Defence strategies of Chrysomela lapponica (Coleoptera: Chrysomelidae) larvae: relative efficacy of secreted and stored defences against insect and avian predators

Defence strategies of Chrysomela lapponica (Coleoptera: Chrysomelidae) larvae: relative efficacy... Abstract Larvae of the leaf beetle Chrysomela lapponica defend themselves by release of repellent secretions, but also store potentially toxic compounds in their body tissues. We addressed the role of major groups of predators in the evolution of these two defence strategies by testing effects of these strategies on the behaviour of insect (wood ant, Formica polyctena) and avian (great tit, Parus major) predators. Ants were repelled by larval defensive secretions, but not by larvae devoid of secretions, larval haemolymph or integument. By contrast, birds rejected larvae devoid of secretions after the first attack; this suggests the presence of non-volatile defensive compounds within the larval body. However, survival was three-fold greater for larvae with intact secretions than for larvae with depleted secretions due to (1) irritating effects of secretions, resulting in frequent release of undamaged prey, and (2) faster avoidance learning and better prey memorability based on contact with secretions. Thus, volatile secretions and non-volatile compounds stored in the body act against birds jointly. Secretions sequestered from host plants were more effective than were autogenously produced secretions. We conclude that insect predators could contribute to the evolution of secreted but not of stored defensive chemicals, whereas bird predation could favour the evolution of both lines of defence. avoidance learning, bird predation, chemical defence, prey memorability, subtribe Chrysomelina, wood ant predation INTRODUCTION Animals have evolved numerous ways of defending themselves against their natural enemies. Among the various anti-predatory strategies, chemical defences are most widespread, and their diversity across species is striking (Blum, 1981; Ruxton, Sherratt & Speed, 2004; Eisner, Eisner & Siegler, 2007). To date, more than 1000 chemical compounds have been reported to have defensive functions in animals (El-Sayed, 2014), causing either toxic or irritating effects or both, or serving a warning function (Ruxton et al., 2004; Eisner et al., 2007). Moreover, animals employ various strategies of production, storage and deployment of these chemicals. The defensive chemicals may be synthesized de novo or can be acquired from the environment, in particular from food plants (Bowers, 1992; Nishida, 2002), and they can be used in different ways before, during and after a predator attack (Ruxton et al., 2004). The intriguing question is how evolution has shaped this diversity of defensive compounds and the strategies of their use. The two major defensive strategies of arthropods differ in the site of storage and mode of deployment of the defensive compounds: (1) they may be accumulated and stored internally in the body, or (2) they may be accumulated in specialized integumental organs, usually ectodermal glands, and released externally upon a predator attack (Blum, 1981; Pasteels, Grégoire & Rowell-Rahier, 1983). The chemicals stored inside the body cause toxic or deterrent effects on the predator only after the prey is consumed or its integument is broken, which means that the prey is seriously injured or killed. This strategy provides protection for other similar prey due to avoidance learning and generalization of predators (Gamberale-Stille & Tullberg, 1999; Skelhorn & Rowe, 2006a, b, c; Svádová et al., 2009). By contrast, externally secreted defences act on the predator’s chemical senses during (or even before) the attack and can thereby protect the prey from being injured or killed (Wiklund & Järvi, 1982; Sillén-Tullberg, 1985; Hotová Svádová et al., 2013). Thus, these two strategies differ in their effects on the survival of individual prey (Skelhorn & Rowe, 2006c). Secretion of defensive compounds is likely to be more costly compared to their storage in body tissues for several reasons. First, secretion usually requires the development of specialized organs, thereby incurring high constitutive costs. Second, considerable volumes of secretions are usually depleted during a predator attack, leading to additional costs of replenishment of lost chemicals and liquids (Bowers, 1992; Higginson et al., 2011). Third, defence externalization incurs ecological costs, expressed as an increased risk of parasitism (Zvereva & Kozlov, 2016). Conversely, the strategy of defence externalization may be advantageous because it can stop a predator’s attack before the prey is fatally damaged (Wiklund & Järvi, 1982; Sillén-Tullberg, 1985). Moreover, birds learn more rapidly to avoid prey possessing externally secreted defences than prey having defences stored in the body (Skelhorn & Rowe, 2006c; Hotová Svádová et al., 2013). A recent meta-analysis (Zvereva & Kozlov, 2016) has also confirmed that secretions released from defensive glands are more effective against naïve vertebrate predators than are stored defences. Thus, secretion of defensive compounds is more costly, but it is also more effective than storing them in body tissues; therefore, both strategies have similar cost–benefit ratios for animals using them. Consequently, both strategies are widely distributed. The high variation observed for the effectiveness of prey defences against different taxonomic and ecological groups of natural enemies (Zvereva & Kozlov, 2016) suggests that the response to chemically defended prey is taxon-specific. In particular, vertebrate and arthropod predators differ considerably in their hunting strategies, perception of chemical signals and sensitivity to toxins (Lagadic & Caquet, 1998; Kaupp, 2010; Boevé et al., 2013). For example, the main target of volatile, non-specific irritants released from defensive glands is suggested to be arthropods, while specific toxins stored in the body are more effective against birds and other vertebrate predators (Pasteels et al., 1983; Aldrich, 1988). Therefore, one hypothesis is that the evolution of different defensive strategies was driven by contrasting selective pressures imposed by various natural enemies (Boevé et al., 2013). Conversely, a meta-analysis (Zvereva & Kozlov, 2016) did not confirm differential effects of the two strategies on vertebrate and invertebrate (largely arthropod) predators in terrestrial ecosystems. The reason may be that non-toxic compounds secreted by prey may cause deterrent effects on invertebrate as well as on vertebrate predators (Conner et al., 2007) and could therefore be as effective as stored toxins against vertebrates. However, the effects of stored versus secreted prey defences on bird responses to natural prey and prey survival have never been compared experimentally. The majority of chemically defended insects use only one of the two above-mentioned major defensive strategies, either secreting chemicals from defensive glands or storing defensive compounds in body tissues (Zvereva & Kozlov, 2016). However, some species accumulate defensive chemicals inside the body (e.g. in haemolymph) and at the same time deploy specialized defensive glands; examples include the larvae of Lymantria dispar (Deml, 2003), some species of Saturniidae (Deml, 2003) and Papilionidae (Pinto, Urzua & Niemeyer, 2011; Priestap et al., 2012), and the leaf beetle Oreina cacaliae (Rowell-Rahier et al., 1995). In all these cases, the same defensive compound was identified both in haemolymph and in secretions. By contrast, in some insect species, the stored and secreted defensive compounds differ. The larvae of many leaf beetles (Chrysomelidae) possess specialized defensive glands, in which they accumulate compounds that are usually deterrents for various natural enemies (Pasteels, Braekman & Daloze, 1988). The larvae of the subtribe Chrysomelina were recently found to possess toxic compounds in the haemolymph; these were identified as isoxazolin glucoside and its 3-nitropropanoyl (3-NPA) esters (Pauls et al., 2016). Defence production is usually costly (Bowers, 1992; Ruxton et al., 2004), and thus the maintenance of two different pathways of biosynthesis would be expected to impose a considerable increase in constitutive allocation costs (i.e. costs associated with possessing a chemical defence). Therefore, the cost of possessing two lines of chemical defence would be supported by natural selection only if their combination provides exceptionally effective protection from enemies. The combination of secreted and stored compounds has been recently proposed to contribute synergistically to protection against diverse predators (Pauls et al., 2016). However, this hypothesis has not been tested to date. The strategies of chemical defence can be classified not only by the storage or secretion of chemicals, but also by the mechanism of defence acquisition: some species synthesize defensive chemicals de novo, while other species sequester plant allelochemicals (Pasteels et al., 1983; Bowers, 1992). The ability of herbivores to sequester plant compounds is tightly linked to their host-plant specialization (Bernays & Graham, 1988), so that sequestration is usually associated with highly specialized herbivore species (Zvereva & Kozlov, 2016). Therefore, comparative study of the costs and benefits of these two methods of defence acquisition is important for understanding the role of predators in the evolution of plant–herbivore interactions. For some groups of herbivores (e.g. cyanogenic butterflies), de novo synthesis is considered to be a phylogenetically derived strategy (Brown & Francini, 1990), whereas in leaf beetles of the subtribe Chrysomelina, the sequestration of chemicals is a derived feature while de novo biosynthesis is ancestral (Pasteels et al., 1988; Termonia et al., 2001; Fürstenberg-Hägg et al., 2014). The energetic costs are generally presumed to be higher for de novo synthesis than for sequestration (Rowell-Rahier & Pasteels, 1986; Bowers, 1992; Fürstenberg-Hägg et al., 2014); however, this difference has not been found in a study comparing several leaf beetle species (Zvereva et al., 2017) or in a meta-analysis involving 22 species of chemically defended insect herbivores (Zvereva & Kozlov, 2016). By contrast, several studies on leaf beetles have demonstrated that defences sequestered from the herbivore’s host plant are more effective against birds (Rowell-Rahier et al., 1995), ants (Zvereva, Kruglova & Kozlov, 2010), bacteria and fungi (Gross, Podsiadlowski & Hilker, 2002) than are autogenously produced defences. This suggests that the evolution of sequestration in leaf beetles has been favoured by the high effectiveness of the sequestered compounds against their major predators. Leaf beetles of the subtribe Chrysomelina, whose larvae use two strategies of anti-predatory defence (repellents and toxins: Pauls et al., 2016) and, in addition, are able to both sequester and synthesize the defensive compounds (Schulz, Gross & Hilker, 1997), provide a unique opportunity to test several hypotheses concerning evolution of anti-predator defences for a prey living in a multi-enemy world. In the present study, we have tested the hypothesis that two strategies of chemical defence in leaf beetle larvae – storage of chemicals in haemolymph and their secretion by specialized exocrine glands – have evolved as protection against the two major groups of predators of these beetles: insects and birds. We predict that secretions released from defensive glands would be most effective against insect predators, whereas compounds stored in the larval body would be protective primarily against birds. We have tested these predictions in several experiments aimed at answering the following questions: (1) Do defensive secretions released by Chrysomela lapponica larvae cause their rejection by wood ants and great tits? (2) Are larvae of C. lapponica that are devoid of secretions still toxic and/or unpalatable to ants and birds? (3) Do stored and secreted defences have different effects on avoidance learning and memorability in birds? We also tested the hypothesis that birds may have been responsible for the specialization of leaf beetles on toxic host plants, because secretion of allelochemicals sequestered from host plants provides better protection from birds when compared to autogenously produced secretions. MATERIAL AND METHODS Prey The leaf beetle C. lapponica is a univoltine species widely distributed in the Palaearctic region. Different populations of this leaf beetle feed on species of Salicaceae or Betulaceae. Adults hibernate in soil and start feeding and copulating on host plants soon after leaf flush. Females lay batches of 35–40 eggs on host plant leaves. Larvae feed in groups for about 1 month and pupate on the host plants. Larvae are black and possess nine pairs of eversible glands on the dorsal side of the thorax and abdomen. When disturbed, the larvae release droplets of defensive secretions from these glands for 1–3 s and then withdraw it back into gland reservoirs. The composition of the defensive secretions in C. lapponica larvae depends on the host plant: larvae feeding on Salix myrsinifolia sequester salicylic glucosides (SGs) to produce salicylaldehyde, while larvae feeding on SG-poor species (e.g. Betulaceae and some willow species, such as S. caprea) autogenously produce iso- and 2-methylbutyric esters as major defensive compounds (Hilker & Schulz, 1994; Termonia et al., 2001; Geiselhardt et al., 2015). For this study, we used a population of C. lapponica from the Kola Peninsula (NW Russia), which feeds in nature mostly on SG-rich S. myrsinifolia, although about 20% of the population was recorded on SG-poor willow species (Zvereva, Kozlov & Neuvonen, 1995). We obtained larvae for the experiments from eggs laid by females collected in the surroundings of Monchegorsk (67.94°N, 32.89°E) in June 2015 and 2016. We reared the larvae in the laboratory at room temperature and under a natural photoperiod in 50-mL glass vials on SG-rich S. myrsinifolia and SG-poor S. caprea. Consequently, the reared larvae either sequestered their defensive secretions from their host plants (S-secretions, hereafter) or produced them autogenously (A-secretions, hereafter) (Geiselhardt et al., 2015). In the experiments, we used larvae that had moulted to the last instar and had reached a fresh weight of about 25 mg. We used mealworms (larvae of Tenebrio molitor, Tenebrionidae) in experiments with adult birds and larvae of blow flies (Calliphora sp.; Calliphoridae, maggots, hereafter) in experiments with ants as well as with juvenile birds as palatable control prey to check the foraging motivation of predators. Predators Predation by ants (Hymenoptera: Formicidae) is often considered an important selective force in the evolution of arthropod chemical defences (Pasteels et al., 1983) and therefore ants are frequently used in studies of anti-predatory effects of various defensive compounds (Zvereva & Kozlov, 2016). Wood ants (Formica polyctena) use carbohydrates (secretions of aphids) and protein sources for nutrition and frequently prey on arthropods. Predation on arthropods becomes particularly important when ants actively search for protein sources to feed their larvae (Lenoir, 2002). We selected F. polyctena for our experiments because this species was most abundant in the localities populated by C. lapponica. Moreover, these ants were observed to attack leaf beetle larvae in nature, especially frequently on willows where ants tended aphids (Zvereva, Kozlov & Rank, 2016). Great tits (Parus major) are predominantly insectivorous birds. They have a wide distribution through the Palaearctic and occur over a range of different woodland types (Cramp & Perrins, 1993). They live in the same habitats as C. lapponica and therefore are likely to encounter the leaf beetle in nature. Great tits are frequently used in studies of insect anti-predatory defences (Lindström et al., 1999; Tullberg, Leimar & Gamberale-Stille, 2000; Exnerová et al., 2015) because they can easily adapt to laboratory conditions and to novel foods. We used wild-caught as well as naïve, hand-reared great tits as predators. We captured adult birds using mist-nets in Prague (50.08°N, 14.24°E) during September and October 2015. These birds were housed individually for 2–5 days before the experiments in plastic cages (50 × 40 × 50 cm) with wire-mesh front walls equipped with perches and water bowls for habituation to the laboratory conditions. The light conditions were set according to the outdoor photoperiod. The birds were fed a diet consisting of mealworms, sunflower seeds and a commercial food mixture (Uni Patee, Orlux). Nestlings (12–15 days old) were taken from nestboxes placed in woods at the outskirts of Prague in May 2016. We took a maximum of two juveniles from a single brood. They had only limited visual experience with food brought by their parents and were naïve with respect to any kind of unpalatable prey. Birds were hand-reared until they were able to feed themselves. Their diet consisted of mealworms, a commercial mixture for hand-rearing of passerines (Handmix, Orlux), and a mixture of boiled eggs and biscuits. Nestlings were kept in artificial nestboxes in small groups until fledged. They were then housed in the same cages as wild-caught birds, and provided with mealworms and food mixtures for insectivorous birds (Oké-bird and Nutribird, Versele-Laga; Uni Patee and Insect Patee, Orlux), vitamins, minerals (Roboran, Combisol) and water. Juveniles were tested when they had reached a stage of full independence (after the 35th day of life). Experiments with ants We conducted experiments with ants on 22 June 2016 in a pine forest (67.58°N, 32.55°E) situated 40 km south of Monchegorsk. We selected two large ant nests (mounds) about 20 m apart from each other. No leaf beetle larvae naturally occurred in the surroundings of the nests: the nearest population of C. lapponica was 10 km from the study site. Thus, all ants were naïve with respect to the prey items used in our experiments. We conducted the experiment using a method developed in an earlier study (Zvereva et al., 2016). We tested the motivation of the ants to use protein food near each mound by offering a maggot as a control prey. We placed each experimental item at about 1 m distance from the mound and about 5 cm from an ant trail, and no closer than 20 cm to a previous item and no earlier than 5 min after removal of previous prey; this ensured that different ants took part in the attack on each prey item. One observer (V.Z.) recorded ant behaviour until the prey item was carried to a nest. We obtained quantitative estimates of the effectiveness of prey defences by recording the number of ants that encountered the prey until one attacked it by attempting to bite (i.e. the number of repelled ants), which is frequently used as a measure of defence effectiveness against ants (Zvereva et al., 2016, and references therein). We used the following prey items in the experiment: (1) control prey (maggots); (2) C. lapponica larvae reared on S. myrsinifolia, with intact secretions; (3) similar larvae devoid of secretions (depleted larvae, hereafter); (4) maggots smeared with secretions of C. lapponica collected during secretion depletion; (5) maggots smeared with the body contents of dissected larva; and (6) integument of depleted larvae with viscera removed. We tested ten prey items (five per mound) of each kind in a random order, with one prey item offered during each trial. Different prey items were distributed evenly during the day to account for diurnal changes in ant activity. Autogenously produced secretions of C. lapponica (larvae reared on S. caprea) were compared with sequestered secretions (larvae reared on S. myrsinifolia) in earlier experiments, with the same ant species (Zvereva et al., 2010), and therefore were not tested here. We obtained depleted larvae by disturbing them with forceps to force them to release secretions, which we immediately collected in glass capillaries; we applied these secretions onto a control prey (maggot) to study the effects of secretions per se. We applied secretions collected from a single larva onto one maggot. We tested for the presence of defensive compounds in the integument by dissecting the depleted larvae, removing their viscera and haemolymph with a scalpel, and then cleaning the integument with a small piece of tissue paper, taking care that no traces of haemolymph were left on the surface. We then applied the removed body content (viscera and haemolymph) to maggots to test for the presence of defensive compounds within the larval body. We prepared all items in the field immediately before offering them to ants. Experiments with birds We carried out experiments with birds at the Faculty of Science, Charles University, Prague. The experiments with wild-caught birds were conducted during September–October 2015, and experiments with naïve, hand-reared birds during July 2016. We tested the birds individually in wooden cages (70 × 70 × 70 cm) with wire mesh walls, equipped with a perch, a dish with water and a rotating feeding tray with six cups. We observed the birds through a one-way glass in the front wall of the cage. The cage was illuminated by a daylight-simulating Biolux Combi 18-W bulb (Osram). Before the experiment, the birds were habituated to the experimental cages, were trained to take food from a glass Petri dish placed on a rotating tray, and were deprived of food for 2 h. Prey items were offered on a green background made from a leaf (white beam Sorbus aria in 2015, willow S. caprea in 2016). We recorded all experiments with birds with a video camera and continuously noted the behavioural elements using Observer XT 8.0 (Noldus). All experiments consisted of two parts: (1) an avoidance-learning session and (2) a memory test carried out 1 day after the avoidance-learning, in which we tested whether the birds remembered their experience with a particular prey. The avoidance-learning session consisted of a series of six consecutive trials, in which the bird was offered one prey item per trial: three trials with control prey and three trials with experimental prey, control and experimental trials in turn, starting from the control trial to check bird foraging motivation. If the bird refused to attack a control prey, we repeated the trial until the prey was consumed. Each trial was terminated after the bird ate the prey; otherwise, it lasted 5 min. The memory test consisted of two trials (the first with a control prey and the second with an experimental prey) conducted in the same way as the learning session; we offered the same experimental prey to the bird as had been offered the day before. During each trial we recorded (1) the latency of first attack; (2) whether the prey item was attacked (touched by beak, pecked or seized), killed/damaged (integument broken and larva is non-motile) and eaten by birds; and (3) duration of discomfort behaviour (cleaning the bill, shaking the head, ruffling the feathers). During the trials and several times during the following day, we recorded whether the birds demonstrated any signs of sickness (decreased activity, vomiting). We studied the different components of prey defences separately and in combination by designing three kinds of experiments. Within each experiment, the birds were randomly assigned to experimental subgroups tested with different types of prey. Each bird participated in only one of the experiments and was tested only with one of the prey types. The sex and age of the birds in experimental subgroups were balanced in all experiments. Reactions of birds to the secretions applied on palatable prey We conducted this experiment to study bird responses to sequestered and autogenous defensive secretions per se. We applied approximately 0.2 µL of secretions (corresponding to the amount of secretions produced by one larva: Geiselhardt et al., 2015) onto half a mealworm painted black with an odourless and non-toxic dye (Jovi S.A.157) to give the prey item a visual resemblance to a leaf beetle larva. The secretions were collected from defensive glands of last-instar larvae reared on either S. myrsinifolia or S. caprea into glass capillaries, which were sealed and kept in a freezer at −18 °C until the experiment. We used a pipette to apply secretions from the capillaries onto a mealworm immediately before the trial. We tested 18 birds with each S-secretions and A-secretions. Reactions of adult birds to larval body We studied whether the larval body itself (lacking the secretions) is unpalatable and/or toxic to birds, and whether host plants consumed by the larvae contribute to these effects, using frozen C. lapponica larvae reared either on S. myrsinifolia or on S. caprea. The larvae were depleted before freezing by soaking up the released secretions with a piece of filter paper. We tested 18 birds with each kind of larva. Half a mealworm (which matched the size of a C. lapponica larva) was used as control palatable prey. Reactions of juvenile hand-reared birds to live larvae We conducted this experiment to study the innate responses of naïve birds as well as the process of avoidance learning and memorability of fully defended leaf beetle larvae. We used three kinds of C. lapponica larvae: larvae reared on S. myrsinifolia (S-secretions intact), larvae reared on S. myrsinifolia (S-secretions depleted) and larvae reared on S. caprea (A-secretions intact). We tested 15 birds with each kind of prey. We accounted for a potential effect of the birds’ neophobia using maggots as a control palatable prey that the birds had not encountered before. We avoided the effects of differences in colour between larvae (black) and maggots (white) by painting the maggots with an odourless and non-toxic black dye. Statistical analysis We compared the numbers of ants repelled by different types of prey by ANOVA, followed by Duncan’s post-hoc test. The data were ln-transformed prior to the analysis to meet the assumptions of normality. In all experiments with birds, we compared attack latencies between control prey (mealworms or maggots) and experimental prey using the Wilcoxon Signed Rank test. In the experiment with secretions applied on a control prey, we analysed the differences in attack latencies and in duration of discomfort behaviour between the two types of secretions across trials by repeated-measures ANOVA. These two variables were sqrt(ln(x))-transformed to fit a normal distribution. In the experiment with live larvae, we used the Kruskal–Wallis test for comparison of first attack latencies between the three kinds of larvae. In experiments with live and frozen larvae, we compared the numbers of birds that attacked, killed (damaged) or ate the prey between the different types of larvae and between larvae and the corresponding control prey using the two-sided Fisher’s exact test. We performed all statistical tests using SAS 9.4 (SAS Institute, 2015). Ethical note We obtained permissions for experiments with wild-caught and hand-reared great tits from the Environmental Department of Municipality of Prague (S-MHMP-83637/2014/OZP-VII-3/R-8/F), Ministry of Agriculture (13060/2014-MZE-17214), and Ministry of the Environment of the Czech Republic (42521/ENV/14–2268/630/14). We ringed birds individually and released them back to the locality of capture within a few days after experimentation. RESULTS Ant predation The prey types we used in the field experiment repelled different numbers of ants before the prey was attacked (F5,53 = 12.23, P < 0.0001). All control prey (maggots) were immediately attacked by the first ant that found it (Fig. 1), indicating a high motivation of the ants in the selected nests to collect protein food. In contrast, up to 20 ants were repelled by intact larvae before the first attack was attempted. When an ant approached the intact larva of C. lapponica and touched the prey with its antennae, the larva released droplets of secretions from its glands, and then slowly drew them back into the glands when the disturbance was over. After contact with secretions, the ant usually retreated and cleaned its antennae. After several encounters with ants, accompanied by the release of secretions, the amount of larval secretions decreased, and this allowed for the first ant bite. Subsequently, other nearby ants joined the attack and the prey was quickly killed and transported to the nest. Maggots with larval secretions applied onto their surfaces also repelled ants; the difference in repellence between maggots coated with secretions and intact larvae was not statistically significant (Fig. 1). Larvae with depleted secretions, maggots coated with larval body content and larval integument alone did not demonstrate significant repellence and were attacked as rapidly as control maggots (Fig. 1). Figure 1. View largeDownload slide Effect of chemical defences of larvae of the leaf beetle Chrysomela lapponica on the number of wood ants (Formica polyctena) (mean+SE; each based on ten items) repelled by different types of prey in the field experiment. LS: larva of C. lapponica with intact sequestered secretions (S-secretions); CS: maggot coated with S-secretions; LD: larva of C. lapponica with depleted S-secretions; CH: maggot coated with larval haemolymph; LI: depleted larva with viscera removed; C: maggot of the same size as a larva (control). Different letters above bars indicate significant (P < 0.05) differences between prey types (Duncan test). Figure 1. View largeDownload slide Effect of chemical defences of larvae of the leaf beetle Chrysomela lapponica on the number of wood ants (Formica polyctena) (mean+SE; each based on ten items) repelled by different types of prey in the field experiment. LS: larva of C. lapponica with intact sequestered secretions (S-secretions); CS: maggot coated with S-secretions; LD: larva of C. lapponica with depleted S-secretions; CH: maggot coated with larval haemolymph; LI: depleted larva with viscera removed; C: maggot of the same size as a larva (control). Different letters above bars indicate significant (P < 0.05) differences between prey types (Duncan test). Reactions of birds to the secretions applied onto palatable prey Of the 108 mealworms with applied C. lapponica secretions, only three (all with S-secretions) were not attacked during the experiment and five were rejected after attack without being eaten (two with A-secretions and three with S-secretions). The two types of secretions differed in some of the effects they had on bird behaviour. The latencies of attacks on mealworms coated with secretions were longer than those on control mealworms in the first trial for both S-secretions (S = 55.5, P = 0.007) and A-secretions (S = 59, P = 0.008) and approached significance in the second trial for S-secretions (S = 40.5, P = 0.056), but not for A-secretions (S = 31.5, P = 0.18). In the third trial, the differences between secretion-coated and control mealworms were not statistically significant (S-secretions: S = 33, P = 0.09; A-secretions: S = 20, P = 0.40) (Fig. 2). Across all the trials (interaction between secretion type and trial number: F2,68 = 0.77, P = 0.47), birds hesitated for a longer period before attacking mealworms coated with S-secretions than before attacking mealworms coated with A-secretions (repeated-measures ANOVA: F1,34 = 3.96, P = 0.05). The duration of discomfort behaviour while handling and after eating the prey decreased with the sequence number of the trial (F2,56 = 15.66, P < 0.0001) and was longer for birds that came into contact with S-secretions than with A-secretions (F1,28 = 4.75, P = 0.038) (Fig. 3); the difference between secretion types did not depend on the trial number (interaction between secretion type and trial number: F2,56 = 0.94, P = 0.40). Figure 2. View largeDownload slide Latencies of attacks of wild-caught adult great tits on mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica when compared with control mealworms during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Points indicate medians, boxes indicate interquartile ranges and whiskers show non-outlier ranges. Asterisks indicate significant (P < 0.05) differences between latencies of attacks on control mealworms and on mealworms with secretions applied (Wilcoxon Signed Rank test). Figure 2. View largeDownload slide Latencies of attacks of wild-caught adult great tits on mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica when compared with control mealworms during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Points indicate medians, boxes indicate interquartile ranges and whiskers show non-outlier ranges. Asterisks indicate significant (P < 0.05) differences between latencies of attacks on control mealworms and on mealworms with secretions applied (Wilcoxon Signed Rank test). Figure 3. View largeDownload slide Duration of discomfort behaviour of wild-caught adult great tits during handling and eating of mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Bars indicate means + SE; asterisks indicate significant (P < 0.05) differences between two types of secretions (repeated-measures ANOVA). Figure 3. View largeDownload slide Duration of discomfort behaviour of wild-caught adult great tits during handling and eating of mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Bars indicate means + SE; asterisks indicate significant (P < 0.05) differences between two types of secretions (repeated-measures ANOVA). Reactions of birds to a larval body lacking secretions First attack latencies were longer toward dead depleted C. lapponica larvae than toward mealworms (S = 356, P = 0.002). After the first attack (in the second and third trials), the numbers of birds that attacked the larvae were significantly reduced when compared to control mealworms (A-secretion larvae: P = 0.003; S-secretion larvae: P < 0.0001) and to the first trial with larvae (A-secretion larvae: P < 0.0001; S-secretion larvae: P < 0.0001), indicating that the birds also learned to avoid the larvae in the absence of released secretions. The numbers of attacking birds did not differ between the groups tested with the two types of larvae (P = 0.46 and P = 1 for the second and third trial, respectively) (Fig. 4). In the memory test, the numbers of attacking birds remained nearly the same as in the last trial of the learning session (P = 1 for both types of larvae) and these numbers did not differ between the two types of larvae (P = 1) (Fig. 4). Figure 4. View largeDownload slide Numbers of wild-caught adult great tits that attacked freshly defrosted Chrysomela lapponica larvae depleted of either sequestered (sample size: 15 birds) or autogenous (14 birds) secretions during three consecutive trials in an avoidance-learning session and in a memory test. Control prey novel to the birds (maggot) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). Figure 4. View largeDownload slide Numbers of wild-caught adult great tits that attacked freshly defrosted Chrysomela lapponica larvae depleted of either sequestered (sample size: 15 birds) or autogenous (14 birds) secretions during three consecutive trials in an avoidance-learning session and in a memory test. Control prey novel to the birds (maggot) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). Of 13 birds that attacked both types of larvae during the first encounter without damaging the integument, ten (76.9%) avoided the larvae during the following encounters, whereas only six of 15 (40%) birds that ate the larvae during the first encounter rejected them later on. None of the birds that consumed the larvae showed any signs of sickness afterwards. Reactions of birds to live larvae The birds attacked all types of C. lapponica larvae during the first encounter at the same rate as they attacked a novel palatable prey (Fig. 5). However, the latencies of attacks on C. lapponica larvae were longer for all types of larvae even during the first encounter when compared to control maggots (depleted larvae: S = 38, P = 0.03; A-secretion larvae: S = 52, P = 0.002; S-secretion larvae: S = 47, P = 0.005). The differences in the latency between the three groups of larvae were not statistically significant (χ2 = 0.95, df = 2, P = 0.62). Figure 5. View largeDownload slide Numbers of hand-reared juvenile great tits that attacked Chrysomela lapponica larvae differing in chemical defence (sequestered, autogenous or depleted; sample size: 15 birds in each group) during three consecutive trials in an avoidance-learning session and in a memory test. Control prey (maggots) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). Figure 5. View largeDownload slide Numbers of hand-reared juvenile great tits that attacked Chrysomela lapponica larvae differing in chemical defence (sequestered, autogenous or depleted; sample size: 15 birds in each group) during three consecutive trials in an avoidance-learning session and in a memory test. Control prey (maggots) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). The number of birds that attacked the larvae in the second trial was considerably lower for S-secretion larvae, both intact and depleted (P = 0.02 for both groups), when compared with attacks on control maggots, but not for A-secretion larvae (P = 0.22, Fig. 5). In the third trial, the numbers of birds attacking the larvae were lower for S-secretion larvae than for either depleted larvae (P = 0.03) or A-secretion larvae (P = 0.03), whereas the probability of attacks was the same for depleted and A-secretion larvae (P = 1). In the memory test, all types of larvae were attacked at a reduced rate when compared to the control palatable prey, but the S-secretion larvae still differed marginally (P = 0.06) from the depleted larvae (Fig. 5). Some larvae that were attacked by birds were released unharmed and survived the attack. Survival of attack was considerably lower for the depleted larvae than for A-secretion larvae (21.6% and 47.2%, respectively, P = 0.03) and S-secretion larvae (21.6% and 67.8%, respectively, P = 0.0003), but survival did not differ between A-secretion and S-secretion larvae (P = 0.13). However, survival in the last trial was higher for S-secretion larvae than for A-secretion larvae due to a decrease in the attack rate (P = 0.04). In the memory test, larval survival did not differ between S-secretion larvae and A-secretion larvae (P = 0.65), but it was higher for larvae with intact secretions (both types combined) than for larvae with depleted secretions (P = 0.017). Of 11 birds that avoided S-secretion larvae in the memory test, ten attacked but did not damage the larvae encountered in the learning trials. Birds ate a significantly lower proportion of larvae with intact secretions when compared with depleted larvae (S-secretion larvae: P = 0.006; A-secretion larvae: P = 0.0004), and only 50% of the depleted larvae were eaten compared to 100% of the control maggots (P < 0.0001). The S-secretion larvae and A-secretion larvae were eaten at the same rate by the birds (P = 0.55). Of the 16 birds that attacked but did not kill any larvae (and thus had no contact with larval body contents) during the avoidance-learning session, 15 avoided attacking the larvae in the memory test. By contrast, of 15 other birds that ate at least one larva, only four avoided attacking them in the memory test (P = 0.05). None of the 17 juvenile birds that partly or completely consumed the larvae showed any signs of sickness during the experiment or the day after. DISCUSSION Effects of leaf beetle defensive secretions on ants and birds Our study provided a direct estimation of the relative effectiveness of different defence strategies of leaf beetle larvae against insect and avian predators. Moreover, detailed observations of predator behaviour revealed the mechanisms by which these defences provide protection from different predators. One suggestion has been that volatile secretions released by the prey are directed mostly against invertebrate predators, and against ants in particular (Boevé & Pasteels, 1985; Boevé et al., 2013). Consistently, ants have been suggested to represent a major selective agent for secreted chemical defences in leaf beetle larvae (Pasteels et al., 1983), whereas the role of birds in the evolution of volatile defences has remained unclear (Pauls et al., 2016). In agreement with an earlier study (Zvereva et al., 2010), we showed that salicylaldehyde-containing secretions released by leaf beetle larvae in response to disturbance by wood ants, and even by their approach, had a repellent effect on the ants. By contrast, larvae with depleted secretions were attacked by ants at the same rate as control prey. The latter result supports the hypothesis that the release of secretions is the only factor responsible for the protection of leaf beetle larvae against ants. However, wood ants collectively foraging in the field overcome the defensive effects of secretions of C. lapponica larvae by social interactions, chemical signalling and olfactory learning, thereby imposing a high mortality on larvae near ant nests (Zvereva et al., 2016). However, the repellence of secretions favours the survival of leaf beetle larvae when ant density is low. Salicylaldehyde-containing secretions were also found to repel some other generalist arthropod predators, such as true bugs (Rank et al., 1998) and ladybirds (Denno, Larsson & Olmstead, 1990), at the same time serving as a search cue for specialist predators and parasitoids (Köpf et al., 1997; Gross et al., 2004; Zvereva & Rank, 2004). Our results show that volatile secretions of C. lapponica provide a direct defensive effect not only against generalist arthropod predators, but also against insectivorous birds, and that the secretions affect bird behaviour in several ways. First, although the birds attacked C. lapponica larvae during the first encounter at the same rate as they attacked a control palatable prey, they hesitated longer before attacking the larvae than when attacking the control prey. This behaviour could not be explained by increased wariness elicited by a visually novel prey (Marples & Kelly, 1999; Marples et al., 2007) or an innate bias (Lindström et al., 1999) against its conspicuous coloration (black on a green background), because the control prey was also novel (and also painted black) in the experiment with naïve juvenile birds. Attack latencies were increased in comparison with palatable prey in experiments with live C. lapponica larvae, as well as when the secretions were applied onto control prey; therefore, we suggest that this effect may be explained by an innate bias against novel odours (Jetz, Rowe & Guilford, 2001). Novel odours are known to elicit hidden innate biases against visually novel and/or conspicuous prey (Marples & Roper, 1996; Rowe & Guilford, 1996; Lindström, Rowe & Guilford, 2001); therefore, a combined effect of visual and olfactory cues may be responsible for the longer attack latencies observed in our experiment. Second, defensive secretions considerably increased the chance that larvae would survive bird attacks: the birds released, alive and unharmed, 59% of the attacked larvae with intact secretions but only 21% of attacked larvae with removed secretions. This was presumably due to the contact deterrent (irritating) effect of the secretions, which forced the birds to quickly drop the attacked prey. The irritating effect of secretions was also evident from the discomfort behaviour of birds after attacking mealworms coated with secretions. These results are in accordance with other studies showing that birds taste and reject chemically defended prey (Skelhorn & Rowe, 2006a,b; Halpin & Rowe, 2010), which can then be released relatively unharmed (Wiklund & Jarvi, 1982; Sillen-Tullberg, 1985; Skelhorn & Rowe, 2009; Hotová Svádová et al., 2013). The fact that a considerable proportion of the prey population can survive predator attack is important for understanding the pathways of chemical defence evolution in insects with externally secreted defences. It also confirms the importance of individual selection in this process (Skelhorn & Rowe, 2006c). Third, attacks on S-secretion larvae, even those that did not involve killing and eating the larvae, induced more effective avoidance learning than did attacks on larvae lacking the secretions. This effect may be explained either by high aversiveness of S-secretions for birds (Lindström, Alatalo & Mappes, 1997; Lindström et al., 2006) or by their potential role as chemical (olfactory or gustatory) signals that the birds associate with prey noxiousness (Guilford et al., 1987; Skelhorn & Rowe, 2005, 2009; Siddall & Marples, 2008). The reduced survival of larvae with depleted secretions in our experiments provides evidence for the high effectiveness of volatile secretions of C. lapponica larvae against bird predators. The secretions enhance innate wariness in bird predators, induce rapid avoidance learning and increase the chance that the prey will survive the attack. Differences between sequestered and autogenous secretions Our field experiment showed that sequestered secretions of C. lapponica larvae effectively repelled wood ants. Earlier experiments that used the same method have also demonstrated that the autogenously produced secretions are repellent for ants (Zvereva et al., 2010). This result is in accordance with laboratory experiments that demonstrated repellent effects of both kinds of secretions on ants (Blum et al., 1972; Hilker & Schulz, 1994). However, secretions containing sequestered salicylaldehyde appeared more effective than autogenously produced secretions containing butyric esters, as they repelled significantly more ants and thus allowed the larvae to survive longer near the ant nest (Zvereva et al., 2010). We also found differences between the effects of sequestered and autogenous secretions in our experiments with birds. First, the birds learned to avoid sequestered secretions much faster when compared to autogenous secretions: S-secretion larvae were avoided even after first attack, whereas all three A-secretion larvae offered during the learning session were attacked at the same rate. Second, contact with sequestered secretions elicited a stronger discomfort reaction (head shaking and beak wiping) in birds than did the contact with the autogenous secretions. These results indicate stronger aversiveness of sequestered secretions than of autogenously produced secretions. Third, sequestered secretions, when applied onto control prey (mealworms), resulted in longer initial attack latencies in the first trial than we observed for autogenous secretions applied in the same way. The absence of a similar effect in an experiment with live larvae probably reflects the fact that the larvae usually discharge their secretions only when attacked. However, in a natural situation, when larvae are closely aggregated, the release of secretions by a single attacked individual may provide increased protection for the group and enhance the dilution effect (Riipi et al., 2001). Once a larva was attacked, we saw no effect of the secretion type on its survival. Both secretions were quite effective and resulted in high survival rates of larvae relative to control prey: about half of attacked larvae were released unharmed. However, the combined effect of the secretions on attack latencies and avoidance learning led to better overall survival of S-secretion larvae compared to A-secretion larvae. At the same time, the experiment with frozen depleted larvae showed no differences between S-secretion larvae and A-secretion larvae, indicating that differences in defence effectiveness between these two kinds of larvae depend exclusively on the composition and/or the amount of secretions. Along with these differences, we found some similarities in the effects of sequestered and autogenous secretions. Larvae with both types of secretions survived attacks by birds at the same rate, and the birds remembered their negative experience equally well for both kinds of defences, as indicated by the similar survival of larvae in the memory test. Thus, both types of secretions provide effective protection against birds. Effects of defensive substances contained in body tissues The 3-NPA esters, which were discovered in the haemolymph of larvae of species belonging to the subtribe Chrysomelina, were found to be a deterrent for ants (Pasteels, Daloze & Rowell-Rahier, 1986; Sugeno & Matsuda, 2002). However, we did not detect any deterrent effect of body content and integument of C. lapponica larvae against wood ants. A similar contradiction was also reported by Reudler et al. (2015): in their experiments, wood ants preferred a solution extracted from the larvae of Parasemia plantaginis over sugar water, although pure defensive compounds found in haemolymph (iridoid glycosides) were deterrent. The authors hypothesized that the attractiveness of nutrients (e.g. proteins) contained in the prey overcomes potential deterrent effects of defensive chemicals (Reudler et al., 2015). We therefore suggest that only externally released secretions provide sufficient protection for C. lapponica larvae against wood ants. Our experiments with great tits indicate that the secretions alone are not what makes the larvae unpalatable: avoidance was also developed against frozen larvae and against live larvae devoid of secretions. Moreover, not only larvae with secretions but also depleted larvae were attacked at a reduced rate in the memory test when compared with the control palatable prey. This indicates that factors other than secretions contribute to the effectiveness of larval defences against bird predators. One of these factors could be the non-volatile isoxazolinone-5-one glucoside and its 3-NPA ester, which are synthesized autogenously during larval life (Pauls et al., 2016). The latter compound is a neurotoxin that causes poisoning of humans and domestic livestock (Beal et al., 1993; Anderson et al., 2005). The effect of this compound on birds has not been studied before, and our results indicate that the amount contained in one larva is not sufficient to cause any signs of poisoning in a bird the size of a great tit. Even three larvae eaten completely, one after another, do not elicit any signs of sickness. If many larvae need to be eaten to cause toxic effects in bird predators, this raises doubts that body toxins could have evolved under selection pressure imposed by birds. Our results also indicate that the factors contributing to larval defence (other than secretions) act upon contact with the larval surface rather than with the body contents, because consumption or even damaging the prey was not necessary for the development of avoidance. The integument of C. lapponica was not analysed separately by Pauls et al. (2016), but our results suggest that defensive compounds are probably deposited in considerable amounts in integument, as even the contact with a larval surface lacking secretions contributed to avoidance learning. Considerable amounts of defensive chemicals were detected in the integument when it has been analysed separately (Montllor, Bernays & Cornelius, 1991; Fürstenberg-Hägg et al., 2014). This was demonstrated for arctiid moths, which accumulated pyrrolizidine alkaloids in the larval integument (von Nickisch-Rosenegk & Wink, 1993), and for the larvae of a papilionid butterfly, Battus polydamas, which uses the integument as the major site of aristolochic acid accumulation (Priestap et al., 2012). These examples indicate that deposition of defensive compounds within the integument is more widespread than previously thought, because it may increase prey survival upon predator attack owing to taste rejection by the predator before the prey is mortally damaged (Skelhorn & Rowe, 2006a; Zvereva & Kozlov, 2016). Thus, externalized defensive compounds of prey seem to be more important for avoidance learning in avian predators than are the defensive compounds stored inside the prey body. This conclusion concurs with the results of previous experimental studies (Skelhorn & Rowe, 2006c; Hotová Svádová et al., 2013) and with the conclusions of a meta-analysis comparing the effectiveness of externally secreted defensive chemicals with chemicals stored in body tissues against vertebrate predators (Zvereva & Kozlov, 2016). The advantages of externalized chemical defences can be illustrated by the existence of numerous externalization mechanisms, which have independently and repeatedly evolved in several insect taxa (Ohkuma et al., 2004). The low effectiveness of defensive compounds contained inside the body for anti-predatory protection of C. lapponica larvae suggests some other biological significance of these compounds. The isoxalinone derivatives are major components of adult defences in the subtribe Chrysomelina and are released from the beetle exocrine glands upon disturbance (Deroe & Pasteels, 1982; Sugeno & Matsuda, 2002; Pasteels et al., 2003), thereby serving as an efficient externalized defence (Sugeno & Matsuda, 2002). These compounds, synthesized autogenously by larvae (Pauls et al., 2016), can be transferred from larvae to adults. This passing of the accumulated larval defensive compounds through metamorphosis has been demonstrated for many insect species, and for chemicals both sequestered by larvae from their food plants and biosynthesized de novo, such as aristolochic acid (Sime, Feeny & Haribal, 2000), pyrrolizidine alkaloids (Rossini et al., 2003) and cyanogenic glucosides (Fürstenberg-Hägg et al., 2014). This strategy allows adults to conserve energy required for reproduction. CONCLUSIONS We have found a strong repellent effect of exocrine volatile secretions of larvae of C. lapponica on both insect (ants) and avian (great tits) predators. While the aversive effects of these secretions on insect predators have already been reported, we have provided the first evidence that the two kinds of larval secretions of Chrysomelina leaf beetles that differ in their chemistry and origin affect bird behaviour and increase prey survival through both taste rejection and enhanced avoidance learning. Nevertheless, salicylaldehyde-containing secretions sequestered from host plants were more effective against both ants and birds than were autogenously produced butyrate-based secretions. This suggests that both insect and avian predators could have contributed to the evolution of the ability to sequester salicylic glucosides from salicaceous host plants in leaf beetles. We have demonstrated that larvae devoid of secretions, although not toxic, are still unpalatable for birds, confirming the anti-predatory function of non-volatile compounds within the larval body. We also found that the two lines of defence – secretion from glands and storage in the body – act jointly against bird predation. However, we failed to detect any defensive effects of larval body content against ants. We conclude that insect predators could have contributed to the evolution of secreted defensive chemicals, but not of stored chemical defences, whereas bird predation could have favoured the evolution of both lines of defences. ACKNOWLEDGEMENTS We are grateful to two anonymous reviewers for their helpful comments on an earlier draft of the manuscript. This work was supported by the Academy of Finland (project 268124) and Czech Science Foundation (P505/11/1459). The authors do not have any conflicts of interest. REFERENCES Aldrich JR. 1988. Chemical ecology of the Heteroptera. Annual Reviews in Entomology  33: 211– 238. Google Scholar CrossRef Search ADS   Anderson RC, Majak W, Rassmussen MA, Callaway TR, Beier RC, Nisbet DJ, Allison MJ. 2005. Toxicity and metabolism of the conjugates of 3-nitropropanol and 3-nitropropionic acid in forages poisonous to livestock. Journal of Agricultural and Food Chemistry  53: 2344– 2350. Google Scholar CrossRef Search ADS   Beal MF, Brouillet E, Jenkins BG, Ferrante RJ, Kowall NW, Miller JM, Storey E, Srivastava R, Rosen BR, Hyman BT. 1993. Neurochemical and histologic characterization of striatal excitotoxic lesions produced by the mitochondrial toxin 3-nitropropionic acid. The Journal of Neuroscience  13: 4181– 4192. Google Scholar CrossRef Search ADS   Bernays E, Graham M. 1988. On the evolution of host specificity in phytophagous arthropods. Ecology  69: 886– 892. Google Scholar CrossRef Search ADS   Blum MS, Brand JM, Wallace JB, Fales HM. 1972. Chemical characterization of defensive secretion of a chrysomelid larva. Life Sciences. Part 2, Biochemistry, General & Molecular Biology  11: 525– 531. Blum MS. 1981. Chemical defenses of arthropods . London: Academic Press. Boevé JL, Pasteels JM. 1985. Modes of defense in nematine sawfly larvae: efficiency against ants and birds. Journal of Chemical Ecology  11: 1019– 1036. Google Scholar CrossRef Search ADS   Boevé JL, Blank SM, Meijer G, Nyman T. 2013. Invertebrate and avian predators as drivers of chemical defensive strategies in tenthredinid sawflies. BMC Evolutionary Biology  13: 198. Google Scholar CrossRef Search ADS   Bowers MD. 1992. The evolution of unpalatability and the cost of chemical defense in insects. In: Roitberg BD, Isman MB, eds. Insect chemical ecology: an evolutionary approach . New York: Chapman and Hall, 216– 244. Brown KSJr, Francini RB. 1990. Evolutionary strategies of chemical defense in aposematic butterflies: cyanogenesis in Asteraceae-feeding American Acraeinae. Chemoecology  1: 52– 56. Google Scholar CrossRef Search ADS   Chomcheon P, Wiyakrutta S, Sriubolmas N, Ngamrojanavanich N, Isarangkul D, Kittakoop P. 2005. 3-Nitropropionic acid (3-NPA), a potent antimycobacterial agent from endophytic fungi: is 3-NPA in some plants produced by endophytes? Journal of Natural Products  68: 1103– 1105. Google Scholar CrossRef Search ADS   Conner WE, Alley KM, Barry JR, Harper AE. 2007. Has vertebrate chemesthesis been a selective agent in the evolution of arthropod chemical defenses? The Biological Bulletin  213: 267– 273. Google Scholar CrossRef Search ADS   Cramp S, Perrins C. 1993. The birds of the Western Palearctic, Vol. 8–9 . Oxford: Oxford University Press. Deml R. 2003. Pyrrolidonyl and pyridyl alkaloids in Lymantria dispar. Zeitschrift fur Naturforschung. C, Journal of Biosciences  58: 860– 866. Denno RF, Larsson S, Olmstead KL. 1990. Role of enemy-free space and plant quality in host-plant selection by willow beetles. Ecology  71: 124– 137. Google Scholar CrossRef Search ADS   Deroe C, Pasteels JM. 1982. Distribution of adult defense glands in chrysomelids (Coleoptera: Chrysomelidae) and its significance in the evolution of defense mechanisms within the family. Journal of Chemical Ecology  8: 67– 82. Google Scholar CrossRef Search ADS   Eisner T, Meinwald J. 1966. Defensive secretions of arthropods. Science  153: 1341– 1350. Google Scholar CrossRef Search ADS   Eisner T, Eisner M, Siegler M. 2007. Secret weapons: defenses of insects, spiders, scorpions, and other many-legged creatures . Cambridge: Belknap Press. El-Sayed AM. 2014. The pherobase: database of pheromones and semiochemicals . Avaliable at: http://www.pherobase.com (accessed 26 January 2018). Exnerová A, Ježová D, Štys P, Doktorovová L, Rojas B, Mappes J. 2015. Different reactions to aposematic prey in 2 geographically distant populations of great tits. Behavioral Ecology  26: 1361– 1370. Google Scholar CrossRef Search ADS   Francis K, Smitherman C, Nishino SF, Spain JC, Gadda G. 2013. The biochemistry of the metabolic poison propionate 3-nitronate and its conjugate acid, 3-nitropropionate. IUBMB Life  65: 759– 768. Google Scholar CrossRef Search ADS   Fürstenberg-Hägg J, Zagrobelny M, Olsen CE, Jørgensen K, Møller BL, Bak S. 2014. Transcriptional regulation of de novo biosynthesis of cyanogenic glucosides throughout the life-cycle of the burnet moth Zygaena filipendulae (Lepidoptera). Insect Biochemistry and Molecular Biology  49: 80– 89. Google Scholar CrossRef Search ADS   Gamberale-Stille G, Tullberg BS. 1999. Experienced chicks show biased avoidance of stronger signals: an experiment with natural colour variation in live aposematic prey. Evolutionary Ecology  13: 579– 589. Google Scholar CrossRef Search ADS   Geiselhardt S, Hilker M, Müller F, Kozlov MV, Zvereva EL. 2015. Inter- and intrapopulation variability in the composition of larval defensive secretions of willow-feeding populations of the leaf beetle Chrysomela lapponica. Journal of Chemical Ecology  41: 276– 286. Google Scholar CrossRef Search ADS   Gross J, Podsiadlowski L, Hilker M. 2002. Antimicrobial activity of exocrine glandular secretion of Chrysomela larvae. Journal of Chemical Ecology  28: 317– 331. Google Scholar CrossRef Search ADS   Gross J, Fatouros NE, Neuvonen S, Hilker M. 2004. The importance of specialist natural enemies for Chrysomela lapponica in pioneering a new hostplant. Ecological Entomology  29: 584– 593. Google Scholar CrossRef Search ADS   Guilford T, Nicol C, Rothschild M, Moore BP. 1987. The biological roles of pyrazines − evidence for a warning odor function. Biological Journal of the Linnean Society  31: 113– 128. Google Scholar CrossRef Search ADS   Halpin CG, Rowe C. 2010. Taste-rejection behaviour by predators can promote variability in prey defences. Biology Letters  6: 617– 619. Google Scholar CrossRef Search ADS   Higginson AD, Delf J, Ruxton GD, Speed MP. 2011. Growth and reproductive costs of larval defence in the aposematic lepidopteran Pieris brassicae. Journal of Animal Ecology  80: 384– 392. Google Scholar CrossRef Search ADS   Hilker M, Schulz S. 1994. Composition of larval secretion of Chrysomela lapponica (Coleoptera, Chrysomelidae) and its dependence on host plant. Journal of Chemical Ecology  20: 1075– 1093. Google Scholar CrossRef Search ADS   Hotová Svádová K, Exnerová A, Kopecková M, Štys P. 2013. How do predators learn to recognize a mimetic complex: experiments with naive great tits and aposematic Heteroptera. Ethology  119: 814– 830. Google Scholar CrossRef Search ADS   Huang LS, Sun G, Cobessi D, Wang AC, Shen JT, Tung EY, Anderson VE, Berry EA. 2006. 3-Nitropropionic acid is a suicide inhibitor of mitochondrial respiration that, upon oxidation by complex II, forms a covalent adduct with a catalytic base arginine in the active site of the enzyme. Journal of Biological Chemistry  281: 5965– 5972. Google Scholar CrossRef Search ADS   Jetz W, Rowe C, Guilford T. 2001. Non-warning odors trigger innate color aversions − as long as they are novel. Behavioral Ecology  12: 134– 139. Google Scholar CrossRef Search ADS   Kaupp UB. 2010. Olfactory signalling in vertebrates and insects: differences and commonalities. Nature Reviews Neuroscience  11: 188– 200. Google Scholar CrossRef Search ADS   Köpf A, Rank NE, Roininen H, Tahvanainen J. 1997. Defensive larval secretions of leaf beetles attract a specialist predator Parasyrphus nigritarsis. Ecological Entomology  22: 176– 183. Google Scholar CrossRef Search ADS   Lagadic L, Caquet T. 1998. Invertebrates in testing of environmental chemicals: are they alternatives? Environmental Health Perspectives  106: 593– 611. Google Scholar CrossRef Search ADS   Lenoir L. 2002. Can wood ants distinguish between good and bad food patches on the forest floor? European Journal of Soil Biology  38: 97– 102. Google Scholar CrossRef Search ADS   Lindström L, Alatalo RV, Mappes J. 1997. Imperfect Batesian mimicry − the effects of the frequency and the distastefulness of the model. Proceedings of the Royal Society of London. Series B, Biological Sciences  264: 149– 153. Google Scholar CrossRef Search ADS   Lindström L, Alatalo RV, Mappes J, Riipi M, Vertainen L. 1999. Can aposematic signals evolve by gradual change? Nature  397: 249– 251. Google Scholar CrossRef Search ADS   Lindström L, Rowe C, Guilford T. 2001. Pyrazine odour makes visually conspicuous prey aversive. Proceedings of the Royal Society of London. Series B, Biological Sciences  268: 159– 162. Google Scholar CrossRef Search ADS   Lindström L, Lyytinen A, Mappes J, Ojala K. 2006. Relative importance of taste and visual appearance for predator education in Müllerian mimicry. Animal Behaviour  72: 323– 333. Google Scholar CrossRef Search ADS   Marples NM, Roper TJ. 1996. Effects of novel colour and smell on the response of naive chicks towards food and water. Animal Behaviour  51: 1417– 1424. Google Scholar CrossRef Search ADS   Marples NM, Kelly DJ. 1999. Neophobia and dietary conservatism: two distinct processes? Evolutionary Ecology  13: 641– 653. Google Scholar CrossRef Search ADS   Marples NM, Quinlan M, Thomas RJ, Kelly DJ. 2007. Deactivation of dietary wariness through experience of novel food. Behavioral Ecology  18: 803– 810. Google Scholar CrossRef Search ADS   Montllor CB, Bernays EA, Cornelius ML. 1991. Responses of two hymenopteran predators to surface chemistry of their prey: significance for an alkaloid-sequestering caterpillar. Journal of Chemical Ecology  17: 391– 399. Google Scholar CrossRef Search ADS   Nishida R. 2002. Sequestration of defensive substances from plants by Lepidoptera. Annual Review of Entomology  47: 57– 92. Google Scholar CrossRef Search ADS   Ohkuma M, Yuzawa H, Amornsak W, Sornnuwat Y, Takematsu Y, Yamada A, Vongkaluang C, Sarnthoy O, Kirtibutr N, Noparatnaraporn N, Kudo T, Inoue T. 2004. Molecular phylogeny of Asian termites (Isoptera) of the families Termitidae and Rhinotermitidae based on mitochondrial COII sequences. Molecular Phylogenetics and Evolution  31: 701– 710. Google Scholar CrossRef Search ADS   Pasteels JM, Grégoire JC, Rowell-Rahier M. 1983. The chemical ecology of defense in arthropods. Annual Reviews in Entomology  28: 263– 289. Google Scholar CrossRef Search ADS   Pasteels JM, Daloze D, Rowell-Rahier M. 1986. Chemical defense in chrysomelid eggs and neonate larvae. Physiological Entomology  11: 29– 37. Google Scholar CrossRef Search ADS   Pasteels JM, Braekman J-C, Daloze D. 1988. Chemical defense in the Chrysomelidae. In: Solivet P, Petitpierre E, Hsiao TH, eds. Biology of Chrysomelidae . Dordrecht: Kluwer, 233– 252. Google Scholar CrossRef Search ADS   Pasteels JM, Termonia A, Daloze D, Windsor DM. 2003. Distribution of toxins in chrysomeline leaf beetles: possible taxonimic inferences. In: Furth DG, ed. Proceedings of the Fifth International Symposium on the Chrysomelidae, 25–27 August 2000, Iguassu Falls, Brazil, XXI International Congress of Entomology. Sofia: Pensoft Publishers, 261– 275. Pauls G, Becker T, Rahfeld P, Gretscher RR, Paetz C, Pasteels J, von Reuss SH, Burse A, Boland W. 2016. Two defensive lines in juvenile leaf beetles; esters of 3-nitropropionic acid in the hemolymph and aposematic warning. Journal of Chemical Ecology  42: 240– 248. Google Scholar CrossRef Search ADS   Pinto CF, Urzua A, Niemeyer HM. 2011. Sequestration of aristolochic acids from meridic diets by larvae of Battus polydamas archidamas (Papilionidae: Troidini). European Journal of Entomology  108: 41– 45. Google Scholar CrossRef Search ADS   Priestap HA, Velandia AE, Johnson JV, Barbieri MA. 2012. Secondary metabolite uptake by the Aristolochia-feeding papilionoid butterfly Battus polydamas. Biochemical Systematics and Ecology  40: 126– 137. Google Scholar CrossRef Search ADS   Rank NE, Kopf A, Julkunen-Tiitto R, Tahvanainen J. 1998. Host preference and larval performance of the salicylate-using leaf beetle Phratora vitellinae. Ecology  79: 618– 631. Google Scholar CrossRef Search ADS   Reudler JH, Lindstedt C, Pakkanen H, Lehtinen I, Mappes J. 2015. Costs and benefits of plant allelochemicals in herbivore diet in a multi enemy world. Oecologia  179: 1147– 1158. Google Scholar CrossRef Search ADS   Riipi M, Alatalo RV, Lindström L, Mappes J. 2001. Multiple benefits of gregariousness cover detectability costs in aposematic aggregations. Nature  413: 512– 514. Google Scholar CrossRef Search ADS   Rossini C, Bezzerides A, Gonzalez A, Eisner M, Eisner T. 2003. Chemical defense: incorporation of diet-derived pyrrolizidine alkaloid into the integumental scales of a moth (Utetheisa ornatrix). Chemoecology  13: 199– 205. Google Scholar CrossRef Search ADS   Rowe C, Guilford T. 1996. Hidden colour aversions in domestic clicks triggered by pyrazine odours of insect warning displays. Nature  383: 520– 522. Google Scholar CrossRef Search ADS   Rowell-Rahier M, Pasteels JM. 1986. Economics of chemical defense in Chrysomelinae. Journal of Chemical Ecology  12: 1189– 1203. Google Scholar CrossRef Search ADS   Rowell-Rahier M, Pasteels JM, Alonso-Mejia A, Brower LP. 1995. Relative unpalatability of leaf beetles with either biosynthesized or sequestered chemical defense. Animal Behaviour  49: 709– 714. Google Scholar CrossRef Search ADS   Ruxton GD, Sherratt TN, Speed MP. 2004. Avoiding attack: the evolutionary ecology of crypsis, warning signals, and mimicry . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   SAS Institute. 2015. SAS/Stat. User’s guide, version 9.4. Cary: SAS Institute. Schulz S, Gross J, Hilker M. 1997. Origin of the defensive secretion of the leaf beetle Chrysomela lapponica. Tetrahedron  53: 9203– 9212. Google Scholar CrossRef Search ADS   Siddall EC, Marples NM. 2008. Better to be bimodal: the interaction of color and odor on learning and memory. Behavioral Ecology  19: 425– 432. Google Scholar CrossRef Search ADS   Sillén-Tullberg B. 1985. Higher survival of an aposematic than of a cryptic form of a distasteful bug. Oecologia  67: 411– 415. Google Scholar CrossRef Search ADS   Sime KR, Feeny PP, Haribal MM. 2000. Sequestration of aristolochic acids by the pipevine swallowtail, Battus philenor (L.): evidence and ecological implications. Chemoecology  10: 169– 178. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2005. Tasting the difference: do multiple defence chemicals interact in Müllerian mimicry? Proceedings of the Royal Society of London. Series B, Biological Sciences  272: 339– 345. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2006a. Taste-rejection by predators and the evolution of unpalatability in prey. Behavioral Ecology and Sociobiology  60: 550– 555. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2006b. Avian predators taste-reject aposematic prey on the basis of their chemical defence. Biology Letters  2: 348– 350. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2006c. Predator avoidance learning of prey with secreted or stored defences and the evolution of insect defences. Animal Behaviour  72: 827– 834. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2009. Distastefulness as an antipredator defence strategy. Animal Behaviour  78: 761– 766. Google Scholar CrossRef Search ADS   Sugeno W, Matsuda K. 2002. Adult secretions of four Japanese Chrysomelinae (Coleoptera: Chrysomelidae). Applied Entomology and Zoology  37: 191– 197. Google Scholar CrossRef Search ADS   Svádová K, Exnerová A, Štys P, Landová E, Valenta J, Fučíková A, Socha R. 2009. Role of different colours of aposematic insects in learning, memory and generalization of naive bird predators. Animal Behaviour  77: 327– 336. Google Scholar CrossRef Search ADS   Termonia A, Hsiao TH, Pasteels JM, Milinkovitch MC. 2001. Feeding specialization and host-derived chemical defense in Chrysomeline leaf beetles did not lead to an evolutionary dead end. Proceedings of the National Academy of Sciences of the USA  98: 3909– 3914. Google Scholar CrossRef Search ADS   Tullberg BS, Leimar O, Gamberale- Stille G. 2000. Did aggregation favour the initial evolution of warning coloration? A novel world revisited. Animal Behaviour  59: 281– 287. Google Scholar CrossRef Search ADS   Von Nickisch-Rosenegk E, Wink M. 1993. Sequestration of pyrrolizidine alkaloids in several arctiid moths (Lepidoptera: Arctiidae). Journal of Chemical Ecology  19: 1889– 1903. Google Scholar CrossRef Search ADS   Wiklund C, Järvi T. 1982. Survival of distasteful insects after being attacked by naive birds: a reappraisal of the theory of aposematic coloration evolving through individual selection. Evolution  36: 998– 1002. Google Scholar CrossRef Search ADS   Zvereva EL, Rank NE. 2004. Fly parasitoid Megaselia opacicornis uses defensive secretions of the leaf beetle Chrysomela lapponica to locate its host. Oecologia  140: 516– 522. Google Scholar CrossRef Search ADS   Zvereva EL, Kozlov MV. 2016. The costs and effectiveness of chemical defenses in herbivorous insects: a meta-analysis. Ecological Monographs  86: 107– 124. Zvereva EL, Kruglova OY, Kozlov MV. 2010. Drivers of host plant shifts in the leaf beetle Chrysomela lapponica: natural enemies or competition? Ecological Entomology  35: 611– 622. Google Scholar CrossRef Search ADS   Zvereva EL, Kozlov MV, Neuvonen S. 1995. Decrease in feeding niche breadth of Melasoma lapponica (Coleoptera: Chrysomelidae) with increase in pollution Oecologia  104: 323– 329. Google Scholar CrossRef Search ADS   Zvereva EL, Kozlov MV, Rank NE. 2016. Does ant predation favour leaf beetle specialization on toxic host plants? Biological Journal of the Linnean Society  119: 201– 212. Google Scholar CrossRef Search ADS   Zvereva EL, Zverev V, Kruglova OY, Kozlov MV. 2017. Strategies of chemical anti-predator defences in leaf beetles: is sequestration of plant toxins less costly than de novo synthesis? Oecologia  183: 93– 106. Google Scholar CrossRef Search ADS   © 2018 The Linnean Society of London, Biological Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Biological Journal of the Linnean Society Oxford University Press

Defence strategies of Chrysomela lapponica (Coleoptera: Chrysomelidae) larvae: relative efficacy of secreted and stored defences against insect and avian predators

Loading next page...
 
/lp/ou_press/defence-strategies-of-chrysomela-lapponica-coleoptera-chrysomelidae-bHg0mO6q0E
Publisher
Oxford University Press
Copyright
© 2018 The Linnean Society of London, Biological Journal of the Linnean Society
ISSN
0024-4066
eISSN
1095-8312
D.O.I.
10.1093/biolinnean/bly045
Publisher site
See Article on Publisher Site

Abstract

Abstract Larvae of the leaf beetle Chrysomela lapponica defend themselves by release of repellent secretions, but also store potentially toxic compounds in their body tissues. We addressed the role of major groups of predators in the evolution of these two defence strategies by testing effects of these strategies on the behaviour of insect (wood ant, Formica polyctena) and avian (great tit, Parus major) predators. Ants were repelled by larval defensive secretions, but not by larvae devoid of secretions, larval haemolymph or integument. By contrast, birds rejected larvae devoid of secretions after the first attack; this suggests the presence of non-volatile defensive compounds within the larval body. However, survival was three-fold greater for larvae with intact secretions than for larvae with depleted secretions due to (1) irritating effects of secretions, resulting in frequent release of undamaged prey, and (2) faster avoidance learning and better prey memorability based on contact with secretions. Thus, volatile secretions and non-volatile compounds stored in the body act against birds jointly. Secretions sequestered from host plants were more effective than were autogenously produced secretions. We conclude that insect predators could contribute to the evolution of secreted but not of stored defensive chemicals, whereas bird predation could favour the evolution of both lines of defence. avoidance learning, bird predation, chemical defence, prey memorability, subtribe Chrysomelina, wood ant predation INTRODUCTION Animals have evolved numerous ways of defending themselves against their natural enemies. Among the various anti-predatory strategies, chemical defences are most widespread, and their diversity across species is striking (Blum, 1981; Ruxton, Sherratt & Speed, 2004; Eisner, Eisner & Siegler, 2007). To date, more than 1000 chemical compounds have been reported to have defensive functions in animals (El-Sayed, 2014), causing either toxic or irritating effects or both, or serving a warning function (Ruxton et al., 2004; Eisner et al., 2007). Moreover, animals employ various strategies of production, storage and deployment of these chemicals. The defensive chemicals may be synthesized de novo or can be acquired from the environment, in particular from food plants (Bowers, 1992; Nishida, 2002), and they can be used in different ways before, during and after a predator attack (Ruxton et al., 2004). The intriguing question is how evolution has shaped this diversity of defensive compounds and the strategies of their use. The two major defensive strategies of arthropods differ in the site of storage and mode of deployment of the defensive compounds: (1) they may be accumulated and stored internally in the body, or (2) they may be accumulated in specialized integumental organs, usually ectodermal glands, and released externally upon a predator attack (Blum, 1981; Pasteels, Grégoire & Rowell-Rahier, 1983). The chemicals stored inside the body cause toxic or deterrent effects on the predator only after the prey is consumed or its integument is broken, which means that the prey is seriously injured or killed. This strategy provides protection for other similar prey due to avoidance learning and generalization of predators (Gamberale-Stille & Tullberg, 1999; Skelhorn & Rowe, 2006a, b, c; Svádová et al., 2009). By contrast, externally secreted defences act on the predator’s chemical senses during (or even before) the attack and can thereby protect the prey from being injured or killed (Wiklund & Järvi, 1982; Sillén-Tullberg, 1985; Hotová Svádová et al., 2013). Thus, these two strategies differ in their effects on the survival of individual prey (Skelhorn & Rowe, 2006c). Secretion of defensive compounds is likely to be more costly compared to their storage in body tissues for several reasons. First, secretion usually requires the development of specialized organs, thereby incurring high constitutive costs. Second, considerable volumes of secretions are usually depleted during a predator attack, leading to additional costs of replenishment of lost chemicals and liquids (Bowers, 1992; Higginson et al., 2011). Third, defence externalization incurs ecological costs, expressed as an increased risk of parasitism (Zvereva & Kozlov, 2016). Conversely, the strategy of defence externalization may be advantageous because it can stop a predator’s attack before the prey is fatally damaged (Wiklund & Järvi, 1982; Sillén-Tullberg, 1985). Moreover, birds learn more rapidly to avoid prey possessing externally secreted defences than prey having defences stored in the body (Skelhorn & Rowe, 2006c; Hotová Svádová et al., 2013). A recent meta-analysis (Zvereva & Kozlov, 2016) has also confirmed that secretions released from defensive glands are more effective against naïve vertebrate predators than are stored defences. Thus, secretion of defensive compounds is more costly, but it is also more effective than storing them in body tissues; therefore, both strategies have similar cost–benefit ratios for animals using them. Consequently, both strategies are widely distributed. The high variation observed for the effectiveness of prey defences against different taxonomic and ecological groups of natural enemies (Zvereva & Kozlov, 2016) suggests that the response to chemically defended prey is taxon-specific. In particular, vertebrate and arthropod predators differ considerably in their hunting strategies, perception of chemical signals and sensitivity to toxins (Lagadic & Caquet, 1998; Kaupp, 2010; Boevé et al., 2013). For example, the main target of volatile, non-specific irritants released from defensive glands is suggested to be arthropods, while specific toxins stored in the body are more effective against birds and other vertebrate predators (Pasteels et al., 1983; Aldrich, 1988). Therefore, one hypothesis is that the evolution of different defensive strategies was driven by contrasting selective pressures imposed by various natural enemies (Boevé et al., 2013). Conversely, a meta-analysis (Zvereva & Kozlov, 2016) did not confirm differential effects of the two strategies on vertebrate and invertebrate (largely arthropod) predators in terrestrial ecosystems. The reason may be that non-toxic compounds secreted by prey may cause deterrent effects on invertebrate as well as on vertebrate predators (Conner et al., 2007) and could therefore be as effective as stored toxins against vertebrates. However, the effects of stored versus secreted prey defences on bird responses to natural prey and prey survival have never been compared experimentally. The majority of chemically defended insects use only one of the two above-mentioned major defensive strategies, either secreting chemicals from defensive glands or storing defensive compounds in body tissues (Zvereva & Kozlov, 2016). However, some species accumulate defensive chemicals inside the body (e.g. in haemolymph) and at the same time deploy specialized defensive glands; examples include the larvae of Lymantria dispar (Deml, 2003), some species of Saturniidae (Deml, 2003) and Papilionidae (Pinto, Urzua & Niemeyer, 2011; Priestap et al., 2012), and the leaf beetle Oreina cacaliae (Rowell-Rahier et al., 1995). In all these cases, the same defensive compound was identified both in haemolymph and in secretions. By contrast, in some insect species, the stored and secreted defensive compounds differ. The larvae of many leaf beetles (Chrysomelidae) possess specialized defensive glands, in which they accumulate compounds that are usually deterrents for various natural enemies (Pasteels, Braekman & Daloze, 1988). The larvae of the subtribe Chrysomelina were recently found to possess toxic compounds in the haemolymph; these were identified as isoxazolin glucoside and its 3-nitropropanoyl (3-NPA) esters (Pauls et al., 2016). Defence production is usually costly (Bowers, 1992; Ruxton et al., 2004), and thus the maintenance of two different pathways of biosynthesis would be expected to impose a considerable increase in constitutive allocation costs (i.e. costs associated with possessing a chemical defence). Therefore, the cost of possessing two lines of chemical defence would be supported by natural selection only if their combination provides exceptionally effective protection from enemies. The combination of secreted and stored compounds has been recently proposed to contribute synergistically to protection against diverse predators (Pauls et al., 2016). However, this hypothesis has not been tested to date. The strategies of chemical defence can be classified not only by the storage or secretion of chemicals, but also by the mechanism of defence acquisition: some species synthesize defensive chemicals de novo, while other species sequester plant allelochemicals (Pasteels et al., 1983; Bowers, 1992). The ability of herbivores to sequester plant compounds is tightly linked to their host-plant specialization (Bernays & Graham, 1988), so that sequestration is usually associated with highly specialized herbivore species (Zvereva & Kozlov, 2016). Therefore, comparative study of the costs and benefits of these two methods of defence acquisition is important for understanding the role of predators in the evolution of plant–herbivore interactions. For some groups of herbivores (e.g. cyanogenic butterflies), de novo synthesis is considered to be a phylogenetically derived strategy (Brown & Francini, 1990), whereas in leaf beetles of the subtribe Chrysomelina, the sequestration of chemicals is a derived feature while de novo biosynthesis is ancestral (Pasteels et al., 1988; Termonia et al., 2001; Fürstenberg-Hägg et al., 2014). The energetic costs are generally presumed to be higher for de novo synthesis than for sequestration (Rowell-Rahier & Pasteels, 1986; Bowers, 1992; Fürstenberg-Hägg et al., 2014); however, this difference has not been found in a study comparing several leaf beetle species (Zvereva et al., 2017) or in a meta-analysis involving 22 species of chemically defended insect herbivores (Zvereva & Kozlov, 2016). By contrast, several studies on leaf beetles have demonstrated that defences sequestered from the herbivore’s host plant are more effective against birds (Rowell-Rahier et al., 1995), ants (Zvereva, Kruglova & Kozlov, 2010), bacteria and fungi (Gross, Podsiadlowski & Hilker, 2002) than are autogenously produced defences. This suggests that the evolution of sequestration in leaf beetles has been favoured by the high effectiveness of the sequestered compounds against their major predators. Leaf beetles of the subtribe Chrysomelina, whose larvae use two strategies of anti-predatory defence (repellents and toxins: Pauls et al., 2016) and, in addition, are able to both sequester and synthesize the defensive compounds (Schulz, Gross & Hilker, 1997), provide a unique opportunity to test several hypotheses concerning evolution of anti-predator defences for a prey living in a multi-enemy world. In the present study, we have tested the hypothesis that two strategies of chemical defence in leaf beetle larvae – storage of chemicals in haemolymph and their secretion by specialized exocrine glands – have evolved as protection against the two major groups of predators of these beetles: insects and birds. We predict that secretions released from defensive glands would be most effective against insect predators, whereas compounds stored in the larval body would be protective primarily against birds. We have tested these predictions in several experiments aimed at answering the following questions: (1) Do defensive secretions released by Chrysomela lapponica larvae cause their rejection by wood ants and great tits? (2) Are larvae of C. lapponica that are devoid of secretions still toxic and/or unpalatable to ants and birds? (3) Do stored and secreted defences have different effects on avoidance learning and memorability in birds? We also tested the hypothesis that birds may have been responsible for the specialization of leaf beetles on toxic host plants, because secretion of allelochemicals sequestered from host plants provides better protection from birds when compared to autogenously produced secretions. MATERIAL AND METHODS Prey The leaf beetle C. lapponica is a univoltine species widely distributed in the Palaearctic region. Different populations of this leaf beetle feed on species of Salicaceae or Betulaceae. Adults hibernate in soil and start feeding and copulating on host plants soon after leaf flush. Females lay batches of 35–40 eggs on host plant leaves. Larvae feed in groups for about 1 month and pupate on the host plants. Larvae are black and possess nine pairs of eversible glands on the dorsal side of the thorax and abdomen. When disturbed, the larvae release droplets of defensive secretions from these glands for 1–3 s and then withdraw it back into gland reservoirs. The composition of the defensive secretions in C. lapponica larvae depends on the host plant: larvae feeding on Salix myrsinifolia sequester salicylic glucosides (SGs) to produce salicylaldehyde, while larvae feeding on SG-poor species (e.g. Betulaceae and some willow species, such as S. caprea) autogenously produce iso- and 2-methylbutyric esters as major defensive compounds (Hilker & Schulz, 1994; Termonia et al., 2001; Geiselhardt et al., 2015). For this study, we used a population of C. lapponica from the Kola Peninsula (NW Russia), which feeds in nature mostly on SG-rich S. myrsinifolia, although about 20% of the population was recorded on SG-poor willow species (Zvereva, Kozlov & Neuvonen, 1995). We obtained larvae for the experiments from eggs laid by females collected in the surroundings of Monchegorsk (67.94°N, 32.89°E) in June 2015 and 2016. We reared the larvae in the laboratory at room temperature and under a natural photoperiod in 50-mL glass vials on SG-rich S. myrsinifolia and SG-poor S. caprea. Consequently, the reared larvae either sequestered their defensive secretions from their host plants (S-secretions, hereafter) or produced them autogenously (A-secretions, hereafter) (Geiselhardt et al., 2015). In the experiments, we used larvae that had moulted to the last instar and had reached a fresh weight of about 25 mg. We used mealworms (larvae of Tenebrio molitor, Tenebrionidae) in experiments with adult birds and larvae of blow flies (Calliphora sp.; Calliphoridae, maggots, hereafter) in experiments with ants as well as with juvenile birds as palatable control prey to check the foraging motivation of predators. Predators Predation by ants (Hymenoptera: Formicidae) is often considered an important selective force in the evolution of arthropod chemical defences (Pasteels et al., 1983) and therefore ants are frequently used in studies of anti-predatory effects of various defensive compounds (Zvereva & Kozlov, 2016). Wood ants (Formica polyctena) use carbohydrates (secretions of aphids) and protein sources for nutrition and frequently prey on arthropods. Predation on arthropods becomes particularly important when ants actively search for protein sources to feed their larvae (Lenoir, 2002). We selected F. polyctena for our experiments because this species was most abundant in the localities populated by C. lapponica. Moreover, these ants were observed to attack leaf beetle larvae in nature, especially frequently on willows where ants tended aphids (Zvereva, Kozlov & Rank, 2016). Great tits (Parus major) are predominantly insectivorous birds. They have a wide distribution through the Palaearctic and occur over a range of different woodland types (Cramp & Perrins, 1993). They live in the same habitats as C. lapponica and therefore are likely to encounter the leaf beetle in nature. Great tits are frequently used in studies of insect anti-predatory defences (Lindström et al., 1999; Tullberg, Leimar & Gamberale-Stille, 2000; Exnerová et al., 2015) because they can easily adapt to laboratory conditions and to novel foods. We used wild-caught as well as naïve, hand-reared great tits as predators. We captured adult birds using mist-nets in Prague (50.08°N, 14.24°E) during September and October 2015. These birds were housed individually for 2–5 days before the experiments in plastic cages (50 × 40 × 50 cm) with wire-mesh front walls equipped with perches and water bowls for habituation to the laboratory conditions. The light conditions were set according to the outdoor photoperiod. The birds were fed a diet consisting of mealworms, sunflower seeds and a commercial food mixture (Uni Patee, Orlux). Nestlings (12–15 days old) were taken from nestboxes placed in woods at the outskirts of Prague in May 2016. We took a maximum of two juveniles from a single brood. They had only limited visual experience with food brought by their parents and were naïve with respect to any kind of unpalatable prey. Birds were hand-reared until they were able to feed themselves. Their diet consisted of mealworms, a commercial mixture for hand-rearing of passerines (Handmix, Orlux), and a mixture of boiled eggs and biscuits. Nestlings were kept in artificial nestboxes in small groups until fledged. They were then housed in the same cages as wild-caught birds, and provided with mealworms and food mixtures for insectivorous birds (Oké-bird and Nutribird, Versele-Laga; Uni Patee and Insect Patee, Orlux), vitamins, minerals (Roboran, Combisol) and water. Juveniles were tested when they had reached a stage of full independence (after the 35th day of life). Experiments with ants We conducted experiments with ants on 22 June 2016 in a pine forest (67.58°N, 32.55°E) situated 40 km south of Monchegorsk. We selected two large ant nests (mounds) about 20 m apart from each other. No leaf beetle larvae naturally occurred in the surroundings of the nests: the nearest population of C. lapponica was 10 km from the study site. Thus, all ants were naïve with respect to the prey items used in our experiments. We conducted the experiment using a method developed in an earlier study (Zvereva et al., 2016). We tested the motivation of the ants to use protein food near each mound by offering a maggot as a control prey. We placed each experimental item at about 1 m distance from the mound and about 5 cm from an ant trail, and no closer than 20 cm to a previous item and no earlier than 5 min after removal of previous prey; this ensured that different ants took part in the attack on each prey item. One observer (V.Z.) recorded ant behaviour until the prey item was carried to a nest. We obtained quantitative estimates of the effectiveness of prey defences by recording the number of ants that encountered the prey until one attacked it by attempting to bite (i.e. the number of repelled ants), which is frequently used as a measure of defence effectiveness against ants (Zvereva et al., 2016, and references therein). We used the following prey items in the experiment: (1) control prey (maggots); (2) C. lapponica larvae reared on S. myrsinifolia, with intact secretions; (3) similar larvae devoid of secretions (depleted larvae, hereafter); (4) maggots smeared with secretions of C. lapponica collected during secretion depletion; (5) maggots smeared with the body contents of dissected larva; and (6) integument of depleted larvae with viscera removed. We tested ten prey items (five per mound) of each kind in a random order, with one prey item offered during each trial. Different prey items were distributed evenly during the day to account for diurnal changes in ant activity. Autogenously produced secretions of C. lapponica (larvae reared on S. caprea) were compared with sequestered secretions (larvae reared on S. myrsinifolia) in earlier experiments, with the same ant species (Zvereva et al., 2010), and therefore were not tested here. We obtained depleted larvae by disturbing them with forceps to force them to release secretions, which we immediately collected in glass capillaries; we applied these secretions onto a control prey (maggot) to study the effects of secretions per se. We applied secretions collected from a single larva onto one maggot. We tested for the presence of defensive compounds in the integument by dissecting the depleted larvae, removing their viscera and haemolymph with a scalpel, and then cleaning the integument with a small piece of tissue paper, taking care that no traces of haemolymph were left on the surface. We then applied the removed body content (viscera and haemolymph) to maggots to test for the presence of defensive compounds within the larval body. We prepared all items in the field immediately before offering them to ants. Experiments with birds We carried out experiments with birds at the Faculty of Science, Charles University, Prague. The experiments with wild-caught birds were conducted during September–October 2015, and experiments with naïve, hand-reared birds during July 2016. We tested the birds individually in wooden cages (70 × 70 × 70 cm) with wire mesh walls, equipped with a perch, a dish with water and a rotating feeding tray with six cups. We observed the birds through a one-way glass in the front wall of the cage. The cage was illuminated by a daylight-simulating Biolux Combi 18-W bulb (Osram). Before the experiment, the birds were habituated to the experimental cages, were trained to take food from a glass Petri dish placed on a rotating tray, and were deprived of food for 2 h. Prey items were offered on a green background made from a leaf (white beam Sorbus aria in 2015, willow S. caprea in 2016). We recorded all experiments with birds with a video camera and continuously noted the behavioural elements using Observer XT 8.0 (Noldus). All experiments consisted of two parts: (1) an avoidance-learning session and (2) a memory test carried out 1 day after the avoidance-learning, in which we tested whether the birds remembered their experience with a particular prey. The avoidance-learning session consisted of a series of six consecutive trials, in which the bird was offered one prey item per trial: three trials with control prey and three trials with experimental prey, control and experimental trials in turn, starting from the control trial to check bird foraging motivation. If the bird refused to attack a control prey, we repeated the trial until the prey was consumed. Each trial was terminated after the bird ate the prey; otherwise, it lasted 5 min. The memory test consisted of two trials (the first with a control prey and the second with an experimental prey) conducted in the same way as the learning session; we offered the same experimental prey to the bird as had been offered the day before. During each trial we recorded (1) the latency of first attack; (2) whether the prey item was attacked (touched by beak, pecked or seized), killed/damaged (integument broken and larva is non-motile) and eaten by birds; and (3) duration of discomfort behaviour (cleaning the bill, shaking the head, ruffling the feathers). During the trials and several times during the following day, we recorded whether the birds demonstrated any signs of sickness (decreased activity, vomiting). We studied the different components of prey defences separately and in combination by designing three kinds of experiments. Within each experiment, the birds were randomly assigned to experimental subgroups tested with different types of prey. Each bird participated in only one of the experiments and was tested only with one of the prey types. The sex and age of the birds in experimental subgroups were balanced in all experiments. Reactions of birds to the secretions applied on palatable prey We conducted this experiment to study bird responses to sequestered and autogenous defensive secretions per se. We applied approximately 0.2 µL of secretions (corresponding to the amount of secretions produced by one larva: Geiselhardt et al., 2015) onto half a mealworm painted black with an odourless and non-toxic dye (Jovi S.A.157) to give the prey item a visual resemblance to a leaf beetle larva. The secretions were collected from defensive glands of last-instar larvae reared on either S. myrsinifolia or S. caprea into glass capillaries, which were sealed and kept in a freezer at −18 °C until the experiment. We used a pipette to apply secretions from the capillaries onto a mealworm immediately before the trial. We tested 18 birds with each S-secretions and A-secretions. Reactions of adult birds to larval body We studied whether the larval body itself (lacking the secretions) is unpalatable and/or toxic to birds, and whether host plants consumed by the larvae contribute to these effects, using frozen C. lapponica larvae reared either on S. myrsinifolia or on S. caprea. The larvae were depleted before freezing by soaking up the released secretions with a piece of filter paper. We tested 18 birds with each kind of larva. Half a mealworm (which matched the size of a C. lapponica larva) was used as control palatable prey. Reactions of juvenile hand-reared birds to live larvae We conducted this experiment to study the innate responses of naïve birds as well as the process of avoidance learning and memorability of fully defended leaf beetle larvae. We used three kinds of C. lapponica larvae: larvae reared on S. myrsinifolia (S-secretions intact), larvae reared on S. myrsinifolia (S-secretions depleted) and larvae reared on S. caprea (A-secretions intact). We tested 15 birds with each kind of prey. We accounted for a potential effect of the birds’ neophobia using maggots as a control palatable prey that the birds had not encountered before. We avoided the effects of differences in colour between larvae (black) and maggots (white) by painting the maggots with an odourless and non-toxic black dye. Statistical analysis We compared the numbers of ants repelled by different types of prey by ANOVA, followed by Duncan’s post-hoc test. The data were ln-transformed prior to the analysis to meet the assumptions of normality. In all experiments with birds, we compared attack latencies between control prey (mealworms or maggots) and experimental prey using the Wilcoxon Signed Rank test. In the experiment with secretions applied on a control prey, we analysed the differences in attack latencies and in duration of discomfort behaviour between the two types of secretions across trials by repeated-measures ANOVA. These two variables were sqrt(ln(x))-transformed to fit a normal distribution. In the experiment with live larvae, we used the Kruskal–Wallis test for comparison of first attack latencies between the three kinds of larvae. In experiments with live and frozen larvae, we compared the numbers of birds that attacked, killed (damaged) or ate the prey between the different types of larvae and between larvae and the corresponding control prey using the two-sided Fisher’s exact test. We performed all statistical tests using SAS 9.4 (SAS Institute, 2015). Ethical note We obtained permissions for experiments with wild-caught and hand-reared great tits from the Environmental Department of Municipality of Prague (S-MHMP-83637/2014/OZP-VII-3/R-8/F), Ministry of Agriculture (13060/2014-MZE-17214), and Ministry of the Environment of the Czech Republic (42521/ENV/14–2268/630/14). We ringed birds individually and released them back to the locality of capture within a few days after experimentation. RESULTS Ant predation The prey types we used in the field experiment repelled different numbers of ants before the prey was attacked (F5,53 = 12.23, P < 0.0001). All control prey (maggots) were immediately attacked by the first ant that found it (Fig. 1), indicating a high motivation of the ants in the selected nests to collect protein food. In contrast, up to 20 ants were repelled by intact larvae before the first attack was attempted. When an ant approached the intact larva of C. lapponica and touched the prey with its antennae, the larva released droplets of secretions from its glands, and then slowly drew them back into the glands when the disturbance was over. After contact with secretions, the ant usually retreated and cleaned its antennae. After several encounters with ants, accompanied by the release of secretions, the amount of larval secretions decreased, and this allowed for the first ant bite. Subsequently, other nearby ants joined the attack and the prey was quickly killed and transported to the nest. Maggots with larval secretions applied onto their surfaces also repelled ants; the difference in repellence between maggots coated with secretions and intact larvae was not statistically significant (Fig. 1). Larvae with depleted secretions, maggots coated with larval body content and larval integument alone did not demonstrate significant repellence and were attacked as rapidly as control maggots (Fig. 1). Figure 1. View largeDownload slide Effect of chemical defences of larvae of the leaf beetle Chrysomela lapponica on the number of wood ants (Formica polyctena) (mean+SE; each based on ten items) repelled by different types of prey in the field experiment. LS: larva of C. lapponica with intact sequestered secretions (S-secretions); CS: maggot coated with S-secretions; LD: larva of C. lapponica with depleted S-secretions; CH: maggot coated with larval haemolymph; LI: depleted larva with viscera removed; C: maggot of the same size as a larva (control). Different letters above bars indicate significant (P < 0.05) differences between prey types (Duncan test). Figure 1. View largeDownload slide Effect of chemical defences of larvae of the leaf beetle Chrysomela lapponica on the number of wood ants (Formica polyctena) (mean+SE; each based on ten items) repelled by different types of prey in the field experiment. LS: larva of C. lapponica with intact sequestered secretions (S-secretions); CS: maggot coated with S-secretions; LD: larva of C. lapponica with depleted S-secretions; CH: maggot coated with larval haemolymph; LI: depleted larva with viscera removed; C: maggot of the same size as a larva (control). Different letters above bars indicate significant (P < 0.05) differences between prey types (Duncan test). Reactions of birds to the secretions applied onto palatable prey Of the 108 mealworms with applied C. lapponica secretions, only three (all with S-secretions) were not attacked during the experiment and five were rejected after attack without being eaten (two with A-secretions and three with S-secretions). The two types of secretions differed in some of the effects they had on bird behaviour. The latencies of attacks on mealworms coated with secretions were longer than those on control mealworms in the first trial for both S-secretions (S = 55.5, P = 0.007) and A-secretions (S = 59, P = 0.008) and approached significance in the second trial for S-secretions (S = 40.5, P = 0.056), but not for A-secretions (S = 31.5, P = 0.18). In the third trial, the differences between secretion-coated and control mealworms were not statistically significant (S-secretions: S = 33, P = 0.09; A-secretions: S = 20, P = 0.40) (Fig. 2). Across all the trials (interaction between secretion type and trial number: F2,68 = 0.77, P = 0.47), birds hesitated for a longer period before attacking mealworms coated with S-secretions than before attacking mealworms coated with A-secretions (repeated-measures ANOVA: F1,34 = 3.96, P = 0.05). The duration of discomfort behaviour while handling and after eating the prey decreased with the sequence number of the trial (F2,56 = 15.66, P < 0.0001) and was longer for birds that came into contact with S-secretions than with A-secretions (F1,28 = 4.75, P = 0.038) (Fig. 3); the difference between secretion types did not depend on the trial number (interaction between secretion type and trial number: F2,56 = 0.94, P = 0.40). Figure 2. View largeDownload slide Latencies of attacks of wild-caught adult great tits on mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica when compared with control mealworms during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Points indicate medians, boxes indicate interquartile ranges and whiskers show non-outlier ranges. Asterisks indicate significant (P < 0.05) differences between latencies of attacks on control mealworms and on mealworms with secretions applied (Wilcoxon Signed Rank test). Figure 2. View largeDownload slide Latencies of attacks of wild-caught adult great tits on mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica when compared with control mealworms during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Points indicate medians, boxes indicate interquartile ranges and whiskers show non-outlier ranges. Asterisks indicate significant (P < 0.05) differences between latencies of attacks on control mealworms and on mealworms with secretions applied (Wilcoxon Signed Rank test). Figure 3. View largeDownload slide Duration of discomfort behaviour of wild-caught adult great tits during handling and eating of mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Bars indicate means + SE; asterisks indicate significant (P < 0.05) differences between two types of secretions (repeated-measures ANOVA). Figure 3. View largeDownload slide Duration of discomfort behaviour of wild-caught adult great tits during handling and eating of mealworms coated with either sequestered or autogenous secretions of larvae of the leaf beetle Chrysomela lapponica during three consecutive trials in an avoidance-learning session (sample size: 18 birds in each group). Bars indicate means + SE; asterisks indicate significant (P < 0.05) differences between two types of secretions (repeated-measures ANOVA). Reactions of birds to a larval body lacking secretions First attack latencies were longer toward dead depleted C. lapponica larvae than toward mealworms (S = 356, P = 0.002). After the first attack (in the second and third trials), the numbers of birds that attacked the larvae were significantly reduced when compared to control mealworms (A-secretion larvae: P = 0.003; S-secretion larvae: P < 0.0001) and to the first trial with larvae (A-secretion larvae: P < 0.0001; S-secretion larvae: P < 0.0001), indicating that the birds also learned to avoid the larvae in the absence of released secretions. The numbers of attacking birds did not differ between the groups tested with the two types of larvae (P = 0.46 and P = 1 for the second and third trial, respectively) (Fig. 4). In the memory test, the numbers of attacking birds remained nearly the same as in the last trial of the learning session (P = 1 for both types of larvae) and these numbers did not differ between the two types of larvae (P = 1) (Fig. 4). Figure 4. View largeDownload slide Numbers of wild-caught adult great tits that attacked freshly defrosted Chrysomela lapponica larvae depleted of either sequestered (sample size: 15 birds) or autogenous (14 birds) secretions during three consecutive trials in an avoidance-learning session and in a memory test. Control prey novel to the birds (maggot) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). Figure 4. View largeDownload slide Numbers of wild-caught adult great tits that attacked freshly defrosted Chrysomela lapponica larvae depleted of either sequestered (sample size: 15 birds) or autogenous (14 birds) secretions during three consecutive trials in an avoidance-learning session and in a memory test. Control prey novel to the birds (maggot) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). Of 13 birds that attacked both types of larvae during the first encounter without damaging the integument, ten (76.9%) avoided the larvae during the following encounters, whereas only six of 15 (40%) birds that ate the larvae during the first encounter rejected them later on. None of the birds that consumed the larvae showed any signs of sickness afterwards. Reactions of birds to live larvae The birds attacked all types of C. lapponica larvae during the first encounter at the same rate as they attacked a novel palatable prey (Fig. 5). However, the latencies of attacks on C. lapponica larvae were longer for all types of larvae even during the first encounter when compared to control maggots (depleted larvae: S = 38, P = 0.03; A-secretion larvae: S = 52, P = 0.002; S-secretion larvae: S = 47, P = 0.005). The differences in the latency between the three groups of larvae were not statistically significant (χ2 = 0.95, df = 2, P = 0.62). Figure 5. View largeDownload slide Numbers of hand-reared juvenile great tits that attacked Chrysomela lapponica larvae differing in chemical defence (sequestered, autogenous or depleted; sample size: 15 birds in each group) during three consecutive trials in an avoidance-learning session and in a memory test. Control prey (maggots) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). Figure 5. View largeDownload slide Numbers of hand-reared juvenile great tits that attacked Chrysomela lapponica larvae differing in chemical defence (sequestered, autogenous or depleted; sample size: 15 birds in each group) during three consecutive trials in an avoidance-learning session and in a memory test. Control prey (maggots) were offered before each trial and were attacked by all birds (data not shown on the figure). Asterisks indicate significant (P < 0.05) differences between bird attacks on larvae and on maggots (Fisher’s Exact Test). The number of birds that attacked the larvae in the second trial was considerably lower for S-secretion larvae, both intact and depleted (P = 0.02 for both groups), when compared with attacks on control maggots, but not for A-secretion larvae (P = 0.22, Fig. 5). In the third trial, the numbers of birds attacking the larvae were lower for S-secretion larvae than for either depleted larvae (P = 0.03) or A-secretion larvae (P = 0.03), whereas the probability of attacks was the same for depleted and A-secretion larvae (P = 1). In the memory test, all types of larvae were attacked at a reduced rate when compared to the control palatable prey, but the S-secretion larvae still differed marginally (P = 0.06) from the depleted larvae (Fig. 5). Some larvae that were attacked by birds were released unharmed and survived the attack. Survival of attack was considerably lower for the depleted larvae than for A-secretion larvae (21.6% and 47.2%, respectively, P = 0.03) and S-secretion larvae (21.6% and 67.8%, respectively, P = 0.0003), but survival did not differ between A-secretion and S-secretion larvae (P = 0.13). However, survival in the last trial was higher for S-secretion larvae than for A-secretion larvae due to a decrease in the attack rate (P = 0.04). In the memory test, larval survival did not differ between S-secretion larvae and A-secretion larvae (P = 0.65), but it was higher for larvae with intact secretions (both types combined) than for larvae with depleted secretions (P = 0.017). Of 11 birds that avoided S-secretion larvae in the memory test, ten attacked but did not damage the larvae encountered in the learning trials. Birds ate a significantly lower proportion of larvae with intact secretions when compared with depleted larvae (S-secretion larvae: P = 0.006; A-secretion larvae: P = 0.0004), and only 50% of the depleted larvae were eaten compared to 100% of the control maggots (P < 0.0001). The S-secretion larvae and A-secretion larvae were eaten at the same rate by the birds (P = 0.55). Of the 16 birds that attacked but did not kill any larvae (and thus had no contact with larval body contents) during the avoidance-learning session, 15 avoided attacking the larvae in the memory test. By contrast, of 15 other birds that ate at least one larva, only four avoided attacking them in the memory test (P = 0.05). None of the 17 juvenile birds that partly or completely consumed the larvae showed any signs of sickness during the experiment or the day after. DISCUSSION Effects of leaf beetle defensive secretions on ants and birds Our study provided a direct estimation of the relative effectiveness of different defence strategies of leaf beetle larvae against insect and avian predators. Moreover, detailed observations of predator behaviour revealed the mechanisms by which these defences provide protection from different predators. One suggestion has been that volatile secretions released by the prey are directed mostly against invertebrate predators, and against ants in particular (Boevé & Pasteels, 1985; Boevé et al., 2013). Consistently, ants have been suggested to represent a major selective agent for secreted chemical defences in leaf beetle larvae (Pasteels et al., 1983), whereas the role of birds in the evolution of volatile defences has remained unclear (Pauls et al., 2016). In agreement with an earlier study (Zvereva et al., 2010), we showed that salicylaldehyde-containing secretions released by leaf beetle larvae in response to disturbance by wood ants, and even by their approach, had a repellent effect on the ants. By contrast, larvae with depleted secretions were attacked by ants at the same rate as control prey. The latter result supports the hypothesis that the release of secretions is the only factor responsible for the protection of leaf beetle larvae against ants. However, wood ants collectively foraging in the field overcome the defensive effects of secretions of C. lapponica larvae by social interactions, chemical signalling and olfactory learning, thereby imposing a high mortality on larvae near ant nests (Zvereva et al., 2016). However, the repellence of secretions favours the survival of leaf beetle larvae when ant density is low. Salicylaldehyde-containing secretions were also found to repel some other generalist arthropod predators, such as true bugs (Rank et al., 1998) and ladybirds (Denno, Larsson & Olmstead, 1990), at the same time serving as a search cue for specialist predators and parasitoids (Köpf et al., 1997; Gross et al., 2004; Zvereva & Rank, 2004). Our results show that volatile secretions of C. lapponica provide a direct defensive effect not only against generalist arthropod predators, but also against insectivorous birds, and that the secretions affect bird behaviour in several ways. First, although the birds attacked C. lapponica larvae during the first encounter at the same rate as they attacked a control palatable prey, they hesitated longer before attacking the larvae than when attacking the control prey. This behaviour could not be explained by increased wariness elicited by a visually novel prey (Marples & Kelly, 1999; Marples et al., 2007) or an innate bias (Lindström et al., 1999) against its conspicuous coloration (black on a green background), because the control prey was also novel (and also painted black) in the experiment with naïve juvenile birds. Attack latencies were increased in comparison with palatable prey in experiments with live C. lapponica larvae, as well as when the secretions were applied onto control prey; therefore, we suggest that this effect may be explained by an innate bias against novel odours (Jetz, Rowe & Guilford, 2001). Novel odours are known to elicit hidden innate biases against visually novel and/or conspicuous prey (Marples & Roper, 1996; Rowe & Guilford, 1996; Lindström, Rowe & Guilford, 2001); therefore, a combined effect of visual and olfactory cues may be responsible for the longer attack latencies observed in our experiment. Second, defensive secretions considerably increased the chance that larvae would survive bird attacks: the birds released, alive and unharmed, 59% of the attacked larvae with intact secretions but only 21% of attacked larvae with removed secretions. This was presumably due to the contact deterrent (irritating) effect of the secretions, which forced the birds to quickly drop the attacked prey. The irritating effect of secretions was also evident from the discomfort behaviour of birds after attacking mealworms coated with secretions. These results are in accordance with other studies showing that birds taste and reject chemically defended prey (Skelhorn & Rowe, 2006a,b; Halpin & Rowe, 2010), which can then be released relatively unharmed (Wiklund & Jarvi, 1982; Sillen-Tullberg, 1985; Skelhorn & Rowe, 2009; Hotová Svádová et al., 2013). The fact that a considerable proportion of the prey population can survive predator attack is important for understanding the pathways of chemical defence evolution in insects with externally secreted defences. It also confirms the importance of individual selection in this process (Skelhorn & Rowe, 2006c). Third, attacks on S-secretion larvae, even those that did not involve killing and eating the larvae, induced more effective avoidance learning than did attacks on larvae lacking the secretions. This effect may be explained either by high aversiveness of S-secretions for birds (Lindström, Alatalo & Mappes, 1997; Lindström et al., 2006) or by their potential role as chemical (olfactory or gustatory) signals that the birds associate with prey noxiousness (Guilford et al., 1987; Skelhorn & Rowe, 2005, 2009; Siddall & Marples, 2008). The reduced survival of larvae with depleted secretions in our experiments provides evidence for the high effectiveness of volatile secretions of C. lapponica larvae against bird predators. The secretions enhance innate wariness in bird predators, induce rapid avoidance learning and increase the chance that the prey will survive the attack. Differences between sequestered and autogenous secretions Our field experiment showed that sequestered secretions of C. lapponica larvae effectively repelled wood ants. Earlier experiments that used the same method have also demonstrated that the autogenously produced secretions are repellent for ants (Zvereva et al., 2010). This result is in accordance with laboratory experiments that demonstrated repellent effects of both kinds of secretions on ants (Blum et al., 1972; Hilker & Schulz, 1994). However, secretions containing sequestered salicylaldehyde appeared more effective than autogenously produced secretions containing butyric esters, as they repelled significantly more ants and thus allowed the larvae to survive longer near the ant nest (Zvereva et al., 2010). We also found differences between the effects of sequestered and autogenous secretions in our experiments with birds. First, the birds learned to avoid sequestered secretions much faster when compared to autogenous secretions: S-secretion larvae were avoided even after first attack, whereas all three A-secretion larvae offered during the learning session were attacked at the same rate. Second, contact with sequestered secretions elicited a stronger discomfort reaction (head shaking and beak wiping) in birds than did the contact with the autogenous secretions. These results indicate stronger aversiveness of sequestered secretions than of autogenously produced secretions. Third, sequestered secretions, when applied onto control prey (mealworms), resulted in longer initial attack latencies in the first trial than we observed for autogenous secretions applied in the same way. The absence of a similar effect in an experiment with live larvae probably reflects the fact that the larvae usually discharge their secretions only when attacked. However, in a natural situation, when larvae are closely aggregated, the release of secretions by a single attacked individual may provide increased protection for the group and enhance the dilution effect (Riipi et al., 2001). Once a larva was attacked, we saw no effect of the secretion type on its survival. Both secretions were quite effective and resulted in high survival rates of larvae relative to control prey: about half of attacked larvae were released unharmed. However, the combined effect of the secretions on attack latencies and avoidance learning led to better overall survival of S-secretion larvae compared to A-secretion larvae. At the same time, the experiment with frozen depleted larvae showed no differences between S-secretion larvae and A-secretion larvae, indicating that differences in defence effectiveness between these two kinds of larvae depend exclusively on the composition and/or the amount of secretions. Along with these differences, we found some similarities in the effects of sequestered and autogenous secretions. Larvae with both types of secretions survived attacks by birds at the same rate, and the birds remembered their negative experience equally well for both kinds of defences, as indicated by the similar survival of larvae in the memory test. Thus, both types of secretions provide effective protection against birds. Effects of defensive substances contained in body tissues The 3-NPA esters, which were discovered in the haemolymph of larvae of species belonging to the subtribe Chrysomelina, were found to be a deterrent for ants (Pasteels, Daloze & Rowell-Rahier, 1986; Sugeno & Matsuda, 2002). However, we did not detect any deterrent effect of body content and integument of C. lapponica larvae against wood ants. A similar contradiction was also reported by Reudler et al. (2015): in their experiments, wood ants preferred a solution extracted from the larvae of Parasemia plantaginis over sugar water, although pure defensive compounds found in haemolymph (iridoid glycosides) were deterrent. The authors hypothesized that the attractiveness of nutrients (e.g. proteins) contained in the prey overcomes potential deterrent effects of defensive chemicals (Reudler et al., 2015). We therefore suggest that only externally released secretions provide sufficient protection for C. lapponica larvae against wood ants. Our experiments with great tits indicate that the secretions alone are not what makes the larvae unpalatable: avoidance was also developed against frozen larvae and against live larvae devoid of secretions. Moreover, not only larvae with secretions but also depleted larvae were attacked at a reduced rate in the memory test when compared with the control palatable prey. This indicates that factors other than secretions contribute to the effectiveness of larval defences against bird predators. One of these factors could be the non-volatile isoxazolinone-5-one glucoside and its 3-NPA ester, which are synthesized autogenously during larval life (Pauls et al., 2016). The latter compound is a neurotoxin that causes poisoning of humans and domestic livestock (Beal et al., 1993; Anderson et al., 2005). The effect of this compound on birds has not been studied before, and our results indicate that the amount contained in one larva is not sufficient to cause any signs of poisoning in a bird the size of a great tit. Even three larvae eaten completely, one after another, do not elicit any signs of sickness. If many larvae need to be eaten to cause toxic effects in bird predators, this raises doubts that body toxins could have evolved under selection pressure imposed by birds. Our results also indicate that the factors contributing to larval defence (other than secretions) act upon contact with the larval surface rather than with the body contents, because consumption or even damaging the prey was not necessary for the development of avoidance. The integument of C. lapponica was not analysed separately by Pauls et al. (2016), but our results suggest that defensive compounds are probably deposited in considerable amounts in integument, as even the contact with a larval surface lacking secretions contributed to avoidance learning. Considerable amounts of defensive chemicals were detected in the integument when it has been analysed separately (Montllor, Bernays & Cornelius, 1991; Fürstenberg-Hägg et al., 2014). This was demonstrated for arctiid moths, which accumulated pyrrolizidine alkaloids in the larval integument (von Nickisch-Rosenegk & Wink, 1993), and for the larvae of a papilionid butterfly, Battus polydamas, which uses the integument as the major site of aristolochic acid accumulation (Priestap et al., 2012). These examples indicate that deposition of defensive compounds within the integument is more widespread than previously thought, because it may increase prey survival upon predator attack owing to taste rejection by the predator before the prey is mortally damaged (Skelhorn & Rowe, 2006a; Zvereva & Kozlov, 2016). Thus, externalized defensive compounds of prey seem to be more important for avoidance learning in avian predators than are the defensive compounds stored inside the prey body. This conclusion concurs with the results of previous experimental studies (Skelhorn & Rowe, 2006c; Hotová Svádová et al., 2013) and with the conclusions of a meta-analysis comparing the effectiveness of externally secreted defensive chemicals with chemicals stored in body tissues against vertebrate predators (Zvereva & Kozlov, 2016). The advantages of externalized chemical defences can be illustrated by the existence of numerous externalization mechanisms, which have independently and repeatedly evolved in several insect taxa (Ohkuma et al., 2004). The low effectiveness of defensive compounds contained inside the body for anti-predatory protection of C. lapponica larvae suggests some other biological significance of these compounds. The isoxalinone derivatives are major components of adult defences in the subtribe Chrysomelina and are released from the beetle exocrine glands upon disturbance (Deroe & Pasteels, 1982; Sugeno & Matsuda, 2002; Pasteels et al., 2003), thereby serving as an efficient externalized defence (Sugeno & Matsuda, 2002). These compounds, synthesized autogenously by larvae (Pauls et al., 2016), can be transferred from larvae to adults. This passing of the accumulated larval defensive compounds through metamorphosis has been demonstrated for many insect species, and for chemicals both sequestered by larvae from their food plants and biosynthesized de novo, such as aristolochic acid (Sime, Feeny & Haribal, 2000), pyrrolizidine alkaloids (Rossini et al., 2003) and cyanogenic glucosides (Fürstenberg-Hägg et al., 2014). This strategy allows adults to conserve energy required for reproduction. CONCLUSIONS We have found a strong repellent effect of exocrine volatile secretions of larvae of C. lapponica on both insect (ants) and avian (great tits) predators. While the aversive effects of these secretions on insect predators have already been reported, we have provided the first evidence that the two kinds of larval secretions of Chrysomelina leaf beetles that differ in their chemistry and origin affect bird behaviour and increase prey survival through both taste rejection and enhanced avoidance learning. Nevertheless, salicylaldehyde-containing secretions sequestered from host plants were more effective against both ants and birds than were autogenously produced butyrate-based secretions. This suggests that both insect and avian predators could have contributed to the evolution of the ability to sequester salicylic glucosides from salicaceous host plants in leaf beetles. We have demonstrated that larvae devoid of secretions, although not toxic, are still unpalatable for birds, confirming the anti-predatory function of non-volatile compounds within the larval body. We also found that the two lines of defence – secretion from glands and storage in the body – act jointly against bird predation. However, we failed to detect any defensive effects of larval body content against ants. We conclude that insect predators could have contributed to the evolution of secreted defensive chemicals, but not of stored chemical defences, whereas bird predation could have favoured the evolution of both lines of defences. ACKNOWLEDGEMENTS We are grateful to two anonymous reviewers for their helpful comments on an earlier draft of the manuscript. This work was supported by the Academy of Finland (project 268124) and Czech Science Foundation (P505/11/1459). The authors do not have any conflicts of interest. REFERENCES Aldrich JR. 1988. Chemical ecology of the Heteroptera. Annual Reviews in Entomology  33: 211– 238. Google Scholar CrossRef Search ADS   Anderson RC, Majak W, Rassmussen MA, Callaway TR, Beier RC, Nisbet DJ, Allison MJ. 2005. Toxicity and metabolism of the conjugates of 3-nitropropanol and 3-nitropropionic acid in forages poisonous to livestock. Journal of Agricultural and Food Chemistry  53: 2344– 2350. Google Scholar CrossRef Search ADS   Beal MF, Brouillet E, Jenkins BG, Ferrante RJ, Kowall NW, Miller JM, Storey E, Srivastava R, Rosen BR, Hyman BT. 1993. Neurochemical and histologic characterization of striatal excitotoxic lesions produced by the mitochondrial toxin 3-nitropropionic acid. The Journal of Neuroscience  13: 4181– 4192. Google Scholar CrossRef Search ADS   Bernays E, Graham M. 1988. On the evolution of host specificity in phytophagous arthropods. Ecology  69: 886– 892. Google Scholar CrossRef Search ADS   Blum MS, Brand JM, Wallace JB, Fales HM. 1972. Chemical characterization of defensive secretion of a chrysomelid larva. Life Sciences. Part 2, Biochemistry, General & Molecular Biology  11: 525– 531. Blum MS. 1981. Chemical defenses of arthropods . London: Academic Press. Boevé JL, Pasteels JM. 1985. Modes of defense in nematine sawfly larvae: efficiency against ants and birds. Journal of Chemical Ecology  11: 1019– 1036. Google Scholar CrossRef Search ADS   Boevé JL, Blank SM, Meijer G, Nyman T. 2013. Invertebrate and avian predators as drivers of chemical defensive strategies in tenthredinid sawflies. BMC Evolutionary Biology  13: 198. Google Scholar CrossRef Search ADS   Bowers MD. 1992. The evolution of unpalatability and the cost of chemical defense in insects. In: Roitberg BD, Isman MB, eds. Insect chemical ecology: an evolutionary approach . New York: Chapman and Hall, 216– 244. Brown KSJr, Francini RB. 1990. Evolutionary strategies of chemical defense in aposematic butterflies: cyanogenesis in Asteraceae-feeding American Acraeinae. Chemoecology  1: 52– 56. Google Scholar CrossRef Search ADS   Chomcheon P, Wiyakrutta S, Sriubolmas N, Ngamrojanavanich N, Isarangkul D, Kittakoop P. 2005. 3-Nitropropionic acid (3-NPA), a potent antimycobacterial agent from endophytic fungi: is 3-NPA in some plants produced by endophytes? Journal of Natural Products  68: 1103– 1105. Google Scholar CrossRef Search ADS   Conner WE, Alley KM, Barry JR, Harper AE. 2007. Has vertebrate chemesthesis been a selective agent in the evolution of arthropod chemical defenses? The Biological Bulletin  213: 267– 273. Google Scholar CrossRef Search ADS   Cramp S, Perrins C. 1993. The birds of the Western Palearctic, Vol. 8–9 . Oxford: Oxford University Press. Deml R. 2003. Pyrrolidonyl and pyridyl alkaloids in Lymantria dispar. Zeitschrift fur Naturforschung. C, Journal of Biosciences  58: 860– 866. Denno RF, Larsson S, Olmstead KL. 1990. Role of enemy-free space and plant quality in host-plant selection by willow beetles. Ecology  71: 124– 137. Google Scholar CrossRef Search ADS   Deroe C, Pasteels JM. 1982. Distribution of adult defense glands in chrysomelids (Coleoptera: Chrysomelidae) and its significance in the evolution of defense mechanisms within the family. Journal of Chemical Ecology  8: 67– 82. Google Scholar CrossRef Search ADS   Eisner T, Meinwald J. 1966. Defensive secretions of arthropods. Science  153: 1341– 1350. Google Scholar CrossRef Search ADS   Eisner T, Eisner M, Siegler M. 2007. Secret weapons: defenses of insects, spiders, scorpions, and other many-legged creatures . Cambridge: Belknap Press. El-Sayed AM. 2014. The pherobase: database of pheromones and semiochemicals . Avaliable at: http://www.pherobase.com (accessed 26 January 2018). Exnerová A, Ježová D, Štys P, Doktorovová L, Rojas B, Mappes J. 2015. Different reactions to aposematic prey in 2 geographically distant populations of great tits. Behavioral Ecology  26: 1361– 1370. Google Scholar CrossRef Search ADS   Francis K, Smitherman C, Nishino SF, Spain JC, Gadda G. 2013. The biochemistry of the metabolic poison propionate 3-nitronate and its conjugate acid, 3-nitropropionate. IUBMB Life  65: 759– 768. Google Scholar CrossRef Search ADS   Fürstenberg-Hägg J, Zagrobelny M, Olsen CE, Jørgensen K, Møller BL, Bak S. 2014. Transcriptional regulation of de novo biosynthesis of cyanogenic glucosides throughout the life-cycle of the burnet moth Zygaena filipendulae (Lepidoptera). Insect Biochemistry and Molecular Biology  49: 80– 89. Google Scholar CrossRef Search ADS   Gamberale-Stille G, Tullberg BS. 1999. Experienced chicks show biased avoidance of stronger signals: an experiment with natural colour variation in live aposematic prey. Evolutionary Ecology  13: 579– 589. Google Scholar CrossRef Search ADS   Geiselhardt S, Hilker M, Müller F, Kozlov MV, Zvereva EL. 2015. Inter- and intrapopulation variability in the composition of larval defensive secretions of willow-feeding populations of the leaf beetle Chrysomela lapponica. Journal of Chemical Ecology  41: 276– 286. Google Scholar CrossRef Search ADS   Gross J, Podsiadlowski L, Hilker M. 2002. Antimicrobial activity of exocrine glandular secretion of Chrysomela larvae. Journal of Chemical Ecology  28: 317– 331. Google Scholar CrossRef Search ADS   Gross J, Fatouros NE, Neuvonen S, Hilker M. 2004. The importance of specialist natural enemies for Chrysomela lapponica in pioneering a new hostplant. Ecological Entomology  29: 584– 593. Google Scholar CrossRef Search ADS   Guilford T, Nicol C, Rothschild M, Moore BP. 1987. The biological roles of pyrazines − evidence for a warning odor function. Biological Journal of the Linnean Society  31: 113– 128. Google Scholar CrossRef Search ADS   Halpin CG, Rowe C. 2010. Taste-rejection behaviour by predators can promote variability in prey defences. Biology Letters  6: 617– 619. Google Scholar CrossRef Search ADS   Higginson AD, Delf J, Ruxton GD, Speed MP. 2011. Growth and reproductive costs of larval defence in the aposematic lepidopteran Pieris brassicae. Journal of Animal Ecology  80: 384– 392. Google Scholar CrossRef Search ADS   Hilker M, Schulz S. 1994. Composition of larval secretion of Chrysomela lapponica (Coleoptera, Chrysomelidae) and its dependence on host plant. Journal of Chemical Ecology  20: 1075– 1093. Google Scholar CrossRef Search ADS   Hotová Svádová K, Exnerová A, Kopecková M, Štys P. 2013. How do predators learn to recognize a mimetic complex: experiments with naive great tits and aposematic Heteroptera. Ethology  119: 814– 830. Google Scholar CrossRef Search ADS   Huang LS, Sun G, Cobessi D, Wang AC, Shen JT, Tung EY, Anderson VE, Berry EA. 2006. 3-Nitropropionic acid is a suicide inhibitor of mitochondrial respiration that, upon oxidation by complex II, forms a covalent adduct with a catalytic base arginine in the active site of the enzyme. Journal of Biological Chemistry  281: 5965– 5972. Google Scholar CrossRef Search ADS   Jetz W, Rowe C, Guilford T. 2001. Non-warning odors trigger innate color aversions − as long as they are novel. Behavioral Ecology  12: 134– 139. Google Scholar CrossRef Search ADS   Kaupp UB. 2010. Olfactory signalling in vertebrates and insects: differences and commonalities. Nature Reviews Neuroscience  11: 188– 200. Google Scholar CrossRef Search ADS   Köpf A, Rank NE, Roininen H, Tahvanainen J. 1997. Defensive larval secretions of leaf beetles attract a specialist predator Parasyrphus nigritarsis. Ecological Entomology  22: 176– 183. Google Scholar CrossRef Search ADS   Lagadic L, Caquet T. 1998. Invertebrates in testing of environmental chemicals: are they alternatives? Environmental Health Perspectives  106: 593– 611. Google Scholar CrossRef Search ADS   Lenoir L. 2002. Can wood ants distinguish between good and bad food patches on the forest floor? European Journal of Soil Biology  38: 97– 102. Google Scholar CrossRef Search ADS   Lindström L, Alatalo RV, Mappes J. 1997. Imperfect Batesian mimicry − the effects of the frequency and the distastefulness of the model. Proceedings of the Royal Society of London. Series B, Biological Sciences  264: 149– 153. Google Scholar CrossRef Search ADS   Lindström L, Alatalo RV, Mappes J, Riipi M, Vertainen L. 1999. Can aposematic signals evolve by gradual change? Nature  397: 249– 251. Google Scholar CrossRef Search ADS   Lindström L, Rowe C, Guilford T. 2001. Pyrazine odour makes visually conspicuous prey aversive. Proceedings of the Royal Society of London. Series B, Biological Sciences  268: 159– 162. Google Scholar CrossRef Search ADS   Lindström L, Lyytinen A, Mappes J, Ojala K. 2006. Relative importance of taste and visual appearance for predator education in Müllerian mimicry. Animal Behaviour  72: 323– 333. Google Scholar CrossRef Search ADS   Marples NM, Roper TJ. 1996. Effects of novel colour and smell on the response of naive chicks towards food and water. Animal Behaviour  51: 1417– 1424. Google Scholar CrossRef Search ADS   Marples NM, Kelly DJ. 1999. Neophobia and dietary conservatism: two distinct processes? Evolutionary Ecology  13: 641– 653. Google Scholar CrossRef Search ADS   Marples NM, Quinlan M, Thomas RJ, Kelly DJ. 2007. Deactivation of dietary wariness through experience of novel food. Behavioral Ecology  18: 803– 810. Google Scholar CrossRef Search ADS   Montllor CB, Bernays EA, Cornelius ML. 1991. Responses of two hymenopteran predators to surface chemistry of their prey: significance for an alkaloid-sequestering caterpillar. Journal of Chemical Ecology  17: 391– 399. Google Scholar CrossRef Search ADS   Nishida R. 2002. Sequestration of defensive substances from plants by Lepidoptera. Annual Review of Entomology  47: 57– 92. Google Scholar CrossRef Search ADS   Ohkuma M, Yuzawa H, Amornsak W, Sornnuwat Y, Takematsu Y, Yamada A, Vongkaluang C, Sarnthoy O, Kirtibutr N, Noparatnaraporn N, Kudo T, Inoue T. 2004. Molecular phylogeny of Asian termites (Isoptera) of the families Termitidae and Rhinotermitidae based on mitochondrial COII sequences. Molecular Phylogenetics and Evolution  31: 701– 710. Google Scholar CrossRef Search ADS   Pasteels JM, Grégoire JC, Rowell-Rahier M. 1983. The chemical ecology of defense in arthropods. Annual Reviews in Entomology  28: 263– 289. Google Scholar CrossRef Search ADS   Pasteels JM, Daloze D, Rowell-Rahier M. 1986. Chemical defense in chrysomelid eggs and neonate larvae. Physiological Entomology  11: 29– 37. Google Scholar CrossRef Search ADS   Pasteels JM, Braekman J-C, Daloze D. 1988. Chemical defense in the Chrysomelidae. In: Solivet P, Petitpierre E, Hsiao TH, eds. Biology of Chrysomelidae . Dordrecht: Kluwer, 233– 252. Google Scholar CrossRef Search ADS   Pasteels JM, Termonia A, Daloze D, Windsor DM. 2003. Distribution of toxins in chrysomeline leaf beetles: possible taxonimic inferences. In: Furth DG, ed. Proceedings of the Fifth International Symposium on the Chrysomelidae, 25–27 August 2000, Iguassu Falls, Brazil, XXI International Congress of Entomology. Sofia: Pensoft Publishers, 261– 275. Pauls G, Becker T, Rahfeld P, Gretscher RR, Paetz C, Pasteels J, von Reuss SH, Burse A, Boland W. 2016. Two defensive lines in juvenile leaf beetles; esters of 3-nitropropionic acid in the hemolymph and aposematic warning. Journal of Chemical Ecology  42: 240– 248. Google Scholar CrossRef Search ADS   Pinto CF, Urzua A, Niemeyer HM. 2011. Sequestration of aristolochic acids from meridic diets by larvae of Battus polydamas archidamas (Papilionidae: Troidini). European Journal of Entomology  108: 41– 45. Google Scholar CrossRef Search ADS   Priestap HA, Velandia AE, Johnson JV, Barbieri MA. 2012. Secondary metabolite uptake by the Aristolochia-feeding papilionoid butterfly Battus polydamas. Biochemical Systematics and Ecology  40: 126– 137. Google Scholar CrossRef Search ADS   Rank NE, Kopf A, Julkunen-Tiitto R, Tahvanainen J. 1998. Host preference and larval performance of the salicylate-using leaf beetle Phratora vitellinae. Ecology  79: 618– 631. Google Scholar CrossRef Search ADS   Reudler JH, Lindstedt C, Pakkanen H, Lehtinen I, Mappes J. 2015. Costs and benefits of plant allelochemicals in herbivore diet in a multi enemy world. Oecologia  179: 1147– 1158. Google Scholar CrossRef Search ADS   Riipi M, Alatalo RV, Lindström L, Mappes J. 2001. Multiple benefits of gregariousness cover detectability costs in aposematic aggregations. Nature  413: 512– 514. Google Scholar CrossRef Search ADS   Rossini C, Bezzerides A, Gonzalez A, Eisner M, Eisner T. 2003. Chemical defense: incorporation of diet-derived pyrrolizidine alkaloid into the integumental scales of a moth (Utetheisa ornatrix). Chemoecology  13: 199– 205. Google Scholar CrossRef Search ADS   Rowe C, Guilford T. 1996. Hidden colour aversions in domestic clicks triggered by pyrazine odours of insect warning displays. Nature  383: 520– 522. Google Scholar CrossRef Search ADS   Rowell-Rahier M, Pasteels JM. 1986. Economics of chemical defense in Chrysomelinae. Journal of Chemical Ecology  12: 1189– 1203. Google Scholar CrossRef Search ADS   Rowell-Rahier M, Pasteels JM, Alonso-Mejia A, Brower LP. 1995. Relative unpalatability of leaf beetles with either biosynthesized or sequestered chemical defense. Animal Behaviour  49: 709– 714. Google Scholar CrossRef Search ADS   Ruxton GD, Sherratt TN, Speed MP. 2004. Avoiding attack: the evolutionary ecology of crypsis, warning signals, and mimicry . Oxford: Oxford University Press. Google Scholar CrossRef Search ADS   SAS Institute. 2015. SAS/Stat. User’s guide, version 9.4. Cary: SAS Institute. Schulz S, Gross J, Hilker M. 1997. Origin of the defensive secretion of the leaf beetle Chrysomela lapponica. Tetrahedron  53: 9203– 9212. Google Scholar CrossRef Search ADS   Siddall EC, Marples NM. 2008. Better to be bimodal: the interaction of color and odor on learning and memory. Behavioral Ecology  19: 425– 432. Google Scholar CrossRef Search ADS   Sillén-Tullberg B. 1985. Higher survival of an aposematic than of a cryptic form of a distasteful bug. Oecologia  67: 411– 415. Google Scholar CrossRef Search ADS   Sime KR, Feeny PP, Haribal MM. 2000. Sequestration of aristolochic acids by the pipevine swallowtail, Battus philenor (L.): evidence and ecological implications. Chemoecology  10: 169– 178. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2005. Tasting the difference: do multiple defence chemicals interact in Müllerian mimicry? Proceedings of the Royal Society of London. Series B, Biological Sciences  272: 339– 345. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2006a. Taste-rejection by predators and the evolution of unpalatability in prey. Behavioral Ecology and Sociobiology  60: 550– 555. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2006b. Avian predators taste-reject aposematic prey on the basis of their chemical defence. Biology Letters  2: 348– 350. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2006c. Predator avoidance learning of prey with secreted or stored defences and the evolution of insect defences. Animal Behaviour  72: 827– 834. Google Scholar CrossRef Search ADS   Skelhorn J, Rowe C. 2009. Distastefulness as an antipredator defence strategy. Animal Behaviour  78: 761– 766. Google Scholar CrossRef Search ADS   Sugeno W, Matsuda K. 2002. Adult secretions of four Japanese Chrysomelinae (Coleoptera: Chrysomelidae). Applied Entomology and Zoology  37: 191– 197. Google Scholar CrossRef Search ADS   Svádová K, Exnerová A, Štys P, Landová E, Valenta J, Fučíková A, Socha R. 2009. Role of different colours of aposematic insects in learning, memory and generalization of naive bird predators. Animal Behaviour  77: 327– 336. Google Scholar CrossRef Search ADS   Termonia A, Hsiao TH, Pasteels JM, Milinkovitch MC. 2001. Feeding specialization and host-derived chemical defense in Chrysomeline leaf beetles did not lead to an evolutionary dead end. Proceedings of the National Academy of Sciences of the USA  98: 3909– 3914. Google Scholar CrossRef Search ADS   Tullberg BS, Leimar O, Gamberale- Stille G. 2000. Did aggregation favour the initial evolution of warning coloration? A novel world revisited. Animal Behaviour  59: 281– 287. Google Scholar CrossRef Search ADS   Von Nickisch-Rosenegk E, Wink M. 1993. Sequestration of pyrrolizidine alkaloids in several arctiid moths (Lepidoptera: Arctiidae). Journal of Chemical Ecology  19: 1889– 1903. Google Scholar CrossRef Search ADS   Wiklund C, Järvi T. 1982. Survival of distasteful insects after being attacked by naive birds: a reappraisal of the theory of aposematic coloration evolving through individual selection. Evolution  36: 998– 1002. Google Scholar CrossRef Search ADS   Zvereva EL, Rank NE. 2004. Fly parasitoid Megaselia opacicornis uses defensive secretions of the leaf beetle Chrysomela lapponica to locate its host. Oecologia  140: 516– 522. Google Scholar CrossRef Search ADS   Zvereva EL, Kozlov MV. 2016. The costs and effectiveness of chemical defenses in herbivorous insects: a meta-analysis. Ecological Monographs  86: 107– 124. Zvereva EL, Kruglova OY, Kozlov MV. 2010. Drivers of host plant shifts in the leaf beetle Chrysomela lapponica: natural enemies or competition? Ecological Entomology  35: 611– 622. Google Scholar CrossRef Search ADS   Zvereva EL, Kozlov MV, Neuvonen S. 1995. Decrease in feeding niche breadth of Melasoma lapponica (Coleoptera: Chrysomelidae) with increase in pollution Oecologia  104: 323– 329. Google Scholar CrossRef Search ADS   Zvereva EL, Kozlov MV, Rank NE. 2016. Does ant predation favour leaf beetle specialization on toxic host plants? Biological Journal of the Linnean Society  119: 201– 212. Google Scholar CrossRef Search ADS   Zvereva EL, Zverev V, Kruglova OY, Kozlov MV. 2017. Strategies of chemical anti-predator defences in leaf beetles: is sequestration of plant toxins less costly than de novo synthesis? Oecologia  183: 93– 106. Google Scholar CrossRef Search ADS   © 2018 The Linnean Society of London, Biological Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Biological Journal of the Linnean SocietyOxford University Press

Published: May 15, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off