Biology Letters
You have accessResearch articles

Rapid learning in a native predator shifts diet preferences towards invasive prey

M. E. Alexander

M. E. Alexander

School of Health and Life Sciences, Institute of Biomedical and Environmental Health Research (IBEHR), University of the West of Scotland, High Street, Paisley PA1 2BE, UK

Contribution: Formal analysis, Methodology, Writing – original draft

Google Scholar

Find this author on PubMed

L. Skein

L. Skein

Department of Botany and Zoology, Centre for Invasion Biology, Stellenbosch University, Stellenbosch, South Africa

Contribution: Data curation, Formal analysis, Methodology, Writing – review & editing

Google Scholar

Find this author on PubMed

T. B. Robinson

T. B. Robinson

Department of Botany and Zoology, Centre for Invasion Biology, Stellenbosch University, Stellenbosch, South Africa

[email protected]

Contribution: Formal analysis, Methodology, Supervision, Writing – review & editing

Google Scholar

Find this author on PubMed


    Biological invasions often exert negative impacts on native communities and can disrupt a range of biotic interactions such as those between predators and prey. For example, when invasive species alter the foraging landscape, native predators can fail to recognize them as profitable prey because of unfamiliarity. This study therefore investigated whether a native predator (rock lobster Jasus lalandii) can develop a new preference for an invasive prey (mussel Semimytilus patagonicus) following conditioning through a short-term exposure. Conditioned lobsters, exposed to only S. patagonicus for a month, demonstrated a significant change in preference for the novel invasive prey, which was found to contrast with non-conditioned lobsters that continued to show predator preferences toward a native mussel (Choromytilus meridionalis). There is therefore potential for native predators such as J. lalandii to adapt and switch towards feeding on an abundant invasive prey, even if they avoid it at first. This indicates that rapid learning can occur in a species exposed to novel food resources and demonstrates that native species can adapt to biological invasions.

    1. Introduction

    Invasions by non-native species can exert negative impacts on a range of biotic interactions [1], including disruption to predator–prey relationships [2]. However, while the effects of non-native predators on native prey communities have been well documented (e.g. [35]), the impacts of non-native prey on native predators are comparatively understudied. Invasive prey species may displace competitively inferior native prey, subsequently altering the native predator prey-base [6]. This can in turn present native predators with several challenges, including the physical handling of morphologically unusual prey [7] or overcoming toxic defence mechanisms of some non-native species [8].

    The ability of a predator to switch from familiar to novel prey is important in the context of efficient resource utilization [9,10], and ‘conditioning’ (i.e. continuous exposure) towards new prey is a mechanism through which this can occur. Such adaptive learning is thought to occur more frequently in generalist predators compared to specialists, which often have fine-tuned adaptations to handle fewer, more specific prey types [11]. Various factors have been proposed to influence conditioning, including past feeding experience [12,13], frequency of prey encounters [14] and predator handling capabilities [11]. This concept of learning by native predators has often been investigated in relation to novel toxic prey [1517]; however, as an adaptive response to invasions by non-toxic prey, it has received less research attention.

    In South Africa, the west coast rock lobster Jasus lalandii is an important subtidal generalist predator with a preference for mussel prey [18,19]. However, the mussel prey-base for this species is changing due to the invasion by Semimytilus patagonicus (formerly S. algosus), which is beginning to form a significant part of the subtidal mussel community that comprised the native species Choromytilus meridionalis and Aulacomya atra [20]. Previous work conducted in this region has shown that J. lalandii avoids S. patagonicus, preferring to select native C. meridionalis [21]. In comparison to native mussels, the invasive species offers the greatest energetic reward and has the weakest shells, making it a profitable prey choice for most mussel predators [21]. Therefore, it is suggested avoidance occurs because of unfamiliarity with the novel species, which many south coast rock lobster populations have yet to encounter.

    The aim of this research was thus to determine whether the avoidance of S. patagonicus by the rock lobster J. lalandii can be overcome through continuous exposure (i.e. conditioning) to this invasive prey. As J. lalandii is a predator with a flexible diet and invasive S. patagonicus is currently the most profitable mussel prey, it was predicted that rock lobsters would switch from feeding on less profitable native mussels to S. patagonicus once they become familiar with this species.

    2. Methods

    Mussels (C. meridonalis, A. atra and S. patagonicus) and rock lobsters J. lalandii of a standardized length were collected from the wild to be used in predation trials (see electronic supplementary material for details on collection sites and sizes of animals used).

    (a) Phase 1: predator conditioning

    During a conditioning period, individual rock lobsters (n = 20) were exposed to S. patagonicus prey only. This was conducted in individual cages (0.045 m3 in volume) situated in the field (see electronic supplementary material) to ensure that rock lobsters were exposed to S. patagonicus in the presence of naturally occurring cues. An initial 7-day starvation period was conducted to allow for acclimatization and to standardize rock lobster hunger level. This was followed by four weeks of conditioning where J. lalandii were fed with S. patagonicus in crushed and whole forms. The total number of mussels (n = 40) provided to each lobster per week was kept constant throughout conditioning. The ratio of crushed to whole mussels was however progressively adjusted as rock lobsters became more familiar with this prey species, with only whole mussels offered in the final week (electronic supplementary material, table S1).

    (b) Phase 2: determining the influence of conditioning on prey preference

    To determine whether conditioning induced a switch in preference from C. meridionalis to S. patagonicus, mussel selection from conditioned lobsters (Phase 1) was compared to that of non-conditioned individuals from the same population. The prey preference of conditioned and non-conditioned J. lalandii (n = 20 per treatment) took place in the laboratory (see electronic supplementary material, information). Feeding trials lasted 7 days after a 7-day acclimatisation and starvation period. Rock lobsters were provided with one of two diet treatments which was either a ‘current diet’ or ‘future diet’ (n = 10 per diet for conditioned and non-conditioned lobsters). Diets comprised different proportions of the three subtidal mussel species of interest reflecting their current and projected occurrence in the field (table 1; [20]). Mussel consumption was tracked daily, and availability was kept constant throughout the trials.

    Table 1. Proportions of native C. meridionalis and A. atra, and invasive S. patagonicus that were presented to rock lobsters in current and future diet scenarios.

    scenario C. meridionalis A. atra S. patagonicus
    current 1 2 1
    future 1 1 2

    (c) Statistical analyses

    All statistical analyses were conducted in R v. 3.1.1 (R Core Team, 2016). The difference in overall prey consumption between non-conditioned and conditioned lobsters was analysed using a generalized linear model (GLM) with a quasi-Poisson error distribution and log link. Chesson selectivity indices [22] were calculated for each prey species (native C. meridionalis and A. atra and invasive S. patagonicus) through the formula:

    where ri/rj is the proportion of a particular species in the diet (consumed), pi/pj the proportion of that species present in the overall habitat (on offer) and n the total number of prey species in the overall habitat (on offer). When α = 1/n neutral selection/the absence of selective predation in rock lobsters is indicated, whereas α < 1/n infers negative selection (avoidance) and α > 1/n positive selection (preference). This selectivity index is appropriate in this case as it accounts for the presence and proportion(s) of multiple species in the feeding landscape.

    Chesson selectivity indices for different mussel prey species in each diet treatment for both conditioned and non-conditioned lobsters were first arscine transformed due to their proportional nature and then assessed using Friedman's ANOVAs. This was followed by Conover post hoc tests to detect differences between Chesson selectivity indices of prey species within each diet treatment. To establish whether conditioning altered selection of a particular prey species, the Chesson selectivity indices of each prey species as selected by conditioned and non-conditioned lobsters in each diet treatment were compared using t-tests with a Bonferroni correction to account for multiple comparisons.

    3. Results

    There was no statistical difference in the overall consumption of prey between non-conditioned and conditioned lobsters (GLM: F1,27 = 0.411, i = 0.53). There was, however, a significant difference in the selection of mussel species by non-conditioned J. lalandii in both diet treatments (current diet: χ22=7.153, p = 0.027, figure 1a; future diet: χ22=7.517, p = 0.023; figure 1b). Here, lobsters showed a positive selection towards the native mussel C. meridionalis regardless of the proportions in which it was present in the overall diet. However, this selection was not found to statistically differ from the novel prey, S. patagonicus, in the current diet (figure 1a). Prey selection by conditioned J. lalandii also differed significantly in both diet treatments (current diet: χ22=6.869, p = 0.032, figure 1c; future diet: χ22=9.867, p = 0.007; figure 1d), and conditioning of J. lalandii to S. patagonicus drove significantly greater selectivity for the invasive species compared to native prey. This selection was evident regardless of the proportions in which prey were offered to J. lalandii.

    Figure 1.

    Figure 1. Median (interquartile range, minimum and maximum) Chesson selectivity indices for three prey species (Choromytilus meridionalis, Aulacomya atra and Semimytilus patagonicus) as predated upon by non-conditioned rock lobsters in (a) current and (b) future diet treatments, and conditioned rock lobsters in (c) current and (d) future diet treatments. Dots represent outliers. Boxes with different letters differ significantly (Conover post hoc with Bonferroni correction). Values above the dashed line = positive selection (preference), on the line = neutral selection, below the line = negative selection (avoidance).

    Chesson selectivity indices for invasive S. patagonicus were significantly greater in conditioned lobsters fed on both the current (t11 = 2.402, p = 0.035; figure 2a) and future diets (t14 = 4.084, p = 0.001; figure 2b). This was accompanied by a decreased preference for the native C. meridionalis, as reflected by the lower selection for this species in conditioned lobsters in both the current (t11 = 3.115, p = 0.009; figure 2a) and future diets (t14 = 2.656, p = 0.019; figure 2b). Finally, the selection index for native A. atra remained low regardless of predator conditioning and regardless of diet treatment (current diet: t11 = 0.706, p = 0.495, figure 2a; future diet: t14 = 0.221, p = 0.828; figure 2b).

    Figure 2.

    Figure 2. Species-specific comparisons between Chesson selectivity index values (mean ± 1 s.e.) for prey (Choromytilus meridionalis, Aulacomya atra and Semimytilus patagonicus) as predated upon by non-conditioned and conditioned Jasus lalandii in the (a) current and (b) future diet treatments. Values above the dashed line = positive selection (preference), values on the line = neutral selection and values below the line = negative selection (avoidance).

    4. Discussion

    The ability of predators to adapt to an altered prey-base as a result of biological invasions is important for ensuring that they will be able to incorporate such prey should native resources decline or become fully displaced [7,23,24]. Therefore, native predator species that are able to do this may be more successful in the face of invasion fronts that move through habitats, displacing native species rapidly [10]. This is likely to be especially true of predators that can learn to adopt to consume new prey across short timescales. As such, we demonstrate in this study that, despite initial avoidance, a native predator, the rock lobster J. lalandii, can come to preferentially select an invasive mussel S. patagonicus through a period of rapid learning.

    After a short conditioning phase of four weeks, rock lobsters changed their diet preferences, as measured using Chesson selectivity indices [22], from the dominant native mussel C. meridonalis to the recent invader S. patagonicus. What was noteworthy was that this occurred irrespective of the proportions of each prey species appearing in the diet, with ‘current’ diets containing a greater number of native mussels while ‘future’ diets contained more invasive mussels. The invasion of S. patagonicus on the west coast of South Africa was first detected in 2009 [25], having recently been found to have spread to the south coast [20]. While this species represents a morphologically similar prey to the native prey-base, previous work reported that J. lalandii avoided it for reasons that are not clear [21]. Native predators, however, could be expected to shift towards feeding on a novel prey when there are low associated search and handling times, and when it offers higher energetic rewards compared to other prey, as predicted by classic foraging theory [26]. It is also noteworthy that no preference was observed in any of the treatment combinations towards the native mussel A. atra. Of the mussels on offer in this study, this species has the greatest shell strength and adductor muscle weight with an intermediate energy content [21] and was therefore likely not selected due to a trade-off between these measures.

    The invasive mussel S. patagonicus is a more profitable prey, offering the greatest energetic reward with the weakest shells [21], and it is therefore possible that the native predator is unfamiliar with other aspects of its presence such as its chemical signature. It is therefore likely that over the conditioning period, J. lalandii identified cues from novel S. patagonicus and through continuous exposure learned to associate it with a profitable food choice, leading to preference development. This ultimately drove rock lobsters to seek out S. patagonicus, despite the presence and relative abundances of co-occurring native C. meridionalis and A. atra in the experimental arenas. Here, a preference for S. patagonicus developed following exclusive exposure to this species for four weeks, however in a natural setting containing more prey species, the development of this preference may be delayed. It is also unknown whether this preference would persist over longer time periods or whether J. lalandii can retain a memory for this new species. However, the strong feeding switch to S. patagonicus observed in this study suggests that it might still be possible for rock lobsters to develop a preference for it in a setting where they will often encounter this species, which is likely given the invasion history of S. patagonicus to date where it rapidly forms dense beds that exclude competitors [27,28].

    Conditioning through continuous exposure can lead to the development of a chemical ‘search image’ for a specific prey, which can subsequently improve the ability to locate and ingest that prey [29]. Such chemoreceptive plasticity has been shown to be important in animals that are omnivorous, long-lived and found in various habitat types, all of which can lead to variation in prey availability [30]. Thus, even though rock lobsters are known to be generalist predators, variability in prey preference at an individual level can be extreme. In addition, it has been suggested that rock lobsters may have a genetic predisposition to act on chemical cues from prey that are profitable [31]. The ability to develop a chemical ‘search image’ for a particular prey (promoted through continuous exposure) can in itself be viewed as a mechanism that enhances the detection and intake of profitable prey and is likely important here with the new preference towards invasive, novel prey.


    The experiments in this study were carried out following Stellenbosch University guidelines for studies on invertebrates. All specimens used in experiments were collected under a research permit granted by the Department of Environmental Affairs and the then Department of Agriculture, Forestry and Fisheries of the Republic of South Africa (permit no. RES2018/32).

    Data accessibility

    The datasets and associated code supporting this article have been uploaded as part of the electronic supplementary material [32].

    Authors' contributions

    M.E.A.: formal analysis, methodology and writing—original draft; L.S.: data curation, formal analysis, methodology and writing—review and editing; T.B.R.: formal analysis, methodology, supervision and writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

    Competing interests

    We declare we have no competing interests.


    This work was supported by a PhD fellowship award to L.S. from the DSI-NRF Centre of Excellence for Invasion Biology.


    Katie Keanly and Nicole Martin are thanked for help in collecting mussels. Andrea Plos is thanked for collecting rock lobsters and for the supply of rock lobster cages for the field component. We are also grateful to False Bay Yacht Club for providing space to maintain rock lobsters in the field.


    Electronic supplementary material is available online at

    Published by the Royal Society. All rights reserved.