Responses of juvenile terrestrial salamanders to introduced (Lithobius forficatus) and native centipedes (Scolopocryptops sexspinosus) by Cari-Ann Hickerson | Papers by Cari-Ann

Journal of Zoology. Print ISSN 0952-8369 Responses of juvenile terrestrial salamanders to introduced (Lithobius forficatus) and native centipedes (Scolopocryptops sexspinosus) C. D. Anthony1, C. A. M. Hickerson2 & M. D. Venesky1 1 Department of Biology, John Carroll University, University Heights, OH, USA 2 Department of Biological, Geological, and Environmental Science, Cleveland State University, Cleveland, OH, USA Keywords Plethodon cinereus; Scolopocryptops sexspinosus; Lithobius forficatus; intraguild predation; introduced species; behaviour; competition. Correspondence Carl D. Anthony, Department of Biology, John Carroll University, University Heights, OH 44118, USA. Fax: 216 397 4482 Email: canthony@jcu.edu Received 19 October 2005; accepted 8 May 2006 doi:10.1111/j.1469-7998.2006.00202.x Abstract When introduced species invade ecosystems, alterations in community structure can emerge from the competitive and predatory interactions that occur between introduced and native guild members. Because a number of recent studies have shown that large predatory invertebrates can both compete with and prey on small vertebrates and because introductions of non-native species may play a role in amphibian declines, the effects of introduced centipedes Lithobius forficatus and native centipedes Scolopocryptops sexspinosus on juveniles of the red-backed salamander Plethodon cinereus were examined. In laboratory arenas, juvenile salamanders exhibited submissive behaviour in response to the odours of both species of centipede. There were no significant differences in salamander response to the two centipede odour treatments, but compared to controls, juveniles of P. cinereus spent significantly more time in escape and in a flattened submissive posture when presented with native centipede odours. Despite significant size differences between centipedes and juvenile salamanders, no predation of salamanders by either species of centipede occurred in any pairings. Juveniles exhibited more chemosensory behaviour towards native centipedes and towards their odours and exhibited marked reductions in aggressive posturing when centipedes were present. Field and laboratory data suggest that juveniles of P. cinereus and centipedes were negatively associated. In laboratory trials, the native centipede excluded juvenile salamanders from cover objects and we found fewer instances of co-occurrence in the field than expected. These studies are the first to examine the behavioural interactions between juveniles of P. cinereus and invertebrate predators, one introduced and one native, of eastern deciduous forest-floor food webs. Introduction Plethodontid salamanders are important components of the forest-floor food webs of eastern temperate forests of North America. Of particular importance are small-bodied and numerically abundant species, such as the red-backed salamander Plethodon cinereus, because their small body mass (usually less than 1 g) allows them to exploit prey species that are not suitable prey for other vertebrates (Pough, 1983). This ability, combined with exceptionally large population size and biomass (Burton & Likens, 1975), makes them important regulators of below-ground food webs. Indeed, a growing body of evidence suggests that the presence or absence of salamanders in forest-floor food webs can have far-reaching effects on invertebrate abundance and leaf-litter decomposition (Wyman, 1998; Walton & Steckler, 2005). Given the important roles that amphibians play in ecosystem function, ecologists have expressed concern over the global declines of these species (Alford & Richards, 1999; 54 Houlahan, Findlay & Schmidt, 2000; Lannoo, 2005). A number of potential causes for the global amphibian decline have been identified, including climate change, emerging disease, habitat loss, ultraviolet radiation, environmental toxins and introduced species (see the citations in Green, 2003). Recent evidence (Highton, 2005) indicates that salamanders of the genus Plethodon are experiencing declines throughout the eastern United States, but it is unclear which, if any, of the above causes play a role in the decline of populations of this genus. Although recent studies have examined the effects of introduced predators on aquatic amphibian populations (Kats & Ferrer, 2003), few researchers have examined the effects of introduced species on terrestrial salamander populations (Ducey et al., 1999; Maerz et al., 2005). Despite their perceived importance in detrital food webs and the potential negative effects that introduced predatory invertebrates have on terrestrial salamanders, we know little of the specific behavioural interactions that occur among c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London C. D. Anthony, C. A. M. Hickerson and M. D. Venesky Interactions between salamanders and centipedes these species. Forest-dwelling centipedes and red-backed salamanders have similar food and habitat requirements and likely compete for resources such as prey and space on the forest floor (Hickerson, Anthony & Wicknick, 2004). Cover objects, such as rocks and logs, provide isolated patches of moisture and associated prey during dry spells between periods of rainfall (Jaeger, 1981). Thus, when the forest floor dries, salamanders may enter into competition with other guild members. Larger centipedes, such as the two species examined in this study, may act as intraguild predators on one another. In laboratory microcosms, Scolopocryptops preyed on Lithobius (Hickerson, Anthony & Walton, 2005) and large individuals of either species have the potential to act as intraguild predators on juvenile red-backed salamanders. Lithobius forficatus, a non-native centipede introduced to the eastern United States from Europe in the 1800s, is associated with human disturbance (Lee, 1980) but has invaded some forested areas (Frund, Balkenhol & Ruszkowski, 1997; Hickerson et al., 2005). This species reaches 26 mm total length (Williams & Hefner, 1928), approximately twice the size of a neonate red-backed salamander. Scolopocryptops sexspinosus is a native resident of forest-floor habitats in the eastern United States and reaches 69 mm total length (Shelley, 2002). Adult red-backed salamanders in the region of this study average 40.5 mm snout–vent length (SVL; Pfingsten & Downs, 1989). Both Lithobius and Scolopocryptops are venomous and are capable of delivering painful bites (Williams & Hefner, 1928). Because red-backed salamanders are territorial (Mathis et al., 1995), simply fleeing from dangerous competitors may not be an adaptive solution to an encounter. Instead, territorial owners are expected to use behavioural displays in an attempt to expel invertebrate intruders. Previous studies, using adult red-backed salamanders, support this hypothesis. In laboratory trials, male redbacked salamanders exhibit aggressive behaviour towards carabid beetles (Gall, Anthony & Wicknick, 2003) and centipedes (Hickerson et al., 2004). We are aware of only two studies that have addressed how juvenile salamanders interact with predatory invertebrates (Ovaska & Smith, 1988; Rubbo et al., 2003) and it is unknown how juveniles of red-backed salamanders P. cinereus interact with native and introduced predatory invertebrates. In this study we addressed the following questions: (1) Do adult centipedes exclude juveniles of P. cinereus from cover objects in the field or in the lab? (2) Do juveniles of P. cinereus respond differently to the odours of, or to laboratory encounters with, either species of centipede? (3) Do adult centipedes of either species prey on juveniles of P. cinereus? 72 ACOs for the presence of salamanders and centipedes. The array was constructed as part of a larger study examining the interactions of salamanders and large invertebrates in forest-floor food webs. The entire array consisted of 288 ACOs and included removal treatments that were not sampled for this study. The ACO array was assembled in early April 2004 and sampling began 2 weeks after ACOs were laid down. We used large (30.5  30.5 cm) ceramic tiles as ACOs and visited the site approximately every 2 weeks through early December 2004. We turned each cover object and recorded the numbers of adults (432 mm SVL) and juveniles (o22 mm SVL) of P. cinereus, the numbers of adults of S. sexspinosus (435 mm) and the numbers of lithobiomorph centipedes over 15 mm total length. We determined whether cover objects yielded salamanders and/ or centipedes at any time during the study and assessed the degree of co-occurrence with w2 tests of independence. For the analysis, we assigned each of the 72 cover objects to one of six categories based on whether salamanders, centipedes, or salamanders and centipedes were found under an ACO during any visit. Thus, if an ACO produced a centipede during visit 1 and a juvenile salamander during visit 12, it was designated as a shared cover object. This is a conservative approach that overestimates the degree of co-occurrence between salamanders and centipedes. In generating the expected distribution, we assumed that the occurrence of each category type was equally probable. Thus the expected distribution was simply the number of ACOs examined divided by the number of categories used in each analysis. General methods for laboratory trials Specimens were collected from three adjacent counties in north-eastern Ohio, USA. We collected juveniles of P. cinereus in October 2004 from mature beech maple forest in northern Summit County. Adults of L. forficatus were collected from a residential area in Cuyahoga County and adults of S. sexspinosus were collected from a mature beech maple forest in Lake County. We housed salamanders and centipedes individually in plastic chambers (17  11  4.5 cm for salamanders; 24  16.5  6 cm for centipedes) on leaf litter under a natural photoperiod at 16.7 Æ 1.1 1C. All specimens were fed Drosophila hydei ad libitum. Centipedes were denied food for 4 days before testing. Salamanders and centipedes were weighed periodically throughout the study and no individuals lost mass during the study. Odour discrimination In December 2004, we examined the responses of juveniles of P. cinereus to the odours of adult centipedes. We exposed juveniles of P. cinereus (mean SVL Æ SE =20.1 Æ 0.55 mm; mean mass Æ SE = 0.143 Æ 0.011 g, n =15) to three substrate odours in separate trials. We used plastic Petri dishes (1.5 cm  14 cm diameter) lined with 15 cm diameter Ahlstroms (Ahlstrom Corporation, Mt. Holly Springs, Pennsylvania, USA) qualitative filter paper for our experimental arenas. Salamanders were carefully placed on 55 Methods Co-occurrence under artificial cover objects (ACOs) in the field During the months of April–December 2004, we examined the degree of co-occurrence of juvenile salamanders and centipedes under ACOs in the field. We sampled an array of c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London Interactions between salamanders and centipedes C. D. Anthony, C. A. M. Hickerson and M. D. Venesky damp circular filter paper on which an adult Scolopocryptops, an adult Lithobius or no centipede had previously resided for 5 days. We allowed salamanders to interact freely with the substrate and any associated odours for 15 min. In experimental treatments (centipede odours), centipedes were removed from substrates less than 60 s before exposure of the salamander to the substrate. We presented salamanders with odour treatments in a randomized block design, where each salamander was exposed to each odour on a different day and equal numbers of salamanders were tested on each odour each day. No salamander was tested more than once in each 5-day period. To avoid temporal bias in the data, treatments and controls were evenly dispersed across test dates and observers (Hurlbert, 1984). During data collection, we made every effort to disguise the treatment type from data recorders. We used the software package EVENT-PC (James C. Ha, University of Washington) to record the frequency and duration of the following behaviours of juveniles of P. cinereus when exposed to the three odour treatments. These behaviours were modified from Jaeger (1984) and Hickerson et al. (2004): flattened (FLAT) – considered a submissive posture, the entire ventral surface of the body and the chin is in contact with the substrate; front trunk raised (FTR) – considered a resting posture; all trunk raised (ATR) – considered an aggressive posture, the legs are extended such that the head, trunk and sometimes the tail are lifted off the substrate; nose tapping the substrate (NTS) – considered an investigative behaviour, contact of the nasolabial cirri to the substrate; nose rubbing the substrate (NTR) – considered an investigative behaviour, the snout is held to and sometimes rubbed on the substrate for several seconds at a time; immobility (IMMOBILE) – considered an antipredator behaviour; escape behaviour (ESCAPE) – defined as circling the periphery of the chamber while pressing the snout or body against the outer edge of the Petri dish; sustained escape (SUST ESCAPE) – defined as the longest interval of escape per trial. Escape behaviour can be considered as either submissive (Wise & Jaeger, 1998) or antipredator behaviour. Comparisons among treatments were made using two-tailed paired t-tests (where the data met the assumptions of parametric statistics) or Wilcoxon signed-ranks tests, a non-parametric equivalent. We reduced a to 0.025 because each data set was used twice in each analysis. We used one-tailed tests when analysing time spent in escape because a previous study (Hickerson et al., 2004) indicated that adults of P. cinereus exhibited increases in escape when exposed to the odours of centipedes. Behavioural interactions between salamanders and centipedes In January and February 2005, we explored the potential for aggression and intraguild predation (IGP) between juvenile salamanders and adult centipedes in laboratory arenas. We paired juveniles of P. cinereus [mean SVL= 19.3 Æ 0.31 (SE) mm; mean mass = 0.116 Æ 0.003 (SE) g, n = 30] with adults of L. forficatus [mean total length (TL) = 27.9 Æ 0.45 (SE) mm; mean mass = 0.146 Æ 0.007 (SE) g, n = 30] and with 56 adults of S. sexspinosus [mean TL = 43.4 Æ 0.56 (SE) mm; mean mass = 0.282 Æ 0.010 (SE) g, n =30] in separate trials. We minimized mass differences between paired animals by sorting animals by mass and then randomly pairing within each of five mass classes. In pairings, salamander SVL was always shorter than centipede TL [(mean Æ SE) difference = 8.6 Æ 0.54 mm for Lithobius/Plethodon pairs and 23.9 Æ 0.61 mm for Scolopocryptops/Plethodon pairs]. Salamanders were always lighter in mass than Scolopocryptops [(mean Æ SE) difference = 0.16 Æ 0.01 g] and lighter than Lithobius in 25 of 30 trials [(mean Æ SE) difference = 0.04 Æ 0.007 g]. We paired each salamander with a centipede of each species in random order. Eight to 10 days passed between pairings and equal numbers of salamanders were paired with Lithobius first and Scolopocryptops second. Centipedes were not used more than once in this experiment, but some of the salamanders used in the odour experiment were reused. We were careful not to pair salamanders with individual centipedes that they had experienced odours from in the previous experiment. To avoid temporal bias in the data, equal numbers of Lithobius and Scolopocryptops trials were run on each test day (Hurlbert, 1984). Observers collected data from approximately equal numbers of each trial type, but it was not possible to conduct these trials in a blind design because the species identity of the centipede was conspicuously apparent. Salamanders and centipedes were tested in circular arenas (as described in the first experiment) on damp (unmarked) filter paper. We carefully placed each salamander and centipede (Scolopocryptops or Lithobius) into the arena and covered each animal with an opaque habituation dish (5.5 cm diameter). After a 5 min acclimation period the dishes were lifted and the trial was started immediately after it was apparent that the salamander was aware of the presence of the centipede (indicated by the salamander turning its head towards the centipede, or the centipede moving across the salamander’s forward field of vision). We used the software package EVENT-PC to record frequency and duration of behaviours of salamanders and centipedes during each 15 min trial. For salamanders, we recorded the same behaviours as listed above in experiment 1 (FLAT, FTR, ATR, NTS, NTR, IMMOBILE, ESCAPE, SUST ESCAPE) and the following additional behaviours: nose tapping the centipede (NTC) – contact of the nasolabial cirri to the centipede (Hickerson et al., 2004); move towards (MT) – salamander approaches the centipede in a direct path that would result in contact if the movement were to continue; flipping (FLIP) – rapid twisting or writhing that carries the salamander away from the centipede in a salutatory manner (Brodie, 1977). Comparisons between the two treatments were made using two-tailed paired t-tests (where the data met the assumptions of parametric statistics) or Wilcoxon signed-ranks tests, a non-parametric equivalent. Competition for ACOs in the laboratory In March 2005, we examined the frequency in which juveniles of P. cinereus co-occurred with adults of either c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London C. D. Anthony, C. A. M. Hickerson and M. D. Venesky Interactions between salamanders and centipedes centipede species under cover objects in experimental arenas. We used small (7.5  5.25  0.9 cm) ceramic tiles as cover objects. A 1 cm length of 0.6 cm diameter surgical tubing was used as a shim to raise one short end of the tile. The tile and shim were placed on a single sheet of 15 cm diameter filter paper in a 22 cm square experimental arena. We dampened the filter paper with 5 mL of spring water before introducing the animals. This provided three microhabitats within the arena: under cover and on damp filter paper (6.3% of the chamber); on damp filter paper only (22% of the chamber); and on the dry surface of the chamber (71.7% of the chamber). As in the previous experiment, salamanders and centipedes were randomly paired within size classes to minimize size differences. In pairings, salamander SVL was always shorter than centipede TL [(mean Æ SE) difference = 8.3 Æ 0.32 mm for Lithobius/ Plethodon pairs and 23.8 Æ 0.58 mm for Scolopocryptops/ Plethodon pairs. Salamanders were always lighter in mass than centipedes [(mean Æ SE) difference = 0.16 Æ 0.009 g for Scolopocryptops/Plethodon pairs and 0.027 Æ 0.003 g for Lithobius/Plethodon pairs]. Each salamander (n =28) was randomly placed in each of three treatments: a control treatment where no centipede was present and two experimental treatments where a Lithobius or a Scolopocryptops was present. Salamanders were not tested more than once in any 4-day period and, although the same centipedes and salamanders from the previous experiment were used, salamanders were not re-paired with individual centipedes that they had interacted with in previous experiments. We introduced a juvenile P. cinereus and either an adult Lithobius or an adult Scolopocryptops simultaneously into the arenas between 16:00 and 17:00 h. Salamanders and centipedes were allowed to interact until 10:30 h the following day when we recorded the location of each animal within the chamber as well as the number and location of salamander faecal pellets. We predicted that, in the control treatment, salamanders would be free to use the cover object, but Date Date that in the presence of centipedes they would be excluded from cover. For each treatment, the position of the salamander at the end of the trial was scored as either under cover (+) or not under cover (À) and a sign test was used to determine if salamander position was influenced by centipede presence. We used a G-test of independence to determine if faecal pellet location was influenced by centipede presence. Here we reasoned that salamanders would be less likely to deposit pellets under cover if they were excluded by centipedes from using cover objects. We used one-tailed tests where we had evidence from field data (this study) and from a laboratory pilot study on ACO use that indicated that centipedes and juvenile salamanders were unlikely to share cover objects. Results Co-occurrence under ACOs in the field We visited the field site 15 times from April to December 2004. Sixty-eight of the 72 ACOs produced salamanders, centipedes or both sometime during the study. Thus the ACOs provided suitable cover. Individuals of P. cinereus were active under cover objects during 13 of 15 visits and centipedes were found during 14 of 15 visits (Fig. 1). A juvenile salamander co-occurred at the same time with a centipede (a lithobiid) in only one instance during the 8-month sampling period. We placed each of the 72 ACOs into one of six categories based on whether salamanders and/or centipedes were present or absent during the sampling period. In this way, potential bias from multiple sampling of the same individuals over time could not affect the sample size or which category an ACO was assigned to. Categories were defined as follows: (1) ACOs that yielded neither centipedes nor juvenile salamanders but did yield adult salamanders (n = 26 ACOs); (2) ACOs that yielded juvenile salamanders but no centipedes (n = 8 ACOs); 1 1 5 10 15 37 1 5 10 15 Co-occurrence Juvenile Plethodon + lithobiid Juvenile Plethodon + Scolopocryptops 12 48 No co-occurrence Scololocryptops, no juvenile Plethodon Lithobiid, no juvenile Plethodon Adult Plethodon, no juv, no centipede Juvenile Plethodon, no centipede Empty ACO 24 60 A C O A C O 36 72 Figure 1 Occurrence of salamanders and centipedes under the 72 artificial cover objects (ACOs) used in the field study. The dates (1–15) correspond to approximately bi-monthly observations made from April to December 2004. Salamanders and centipedes co-occurred if they utilized the same ACO anytime during the study. For example, ACO #10 was considered a shared cover object because it was utilized by a juvenile salamander on observation date 10 and by a lithobiid centipede on date 12. One case of simultaneous co-occurrence of a juvenile salamander and a lithobiid is indicated by the filled square. There were no cases of juvenile salamanders simultaneously sharing a cover object with a Scolopocryptops. c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London 57 Interactions between salamanders and centipedes C. D. Anthony, C. A. M. Hickerson and M. D. Venesky Mean time (s) spent in the submissive posture - FLAT Table 1 Behaviour of juveniles of Plethodon cinereus in odour treatments Behaviour FLAT FTR ATR IMMOBILE ESCAPE SUST ESCAPE NTS NTR Control, n = 15 3.1 (2.4) 513.1 (63.2) 177.3 (38.7) 742.1 (39.1) 32.6 (12.6) 10.2 (3.6) 18.6 (4.5) 34.3 (11.1) Lithobius, n = 15 10.0 (6.5) 429.6 (80.0) 233.6 (50.4) 782.1 (25.6) 78.0 (34.2) 23.7 (9.5) 15.6 (5.4) 13.3 (4.1) Scolopocryptops, n = 15 25.8 (13.3) 400.6 (72.5) 202.5 (50.6) 725.9 (29.4) 105.2 (39.4) 37.3 (13.1) 18.4 (4.1) 36.0 (11.3) 45 40 35 30 25 20 15 10 5 0 (a) b ab a Control Lithobius Scolopocryptops Mean time (s) spent in sustained escape behaviour 60 (b) 50 40 30 20 10 0 b Juvenile salamanders were exposed to control (no odour) and experimental (centipede odour) substrates. All behaviours were timed (s) with the exception of nose tap substrate (NTS), which was recorded as a frequency. Values are means (SE). See text for descriptions of the behaviours. FLAT, flattened; FTR, front trunk raised; ATR, all trunk raised; IMMOBILE, immobility; ESCAPE, escape behaviour; SUST ESCAPE, sustained escape; NTR, nose rubbing the substrate. ab a Control Lithobius Scolopocryptops a (3) ACOs that yielded S. sexspinosus and juvenile salamanders (n = 1 ACO); (4) ACOs that yielded lithobiids and juvenile salamanders (n = 6 ACOs); (5) ACOs that yielded Scolopocryptops and no juvenile salamanders (n = 10 ACOs); and (6) ACOs that yielded lithobiids and no juvenile salamanders (n = 23 ACOs). Three ACOs that yielded neither centipedes nor salamanders of any species or size class were excluded from the analysis. We found that more cover objects housed adult salamanders, but not juveniles or 2 centipedes, than expected (category one: w2 test; w5 ¼ 40:0, Po0.001). When we partitioned these cover objects out from the analysis, we found significantly more cases of non-native centipedes occurring alone (category six: w2 test; 2 w4 ¼ 28:04, Po0.001). The native centipede never cooccurred with juvenile salamanders, but this result was not 2 statistically significant (category three: w2 test; w3 ¼ 7:16, P= 0.072). We overestimated the degree of co-occurrence between juvenile salamanders and centipedes by combining independent observations of individuals throughout the duration of the experiment. We interpret the lack of co-occurrence of juvenile salamanders and centipedes under these conservative restrictions as evidence that these species are negatively associated on the forest floor. Mean time (s) spent in escape behaviour 160 140 120 100 80 60 40 20 0 (c) a a Control Lithobius Scolopocryptops Figure 2 Behaviour of juvenile salamanders when exposed to controls (n =15), odours of introduced centipedes (Lithobius, n = 15) and odours of native centipedes (Scolopocryptops, n = 15). Mean time spent in (a) the submissive posture FLAT (flattened), (b) sustained escape and (c) escape behaviour by salamanders exposed to the three odour treatments. Different letters above bars indicate statistically significant differences. t14 = 2.01, P= 0.032, one tailed; Fig. 2c) when on Scolopocryptops substrates, compared with controls, as well. No significant differences in salamander behaviour were detected between centipede odour treatments (Table 1, Fig. 2). Odour discrimination experiment We found little evidence that juveniles of P. cinereus detect substrate odours of the introduced centipede L. forficatus. In no instances did salamanders behave differently on Lithobius substrates compared with controls (Table 1, Fig. 2). In contrast, when on Scolopocryptops substrates juveniles of P. cinereus spent significantly more time in sustained escape behaviour (Wilcoxon signed-ranks test; t= 2.11, n = 15, P= 0.011, one tailed; Fig. 2b) and in FLAT (Wilcoxon signed-ranks test; t= 1.97, n= 15, P= 0.010, two tailed; Fig. 2a) compared with control substrates. Salamanders tended to spend more time in ESCAPE (paired t-test; 58 Behavioural interactions between salamanders and centipedes We observed no instances of IGP in behavioural pairings. Centipedes occasionally chased and appeared to bite salamanders and, in several cases, salamanders bit centipedes, but no injuries were observed. In no cases were bites by any species held for extended periods and we have no evidence that bitten salamanders were envenomated by centipedes. When paired with centipedes, salamanders spent a large portion of the trial in escape behaviour, usually climbing to the top edge of the Petri dish at some point during each trial. We did not observe this climbing behaviour in the odour trials. Compared to the odour trials, salamanders spent c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London C. D. Anthony, C. A. M. Hickerson and M. D. Venesky Interactions between salamanders and centipedes Table 2 Behaviour of juveniles of Plethodon cinereus when paired with native and introduced centipedes Behaviour of salamander FTR ATR MT IMMOBILE ESCAPE SUST ESCAPE NTS NTR NTC FLIP Lithobius, n = 30 206.1 (45.4) 49.9 (16.0) 63.3 (11.0) 834.6 (7.7) 469.3 (48.5) 219.4 (33.0) 2.7 (0.63) 0.31 (0.15) 1.0 (0.26) 0.47 (0.19) Scolopocryptops, n =30 129.8 (26.3) 92.1 (23.7) 53.9 (8.9) 818.7 (7.9) 532.0 (36.2) 261.9 (34.6) 4.1 (0.80) 1.84 (0.83) 1.1 (0.25) 0.40 (0.14) Test statistic T =0.99 T =2.45 t =0.76 t =1.53 t =1.03 t =0.85 t =1.39 T =2.17 t =0.26 T =0.61 P 0.33 NS 0.014à 0.46 NS 0.14 NS 0.31 NS 0.41 NS 0.17 NS 0.03à 0.79 NS 0.54 NS Juveniles of P. cinereus were paired with either an adult of Lithobius forficatus (the introduced centipede) or an adult of Scolopocryptops sexspinosus (the native centipede). Nose tap substrate (NTS), nose tap centipede (NTC) and flipping by the salamander (FLIP) were recorded as frequencies. All other behaviours were timed. Values are means (SE). See text for descriptions of the behaviours. ÃIndicates significance at a = 0.05. MT, move towards; FTR, front trunk raised; ATR, all trunk raised; IMMOBILE, immobility; ESCAPE, escape behaviour; SUST ESCAPE, sustained escape; NTR, nose rubbing the substrate; NS, not significant. c. 10 times as much time in escape behaviour when centipedes were present and approximately one-third as much time in the aggressive posture ATR when centipedes were present (Table 2). The submissive behaviour FLAT was observed in only three trials. We detected few differences in salamander behaviour between centipede treatments. Juveniles of P. cinereus spent significantly more time in NTR (Wilcoxon signed-ranks test; t =2.17, n = 30, P= 0.03, two-tailed; Fig. 3a) and in ATR (Wilcoxon signed-ranks test; t = 2.45, n = 30, P= 0.014, two-tailed; Fig. 3b) when paired with native centipedes, but salamanders exhibited similar levels of other behaviours in both treatments (Table 2). 3 Mean time spent in the chemosensory behaviour nose-tap-rub (NTR) (a) 2.5 b 2 1.5 1 a 0.5 0 Competition for ACOs in the laboratory Scolopocryptops and Lithobius were found under cover in the experimental chambers in 100 and 82.14% of trials, respectively. There was a weak but significant effect of the presence of the native centipede on cover use by juvenile P. cinereus. When paired with Scolopocryptops, juvenile salamanders were significantly more likely to be found either out from under the cover object (14.3% on the filter paper; 7.1% in the dry portion of the chamber) or only partially under the cover object (21.4% of salamanders; sign test; statistical n = 14, Po0.05, one tailed; Fig. 4) and there were significantly fewer pellets found under cover objects in this treatment compared to the control and Lithobius treatments (G-test of independence; G = 36.13, n = 28, Po0.001; Fig. 5). No effect of introduced centipede presence on cover object use by salamanders was detected (sign test; statistical n = 7, P40.5, one tailed). Despite the negative effect of centipede presence on cover object use by salamanders, by the end of the experiment salamanders shared cover objects with native and introduced centipedes in 60.7 and 53.5% of trials, respectively. As in the behavioural pairings, we observed no evidence of IGP between the species tested, even with the extended length (at least 17.5 h) of the trials. 140 Time spent in the aggressive behaviour all-trunk-raised (ATR) Lithobius Scolopocryptops (b) 120 100 80 a 60 40 20 0 Lithobius Scolopocryptops b Figure 3 Responses of juvenile salamanders when paired with introduced (Lithobius, n = 30) and native (Scolopocryptops, n = 30) centipedes. Juvenile salamanders spent significantly more time in investigative and aggressive behaviours when paired with native centipedes. Mean time spent in (a) the investigative behaviour nose tap rub (NTR) and (b) the aggressive posture all-trunk raised (ATR) by salamanders. Different letters above bars indicate statistically significant differences. c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London 59 Interactions between salamanders and centipedes C. D. Anthony, C. A. M. Hickerson and M. D. Venesky 100% 75% 50% 25% 0% Control Dry portion of chamber Cover Lithobius Scolopocryptops In dry portion of chamber On damp filter paper Partially under cover Completely under cover Damp filter paper Figure 4 Location of salamanders in laboratory arenas in the cover object experiment. Juvenile salamanders were less likely to be found completely under the cover object when paired with native (Scolopocryptops, n = 28) centipedes. When paired with the introduced (Lithobius, n = 28) centipedes, cover object use by salamanders did not differ significantly from controls (n = 28). 80 Percentage of fecal pellets under cover 70 60 50 40 30 20 10 0 Control Lithobius Scolopocryptops Figure 5 Location of faecal pellets left by juvenile salamanders in the laboratory cover object experiment. Juveniles were placed in arenas alone or with either a native (Scolopocryptops) or an introduced (Lithobius) centipede. After spending the night in an arena, juvenile salamanders were significantly less likely to deposit faecal pellets under cover objects when paired with native (Scolopocryptops) centipedes. Discussion In laboratory arenas, juvenile salamanders exhibited submissive behaviour when exposed to the odours of both species of centipedes, but we detected no significant differences in their responses to centipede odour treatments. Juveniles of P. cinereus spent significantly more time in escape beha60 viour and in a flattened (submissive) posture when presented with native centipede odours. The inability to detect odours of introduced centipedes may be due to lack of an innate avoidance of non-native species or lack of exposure to nonnative centipedes. At our field sites, the non-native species was rare compared with the native species; thus there may have been few chances for learned avoidance of introduced centipedes by juvenile salamanders. These results are consistent with those of Murray, Roth & Wirsing (2004), who argued that predator avoidance behaviour tended to be learned, not innate, in several western North American amphibian species. Alternatively, lithobiid centipedes may not pose a predatory threat to juveniles of P. cinereus. Despite significant size and mass differences between centipedes and salamanders, we found no evidence of IGP of juvenile salamanders by either species of centipede in any laboratory pairings. Previous studies pairing large wolf spiders Gladicosa pulchra with juvenile spotted salamanders Ambystoma maculatum and with juvenile ground skinks Scincella lateralis reported significant predation on both vertebrate species (Rubbo et al., 2001, 2003). Our centipedes were treated similarly to the spiders used in the studies by Rubbo and colleagues. Centipedes were denied food before testing and were paired with salamanders in small, structurally simple arenas. The lack of predation in our study could be a result of noxious skin secretions of the salamanders (Brodie, 1977) or differences in prey handling between spiders and centipedes. Centipedes kill large prey through venom injection before using their mandibles to macerate the prey item (Lewis, 1981). Spiders often begin the digestion process externally via venom and regurgitated stomach enzymes (Foelix, 1996). It is possible that spiders are effectively able to avoid adhesive and/or noxious skin secretions by breaking down the secretion before ingestion. Centipedes, however, may not prey on salamanders because they may be unable to insert their forcipules or chew through the adhesive skin secretions of salamanders. Because our centipedes were denied food before testing, continued to feed and gain mass in the months following testing, and in one experiment were held with their potential vertebrate prey overnight, we are confident that these two species of centipede do not consume juvenile red-backed salamanders. IGP is an important behavioural attribute of food webs and it has the potential to add significant complexity to food-web interactions. IGP can be categorized as symmetrical (looping; Polis, Myers & Holt, 1989), in which species A and B are mutual predators of one another, or asymmetrical, in which species A always preys on species B. IGP in each of these categories can be influenced by ontogenetic changes in size and resulting vulnerability to predation. IGP can also be influenced by prey handling and antipredator behaviour of IG predators and prey. It is often assumed that symmetrical IGP occurs as a result of ontogenetic reversal of predation, such that adults of species A eat juveniles of species B and adults of B eat juveniles of A (Polis et al., 1989). With regard to salamander and centipede interactions, it is further assumed that because both groups are generalist predators that experience large changes in size Percentage of salamanders c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London C. D. Anthony, C. A. M. Hickerson and M. D. Venesky Interactions between salamanders and centipedes through ontogeny, and because centipedes are venomous, the likelihood of symmetrical IGP is high. However, we found no evidence of predation by adult centipedes on juvenile salamanders, and we caution that researchers should not assume symmetrical IGP is occurring based simply on differences in body size and trophic membership. Ducey et al. (1999) reported similar results when examining trophic interactions between small terrestrial vertebrates (salamanders and small snakes) and an introduced predatory flatworm. Despite the larger size of predators used in the experiment, flatworms were rejected as suitable prey in most cases. Juvenile salamanders exhibited increased ATR towards native centipedes and exhibited more chemosensory behaviour towards native centipedes and their odours. The aggressive behaviour ATR has been well studied in adult Plethodon (Mathis et al., 1995), and Hickerson et al. (2004) concluded that ATR functions as a threat display when exhibited by adults of P. cinereus towards native centipedes. The function of ATR in juvenile behaviour is less clear. In our trials, the time spent in this behaviour decreased drastically when centipedes were present compared with the odour trials where they were not. In the presence of centipedes, juvenile salamanders shifted from ATR to escape behaviour or to immobility. If ATR in juveniles is a threat display that functions in a competitive context, then the lack of ATR exhibited towards non-native centipedes may be due to inexperience (Murray et al., 2004) or it may result from decreased niche overlap between L. forficatus and red-backed salamanders. Lithobius forficatus is known to be omnivorous during periods of reduced prey availability (Lewis, 1965) and the species occurs more commonly in disturbed habitats (Lee, 1980) than does P. cinereus. Our field and laboratory data suggest that juveniles of P. cinereus and centipedes avoid one another. In laboratory trials, the native centipede excluded juvenile salamanders from cover and we found fewer instances of co-occurrence of these two species in the field than expected. Hickerson et al. (2004) reported similar responses by adults of P. cinereus towards native centipedes in the field. Faecal deposition patterns by juvenile salamanders, in our shared cover object experiment, are consistent with this result. Juveniles deposited pellets under cover objects when placed in arenas alone and when paired with introduced centipedes, but they deposited pellets exclusive of cover when paired with native centipedes. This suggests that salamanders either were unable to access space beneath cover or were marking an area exclusive of the cover object when paired with native centipedes. Adults of P. cinereus deposit pheromones on faecal pellets (Simons, Felgenhauer & Thompson, 1999) and territory intruders use pellets to gain information regarding the competitive abilities of territory holders (Mathis, 1990). The function of faecal pellet placement and/or marking by juvenile salamanders is unknown, but juveniles of P. cinereus do possess active post-cloacal glands (Simons, Jaeger & Felgenhauer, 1995), the source of territorial pheromones in red-backed salamanders (Simons & Felgenhauer, 1992). These studies are the first to examine behavioural interactions between juveniles of P. cinereus and introduced and native invertebrate predators. As amphibian populations experience global population declines (Lannoo, 2005), it is increasingly important to strive to understand the factors responsible for these declines. We found no evidence that centipedes (introduced or native) prey on terrestrial salamanders, but we cannot discount negative competitive effects from such species. Predatory macrofauna such as spiders, beetles and centipedes reach large biomasses (Lewis, 1965; Wise & Chen, 1999; Scheu et al., 2003) and likely interact strongly with intermediate vertebrate predators such as red-backed salamanders. Recent behavioural studies indicate that salamanders recognize and respond to some of these invertebrate predators in predictable ways (Gall et al., 2003; Hickerson et al., 2004; this study). We are currently conducting field removals of forest-floor predators in the hope of providing a clearer picture of the complex interactions that occur within forest-floor food webs. Acknowledgements We thank J. Keiper of The Cleveland Museum of Natural History and The Ohio Conservation Alliance for funding portions of this study. The Cuyahoga Valley National Park granted permission to conduct fieldwork and provided access to the Woodlake Environmental Field Station. The field portions of this study were conducted under National Park Service scientific research permit number CUVA2004-SCI-0010 to C. Hickerson. The John Carroll University IACUC approved field and laboratory studies under protocol number JCU503. C. Anthony was supported by a JCU Faculty Summer Fellowship. Comments provided by an anonymous reviewer helped clarify portions of the manuscript. K. Ivanov, O. Lockhart, J. Loew and J. Milligan assisted in field site set-up and data collection. References Alford, R.A. & Richards, S.J. (1999). Global amphibian declines: a problem in applied ecology. Annu. Rev. Ecol. Syst. 30, 133–165. Brodie, E.D. Jr. (1977). Salamander antipredator postures. Copeia 1977, 523–535. Burton, T.M. & Likens, G.E. (1975). Salamander populations and biomass in the Hubbard Brook Experimental Forest, New Hampshire. Copeia 1975, 541–546. Ducey, P.K., Messere, M., LaPoint, K. & Noce, S. (1999). Lumbricid prey and potential herpetofaunal predators of the invading terrestrial flatworm Bipalium adventitium (Turbellaria: Tricladida; Terricola). Am. Midl. Nat. 141, 305–314. Foelix, R.F. (1996). Biology of spiders. 2nd edn. Oxford: Oxford University Press. Frund, H.C., Balkenhol, B. & Ruszkowski, B. (1997). Chilopoda in forest habitat-islands in north-west Westphalia, Germany. Entomol. Scand. 51, 107–114. 61 c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London Interactions between salamanders and centipedes C. D. Anthony, C. A. M. Hickerson and M. D. Venesky Gall, S.B., Anthony, C.D. & Wicknick, J.A. (2003). Are behavioral interactions between salamanders and beetles indicative of a guild relationship? Am. Midl. Nat. 149, 363–374. Green, D.M. (2003). The ecology of extinction: population fluctuation and decline in amphibians. Biol. Conserv. 11, 331–343. Hickerson, C.M., Anthony, C.D. & Walton, B.M. (2005). Edge effects and intraguild predation in native and introduced centipedes: evidence from the field and from laboratory microcosms. Oecologia 146, 110–119. Hickerson, C.M., Anthony, C.D. & Wicknick, J.A. (2004). Behavioural interactions between salamanders and centipedes: competition in divergent taxa. Behav. Ecol. 15, 679–686. Highton, R. (2005). Declines of eastern North American woodland salamanders (Plethodon). In Amphibian declines: the conservation status of United States species: 34–46. Lannoo, M. (Ed.). California: University of California Press. Houlahan, J.E., Findlay, C.S. & Schmidt, B.R. (2000). Quantitative evidence for global amphibian population declines. Nature 404, 752–755. Hurlbert, S.H. (1984). Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54, 187–211. Jaeger, R.G. (1981). Dear enemy recognition and the costs of aggression between salamanders. Am. Nat. 117, 962–974. Jaeger, R.G. (1984). Agonistic behaviour of the red-backed salamander. Copeia 1984, 309–314. Kats, L.B. & Ferrer, R.P. (2003). Alien predators and amphibian declines: review of two decades of science and the transition to conservation. Divers. Distrib. 9, 99–110. Lannoo, M. (2005). Amphibian declines: the conservation status of United States species. California: University of California Press. Lee, R.E. Jr. (1980). Summer microhabitat distribution of some centipedes in a deciduous and coniferous community of central Ohio (Chilopoda). Entomol. News 91, 1–6. Lewis, J.G.E. (1965). The food and reproductive cycles of the centipedes Lithobius variegatus and Lithobius forficatus in a Yorkshire woodland. Proc. Zool. Soc. Lond. 144, 269–283. Lewis, J.G.E. (1981). The biology of centipedes. Cambridge: Cambridge University Press. Maerz, J.C., Karuzas, J.M., Madison, D.M. & Blossey, B. (2005). Introduced invertebrates are important prey for a generalist predator. Divers. Distrib. 11, 83–90. Mathis, A. (1990). Territorial salamanders assess sexual and competitive information using chemical signals. Anim. Behav. 40, 953–962. Mathis, A., Jaeger, R.G., Keen, W.H., Ducey, P.K., Walls, S.C. & Buchanan, B.W. (1995). Aggression and territoriality by salamanders and a comparison with the territorial behavior of frogs. In Amphibian biology, Vol. 2, social behavior: 636–676. Heatwole, H. & Sullivan, B. (Eds). Chipping Norton: Surrey, Beatty, and Sons. Murray, D.L., Roth, J.D. & Wirsing, A.J. (2004). Predation risk avoidance by terrestrial amphibians: the role of prey experience and vulnerability to native and exotic predators. Ethology 110, 635–647. Ovaska, K. & Smith, M.A. (1988). Predatory behaviour of two species of ground beetles (Coleoptera: Carabidae) toward juvenile salamanders, Plethodon vehiculum. Can. J. Zool. 66, 599–604. Pfingsten, R.A. & Downs, F.L. (1989). Salamanders of Ohio. Ohio Biol. Surv. Bull. New Series Vol. 7, No. 2. Polis, G.A., Myers, C.A. & Holt, R.D. (1989). The ecology and evolution of intraguild predation: potential competitors that eat each other. Annu. Rev. Ecol. Syst. 20, 297–330. Pough, H.F. (1983). Amphibians and reptiles as low-energy systems. In Behavioural energetics: the cost of survival in vertebrates: 141–188. Aspey, W.P. & Lustick, S.I. (Eds). Ohio: Ohio State University Press. Rubbo, M.J., Townsend, V.R., Smyers, S.D. & Jaeger, R.G. (2001). The potential for invertebrate–vertebrate intraguild predation: the predatory relationship between wolf spiders (Gladicosa pulchra) and ground skinks (Scincella lateralis). Can. J. Zool. 79, 1465–1471. Rubbo, M.J., Townsend, V.R., Smyers, S.D. & Jaeger, R.G. (2003). An experimental assessment of invertebrate/vertebrate predation: the interaction between wolf spiders (Gladicosa pulchra) and terrestrial salamanders (Ambystoma maculatum). J. Zool. (Lond.) 261, 1–5. Scheu, S., Albers, D., Alphei, J., Buryn, R., Klages, U., Migge, S., Platner, C. & Salamon, J.-A. (2003). The soil fauna community in pure and mixed stands of beech and spruce of different age: trophic structure and structuring forces. Oikos 101, 225–238. Shelley, R.M. (2002). A synopsis of the North American centipedes of the order Scolopendromorpha (Chilopoda). Memoir no. 5, Virginia Museum of Natural History. Simons, R.R. & Felgenhauer, B.E. (1992). Identifying areas of chemical signal production in the red-backed salamander, Plethodon cinereus. Copeia 1992, 776–781. Simons, R.R., Felgenhauer, B.E. & Thompson, T. (1999). Description of the postcloacal glands of Plethodon cinereus, the red-backed salamander, during bouts of scent marking. J. Morphol. 242, 257–269. Simons, R.R., Jaeger, R.G. & Felgenhauer, B.E. (1995). Juvenile terrestrial salamanders have active postcloacal glands. Copeia 1995, 481–483. Walton, B.M. & Steckler, S. (2005). Contrasting effects of salamanders on forest-floor macro- and mesofauna in laboratory microcosms. Pedobiologia 49, 51–60. Williams, S.R. & Hefner, R.A. (1928). The millipedes and centipedes of Ohio. Ohio Biol. Surv. Bull. 33, 140–143. Wise, D.H. & Chen, B. (1999). Impact of intraguild predators on survival of a forest-floor wolf spider. Oecologia 121, 129–137. Wise, S.E. & Jaeger, R.G. (1998). The influence of tail autotomy of agonistic behaviour in a territorial salamander. Anim. Behav. 55, 1707–1716. Wyman, R.L. (1998). Experimental assessment of salamanders as predators of detrital food webs: effects on invertebrates, decomposition and the carbon cycle. Biodivers. Conserv. 7, 641–650. 62 c c Journal of Zoology 271 (2007) 54–62  2006 The Authors. Journal compilation  2006 The Zoological Society of London