SEARCH

SEARCH BY CITATION

Keywords:

  • human–wildlife interaction;
  • wildlife conservation management;
  • feeding behaviour;
  • maladaptive behaviour;
  • social learning;
  • bottlenose dolphin

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Although harmful human–wildlife interactions involving anthropogenic food sources are a significant issue for wildlife conservation, few studies have addressed whether social learning may influence how animals learn to use anthropogenic foods. We examined a long-term (1993–2003) human–wildlife interaction involving the illegal feeding of bottlenose dolphins (Tursiops aduncus) by recreational fishers in south-western Australia. We developed predictor variables for whether dolphins learned to accept food handouts from human provisioners, based on biological (age-class, sex) and behavioural (ranging and association patterns) data for a population of 74 dolphins. Two variables provided clear predictors for whether dolphins became conditioned to food handouts: the use of areas with high densities of recreational boats (BOAT) and the average coefficient of association with previously conditioned dolphins (ASSOC). An individual was more likely to become conditioned when it spent more time in high boat density areas and when it spent more time with other conditioned dolphins. When considering all the models available, there was strong weight of evidence for the effects of ASSOC and BOAT on the response variable. We were unable to detect any effects of age-class and sex with the available statistical power. These findings suggest that social learning can facilitate the acquisition of undesirable and maladaptive behaviours in wildlife, and indicate the value of long-term individual-specific data for the conservation management of wildlife engaging in undesirable interactions with humans.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Human–wildlife interactions involving anthropogenic foods (e.g. crops, fish captured with hooks or traps, food waste, provisioned foods) often result in death or injury (to humans or to wildlife), the culling of ‘problem’ wildlife, property damage, and economic loss (e.g. Bryant, 1994; Conover et al., 1995; Herrero, 2001; McGrath, 2001; McNay, 2002; Osborn, 2002; Gilman et al., 2006). Management of these interactions requires information on how animals learn to use anthropogenic foods, both to develop effective mitigation measures and to predict which animals are likely to acquire these undesirable behaviours (Whittaker & Knight, 1998; Taylor & Knight, 2003; Breck et al., 2008; Mazur & Seher, 2008). Many taxa that use anthropogenic foods, including elephants, cetaceans, canids, ursids and primates, have the capacity to acquire novel foraging behaviours in part through social learning (Box & Gibson, 1999; Galef & Giraldeau, 2001; Galef & Laland, 2005; Breck et al., 2008; Mazur & Seher, 2008). Despite this, few studies have examined whether social learning can influence how wild animals learn to use anthropogenic foods (Rendell & Whitehead, 2001a,b; Whitehead et al., 2004; Breck et al., 2008; Hoppitt & Laland, 2008; Mazur & Seher, 2008). Here, we examine whether social learning influenced bottlenose dolphins (Tursiops aduncus) that learned to accept food from recreational fishers in Western Australia, and whether predictor variables could be developed to indicate whether animals would learn to acquire anthropogenic foods, based on individual-specific data (Finn, 2005; Finn, Donaldson & Calver, 2008).

How can social learning influence how animals learn about anthropogenic foods?

Animals that learn to eat anthropogenic foods are often described as ‘food-conditioned’ (Breck et al., 2008; Mazur & Seher, 2008), ‘conditioned’ (Finn et al., 2008) or ‘conditioned to human interaction by food reinforcement’ (Samuels & Bejder, 2004), on the basis that they have acquired food-based behaviours through operant conditioning. In this context, operant conditioning is a learning process in which animals learn about anthropogenic food sources through: (1) their exposure to certain food-related human stimuli (e.g. campsites, human provisioners, fishing gear); (2) the utilization of particular behaviours in response to those stimuli (e.g. opening a lid, solicitous behaviours); (3) a food reward that positively reinforces these behaviours (e.g. food scraps) (Reynolds, 1975; McFarland, 1981; Whittaker & Knight, 1998; Samuels & Bejder, 2004; Young & Cipreste, 2004; Breck et al., 2008). Operant conditioning with food reinforcement implies a determinant role for human-related factors in acquiring novel behaviours. In natural environments, the deliberate or inadvertent presentation of a food source exposes animals to food-related stimuli and, for the period that it is available, sustains reinforcement for behaviours developed in response to these stimuli. Where feeding is deliberate, human provisioners can selectively reinforce certain conditioned behaviours, such as solicitous gestures (Mann, 2001; Durden, 2005; Finn et al., 2008).

The determinant role of human-related factors in the conditioning process suggests that animals may typically acquire behaviours for feeding on anthropogenic foods largely or solely through individual asocial learning (i.e. through their own independent experience of an environment containing anthropogenic food sources). Asocial learning is likely to be sufficient for the conditioning of species that receive limited maternal care and are solitary as juveniles and adults, and thus have few opportunities for observations of conspecifics interacting with anthropogenic foods. However, social learning could influence conditioning in more social species, particularly if social learning also influences the acquisition of the information and skills underlying natural behaviours (Whiten & Ham, 1992; Giraldeau, Caraco & Valone, 1994; Heyes & Galef Jr, 1996; Laland, Richerson & Boyd, 1996; Boran & Heimlich, 1999; Byrne, 1999; Galef & Giraldeau, 2001; Osborn, 2002; Whitehead et al., 2004; Breck et al., 2008; Hoppitt & Laland, 2008; Laland, Atton & Webster, 2011; Slagsvold & Wiebe, 2011; Thornton & Clutton-Brock, 2011). Social learning may be particularly important in species with extended periods of juvenile dependence and social structures involving long-term relationships between individuals (e.g. bottlenose dolphins: Wells, Scott & Irvine, 1987; Connor, Smolker & Richards, 1992; Connor et al., 2000; Rendell & Whitehead, 2001a; Lusseau, 2003; Krützen et al., 2005; Sargeant et al., 2005, 2007; Sargeant & Mann, 2009).

Further, as operant conditioning is a learning process, animals may integrate information and skills that are individually and socially acquired, making the ultimate acquisition of a learned behaviour the result of both individual experience and social influences (Galef & Giraldeau, 2001; Sargeant & Mann, 2009). For learned behaviours involving anthropogenic foods, social learning could occur through mechanisms ranging from stimulus or local enhancement to imitative learning (Whiten & Ham, 1992; Zentall, 2006; Hoppitt & Laland, 2008). These considerations suggest how social learning could play a facilitative or supplementary role in the acquisition of learned behaviours involving the use of anthropogenic foods, and thus potentially influence which individuals within a population become conditioned to human interaction (Rendell & Whitehead, 2001a,b). This view avoids the either/or dichotomy between asocial and social learning that sometimes characterizes debate over concepts such as culture and tradition (e.g. Heyes & Galef Jr, 1996; Rendell & Whitehead, 2001a,b; Whitehead et al., 2004; Krützen et al., 2005; Laland & Janik, 2006; Sargeant et al., 2007).

Can individual-specific data predict which individuals will learn to acquire anthropogenic foods?

Many attributes of potential interest for wildlife conservation and management are extremely difficult to measure in field-based research (Rendell & Whitehead, 2001a; Whitehead 2010). However, long-term studies that identify and monitor individual animals can allow for individual-specific traits to be assessed as predictors for whether or not animals will learn to acquire anthropogenic foods. These include measures or indices relating to: (1) exposure to anthropogenic foods, food-related stimuli and human provisioners (e.g. ranging patterns within areas containing anthropogenic food sources); (2) opportunities for social learning (e.g. patterns of association between individuals); (3) behavioural propensities, including those related to age, sex and reproductive status.

Human–dolphin interactions and anthropogenic food sources

Both bottlenose dolphin species (common bottlenose dolphin Tursiops trunactus and Indo-Pacific bottlenose dolphin T. aduncus) may engage in interactions with humans that are illegal, harmful or undesirable, based on their learning to accept food from recreational fishers or to obtain fish from fishing gear (depredation) (e.g. Bryant, 1994; Samuels & Bejder, 2004; Durden, 2005; Finn et al., 2008; Donaldson, Finn & Calver, 2010). Such interactions are particularly prevalent when dolphins exhibit long-term fidelity to (or seasonal occupancy of) coastal or estuarine areas that have high levels of human use, and thus may interact frequently with recreational users, tourism operators and commercial fishermen (Wells et al., 1987; Connor et al., 2000; Samuels, Bejder & Heinrich, 2000; Bejder & Samuels, 2003; Samuels et al., 2003).

Finn et al. (2008) documented a long-term (1993–2003) human–wildlife interaction based on recreational fishers illegally feeding members of a resident bottlenose dolphin (T. aduncus) community in Cockburn Sound, Western Australia. They found that the number of dolphins exhibiting behaviours indicative of conditioning to human interaction by food reinforcement increased over time, from 1 in 1993 to 14 by mid-2003. The availability of long-term ecological and behavioural information for known individuals provided a unique opportunity to examine whether social learning influenced the apparent conditioning of dolphins to human interaction in Cockburn Sound.

Study aims

Here we developed predictor variables for whether dolphins in Cockburn Sound became conditioned, based on environmental (distribution of recreational boats), biological (age-class and sex) and behavioural (ranging and association patterns) data, and examined the contribution of these variables using a generalized linear model. Our focus was whether the conditioning of individual dolphins was associated with four variables: (1) sex; (2) age-class; (3) use of areas with high densities of recreational boats prior to becoming conditioned; (4) patterns of associations prior to becoming conditioned. We used the findings from the model to examine the following hypotheses:

  1. Exposure to human provisioners H0: Whether a dolphin became conditioned showed no association with the frequency with which individuals utilized areas with high densities of human provisioners;
  2. Behavioural propensities H0: Whether a dolphin became conditioned showed no association with age-class or sex; and
  3. Social learning H0: Whether a dolphin became conditioned showed no relationship with individual patterns of association.

We consider the findings to illustrate how long-term and individual-specific data can be used to identify key factors associated with harmful human–wildlife interactions and, for social species, to examine the potential influence of social learning.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study area and population

Cockburn Sound (32°12' S, 115°44' E) is an embayment on the southwestern coast of Australia. It covers an area of approximately 124 km2 and represents an important recreational fishing area for the city of Perth, Western Australia, a metropolitan area of more than 1.2 million people. From 1993 to mid-1997, a community of 74 dolphins (excluding calves) was identified residing year-round in Cockburn Sound (Finn et al., 2008; Donaldson et al., 2010). Additional study of the community was conducted 2000–2003 (Finn, 2005). Members of the community showed long-term site fidelity and ranged largely within Cockburn Sound, although individual home ranges varied. Identical methods were used during the two study periods and individuals were photo-identified, sexed and classified in age classes (Finn et al., 2008).

Information was also collected on human–wildlife interactions involving the illegal feeding of dolphins in Cockburn Sound, including the identity and behaviours of dolphins that interacted with humans (Finn, 2005; Finn et al., 2008; Donaldson et al., 2010). The combination of data from the two studies allowed us to identify long-term (10 + years) trends in the number of dolphins that exhibited conditioned behaviours (i.e. behaviours indicative of conditioning to human interaction by food reinforcement) and to determine that while only one individual was observed to interact with recreational fishers for food in 1993, at least 14 did so by mid-2003 (Finn et al., 2008). These dolphins consisted of eight adults and six independent sub-adults and 11 of the 13 dolphins of known sex were male (Finn et al., 2008). Twelve of these dolphins were confirmed to have first exhibited conditioned behaviours between 1995 and 2005, with the two exceptions being: (1) an adult male (‘Touch’) considered conditioned at the start of the study in 1993 and (2) a sub-adult that exhibited conditioned behaviours when first identified in 2003.

Finn et al. (2008) used fixed-width transect sampling to determine the density of recreational vessels across ten locations in Cockburn Sound, including the two main seagrass habitats within the embayment. Transect sampling was conducted from June 2000 to April 2001. Boats were recorded as present if they were anchored or drifting within 200 m of the transect line. Vessels that were moving were not recorded as present as they may have been transiting through the habitat. Boat density differed significantly across the 10 locations, with the highest densities observed in the two seagrass habitats (2.2–2.5 boats km−2) and a nearshore area (0.5 boats km−2) (Finn et al., 2008). Finn et al. (2008) also found a significant correlation between encounter rates for conditioned dolphins during transect sampling and densities of recreational boats (i.e. the higher the density of boats, the greater the likelihood of encountering a conditioned dolphin). A summary of the characteristics of dolphins exhibiting conditioned behaviours between 1993 and 2003 is in Finn (2005), and a map of locations of dolphins exhibiting conditioned behaviours during transect sampling is in Finn et al. (2008).

Finn (2005) documented two findings suggesting the potential for social learning to influence the conditioning of dolphins to human interaction in Cockburn Sound. Firstly, individual-specific association data indicated that: (1) associations existed between 13 of the 14 dolphins that exhibited conditioned behaviours and that (2) for most individuals, these associations were confirmed to exist prior to these individuals exhibiting the behaviours indicative of conditioning (Finn, 2005). Secondly, observational data collected ad libitum from 1993 to 2003 included evidence of specific contexts in which social learning could have occurred, such as observations of conditioned dolphins receiving food from humans while unconditioned individuals who later became conditioned were present (Finn, 2005). Six of the conditioned dolphins, for example, were observed within 10 m of a feeding interaction involving the community's first conditioned animal (the adult male Touch) on at least one occasion prior to their becoming conditioned.

Behavioural methodology

During sightings of dolphins, we collected behavioural data using a behavioural survey based on a 5-minute scan sample of group composition and activity (Mann, 1999). We used these data to calculate the half-weight coefficient of association (COA), a measure of the association between two individuals based on the frequency with which they are observed together (Cairns & Schwager, 1987; Smolker et al., 1992). We calculated the COAs between all individuals within the Cockburn Sound community using 927 behavioural surveys from 1993 to 1997.

Identification of conditioned dolphins

We classified dolphins from Cockburn Sound as conditioned to human interaction through food reinforcement if two requirements were fulfilled.

The first requirement was observations of an individual exhibiting behaviours identified as being indicative of conditioning to human interaction through food reinforcement (Samuels & Bejder, 2004; Finn et al., 2008). Following Finn et al. (2008), we classified individuals as exhibiting ‘conditioned’ behaviours if they: (1) exhibited an active, direct approach to a stationary vessel or jetty; (2) remained in close proximity to the vessel or people on the jetty; (3) exhibited solicitous behaviours while in close proximity to vessels or people on a jetty; (4) accepted food from humans if offered.

The second requirement was observational data indicating a ‘pre-conditioning’ interval for each individual, that is, a series of observations, typically over the first several years of research, in which a dolphin was not observed to exhibit behaviours indicative of conditioning. We use the term ‘unconditioned’ to describe dolphins that did not meet these two requirements. This classification scheme is restrictive as it excluded individuals that exhibited behaviours indicative of conditioning to human interaction, but for which a series of pre-/post-conditioning observations did not exist.

Statistical analysis

We used a generalized linear model to examine the contribution of four variables to whether dolphins in Cockburn Sound became conditioned to human interaction through food reinforcement: age-class; sex; use of high-boat density areas; and association(s) with conditioned dolphins prior to their becoming conditioned (Finn, 2005). Our analysis included 72 individuals of the 74 dolphins identified as Cockburn Sound residents for the period 1993–97, excluding one individual that disappeared towards the start of this period and one dolphin (Touch) who was considered conditioned at the start of our study in 1993 (Finn et al., 2008). It also excluded the conditioned dolphin Ladder, who was observed for only two behavioural surveys in 1996 and thus was not confirmed as a Cockburn Sound resident. We excluded calves from our analyses because: we did not include calves in our determination of the Cockburn Sound community; no calves were observed engaging in these feeding interactions; and their dependent status complicates association analyses (Finn, 2005; Finn et al., 2008).

Given our hypotheses, we assessed whether the following independent variables influenced whether dolphins became conditioned (COND): (1) dolphin age-class (AGE); (2) dolphin sex (SEX); (3) the proportion of sightings (i.e. behavioural surveys) for a dolphin that were located within high-boat density areas in Cockburn Sound (BOAT); (4) the average association index of an individual with conditioned individuals (for dolphins that became conditioned, this index was calculated from data only from the time period when they were still classified as ‘unconditioned’) (ASSOC). We fitted generalized linear models using a binomial distribution for errors and selected models using Akaike Information Criteria (AICc). We did not detect any collinearity issues between the independent variables. A full explanation of the predictors used in this study follows.

Factors 1 and 2 – age-class (AGE) and sex (SEX)

Data for the age-class (AGE) and sex (SEX) of individuals are based on field-sexing and age estimation of individuals from 1993 to 1997 (Finn, 2005; Finn et al., 2008). We determined sex from observations of genital areas or consistent presence of a calf in infant position, in conjunction with association patterns. We assigned individuals to one of two age classes (adult, sub-adult) based on physical (body length, size) and behavioural attributes, including consistent presence of a calf (Gibson & Mann, 2008). For sub-adult-sized males, we assigned them as sub-adult if they showed strong associations with other sub-adults and weak or non-existent associations with adult males. One approximately adult-sized male was designated ‘sub-adult’ based on his strong and stable associations with sub-adult males, and four females were designated ‘adult or subadult’.

Factor 3 – use of high-boat density areas (BOAT)

We considered the seagrass habitats in Cockburn Sound to be high-boat density areas following Finn et al. (2008) and as described previously. To examine individual patterns in the use of seagrass areas, we determined the frequency with which individual dolphins were observed in these habitats between 1993 and 1996 (the number of behavioural surveys in which the individual was observed in seagrass areas divided by the total number of surveys in which an individual was observed, Fig. 1). This time period provided a useful baseline to assess ranging patterns because provisioning interactions were not well established in Cockburn Sound at this time (i.e. there were only four conditioned dolphins by the end of 1996) (Finn et al., 2008).

figure

Figure 1. Frequency distribution for the proportion of the behavioural surveys for each individual (1993–1997) in which the dolphin was observed in areas of high-boat density (n = 74 dolphins).

Download figure to PowerPoint

Factor 4 – average association index with individuals that were previously conditioned (ASSOC)

We used COA data to develop a variable providing a proxy for the propensity that an individual would spend time with conditioned conspecifics (ASSOC). We calculated the value for this variable by determining the average COA of an individual to all conditioned dolphins, using COA data from the first 5 years of the study as described previously. To determine average COAs for individuals that became conditioned during the study, we used COA data for years before they were conditioned. This ensured that ASSOC included only relevant COA data, and excluded COAs from periods when both individuals were already conditioned.

We chose average COA rather than summed COA to estimate the probability that the individual of interest is present when other dolphins perform the conditioned behaviour, because of the risk that using summed COA would overestimate an individual's exposure to the behaviour. An individual that has high COAs with several conditioned associates will necessarily spend time with more than one conditioned associate together, and will likely be exposed to the conditioned behaviour from two or more of these associates during a single vessel encounter. Although the opportunities for exposure to conditioned behaviours during a vessel encounter will be similar regardless of the number of conditioned associates present, in that during any encounter the naïve individual will be exposed to repeated solicitous behaviours and acceptance of any food (whether by one or more associates) for the duration of the encounter, for encounters involving more than one conditioned associate the use of summed COA would overestimate the probability of exposure, by double- (or triple-) counting such single exposure situations. While average COA is interpretable readily as meaning an average exposure, summed COA confounds simultaneous exposure to conditioned behaviours with independent exposures.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Identification of conditioned dolphins

We classified 11 dolphins from Cockburn Sound as conditioned during the study, based on the defined requirements. A further three dolphins exhibited behaviours indicative of conditioning to human interaction by food reinforcement, but were not classified as conditioned because: (1) there were no preconditioning observations for them (the dolphin Touch) or (2) observations were limited (Ladder, n = 2 behavioural surveys; Dharma, n = 1 behavioural survey).

Generalized linear model

The best-fitting model (Table 1) included an effect of the use of high-boat density areas (BOAT) and the average coefficient of association with conditioned conspecifics (ASSOC) on the likelihood that an individual became conditioned (Fig. 2; Table 2). An individual was more likely to become conditioned when it spent more time in the high-boat density area and when it spent more time with other individuals that were already conditioned. The second best-fitting model (Table 1) also included an effect of age-class; however, this effect was not significant. Finally, there was less support for other models emerging from the hypotheses (Table 1, ΔAICc > 2). When considering all the models emerging from the hypotheses, there was strong weight of evidence for the effects of ASSOC and BOAT on the response variable (cumulative AICc weights for BOAT: 0.90 and for ASSOC: 0.95). Any effects of age and sex were sufficiently small as to be undetectable in the generalized linear model with the available statistical power.

figure

Figure 2. The partial effect of (a) BOAT and (b) ASSOC on the likelihood that an individual became conditioned given the best-fitting model (Table 2). The figure presents the estimated effect along with the 95% confidence interval as well as the spread of the data along the x-axis.

Download figure to PowerPoint

Table 1. Model selection process using AICc
ModelAICcΔAICcAICc weight
  1. There is less support for models that have ΔAICc > 2 compared with the best-fitting model and little support for models that have ΔAICc > 4. Models were developed from the hypotheses presented in the Introduction. A measure of support for the influence of independent variables on the response variable can be estimated by summing the AICc weights of all models in which the independent variables were included. The closer to 1 this cumulative weight is, the more support there is for this effect. BOAT, frequency in which an individual used seagrass (high-boat density) areas; ASSOC, coefficient of association of individuals to conditioned dolphins; AGE, age-class; SEX, sex; AIC, Akaike Information Criteria.

Constant59.721.1< 0.001
BOAT43.14.50.043
ASSOC41.42.80.100
AGE6122.4< 0.001
SEX5314.4< 0.001
BOAT + ASSOC38.600.404
BOAT + ASSOC + AGE390.40.331
BOAT + ASSOC + AGE + SEX412.40.122
Table 2. Details of the best-fitting generalized linear model estimating the likelihood that an individual became conditioned given the time it spent in the high-boat density area and with other individuals that were already conditioned (deviance 32.58 on 70 d.f.)
 CoefficientSEz-statisticP
Intercept−4.91.04−4.75< 0.0001
BOAT14.926.922.160.031
ASSOC0.150.072.170.030

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study addressed two key questions for studies of human–wildlife interactions involving anthropogenic food sources, and for applied wildlife conservation management: (1) whether social learning could influence how animals learn to acquire anthropogenic food sources, and (2) whether individual-specific data could be used to predict which individuals would learn to use those food sources. Generalized linear models suggested that two attributes provided clear predictors for whether dolphins learned to accept food from human provisioners: the use of areas in which high densities of boats occurred, and association with a previously conditioned dolphin.

Identifying the predictors of an animal's risk of anthropogenic feeding interactions will be of considerable value to wildlife managers. It is not surprising that the first attribute, exposure to human provisioners, emerged as a significant predictor of whether dolphins became conditioned, given that: (1) animals learn readily through operant conditioning and (2) human provisioners provided the stimuli, food rewards, and schedules of reinforcement needed to develop and sustain conditioned behaviours among dolphins in Cockburn Sound from 1993 to 2003 (Finn et al., 2008). This finding indicates that individual-specific ranging data can help managers to identify those individuals at greatest risk of learning how to use anthropogenic food sources, particularly if data on the distribution of those food sources are available or ‘hotspots’ of human–wildlife interactions occur (Samuels & Bejder, 2004; Finn et al., 2008). However, ranging data (such as those used in this study) are only indices for exposure to human provisioners, and more refined estimates of potential exposure could be obtained through sustained observations of known individuals (e.g. focal individual follows – Altmann, 1974; Mann, 1999), recording individual visitation to areas where food sources and provisioners occur, or through other individual-based sampling approaches.

While this study found that some individual attributes were effective predictors, the influence of others was difficult to interpret. For example, although our analysis did not indicate that either age-class or sex was a significant predictor variable for whether dolphins became conditioned, we are cautious about dismissing the potential influence of propensities associated with age and or sex. We suggest that the potential contributions of these variables are unresolved because of the small and sex-biased sample sizes, and because other studies suggest that age and sex may be important factors influencing whether wild animals interact with humans in other contexts (e.g. Mann & Smuts, 1999; Mann et al., 2000; Samuels et al., 2000; Constantine, 2001; Fuentes & Gamerl, 2005).

With regard to our finding that dolphins were more likely to learn to accept food from humans when they spent more time with other conditioned dolphins, we emphasize that examinations of the influence of social learning in contexts like Cockburn Sound should be conservative, because of the way in which humans can orchestrate the conditioning of animals to human interaction through food reinforcement, and the challenges of field-based studies to provide definitive proof of the social transfer of information and skill (Rendell & Whitehead, 2001a,b). However, three of our findings indicate that the independent acquisition of behaviours indicative of conditioning is unlikely for at least some dolphins in Cockburn Sound, and thus that a facilitative role for social learning is plausible and parsimonious. Firstly, Finn (2005) documented the presence of both conditioned and unconditioned individuals during feeding interactions, indicating specific contexts in which the social transfer of information could have occurred. Secondly, association data from 1993 to 1997 demonstrate that all individuals who became conditioned had some level of association with at least one other previously conditioned dolphin, and in many cases these association levels were high (Finn, 2005). Finally, this study found that the amount of time spent with conditioned conspecifics was a significant predictor of whether dolphins became conditioned, even after accounting for the effect of using high-boat density areas. Horizontal transmission of social information may be indicated, given that no conditioned individuals were dependent calves. We note also the potential to apply network models, such as network-based diffusion analysis, to datasets to investigate social learning (Hoppitt, Boogert & Laland, 2010). We are interested in the application of these methods to investigate harmful human–wildlife interactions such as those reported here, particularly as we acquire datasets with detailed timelines for behaviour acquisition.

Conservation management implications

These findings support predictions about the influence of social learning in human–wildlife interactions (Whitehead et al., 2004). In particular, this study suggests that, while exposure to human provisioners and other anthropogenic food sources may provide sufficient cause for wild animals to learn how to use anthropogenic foods, some form of social learning can have a facilitative function, and thus may influence the rate and direction in which behaviours associated with conditioning to human interaction are acquired within a population. In these situations, individual-specific association data may help wildlife managers to predict how quickly an undesirable behaviour will spread within a population, and help to identify and target the individuals and groups at greatest risk of learning this behaviour (Whitehead et al., 2004). Social information sometimes leads to incorrect cost-benefit decisions for animals (Rieucau & Giraldeau, 2011), and given recent evidence that dolphins conditioned to taking food from humans in Cockburn Sound have higher risks of boat strikes and entanglements (Donaldson et al., 2010), this study indicates the potential for social relationships and social learning to facilitate the transmission of harmful or maladaptive behaviours.

This study also shows the potential value of longitudinal and individual-specific data for managing human–wildlife interactions (Bejder et al., 2006, 2009), particularly those involving high-conservation value species, or species that may present a potential threat to humans and human infrastructure (e.g. crops – Pienkowski et al., 1998). Such studies can be useful not only to indicate which animals may become involved in harmful or undesirable interactions. They can also be used to improve decisions about conservation or management actions that may be controversial (e.g. aversive conditioning, culling and translocation), by guiding managers to focus their actions geographically or narrow the suite of individuals to be targeted, thereby reducing costs and impacts to non-target individuals (e.g. Butler et al., 2008). Finally, the role of human provisioners observed in this study emphasizes that human behaviours can be key determinants of harmful human–wildlife interactions, and suggests that efforts to change human behaviours are often the best management option.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

For their financial, technical or logistical support, we thank BP Kwinana Refinery, Stuart Bradley, Bob Wyburn, Bob Hammond, Ian Potter Foundation, Orbital Engine Corporation, Jane's Marine, Rockingham Wild Encounters, Western Australia Environmental Protection Authority, Oceanica, Cockburn Sound Powerboat Association, American Museum of Natural History, Rhode Island Zoological Society, the Western Australia Department of Fisheries, and the Western Australia Department of Environment and Conservation. Field research was conducted under the conditions of licenses, authorities and permits from: Western Australia Department of Environment and Conservation, Western Australia Department of Health, and Murdoch University Animal Ethics Committee.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Altmann, J. (1974). Observational study of behavior: sampling methods. Behaviour 49, 227267.
  • Bejder, L. & Samuels, A. (2003). Evaluating the effects of nature-based tourism on cetaceans. In Marine mammals: fisheries, tourism, and management issues: 229256. Gales, N. , Hindell, M. & Kirkwood, R. (Eds). Collingwood: CSIRO Publishing.
  • Bejder, L., Samuels, A., Whitehead, H., Finn, H. & Allen, S. (2009). Impact assessment research: use and misuse of habituation, sensitisation and tolerance to describe wildlife responses to anthropogenic stimuli. Mar. Ecol. Prog. Ser. 395, 177185.
  • Bejder, L., Samuels, A., Whitehead, H. & Gales, N. (2006). Interpreting short-term behavioral responses to disturbance within a longitudinal perspective. Anim. Behav. 72, 11491158.
  • Boran, J.R. & Heimlich, S.L. (1999). Social learning in cetaceans: hunting, hearing, and hierarchies. In Mammalian social learning: comparative and ecological perspectives. Symposia of the Zoological Society of London: 72. 282307. Box, H.O. & Gibson, K.R. (Eds). Cambridge: Cambridge University Press.
  • Box, H.O. & Gibson, K.R. (Eds) (1999). Mammalian social learning: comparative and ecological perspectives. Symposia of the Zoological Society of London 72. Cambridge: Cambridge University Press.
  • Breck, S.W., Williams, C.L., Beckmann, J.P., Matthews, S.M., Lackey, C.W. & Beecham, J.J. (2008). Using genetic relatedness to investigate the development of conflict behavior in black bears. J. Mammal. 89, 428434.
  • Bryant, L. (1994). Report to congress on results of feeding wild dolphins: 1989–1994. Silver Springs: Office of Protected Resources, National Marine Fisheries Service.
  • Butler, J.R.A., Middlemas, S.J., McKelvey, S.A., McMyn, I., Leyshon, B., Walker, I., Thompson, P.M., Boyd, I.L., Duck, C., Armstrong, J.D., Graham, I.M. & Baxter, J.M. (2008). The moray firth seal management plan: an adaptive framework for balancing the conservation of seals, salmon, fisheries and wildlife tourism in the UK. Aquat. Conserv. 18, 10251038.
  • Byrne, R.W. (1999). Cognition in great ape ecology: skill-learning opens up foraging opportunities. In Mammalian social learning: comparative and ecological perspectives. Symposia of the zoological society of London 72: 333350. Box, H.O. & Gibson, K.R. (Eds). Cambridge: Cambridge University Press.
  • Cairns, S.J. & Schwager, S. (1987). A comparison of association indices. Anim. Behav. 3, 14541469.
  • Connor, R.C., Smolker, R.A. & Richards, A.F. (1992). Two levels of alliance formation among bottlenose dolphins (Tursiops sp.). P. Natl. Acad. Sci. USA. 89, 987990.
  • Connor, R.C., Wells, R.S., Mann, J. & Read, A.J. (2000). The bottlenose dolphin: social relationships in a fission-fusion society. In Cetacean societies: field studies of dolphins and whales: 91126. Mann, J. , Connor, R.C.C. , Tyack, P.L. & Whitehead, H. (Eds). Chicago: University of Chicago Press.
  • Conover, M.R., Pitt, W.C., Kessler, K.K., DuBow, T.J. & Sanborn, W.A. (1995). Review of human injuries, illnesses, and economic losses caused by wildlife in the United States. Wildl. Soc. Bull. 23, 407414.
  • Constantine, R. (2001). Increased avoidance of swimmers by wild bottlenose dolphins (Tursiops truncatus) due to long-term exposure to swim-with-dolphin tourism. Mar. Mamm. Sci. 17, 689702.
  • Donaldson, R., Finn, H. & Calver, M. (2010). Illegal feeding increases risk of boat-strike and entanglement in bottlenose dolphins. Pacific Conserv. Biol. 16, 157161.
  • Durden, W.N. (2005). The harmful effects of inadvertently conditioning a wild bottlenose dolphin (Tursiops truncatus) to interact with fishing vessels in the Indian River Lagoon. Aquat. Mamm. 31, 413419.
  • Finn, H.C. (2005). Conservation biology of bottlenose dolphins (Tursiops sp.) in Perth metropolitan waters. PhD thesis, Murdoch University, Perth, Western Australia.
  • Finn, H.C., Donaldson, R. & Calver, M.C. (2008). Feeding Flipper: a case study of a human-dolphin interaction. Pacific Conserv. Biol. 14, 215225.
  • Fuentes, A. & Gamerl, S. (2005). Disproportionate participation by age/sex classes in aggressive interactions between long-tailed macaques (Macaca fascicularis) and human tourists at Padangtegal Monkey Forest, Bali, Indonesia. Am. J. Primatol. 66, 197204.
  • Galef, B.G. & Laland, K.N. (2005). Social learning in animals: empirical studies and theoretical models. BioScience 55, 489499.
  • Galef Jr, B.G. & Giraldeau, L.A. (2001). Social influences on foraging in vertebrates: causal mechanisms and adaptive functions. Anim. Behav. 61, 315.
  • Gibson, Q.A. & Mann, M. (2008). The size, composition and function of wild bottlenose dolphin (Tursiops sp.) mother-calf groups in Shark Bay, Australia. Anim. Behav. 76, 389405.
  • Gilman, E., Brothers, N., McPherson, G. & Dalzell, P. (2006). A review of cetacean interactions with longline gear. J. Cetacean Res. Manage. 8, 215223.
  • Giraldeau, L.-A., Caraco, T. & Valone, T.J. (1994). Social foraging: individual learning and cultural transmission of innovations. Behav. Ecol. 5, 3543.
  • Herrero, S. (2001). Bear attacks: their causes and avoidance. Revised edition. Toronto: M & S Publishing.
  • Heyes, C.M. & Galef Jr, B.G. (Eds) (1996). Social learning in animals: the roots of culture. San Diego: Academic Press.
  • Hoppitt, W., Boogert, N.J. & Laland, K.N. (2010). Detecting social transmission in networks. J. Theor. Biol. 263, 544555.
  • Hoppitt, W. & Laland, K.N. (2008). Social processes influencing learning in animals: a review of the evidence. Adv. Study Behav. 38, 105165.
  • Krützen, M., Mann, J., Heithaus, M.R., Connor, R.C., Bejder, L. & Sherwin, W.B. (2005). Cultural transmission of tool use in bottlenose dolphins. Proc. Natl. Acad. Sci. USA 102, 89398943.
  • Laland, K.N., Atton, N. & Webster, M.M. (2011). From fish to fashion: experimental and theoretical insights into the evolution of culture. Phil. Trans. R. Soc. B. 366, 958968.
  • Laland, K.N. & Janik, V. (2006). The animal cultures debate. Trends Ecol. Evol. 21, 542547.
  • Laland, K.N., Richerson, P.J.R. & Boyd, R. (1996). Developing a theory of animal social learning. In Social learning in animals: the roots of culture: 129154. Heyes, C.M. & Galef Jr., B.G. (Eds). London: Academic Press.
  • Lusseau, D. (2003). Male and female bottlenose dolphins Tursiops sp. have different strategies to avoid interactions with tour boats in Doubtful Sound, New Zealand. Mar. Ecol. Prog. Ser. 257, 267274.
  • Mann, J. (1999). Behavioral sampling methods for cetaceans: a review and a critique. Mar. Mamm. Sci. 15, 102122.
  • Mann, J. (2001). Cetacean culture: definitions and evidence. Behav. Brain Sci. 24, 343.
  • Mann, J., Connor, R.C., Barre, L.M. & Heithaus, M.R. (2000). Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behav. Ecol. 11, 210219.
  • Mann, J. & Smuts, B.B. (1999). Behavioral development in wild bottlenose dolphin newborns Tursiops sp. Behaviour 136, 529566.
  • Mazur, R. & Seher, V. (2008). Socially learned foraging behaviour in wild black bears, Ursus americanus. Anim. Behav. 75, 15031508.
  • McFarland, D. (1981). Conditioning. In The Oxford companion to animal behaviour: 9798. McFarland, D. (Ed.). Oxford: Oxford University Press.
  • McGrath, C. (2001). The Fraser Island dingo case. Environ. Plann. Law J. 18, 269272.
  • McNay, M.E. (2002). Wolf-human interactions in Alaska and Canada: a review of the case history. Wildl. Soc. Bull. 30, 831843.
  • Osborn, F.V. (2002). Capsicum oleoresin as an elephant repellent: field studies in the communal lands of Zimbabwe. J. Wildl. Manage. 66, 674677.
  • Pienkowski, M.W., Watkinson, A.R., Kerby, G., Naughton-Treves, L., Treves, A., Chapman, C. & Wrangham, R. (1998). Temporal patterns of crop-raiding by primates: linking food availability in croplands and adjacent forest. J. Appl. Ecol. 35, 596606.
  • Rendell, L. & Whitehead, H. (2001a). Culture in whales and dolphins. Behav. Brain Sci. 24, 309324.
  • Rendell, L. & Whitehead, H. (2001b). Cetacean culture: still afloat after the first naval engagement of the culture wars. Behav. Brain Sci. 24, 360382.
  • Reynolds, G. (1975). A primer of operant conditioning. Illinois: Scott, Foresman, and Co.
  • Rieucau, G. & Giraldeau, L.-A. (2011). Exploring the costs and benefits of social information use: an appraisal of current experimental evidence. Phil. Trans. R. Soc. B. 366, 949957.
  • Samuels, A. & Bejder, L. (2004). Chronic interactions between humans and free-ranging bottlenose dolphin near Panama City, Florida, USA. J. Cetacean Res. Manage. 6, 6977.
  • Samuels, A., Bejder, L., Constantine, R. & Heinrich, S. (2003). Swimming with wild cetaceans, with a special focus on the Southern Hemisphere. In Marine mammals: fisheries, tourism, and management issues: 277303. Gales, N. , Hindell, M. & Kirkwood, R. (Eds). Collingwood: CSIRO Publishing.
  • Samuels, A., Bejder, L. & Heinrich, S. (2000). A review of the literature pertaining to swimming with wild dolphins. Report to the Marine Mammal Commission. Bethesda, MD, USA.
  • Sargeant, B.L. & Mann, J. (2009). Developmental evidence for foraging traditions in wild bottlenose dolphins. Anim. Behav. 78, 715721.
  • Sargeant, B.L., Mann, J., Berggren, P. & Krützen, M. (2005). Specialization and development of beach hunting, a rare foraging behavior, by wild bottlenose dolphins (Tursiops sp.). Can. J. Zool. 83, 14001410.
  • Sargeant, B.L., Wirsing, A.J., Heithaus, M.R. & Mann, J. (2007). Can environmental heterogeneity explain individual foraging variation in wild bottlenose dolphins (Tursiops sp.)? Behav. Ecol. Sociobiol. 61, 679688.
  • Slagsvold, T. & Wiebe, K.L. (2011). Social learning in birds and its role in shaping a foraging niche. Phil. Trans. R. Soc. B. 366, 969977.
  • Smolker, R., Richards, A.F., Connor, R.C. & Pepper, J. (1992). Sex differences in patterns of association among Indian Ocean bottlenose dolphins. Behaviour 123, 3869.
  • Taylor, A.R. & Knight, R.L. (2003). Behavioral responses of wildlife to human activity: terminology and methods. Wildl. Soc. Bull. 31, 12631271.
  • Thornton, A. & Clutton-Brock, T. (2011). Social learning and the development of individual and group behaviour in mammal societies. Phil. Trans. R. Soc. B. 366, 978987.
  • Wells, R.S., Scott, M.D. & Irvine, A.B. (1987). The social structure of free-ranging bottlenose dolphins. In Current mammalogy: 247306. Genoways, H.H. (Ed.). New York: Plenum.
  • Whitehead, H. (2010). Conserving and managing animals that learn socially and share cultures. Learn. Behav. 38, 329336.
  • Whitehead, H., Rendell, L., Osborne, R.W. & Würsig, B. (2004). Culture and conservation of non-humans with reference to whales and dolphins: review and new directions. Biol. Conserv. 120, 431441.
  • Whiten, A. & Ham, R. (1992). On the nature and evolution of imitation in the animal kingdom: reappraisal of a century of research. Adv. Study Behav. 21, 239283.
  • Whittaker, D. & Knight, R.L. (1998). Understanding wildlife responses to humans. Wildl. Soc. Bull. 26, 312317.
  • Young, R.J. & Cipreste, C.F. (2004). Applying animal learning theory: training captive animals to comply with veterinary and husbandry procedures. Anim. Welf. 13, 225232.
  • Zentall, T.R. (2006). Imitation: definitions, evidence, and mechanisms. Anim. Cognit. 9, 335353.