SEARCH

SEARCH BY CITATION

Keywords:

  • animal movement;
  • chemical camouflage;
  • chemical cue;
  • information use;
  • search tactics

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. Olfactory predator search processes differ fundamentally to those based on vision, particularly when odour cues are deposited rather than airborne or emanating from a point source. When searching for visually cryptic prey that may have moved some distance from a deposited odour cue, cue context and spatial variability are the most likely sources of information about prey location available to an olfactory predator.

2. We tested whether the house mouse (Mus domesticus), a model olfactory predator, would use cue context and spatial variability when searching for buried food items; specifically, we tested the effect of varying cue patchiness, odour strength, and cue–prey association on mouse foraging success.

3. Within mouse- and predator-proof enclosures, we created grids of 100 sand-filled Petri dishes and buried peanut pieces in a set number of these patches to represent visually cryptic ‘prey’. By adding peanut oil to selected dishes, we varied the spatial distribution of prey odour relative to the distribution of prey patches in each grid, to reflect different levels of cue patchiness (Experiment 1), odour strength (Experiment 2) and cue–prey association (Experiment 3). We measured the overnight foraging success of individual mice (percentage of searched patches containing prey), as well as their foraging activity (percentage of patches searched), and prey survival (percentage of unsearched prey patches).

4. Mouse foraging success was highest where odour cues were patchy rather than uniform (Experiment 1), and where cues were tightly associated with prey location, rather than randomly or uniformly distributed (Experiment 3). However, when cues at prey patches were ten times stronger than a uniformly distributed weak background odour, mice did not improve their foraging success over that experienced when cues were of uniform strength and distribution (Experiment 2).

5. These results suggest that spatial variability and cue context are important means by which olfactory predators can use deposited odour cues to locate visually cryptic prey. They also indicate that chemical crypsis can disrupt these search processes as effectively as background matching in visually based predator–prey systems.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The outcomes of predator–prey interactions have a major influence on both the ecology of predators and the dynamics of their prey (Tinbergen, Impekoven & Franck 1967; Lima 2002). To better understand how predators find their prey, many recent studies have modelled optimum predator movement strategies, including variations of random and Lévy walks, nearest-neighbour and trajectory-directed searching (e.g. Higgins & Strauss 2004; Viswanathan et al. 1999; Bartumeus 2009; Reynolds 2010). However, there is abundant evidence for strong selection on enhanced sensory capabilities in the game of detection and evasion between predator and prey, meaning that cues of prey activity are likely to be of fundamental importance in the search strategies of predators (Conover 2007).

Understanding the differences between direct and indirect cues of prey activity is essential to understanding their role in predator search. Much of the work modelling predator movements considers scenarios where detection and localization of prey occur simultaneously (Ruxton 2009), i.e. where predators rely upon direct visual detection of prey (e.g. Higgins & Strauss 2004; Viswanathan et al. 1999; Bartumeus 2009; Reynolds 2010; Tinbergen, Impekoven & Franck 1967). These models represent movement patterns that maximize rates of prey encounter. However, active prey often also emit abundant indirect cues, for which predator movement strategies that maximize information gain may be more profitable (e.g. Vergassola, Villermaux & Shraiman 2007). For example, UV sensing raptors forage in patches with high concentrations of UV visible urine marks secreted by their vole prey (Koivula & Korpimaki 2001); the searching raptors aim to increase information gain which inevitably precedes prey encounter.

Olfactory detection of prey is a particularly widespread, indirect mode of searching for prey used by a diverse range of predators. Animals inescapably emit body odours that provide clues to their presence and location, and many predatory animals are physiologically well equipped to exploit these cues (Conover 2007). Many potential prey species also intentionally use chemical communication for social or sexual intraspecific signalling, and predators can eavesdrop on these cues as illegitimate receivers (Zuk & Kolluru 1998). Often, terrestrial mammals deposit odour cues onto surfaces in their environment (either intentionally when signalling or unintentionally when resting), and then move elsewhere (e.g. see Hughes & Banks 2010). These deposited olfactory cues may persist long after the donor animal’s departure, thus encountering an odour cue is only the first step in the search process of olfactory predators. They must then use information within the spatial context of that deposited odour cue to locate their prey, a very specific type of search process, which to our knowledge has not been studied within the literature covering animals tracking either aquatic (e.g. Atema 1996; Ide et al. 2006; Webster & Weissburg 2009), or airborne (e.g. Carde & Willis 2008; Reynolds et al. 2009; Riffell, Abrell & Hildebrand 2008) odour plumes, or in visually based models of predator search. As all cues are spatially variable, information about prey location is likely embedded within different features of this variation (Plotnick 2007; Vergassola, Villermaux & Shraiman 2007). Indeed, a lack of variation in potential cues provides no information to searching predators; this is the basis of camouflage or background matching (Stevens & Merilaita 2009), which has been extensively studied in visual systems (e.g. Kettlewell 1955; Ruxton, Sherratt & Speed 2004; Merilaita & Lind 2005). More recently, olfactory background matching has been recognized in invertebrate systems; for example, Mechanitis polymnia larvae are defended against predatory ants by the close match between the chemical make-up of their cuticular lipids and those of their host plant (Henrique, Portugal & Trigo 2005). Similarly, Biston robustum caterpillars are visually cryptic on their host plant, but their cuticular chemicals are also very similar to those of the plant; this chemical crypsis prevents ant attack even after antennal contact has been made (Akino, Nakamura & Wakamura 2004). Additionally, chemical camouflage of female moth pheromone plumes by broadcast artificial pheromones effectively disrupts moth mating systems (Carde & Minks 1995), suggesting that, as in visual systems, olfactory cues must be sufficiently spatially variable to be distinguished from their background by a searcher.

Spatial variability, or ‘patchiness’ of an odour cue may simply represent a clumped spatial distribution of the odour cue, or it may result from differences in cue strength. Patchiness and variable strength of odour cues may be attributed to repeated visitation or low mobility of individuals (e.g. at a nest site; Banks, Norrdahl & Korpimaki 2000), from multiple animals (e.g. through intraspecific signal receiving; Hughes, Kelley & Banks 2009), or prolonged exposure to the environment (as aged cues vary in strength and chemical composition; e.g. Buesching, Waterhouse & Macdonald 2002). For a foraging predator, therefore, patchy, strong prey odour cues potentially equate to a higher probability of prey encounter. A tight spatial association between cues and prey is also predicted to be necessary for predators to associate cues with prey location (Pearce 1997), as cue–prey association is a measure of the cue: reward ratio, or how reliably cues indicate prey presence.

We tested whether these three factors – (i) patchiness and (ii) strength of odour cues, and (iii) cue–prey association – provide information to foraging predators using olfaction to locate prey. Optimal foraging theory (Charnov 1976; MacArthur & Pianka 1966) suggests that a predator should be able to improve its foraging success by using such information to update its search strategy (Dall et al. 2005). Whilst optimal foraging theory has been criticized for being too simplistic (Pierce & Ollason 1987), it is reasonable to assume that predators behave so as to maximize energy intake for minimum cost. Such costs may be measured in time, energy expenditure, or other ‘effort’ (Emlen 1966). This ‘decision-making’ process for a foraging predator can be summarized as the reward for effort ratio, or the overall ‘foraging success’ of the predator. For generalist predators, search strategies and prey preferences are predicted to change when preferred prey become scarce (Dukas & Kamil 2001) and foraging success falls.

In three experiments, we attempted to manipulate the foraging success of a model olfactory predator (Mus domesticus) by varying the application of prey odour cues to the environment. Peanuts, a highly valuable ‘prey’ item for mice, were hidden in sand to exclude visual detection. Peanut oil provided prey odour cues, and we varied the spatial distribution of odours to test our prediction that the mice would use this variation to locate the peanuts and improve their foraging success. Rodents were chosen as a model predator as they are known to use a highly developed olfactory system to locate food (Slotnick 2001), including both mobile and immobile prey, and inanimate vegetable matter. As most models of predator search are based on immobile prey, and we assume that mice use olfaction in a similar way to search for all food items, we felt the choice of peanuts as a model prey item was justified by the need to preclude the visual, auditory and vibrotactile cues that live prey such as invertebrates would emit (similarly, to limit foraging kiwis to olfactory search modes, Cunningham, Castro & Potter 2009, killed and buried their mealworm prey).

To test the importance of (i) patchiness in an odour cue, we varied the application of prey odours to simulate varying degrees of conspicuousness to an olfactory predator. We applied single-strength odour only to prey locations (making prey conspicuous), distributed prey odour uniformly on all patches (making prey match their background) or applied no additional odour (control). Thus, if mice use the patchiness of the odour cue to locate prey, then where odour cues are patchy they should be able to increase their foraging success above that expected if searching at random. Where cues match their background, they become effectively camouflaged to a foraging predator, and our hypothesis predicts that mice should only find peanut prey at random, as no information about prey location would be available to them.

To test the importance of (ii) odour strength, we created patchiness in odour concentration by altering the prey cue–background odour strength ratio. If mice direct their search towards locations with stronger odour as a means of locating prey, they should achieve higher foraging success in treatments where prey odours are stronger than background odours. When background odour strength approaches that of the prey odour, the cue becomes uniform and foraging success is predicted to be that of a random search, as in (i). To test whether mice would switch foraging strategies when searching for preferred prey became too costly, in this experiment we also provided pearl barley grains as easily available, but lower value alternate prey.

In our third experiment, we assessed the importance of (iii) cue–prey association for mouse foraging success. We used kernel analysis to determine the spatial scale at which mice were foraging in the second experiment, and then used this information to manipulate the level of spatial association between odour cues and ‘prey’ in the third experiment. Our hypothesis, that spatial association between cue and prey is important for predators to associate information in odour cues with probable prey location, predicts that where the cue–prey association is low mice should experience decreased foraging success, as lower reward rates will weaken cue-following behaviours. Conversely, where the association is strong, mice should be more successful.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Field Methods

This study was conducted in nine 15 × 15 m outdoor field enclosures at the Mallee Research Station, Walpeup, in the mallee wheat lands of north-western Victoria, Australia. Enclosure area was slightly smaller than the home range of mice in the breeding season (0·27 ha, Chambers, Singleton & Krebs 2000), and mice were able to cover the enclosure in a single night. Enclosures were mouse- and predator-proof, surrounded by two layers of smooth, galvanized steel, c. 1·5 m high and buried to a depth of at least 50 cm, preventing entry or escape of small mammals by either climbing or digging, and covered by wire mesh (5 cm aperture, 2·5 m high) to prevent entry by predatory birds or mammals (measurements in Hughes & Banks 2010; for construction details see Barker, Singleton & Spratt 1991). Thus, we were confident that the food used in our experiments was consumed only by animals placed into the enclosures by ourselves. The enclosures contained vegetation representative of the surrounding agricultural fields (low grasses, wheat and barley). The interior perimeter of each enclosure was mowed, creating a 70-cm buffer zone to discourage movement around the edges of the enclosure, and to further prevent the possibility of escape via climbing.

Mice were trapped along vegetated fence-lines adjacent to agricultural crops (wheat and canola) and in hay stores around the Mallee Research Station (from approximately eight different areas) using small Elliot traps (33 × 10 × 9 cm, Elliot Scientific Equipment, Upwey Victoria). Traps were baited with wheat and sunflower seeds rather than peanuts to prevent the association of peanuts with the trapping experience. Mice were trapped as needed and held in captivity for no longer than 2 weeks to facilitate normal foraging behaviour in enclosures. Subadult mice (<11 g) and adult females that were noticeably pregnant were excluded from the study.

Each enclosure contained 100 possible food patches, comprising of Sarstedt plastic Petri dishes (90 mm diameter, 10 mm depth) filled with sifted dry sand collected from local road verges. Petri dishes served as shallow foraging patches to allow the mice to rapidly find nuts with limited foraging, such that a visit to the site equalled ‘mortality’ of the simulated prey. Patches were laid out in 10 × 10 grids 1·3 m apart and 1·2 m from the buffer zone; walking paths were designated between grid lines to minimize trampling of vegetation when checking patches.

As mice use odour to communicate and readily scent mark novel surfaces (Desjardins, Maruniak & Bronson 1973) and food sources (Hurst 1993), care was taken to prevent cross-contamination of odour treatments. Dish marking behaviour was checked by feeding captive mice (n = 13) with sunflower seeds dyed with sodium fluorescein. This dye stains urine, which can then be detected using a UV light. Only one scent mark was found and it is likely that mice in this arid region may scent mark very little (N. Hughes pers. comm.). Despite this very low incidence of scent marking, additional sets of dishes were placed underneath food patches so that exchange of dishes and treatments would remove residual scent marks and prevent contamination of the bare ground. Petri dishes were carefully washed between treatments to remove any mouse scent marks and traces of peanut odour before re-use.

In each treatment, either 15 (Experiments 1 and 2) or nine (Experiment 3) of the dishes in each enclosure were randomly assigned as targets and received two pieces of crushed peanuts (c. 0·15 g), the prey. This amount was chosen to make the targets profitable and foraging selective, as initial trials using one piece of peanut showed that the mice were visiting a high proportion of all dishes. Peanut pieces were buried in the sand to prevent the use of visual cues in foraging and to prevent raiding by ants. Peanuts were chosen because (i) they are high in protein and mice are protein limited in this ecosystem (Singleton et al. 2001); (ii) small birds able to fit through the mesh around the enclosures were not attracted to them as a food source; (iii) peanut oil is readily available to provide prey odour; and (iv) mice are attracted to peanut oil (JPB pers. observ.). Enclosures also contained invertebrate prey, green foliage and spilled grain from previous experiments as additional food sources.

Extra virgin cold-pressed peanut oil (Biogenic brand) provided prey odour and hence cues of prey presence. The standard odour treatment comprised 200 mL of peanut oil mixed with 2·5 L of sand, as this provided a sandy texture without allowing the mice to use the oil as a food source. Mice were randomly assigned to enclosures and enclosures randomly assigned to individual treatments. A single mouse was released into each enclosure at dusk, as mouse foraging activity peaks in the hours immediately after sunset (Sutherland & Singleton 2003). Treatments began on the animals’ first night in the enclosure, as initial trials allowing acclimatization showed that the foraging activity of the mice did not change between first and subsequent nights in the enclosure (F1,6 = 0·34, = 0·59). Foraging activity (percentage of dishes visited) and prey survival (percentage of prey dishes surviving) were recorded the next morning at dawn. Deriving our equation from the optimality model, which assumes that predators behave so as to maximize energy intake for minimum cost, we defined foraging success as the percentage of dishes foraged that were targets, such that:

  • image

where foraging success represents the profitability of hunting for a certain prey type.

Data from eight mice were excluded from experiments because of weather or failure to forage (visited fewer than 5% of dishes). N = 8 in each treatment for Experiments 1 (total n = 24 mice) and 3 (total n = 32 mice) and n = 6–8 for Experiment 2 (total n = 43 mice); approximately even sex ratios were used in each experiment. Each mouse received a single treatment before being trapped out of the enclosures on subsequent nights.

Odour treatments

Experiment 1 tested the prediction that odour cues must be spatially variable to be useful to a foraging predator. Two odour treatments and a control were used (Fig. 1a). In patchy treatments, ‘prey odour’ was applied only to target dishes containing peanut prey. In uniform odour treatments, prey odour was applied to every dish in an enclosure, so that target dishes containing peanut prey would smell no different to dishes without prey. A uniform odour field in this sense is not something mice would likely encounter in a natural situation; however, we felt it was a useful experimental test of the effect of an olfactory cue matching its background. In control treatments, no additional odour was applied to any dishes; this was to test whether the mice could use the relatively weak odour cues of the peanuts themselves to direct their foraging.

image

Figure 1.  Schematic representation of the experimental design for (a) Experiment 1, testing for the effect of patchiness in an odour cue, (b) Experiment 2, testing for the effect of both an odour gradient in cue strength and the presence of alternate prey, and (c) Experiment 3, testing for the effect of variation in spatial association between cue and prey.

Download figure to PowerPoint

In Experiment 2, we used three different odour treatments (Fig. 1b) to determine whether mice use variation in the strength of an odour cue to locate food. Target dishes containing the peanut prey were always filled with the standard peanut oil sand mixture and thus were associated with the stronger peanut scent in all three treatments. Patchy and uniform odour treatments were repeated; control treatments were not used in Experiment 2, as results for control treatments in Experiment 1 showed that the mice could not use the cues of the peanuts themselves. The third treatment had a weak odour background: non-target dishes comprised sand with a 1 : 10 dilution of the standard odour treatment, that is, 250 mL of standard odour sand mixed with 2·25 L of plain sand. Alternate prey was provided in half the treatments via a second, internal grid of 15 Petri dishes. Pearl barley was chosen as low-value alternate prey because it contains approximately half the protein of peanuts. Each dish contained six grains of barley which provided half the protein of a single target dish. Barley dishes were placed an even distance from the sand dishes and barley was not hidden in sand to provide a conspicuous but lower quality alternate prey. The amount of alternate prey provided was not intended to represent a significant increase in the amount of alternate prey already available to the mice in the enclosures; rather to allow measurement of the animals’ use of alternate prey. To prevent raiding by ants and small birds, barley grains were lightly adhered to empty Petri dishes using double-sided tape and consumption was recorded at dawn. We were confident that barley was only consumed by mice, as they left teeth marks in the tape when removing barley grains.

Visual inspection of kernels created from the data in Experiment 2 using the ArcGIS Animal Movement package (ArcView GIS 3.3, ESR Institute 1992) suggested that mouse foraging was clumped at the 3 × 3 dish scale (c. 2·6 × 2·6 m); thus, the camouflage effect of uniform odour should only be effective at scales larger than this (Danell, Edentius & Lundberg 1991). The limited size of the enclosures made a test of this predicted camouflage effect from 4 × 4 odour patches impractical, as few dishes would remain odourless. We therefore designed Experiment 3 to test the converse prediction that a smaller, 2 × 2 application of odour around the targets would act as a larger patchy cue, attracting the mice and hence lowering target survival (compared to a larger-scale application of odour). Thus, cue and prey were intended to be very closely associated in this treatment, as odour was clumped around the targets.

Nine 2 × 2 patches were allocated within each enclosure, spaced two dishes apart so as to ensure independence of patches, and a single target was randomly allocated within each patch. In the patchy 2 × 2 treatment, odour was applied to the dishes of this 2 × 2 patch only (Fig. 1c). This spacing resulted in a dispersed distribution of the targets, which may have affected foraging; thus, targets were randomly assigned within the 2 × 2 patches in all treatments. As in the previous experiment, dishes containing target prey were always associated with odour.

To test whether a random, small-scale patchy distribution of odour cues would reduce cue–prey association, and hence the predator’s foraging success, odour was applied to nine empty dishes chosen randomly as well as on targets, such that only half of the odour dishes contained prey (Fig. 1c). This reduced the cue–prey association by 50%. Patchy and uniform odour treatments were repeated for comparison, as they represent opposite ends of the spatial scale of association between odour cue and prey presence. In patchy treatments, odour was closely spatially associated with the prey as 100% of odour dishes contained prey (Fig. 1c). In uniform odour treatments, the association between cue and prey was very low, as only 9% of odour dishes contained prey (Fig. 1c).

Statistical Analyses

One-way analysis of variance (anova) tested for treatment effects on foraging success, target survival and overall foraging activity in Experiments 1 and 3. For Experiment 2, two-factor anovas were used to detect differences in foraging success, target survival, and foraging activity, between background odour treatments (factor 1), and between treatments with and without alternate prey (factor 2). We then tested the prediction that repeated encounters with an alternate prey type will reduce predation on target prey (Dukas & Kamil 2001) by examining the relationship between predation on peanuts and barley (using Pearson’s correlation coefficient, r), and tested for differences in barley consumption between odour treatments using a one-factor anova. All statistics, including tests for the assumptions of anova, were performed using jmp statistical software (v5.1.2, SAS Institute 2004).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Mice readily foraged for peanuts in treated and in untreated sand, and in Experiment 1, 94·2% of target dishes that were visited had the peanuts removed; thus, visitation to a target dish was assumed to equal mortality of the ‘prey’. Inspection of the spatial distribution of visitation confirmed that mice were able to move around the entire enclosure in a single night.

Experiment 1: Patchiness of an Odour Cue

Mice in patchy odour treatments had almost twice the foraging success of mice in both control and uniform odour treatments, which approximated that of a random search [27·3% vs. 15·9% and 13·4% respectively; Fig. 2a, < 0·01, Table 1; Tukey’s Honest Significant Difference (HSD) post hoc test patchy > uniform = control]. A forager searching at random in this experiment would experience a foraging success of c. 15%, as 15 of the 100 dishes contained peanut prey. There was a trend towards greater prey survival in the uniform odour treatment (55·8%), compared to the patchy treatment (39·2%, Fig. 2b), which is consistent with the reduced predator foraging success, although this difference was not significant (= 0·41, Table 1). Post hoc power analysis revealed a minimum detectable difference in target survival of 46·1% between treatments for our design. There was no difference in the percentage of dishes visited between treatments (Fig. 2c).

image

Figure 2.  The mean percentage (±SE) of (a) dishes visited by mice that contained prey, (b) surviving peanut ‘prey’ dishes, and (c) dishes visited by mice in patchy, uniform and control odour treatments in Experiment 1. Common letters indicate treatment groups which were not significantly different (Tukey’s HSD, > 0·05). The dashed line represents the foraging success expected for a forager searching at random (15%).

Download figure to PowerPoint

Table 1.   The results of one and two-way anovas testing for the effects of (1) the patchiness of an odour cue, (2) the strength of an odour cue relative to its background and (3) the level of association between cue and prey on mouse foraging success, target survival and mouse foraging activity
 SourceSSd.f.FP
(1) Patchiness of an odour cue
   Foraging successTreatment875·9210·00<0·01
Error915·921  
   Target survivalTreatment148420·930·41
Error1682821  
   Foraging activityTreatment18320·290·75
Error670021  
(2) Variation in odour strength
  Foraging successBackground odour strength2514·0526·66<0·01
Alternate prey657·3313·480·07
Background odour strength × Alternate prey531·1821·410·26
Error6981·0237  
  Target survivalBackground odour strength3020·0722·320·11
Alternate prey1419·8712·180·15
Background odour strength × Alternate prey40·4920·030·10
Error24124·5137  
  Foraging activityBackground odour strength472·4621·750·19
Alternate prey43·0510·320·58
Background odour strength × Alternate prey246·9220·920·41
Error4357·3837  
(3) Cue–prey association
  Foraging successTreatment781·2035·71<0·01
Error1277·0428  
  Target survivalTreatment9259·2633·060·04
Error28209·8828  
  Foraging activityTreatment4734·63312·21<0·01
Error3619·2528  

Experiment 2: Strength of an Odour Cue

Background odour strength significantly altered mouse foraging success (< 0·01, Table 1; Tukey’s HSD post hoc test patchy > weak background = uniform). In patchy treatments without alternate prey (barley), cues were strongly associated with target prey and mouse foraging success was more than double that expected if searching at random (32·9% vs. 15%, Fig. 3a). Adding barley to patchy treatments, applying weak background odour with or without barley, and applying uniform odour without barley all reduced foraging success to approximately random (Fig. 3a). A combination of uniform background odour and the presence of alternate prey further reduced foraging success to even lower than expected for random foraging (Fig. 3a), although this difference was not statistically significant. Thus, there was also a trend for alternate prey to reduce foraging success (= 0·07, Table 1), although this effect appeared to differ depending on background odour strength (Fig. 3a). The interaction term was not significant, however (= 0·26, Table 1), and power analysis revealed that with the typical sample size of seven this experiment could reveal treatment differences exceeding 26·8%. Unexpectedly, neither background odour strength nor alternate prey presence made any difference to prey survival (Fig. 3b, Table 1) or mouse foraging activity (Fig. 3c, Table 1), and no interactions were found between factors (Table 1).

image

Figure 3.  The mean percentage (±SE) of (a) dishes visited by mice that contained prey, (b) surviving peanut ‘prey’ dishes, and (c) dishes visited by mice in patchy, weak background, and uniform odour treatments in Experiment 2, with (white bars) and without (grey bars) alternate prey (barley dishes) available. Common letters indicate treatment groups which were not significantly different (Tukey’s HSD, > 0·05). The dashed line represents the foraging success expected for a forager searching at random (15%).

Download figure to PowerPoint

Background odour did not affect the amount of barley consumed (F2,19 = 0·36; = 0·70) nor was any correlation found between barley consumption and target prey survival in either patchy (= −0·07) or weak background odour treatments (= −0·09). However, in the uniform odour treatment, higher barley consumption was strongly correlated with lower peanut prey survival (= −0·92).

Experiment 3: Cue–Prey Association

The foraging success of mice was strongly influenced by the spatial scale and pattern of odour distribution and hence the level of cue–prey association (< 0·01, Table 1). In the patchy treatment, mice again more than doubled (20·73%, Fig. 4a) the predicted random foraging success (9% in this experiment). Mouse foraging success under all other patterns of odour application was significantly lower (< 0·01, Table 1, Tukey’s HSD post hoc test patchy >patchy 2 × 2 = random = uniform), and similar to that expected for a random forager (Fig. 4a).

image

Figure 4.  The mean percentage (±SE) of (a) dishes visited by mice that contained prey, (b) surviving peanut ‘prey’ dishes, and (c) dishes visited by mice in patchy, patchy 2 × 2, random and uniform odour treatments in Experiment 3. Common letters indicate treatment groups which were not significantly different (Tukey’s HSD, > 0·05). The dashed line represents the foraging success expected for a forager searching at random (9%).

Download figure to PowerPoint

Prey survival varied according to the odour treatment (= 0·04, Table 1). Very few prey survived in the patchy treatment (15·3%), and survival was more than twofold higher in the random (40·3%) and uniform (31·9%) odour treatments (Fig. 4b). Unexpectedly, the survival rates in the patchy 2 × 2 treatment (62·5%) were approximately four times higher than in patchy treatments (Fig. 4b). A post hoc Tukey’s HSD test showed significant differences in prey survival between patchy and uniform treatments, but neither patchy 2 × 2 nor random were significantly different from either patchy or uniform (Fig. 4b). Similarly, background odour treatment also affected foraging activity (< 0·01, Table 1). In this experiment, mice were significantly more active (visited more dishes) in the uniform odour treatment than in any other treatment (Fig. 4c, Tukey’s HSD test).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Mice achieved high foraging success where odour cues were patchy rather than uniform, and where cues were tightly associated with prey location. However, they did not use differences in odour strength to direct their search and increase their foraging success, even when cues at prey patches were ten times stronger than background odour.

Patchiness in an Odour Cue

All three experiments consistently showed that our model olfactory predators, mice, used patchy odour cues to locate visually cryptic peanut ‘prey’, directing their search effort more accurately to target dishes, and doubling their foraging success in patchy odour treatments above that expected if searching at random, across all three experiments (Figs 2a, 3a, and 4a). This result supports our hypothesis that patchiness in an odour cue is a crucial form of information about prey location that is available to, and used by, searching olfactory predators. Conversely, when odour cues matched their background, mouse foraging success was consistently reduced to that expected if searching at random (uniform odour treatments, Figs 2a, 3a, and 4a), further confirming that the theory of camouflage developed in visual systems (Ruxton, Sherratt & Speed 2004; Stevens & Merilaita 2009) applies equally to olfactory cue detection (Ruxton 2009).

This result echoes the findings of Cunningham, Castro & Potter (2009), who buried killed mealworms in plain soil or soil into which they had mixed powdered freeze-dried mealworms, with the result that New Zealand brown kiwis (Apteryx mantelli) had much lower success at finding their prey when their sense of smell was ‘overwhelmed’ by the uniformly prey-scented soil. Such findings provide evidence that olfactory or chemical camouflage works through exactly the same mechanisms as visual camouflage – the cue of interest matches its background and cannot be picked out by the searcher (Stevens & Merilaita 2009).

Cue–Prey Association

Importantly, Experiment 3 revealed that the consistently high foraging success of mice in patchy odour treatments was probably also attributed to the tight coupling of the odour cue with the reward in these treatments, and that even minor disruption to this association can alter foraging success. In our patchy treatments, the odour cue was always paired with prey presence; thus, cue use was rewarded and reinforced with high foraging success. In the random treatments in Experiment 3, only 50% of the cues had targets, weakening the association between prey and cue, and mouse foraging success was reduced to no better than that expected if foraging at random. The same results were obtained for the uniform odour treatments, which had a very low association between the odour cues and target prey (9% of odour dishes contained targets). Where cue–prey association is low, cue-directed searching behaviours are not reinforced with rewards (Pearce 1997). Our results strongly suggest that odour cues must be both patchy and tightly associated with prey presence to be useful in directing predator search. Where cues were uniform or prey–cue association was low, mice did not use odour cues to locate prey, either because target odour cues were indistinguishable from the background or because the cue investigation behaviours were weakened by lowered reward rates, making this foraging strategy unprofitable.

Odour Cues and the Scale of Predator Search

We specifically assessed the scale at which the mice in our experiments were making foraging decisions when planning the patterns of odour distribution for Experiment 3. Nevertheless, the results obtained for the patchy 2 × 2 treatment in Experiment 3 did not match our predictions, which were based on our finding that the mice in Experiment 2 were foraging at the 3 × 3 dish scale. If the mice in Experiment 3 were assessing patches at the 3 × 3 dish scale, then the patchy 2 × 2 odour distribution should have appeared patchy to them, drawing them to the location of the prey. However, this pattern of odour effectively ‘camouflaged’ the prey patches, resulting in low foraging success, and suggesting that a 2 × 2 dish distribution of odour appeared uniform to these mice, i.e. they were making foraging decisions at the individual dish level as well as a larger scale (see Danell, Edentius & Lundberg 1991; and Hjältén, Danell & Lundberg 1993; for examples of other studies in which foragers undertook patch assessment at multiple scales). This result confirms the importance of considering the spatial scale most appropriate to the study animal, as the definition of a ‘patch’ is critical to foraging ecology research, and inextricably reliant upon the scale of assessment – a distribution that appears ‘patchy’ at one scale, may look uniform at another (Danell, Edentius & Lundberg 1991; Hjältén, Danell & Lundberg 1993; Wiggins et al. 2006; Hebblewhite & Merrill 2007). In terms of cue–prey association, assessment at the individual dish level would mean that in patchy 2 × 2 treatments mice would experience a 25% reward rate when investigating odour cues; thus, this odour treatment reduced the spatial association between cue and prey even further than in random treatments.

Variation in Odour Strength

Interestingly, mice in Experiment 2 did not use the difference in odour strength between strong target and weak background cues to direct their search, as foraging success did not differ between weak background and uniform odour treatments (Fig. 3a). It is possible that the mice interpreted the presence of the odour, in any concentration, as a cue to prey presence or that 1 : 10 is not a sufficient dilution of the odour cue, although given the highly developed rodent olfactory sensory system (Slotnick 2001), this seems unlikely. It is also possible that variation in cue age is more informative than concentration when searching for prey (e.g. the age of cues is important in social communication between cats; De Boer 1977; and the age of cues affects prey response to predator odours; Peacor 2006), as freshly deposited cues equate to more recent visitation and so a higher probability of prey encounter. As peanuts are inanimate, such a consideration would not strictly apply here, but is probably important for mice and other olfactory predators when searching for live prey. Aged olfactory cues change not only in concentration but in composition because some components dispersing or deteriorating faster than others (Buesching, Waterhouse & Macdonald 2002; Peacor 2006). Changes in concentration unaccompanied by changes in composition because of ageing are probably rare in nature, except perhaps where a larger animal may deposit a stronger odour cue (through volume alone) than a smaller animal. Most means of increasing concentration will also increase the freshness of the cue (e.g. repeated marking by multiple or individual animals, or higher prey activity levels in an area). In these experiments, all cues were of the same age and varied in concentration alone, something these wild mice may not naturally encounter.

Another possibility is that over the course of one night, mice may have been unable to learn the association between the stronger scent and the prey in this experimental protocol. A longer timeframe (perhaps several nights), providing more frequent encounters of the pairing of prey with strong scent, may facilitate discrimination between different odour strengths (Pearce 1997). The question of whether olfactory predators use spatial variability in odour concentration when searching for prey remains unresolved and will require careful disentanglement of the effects of cue age and strength. Other methods which reveal the order in which mice visited patches (such as spool-and-line tracking) could also help to resolve this.

Target Survival and Foraging Activity

Our results suggest that uniform odour increased survival rates for the ‘prey’, as prey survival was significantly higher in uniform than in patchy treatments in Experiment 3 (Fig. 4b), and tended to be higher in uniform treatments in both Experiments 1 and 2. Uniform odour increased search costs for the mice, which would be expected to increase prey survival rates. However, it was possible for prey survival to remain unaffected by uniform treatments if mice simply increased their search effort. The small size of the enclosures may have meant that mice paid few costs to investigate many dishes even when success rates were low. Mice did visit more dishes in uniform treatments in Experiment 3 and tended to visit more dishes in uniform treatments in Experiment 2, but foraging activity levels did not differ between treatments in Experiment 1.

Alternate Prey and the Limited Attention Hypothesis

Background matching is a strategy that works most effectively against generalist predators, which may give up searching for target prey when it becomes too costly and instead turn their attention to alternate prey. Because it is difficult to search efficiently for both prey types at once, they must ‘switch’ to a new search image (for the alternate prey) (Dukas & Kamil 2001). This ‘limited attention’ hypothesis predicts a positive relationship between the survival of peanut ‘prey’ and the consumption of alternate barley prey in Experiment 2. However, no positive relationship was found between barley consumption and peanut prey survival in any of the treatments. There was, however, a strong negative relationship between these factors when the odour distribution was uniform, confirming that mice were foraging at random in these treatments, resulting in equivalent encounter rates for both prey types (Ruxton 2005). Overall these results indicate that there was no limited attention being divided between target and alternate prey for these mice.

Future Directions

A field test of the ideas about olfactory predator search processes put forward in this paper would demonstrate whether free-living predatory mammals that rely on olfaction to hunt (e.g. dogs, foxes) use spatial variation in odour cues to locate their prey. Although there is abundant anecdotal evidence that predators use odours in the search for prey (e.g. the sensory capabilities and brain morphology of predators; Barton, Purvis & Harvey 1995; and in the development of trap baits; Saunders & Harris 2000), there is limited field data on how predators use the spatial distribution, and the context of deposited olfactory prey cues to improve their foraging success. Recently, Leighton, Horrocks & Kramer (2009) showed that mongooses searching for buried turtle nests rely on olfactory cues working in concert with other cues such as surface disturbance to the sand caused by burying (a visual cue), or looser sand above a buried object (a tactile cue), and it is likely that most predators also use a combination of several sensory modes when locating prey (Conover 2007). Here, we chose a study system that allowed us to examine the use of olfaction in particular, but further study into how olfaction works synergistically with the other senses when using cues to locate prey, and into the roles and preferences given to each sensory mode by different guilds of predators would be informative (for example, Colton & Hurst 2010, show that Gadus macrocephalus and Theragra chalcogramma larvae use light gradients but not chemical cues to locate patchily distributed prey).

The use of olfaction to locate prey is widespread, but the particular spatial and temporal properties of odour cues pose particular challenges for their exploitation. Future attempts to model predator search should therefore incorporate the use of variability within prey cues by searching predators and consider the important differences between search processes based on each sensory system. Our results also suggest that such models could be improved by including the role of unsuccessful search efforts in predator search strategies, which are strongly influenced by the scale of patch assessment. Methods for mapping exact paths such as spool-and-line techniques (e.g. Cox, Cox & Dickman 2000) or use of video footage (e.g. Klaassen, Nolet & Van Leeuwen 2007) would provide valuable insight into predator decision processes and foraging behaviour as cue: prey associations build up or break down, and potentially reveal foraging ‘rules’ or search tactics used by foragers. Such promising lines of further enquiry will provide important steps towards ‘putting predators back into predator–prey interactions’ (sensuLima 2002).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors thank the Mallee Research Station for their assistance and access to mouse enclosures; N. Hughes, C. Price, A. Lothian and A. Sabella for field assistance; and N. Hughes, C. Price and two anonymous referees for comments on this manuscript. This research was approved by the UNSW Animal Care and Ethics Committee and supported by ARC Discovery Grants DP0881455 and DP0877585 to PBB.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Akino, T., Nakamura, K. & Wakamura, S. (2004) Diet-induced chemical phytomimesis by twig-like caterpillars of Biston robustum Butler (Lepidoptera: Geometridae). Chemoecology, 14, 165174.
  • Atema, J. (1996) Eddy chemotaxis and odor landscapes: exploration of nature with animal sensors. The Biological Bulletin, 191, 129.
  • Banks, P.B., Norrdahl, K. & Korpimaki, E. (2000) Nonlinearity in the predation risk of prey mobility. Proceedings of the Royal Society B: Biological Sciences, 267, 16211625.
  • Barker, S.C., Singleton, G.R. & Spratt, D.M. (1991) Can the nematode Capillaria hepatica regulate abundance in wild house mice – results of enclosure experiments in southeastern Australia. Parasitology, 103, 439449.
  • Barton, R.A., Purvis, A. & Harvey, P.H. (1995) Evolutionary radiation of visual and olfactory brain systems in primates, bats and insectivores. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 348, 381392.
  • Bartumeus, F. (2009) Behavioral intermittence, Levy patterns, and randomness in animal movement. Oikos, 118, 488494.
  • Buesching, C.D., Waterhouse, J.S. & Macdonald, D.W. (2002) Gas-chromatographic analyses of the subcaudal gland secretion of the European badger (Meles meles) Part II: time-related variation in the individual-specific composition. Journal of Chemical Ecology, 28, 5769.
  • Carde, R.T. & Minks, A.K. (1995) Control of moth pests by mating disruption – successes and constraints. Annual Review of Entomology, 40, 559585.
  • Carde, R.T. & Willis, M.A. (2008) Navigational strategies used by insects to find distant, wind-borne sources of odor. Journal of Chemical Ecology, 34, 854866.
  • Chambers, L.K., Singleton, G.R. & Krebs, C.J. (2000) Movements and social organization of wild house mice (Mus domesticus) in the wheatlands of northwestern Victoria, Australia. Journal of Mammalogy, 81, 5969.
  • Charnov, E.L. (1976) Optimal foraging, marginal value theorem. Theoretical Population Biology, 9, 129136.
  • Colton, A.R. & Hurst, T.P. (2010) Behavioral responses to light gradients, olfactory cues, and prey in larvae of two North Pacific gadids (Gadis macrocephalus and Theragra chalcogramma). Environmental Biology of Fishes, 88, 3949.
  • Conover, M.R. (2007) Predator–Prey Dynamics: The Role of Olfaction. CRC Press, Boca Raton.
  • Cox, M.P.G., Cox, W.G. & Dickman, C.R. (2000) Use of habitat by the black rat (Rattus rattus) at North Head, New South Wales: an observational and experimental study. Austral Ecology, 25, 375385.
  • Cunningham, S.J., Castro, I. & Potter, M.A. (2009) The relative importance of olfaction and remote touch in prey detection by North Island brown kiwis. Animal Behaviour, 78, 899905.
  • Dall, S.R.X., Giraldeau, L.-A., Olsson, O., McNamara, J.M. & Stephens, D.W. (2005) Information and its use by animals in evolutionary ecology. Trends in Ecology & Evolution, 20, 187193.
  • Danell, K., Edentius, L. & Lundberg, P. (1991) Herbivory and tree stand composition: moose patch use in winter. Ecology, 72, 8.
  • De Boer, J.N. (1977) The age of olfactory cues functioning in chemocommunication among male domestic cats. Behavioural Processes, 2, 209225.
  • Desjardins, C., Maruniak, J.A. & Bronson, F.H. (1973) Social rank in house mice – differentiation revealed by ultraviolet visualization of urinary marking patterns. Science, 182, 939941.
  • Dukas, R. & Kamil, A.C. (2001) Limited attention: the constraint underlying search image. Behavioral Ecology, 12, 192199.
  • Emlen, J.M. (1966) The role of time and energy in food preference. The American Naturalist, 100, 611617.
  • ESRI (Environmental Systems Research Institute) (1992) ArcView GIS Version 3.3. ESRI, Redlands, California.
  • Hebblewhite, M. & Merrill, E. (2007) Multiscale wolf predation risk for elk: does migration reduce risk? Oecologia, 152, 377387.
  • Henrique, A., Portugal, A. & Trigo, J. (2005) Similarity of cuticular lipids between a caterpillar and its host plant: a way to make prey undetectable for predatory ants? Journal of Chemical Ecology, 31, 25512561.
  • Higgins, C.L. & Strauss, R.E. (2004) Discrimination and classification of foraging paths produced by search-tactic models. Behavioral Ecology, 15, 248254.
  • Hjältén, J., Danell, K. & Lundberg, P. (1993) Herbivore avoidance by association – vole and hare utilization of woody-plants. Oikos, 68, 125131.
  • Hughes, N.K. & Banks, P.B. (2010) Interacting effects of predation risk and signal patchiness on activity and communication in house mice. Journal of Animal Ecology, 79, 8897.
  • Hughes, N., Kelley, J. & Banks, P. (2009) Receiving behaviour is sensitive to risks from eavesdropping predators. Oecologia, 160, 609617.
  • Hurst, J.L. (1993) The priming effects of urine substrate marks on interactions between male house mice, Mus musculus domesticus Schwarz & Schwarz. Animal Behaviour, 45, 5581.
  • Ide, K., Takahashi, K., Nakano, T., Sato, M. & Omori, M. (2006) Chemoreceptive foraging in a shallow-water scavenging lysianassid amphipod: role of amino acids in the location of carrion in Scopelocheirus onagawae. Marine Ecology-Progress Series, 317, 193202.
  • Kettlewell, H.B.D. (1955) Recognition of appropriate backgrounds by the pale and black phases of Lepidoptera. Nature, 175, 943944.
  • Klaassen, R.H.G., Nolet, B.A. & Van Leeuwen, C.H.A. (2007) Prior knowledge about spatial pattern affects patch assessment rather than movement between patches in tactile-feeding mallard. Journal of Animal Ecology, 76, 2029.
  • Koivula, M. & Korpimaki, E. (2001) Do scent marks increase predation risk of microtine rodents? Oikos, 95, 275281.
  • Leighton, P.A, Horrocks, J.A. & Kramer, D.L. (2009) How depth alters detection and capture of buried prey: exploitation of sea turtle eggs by mongooses. Behavioral Ecology, 20, 12991306.
  • Lima, S.L. (2002) Putting predators back into behavioral predator–prey interactions. Trends in Ecology & Evolution, 17, 7075.
  • MacArthur, R.H. & Pianka, E.R. (1966) On optimal use of a patchy environment. The American Naturalist, 100, 603609.
  • Merilaita, S. & Lind, J. (2005) Background-matching and disruptive coloration, and the evolution of cryptic coloration. Proceedings of the Royal Society B-Biological Sciences, 272, 665670.
  • Peacor, S. (2006) Behavioural response of bullfrog tadpoles to chemical cues of predation risk are affected by cue age and water source. Hydrobiologia, 573, 3944.
  • Pearce, J.M. (1997) Animal Learning and Cognition: An Introduction. Psychology Press, East Sussex.
  • Pierce, G.J. & Ollason, J.G. (1987) 8 reasons why optimal foraging theory is a complete waste of time. Oikos, 49, 111118.
  • Plotnick, R.E. (2007) Chemoreception, odor landscapes, and foraging in ancient marine landscapes. Palaeontologia Electronica, 10, 111.
  • Reynolds, A.M. (2010) Animals that randomly reorient at cues left by correlated random walkers do the Lévy walk. The American Naturalist, 175, 607613.
  • Reynolds, A.M., Swain, J.L., Smith, A.D., Martin, A.P. & Osborne, J.L. (2009) Honeybees use a Levy flight search strategy and odour-mediated anemotaxis to relocate food sources. Behavioral Ecology and Sociobiology, 64, 115123.
  • Riffell, J.A., Abrell, L. & Hildebrand, J.G. (2008) Physical processes and real-time chemical measurement of the insect olfactory environment. Journal of Chemical Ecology, 34, 837853.
  • Ruxton, G.D. (2005) Increasing search rate over time may cause a slower than expected increase in prey encounter rate with increasing prey density. Biology Letters, 1, 133135.
  • Ruxton, G.D. (2009) Non-visual crypsis: a review of the empirical evidence for camouflage to senses other than vision. Philosophical Transactions of the Royal Society B: Biological Sciences, 364, 549557.
  • Ruxton, G.D., Sherratt, T.N. & Speed, M.P. (2004) Avoiding Attack: The Evolutionary Ecology of Crypsis, Warning Signals and Mimicry, Oxford University Press, Oxford.
  • SAS Institute (2004) JMP Version 5.1.2. SAS Institute Inc., Cary, NC.
  • Saunders, G. & Harris, S. (2000) Evaluation of attractants and bait preferences of captive red foxes (Vulpes vulpes). Wildlife Research, 27, 237243.
  • Singleton, G., Krebs, C.J., Davis, S., Chambers, L. & Brown, P. (2001) Reproductive changes in fluctuating house mouse populations in southeastern Australia. Proceedings of the Royal Society of London Series B-Biological Sciences, 268, 17411748.
  • Slotnick, B. (2001) Animal cognition and the rat olfactory system. Trends in Cognitive Sciences, 5, 216222.
  • Stevens, M. & Merilaita, S. (2009) Animal camouflage: current issues and new perspectives. Philosophical Transactions of the Royal Society B: Biological Sciences, 364, 423427.
  • Sutherland, D.R. & Singleton, G.R. (2003) Monitoring activity patterns and social interactions of small mammals with an automated event-recording system: wild house mice (Mus musculus) as a case study. Rats, Mice and People: Rodent Biology and Management (eds G.R.Singleton, L.A.Hinds, C.J.Krebs & D.M.Spratt), pp. 159164. Australian Centre for Agricultural Research, Canberra.
  • Tinbergen, N., Impekoven, M. & Franck, D. (1967) An experiment on spacing-out as a defence against predation. Behaviour, 28, 307321.
  • Vergassola, M., Villermaux, E. & Shraiman, B.I. (2007) ‘Infotaxis’ as a strategy for searching without gradients. Nature, 445, 406409.
  • Viswanathan, G.M., Buldyrev, S.V., Havlin, S., da Luz, M.G.E., Raposo, E.P. & Stanley, H.E. (1999) Optimizing the success of random searches. Nature, 401, 911914.
  • Webster, D.R. & Weissburg, M.J. (2009) The hydrodynamics of chemical cues among aquatic organisms. Annual Review of Fluid Mechanics, 41, 7390.
  • Wiggins, N.L., McArthur, C., Davies, N.W. & McLean, S. (2006) Spatial scale of the patchiness of plant poisons: a critical influence on foraging efficiency. Ecology, 87, 22362243.
  • Zuk, M. & Kolluru, G.R. (1998) Exploitation of sexual signals by predators and parasitoids. The Quarterly Review of Biology, 73, 415437.