: a natural enemy (the intraguild (IG) predator) that negatively affects a second natural enemy (the IG prey) by eating it and the shared (extraguild) prey (Rosenheim et al., 1995;Lucas, 2005).
IGP appears to be widespread in communities of biological control agents and may affect the distribution and abundance of IG competitors and modulate the dynamics of a pest population (Polis et al., 1989;Polis & Holt, 1992;Rosenheim et al., 1993Rosenheim et al., , 1995;;Lucas et al., 1998;Rosenheim, 1998;Colfer & Rosenheim, 2001).IGP is characterized by evaluating its intensity (the level of IGP in a specific combination of predators), its direction (identity of the predator and prey) and its symmetry (in asymmetrical IGP one species consistently preys upon the other, while in symmetrical IGP both species prey equally upon each other) (Lucas et al., 1998;Lucas, 2005).
If an IG predator encounters both IG prey and extraguild prey in the same patch, it is possible that the intensity of IGP depends on the abundance of extraguild prey.A model has been proposed in which the IG predator can switch from feeding exclusively on the most profitable prey (usually the extraguild prey) to feeding on the less profitable prey (the IG prey) when the encounter rate with the profitable prey falls below a certain threshold value (Křivan & Diehl, 2005).This model predicts that the coexistence of IG predators and IG prey is increased, even if the relative abundances of the organisms involved fluctuate over time (Law & Blackford, 1992;Křivan & Diehl, 2005).In biological control, if the IG predator is the superior competitor, theory predicts that the IG prey will always be excluded since the IG predator reduces the pest densities to lower densities than the IG prey, without IGP having a negative effect on biological control (Janssen et al., 2006).If this is correct then there is no advantage in using two species of natural enemies that are IGPs.

Extraguild prey consumption
We found that the number of aphids consumed by A. bipunctata significantly increased with the increase in the number of prey (F 2,72 = 117, P < 0.001; Fig. 1).
We also found a significant difference in the number of aphids consumed by M. pygmaeus in the presence of A. bipunctata at the three different aphid densities (F 2,72 = 4.5, P < 0.05), with the results for T10 and T20 treatments differing from T5 (Fig. 1).
The total number of aphids consumed by the two predators in the T5, T10 and T20 treatments was, respectively, 4.93, 9.58 and 18.1, which are very close to the total number of prey items in each of these treatments.

DISCUSSION AND CONCLUSIONS
The organisms that are involved in IGPs do not have well defined roles because several factors, such as their relative size, mobility, density, feeding specificity and morphology, influence whether they are predators or prey (Polis et al., 1989;Lucas et al., 1998;Burgio et al., 2005;Lucas, 2005;Fréchette et al., 2007).Based on the definition proposed by Lucas (2005), we demonstrated that there is unidirectional sensu stricto IGP between A. bipunctata and M. pygmaeus: when IGP occurs, A. bipunctata is always the IG predator that kills and eats M. pygmaeus, the IG prey.Similar to our results, Lucas & Alomar (2000, 2001) show that M. pygmaeus is frequently an IGP with Dicyphus tamaninii Wagner (Heteroptera: Miridae) as the IG prey.
The small volume of the arena used in the present experiment may have altered the behaviour and interaction between the species.It is known that the size of the arena has a clear effect on the intensity of IGP, with smaller and simpler arenas generating higher levels of IGP by increasing predator encounter rates and/ 80-90% rh and a 18L: 6D photoperiod, feeding on broad bean plants (Vicia faba c.v. Agua dulce).
Macrolophus pygmaeus individuals came from a colony kept under laboratory conditions on potted tomato and broad bean plants, at room temperature and under a 18L: 6D photoperiod.Sterilized Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) eggs, were used as a supplemental diet.Adalia bipunctata was laboratory reared on A. pisum.The E. kuehniella eggs, M. pygmaeus and A. bipunctata were purchased from Koppert Biological Systems.

Experimental procedures
All the experiments were carried out in plastic cylinders (diameter: 5.5 cm, height: 7.5 cm, volume: 150 ml) with a mesh covered ventilation hole (4.5 cm in diameter) in the screw-top.Each cylinder contained a fresh cut broad bean leaf with the stalk inside an Eppendorf tube filled with water and sealed with parafilm to prevent desiccation.Cohorts of 2 nd instar A. pisum were used in the experiments.Four different experimental treatments (T), always containing one M. pygmaeus adult female plus one A. bipunctata 4 th instar larva, were set up as follows: T0: without A. pisum nymphs (N = 28 replicates); T5: with five A. pisum nymphs (N = 29); T10: with ten A. pisum nymphs (N = 26); T20: with twenty A. pisum nymphs (N = 20).For each experimental treatment, M. pygmaeus, A. bipunctata and A. pisum (if any) were gently transferred using a paintbrush into the cylinders with a broad bean leaf.The cylinders were kept in an environmental chamber at 22°C, 80-90% rh and a 18L: 6D photoperiod.After 24 h the cylinders were inspected in order to determine if IGP occurred; the number of aphids consumed by A. bipunctata and by M. pygmaeus was also recorded (in the latter case aphid skins remained after predator feeding).

Data analyses
The numbers of aphids consumed by A. bipunctata and M. pygmaeus at the three prey densities (5, 10 and 20 aphids) were separately analyzed using a one-way ANOVA model with treatment as the main factor.The number of replicates in which IGP occurred was analyzed using a Chi-square contingency table to determine if the intensity of IGP depended on the number of aphids.The analyses were done using R.3.1.1 software (R Core Team, 2014).or by reducing refuge patches (Chang, 1996;Hindayana et al., 2001;Janssen et al., 2006;Ingels & De Clercq, 2011).However, the shape and dimensions of the arena allowed the mirid to escape predation by performing short flights, which reveals their potential effect on the predatory interactions between species.The recorded unidirectional IGP is possibly due to A. bipunctata being larger and more voracious than M. pygmaeus.
It is known that some hemipteran and syrphid species are able to prey on M. pygmaeus in the absence and in some cases also in the presence of the extraguild prey (Jakobsen et al., 2004;Fréchette et al., 2007;Perdikis et al., 2014).As proposed by Lucas et al. (1998), it is possible that the extraguild prey density may affect the outcome of IGP.In the present experiment we recorded no IGP between A. bipunctata and M. pygmaeus when extraguild prey was abundant.The IG predator switched from feeding exclusively on aphids to feeding on IG prey when the encounter rate with the extraguild prey was low.These results fit with the assumptions of a model proposed by Křivan & Diehl (2005) that predict the coexistence of IG predators and prey.Further, the intensity of IGP is a function of the extraguild prey abundance.In the theoretical scenario proposed by Lucas et al. (1998), this kind of relationship between extraguild prey density and IGP is expected to occur for two predators that forage for prey randomly, in which their searching behaviour does not influence their chance of meeting.
Based on our previous experiments (data unpubl.)we know that under the same experimental conditions, when ten or twenty 2 nd instar nymphs of A. pisum are offered to M. pygmaeus alone, the mirid bug consumed about five and seven prey items, respectively.In the present study, M. pygmaeus consumed about one aphid in the T10 and T20 experimental treatments.Moreover, we observed that A. bipunctata consumes the skins remaining after M. pygmaeus feeds on an aphid.Thus it is possible that the number of aphids consumed by M. pygmaeus was underestimated in our experiment.Nevertheless, in this system A. bipunctata could be considered the superior competitor.
Even if M. pygmaeus and A. bipunctata are not used together in aphid biological control, nowadays the propensity to include predatory mirids in biological control programs against aphids is more likely, especially in solanaceous plant crops (Pérez-Hedo & Urbaneja, 2014).Combining both predators may be the better strategy for crop protection because IGP was only substantial when the shared resource (aphids) was limited and because they differ in some important features.Coccinellids feed primarily on aphids and only opportunistically upon other prey (Hodek & Honěk, 1996).As a consequence, the release of the larvae usually does not result in the establishment of self-perpetuating populations (Powell & Pell, 2007) since coccinellids lay eggs in patches with aphid colonies at an early stage of development, and their developmental time is similar to the average duration of an aphid colony (Dixon & Hemptinne, 2003).Mirids are broad generalists and polyvoltine species and their density is not only influenced by the density of a specific prey (Harmon & Andow, 2004).For all these reasons coccinellids and mirids can exert a complementary action in biological control.

Fig. 1 .
Fig. 1.Number of aphids (mean ± standard errors) consumed by A. bipunctata (dark grey) and by M. pygmaeus (light grey) in each of the three treatments.Treatments: T5: one adult female of M. pygmaeus (Mp) plus one 4 th instar larva of A. bipunctata (Ab) plus five nymphs of A. pisum (Ap) (N = 29); T10: one Mp plus one Ab plus ten Ap (N = 26); T20: one Mp plus one Ab plus twenty Ap (N = 20).Total predation is the total number of aphids consumed by both predators.

Fig. 2 .
Fig. 2. Percentage of the replicates in each treatment in which IGP of M. pygmaeus by A. bipunctata was recorded.Treatments: T0, one adult female of M. pygmaeus (Mp) plus one 4 th instar larva of A. bipunctata (Ab) and no nymphs of A. pisum (Ap) (N = 28); T5: one Mp plus one Ab plus five nymphs of Ap (N = 29); T10: one Mp plus one Ab plus ten Ap (N = 26); T20: one Mp plus one Ab plus twenty Ap (N = 20).