Chemicals are the major means of control used against slugs, which are serious pests of various crops. To increase the sustainability of farming practices, alternatives that do not harm nontarget organisms are necessary. One area of investigation focuses on the capacities of prey to perceive their predators, which enables them to display antipredator behaviours. This study presents initial evidence of the potential effectiveness of using chemical cues from a predatory ground beetle to protect young oilseed rape shoots against a worldwide pest, the slug Deroceras reticulatum (Müller, 1774) (Stylommatophora: Agriolimacidae). A two-choice assay was used to test whether chemical cues from Carabus nemoralis Müller, 1764 (Coleoptera: Carabidae) could impede the foraging of D. reticulatum on young oilseed rape shoots. Significantly fewer cotyledons were consumed when chemical cues from the ground beetle were present compared with the control area, where slugs were mainly found. Chemical cues from predatory ground beetles appear to be a promising solution for protecting at-risk crops from the depredations of pest slugs.
Many studies have explored the use of natural enemies for managing slug populations (Barker, 2004). Among these predators, carabid (or ground) beetles occupy a prominent place (Kromp, 1999; Symondson, Sunderland & Greenstone, 2002), because they are naturally occurring in most agroecosystems and display useful features of generalist predators (Ekschmitt & Wolters, 1997; Symondson et al., 2002; Symondson, 2004). Most of this research has examined direct effects of predation by carabids, i.e. regulation of the slug population through direct consumption (Symondson, 1993; Bohan et al., 2000; Symondson et al., 2006). However, predation may also have indirect, nonlethal effects on prey species, which can substantially affect their fitness (Brown, Laundré & Gurung, 1999; Preisser, Bolnick & Benard, 2005; Peckarsky, 2007) by altering their morphology, life-history traits and behaviours (Lima & Dill, 1990; Werner & Peacor, 2003). In other words, prey species shift their resource allocation in order to mitigate risks posed by predation attempts. In particular, avoiding areas visited by predators is a frequent means of reducing predation risks. This can be done through the use of indirect cues by prey, such as semiochemicals (Apfelbach et al., 2005). In this regard, Armsworth et al. (2005) have demonstrated that the grey field slug Deroceras reticulatum (Müller, 1774) detects and avoids areas previously explored by the generalist carabid beetle Pterostichus melanarius (Illiger, 1798), indicating that slugs adopt antipredator behaviour in the presence of kairomones left by the carabid. In addition to generalist carabids, there are also specialist malacophagous ground beetles. Laboratory and field studies, as well as specialized morphological and behavioural features, have stressed their superiority in hunting and eating slugs, compared with small carabids. They are thus expected to exert stronger selection pressure on these prey, making avoidance of these predators even more critical. Carabus nemoralis Müller, 1764 is such a predator: gut contents reveal slugs remain in C. nemoralis (Tod, 1973; Hatteland et al., 2011, 2013), while Digweed (1994) demonstrated that this ground beetle orientates to D. reticulatum mucus and, in a study by Ayre (1995), C. nemoralis was one of the most efficient of specialist predators in killing D. reticulatum as well as Arion lusitanicus (see Hatteland, 2010). In a recent study, Bursztyka et al. (2013) showed that D. reticulatum actively avoids potential refuges that have been previously sprayed with cuticular extract from C. nemoralis. This modification of such important self-maintenance behaviour for slug survival (Rollo & Wellington, 1977; Grewal et al., 2001) has been observed in juvenile D. reticulatum deprived of any previous exposure to predators of any kind. This suggests that C. nemoralis cuticular extracts may contain semiochemicals, which play a role in the avoidance behaviour. Predatory cues are thus able to alter important behavioural processes such as sheltering.
This study investigated whether chemical cues from C. nemoralis could protect oilseed rape (Brassica napus) seedlings from D. reticulatum, the most widespread and abundant pest slug in Western Europe (Moens & Glen, 2002), which can be exceedingly damaging during the early growth stages of this crop. This vulnerability has been exacerbated since the 1980s with the widespread planting of a ‘double-low’ variety for human consumption (Moens & Glen, 2002). This variety contains low concentrations of both erucic acid and glucosinolates, two substances that normally participate in the defence of brassicas against polyphagous plant feeders (Blau, Feeny & Contardo, 1978; Glen, Jones & Fieldsend, 1990). In a two-choice test, starved slugs were allowed to feed on young shoots of oilseed rape equitably distributed between two foraging area. The entrance to each foraging area, which the slugs had to cross to gain access to the oilseed rape shoots, was treated with either a cuticular extract from C. nemoralis or a control solution. It was hypothesized that slugs would preferentially feed on the foraging area treated with the control solution.
MATERIAL AND METHODS
For the purpose of the experiment, slugs (Deroceras reticulatum) were caught from homogeneous organic grassland and supplied by Arbiotech (Ille-et-Vilaine, Brittany, France) in early April 2014, since D. reticulatum was quite difficult to obtain in the local area at this time of year. The slugs were placed in Pyrex dishes (25 by 20 cm, 8.5 cm deep) lined with damp paper towel and kept at 75 ± 5% relative humidity (RH) to ensure a moist environment (Hommay, 2001). Pyrex dishes were stored in a climatic chamber at 11 ± 0.5 °C, to limit the development and spread of disease among slugs (Armsworth et al., 2005), with a L:D cycle of 11:13 h (light from 0900–2000). For 1 week before the trial, slugs were maintained on lettuce, and rabbit pellets (Coqtel®) were added as a convenient and standardized dietary supplement. Since terrestrial gastropods can show rapid learning for a particular food item after previous experience (Teyke, 1995; Cook et al., 2000), oilseed rape shoots were also added to the diet of the tested slugs to enhance feeding intensity on this plant during the trial.
Carabus nemoralis beetles were captured during spring 2013 in deciduous forests near Rennes-les-Bains (Aude, France). Only females were caught at this time and were stored in plastic boxes (22 by 13 cm, by 9 cm deep) on moist peat and a layer of foam in a climatic room (20 ± 0.5 °C and 70 ± 2% RH) with a L:D cycle of 9:15 h (light from 0800 to 1700), corresponding to standard rearing conditions for many ground-beetle species (Symondson, 1994; Assmann, 2003; Giglio et al., 2009). They were fed every day with Xeropicta derbentina snails, collected in the vicinity of the laboratory (Saint-Saturnin-lès-Apt, Vaucluse, France).
Young shoots of oilseed rape served as baits. The oilseed rape seeds were grown in a glass chamber with a sodium light (D:L 9:15 h) hung 40 cm above the seeds. The sodium light heated the chamber to a temperature of 28 ± 1.5 °C during the day, while the temperature dropped to about 12 ± 1 °C during the night. Oilseed rape seeds were sown on a fully hydrated vermiculite layer (about 1 cm thick) in a rectangular plastic box (12 by 5 cm, by 4 cm deep). Under such conditions, a growth period of about 4 days is needed to reach a suitable stage.
The methodology employed to prepare the cuticular extract was similar to that described by Bursztyka et al. (2013). One female, weighing 0.6 g, was placed in a clean glass beaker for 48 h to remove any digestive residues in the digestive tract. The beetle was then transferred to a new clean glass beaker and placed in a refrigerator at 4 °C for 4 h. At this low temperature the beetle enters a state of torpor and can then be transferred to a freezer at −80 °C for a further 4 h, avoiding the undesirable release of defensive secretions. Once dead, the beetle was placed in a new glass vial filled with 12 ml of pure ethanol (99.8%), equivalent to 20 ml of ethanol per gram of live weight. This stock solution was stored at 4 °C and gently shaken twice a week for 2 weeks. The beetle was then removed using clean metallic forceps. Working solutions were then created by placing 2 ml of stock solution in a new clean glass vial, stored at 4 °C for the duration of the trial. Pure ethanol, stored under the same conditions, served as a control solution.
The experiment was conducted using rectangular aluminium trays. Each tray measured 392 mm long, 77 mm wide and 35 mm deep, and had three parts (Fig. 1): a central rectangular crawling surface with a foraging area at either end. The crawling surface was covered with a piece of absorbent cellulose pad (Molinea® underpads, Hartmann Australasia) moistened with 20 ml of tap water upon which the slugs could move easily; the foraging areas were made from pieces of immunology plates (Dutscher cat. no. 055173) with eight rows of six wells each. Three young oilseed rape shoots were selected with their cotyledons completely unfolded, corresponding to the10th phenological stage of the BBCH scale (Meier, 2001). The shoots were placed in the fourth row of wells away from the crawling zone (i.e. roughly at the centre of the foraging zone), in three wells filled with hydrated vermiculite and separated from each other by a blank well, while all remaining wells were left empty. All inner walls were coated with Fluon® (AGC Chemicals Europe) both to prevent the slugs from escaping during the trials and to compel them to cross the treated thresholds providing access to the foraging areas. Treatments were applied at the junction of the crawling surface and the foraging areas, acting as a chemical fence which slugs had to cross in order to reach the baits behind. These thresholds were made from strips of Whatman paper (grade 1), 76 mm long by 26 mm wide, placed on a clean microscope glass slide of the same size in order to protect them from the moisture of the underlying pad. According to experiments by Armsworth et al. (2005), Fluon® has no effects on the behaviour of the tested animals.
Each treatment (the cuticular extract and a blank, i.e. a strip of Whatman paper moistened with 100 µl distilled water) was tested at the same time, in a separate device, against a control (the vehicle employed, i.e. pure ethanol). The foraging areas were distinguished according to the nature of the treatment applied to their thresholds, i.e. cuticular extract, blank or control.
Mature adult slugs, weighing between 0.3 and 0.5 g, were used for the experiment, as by Armsworth et al. (2005). This selected class size is representative of that found in crops and, along with a smaller class, it is more widespread in crops than the largest class (Archard et al., 2004; Howlett, 2005); it is therefore responsible for most of the damage. Slugs make use of trail following for various purposes (Ng et al., 2013), hence one slug was employed at a time to avoid interindividual choice bias. The tested slug was starved for 24 h, in order to enhance the motivational state and to reduce variation in food consumption (Rollo, 1988). One hour prior to the beginning of a replicate, the slug was placed in a habituation zone bounded by a circular plastic cylinder, 52 mm in diameter and 50 mm high, whose inner surface was coated with Fluon®. This allowed the slug to become accustomed to the device. 100 µl of the treatment was applied with a micropipette to the strips of Whatman paper on the glass microscope slides. Ethanol was allowed to evaporate at room temperature (about 17 °C) for 10 min; at 1800 h the plastic cylinder of the habituation zone was removed from each device, marking the start of a replicate, which lasted until 0900 the next morning. All tests were blind. After each replicate, the devices and the glass slides were thoroughly washed with detergent and hot water and then dried at 200 °C for 2 h. All disposable parts (strips of Whatman paper, pads and baits) were replaced. New slugs were used for each replicate and the positions of treatments were swapped with the control to avoid any positional bias.
Batches of four replicates with the cuticular extract against the control, and four replicates with the blank against the control solution, were performed over six nights; two others replicates were conducted during an additional night. A total of 26 replicates were thus performed with the cuticular extract against the control and 25 with the blank against the control (one replicate was excluded because no choice was made by the tested slug).
First, comparisons were made between the mean number of cotyledons consumed in the foraging areas of each of the two-choice test conditions (cuticular extract vs control and control vs blank). While the consumption of a small part of the cotyledon may not preclude the survivorship of the plant, partially damaged cotyledons (i.e. cotyledons with only a trace of consumption) were considered consumed to avoid the subjectivity of an assessment using a visual damage scale (as usually employed to assess damage on true leaves with larger surface areas; Frank, 1998a). As normality was not ascertained, the Wilcoxon matched-pairs signed-ranks test was used. The total number of slugs found in each foraging area was also compared using a binomial test. All data analyses were carried out using Statistica v. 10.0 software.
Significantly fewer cotyledons were consumed in the foraging area treated with the cuticular extract from Carabus nemoralis than in the area treated with the control solution (P = 0.0280). There was no significant difference (P = 0.8295) between the foraging areas treated with the control solution and those treated with the blank (Fig. 2).
Slugs were found significantly more often in the foraging area treated with the control solution than in those treated with the cuticular extract (P = 0.0060) (Fig. 3). The number of slugs found in the foraging area was almost the same in the control and blank conditions (12 and 13 respectively, P = 0.8415). Significantly fewer slugs were found in the foraging area when it was treated with the cuticular extract than when Whatman paper alone was used alone (P = 0.0327).
The experiment demonstrates that cuticular extract from Carabus nemoralis influences the grazing behaviour of Deroceras reticulatum. Significantly fewer cotyledons of oilseed rape were consumed in the foraging area with a threshold treated with C. nemoralis cuticular extract than in the control area. Chemical cues from C. nemoralis have already been shown to be effective in discouraging slugs from entering treated shelters (Bursztyka et al., 2013). While sheltering is an important component in slugs' survivorship (Rollo & Wellington, 1977; Grewal et al., 2001), foraging is obviously a more vital self-maintenance behaviour, so that interfering with it should be more difficult. The chemical cues were nevertheless successful in modifying foraging behaviour. The deterrent effect was not only observed in the reduced consumption of rape shoots in the treated plots, but also in the final position occupied by tested slugs at the end of the replicates. The chemical cues were potent in preventing slugs from reaching the protected foraging area, as shown by the greater number of slugs found in the control and blank foraging areas. As demonstrated by Bursztyka et al. (2013), this response is specific and not provoked simply by a new odour. Deroceras reticulatum is an epigeic feeder (Frank, 1998b), thus the treatment application method used here was particularly suitable to this species since slugs must crawl over the surface of the treated paper (rather than beneath it), as they would over the soil surface of a field. Furthermore, C. nemoralis is a ground-dwelling beetle that may leave residual chemical cues that are perceived by slugs, as observed with the generalist carabid predator Pterostichus melanarius (Armsworth et al., 2005). The application method thus fits with the ecology of both the prey and the predator, an important characteristic that influences the ability of prey to detect predators (Downes & Shine, 1998; Head, Keogh & Daughty, 2002).
In considering their methodological experimental procedure, Armsworth et al. (2005) pointed out that the antipredator behaviour displayed by D. reticulatum should be a response to kairomones left by P. melanarius while the beetles are moving. In other words, the chemical cues of interest to the slugs are presumably deposited passively from the cuticular exoskeleton of the beetles while they were crawling on the experimental surface. For this reason, we employed ethanol as the solvent in our experiment because, unlike other more usual extraction procedures using hexane or dichloromethane for the cuticular extract, it has affinity for hydrophobic and hydrophilic molecules and is not toxic to slugs. In similar approaches, ethanol extracts from cockroaches (Rollo, Czyzwska & Borden, 1994; Rollo, Borden & Casey, 1995) and isopods (Yao et al., 2009) proved effective to elute the repellent materials active against conspecifics.
The present study lays the groundwork for further research, which should focus first on identifying the compounds involved in the avoidance displayed by D. reticulatum, in order to consider the feasibility of creating a product based on these compounds for a practical application in pest management. Field experiments should be carried out to assess the efficacy of this semiochemical-based treatment in precluding severe damage to winter rape and other at-risk crops (such as winter wheat) and to evaluate this new approach as an additional tool for integrated pest management as part of a push-and-pull strategy. The chemical baits usually employed against slugs are less palatable than the crops they are supposed to protect, limiting their efficiency (Bailey, 2002). We can hence suggest that foraging slugs could be deflected from their favourite seedlings to the less palatable toxic baits—whose lethal potency would be increased as a result of the application of the deterrent semiochemical in the immediate surroundings of the seedlings. Similarly, slugs may be led to seek refuge (Bursztyka et al., 2013) and to forage in field margins, where they may be subject to higher risks of predation since field margins are reservoirs for predatory arthropods (Dennis & Fry, 1992). However, care should be taken to assess any side effects of these chemical cues on other useful species. For instance, ground beetles seem to locate sparsely and randomly distributed prey, such as slugs, through optimal Lévy-flight searching patterns, executed by avoiding odour trail previously left by themselves or conspecifics (Reynolds, 2007; Guy et al., 2008). It is believed that D. reticulatum detects these pheromone trails to its own benefit to avoid P. melanarius (Armsworth et al., 2005; Guy et al., 2008). It is therefore likely that the compounds in the cuticular extract from C. nemoralis involved in the deterrent effect on D. reticulatum have a similar purpose for these ground beetles. Consequently, a large-scale application of such semiochemicals may disturb the foraging behaviour of C. nemoralis, leading to a decrease of their predation upon slugs in the treated area. Nevertheless, such drawbacks should not be overestimated, since C. nemoralis is less common in arable crops than carabids from the tribe Pterostichini, such as P. melanarius.
Finally, it would be interesting to test the response of other pest slugs. For instance, C. nemoralis is an efficient predator of Arion lusitanicus (see Hatteland, 2010), which has become a major pest in central and northern Europe (Grimm, Paill & Kaiser, 2000; Hatteland et al., 2010). It is therefore likely that chemical cues from C. nemoralis could deter this slug from foraging in a similar way as demonstrated in this study with D. reticulatum. Further experiments should be conducted on slug species which feed below ground, like Deroceras laeve, against which surface application may be ineffective.
We are grateful to Julien Delnatte for his help in catching Carabus nemoralis and his invaluable knowledge of entomology, Julie Fourrier for her enthusiastic assistance, Arbiotech (Saint Gilles, FR) for providing slugs and Myriam Robejean for her help in the rearing of the beetles and slugs. We are also grateful to Eve Landen for correcting the manuscript and providing useful comments.