Behavioral Ecology Vol. 14 No. 1: 74-79
© 2003 International Society for Behavioral Ecology
Analysis of multiple aspects of a repeated signal: power and rate of rapping during shell fights in hermit crabs
School of Biology and Biochemistry, Queen's University Belfast, Belfast, UK
Address correspondence to M. Briffa, School of Biology and Biochemistry, Queen's University Belfast, Medical Biology Centre, 97 Lisburn Road, Belfast BT7 1AE, UK. E-mail: m.briffa{at}qub.ac.uk.
Received 29 January 2002; revised 17 May 2002; accepted 17 May 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Repeated activities used by animals during contests are assumed to act as signals advertising the quality of the sender. However, their exact functions are not well understood and observations fit only a limited set of the predictions made by models of signaling systems. Experimental studies of contest behavior tend to focus on analysis of the rate of signaling, but individual performances may also vary in magnitude. Both of these features can vary between outcomes and within contests. We examined changes in the rate and power of shell rapping during shell fights in hermit crabs. We show that both rate and power decline during the course of the encounter and that the duration of pauses between bouts of shell rapping increases with an index of the total effort put into each bout. This supports the idea that the vigor of shell rapping is regulated by fatigue and could therefore act as a signal of stamina. By examining different interacting components of this complex activity, we gain greater insight into its function than would be achieved by investigating a single aspect in isolation.
Key words: aggression, hermit crabs, multicomponent signals, Pagarus bernhardus, power, repeated signal.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Contests between animals are often settled by the use of repeated activities that have been interpreted as signals. A key concept is the idea that contestants assess signals to estimate the likely benefit of engaging in or continuing with a contest. In particular, contestants are expected to assess the relative fighting ability of their opponent. Models of contest behavior based on these ideas have developed from relatively simple game theoretical models, which predict fight outcomes and the roles that contestants should adopt (Maynard Smith and Parker, 1976
; Maynard Smith and Price, 1973; Parker, 1974; Parker and Rubenstein, 1981). More complex models predict how patterns of behavior should vary during the course of a contest (Enquist and Leimar, 1983, 1987; Enquist et al., 1990; Payne, 1998; Payne and Pagel, 1996, 1997). Observed behavior patterns, however, tend to meet only a limited number of the predictions of the various models (e.g., Briffa and Elwood, 2000b; Enquist et al., 1990; Turner and Huntingford, 1987), and the precise functions of these activities remain unclear. Furthermore, recent analyses of the effects of contestant weight (an index of fighting ability) on contest duration (Bridge et al., 2000; Taylor et al., 2001) have questioned the role of assessment in settling contests. Evidence from studies of shell fighting in hermit crabs (Briffa and Elwood, 2000a,b,c, 2001a,b; Briffa et al., 1998), however, suggests that the key aggressive activity used by attackers (see below) might advertise the sender's stamina. Defenders are assumed to assess the effort used by attackers, and fatigue is expected to mediate the rate of performance.
Where repeated activities have been investigated, the focus has been on the analysis of rates of signaling and how these vary both during encounters and between outcomes, but variation may also occur in magnitude (Enquist and Leimar, 1983). In the case of a repeated vocal display, for example, both the rate of performance and the sound intensity of the vocalizations might be expected to vary during the course of an encounter. Thus, the magnitude of individual performances, in addition to the temporal structure, might contain information about the sender. These are distinct from features that may show variation between individuals but are constant during the encounter. An example of the latter is the frequency (pitch) of acoustic signals used in aggressive encounters by orthopterans, which may vary between indi viduals but is produced by tegminal stridulation and is there fore fixed within encounters (e.g., Mason, 1996). Frequency is therefore unlikely to reflect the effort used by the sender in a similar manner to that of signaling rate and magnitude.
Hermit crabs engage in shell fights over ownership of the gastropod shells that they inhabit. The key agonistic activity of shell rapping involves the attacker bringing its own shell rapidly and repeatedly into contact with that of the defender in a series of bouts. (For a full description of shell fighting, seeElwood and Briffa, 2001.) The temporal structure of these fights is well understood. Successful attackers perform more raps in each bout and leave shorter pauses between the bouts than do those that give up without first evicting the defender from its shell (Briffa et al., 1998). They increase the number of raps in each bout during the final bouts of the fight, whereas losing attackers perform a decreasing number of raps in the final bouts (Briffa et al., 1998). The speed of rapping within bouts decreases from the start of each bout until the bout is terminated by a pause and also decreases from bout to bout (Briffa and Elwood, 2000a). Fatigue, therefore, appears to play a key role in regulating the rate of rapping (Briffa and Elwood, 2001a), and we have suggested that defenders are able to assess the fighting ability of attackers by monitoring these changes in the vigor of rapping (Briffa and Elwood, 2000a,b,c, 2001a,b; Briffa et al., 1998; Elwood and Briffa, 2001).
"Vigor" has thus been used to denote the rate of rapping during these contests; attackers that rap with high vigor being more likely to win (Briffa et al., 1998). Total vigor, however, in this and other signaling systems should be a combination of the temporal pattern and the magnitude of the individual performances. A study where the raps were damped by applying an elastic coating to the surface of shells sug gestedthat the force of rapping also influences the likelihood of the defender being evicted (Briffa and Elwood, 2000c). Inthat study, however, the temporal pattern was altered by thistreatment, probably because the attacker attempted tocompensate for the reduced power of impact caused by the elastic coating. Thus, the power of rapping appeared to influence fight outcomes, but it was not clear whether this was due to power per se or a result of disruption to the temporal pattern of rapping.
The aim of the present study, therefore, was to investigate effects of the power of rapping without interfering with the shell fight. We directly measured the intensity of the rapping sound to provide an index of the power supplied to each individual rap. We aimed to determine (1) the role of the power of shell rapping in decisions about fight outcomes and (2) whether rapping varies during the course of the encounter in a manner consistent with the idea that fatigue regulates vigor. If fatigue regulates the power in addition to the rate, we would expect to see decreasing power as the fight progresses both within bouts and from bout to bout. It is likely that defenders would be able to assess this information, and attackers that show the least reduction in power should be the most successful. If, however, the power of rapping remains constant for the duration of the fight or increases, this would suggest that temporal vigor is traded off in favor of maintaining power. If this were the case, it would be more likely that theprimary function of rapping behavior depends on a series of consistently powerful blows rather than on a high rate.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We collected littoral specimens of the common European hermit crab Pagurus bernhardus weekly from Ballywalter, on the coast of the Ards peninsula, Northern Ireland, between October 2000 and January 2001. Crabs were held in groups of 75–100 in 60 cm x 30 cm plastic tanks, which were filled with aerated sea water at 10°C to a depth of 10 cm, and they were fed twice weekly with chopped whitebait. We used crabs in one contest only, within 1 week, and then returned them to the sea. We removed the crabs from their shells by carefully cracking the shells open in a bench vice and sexed them; we used only males in the experiments. We supplied females with new shells and returned them to the sea, thus avoiding sex differences in fighting behavior that have been noted in previous studies (Neil and Elwood, 1985
). We used only male crabs that were free from obvious parasites, loss of appendages, and recent molt. Male crabs were allocated to pairs, such that each pair contained a small crab and a large crab.
The relative weight difference (RWD), calculated by RWD= 1 – (weight of small crab/weight of large crab), was varied between the pairs to provide a range of relative weight differences. The larger crab of each pair was provided with a Littorina obtusata shell that was 50% of its preferred shell weight, and the small crab of each pair was supplied with an L.obtusata shell that was 100% adequate for the large crab. Preferred shell weights were determined by previous choice experiments (Jackson, 1988). To stage fights we introduced the two crabs of each pair to a 90-mm crystallizing dish filled to a depth of 5 cm with aerated seawater at 10°C and containing a 1-cm deep layer of cleaned sand. The dish was placed behind the one-way mirror of an observation chamber such that the observer could see the crabs, but the crabs could not see the observer.
To obtain an index of the power of rapping, we analyzed the intensity of the rapping sound. Under ideal conditions sound intensity is related to the acoustic power (energy transferred/time) by the spherical spreading law: P = 4
r2i, where P = power (W), r = distance from the source (m), and i = sound intensity (W/m2). Power, in turn, will be proportional to the force of impact received by defenders and thus the effort supplied by attackers. The accuracy of sound intensity measurements in absolute terms will have been compromised by deviations from the ideal conditions assumed by the above equation; for instance, sound intensity may have been altered by reflection of sound waves from the wall of the crystalliz ingdish, by the acoustic properties of the dish, by sand and shells, and by the plane of orientation of the shells toward thehydrophone. Within fights, however, these complicating variables would have been constant because the crabs did not move position once the fight had been initiated. Furthermore, a large sample size was used to ensure that these variables did not influence comparisons between fights. We measured the average sound intensities of the first and last four raps in each of the first and last four bouts of rapping. In addition, the total number of raps, total number of bouts, and the durations of pauses were measured from the waveform (see Figure 1).
|
Measuring sound intensity
We recorded the rapping sound onto cassette tape using a WM-D6C Sony Professional Walkman via a hydrophone suspended in the water of the test arena close to the glass side of the arena such that the hydrophone faced into the center of the arena. The sea water was not aerated during observations. The hydrophone was constructed from a crystal microphone (Maplin LB68Y) protected by a thin layer of latex. We tested the frequency response of the hydrophone by playing a range of pure tones of differing frequency (0.1–10 khz) through a loudspeaker directed at the hydrophone. These sine waves were generated using Cool Edit Pro 1.2 (running on a PC), and the response of the microphone was assessed by measuring root mean square (RMS) sound pressure using Canary 1.2.4 Bio-acoustic Workstation (run on an Apple Macintosh) (Charif et al., 1995
). The response of the hydrophone was linear between 0.5 and 7 kHz. During thefight, a written record was made of each bout, and the outcome of the contest recorded. We removed the crabs from the arena and noted the distance from the point of impact between the two shells and the center of the hydrophone.
We analyzed the rapping sound from each fight using Canary. The RMS sound intensity of the waveform of each rap was calculated by RMS sound intensity = 
where X = the sum of values of sound pressure, and N = the number of samples, determined by the sampling rate of the computer's sound card (44050 Hz). A digital amplitude resolution of 16 bits was used to ensure accurate amplitude measurements. We used RMS values because they provide a more accurate measure of sound intensity for broad-band recordings than do peak intensity values. All sound intensity values were expressed in decibel units (dB; re 0.6469 aW/ m2). Individual raps were readily discernible from background noise levels, and for each rap average RMS sound intensity was measured for the period from the first point of deviation from background noise levels until the amplitude of the waveform returned to normal levels. See Figure 1 for an example waveform and descriptions of the measurements taken.
To calibrate the sound files analyzed with Canary, a pure-tone sine wave with a frequency of 3.25 kHz (the midpoint of the linear range of the hydrophone's frequency response) was produced using Cool Edit Pro. To set the sound intensity of the calibration tone, it was played through a loudspeaker connected to the sound card of the PC and submerged in sea water at a distance of 1 cm from the microphone of a Bruel and Kjaer sound pressure level meter (model 2231). The sea water used for calibration purposes was maintained at the same temperature (10°C), salinity (30 ppt), and oxygen content (80–85% saturated at 10°C) as that used during the recordings of shell fights. The microphone was connected to the sound pressure level meter, external to the seawater container, via a pre-amp cable (Bruel and Kjaer AO 0027), and both the loudspeaker and microphone were protected by a latex covering. The amplitude of the 3.5 kHz tone was adjusted using the software until the sound pressure level meter registered an RMS sound pressure of 50 dB. We used this signal to calibrate recordings of fights by playing it through the hydrophone at a distance of 1 cm under sea water and recording it onto tape. The tape was then played through the Canary software and used to produce a calibration document (Charif et al., 1995). The same tone was played with the same volume settings at a range of distances from the hydrophone between 2 and 9 cm at 1-cm intervals. We used each of these recordings to produce a separate calibration document, and in each case the RMS sound intensity of the calibration document was set at 50 dB (re 0.6469 aW/m2). When a fight was initiated, we noted the distance from the hydrophone and used the appropriate calibration document (to the nearest 1-cm distance) to calibrate the recording.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We staged 156 encounters that produced 75 fights where the attacker performed shell rapping. Of these, 57 resulted in the defender being evicted, and in 18 cases the defender retained its shell. To determine how the sound intensity of raps varied over the course of fights, we performed four separate repeated-measures ANOVAs. Each analysis had two within factors, bout number and rap number, and one between factor, the outcome of the encounter (defender evicted or not). Because fights comprised different numbers of bouts of rapping, the first and last four bouts of rapping were analyzed. Similarly, the bouts varied in the number of raps that they contained, and the sound intensities of the first and last four raps of these bouts were analyzed. Thus, the first analysis dealt with the first four raps in each of the first four bouts and the second with the last four raps in each of these bouts. The third and fourth analyses were of the first four raps of the last four bouts and the last four raps of the last four bouts, respectively. We could not perform a single overall ANOVA on the data set because some fights comprised fewer than eight bouts and some bouts fewer than eight raps, hence leading to the repeated inclusion of particular data points. To maintain an experiment-wide significance level of.05, sequential Bonferroni adjustments (Dunn-
idák method; see Sokal and Rohlf, 1995) were made to the significance levels associated with each of the comparisons made in these four separate tests, and adjusted probability values are quoted throughout.
First four bouts
The sound intensity of the first four raps in each of the first four bouts was greater for attackers that evicted the defender than for those that failed to evict (F1,43 = 12.4, p =.016; Figure 2a). Overall, the sound intensity decreased from bout to bout (F3,129 = 4.1, p =.016) and a nonsignificant trend for an interaction between bout number and outcome (F3,129 = 2.2, p =.096) suggests that this decrease occurred in fights in which the defender was not evicted (Figure 2a). Within bouts, there was no change in the sound intensity of the raps when all fights are examined, but an interaction effect shows that there was a marked decline in fights in which the defender was not evicted (F3,129 = 5.6, p =.012; Figure 2a). The sound intensity of the last four raps of these bouts was also greater in fights in which the defender was evicted (F1,43 = 12.6, p =.012; Figure 2b), but there was no significant variation either between (F3,129 = 2.1, p =.110) or within bouts (F3,129 = 1.86, p =.139).
|
Last four bouts
The sound intensity of the first four raps in the last four bouts was greater in fights where the defender was evicted (F1,40 = 6.8, p =.050; Figure 2c). The sound intensity decreased from bout to bout (F3,120 = 4.2, p =.012), and a significant interaction effect between bout number and outcome(F3,120= 3.9, p =.012) indicates that this decrease was greater in fights in which the defender was not evicted (Figure 2c). Within bouts the sound intensity decreased (F3,120 = 3.7, p =.012) and an interaction effect, again, shows that this decrease is much more marked in fights where the defender was not evicted (F3,120 = 3.1, p =.016; Figure 2c). The sound intensity of the last four raps of these final four bouts was also greater in fights where the defender was evicted (F1,40 = 6.9, p =.025), and there was a decrease in sound intensity within bouts (F1,137 = 3.1, p =.029; Figure 2d). There was no difference, however, in the extent of this decrease between the two outcomes (F3,120 = 1.2, p =.285), and sound intensity did not vary between bouts (F3,120 = 1.3, p =.283).
Associations between sound intensity, rate of rapping, and size of crabs
Associations between sound intensity and parameters of the rate of rapping were analyzed for each of the first two bouts of rapping. In both bouts the sound intensity of the raps was not associated with the number of raps in the bout (bout 1: r73 =.06, p >.1; bout 2: r63 = -.04, p >.1) or with the duration of the pause following the bout (bout 1: r61 = –.08, p >.1; bout 2: r56 =.11, p >.1), but in both cases there was a positive association between sound intensity and the average duration of gaps within the bout (bout 1: r73 =.25, p <.05; bout 2: r56 =.11, p <.005; Figure 3). In both of these bouts attackers that evicted the defender performed raps of greater sound intensity than did those that gave up (bout 1: t73 = 2.9, p <.005; bout 2: t63 = 3.4, p <.002; Figure 3) and left gaps of greater duration than those that gave up (bout 1: t73 = 2.0, p<.05; bout 2: t63 = 2.6, p <.02). To test that the associations between sound intensity and gap duration reported above for these bouts were not due to these differences in sound intensity and gap durations between outcomes, we performed ANCOVAs to determine the effects of gap duration (covar iate) and outcome (factor) on the average sound intensity of raps in bouts 1 and 2. In the case of bout 1, a significant interaction effect (F1,71 = 7.3, p =.009) indicated a lack of homogeneity between slopes for the two outcomes such that further analysis using ANCOVA was not possible. Calculation of the correlation coefficient between gap duration and average sound intensity separately for each outcome demonstrated a significant positive association for fights in which the attacker gave up (r16 =.57, p <.01). In the case of the second bout there was no interaction effect, but there were significant effects of outcome (F1,61 = 3.7, p <.045) and gap duration (F1,61 = 6.9, p <.011) indicating that sound intensity and gap duration are correlated after controlling for the effect of outcome.
|
To estimate the total energy expenditure in these two bouts, we multiplied the number of raps by the average sound intensity. There was no significant association between this index for the first bout and the duration of the first pause (r73 = –.16, p >.1), but the index for the second bout and theduration of the second pause were positively correlated (r56 =.28, p <.05). There was no association between sound intensity and relative weight difference (bout 1: r73 =.07, p >.1; bout 2: r63 =.05, p >.1) or the size of defenders (bout 1: r73 = 1.61, p >.1; bout 2: r63 = 1.17, p >.1). There was, however, a positive association between the weight of attackers and the average sound intensity of raps in the first bout (r73 = 0.23, p <.05).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
It is clear that the power of shell rapping by attackers is important in evicting defenders. As the fight progresses, however, the power of rapping decreases both within and between bouts, particularly in the case of unsuccessful attackers. Previous studies showed that temporal vigor also declines within bouts, with increasing gaps between raps (Briffa and Elwood, 2000a
), and between bouts with fewer raps per bout, increasing gap duration (Briffa and Elwood, 2000a) and longer pauses between bouts (Briffa et al., 1998). These findings, combined with those of increased lactate due to performing rapping (Briffa and Elwood, 2001a) and reduced vigor because of low oxygen (Briffa and Elwood, 2000b), indicate that both aspects of this activity, temporal rate and magnitude, are constrained by fatigue. Thus, only crabs that are in good condition, and presumably of high quality, are able to produce and maintain a high-power signal with a high repetition rate. Furthermore, it was these high-quality individuals that were successful in fights; victorious attackers rapped harder and showed less decline in power both within and between bouts of rapping than did those thatfailed to win. Briffa and Elwood (2000c) found that experimental damping of raps reduced the chance of victory. This treatment, however, also reduced the number of raps per bout. It was suggested that attackers were trying to increase the power but at a cost of increased fatigue, and it was not clear from that study whether power or temporal vigor was influencing rates of eviction. The present data, however, demonstrate that power may have a key role in determining fight outcomes.
Courting male wolf spiders, Hygrolycosa rubrofasiata, drum their abdomen on the ground in a manner that in many respects is similar to shell rapping (Rivero et al., 2000). Females select males that maintain a high rate and power, presumably because drumming is an honest signal of quality. Thus, although the functions of shell rapping and abdominal drumming are different, both seem to be constrained by fatigue and both seem to reflect the quality of the signaler.
In contrast to the above, however, there was a positive correlation between the duration of gaps between raps within a bout and the force of the raps. Furthermore, victorious attackers not only hit harder but left longer gaps—that is, the intra-bout temporal vigor was lower in victorious attackers. This suggests that it takes longer to produce the movement for a powerful rap possibly because it requires a greater distance between the two shells as the attacker moves them together (i.e., a longer swing). This can be explained by the size of the attacker because this distance would be affected by leg length, and large attackers hit harder. In contrast, the duration of these gaps also appears to be affected by fatigue because gap duration increases during the fight (Briffa and Elwood, 2000a) without any commensurate increase in power (power, in fact, decreases). We have shown previously that attackers that can give many raps per bout and only take short pauses (period between bouts) are likely to be victorious (Briffa et al., 1998). The apparently contradictory finding that victorious crabs leave longer gaps, however, does not imply that this aspect of vigor is unimportant. Indeed, as noted above, crabs seem to rap as rapidly as possible commensurate with sufficient power, and the rate of rapping declines only when fatigue sets in.
Significant changes in the force of rapping were mainly evident during the first four raps of each bout; change in power over the last four raps was seen in the last four bouts but not during the first four bouts (Figure 2d). During the final raps of each bout, the rate of rapping levels off after being progressively reduced during the initial raps (Briffa and Elwood, 2000a). Thus, at the start of encounters, bouts of rapping are characterized by an initial period of vigorous rapping (high rate and high force) that is reduced by fatigue until a sustainable plateau level, with reduced rate and force, is reached. Rapping continues at this level until the bout is terminated by a pause. During the final bouts, the force of rapping continues to decrease until the bout is terminated, suggesting that the accumulated effects of fatigue are greater during this phase. It was previously suggested (Briffa and Elwood, 2000a,b,c; Briffa et al., 1998) that pauses offer recovery periods between bouts of rapping, and the present data support this argument and the findings that (1) power decreases and (2) the power index (number of raps x average power) is positively correlated with pause duration support this argument. That is, the pause duration was positively correlated with an index of the expenditure of energy in the preceding bout. The total vigor of these fights is thus a complex interaction of four components: power, number of raps per bout, interbout pauses and intra-bout gaps. In other signaling systems, where the signal is performed repeatedly and with a boutlike structure, an interaction between these four components must also determine the level of the signal.
Multicomponent or multimodal signals, where multiple signaling modes are used together, have been discussed by Rowe (1998). A signal comprising two or more modalities (e.g., sound plus an olfactory cue) is assumed to be more effective than a signal comprising a single channel (Rowe, 1998). We have shown that the apparently simple signal of shell rapping, which uses a single channel, in fact comprises several interacting components that influence receivers. Thus, signals may be described as being multicomponent on two levels—first in terms of the separate modalities used and second in the detailed pattern of timing and magnitude of each individual mode. As evidenced by studies of shell rapping, interactions between components is complex even at the level of a single channel. This may make identification of the key aspect of the signal difficult. Indeed, the present finding that victorious crabs left long intra-bout gaps cannot reasonably be interpreted as suggesting that long gaps per se are required for an eviction, particularly as gap duration increases with fatigue (Briffa and Elwood, 2000a). Long gaps do, however, seem to be necessary to perform powerful raps. Thus, it seems that power is a key component of the signal as long as this is coupled with vigorous rapping in terms of a high number of raps per bout and short pauses.
In situations where the signaler must advertise its quality such as during courtship and fighting, individuals able to maintain a high rate of signaling are predicted to be suc cessful. This is borne out, for example, in studies of roaring in red deer, Cervus elaphus (Clutton-Brock and Albon, 1979) and shell fighting in hermit crabs (Briffa et al., 1998), but it has not been shown in all studies. The present study and that of Rivero et al. (2000) on courtship in wolf spiders both include data on the rate and magnitude of signaling. Both studies show that magnitude and rate contribute to the overall level of the signal and thus influence the outcome of interactions. In both of these cases production of the signal involves close proximity between sender and receiver, and power can, therefore, advertise the quality of signalers. In other cases, when there is a greater distance between senders and receivers, it may not be possible for receivers to assess quality by monitoring signal power, as this feature will degrade in proportion to the separating distance. Furthermore, the repeated impacts in hermit crab contests could have direct effects, possibly on the abdominal muscles of the defenders, so as to weaken their hold on the shell (Briffa and Elwood, 2000c). In hermit crab fights, therefore, it is less likely than in other contests that the key activity is a pure signal such that, along the lines suggested by Grafen (1990) and (Enquist, 1985), this signaling activity could have direct detrimental effects on receivers, presumably in proportion to the power of rapping. Nevertheless, the pattern of declining power demonstrated here is entirely consistent with the idea that fatigue regulates the vigor of this activity such that defenders are able to assess the stamina of attackers. Examination of the different interacting components of this activity provides greater insight to its function than would investigation of individual aspects in isolation. This should apply equally to activities performed by other animals that apparently act as signals.
|
ACKNOWLEDGEMENTS |
|---|
This work was funded by the UK Biotechnology and Biological Sciences Research Council. We thank John Prenter and two anonymous referees for their constructive comments on the manuscript.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bridge AP, Elwood RW, Dick JTA, 2000. Imperfect assessment and limited information preclude optimal strategies in male–male fights in the orb-weaving spider Metellina mengei. Proc R Soc Lond B 267:273-279.[Medline]
Briffa M, Elwood RW, 2000a. Analysis of the finescale timing of repeated signals: does shell rapping in hermit crabs signal stamina? Anim Behav 59:159-165.[CrossRef][Web of Science][Medline]
Briffa M, Elwood RW, 2000b. Cumulative or sequential assessment during hermit crab shell fights: effects of oxygen on decision rules. Proc R Soc Lond B 267:2445-2452.[Medline]
Briffa M, Elwood RW, 2000c. The power of shell rapping influences rates of eviction in hermit crabs. Behav Ecol 11:288-293.
Briffa M, Elwood RW, 2001a. Motivational change during shell fights in the hermit crab Pagurus bernhardus. Anim Behav 62:505-510.[CrossRef]
Briffa M, Elwood RW, 2001b. Decision rules, energy metabolism and vigor of hermit crab fights. Proc R Soc Lond B 268:1841-1848.[Medline]
Briffa M, Elwood RW, Dick JTA, 1998. Analysis of repeated signals during shell fights in the hermit crab Pagurus bernhardus. Proc R Soc Lond B 265:1467-1474.
Charif RA, Mitchell S, Clark CW, 1995. Canary 1.12 user's manual. Ithaca, New York: Cornell Laboratory of Ornithology.
Clutton-Brock TH, Albon SD, 1979. The roaring of red deer and the evolution of honest advertisement. Behaviour 69:145-170.[CrossRef][Web of Science]
Elwood RW, Briffa M, 2001. Information gathering and communication during agonistic encounters: a case study of hermit crabs. Adv Stud Behav 30:53-97.[CrossRef]
Enquist M, 1985. Communication during aggressive interactions with particular reference to variation in choice of behavior. Anim Behav 33:1152-1161.[CrossRef][Web of Science]
Enquist M, Leimar O, 1983. Evolution of fighting behavior: decisionrules and assessment of relative strength. J Theor Biol 102:387-410.[CrossRef]
Enquist M, Leimar O, 1987. Evolution of fighting behavior: the effect of variation in resource value. J Theor Biol 127:187-205.[CrossRef]
Enquist M, Leimar O, Ljunberg T, Mallner Y, Segerdahl N, 1990. A test of the sequential assessment game: fighting in the cichlid fish Nannacara anomala. Anim Behav 40:1-14.
Grafen A, 1990. Biological signals as handicaps. J Theor Biol 144:517-546.[CrossRef][Web of Science][Medline]
Jackson NW, 1988. Information gathering and decision making during shell selection by the hermit crab Pagurus bernhardus (PhD thesis). Belfast: Queen's University Belfast.
Mason AC, 1996. Territoriality and the function of song in the primitive acoustic insect Cyphoderris monstrosa (Orthoptera: Haglidae). Anim Behav 51:211-224.[CrossRef]
Maynard Smith J, Parker GA, 1976. The logic of asymmetric contests. Anim Behav 24:159-175.[CrossRef]
Maynard Smith J, Price GR, 1973. The logic of animal conflict. Nature 246:15-18.[CrossRef]
Neil SJ, Elwood RW, 1985. Behavioural modifications during egg-brooding in the hermit crab Pagurus bernhardus. J Exp Mar Biol Ecol 110:275-297.
Parker GA, 1974. Assessment strategy and the evolution of fighting behavior. J Theor Biol 47:223-243.[CrossRef][Web of Science][Medline]
Parker GA, Rubenstein DI, 1981. Role assessment reserve strategy and acquisition of information in asymmetric animal contests. Anim Behav 29:221-240.[CrossRef]
Payne RJH, 1998. Gradually escalating fights and displays: the cumulative assessment model. Anim Behav 56:651-662.[CrossRef][Web of Science][Medline]
Payne RJH, Pagel M, 1996. Escalation and time costs in displays of endurance. J Theor Biol 183:185-193.[CrossRef]
Payne RJH, Pagel M, 1997. Why do animals repeat displays? Anim Behav 54:109-119.[CrossRef][Web of Science][Medline]
Rivero A, Alatalo RV, Kotiaho JS, Mappes J, Parri S, 2000. Acoustic signaling in a wolf spider: can signal characteristics predict male quality? Anim Behav 60:187-194.[CrossRef][Web of Science][Medline]
Rowe C, 1998. Receiver psychology and the evolution of multicomponent signals. Anim Behav 58:921-931.
Sokal RR, Rohlf FJ, 1995. Biometry, 3rd ed. New York: W.H. Freeman.
Taylor PW, Hasson O, Clark DL, 2001. Initiation and resolution of jumping spider contests: roles for size, proximity, and early detection of rivals. Behav Ecol Sociobiol 50:403-413.[CrossRef][Web of Science]
Turner FG, Huntingford FA, 1987. A problem for game theory analysis: assessment and intention in male mouthbrooder contests. Anim Behav 34:961-970.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Briffa Signal residuals during shell fighting in hermit crabs: can costly signals be used deceptively? Behav. Ecol., May 1, 2006; 17(3): 510 - 514. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||