Behavioral interactions between salamanders and centipedes: competition in divergent taxa

Transcription

1 Behavioral Ecology Vol. 15 No. 4: DOI: /beheco/arh064 Behavioral interactions between salamanders and centipedes: competition in divergent taxa Cari-Ann M. Hickerson, Carl D. Anthony, and Jill A. Wicknick Department of Biology, John Carroll University, University Heights, OH 44118, USA Red-backed salamanders, Plethodon cinereus, use territorial advertisement in the form of agonistic displays and pheromonal scent marking as a mechanism for intraspecific interference competition. Although ecological and behavioral interactions among species of salamanders have been well studied, little is known about the interactions between territorial P. cinereus and other ecologically similar species, such as large predatory invertebrates. Our field data indicate that P. cinereus and a large syntopic centipede, Scolopocryptops sexspinosus, exhibit negative spatial associations in natural habitats, possibly indicating interspecific territoriality. Only seven instances of salamander/centipede co-occurrence were recorded from a field sample of 247 occupied cover objects. Cover object size was positively correlated with salamander SVL (tip of the snout to the anterior end of the cloaca), but there was no correlation of cover object size to centipede length. Data on the ability of P. cinereus to differentiate among chemicals on the substrate suggest that visual cues are not necessary to elicit a territorial response from intruding salamanders. Although in laboratory trials salamanders behaved similarly toward intruders of both species, biting was directed only toward centipedes. Salamanders spent significantly more time approaching centipedes than they did approaching other salamanders. Approach behavior was often associated with nose tapping and may be an investigative, rather than aggressive, behavior. We suggest that territorial P. cinereus respond similarly to intruding salamanders and centipedes, but that they escalate more readily to biting centipedes because S. sexspinosus is sightless and thus unable to respond to visual signals. Key words: aggression, centipede, Plethodon cinereus, salamander, Scolopocryptops sexspinosus, spatial distribution. [Behav Ecol 15: (2004)] Interspecific competition is an important ecological factor that defines community and guild structure. In addition, selection for reduced interspecific competition results in character displacement when ecologically similar species coexist. Thus, studies of competitive interactions among species are crucial to our understanding of ecological and evolutionary processes. Most studies of interspecific competition consider sympatric congeners because it is assumed that closely related species tend to compete more strongly (Darwin, 1859; Jaksic, 1981). Guilds have been described within birds (Holmes et al., 1979; Root, 1967), insects ( Joern and Lawlor, 1981), salamanders (Hairston, 1980), desert lizards (Pianka, 1980), and mammals (MacMahon, 1976). Although most of the focus in the investigation of guild structures remains within closely related groups, a few researchers have examined interactions between unrelated taxa. Introduction and removal experiments have shown that insects and frogs compete for pond periphyton (Morin et al., 1988) and that lizards compete with spiders for insect prey (Schoener and Spiller, 1987). Studies examining guilds encompassing different phyla could help to assess the importance of competition in community structure. Interactions among species that comprise guilds may range from exploitative competition, in which one individual is more efficient at acquiring resources than another, to interference competition, in which one individual physically excludes another from access to resources. Territoriality, a type of interference competition, is one way in which species compete for resources such as food or space. It is an important mechanism mediating competition within and between Address correspondence to C. D. Anthony. canthony@jcu. edu. C. M. Hickerson is now at Department of Biological, Geological, and Environmental Science, Cleveland State University, Cleveland, Ohio 44115, USA. J. A. Wicknick is now at University of Montevallo, Department of Biology, Montevallo, Alabama 35115, USA. Received 30 March 2003; revised 8 September 2003; accepted 11 October salamander species (Anthony et al., 1997; Gergits, 1982; Nishikawa, 1985; Wrobel et al., 1980). Terrestrial salamanders remain on the surface and forage in leaf litter as long as the surface remains wet but will move under rocks and logs as conditions become too dry for them to freely roam on the forest floor. For salamanders, territoriality is adaptive because moisture and prey can become concentrated under cover objects. Therefore, territories beneath rocks and logs on the forest floor allow salamanders access to the surface for longer periods, and they become refugia for isolated prey populations as the forest dries. Extensive laboratory and field evidence shows that the red-backed salamander, Plethodon cinereus (Caudata: Plethodontidae), is territorial (Gergits and Jaeger, 1990; Horne and Jaeger, 1988; Jaeger, 1981; Jaeger et al., 1982). Individuals of P. cinereus exhibit site tenacity, advertise their presence in and are able to defend an area, and expel intra- and interspecific intruders from that area. P. cinereus and some members of the class Chilopoda occupy the same microhabitat, consume similar types of prey, and are similar in body size. Both P. cinereus and the centipede Scolopocryptops sexspinosus (Scolopendromorpha: Scolopocryptopidae) are widely distributed in eastern North America (Petranka, 1998; Shelley, 2002) and can be found living in soil and humus and under rocks, bark, and rotting logs on the forest floor (Lewis, 1981; Petranka, 1998). Blower (1955) described S. sexspinosus as a surface roamer that moves about in the litter when moisture is abundant and later retreats to cover objects. This use of microhabitat is similar to that of P. cinereus (Fraser, 1976). Roberts (1956) examined the gut contents of five species of Lithobius centipedes from an English woodland and found that during March and April approximately 70% contained Collembola remains, a common prey item for P. cinereus (Jaeger, 1990). In addition to consuming similar types of prey, the foraging tactics adopted by both salamanders and centipedes are quite similar. Similar to P. cinereus, centipedes emerge to hunt mostly at night and in periods of high moisture and cooler temperatures (Summers and Uetz, 1979). Laboratory studies of salamanders and centi- Behavioral Ecology vol. 15 no. 4 Ó International Society for Behavioral Ecology 2004; all rights reserved.

2 680 Behavioral Ecology Vol. 15 No. 4 pedes have demonstrated a comparable switch in behavioral search tactics. At high prey densities, salamanders ( Jaeger and Barnard, 1981; Jaeger et al., 1982) and centipedes (Formanowicz and Bradley, 1987) adopt a sit and wait tactic, whereas at low prey densities, search activity increases. It is not known if centipedes are territorial, but they bite (Lewis, 1981; Shelley, 2002; Williams and Hefner, 1928), and could potentially injure or kill a similarly sized salamander. Thus, there is potential for exploitative or interference competition between P. cinereus, and temperate forest centipedes. The purpose of the present study was to examine ecological and behavioral interactions between the red-backed salamander, P. cinereus, and an ecologically similar centipede species, S. sexspinosus. Although P. cinereus and S. sexspinosus occupy the same microhabitat, it is unknown whether they negatively associate within that microhabitat. The first part of the present study included the collection of field data on the microdistribution of these two species. We questioned whether the two species co-occur under the same cover objects on the forest floor or whether they exhibit negative associations. Classic studies of species removal in zones of overlap give strong support for competitive interactions between species and show the importance of spatial distribution in determining species interactions (Connell, 1961; Hairston, 1980). The second part of the present study was designed to resolve the following question: Can members of P. cinereus differentiate between a substrate formerly occupied by a centipede and one formerly occupied by a common prey species? The ability of P. cinereus to recognize differences in chemicals (pheromones) deposited on a substrate is thought to give intruding salamanders information necessary to make decisions about whether or not to increase agonistic behavior in a given encounter with a territorial resident (Simons et al., 1997). Littlewood and Blower (1987) found that in laboratory trials the common brown centipede, Lithobius forficatus, deposits a chemical mark on damp filter paper that affects the behavior of conspecifics. The pheromone is apparently sexspecific and may function in intraspecific chemical communication. The final aspect of the present study was to examine aggressive behavior of P. cinereus toward centipedes. Individuals of P. cinereus have been shown to exhibit aggressive behavior toward conspecifics (see Jaeger, 1984), congenerics (see Jaeger et al., 1998), ambystomatid salamanders (Ducey et al., 1993), and carabid beetles (Gall et al., 2003). Aggressive behavior between species that are thought to compete may indicate a behavioral mechanism of interference competition. Microhabitat separation, the ability to use chemoreception to identify threatening intruders, and aggressive territorial behavior are all suggestive of competitive interactions between species. We therefore proposed the following hypotheses: (1) P. cinereus and S. sexspinosus will exhibit negative spatial associations on the forest floor, and (2) individuals of P. cinereus will behave similarly to conspecifics and to centipedes. Specifically, we predicted that residents of P. cinereus would exhibit similar levels of aggressive behavior toward conspecific and centipede intruders and that resident salamanders would exhibit similar levels of aggressive behavior in response to the odors of both species. METHODS Distribution in the natural habitat To determine distribution patterns of P. cinereus and S. sexspinosus, field data were collected from 10 forested sites in northeastern Ohio (Cuyahoga County) during September and October We overturned 100 cover objects (logs, rocks, and bark greater than 25 cm 2 ) at each of the 10 sites. For each cover object, we noted whether adult P. cinereus and adult S. sexspinosus co-occurred, occurred alone, or were absent. The absence category consisted of cover objects that had juveniles of either study species, or other species of salamanders or centipedes beneath them. Cover objects that did not meet these criteria were excluded from the analysis; thus, we only included cover objects that were suitable for use by salamanders and centipedes. Frequency of P. cinereus/ S. sexspinosus co-occurrence was compared to an expected frequency by using a two-tailed chi-square test with a ¼ 0.05 (Zar, 1999). Territory quality Individuals should compete for high-quality territories when resources are limited. Larger cover objects are thought to be superior because they hold moisture and associated prey for longer time periods, and larger salamanders are thought to be better territory holders owing to the advantages that large size confer during territorial conflicts (Mathis, 1990a). Therefore, the largest salamanders should be found in the best territories (i.e., under the largest cover objects). On 7, 11, and 16 June 2001 we recorded the sizes of cover objects and associated salamanders and centipedes in a forested area in Independence, Cuyahoga County, Ohio ( N; W). We worked systematically through the site to avoid measuring cover objects more than once. Each cover object occupied by individuals of P. cinereus or S. sexspinosus was measured (greatest length times greatest width). Salamanders were measured from the tip of the snout to the anterior end of the cloaca (SVL). Centipedes were measured from the tip of the upper mandible to the posterior end of the last body segment. Relationships between cover object size and salamander SVL or centipede body length were analyzed by using two-tailed Spearman rank correlation a ¼ 0.05 (Mathis, 1990a; Zar, 1999). To ensure independence of samples, cover objects that contained more than one individual were not used in either analysis. Recognition of chemicals on the substrate Twelve adult males of P. cinereus (more than 32 mm SVL; Pfingsten and Downs, 1989) and 12 presumed adults (unsexed) of S. sexspinosus (more than 32 mm) were collected from Independence, Cuyahoga County, Ohio ( N; W) in June We assumed that centipedes were adults because they were similar in size to brooding females that we observed in the field. Individuals of both species were housed separately in 15-cm-diameter Petri dishes (1.6 cm deep) lined with damp filter paper. Chambers were maintained in a controlled environment at C and 12-h light/12-h dark photoperiod. Each individual was fed fruit flies (Drosophila virilis) and allowed to mark territories with pheromones for 5 days. Flies not eaten from experimental containers were removed 2 days before testing. Flies were placed on control substrates at the same time that experimental animals were fed and were removed the day of testing, just before the trial in which they were used. Flies were removed from the experimental containers earlier than from the control containers to ensure that the scent of the flies had time to dissipate from experimental containers. This was to avoid confusing the salamanders with the scent of prey in addition to the scent of conspecifics or centipedes. On day 6, individuals of P. cinereus were randomly tested in one of three trial types. There were two experimental conditions, P. cinereus on a conspecific s substrate or P. cinereus on a centipede s substrate, and a control condition (substrate occupied by D. virilis alone). Twelve trials were conducted for each of the

3 Hickerson et al. Behavioral interactions between salamanders and centipedes 681 three conditions from h in February Salamanders were observed under indirect lighting (two 15- W GE fluorescent Cool White bulbs, 3.0 lux). Treatments and controls were evenly dispersed across test dates (Hurlbert, 1984). Comparisons among treatments were made using twotailed paired Wilcoxon signed-rank test. We reduced a to because each data set was used twice in each analysis. Each salamander was placed on one of the three types of substrates under an opaque, circular habituation dish (5 cm diameter by 1.6 cm deep) for 5 min. The dish was then lifted, and the salamander was permitted to react freely to the scent on the rest of the substrate. Individuals whose substrates were being used in a trial were held in a clean holding chamber lined with damp filter paper. Salamanders whose territory was used in a trial and who were to be tested the same day remained in the holding chamber until they were tested. This avoided cross-contamination of another salamander s scent onto the home territory. Each salamander was tested in the control condition and each of the two experimental conditions. Frequency and duration of the following behaviors were recorded for P. cinereus: (1) All trunk raised (ATR) was considered an aggressive posture; the legs are extended such that the head, trunk, and sometimes tail are lifted off of the substrate ( Jaeger, 1984). (2) Flattened (FLAT) was considered a submissive posture; the entire ventral surface of the body is in contact with the substrate ( Jaeger, 1984). (3) Nose tapping the substrate (NTS) was considered an investigative behavior, contact of the nasolabial cirri to the substrate. Often the salamanders would hold their snout to the substrate for several seconds at one time, and sometimes to fecal pellets. (4) Escape (E) behavior was defined as circling the periphery of the chamber while pressing the snout or body upward against the Petri dish lid at the edge of the testing arena. This is considered a submissive behavior (Wise and Jaeger, 1998). Figure 1 Relationship between cover object size (cm 2 ) and (a) SVL (mm) of Plethodon cinereus (R 2 ¼.321, p ¼.004) and (b) total body length (mm) of Scolopocryptops sexspinosus (R 2 ¼.118, p ¼.275). Aggressive behavior Adult male salamanders (P. cinereus, n ¼ 30) and adult centipedes (S. sexspinosus, n ¼ 30) were collected from the same forested area described above and tested for agonistic interactions in June Individuals of each species were collected and housed separately in 15-cm-diameter (1.6 cm deep) clear plastic Petri dishes lined with damp filter paper. Both species were fed Drosophila virilis ad libitum. Chambers were maintained in a controlled environment at C and 12-h light/12-h dark photoperiod. Salamanders and centipedes were weighed and measured prior to testing (mean SVL and mass of salamanders ¼ mm, SE ¼ 0.51 and 1.09 g, SE ¼ 0.03, mean body length and mass of centipedes ¼ mm, SE ¼ 0.78, and g, SE ¼ 0.01). Before experimental interactions, individuals were kept in separate containers for 5 days, enough time to allow marking of the substrate with pheromones ( Jaeger, 1981). Thirty resident salamanders were randomly paired once with an intruding conspecific and once with an intruding centipede, for a total of 60 trials. Trials were run from h in June and July On day 6 an intruding centipede or conspecific was placed in a resident salamander s chamber, under an opaque circular habituation dish (5 cm diameter by 1.6 cm deep). The resident salamander was also lifted and replaced in its own chamber under an opaque habituation dish, to control for handling. After a 5-min habituation period, both habituation dishes were lifted, allowing interaction between resident and intruder. Behavioral interactions were observed and recorded for both resident and intruder for 15 min under the lighting conditions described above. The first interactive behavior (either look toward or move toward) signaled the start of the 15-min trial period. Each salamander was used as a resident and an intruder in intraspecific trials but only as a resident in interspecific trials. One week was allowed to pass between use as a resident or an intruder, and no salamander was paired with the same individual more than once. Size asymmetries were minimized to reduce the fighting advantage by the larger animal (Mathis, 1990a). This was accomplished by randomly pairing individuals within three size classes. Paired individuals differed by a mean of 2.44 mm (6 1.8 mm). The frequency and duration of the following behavior patterns (as defined in Jaeger, 1984) were recorded for P. cinereus residents. The following were aggressive behaviors: all trunk raised (ATR), as defined in previous section; look toward (LT), salamander turns its head in the direction of the other animal; move toward (MT), salamander approaches the other animal in a direct path that would result in contact if the movement were to continue; and biting (BITE), closing of the jaws around any part of the other animal s body and immediately releasing the grip. The following were considered submissive behaviors: flattened (FLAT), as defined in previous section. Look away (LA) salamander turns its head away from the other animal; and move away (MA), salamander increases the distance between itself and the other animal. Other behaviors were as follows: nose tapping substrate (NTS), as defined in previous section; nose tapping animal (NTA), contact of nasiolabial cirri to other animal s body; front trunk raised (FTR), considered a resting posture, the head and anterior half of the trunk are raised off the substrate; walk on (WO), salamander makes contact with and moves over the other animal; and walk under (WU), salamander makes bodily contact with the other animal, moving or nudging under it.

4 682 Behavioral Ecology Vol. 15 No. 4 Figure 2 Behavior of P. cinereus when placed on substrates previously occupied by a conspecific (Pc indicates Plethodon cinereus), a centipede (Ss indicates Scolopocryptops sexspinosus) or on a control substrate (fly: Drosophila virilis). (a) Mean time (s) spent in aggressive ATR posture. (b) Mean frequency of aggressive ATR posture. (c) Mean time (s) spent nose tapping the substrate (NTS). (d) Mean frequency of nose taps to substrate (NTS). (e) Mean frequency of escapes. Lowercase letters above the bars indicate statistical differences at p ¼.025. Statistical comparisons were made between the behavior of resident salamanders in intraspecific trails and resident salamanders in interspecific trials by using paired Wilcoxon signed-rank tests (two-tailed). Paired tests were used because each salamander was tested once with a conspecific and once with a centipede intruder. We compared the behavior of resident salamanders in intraspecific trials to their behavior when they were tested as intruders by using paired Wilcoxon signedrank tests. One-tailed tests were used for this comparison because previous studies support the prediction that resident salamanders are more aggressive and less submissive toward conspecific intruders (Anthony et al., 1997; Jaeger, 1984; Mathis, 1990b;). a was set at 0.05 for all comparisons (Zar, 1999). An aggression index using the aggressive behaviors MT and LT and the submissive behaviors MA and LA was calculated as [(MT þ LT) ÿ (MA þ LA)] (Mathis et al., 2000) and used in all of the above comparisons. RESULTS Distribution in the natural habitat and territory quality Weather conditions varied among sampling trips (temperature range ¼ 9 Cÿ21 C), but both species were active on the surface during each sampling day (mean number of salamanders ¼ 18, SE ¼ 3.5; mean number of centipedes ¼ 15.5,

5 Hickerson et al. Behavioral interactions between salamanders and centipedes 683 SE ¼ 5.1). Of 247 occupied cover objects sampled, we found that there were only seven instances in which adults of P. cinereus and S. sexspinosus co-occurred under the same cover object. This is significantly less than would be expected by chance (v 2 ¼ 65.4, p,.001). Salamanders were found alone under 85 cover objects, centipedes were found alone under 79 cover objects, and 76 cover objects housed neither adult P. cinereus or adult S. sexspinosus. Salamanders co-occurred under three cover objects, and centipedes co-occurred under five cover objects. These data were not included in the analysis. Cover object size ranged from cm 2 (mean ¼ cm 2, n ¼ 55) and was positively correlated with SVL of P. cinereus (p ¼.004) (Figure 1a). However, body length of S. sexspinosus showed no correlation with cover object size (p ¼.275) (Figure 1b). Recognition of chemicals on the substrate Salamanders spent significantly more time in ATR when placed on substrates previously occupied by either conspecifics (z ¼ 2.71, p ¼.007) or centipedes (p ¼.006) (Figure 2a and Table 1). Frequency of ATR did not significantly differ across treatments (Figure 2b). P. cinereus spent significantly less time nose tapping the substrate of flies than that of other salamanders (z ¼ 2.51, p ¼.01) (Figure 2c). Time spent nose tapping the substrate of centipedes was not significantly different from time spent nose tapping salamander (z ¼ 0.36, p ¼.72) or fly (z ¼ 1.41, p ¼.16) substrates (Figure 2c). Differences in frequency of nose taps were not significant (z ¼ 1.81, p ¼.07) between salamander and fly substrates (Figure 2d); there was no significant difference in the frequency of nose taps between salamander and centipede (z ¼ 0.47, p ¼.64) or centipede and fly (z ¼ 1.49, p ¼.14) substrates. There was a tendency for P. cinereus to exhibit higher levels of escape behavior when exposed to a centipede s substrate compared with a conspecific s substrate (z ¼ 1.95, p ¼.05) (Figure 2e), and frequency of escape by salamanders was significantly higher on centipede substrates compared with the control (z ¼ 2.39, p ¼.02) (Figure 2e). The mean number of escapes was nearly double on centipede substrate compared with salamander trials and control trials (Table 1). Aggressive behavior Comparisons of the aggression indices revealed that P. cinereus was more aggressive as a resident than as an intruder (T ¼ 2.70, p ¼.012), but there were no differences in time spent in ATR (T ¼ 0.56, p ¼.58) or FLAT (T ¼ 0.30, p ¼.77) between residents and intruders. The response of residents of P. cinereus toward intruding conspecifics was comparable to intruding centipedes, with no differences in the aggression index (T ¼ 0.43, p ¼.67) (Figure 3a) or time spent in ATR (z ¼ 0.25, p ¼.80) (Figure 3b). Residents of P. cinereus spent significantly more time approaching (MT) centipede intruders than conspecific intruders (z ¼ 2.10, p ¼.037) (Figure 3c). Interestingly, the only biting that occurred in any trials was P. cinereus residents biting centipede intruders (mean ¼ 0.23, z ¼ 2.33, p ¼.02) (Figure 3d). Biting occurred in 8% of interspecific trials. Salamander residents spent on average half the amount of time in submissive (FLAT) position during interspecific trials compared with intraspecific pairings (z ¼ 1.54, p ¼.07) (Table 2). Residents of P. cinereus spent more time WO (z ¼ 2.77, p ¼.006) and WU (z ¼ 2.37, p ¼.018) salamander intruders than centipede intruders (Figure 3e,f). Qualitative summary of centipede behavior The centipedes were typically unresponsive to salamander movement and usually remained still unless contact was made Table 1 Behaviors of P. cinereus when placed on substrates previously occupied by conspecifics, centipedes, or on the control substrates that were previously occupied by fruit flies (Drosophila virilis) Behavior: P. cinereus (n ¼ 12) S. sexspinosus (n ¼ 12) flies (n ¼ 12) Time (s) spent in ATR (60.8) (51.4) 96.0 (36.1) NTS 21.4 (3.68) 21.2 (5.40) 10.8 (2.83) ESCAPE 26.8 (11.4) 35.9 (12.3) 20.4 (8.07) FLAT 73.4 (47.5) 26.7 (12.9) 14.8 (9.00) Frequency of ATR 7.08 (1.69) 9.25 (1.94) 6.42 (2.17) NTS 15.4 (2.25) 17.3 (2.80) 11.3 (9.32) ESCAPE 2.58 (0.97) 4.58 (1.25) 2.25 (2.90) FLAT 0.50 (0.26) 0.75 (0.28) 0.42 (0.15) Means, standard errors (parentheses), and ranges are presented. See text and Fig. 2 for significant differences. by the salamander. During all trials, centipedes frequently groomed their antennae, especially following contact with the salamander or its fecal pellets. When moving, centipedes circled the edge continuously and infrequently (10% of trials) crossed through the middle of the test chamber. There were several instances when centipedes appeared to sample the shed skin or fecal pellets of a salamander with their antennae. When centipedes walked into salamanders, making contact with their antennae, they often immediately reversed direction, stopping along the edge at the opposite side of the chamber. There were 15 instances (25% of trials) in which the centipedes walked on the salamanders. This action most often resulted in a flip by the salamander in which it spun its entire body approximately 180 degrees, but sometimes resulted in bites by the salamander. Both P. cinereus and S. sexspinosus displayed escape behavior after bites. DISCUSSION Salamanders and centipedes were negatively associated in our field sites. Only seven instances of salamander/centipede cooccurrence were recorded from a field sample of 247 occupied cover objects. Negative spatial association is suggestive of competition for cover objects on the forest floor ( Jaeger, 1979). Marvin (1998) found that, in sympatry, individuals of P. glutinosis and P. kentucki were usually found alone under cover objects, and he concluded that these species compete for space. Alternatively, negative spatial associations may result from differences in microhabitat use. Although we observed no obvious differences in microhabitat between species, variables such as soil temperature, moisture, and ph were not quantified in the present study. It is unclear what factors determine territory (cover object) quality for these species. Several researchers have examined the possibility that large cover object size defines a superior quality territory because large objects hold moisture longer, providing refuge for invertebrate prey. Our findings were similar to those of Mathis (1990a) for P. cinereus. She found

6 684 Behavioral Ecology Vol. 15 No. 4 Figure 3 Behavior of residents of P. cinereus during pairings with intruding centipedes. (a) Mean aggression index (see text). (b) Mean time (s) spent in aggressive all trunk raised (ATR) posture. (c) Mean time (s) spent approaching (MT) centipede intruders. (d) Mean frequency of biting (BITE) intruding centipedes. (e) Mean time (s) spent walking on (WO) intruding centipedes. (f) Mean time (s) spent walking under (WU) intruding centipedes. Lowercase letters above bars indicate statistical differences at p ¼.05. that SVL was positively correlated with cover object area and larger salamanders were better competitors. Other researchers have found no such correlations between SVL and cover object size in salamanders (Anthony et al., 2002; Faragher and Jaeger, 1997; Gabor, 1995; Quinn and Graves, 1999). For centipedes, we found no relationship between cover object area and resident size. It is possible that cover object size may not be as important for centipedes as it is for salamanders because centipedes have the ability to burrow through soil underground and can perhaps locate enough prey and keep from desiccating without the use of large cover objects. However, Fraser (1976) suggested that when salamanders are forced to retreat underground, their foraging opportunities are limited. If centipedes prey on similar invertebrates as salamanders and if those prey are limited underground, cover object quality should be important for both species. Males of P. cinereus responded similarly to centipede substrates and to substrates of other male conspecifics, but had a tendency to attempt escapes more frequently on substrates with centipede odors. The scent of the centipedes or of male conspecifics was sufficient to cause salamanders placed on these marked substrates to display a threat posture. Horne and Jaeger (1988) established that females of P. cinereus only needed olfactory cues left on fecal pellets to display threat postures. Jaeger et al. (1986) found that males of P. cinereus did not differ in the amount of time spent in threat postures in the presence of their own fecal pellets compared with the fecal pellets of conspecifics. Therefore, they concluded that for males a visual display is necessary to release a threat posture. Our conclusions differ from those of Jaeger et al. (1986). Not only did male P. cinereus spend more time in ATR (a threat posture) when on substrates of conspecifics, they behaved similarly when placed on a centipede substrate. Thus, males of P. cinereus in this study detected and responded to both salamander and centipede odors in a way consistent with territorial response to a conspecific intruder.

7 Hickerson et al. Behavioral interactions between salamanders and centipedes 685 Table 2 Behaviors of residents of P. cinereus when paired with conspecific and centipede intruders. Behavior Individuals of P. cinereus tended to exhibit escape behavior in Petri dishes that had been occupied by S. sexspinosus more often than during other trials. Lopez and Martin (2001) reported that amphisbaenians were also able to discriminate between snake and scolopendrid centipede odors, and Littlewood and Blower (1987) provided evidence for intraspecific chemical communication in the centipede, Lithobius forficatus. Thus, some species of centipedes produce recognizable odors. In the present study, salamanders tended to move away from centipede odors but not away from odors of conspecifics. From pheromonal markers of conspecifics, both resident and intruding salamanders are able to assess the size and sex of one another and determine if the contest should escalate (Mathis, 1990b). Perhaps individuals of P. cinereus are able to determine if intruding centipedes pose a threat by having the ability to recognize the chemicals deposited by those individual centipedes. In behavioral trials, individuals of P. cinereus behaved in a similar manner toward intruding conspecifics and intruding centipedes, exhibiting similar levels of aggressive behavior. However, resident salamanders spent more time MT centipede intruders than conspecific intruders and biting by salamanders was only observed in interspecific trials. We believe that the approach (MT) of individuals of P. cinereus toward centipedes was investigative in nature rather than aggressive. Wrobel et al. (1980) stated that for P. cinereus, an encounter begins when one individual approaches another, initiating a sequence of aggressive behavior patterns. In the present study, residents did not follow an approach with aggressive behavior when paired with centipedes, but rather the salamander approached with hesitation by slowly rocking the entire body forward without movement of the limbs until the forward motion forced the salamander to very slowly pick up a limb and take a step in the forward direction. The approach was usually followed by NTA and NTS. Once nose tapped by the salamander, the centipede would usually flee in the opposite direction. It is possible that the aggression index of salamanders paired with centipedes was elevated because of the increased duration of the MT component of the index. However levels of ATR were consistent with the aggression index in conspecific and heterospecific trials. We observed very few instances of biting during the present study. All biting that occurred was directed toward centipedes by salamanders in interspecific trials. Centipedes have been recorded in the guts of some plethodontid salamanders (Fraser, 1976) and Hoffman (1954) reported a juvenile S. sexspinosus in the gut of an adult P. richmondi. However, we believe the relationship between adults of P. cinereus and adults of S. sexspinosus is most likely not a predator/prey relationship because the size asymmetry between the two species is too small for a salamander to consume an adult S. sexspinosus. Perhaps most interesting is that all members of the centipede family Scolopocryptopidae lack ocelli and are considered sightless. S. sexspinosus does not have good visual acuity and can only respond minimally to changes in light and dark (Lewis, 1981). P. cinereus usually resolves territorial contests by using visual displays and bites only as a last resort ( Jaeger et al., 1982). In a recent study in which individuals of P. cinereus were paired with carabid beetles, there was no biting by P. cinereus of conspecific intruders or intruding beetles; both have fairly good vision (Gall et al., 2003). We hypothesize that P. cinereus is forced to escalate to biting in contests with S. sexspinosus because these centipedes are unresponsive to visual threat displays. The data from all three components of this study are supportive of competitive interactions between the centipede S. sexspinosus and the territorial salamander P. cinereus. Salamanders and centipedes exhibited negative spatial distributions, salamanders exhibited escape behavior when presented with centipede odors, and resident salamanders responded similarly to centipede and salamander intruders. Gall et al. (2003) argued that through combinations of territorial and predatory behavior, carabid beetles may have negative effects on salamander populations. We concur that such interactions between small vertebrates and large predatory invertebrates may be important determinants of community composition of forest floor predators. We thank the Cengic family for allowing us to collect on their property and Case Western Reserve University for providing access to Squire Valleevue Experimental Farm. The Ohio Department of Natural Resources granted collecting permits. Paul Cengic, Jennifer Fields, Jessica Hickey, and Moria Nagy provided assistance in the field. This research was conducted with prior approval of the Institutional Animal Care and Use Committee at John Carroll University (IACUC No ). The manuscript benefited from the comments of two anonymous reviewers. REFERENCES Resident Intraspecific (n ¼ 30) Intraspecific (n ¼ 30) Intruder Interspecific (n ¼ 30) ATR (49.3) (53.4) (47.1) LT 4.40 (0.39) 4.20 (0.43) 2.9 (0.38) MT 95.2 (12.8) (21.7) 77.4 (15.5) WO 39.1 (11.3) 3.30 (2.60) 35.4 (11.2) BITE (0.09) FLAT 36.2 (17.6) 15.4 (13.5) 30.8 (11.6) LA 1.10 (0.23) 1.0 (0.23) 0.73 (0.23) MA 34.3 (11.2) 32.5 (11.1) 41.5 (8.73) WU 7.27 (2.81) (2.27) NTS 9.23 (1.66) 8.23 (1.40) 10.4 (1.73) NTA 2.70 (0.57) 2.67 (0.62) 1.77 (0.42) FTR (43.8) (51.7) (42.4) The left two columns show mean resident data when paired with intraspecific intruders, or interspecific intruders. The right column gives means for P. cinereus as intruders paired with intraspecific residents. Means, standard errors (parentheses), and ranges are presented. LT, LA, NTS, NTA, and BITE were recorded as frequency of occurrence. All other data were recorded in seconds. Anthony CD, Wicknick JA, Jaeger RG, Social interactions in two sympatric salamanders: effectiveness of a highly aggressive strategy. Behaviour 134: Anthony CD, Wicknick JA, Jaeger RG, Site tenacity and homing in the Rich Mountain salamander, Plethodon ouachitae (Caudata: Plethodontidae). Southwest Nat 47: Blower JG, Millipedes and centipedes as soil animals. Soil Zool Univ Nottingham Proc

8 686 Behavioral Ecology Vol. 15 No. 4 Connell JH, The influence of interspecific competition and other factors on distribution of the barnacle Chthamalus stellatus. Ecology 42: Darwin C, The origin of species by means of natural selection, or the preservation of favored races in the struggle for life. New York: New American Library. Ducey PK, Schramm K, Cambry N, Interspecific aggression between the sympatric salamanders, Ambystoma maculatum and Plethodon cinereus. Am Mid Nat 131: Faragher SG, Jaeger RG, Distributions of adult and juvenile redback salamanders: testing new hypotheses regarding territoriality. Copeia 1997: Formanowicz Jr, DR, Bradley PJ, Fluctuations in prey density: effects on the foraging tactics of scolopendrid centipedes. Anim Behav 35: Fraser DF, Empirical evaluation of the hypothesis of food competition in salamanders of the genus Plethodon. Ecology 57: Gabor CR, Correlational test of Mathis hypothesis that bigger salamanders have better territories. Copeia 1995: Gall SB, Anthony CD, Wicknick JA, Are behavioral interactions between salamanders and beetles indicative of a guild relationship? Amer Mid Nat 149: Gergits WF, Interference competition and territoriality between the terrestrial salamanders Plethodon cinereus and Plethodon shenandoah (MS thesis). Albany, New York: State University of New York at Albany. Gergits WF, Jaeger RG, Site attachment by the red-backed salamander, Plethodon cinereus. J Herpetol 24: Hairston NG, Evolution under interspecific competition: field experiments on terrestrial salamanders. Evolution 34: Hoffman RL, Distributional records of two species of Plethodon in the southern Appalachians. Herpetologica 10: Holmes RT, Bonney RE Jr, Pacala SW, Guild structure of the Hubbard Brook bird community: a multivariate approach. Ecology 62: Horne EA, Jaeger RG, Territorial pheromones of female redbacked salamanders. Ethology 78: Hurlbert SH, Pseudoreplication and the design of ecological field experiments. Ecol Monogr 54: Jaeger RG, Seasonal spatial distributions of the territorial salamander Plethodon cinereus. Herpetologica 35: Jaeger RG, Dear enemy recognition and the costs of aggression between salamanders. Am Nat 117: Jaeger RG, Agonistic behavior of the red-backed salamander. Copeia 1984: Jaeger RG, Territorial salamanders evaluate size and chitinous content of arthropod prey. In: Behavioural mechanisms of food selection (Hughes, RN, ed). Heidelberg: Springer-Verlag; Jaeger RG, Barnard DE, Foraging tactics of a terrestrial salamander: choice of diet in a structurally simple environment. Am Nat 117: Jaeger RG, Gabor CR, Wilbur HM, An assemblage of salamanders in the southern Appalachian mountains: competitive and predatory behavior. Behaviour 135: Jaeger RG, Goy JM, Tarver M, Marquez CE, Salamander territoriality: pheromonal markers as advertisement by males. Anim Behav 34: Jaeger RG, Kalvarsky D, Shimizu N, Territorial behavior of the red-backed salamander: expulsion of intruders. Anim Behav 30: Jaksic FM, Abuse and misuse of the term guild in ecological studies. Oikos 37: Joern A, Lawlor LR, Guild structure in grasshopper assemblages based on food and microhabitat resources. Oikos 37: Lewis JGE, The biology of centipedes. London: Cambridge University Press. Littlewood PMH, Blower JG, The chemosensory behavior of Lithobius forficatus: evidence for a pheromone released by the coxal organs (Myripoda: Chilopoda). J Zool 211: Lopez P, Martin J, Chemosensory predator recognition induces specific defensive behaviors in a fossorial amphisbaenian. Anim Behav 62: MacMahon JA, Species and guild similarity of North American desert mammal faunas: a functional analysis of communities. Evol Desert Biota Marvin GA, Interspecific aggression and spatial relationships in the salamanders Plethodon kentucki and Plethodon glutinosus: evidence of interspecific competition. Can J Zool 76: Mathis A, 1990a. Territoriality in a terrestrial salamander: the influence of resource quality and body size. Behaviour 112: Mathis A, 1990b. Territorial salamanders assess sexual and competitive information using chemical signals. Anim Behav 40: Mathis A, Schmidt DW, Medley KA, The influence of residency status on agonistic behavior of male and female Ozark zigzag salamanders, Plethodon angusticlavius. Am Mid Nat 143: Morin PJ, Lawler SP, Johnson EA, Competition between aquatic insects and vertebrates: interaction strength and higher order interactions. Ecology 69: Nishikawa KC, Competition and the evolution of aggressive behavior in two species of terrestrial salamander. Evolution 39: Petranka JW, Salamanders of the United States and Canada. Washington: Smithsonian Institution Press. Pfingsten RA, Downs FL, Salamanders of Ohio. Ohio Biol Surv Bull New Series 7: Pianka ER, Guild structure in desert lizards. Oikos 35: Quinn VS, Graves BM, Space use in response to conspecifics by the red-backed salamander (Plethodon cinereus, Plethodontidae, Caudata). Ethology 105: Roberts RB, An ecological study of the arthropods of a mixed beech-oak woodland, with particular reference to Lithobiidae (PhD dissertation). Highfield: University of Southampton. Root RB, The niche exploitation pattern of the blue-gray gnatcatcher. Ecol Monogr 37: Schoener TW, Spiller DA, Effect of lizards on spider populations: manipulative reconstruction of a natural experiment. Science 236: Shelley RM, A Synopsis of the North American centipedes of the order Scolopendromorpha (Chilipoda). Martinsville, Virginia: Virginia Museum of Natural History, memoir no. 5. Simons RR, Jaeger RG, Felgenhauer BE, Competitor assessment and area defense by territorial salamanders. Copeia 1997: Summers G, Uetz GW, Microhabitats of woodland centipedes in a streamside forest. Am Mid Nat 102: Williams SR, Hefner RA, The millipedes and centipedes of Ohio. Ohio Bio Surv Bull 33: Wise SE, Jaeger RG, The influence of tail autotomy on agonistic behavior in a terrestrial salamander. Anim Behav 55: Wrobel DJ, Gergits WF, Jaeger RG, An experimental study of interference competition among terrestrial salamanders. Ecology 61: Zar JH, Biostatistical analysis, 4th ed. Engelwood Cliffs, New Jersey: Prentice Hall.