Capture efficiency and trophic adaptations of a specialist and generalist predator: A comparison

Abstract Specialist true predators are expected to exhibit higher capture efficiencies for the capture of larger and dangerous prey than generalist predators due to their possession of specialized morphological and behavioral adaptations. We used an araneophagous spider (Lampona murina) and a generalist spider (Drassodes lapidosus) as phylogenetically related model species and investigated their realized and fundamental trophic niches and their efficacy with respect to prey capture and prey handling. The trophic niche of both species confirmed that Lampona had a narrow trophic niche with a predominance of spider prey (including conspecifics), while the niche of Drassodes was wide, without any preference. DNA analysis of the gut contents of Lampona spiders collected in the field revealed that spiders form a significant part of its natural diet. Lampona captured significantly larger prey than itself and the prey captured by Drassodes. As concerns hunting strategy, Lampona grasped the prey with two pairs of legs possessing scopulae, whereas Drassodes immobilized prey with silk. Lampona possess forelegs equipped with scopulae and a thicker cuticle similar to other nonrelated araneophagous spiders. Lampona fed for a longer time and extracted more nutrients than Drassodes. We show that specialized behavioral and morphological adaptations altogether increase the hunting efficiency of specialists when compared to generalists.


| INTRODUCTION
Optimal foraging theory assumes that the choice of the food depends on the net energetic gain. Generalists spend less energy on foraging, as they readily accept most of the food they encounter. Their handling time is relatively short. Specialists spend more time and energy on searching for food, the handling time is relatively long, and the energy gain from a single food item is relatively big (Davies, Krebs, & West, 2012;Townsend, Begon, & Harper, 2003).
Trophic adaptations have been studied mostly in herbivores, parasites, and parasitoids (Futuyma & Moreno, 1988). A dichotomy in foraging between generalists and specialists has been shown mainly in herbivores (Bernays, Singer, & Rodrigues, 2004). True predators (i.e., predators that hunt more than one prey item during their lifetimes) have not been studied intensively in this regard. It appears that true predators are less frequently specialized than herbivores, parasites, and parasitoids (Thompson, 1994), as the great majority of them are generalists, catching prey smaller than themselves (Griffiths, 1980). Specialist predators seem to specialize on relatively large prey (e.g., Bulbert, Herberstein, & Cassis, 2014;Mori & Vincent, 2008;Pekár, Šedo, Líznarová, Korenko, & Zdráhal, 2014;Yamada & Boulding, 1998). Such prey is overcome by the use of specialized adaptations, especially by the ability to perform effective prey capture behaviors or by the possession of the right 'tools' to capture prey (Ferry-Graham, Bolnick, & Wainwright, 2002).
Such behavioral adaptations include a variety of hunting tactics which are effective in hunting successfully for different prey types (e.g., Harland & Jackson, 2006). Specialists usually possess hunting strategies dedicated to subduing their focal prey. For example, webinvading araneophages deceive their victims by imitating ensnared prey (e.g., Jackson, 1990) or the male's courtship signals (Jackson & Wilcox, 1990). Myrmecophagous spiders preying on ants use one of three different approaches to capture ants depending on their size and dangerousness (Pekár & Toft, 2015).
Morphological traits may also affect handling effectiveness. For example, specialized crabs feeding on molluscs possess broad claws with dull protuberances, which are well adapted to exerting greater pressure and to breaking hardened shells (Yamada & Boulding, 1998). Snakes specialized on frogs are able to open their jaws wider for the easier swallowing of such bulky prey (Mori & Vincent, 2008).
Crustaceophagous spiders of the genus Dysdera have evolved variously shaped fangs to penetrate the defences of woodlice (Řezáč, Pekár, & Lubin, 2008). Araneophagous spiders of the genus Palpimanus use stout legs with scopulae to firmly grasp their prey. They also have a thick cuticle preventing counter attack by the dangerous prey they catch (Pekár, Šobotník, & Lubin, 2011).
The aim of this study was to compare the capture efficiencies and associated adaptations of a specialist and a generalist predator. We chose spiders as our model system, as both generalist and specialist species can be found among these true predators (Pekár, Coddington, & Blackledge, 2012). First, we investigated the trophic niche of two phylogenetically related species, Lampona murina L. Koch, 1873 (Lamponidae) (Figure 1a), and Drassodes lapidosus (Walckenaer, 1802) (Gnaphosidae) (Platnick, 2000) (Figure 1b), to support our expectation that they represent a specialist and a generalist, respectively. Lampona is an Australian genus reported to be araneophagous, as it was often observed invading the webs of other spiders (Platnick, 2000). Drassodes is a Palearctic genus also reported to occasionally catch spiders (Chinery, Morris, & Hughes, 1979). We expected Lampona to have higher capture and handling efficiencies than Drassodes due to its possession of specialized behavioral and morphological hunting adaptations. All spiders used in laboratory experiments were kept in plastic vials containing moisturized gypsum, which were placed in a chamber at a constant temperature (27 ± 1°C) and under a LD regime (16:8).

| Spiders
Spiders were fed at least once a week with a cricket or a spider or were allowed to consume the prey accepted in trials. Experiments were performed from April to December 2014.
All analyses were performed using R environment (R Core Team, 2014).

| Fundamental niche
To investigate the fundamental trophic niche, acceptance trials were performed. A total of 55 individuals of Lampona and 35 individuals of Drassodes were used. Spiders were starved for 1 week before using them in trials. Individuals were placed singly in Petri dishes (diameter 5 cm). The trials began after at least 1 hr of acclimation. Each of 13 prey types (Table 1) was offered to a spider in a randomized order.
If the prey was not attacked within 10 min of coming into contact with the predator, it was replaced with a different prey type. The trial ended when a spider killed and consumed a prey. If a spider did not accept any prey type, it was considered unmotivated to eat (i.e., satiated or preparing to molt), and data concerning such individuals were rejected. Trials were performed approximately in one-week intervals for each individual. Each prey type was offered to individual spiders only once.
The acceptance (i.e., the relative frequencies of acceptance) of each prey type was compared to the average prey acceptance of Drassodes and Lampona using a linear model with arcsine square root transformation of the data and sum contrasts. The standardized Levins' index (Hurlbert, 1978) was used to calculate the fundamental trophic niche breadth of both Lampona and Drassodes using the following formula: where p j is the proportion of individuals accepting the jth prey, and n is the total number of prey types. Values of B A higher than 0.6 indicate a wide niche, and values below 0.4 indicate a narrow niche (Novakowski, Hahn, & Fugi, 2008). Twelve prey types instead of 13 were used to calculate the index, as all spiders (Lycosidae and Thomisidae) were pooled and considered as a single category. Pianka's index (O jk ) (Pianka, 1973) was used to calculate niche overlap using the following formula: where p ij is the proportion of the ith prey in the diet of jth species, p ik is the proportion of the ith prey in the diet of kth species, and n is the total number of prey types. Values close to one indicate high overlapping; a zero value represents excluded niches.

| Realized niche of L. murina
The realized trophic niche was investigated only in the specialist (Lampona), as examining the realized niche in the generalist (Drassodes).
We assumed that the breadth of the realized niche of the generalist would be wide and only wanted to confirm the narrow niche in the specialist. The niche breadth was investigated using gut contents and analyzed using next-generation sequencing (NGS). General invertebrate primers (Zeale, Butlin, Barker, Lees, & Jones, 2011) were used for the amplification of prey DNA from the guts of Lampona spiders. As these primers allow the amplification of a wide range of taxa, including spiders, and the predators' DNA prevails over prey DNA in the predators' opisthosoma, it was necessary to design blocking oligos (Vestheim & Jarman, 2008) which would prevent the amplification of Lampona spiders' DNA but allow the amplification of their potential prey, including closely related spiders. DNA was extracted from the tibia of two  (Folmer, Black, Hoeh, Lutz, & Vrijenkoek, 1994)   Genetic Analyzer (Applied Biosystems). Sequences were assembled in Sequencher 4.8 (Gene Codes Corporation, Ann Arbor, MI) and aligned using ClustalW (Thompson, Higgins, & Gibson, 1994) implemented in MEGA 5.1 (Tamura et al., 2011). A reverse blocking primer was designed using Amplicon.b08 (Jarman, 2004): A unique section of 23 bases, typical for Lampona only, which was situated inside the fragment amplified using general invertebrate primers and closer to its 3′ end, was selected and modified with C3 spacer. The blocking primer (BlkLamp, 5′-CGTCACCTAATAATCTACCGGAC-3′) was tested with all potential prey samples and with Lampona spiders both to ensure that none of the potential prey would be blocked and to unify optimal PCR conditions.
In two adult males, ten adult females, and 35 juveniles of L. murina, prey DNA was extracted from their opisthosoma using the same approach as above, with a change in the final step when only 50 μl Each sample was PCR-amplified using a unique combination of primers tagged with MID identifiers (10 base-long barcoding sequences; we used seven different MIDs on the forward primer and another seven MIDs on the reverse primer, which allowed us to assign all DNA reads to the predator individuals). PCR products were detected by electrophoresis in 2% GoodView-stained agarose gels. When no PCR product was detected, PCR was repeated until DNA was available. PCR products were purified using QIAquick PCR Purification Kit (Qiagen) according to the manufacturer's protocols. The concentration of each PCR product was measured using a NanoDrop 8000 UV-Vis Spectrophotometer (Thermo Scientific). 5 μl of each 50 μg/μl PCR product was pooled into the same sterile vial and sent for sequencing.

Enrichment (emPCR) and sequencing on an Ion Proton System with an
Ion 316 chip with 400-base read-length chemistry were provided by the SEQme company (Dobříš, Czech Republic).
The sequencing output was processed using the Galaxy platform (https://usegalaxy.org/), BioEdit 7.2.5 (Hall, 1999), MEGA 5.10 (Tamura et al., 2011), fastx-toolkit, and the EMBOSS package (Rice, Longden, & Bleasby, 2000). Reads were split according to their MIDs, resulting in files corresponding to predator individuals. Then, forward and reverse primers were removed, and reads were filtered according to their length; reads shorter than 120 bases were removed. The reads were collapsed, and rare haplotypes (containing <2 identical reads) were removed. Sequences which contained stop codons were removed; the remaining were aligned using MAFFT (Katoh & Standley, 2013), and the sequences with indels causing frameshifts were also removed.

| Capture efficiency and capture behavior
To compare the hunting efficiencies of Lampona and Drassodes for differently sized prey, spiders (Pardosa, Alopecosa, Lycosidae, Araneae) of various sizes were offered to both species in a similar manner as in the acceptance trials. Individuals of Lampona and Drassodes were placed singly in Petri dishes, and after acclimation, the prey was offered. If the prey was not accepted within 10 min of the predator and prey coming into contact, it was replaced by a smaller one. The fangs of large

| Morphology of trophic traits
Ten individuals of Lampona and Drassodes used in laboratory trials were thereafter stored in 75% alcohol and used to measure morphological traits: the length of the prosoma; the length of both segments of the chelicerae; the length of the posterior and anterior spinnerets; the length of the tarsus and metatarsus of the first pair of legs; the area and density of scopula hairs; and the thickness of the cuticle on the prosoma, the opisthosoma, and the legs (femur).
Thickness of the cuticle was measured from thin tissue cuts prepared with scalpel. The cuts were made perpendicularly to the surface in the central section of the prosoma, opisthosoma, and femur. LEICA EZ5 and OLYMPUS CX31 stereomicroscopes were used for measurement, both equipped with an ocular micrometer. The obtained data were modeled by means of a GLM model with gamma distribution and a logarithmic link (Pekár & Brabec, 2016).

| Prey consumption
The consumption time and net weight gain were observed in eight individuals of Lampona and ten individuals of Drassodes. Lycosid spiders were used as prey. After capture of the prey, the individuals were checked every 15 min to determine whether they were feeding on the prosoma or opisthosoma of the prey. Feeding on the leg was considered as feeding on the prosoma. All predator and prey individuals were weighed using a Sartorius balance with a precision of 0.001 mg before and after feeding. Prey consumption time was modeled by means of a GLM model with gamma distribution and a logarithmic link (Pekár & Brabec, 2016). The type of predator and consumed body part were used as factors, and the initial weights of both the predator and the prey were used as covariates. The proportion of individuals feeding on the prosoma of the prey throughout the whole time of consumption was modeled using a logit model with binomial distribution and GEE due to the occurrence of paired measurements. The working correlation structure was exchangeable. Net weight gain was analyzed in a similar manner as prey consumption time. The type of predator was used as a factor, and the weights of both the predator and the prey were used as covariates.

| Trophic niche
Lampona accepted eight of thirteen offered prey types in the laboratory. The average prey acceptance of all thirteen prey types was 27%.
It did not accept woodlice, ants, beetles, true bugs, or caterpillars. At a significantly lower frequency, it accepted cockroaches and fruit flies (contrasts, p < .01). Termites were accepted at a slightly but significantly higher frequency than average (contrast, p = .02). Three prey types were accepted at a significantly higher frequency: crickets, wolf spiders, and crab spiders (contrasts, p < .0001) (Figure 2).
Drassodes accepted ten of thirteen offered prey types. The average prey acceptance of all thirteen prey types was 66%. It did not accept woodlice, ants, or beetles. Eight prey types were accepted at a significantly higher frequency: springtails, cockroaches, termites, crickets, leafhoppers, fruit flies, wolf spiders, and crab spiders (contrasts, p < .001). Only true bugs were accepted at a significantly lower frequency (contrast, p = .001).

| 2761
MICHÁLEK Et aL.  Table S1). The number of spiders was significantly higher compared to other prey pooled (χ 2 1 = 9.6, p = .002). A high number of the spider prey were conspecifics or congenerics (Lampona murina or Lampona sp.). The realized niche breadth of Lampona was even narrower (B A = 0.16) compared to the fundamental niche breadth.

| Capture efficiency and capture behavior
Capture success with respect to spider prey decreased with the relative prey/predator size ratio both in Lampona and Drassodes (GEE, χ 2 1 = 32.5, p < .0001), but it differed between them (GEE, χ 2 1 = 4.9, p = .03). Lampona was more successful in handling larger spider prey than Drassodes. Lampona achieved 50% capture success with prey approximately 1.5 times larger than itself, while Drassodes achieved the same success with smaller prey, 1.25 times larger than itself (Figure 4).
The key element in the predatory behavior of Lampona was grabbing the prey with the scopula on its tarsi and metatarsi. The prey was first grabbed by the predator's forelegs, then bitten, and held firmly until it was paralyzed ( Figure 5, Video S1). Drassodes used a different tactic: It wrapped the prey in silk to immobilize it ( Figure 6, Video S2). The probabilities of transitions during hunting sequence are illustrated in flow diagrams (Figure 7). Lampona was more precise in selecting the location of the bite. The prey was bitten more often on the legs (79%, N = 24) than on the body (prosoma or opisthosoma) (χ 2 1 = 8.7, p = .01). Drassodes bit the prey more often on the body than on the legs (64%, N = 14); however, this difference was not significant (χ 2 1 = 1.2, p = .29). The Shannon entropy of behavioral sequences differed significantly between Lampona and Drassodes: The entropy estimate for Lampona sequences was 3.93 (95% confidence F I G U R E 2 Comparison of the relative frequencies (percentage) of prey accepted by Drassodes and Lampona in the laboratory. Full horizontal line shows the overall mean of prey acceptance for Lampona, dashed line for Drassodes F I G U R E 3 Natural prey in the gut of 36 Lampona individuals obtained from DNA analysis interval 3.24 -5.29), while for Drassodes, it was 7.08 (95% confidence interval 6.60 -8.69); therefore, the behavior of Lampona was more stereotypical.

| Prey consumption
Both Lampona and Drassodes fed longer on the prosoma of the prey than on the opisthosoma (GLM, F 1,34 = 3.9, p = .06), but the time spent feeding on each part of the prey differed between the spiders: It was significantly shorter for Drassodes (GLM, F 1,33 = 5.8, p = .02).

| DISCUSSION
As expected, Lampona is an araneophagous specialist as it accepted spiders more frequently than other prey, similar to other araneophagous species studied so far (Cerveira & Jackson, 2005;Guseinov, Cerveira, & Jackson, 2004;Jackson & Blest, 1982;Kloock, 2001;Li, Jackson, & Barrion, 1999;Pekár et al., 2011). Lampona accepted a few other insect species, although at a lower frequency. By contrast, Drassodes did not show a preference for any prey; thus, our results support our expectation it is a euryphagous generalist.
In nature, Lampona hunted mainly spiders, as confirmed by the DNA analysis of gut contents. Surprisingly, a large proportion of natural prey was composed of spiders of the same genus or even the same species. Yet cannibalism is not uncommon among araneophagous spiders in the form of sexual or juvenile cannibalism, or oophagy (Clark & Jackson, 1994;Clark, Jackson, & Waas, 1999;Jackson & Pollard, 1997). We confirmed that some of the sequences which were assigned to Lampona belonged indeed to the prey and not to the studied pred- proportion of sequences even to order level, as they were assigned to two or more different insect orders with an identical percentage of sequence identity. This may have been because of rapid prey DNA degradation in predators' guts. Perhaps the prey items had been eaten accidentally much earlier than the nondegraded prey or had been eaten by predators which were subsequently consumed by Lampona.
Studies on detection of prey DNA in the guts of invertebrate predators showed secondary predation represents a potential significant source of error in the interpretation of data (Hosseini, Schmidt, & Keller, 2008;Sheppard et al., 2005). Therefore, we assume the presence of this DNA is most probably the result of the secondary predation.
Specialized Lampona was often able to overcome larger prey than the generalist Drassodes, which is in agreement with other studies comparing specialist and generalist predators (Mori & Vincent, 2008;Yamada & Boulding, 1998). Specialists may hunt larger prey in order to decrease the number of risky foraging events when hunting dangerous prey such as spiders. The capture of bigger prey was also observed in other araneophagous spiders (e.g., Eberhard, 1983;Penney & Gabriel, 2009). Here, we provide evidence that a specialist predator is more successful in hunting larger dangerous prey than a generalist.
The ability to hunt larger prey in specialists is facilitated by specialized adaptations. Indeed, Lampona possesses a number of morphological adaptations for handling spider prey. It first uses two pairs of legs equipped with scopulae on the tarsus and metatarsus to capture prey.
These adhesive patches are an effective hunting tool and represent an alternative strategy to the use of costly silk common in other cursorial spiders (Wolff, Nentwig, & Gorb, 2013 Specialists should be more successful than generalists, as they are better adapted to subduing specific prey or foraging in a specific environment and, therefore, they are more precise. This was confirmed in a comparison of specialist and generalist garter snakes foraging in water (Drummond, 1983). Moreover, in the case of dangerous prey, any mistakes could have a significant impact on predator survival (Mukherjee & Heithaus, 2013). As a result, specialization may lead to greater precision in prey capture (Ferry-Graham et al., 2002). Drassodes usually spent more time subduing the prey, as the hunting sequences were longer due to the occasional repetition of behaviors. The strategy of Lampona, as a specialist, was more stereotypical and precise. Precision also plays an important role in the strategy of other araneophagous spiders, like those of the genus Portia, who choose the location of the strike depending on the dangerousness of the particular prey (Harland & Jackson, 2006).
A specialist should also maximize the utilization of the prey (Townsend et al., 2003). Indeed, Lampona spent more time feeding and gained more nutrients in comparison with Drassodes. Feeding on one prey type during the life cycle may lead to challenges with respect to limitations on the sources of nutrients. This issue may be resolved by selective feeding on body parts of a prey with different content of proteins and lipids (Pekár, Mayntz, Ribeiro, & Herberstein, 2010).
However, Lampona did not select only one body part; it fed both on the prosoma and opisthosoma without distinction. Alternatively, nutrient variability may be maintained by the occasional consumption of alternative prey, as araneophagous spiders sporadically hunt insects. Proteins contained mainly in the prosoma seem to be a more important component in the diet of spiders (Blamires, Hochuli, & Thompson, 2009;Mayntz & Toft, 2001). Lampona usually began feeding on the prosoma of spider prey and moved to the ophistosoma later. Drassodes consumed the prey in a similar way. Other spider generalists consuming insect prey fed in a similar manner (Haynes & Sisojevi, 1966;Kim, Krafft, & Choe, 2005;Pollard, 1989). Alternatively, choice of the prosoma may be connected with mechanics of feeding. The hard prosoma is more suitable for suction than the softer opisthosoma that may collapse under suction pressure (Pollard, 1989(Pollard, , 1990. Spiders may represent a high-quality food source for other spiders in general (Wise, 1995). Therefore, spiders probably represent an optimal prey for all spiders. A study on the araneophagous Portia quei (Toft, Li, & Mayntz, 2010) revealed it to be a behavioral specialist but a metabolical generalist, as it shows versatility in the utilization of prey with varied proportions of nutrients, which is a trait more inherent in euryphagous generalists. Behavioral adaptations for hunting spiders may have evolved without metabolical restrictions and without sacrificing the possibility of utilizing alternative prey.
In conclusion, predators specialized on dangerous prey possess adaptations allowing them to better handle their focal prey. Generalist predators, that is, those which occasionally hunt similar prey, may also possess alternative generalized adaptations for handling dangerous prey. However, the effectiveness of such generalized adaptations will be lower than the effectiveness of more specialized ones.

ACKNOWLEDGMENTS
This study was supported by a grant from the Czech Science Foundation (GA15-14762S). We would like to thank Maria Guadalupe Corcobado Márquez for her assistance in collecting Lampona spiders in the field, and Luis Fernando García Hernández both for his assistance in collecting prey and for comments concerning the experimental design and data analysis. We would also like to thank two anonymous reviewers for their constructive comments on this manuscript.

CONFLICT OF INTEREST
None declared.

DATA ACCESSIBILITY
Sequences used for designing primers are archived in the GenBank sequence database. Prey sequences obtained from L. murina are available in Online Supporting Information (Table S2).