TY - JOUR
AU - Pekár,, Stano
AB - Abstract Predators that prey on dangerous species have evolved effective capture traits. In spiders, venom and silk represent distinct substances associated with prey capture. However, the use of such adaptations comes at a cost. Based on a possible trade-off, the use of only one effective capture mechanism should be optimized if a predator is to specialize on a single type of dangerous prey. We investigated hunting strategies in two Callilepis spp. and one Nomisia species, Nomisia exornata, closely related ant-eating spiders from the family Gnaphosidae. We hypothesized that specialized Callilepis spiders would adopt a more stereotyped capture strategy compared with less specialized Nomisia. We also expected that Callilepis would rely on only one subjugation mechanism. To test this, we compared their hunting efficacy and hunting strategies, with an emphasis on venom vs. silk utilization. Nomisia restrained ants with silk (then bit them), whereas Callilepis relied solely on its venom. This was also reflected in trophic traits connected with silk and venom utilization; Callilepis had larger venom glands than Nomisia, whereas Nomisia had more piriform silk glands than Callilepis. Callilepis was more effective because it subdued prey more quickly, presumably owing to ant-specific venom. Callilepis and Nomisia handled ants from two subfamilies with different degrees of success; Callilepis was more successful with Formicinae ants, whereas Nomisia handled Myrmicinae ants better. We show that sole reliance on venom allows Callilepis to be more efficient in overcoming ants than Nomisia, which uses both silk and venom. However, such specific adaptations might restrict specialized predators from exploiting alternative prey. Araneae, Callilepis, glands, Nomisia, prey capture, specialist INTRODUCTION According to the optimal foraging theory, foraging and prey choice are associated with benefits and costs (Davies et al., 2012). Predators that prey on dangerous prey often expend considerable amounts of energy on overcoming their prey and less energy on search and pursuit (Griffiths, 1980). A predator’s energy should therefore be invested in weaponry efficient at restraining prey. Such efficient weapons are represented by the venom of many venomous animals or the silk of spiders. Nevertheless, both are products of metabolism as secretions of specialized glands and are associated with costs. It has been proposed that synthesis of venom is metabolically and ecologically costly, because it is costly to replenish venom after depletion, and venomous predators are dependent on venom in order to subdue prey (Morgenstern & King, 2013). Several studies on snakes and one on scorpions showed that venom depletion leads to an increase in metabolic rate (McCue, 2006; Nisani et al., 2007; Pintor et al., 2010). Furthermore, venom metering (i.e. injection of a volume of venom conditional on prey agility and size) has been reported for snakes, spiders and scorpions, i.e. taxa with independently evolved venom systems (Morgenstern & King, 2013). This is an evidence for a high cost of venom, because venomous animals try to use no more than necessary. In addition, an ecological cost is associated with the time needed to produce venom or the time spent without adequate venom stores (Young et al., 2002; Hayes, 2008; Young, 2008). Likewise, silk is an expensive product to secrete. In web-building spiders, the construction of a web represents a considerable initial investment in this predation strategy, because it also leads to an increase in metabolic rate (Ford, 1977). In addition, orb-weaving spiders eat their web every day to reduce the costs of silk production (Foelix, 2010). The synthesis of dragline silk produced by spiders also requires significantly more ATP than the synthesis of silks produced by herbivorous lepidopteran insects owing to different proportions of amino acids in the silks and costs of the associated metabolic pathways (e.g. lepidopteran glycine = 8.5 ATP, spider glycine = 14.5 ATP) (Craig et al., 1999). Spiders are the most diverse taxon of terrestrial predators (Coddington & Levi, 1991) and have evolved a great variety of predatory strategies, because approximately half of the species primarily use webs to catch prey, whereas the other half capture prey by gripping it with the forelegs and using envenomation (Cardoso et al., 2011). Silk and venom thus represent two distinct capture traits typical for spiders. Given that both venom and silk are composed of proteins and, therefore, amino acids, there may be a trade-off in the utilization of these substances. But there is a large variability in the chemical composition of spider venom and silk (Andersen, 1970; Work & Young 1987; Kuhn-Nentwig et al., 2011), which makes the comparison of the associated costs more complex. Most spiders seem to use both venom and silk during the prey capture, albeit in differing proportions (Olive, 1980). However, some prey-specialized spiders, such as ant-eating zodariids, rely on potent venom only (Pekár et al., 2014). Also, many other cursorial spiders, such as spiders from the superfamily Lycosoidea, do not rely on silk during the prey capture (Eggs et al., 2015). Alternatively, venom has been found to be secondarily lost in uloborid spiders, which use only silk to wrap their prey during capture (King, 2004). Given ‘a jack of all trades, master of none’ trade-off, the utilization of one effective mechanism to subdue prey should be more optimal than the use of two complementary mechanisms. We hypothesize that this will be pronounced in specialists hunting dangerous prey as a result of greater specialization to increase the precision of an attack, to reduce escape of prey and to lower the associated costs. To test this, we investigated prey capture in two phylogenetically related spider genera of the family Gnaphosidae. Gnaphosids are bold predators able to subdue large and hazardous prey with the use of piriform silk. Gluey piriform silk is used by spiders as an anchorage for other silk products (Wolff et al., 2015), but gnaphosids use it as a hunting sticky tape (Wolff et al., 2017). Here, we focused on Callilepis and Nomisia, two ant-eating genera that use different hunting strategies. Callilepis spiders are reported to be ant specialists using a different hunting strategy from other gnaphosids, in that they do not use silk for prey immobilization (Heller, 1976; Borovsky, 2012). Nomisia spiders hunt ants with the use of silk to immobilize them (Soyer, 1943) and accept other prey types besides ants (O. Michálek, unpublished). Given that both spider genera use different strategies to subdue dangerous prey, we investigated the hunting strategies of these spiders in detail; we compared their hunting efficacies and the time investment in behaviours associated with venom vs. silk utilization. In addition, we compared the morphological traits connected with silk and venom production. We hypothesized that the more specialized spider would show more stereotyped (i.e. following the same sequence with almost no variations) capture behaviour compared with the less specialized spider and that the specialist would rely on one subjugation mechanism, which would be reflected in the morphology (e.g. gland size) of the connected trophic traits. MATERIAL AND METHODS Spiders Two species of Callilepis were collected on the forest edge at two sites. Callilepis nocturna (Linnaeus, 1758) and a few Callilepis schuszteri (Herman, 1879) spiders of various stages (mean ± SEM prosoma length = 1.61 ± 0.37 mm) were collected in the valley of the Größer Dürrenbach river, between Villach and Klagefurt, Austria in June 2015. Nomisia exornata (C. L. Koch, 1839) spiders of various stages (mean ± SEM prosoma length = 2.06 ± 0.45 mm) were collected near Serpa, southern Portugal in October 2015 and 2017. Spiders were identified using the online spider identification key (Nentwig et al., 2018). In laboratory experiments, juveniles of Callilepis were occasionally used because the number of adults was too low (53%, N = 32); thus, identification to species level was not possible, and some data were pooled as Callilepis spp. Spiders used in laboratory experiments were kept in plastic vials containing moisturized gypsum and placed in a chamber at a constant temperature (22 ± 1 °C) and under a light–dark regime (16 h–8 h) to standardize the metabolic rate and the satiation of spiders. Spiders were fed once a week with an ant or were allowed to consume the prey accepted in laboratory trials. Experiments were performed from July 2015 to October 2017. All statistical analyses were performed within the R environment (R Core Team, 2017). Capture behaviour To compare the hunting strategies of both genera, capture sequences were recorded using a high-speed camera (IDT MotionXtra N3), using 500 frames/s for Callilepis spp. and a lower frame rate (100 or 200 frames/s) for N. exornata in order to record the whole hunting sequence. A high-speed camera was used because the hunting actions of both Callilepis spp. and N. exornata were very rapid; prey capture took only a few seconds. To choose a suitable ant prey, we performed a preliminary experiment in which we used ants of several genera as a prey item (Lasius, Formica, Tetramorium, Messor and Tapinoma) to observe whether there were differences in the use of hunting strategy. Between ten and 20 ants of each genus were offered to individuals of each spider species in Petri dishes. The capture strategy did not vary for different ant genera in either spider species. Therefore, ants of the genus Tapinoma were used as prey for Callilepis spp. and ants of the genus Messor for N. exornata for more detailed observation, because these ants represented the prey with a similar relative size ratio (total ant length to spider prosoma length) compared with the given spider species (mean ± SEM Tapinoma/Callilepis relative size ratio 2.50 ± 0.43, and mean ± SEM Messor/Nomisia relative size ratio 2.66 ± 0.43). Spiders were placed individually in plastic cups (diameter 3.5 cm, height 5 cm) with gypsum on the bottom and a layer of butter on the walls to prevent escape. Each prey item was introduced after 1 h of acclimation. In total, 27 hunting videos involving Callilepis spp. and 22 videos involving N. exornata were obtained. Individual spiders were used only once. In the videos, the following types of behaviours were distinguished: (1) approach, i.e. the prey or the predator moved towards the other; (2) touching, i.e. the predator gently touched the prey with its first pair of legs; (3) orientation, i.e. the predator turned to face the direction in which the prey was situated; (4) immobility, i.e. the predator stopped on the spot and remained without performing any other activity; (5) wrapping, i.e. the predator ran around the prey and released silk, immobilizing the prey in the process; (6) biting, i.e. the predator delivered a bite to the prey; (7) release, i.e. the prey was released from the chelicerae; and (8) feeding, i.e. the predator started to consume the prey. Using this ethogram, transition matrices were created with JWatcher software (Blumstein et al., 2006). Then, flow diagrams for each spider genus were made. The frequencies of bites on different body parts (leg or antenna) were compared between spiders using a generalized linear model (GLM) with a binomial distribution and the logit link function (Pekár & Brabec, 2016). The spider genus was used as the factor, and the relative prey/predator size was the covariate. The duration of contact with the prey (from the first approach to the release of the ant) was compared between both spiders using a GLM with the Gamma distribution and a logarithmic link. Here, the type of predator and the bite site were used as factors, and the relative prey/predator size was a covariate. We also compared the time that ants were held in chelicerae using a GLM with the Gamma distribution and the same factors and covariate as in the previous model. In N. exornata, we also measured the time that the spider spent wrapping the prey with silk. Time measurements were obtained from recorded hunting sequences using Kinovea software (Kinovea; v.0.8.15; Kinovea open source project, https://www.kinovea.org). To measure the stereotypy of hunting behaviour, we used the following formula of Shannon entropy: H = −∑piln(pi), where pi is the proportion of ith behaviour. Higher entropy values indicate more complex and flexible behaviour, whereas lower values indicate stereotyped behaviour. Entropy estimates along with 95% confidence intervals (CI95) were calculated from the transition matrices by bootstrapping with 1000 replicates for both Callilepis spp. and N. exornata. Capture efficiency To compare the hunting efficiencies of Callilepis spp. and N. exornata for prey of different sizes, Formica and Messor ants of various sizes (Formica mean ± SEM body length = 5.52 ± 0.82 mm; Messor mean ± SE body length = 5.54 ± 1.16 mm) were offered to both spiders. Formica and Messor ants were chosen as representatives of the two ant subfamilies (Formicinae, Myrmicinae) that are preferred by Callilepis spp. and N. exornata, respectively, as their natural prey (L. Petráková, unpublished). Individuals of Callilepis spp. and N. exornata were placed singly in Petri dishes (diameter 5 cm), and after 1 h acclimation a single ant randomly selected from the two species was offered. If the ant was not accepted within 1 h, it was replaced by the ant with smaller relative ant/spider size ratio. The trial ended when the spider killed and consumed the ant. If a spider did not accept any subsequent ant, it was considered unmotivated to eat (i.e. satiated or preparing to moult), and data concerning such individuals were rejected for that day. Another trial with the same spider was performed 1 week later. The lengths of the prosoma of all spiders and the total body lengths of ants were measured under a LEICA EZ5 stereomicroscope with an ocular micrometer before experiments. In total, 30 trials (17 Formica ants, 13 Messor ants) with 23 individuals of Callilepis spp. and 37 trials (17 Formica ants, 20 Messor ants) with 19 individuals of N. exornata were performed. The difference in hunting success was analysed using generalized estimating equations (GEEs) from the geepack package (Halekoh et al., 2006). The GEE is an extension of the GLM for correlated data. It was used because there were repeated measurements on each individual spider. A GEE with a binomial distribution and the logit link function was used. An AR1 correlation matrix was used to account for the temporal replications (Pekár & Brabec, 2018). The acceptance of an ant was the response variable; spider and ant genus were factors, and relative ant/spider size ratio was a covariate. Morphological trophic traits The venom glands from nine adult female individuals of C. schuszteri and nine adult female individuals of N. exornata were dissected. Spiders were first immobilized with CO2, and the glands were placed into a drop of the physiological solution NaCl 0.9% on a glass slide. The dimensions of the glands [widths (2r) and lengths (d)] were measured using an ocular micrometer attached to an Olympus SX stereomicroscope. The volume of the gland (V) was estimated by assuming a cylindrical shape (V = dπr2). The length of the prosoma was measured for each individual. The anterior lateral spinnerets along with associated silk glands from four adult female individuals of C. schuszteri and five adult female individuals of N. exornata were dissected. The number of piriform glands, the number of major ampulate glands, and the length and width of the secretory part of the piriform glands were measured. The dimensions of the glands were measured using ImageJ software (Schneider et al., 2012) on photographs taken with a camera attached to a stereomicroscope (Fig. 5C, D). The volume of each of the piriform glands was estimated in the same manner as for the venom glands. We estimated the volume only of the piriform glands, because these are the glands used to restrain prey (Wolff et al., 2017). The length of the prosoma was also measured for each individual. The relative volumes of venom and silk glands and the relative number of silk glands (i.e. normalized to prosoma length) were compared between species using linear models (LMs). RESULTS Capture behaviour The predatory behaviour of Callilepis spp. began with a brief tapping of the ant’s antennae with its first pair of legs, followed by a rapid bite to the antenna base and release (Fig. 1A–D; Supporting Information, Video S1). Nomisia exornata used a very different tactic: first, it immobilized the prey by running around the prey, while simultaneously turning the abdomen and spinnerets towards the prey and releasing piriform silk onto the prey; and then delivered a bite (Fig. 1E–H; Supporting Information, Video S2). Callilepis spp. was slightly more consistent in selecting the location of the bite than N. exornata (GLM, F1,48 = 40.2, P = 0.05); the prey was bitten more often on the antenna (93%, N = 27) than on the leg (7%). When the ant was bitten on the antenna, it was always on its base. Nomisia exornata also bit the prey on the antenna in most cases. However, unlike Callilepis spp., it bit the ant on the distal part of the antenna and, in 27% of cases, the ant was also bitten on the distal part of the leg (N = 22). The prey size did not affect selection of the bite site (GLM, F1,47 = 39.3, P = 0.33). Figure 1. View largeDownload slide Elements of the predatory behaviour of Callilepis spp. (A–D) and Nomisia exornata (E–H) in detail. A, Callilepis approaches the ant and raises its forelegs. B, it gently touches the antennae of the ant with the first pair of legs. C, it lunges forwards and bites the ant at the base of antenna (white arrow). D, the prey is released, and Callilepis waits nearby until the ant is paralysed. E, Nomisia approaches the ant. F, it runs around the ant, turning its abdomen and spinnerets toward the ant (white arrow), and releases silk, immobilizing the ant in the process. G, it bites the immobilized ant on the leg (white arrow). H, the prey is released, and Nomisia waits until the ant is paralysed. Figure 1. View largeDownload slide Elements of the predatory behaviour of Callilepis spp. (A–D) and Nomisia exornata (E–H) in detail. A, Callilepis approaches the ant and raises its forelegs. B, it gently touches the antennae of the ant with the first pair of legs. C, it lunges forwards and bites the ant at the base of antenna (white arrow). D, the prey is released, and Callilepis waits nearby until the ant is paralysed. E, Nomisia approaches the ant. F, it runs around the ant, turning its abdomen and spinnerets toward the ant (white arrow), and releases silk, immobilizing the ant in the process. G, it bites the immobilized ant on the leg (white arrow). H, the prey is released, and Nomisia waits until the ant is paralysed. The Shannon entropy of behavioural sequences (Fig. 2) differed significantly between Callilepis spp. and N. exornata: the entropy estimate for Callilepis spp. sequences was 2.39 (CI95 = 2.07, 3.08), whereas for N. exornata it was 5.59 (CI95 = 4.89, 7.38); therefore, the behaviour of Callilepis spp. was more stereotypical, i.e. sequences were more similar, with lower variation between individual sequences. Figure 2. View largeDownload slide Flow diagrams of the prey capture behaviour of Callilepis spp. (A) and Nomisia exornata (B). Transition probabilities are shown for each transition. Figure 2. View largeDownload slide Flow diagrams of the prey capture behaviour of Callilepis spp. (A) and Nomisia exornata (B). Transition probabilities are shown for each transition. The duration of total hunting activity was significantly shorter for Callilepis spp. (GLM, F1,47 = 142.6, P < 0.0001); the mean hunting time was 1.18 s (CI95 = 0.99, 1.43) for Callilepis spp., whereas it was 6.66 s (CI95 = 5.45, 8.25) for N. exornata. The mean duration of prey wrapping for N. exornata was 1.34 s (CI95 = 1.00, 1.86). The mean duration of the bite was also significantly shorter for Callilepis spp. (GLM, F1,47 = 294.5, P < 0.0001); it took 0.24 s (CI95 = 0.20, 0.30) for Callilepis spp., and 3.95 s (CI95 = 3.20, 4.97) for N. exornata (Fig. 3). Furthermore, the duration of the bite was significantly influenced by the interaction between the type of predator and the bite site (GLM, F1,45 = 4.1, P = 0.05). Callilepis spp. spent less time biting the leg (0.10 s, CI95 = 0.05, 0.23) than biting the antenna (0.25 s, CI95 = 0.21, 0.31), whereas N. exornata spent more time biting the leg (4.28 s, CI95 = 2.90, 6.70) than biting the antenna (3.83 s, CI95 = 3.00, 5.00). Figure 3. View largeDownload slide Comparison of the total hunting activity (A) and the time spent biting an ant (B) by Callilepis spp. and Nomisia exornata. Bars are means, and vertical lines represent 95% confidence intervals. Figure 3. View largeDownload slide Comparison of the total hunting activity (A) and the time spent biting an ant (B) by Callilepis spp. and Nomisia exornata. Bars are means, and vertical lines represent 95% confidence intervals. Capture efficiency The capture success on ants changed differently in Callilepis spp. and N. exornata with the relative prey/predator size ratio and type of ant prey (GEE, χ21 = 5.0, P = 0.03). Callilepis spp. was more successful in handling larger Formicinae (Formica) ants than N. exornata (Fig. 4A). Callilepis spp. captured Formica ants with a 50% success rate at an ant body length/spider prosoma length ratio of 8.52, whereas N. exornata achieved a similar success at a ratio of 2.17. However, N. exornata was more effective in handling large Myrmicinae (Messor) ants (Fig. 4B); it captured Messor ants with a 50% success rate at an ant body length/spider prosoma length ratio of 8.08, whereas Callilepis spp. achieved similar success at a ratio of 3.87. Figure 4. View largeDownload slide Comparison of the capture success of Callilepis spp. and Nomisia exornata on Formica ants (A) and Messor ants (B) of various relative sizes (prey/predator body size ratio). Estimated logit models are shown. Figure 4. View largeDownload slide Comparison of the capture success of Callilepis spp. and Nomisia exornata on Formica ants (A) and Messor ants (B) of various relative sizes (prey/predator body size ratio). Estimated logit models are shown. Morphological trophic traits The relative sizes of venom glands differed significantly between N. exornata and C. schuszteri (LM, F1,16 = 35.8, P < 0.0001); venom glands of C. schuszteri were 1.65 times larger than those of N. exornata (Fig. 5A). With regard to the spinning apparatus, C. schuszteri and N. exornata differed significantly in their relative number of piriform glands (LM, F1,16 = 23.4, P < 0.001; Fig. 5B); C. schuszteri had two to four glands connected to each spinneret (Fig. 5C), whereas N. exornata had five glands connected to each spinneret (Fig. 5D). Both C. schusteri and N. exornata had one functional major ampulate gland connected to each spinneret. There was no significant difference in the relative volume of a single piriform gland between C. schuszteri and N. exornata (LM, F1,54 = 1.5, P = 0.22; Fig. 5A). Figure 5. View largeDownload slide A, B, comparison of the relative volumes of the venom glands and piriform silk glands (A) and the relative number of piriform silk glands (B) of Callilepis schuszteri and Nomisia exornata. Bars are means, and vertical lines represent 95% confidence intervals. C, D, images show the dissected anterior lateral spinnerets along with silk glands of C. schuszteri (C) and N. exornata (D). Abbreviations: als, anterior lateral spinneret; ma, major ampulate gland; pi, pirform gland. Figure 5. View largeDownload slide A, B, comparison of the relative volumes of the venom glands and piriform silk glands (A) and the relative number of piriform silk glands (B) of Callilepis schuszteri and Nomisia exornata. Bars are means, and vertical lines represent 95% confidence intervals. C, D, images show the dissected anterior lateral spinnerets along with silk glands of C. schuszteri (C) and N. exornata (D). Abbreviations: als, anterior lateral spinneret; ma, major ampulate gland; pi, pirform gland. DISCUSSION A predator that hunts dangerous prey cannot make mistakes, because any inaccuracy could have a high impact on predator survival (Mukherjee & Heithaus, 2013). Specialization on such prey thus frequently leads to greater accuracy and stereotypy in prey capture (Ferry-Graham et al., 2002). Such precision in hunting behaviour can be observed in some araneophagous spiders (Michálek et al., 2017; García et al., 2018). In this comparative study, we revealed that the hunting strategy of strictly ant-eating Callilepis spiders was also very conservative and stereotyped when compared with the relative less specialized N. exornata spiders. Nomisia exornata used both silk and venom to subdue ants, whereas Callilepis spiders used only venom. Heller (1976) noted that Callilepis spiders are not able to envenomate ants that have had the antennae removed, although in the present study, we observed two cases of leg biting. However, in one of these cases the leg of the ant was in close proximity to its antenna, and in the second case the Callilepis spider almost immediately changed the bite site to the antenna. Evidence gathered in the present study suggests that Callilepis spiders are more specialized, because their hunting strategy is less complex. Also, Callilepis spiders need to be more precise, because the ants are not immobilized with silk and thus remain dangerous. In cotrast, N. exornata is less specialized; its hunting strategy is more complex and generalized, because it is also used for non-ant prey, such as spiders and beetles, whereas Callilepis spiders are not able to subdue alternative prey (O. Michálek, unpublished). It took N. exornata a relatively long time to subdue ants. Most apparently, the ant was held in the chelicerae for a considerable period. Spiders can adjust the amount of venom injected (Wigger et al., 2002) while holding prey in the chelicerae (Boevé, 1994; Morgenstern & King, 2013). However, long envenomation represents a greater risk, particularly when subduing a dangerous prey, because it has a longer time to retaliate. Predators can minimize this risk behaviourally by minimizing contact or shortening the handling time and by selecting the direction and position of an attack (Mukherjee & Heithaus, 2013). For example, ant-specialized Zodarion spiders bite ants on the most extended leg (Pekár, 2004). This behaviour might lower the risk even more, because the spider keeps a greater distance from a dangerous prey. Callilepis and Nomisia dealt with this task in a different way. Nomisia exornata reduced the risk by first restraining the prey with silk, then biting the ant on the distal part of the antenna or leg. However, silk production and associated behaviours represent additional costs. Furthermore, envenomation still plays a significant role in N. exornata, because the time spent biting was longer than the time spent wrapping. In contrast, Callilepis spiders used only venom. Given that the bite delivered by Callilepis spiders was very short, we suppose its venom to be especially potent towards ant prey. It is possible that the venom of specialist spiders is tailored more closely to their specific prey taxon (Kuhn-Nentwig et al., 2011). The venom of specialists is less diversified in its composition (Pekár et al., 2018a). It has been confirmed that the venom composition of Conus snails is connected to the level of specialization, because the venom of specialized Conus snails contains fewer conotoxins than that of generalist species of the same genus (Remigio & Duda, 2008). The paralysis latency of ant prey is shorter in C. nocturna than in N. exornata (Pekár et al., 2018b), despite the fact that the duration of the bite is also much shorter in C. nocturna. Callilepis thus possesses a very potent venom that is less diversified in its composition (Pekár et al., 2018a) and presumably less costly. But we were not able to estimate the real costs of the venom and the silk (e.g. energy cost) in our study. More research is needed to make a more detailed comparison between the costs of the syntheses of these substances. The bite of Callilepis spiders was delivered to the base of the ant’s antenna. This bold behaviour probably also facilitates quicker immobilization, because the venom is injected close to nerve ganglions in the head capsule of the ant. The spider Oecobius annulipes Lucas, 1859 also bites ants at the base of the antenna, but in this case the ants are first immobilized with silk (Glatz, 1967). Callilepis spiders tapped approaching ants on the head or antennae before biting them, presumably to identify the bite site. Biting the antennae had, in particular, a significant effect on the response of Formicinae, which are more agile than Myrmicinae. The bitten Formicinae ant moved in circles, which meant that it could not escape after release by the spider (Supporting Information, Video S3). Wrapping in silk also prevents the escape of prey. Although similar touching behaviour was observed in N. exornata in several cases, this spider also touched the ant on other body parts. Prey immobilization with silk is a common strategy of gnaphosid spiders. Morphological and functional modification of the spinning apparatus allows them to subdue large and dangerous prey, such as spiders (Wolff et al., 2017). However, it appears that the use of silk for immobilization is not advantageous for specialist spiders. Araneophagous Lampona murina L. Koch, 1873 uses not silk but venom for prey capture (Michálek et al., 2017). Wolff et al. (2017) argue that araneophagy might have evolved earlier than spinneret modification in Gnaphosidae. However, ant-specialized Callilepis spiders do not use silk at all, whereas less specialized N. exornata spiders do. Given that Callilepis spiders rely only on venom, the venom glands are larger than in N. exornata. Alternative capture strategies or dietary shifts may lead to morphological and physiological alterations, such as reduced venom glands in some snakes or uloborid spiders (King, 2004; Fry et al., 2008). Likewise, Callilepis spiders may have evolved atrophied spinning apparatus in order to allow greater investment in the venom system. Here, we found that the volume of the piriform glands does not differ between C. schuszteri and N. exornata, but C. schuszteri had fewer piriform glands than N. exornata. The number of piriform glands of both spider species was lower compared with some other gnaphosids that possess up to ten piriform glands per spinneret (Wolff et al., 2017). Swathing with silk probably represents an efficient generalized hunting strategy towards dangerous prey in gnaphosid spiders, but it is not used on harmless prey because it is too costly (Wolff et al., 2017). Predators specialized exclusively on dangerous prey may thus prefer investment in other means of weapons. The number and the volume of silk glands in N. exornata are intermediate among gnaphosids, because these spiders rely on both mechanisms to subdue prey. In C. schuszteri, the number of silk glands seems to be reduced to the minimum, because total reduction is probably not possible owing to phylogenetic constraints. Although a study on wandering and web-building Tetragnatha spider species has shown that they do not differ in the amount of venom (Binford, 2001), here we discovered that C. schuszteri has larger venom glands than silk-using N. exornata. Callilepis spiders and N. exornata accepted several ant genera from different subfamilies as a prey in the laboratory (O. Michálek, unpublished). But their hunting successes differed with increasing ant size; Callilepis handled large Formicinae ants more efficiently, whereas Nomisia was more successful with larger Myrmicinae ants. Given that the defences of these two ant subfamilies differ markedly (Formicinae use agility and formic acid, whereas Myrmicinae use stings and powerful mandibles; Buschinger & Maschwitz, 1984), the hunting strategies of the two spider genera in question seem to be adapted to overcome the defences of the preferred prey. The hunting strategy of Callilepis spiders might be specially tuned to subdue Formicinae ants, which were more effectively captured than Myrmicinae ants and preferred as a natural prey (L. Petráková, unpublished). Cuticle thickness varies among ants; Myrmicinae ants (e.g. Messor, Tetramorium) have, on average, relatively thicker cuticles than Formicinae ants (e.g. Lasius, Camponotus) (Peeters et al., 2017). Perhaps it is difficult for Callilepis to penetrate such thicker cuticles with its swift bite; therefore, it has higher success with less sclerotized ants. The use of silk might be a more efficient strategy against Myrmicinae ants, which were indeed subdued by N. exornata more efficiently than Formicinae ants and were preferred as a natural prey (L. Petráková, unpublished). Also, the use of silk appears to be safer. We observed at least two attacks on Callilepis spiders by Formica and Camponotus ants, resulting in the loss of a leg or even death (Supporting Information, Video S3). Meanwhile, no N. exornata spiders were killed by ants. The capture strategy of specialized predators should be fine-tuned towards their focal prey. Strict specialization on a certain prey type may enhance the pronounced utilization of one strategy (and subjugation mechanism), allowing a reduction in the energy needed to subdue prey. In our study, both spider genera were able to subdue ants, but Callilepis was more efficient, because it required less time to overcome an ant and it relied only on its venom, in contrast to N. exornata, which used both venom and silk. Thus, Callilepis saves energy, because it does not have to produce silk for prey immobilization. Nevertheless, the strategy of N. exornata is safer, because silk-restricted ants cannot retaliate, and more universal, because N. exornata is able to subdue alternative prey. Our results are, therefore, in accordance with assumptions concerning prey capture optimization in specialized predators. However, such specific adaptations restrict a predator from exploiting alternative prey. Indeed, Callilepis was not as successful at subduing Myrmicinae ants compared with Formicinae ants. Nomisia exornata maintained the ability to capture alternative prey, with or without the use of silk depending on the degree of danger posed by the prey (Wolff et al., 2017). SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher's web-site: Video S1. Capture of an ant by Callilepis sp. recorded using a high-speed camera (IDT MotionXtra N3) at 500 frames/s. Video S2. Capture of an ant by Nomisia exornata recorded using a high-speed camera (IDT MotionXtra N3) at 100 frames/s. Video S3. Prey capture by Callilepis sp. ACKNOWLEDGEMENTS We would like to thank S. Korenko, C. Komposch, S. Volker, S. Aurenhammerand and R. Borovsky for help with collection of spiders in the field. We also thank three anonymous reviewers for helpful comments. This work was supported by the Czech Science Foundation (grant number GA15-14762S). M.Ř. was supported by the Czech Science Foundation (grant number 17-10976S). O.M. acquired behavioural data, analysed all data and wrote the manuscript. M.Ř. acquired the data on silk glands. E.L. collected the specimens for testing and acquired part of the behavioural data. W.O.C.S. provided editorial advice. S.P. conceived the study, collected the specimens for testing, acquired the data on venom glands and contributed to writing the manuscript. All the authors helped to draft the manuscript and gave final approval for publication. REFERENCES Andersen SO . 1970 . Amino acid composition of spider silks . Comparative Biochemistry and Physiology 35 : 705 – 711 . Google Scholar Crossref Search ADS Binford GJ . 2001 . Differences in venom composition between orb-weaving and wandering Hawaiian Tetragnatha (Araneae) . Biological Journal of the Linnean Society 74 : 581 – 595 . Google Scholar Crossref Search ADS Blumstein DT , Evans CS , Daniels JC . 2006 . JWatcher 1.0 . Available at: http://www.jwatcher.ucla.edu. Boevé JL . 1994 . Injection of venom into an insect prey by the free hunting spider Cupiennius salei (Araneae, Ctenidae) . Journal of Zoology 234 : 165 – 175 . Google Scholar Crossref Search ADS Borovsky V . 2012 . Beobachtungen zur Räuber-Beute-Beziehung zwischen Callilepis spp. (Araneae: Gnaphosidae) und Camponotus vagus (Hymenoptera: Formicidae) . Carinthia II 202./122 : 713 – 722 . Buschinger A , Maschwitz U . 1984 . Defensive behavior and defensive mechanisms in ants . In: Hermann HR , ed. Defensive mechanisms in social insects . New York : Praeger , 95 – 150 . Cardoso P , Pekár S , Jocqué R , Coddington JA . 2011 . Global patterns of guild composition and functional diversity of spiders . PLoS One 6 : e21710 . Google Scholar Crossref Search ADS PubMed Coddington JA , Levi HW . 1991 . Systematics and evolution of spiders (Araneae) . Annual Review of Ecology and Systematics 22 : 565 – 592 . Google Scholar Crossref Search ADS Craig CL , Hsu M , Kaplan D , Pierce NE . 1999 . A comparison of the composition of silk proteins produced by spiders and insects . International Journal of Biological Macromolecules 24 : 109 – 118 . Google Scholar Crossref Search ADS PubMed Davies NB , Krebs JR , West SA . 2012 . An introduction to behavioural ecology , 4 th edn. Oxford : Wiley-Blackwell . Eggs B , Wolff JO , Kuhn‐Nentwig L , Gorb SN , Nentwig W . 2015 . Hunting without a web: how lycosoid spiders subdue their prey . Ethology 121 : 1166 – 1177 . Google Scholar Crossref Search ADS Ferry-Graham LA , Bolnick DI , Wainwright PC . 2002 . Using functional morphology to examine the ecology and evolution of specialization . Integrative and Comparative Biology 42 : 265 – 277 . Google Scholar Crossref Search ADS PubMed Foelix R . 2010 . Biology of spiders , 3 rd edn. New York : Oxford University Press . Ford MJ . 1977 . Energy costs of the predation strategy of the web-spinning spider Lepthyphantes zimmermanni Bertkau (Linyphiidae) . Oecologia 28 : 341 – 349 . Google Scholar Crossref Search ADS PubMed Fry BG , Scheib H , van der Weerd L , Young B , McNaughtan J , Ramjan SF , Vidal N , Poelmann RE , Norman JA . 2008 . Evolution of an arsenal: structural and functional diversification of the venom system in the advanced snakes (Caenophidia) . Molecular & Cellular Proteomics 7 : 215 – 246 . Google Scholar Crossref Search ADS García LF , Viera C , Pekár S . 2018 . Comparison of the capture efficiency, prey processing, and nutrient extraction in a generalist and a specialist spider predator . Die Naturwissenschaften 105 : 30 . Google Scholar Crossref Search ADS PubMed Glatz L . 1967 . Zur biologie und morphologie von Oecobius annulipes Lucas (Araneae, Oecobiidae) . Zeitschrift für Morphologie der Tiere 61 : 185 – 214 . Google Scholar Crossref Search ADS Griffiths D . 1980 . Foraging costs and relative prey size . American Naturalist 116 : 743 – 752 . Google Scholar Crossref Search ADS Halekoh U , Højsgaard S , Yan J . 2006 . The R package geepack for generalized estimating equations . Journal of Statistical Software 15 : 1 – 11 . Google Scholar Crossref Search ADS Hayes WK . 2008 . The snake venom-metering controversy: levels of analysis, assumptions, and evidence . In: Hayes WK , Beaman KR , Caldwell MD , Bush SP , eds. The biology of the rattlesnakes . Loma Linda : Loma Linda University Press , 191 – 220 . Heller G . 1976 . Zum Beutefangverhalten der ameisenfressenden Spinne Callilepis nocturna (Arachnida: Araneae: Drassodidae) . Entomologica Germanica 3 : 100 – 103 . King GF . 2004 . The wonderful world of spiders: preface to the special Toxicon issue on spider venoms . Toxicon 43 : 471 – 475 . Google Scholar Crossref Search ADS PubMed Kuhn-Nentwig L , Stöcklin R , Nentwig W . 2011 . Venom composition and strategies in spiders: is everything possible ? Advances in Insect Physiology 40 : 1 – 86 . Google Scholar Crossref Search ADS McCue MD . 2006 . Cost of producing venom in three North American pitviper species . Copeia 2006 : 818 – 825 . Google Scholar Crossref Search ADS Michálek O , Petráková L , Pekár S . 2017 . Capture efficiency and trophic adaptations of a specialist and generalist predator: a comparison . Ecology and Evolution 7 : 2756 – 2766 . Google Scholar Crossref Search ADS PubMed Morgenstern D , King GF , 2013 . The venom optimization hypothesis revisited . Toxicon 63 : 120 – 128 . Google Scholar Crossref Search ADS PubMed Mukherjee S , Heithaus MR . 2013 . Dangerous prey and daring predators: a review . Biological Reviews of the Cambridge Philosophical Society 88 : 550 – 563 . Google Scholar Crossref Search ADS PubMed Nentwig W , Blick T , Gloor D , Hänggi A , Kropf C . 2018 . Spiders of Europe . Version 10.2018. Available at: www.araneae.nmbe.ch. Nisani Z , Dunbar SG , Hayes WK . 2007 . Cost of venom regeneration in Parabuthus transvaalicus (Arachnida: Buthidae) . Comparative Biochemistry and Physiology. Part A, Molecular & Integrative Physiology 147 : 509 – 513 . Google Scholar Crossref Search ADS PubMed Olive CW . 1980 . Foraging specialization in orb-weaving spiders . Ecology 61 : 1133 – 1144 . Google Scholar Crossref Search ADS Peeters C , Molet M , Lin CC , Billen J . 2017 . Evolution of cheaper workers in ants: a comparative study of exoskeleton thickness . Biological Journal of the Linnean Society 121 : 556 – 563 . Google Scholar Crossref Search ADS Pekár S . 2004 . Predatory behavior of two European ant-eating spiders (Araneae, Zodariidae) . Journal of Arachnology 32 : 31 – 41 . Google Scholar Crossref Search ADS Pekár S , Bočánek O , Michálek O , Petráková L , Haddad CR , Šedo O , Zdráhal Z . 2018a . Venom gland size and venom composition—essential trophic adaptations of venomous predators: a case study using spiders . Molecular Ecology 27 ( 21 ): 4257 – 4269 . Google Scholar Crossref Search ADS Pekár S , Brabec M . 2016 . Modern analysis of biological data. Generalized linear models in R . Brno : MUNI Press . Pekár S , Brabec M . 2018 . Generalized estimating equations: a pragmatic and flexible approach to the marginal GLM modelling of correlated data in the behavioural sciences . Ethology 124 : 86 – 93 . Google Scholar Crossref Search ADS Pekár S , Líznarová E , Bočánek O , Zdráhal Z . 2018b . Venom of prey-specialized spiders is more toxic to their preferred prey: a result of prey-specific toxins . Journal of Animal Ecology 87 : 1639 – 1652 . Google Scholar Crossref Search ADS Pekár S , Šedo O , Líznarová E , Korenko S , Zdráhal Z . 2014 . David and Goliath: potent venom of an ant-eating spider (Araneae) enables capture of a giant prey . Die Naturwissenschaften 101 : 533 – 540 . Google Scholar Crossref Search ADS PubMed Pintor AF , Krockenberger AK , Seymour JE . 2010 . Costs of venom production in the common death adder (Acanthophis antarcticus) . Toxicon 56 : 1035 – 1042 . Google Scholar Crossref Search ADS PubMed R Core Team. 2017 . R: a language and environment for statistical computing . Vienna : R Foundation for Statistical Computing . Available at: http://www.R-project.org/. Remigio EA , Duda TF Jr . 2008 . Evolution of ecological specialization and venom of a predatory marine gastropod . Molecular Ecology 17 : 1156 – 1162 . Google Scholar Crossref Search ADS PubMed Schneider CA , Rasband WS , Eliceiri KW . 2012 . NIH Image to ImageJ: 25 years of image analysis . Nature Methods 9 : 671 – 675 . Google Scholar Crossref Search ADS PubMed Soyer B . 1943 . Contribution a l’étude éthologique et écologique des Araignées de la provence occidentale. I. Quelques Araignées myrmécophages des environs de Marseille . Bulletin du Muséum National d’histoire Naturelle 13 : 51 – 55 . Wigger E , Kuhn-Nentwig L , Nentwig W . 2002 . The venom optimisation hypothesis: a spider injects large venom quantities only into difficult prey types . Toxicon 40 : 749 – 752 . Google Scholar Crossref Search ADS PubMed Wolff JO , Grawe I , Wirth M , Karstedt A , Gorb SN . 2015 . Spider’s super-glue: thread anchors are composite adhesives with synergistic hierarchical organization . Soft Matter 11 : 2394 – 2403 . Google Scholar Crossref Search ADS PubMed Wolff JO , Řezáč M , Krejčí T , Gorb SN . 2017 . Hunting with sticky tape: functional shift in silk glands of araneophagous ground spiders (Gnaphosidae) . Journal of Experimental Biology 220 : 2250 – 2259 . Google Scholar Crossref Search ADS PubMed Work RW , Young CT . 1987 . The amino acid compositions of major and minor ampullate silks of certain orb-web-building spiders (Araneae, Araneidae) . Journal of Arachnology 15 : 65 – 80 . Young BA . 2008 . Perspectives on the regulation of venom expulsion in snakes . In: Hayes WK , Beaman KR , Caldwell MD , Bush SP , eds. The biology of the rattlesnakes . Loma Linda : Loma Linda University Press , 181 – 190 . Young BA , Lee CE , Daley KM . 2002 . Do snakes meter venom ? Bioscience 52 : 1121 – 1126 . 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/open_access/funder_policies/chorus/standard_publication_model)
TI - Silk versus venom: alternative capture strategies employed by closely related myrmecophagous spiders
JF - Biological Journal of the Linnean Society
DO - 10.1093/biolinnean/bly181
DA - 2019-02-28
UR - https://www.deepdyve.com/lp/oxford-university-press/silk-versus-venom-alternative-capture-strategies-employed-by-closely-q4NtY7SJTc
SP - 545
VL - 126
IS - 3
DP - DeepDyve
ER -