Anim. Behav., 1997, 53, 733–747
Assessment strategies in the contests of male crickets, Acheta domesticus (L.)
MACE A. HACK
Department of Biology, University of California San Diego
(Received 19 December 1995; initial acceptance 9 February 1996;
final acceptance 25 May 1996; MS. number: 5115)
Abstract. Game theoretical models predict that the assessment of relative fighting ability and motivation
is a process fundamental to resolving most contests. Demonstrations of assessment must (1) identify
characters associated with fighting success and (2) establish a correlation between opponent asymmetry in
these characters and the costs of fighting. Pair-wise contests between male house crickets revealed that
winners were generally heavier than their opponents, although this effect varied with the degree of
asymmetry in mass and the presence or absence of burrows. Prior burrow residency and initiating a
fighting bout provided additional, but small advantages in fighting success. Fight winners performed a
larger repertoire of agonistic tactics, more total acts, and escalated more frequently to energetically costly
tactics than did their opponents. As a result, the winner’s total energy expenditure usually exceeded the
loser’s. In accordance with the core prediction of assessment models, the cumulative energetic costs of
combat for both opponents increased with decreases in asymmetry of mass and energy expenditure rate.
These results suggest that house crickets resolve contests by assessing asymmetries in both body size and
their relative use of costly tactics. The relative energetic costs incurred by combatants may reliably signal
relative energy reserves and contribute to the active assessment of fighting ability, rather than simply
? 1997 The Association for the Study of Animal Behaviour
accrue as a by-product of combat.
The outcome of contests over mates and other
critical resources can substantially affect a male’s
reproductive success (reviewed by Huntingford &
Turner 1987). However, costs associated with
fighting favour the assessment of resource value
and relative fighting ability over indiscriminately
attacking all opponents (Parker 1974; Maynard
Smith & Parker 1976; Hammerstein & Parker
1982). Despite a strong theoretical and empirical
basis for the general role of assessment in settling
conflict (Maynard Smith 1982), how individuals
acquire and use information during contests
remains unresolved (Grafen & Johnstone 1993). A
better understanding of the assessment process
may provide new insights to traditional concerns
in the study of agonistic behaviour, such as the
function of large agonistic repertoires (Andersson
1980) and the rules governing the sequence of
behaviours during combat (Caryl 1979).
Several models have investigated the assessment process by allowing combatants to acquire
Correspondence and present address: M. A. Hack,
Department of Ecology and Evolutionary Biology,
Princeton University, Princeton, New Jersey 085441003, U.S.A. (email: [email protected]).
0003–3472/97/040733+15 $25.00/0/ar960310
information repeatedly as a contest proceeds
(Parker & Rubenstein 1981; Enquist & Leimar
1983, 1987; Leimar & Enquist 1984). In the most
developed example of these ‘sequential assessment’ games, combatants choose which of several
agonistic tactics to use per assessment round
(Enquist et al. 1990). It is assumed that tactics
differ in the type or quantity of information they
provide assessors and the costs they impose. A
contest then consists of a series of assessment
rounds that ends when one combatant has
sufficient certainty about its lesser fighting ability that further costly assessment is not justified.
In contrast to earlier contest models (e.g.
Hammerstein & Parker 1982), this model predicts that assessment is costly, and that the
sequence of tactics will determine the total costs
and information accrued (Enquist et al. 1990).
To test these ideas one needs to examine contest
costs in a currency likely to vary between tactics,
such as energy expenditure or injury risk.
To understand the assessment process one also
needs to know what information combatants need
to acquire, or which factors affect their fighting
success. These factors usually pertain to either an
? 1997 The Association for the Study of Animal Behaviour
733
734
Animal Behaviour, 53, 4
individual’s motivation or its ‘resource-holding
potential’ (RHP sensu Maynard Smith & Parker
1976, or intrinsic and extrinsic components of
fighting ability combined). Intrinsic fighting ability often derives from body size (reviewed by
Huntingford & Turner 1987), yet other factors
can be important, such as endurance (CluttonBrock & Albon 1979; Marden & Waage 1990),
speed (Garland et al. 1990) and aggressiveness
(Barlow et al. 1986). Recent fighting success or
failure can further affect intrinsic fighting ability
through apparently psychophysiological mechanisms (review by Chase et al. 1994). Extrinsic
aspects of fighting ability, such as prior ownership
of a resource, provide a strategic advantage in
some contests (Parker et al. 1974). Finally, motivation, or a resource’s value to a combatant, often
affects its fighting success (review by Enquist &
Leimar 1987). The large agonistic repertoires of
many species, containing tactics in several sensory
modalities, suggest a capacity in these species to
assess multiple RHP components, and possibly
relative motivation (Enquist 1985).
Field crickets (Orthoptera: Gryllidae) represent
excellent models for the study of contest behaviour. Frequent competition between males for
burrows, mates and other resources (Loher &
Dambach 1989) has presumably driven the evolution of a large repertoire of agonistic tactics
(Alexander 1961). However, whether male
crickets employ assessment strategies to resolve
contests remains untested. Field and laboratory
studies have identified some components of male
RHP in the gryllids, including body size (Dixon &
Cade 1986; Simmons 1986; Souroukis & Cade
1993; but see Burk 1983), sexual maturity (Dixon
& Cade 1986), burrow residency (Burk 1983;
Simmons 1986) and recent fighting history
(Alexander 1961; Burk 1983; Simmons 1986;
Adamo & Hoy 1995). Temporally variable
characters, such as physical condition, may
also determine fighting ability since dominance
relationships tend to fluctuate over several days
(Alexander 1961; Dixon & Cade 1986).
In this study I first sought to identify possible
components of RHP by comparing the morphology and behaviour of contest winners and
losers. I also monitored individual fighting success
in successive contests with the same opponent in
order to gauge temporal consistency in RHP. I
then asked which components of RHP males
actually assess during contests by examining the
costs of fighting as opponent asymmetry in these
components varies. In contrast to most prior tests
for assessment, I quantified contest cost as energy
expenditure, since this is a more accurate measure
than duration when the rate of cost accrual
depends on the type of tactic employed. Injuries
rarely occur in house cricket contests, yet opponents often pause for several minutes after
escalated combat and pump their abdomens as if
recovering from exhaustion. These observations
suggest that energy expenditure is a meaningful
currency for measuring the costs of fighting in
A. domesticus (see also Hack, in press).
METHODS
General Experimental Conditions
I used male house crickets from laboratory
stocks maintained at 23–26)C on a 12.5:11.5 h
light:dark schedule. All crickets were provided
with food (2:1 mixture of dry cat and rabbit
pellets) and water ad libitum. Prior to reaching
sexual maturity, both sexes lived together in
38-litre pails at densities of 50–100 individuals.
To preserve males in good physical condition, I
removed them several days prior to, or within 24 h
following, their imaginal moult, and housed them
singly in 0.5-litre containers.
This study entailed staging pair-wise matches
between males differing in qualities hypothesized
to affect fighting success. Variance between males
in their valuation of a contested resource was
purposely minimized to exclude motivation as a
determinant of fighting success. In total, I
observed the fighting behaviour of 111 unique
pairs (N=156 males) according to three experimental designs (Table I). Each match occurred in
one of two arena types: (1) a 5-litre plastic box
(AsymNR), or (2) a 28-litre glass tank (SymNR
and AsymR). The walls of each arena were
covered to prevent disturbance from observer
movement, and the floor of each arena was spread
with sand to simulate a natural substrate. Males
had access to food and water prior to entering the
arena, but neither was provided in the arena. In all
matches, I placed both males into the arena
simultaneously and gave them several minutes to
settle before beginning observation. Males then
interacted freely for 15 or 30 min, depending
on the experiment (AsymNR, SymNR: 30 min,
AsymR: 15 min). Each match typically contained
Hack: Assessment in cricket contests
735
Table I. Summary of experimental conditions
Character (X&) and asymmetry range
Experiment
No. of
pairs
AsymR
AsymNR
SymNR
43
38
10+20
Mass (mg)
376.5&61.5
284.7&39.9
303.6&41.6
Adult age (days)
0–0.54
0–0.47
0–0.08
27.6&17.1
11.7& 3.9
16.7& 4.6
0–1.56
0–0.46
0–0.08
Resource
Burrows
None
None
AsymR=males asymmetric in mass and age/contested resource present; AsymNR=males
asymmetric in mass and age/no contested resource present; SymNR=males symmetric in
mass and age/no contested resource present. For SymNR matches, opponents in the initial
10 pairs were re-paired with new opponents to test for the effects of prior fighting experience
on fighting success.
multiple bouts of fighting; by definition, each bout
began with physical contact between opponents
and ended with the clear retreat of one opponent
by at least two body lengths. Draws, or ambiguous outcomes to fighting bouts, rarely occurred
(less than 6% of all bouts for each experiment). I
quantified fighting success in two ways, depending
on the analysis: (1) the winner of each bout, or (2)
the winner of a majority of bouts per match
(‘overall winner’). In nearly all cases, an individual
of superior fighting ability clearly emerged within
the duration of observation (see Results).
Methodological details for tests of specific
opponent asymmetries potentially determining
fighting success are given below. Asymmetry for
continuously distributed characters was calculated
as the relative difference between opponents
(intra-pair difference/mean value for pair), and
thus scaled for differences between pairs in mean
character size.
Tests of Specific Asymmetries
Body mass and age
Two experiments, asymmetry/no resource
(AsymNR) and asymmetry/resource (AsymR),
tested the influence of mass and age asymmetries
on relative fighting success (Table I). Body mass in
A. domesticus is strongly correlated with other
measures of size, including head width (r=0.800,
N=114, P<0.001), pronotum width (r=0.725,
N=24, P<0.001) and hindleg femur length
(r=0.719, N=114, P<0.001). Mass seemed to be
the most relevant measure of size since highly
escalated fights are won by the individual best able
to lift and throw its opponent. I quantified mass to
&1 mg prior to each match, and used days since
final eclosion as a measure of adult age. Males
were always at least 5 days past eclosion to
eliminate sexual immaturity as a factor affecting
fighting success (Dixon & Cade 1986). I paired
males to produce a range of asymmetry in mass
and age (Table I). No male fought in more than
one match during the AsymNR experiment, but
20 of 56 males in the AsymR experiment fought in
two matches. These males were always paired with
new opponents and given at least 6 days between
successive matches.
Agonistic behaviour
Consistent behavioural differences between
bout winners and losers potentially reflect asymmetries in RHP characters, such as endurance or
aggressiveness. I recorded with a Canon VC-30
colour video-camera the agonistic behaviour of
opponents (see Table II for repertoire) during a
randomly chosen subset of the AsymR (N=20 of
43) matches and the first 10 SymNR matches
(excluding subsequent matches involving the same
individuals, see below). I then quantified to a
resolution of 0.1 s each individual’s sequence of
acts per fighting bout.
With repeated bouts fought between the same
opponents, it is possible that observed behavioural differences between opponents result from
differences in fighting success (i.e. dominance)
rather than cause them (Francis 1988). Consideration of only the first two bouts of each match
minimized the influence of dominance, or recent
success, on fighting behaviour. I did not use first
bouts alone in this analysis because (1) several
pairs had only a brief first encounter that was then
followed by a longer, truly agonistic interaction,
and (2) the overall match winner won the first
Animal Behaviour, 53, 4
736
Table II. Mass-specific energetic costs for the agonistic
tactics of A. domesticus
Mean O2
consumption& (ìl/g)
Low injury risk
(no physical contact)
Stridulate (SD)
Mandible flare (MF)
Shake (SK)
0.02&0.009
0.08&0.03
0.49&0.08
Moderate injury risk
(light, intermittent physical contact)
Head butt (HB)
0.37&0.03 (estimated)
Fore punch (FP)
0.37&0.03 (estimated)
Mandible spar (MS)
0.45
Antennae lash (AL)
0.68&0.14
Stridulation lash (SL)
0.70
High injury risk
(strong, sustained physical contact)
Kick (KC)
0.37&0.03 (estimated)
Head charge (HC)
0.61&0.04
Mandible lunge (ML)
0.61&0.04
Wrestle (WR)
0.83&0.15
Initiation/Retreat (per step)
0.37&0.03
Energy expenditures are per s for behavioural states
(MF, MS, WR) or per repetition for discrete behaviours
(all others). All values are net, representing the amount
of oxygen consumed above that required at rest for
metabolic maintenance (Hack, in press). For MS and
SL, standard errors are not available since energy expenditures were derived from measurements of the tactic’s
constituent movements.
bout in 80.0% of matches, but won the second
bout in 93.3% of matches, indicating the importance of the second bout to some opponents in
their assessment of the best fighter. The cumulative duration of both bouts constituted on
average the first 37.9 s of combat (=32.5,
range=2–153), and one of the two was usually the
longest and most costly of the match (25 of 30
matches). To exclude from analysis behavioural
differences between opponents that resulted from
differences in fighting success, I considered only
those acts that occurred prior to the retreat of
one combatant. I used cumulative measures
across both bouts per match in the statistical
analysis of behavioural differences to avoid
pseudoreplication.
I measured the following behavioural characters, averaged across the first two bouts of each
match, for both the overall match winner and
loser: (1) number of distinct tactics performed, (2)
maximal rate of energy expenditure as determined
by the most energetically costly tactic performed
(Table II), (3) total number of repetitions of each
tactic performed, and the total frequencies of
(4) tactical transitions, (5) escalations and (6)
de-escalations. A tactical transition occurred
when an individual performed a tactic differing
from the last tactic used by either opponent, that
is, an individual had a low frequency of tactical
transitions if it responded in kind, or not at all, to
its opponent’s behaviour rather than initiate the
use of a new tactic. Neither bout initiation nor
retreat constituted a tactical transition in the
analysis. Some tactical transitions led to changes
in the rate of cost accrual; escalations and
de-escalations represent tactical transitions to
behaviours of higher and lower energetic cost,
respectively. For three tactics, head butt, fore
punch and kick, energetic costs could not be
measured. A conservative estimated cost, equivalent to that for a single step (Table II), was instead
used for each of these tactics in the determination
of escalations and de-escalations.
The relative energetic costs of tactics (Table II)
are taken from a prior study in which I used
flow-through respirometry to measure the net
oxygen consumed per repetition of each agonistic
tactic (Hack, in press). I used these per tactic costs
in combination with the quantified frequencies of
each tactic, to calculate the cumulative oxygen
consumption during combat for each opponent.
The three tactics whose costs I could not measure
(see above) were excluded from the calculation of
cumulative cost. The remaining tactics used in the
calculations accounted for approximately 90% of
the actual oxygen consumption of both opponents
(Hack, in press). Calculated oxygen consumptions
represent mass-specific, net (consumed in excess
of the minimum required for metabolic maintenance) values at STPD (0)C, 760 torr, dry
conditions).
Burrow residency and bout initiation
Asymmetry between opponents in two characters, burrow residency and bout initiation, varied
from bout to bout during a single match, and
possibly represented extrinsic components of
RHP. To examine the effect of residency on
fighting success, the arena for all AsymR matches
contained two identical burrows made from
semi-tubular pieces of cardboard and placed at
Hack: Assessment in cricket contests
opposite ends of the tank. Male A. domesticus
readily defend burrows of this type and use them
as calling sites. Since, in general, successful fighters are also likely to become burrow residents, the
provision of two burrows ensured that the less
successful male was observed as a burrow resident
in some bouts. This allowed me to test residency
independently of other RHP components. To test
the effect of bout initiation on fighting success,
I surveyed bouts in all three experiments and
noted which opponent initiated contact to begin a
fighting bout.
737
Table III. The relationship between fighting success and
relative body size for matches with and without contested burrows (AsymR and AsymNR experiments,
respectively)
No. of matches
won by
Asymmetry
Heavier
male (%)
Lighter
male
G
AsymR
<0.10
>0.10
8 (47.0)
20 (80.0)
9
5
0.06
9.64**
AsymNR
<0.10
>0.10
7 (41.2)
14 (66.7)
10
7
0.53
2.38
Consistency of Relative Fighting Success
I assessed the temporal stability of relative
fighting success over two intervals. Continuous
observation of pairs over 15–30 min provided a
short-term assay of consistency in fighting success.
Additionally, in the AsymNR experiment, I examined how frequently the previous overall winner
was also the overall winner in a subsequent match
with the same opponent after 24 or 96 h of
isolation (N=19 pairs for each treatment).
A further experiment (SymNR) tested whether
a male’s relative fighting success remained consistent against a novel opponent. Pairs of males,
matched in body mass and adult age (less than
0.08 relative difference, Table I), interacted for
30 min in an arena without burrows and established a clear difference in relative fighting success.
I then paired each of these males with new opponents (also mass and age matched, but without
prior fighting experience for several days) and
allowed them to interact for a further 30 min.
Prior winners and losers alternated in fighting
either immediately or 30 min after their first
match. I conducted 10 complete sets of matches
(30 matches total) with 24 males. Fifteen males
participated in more than one match set, but each
was always paired with a new opponent and
isolated for several days between matches.
DETERMINANTS OF FIGHTING
SUCCESS
Results
Body mass and age
Heavier male house crickets generally defeated
their opponents (61.3% of 80 AsymR and
AsymNR matches combined, one AsymR match
excluded for lack of asymmetry in mass, log-
G=log-likelihood ratio test: **P<0.01.
likelihood ratio G1 =4.08, P<0.05). However, the
correlation between relative body mass and fighting success varied according to the degree of
asymmetry in mass between opponents and the
presence (AsymR) or absence (AsymNR) of contested burrows. In both experimental contexts, the
fighting success advantage of heavier opponents
was observed only for asymmetries in mass
greater than 0.10 (Table III). Heavier opponents
actually suffered a slight disadvantage in fighting
success at levels of asymmetry less than 0.10. For
both levels of asymmetry, the heavier opponent’s
advantage in fighting success tended to be larger
in matches fought over burrows than in matches
with no burrows present (Table III).
If relative mass affects a male’s ability to overpower and throw its opponent, we might expect
matches that escalate to wrestling to be won each
time by the heavier opponent. This was not
observed. Indeed, relative body size tended to be a
better predictor of the winner in matches that
never escalated to wrestling than in those that did
(77.8% of 27 matches without versus 50.0% of 22
matches with wrestling, Fisher’s exact P=0.07,
three matches and all AsymNR matches excluded
owing to lack of act sequence data, one match
excluded for lack of asymmetry in mass).
A total of 35 matches, pooled across both
AsymNR and AsymR experiments, involved an
asymmetry in age between opponents. Younger
opponents won 45.7% of these matches (G1 =0.26,
P>0.75). However, concurrent asymmetry in mass
between opponents potentially confounds detecting a strong effect of relative age on fighting
ability. Restricting the analysis to only those
Animal Behaviour, 53, 4
738
Table IV. Summary of behavioural differences between the overall match winner and loser
observed during the initial fighting bouts of each match
Mean
Behavioural
character
No. of tactics
Maximum rate of energy
expenditure (ìl O2/g)
Total acts
Tactical transitions
Escalations
De-escalations
Winner
Loser
4.1
3.1
0.526
67.3
12.2
6.1
6.2
0.419
59.2
10.0
4.0
5.3
Mean
difference
&
Paired t29
&0.2
4.86****
1.0
0.107&0.03
8.1 &3.5
2.2 &1.6
2.1 &0.6
0.9 &0.9
4.00***
2.29*
1.42
3.78***
1.6
Data from asymmetry/resource (N=20) and symmetry/no resource (N=10) matches were
pooled after repeated measures ANOVAs revealed no direct effects of experiment nor
experiment * overall status interactions for any behavioural character comparisons. Data
were ln-transformed for the ANOVA when required to reduce heteroscedasticity among
factor cells. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.
matches involving an asymmetry in mass less than
0.10 revealed that younger opponents won 68.8%
of 16 matches, although this trend was not
statisically significant (G1 =2.31, P<0.15).
Behavioural correlates of fighting success
In the initial stage of each match (first two
bouts), the overall match winner performed an
average of one more tactic per bout and nearly
14% more agonistic acts than the overall match
loser (Table IV). In addition, overall winners
escalated, or switched to more energetically costly
tactics, at a greater frequency than overall losers
(Table IV). As a consequence, overall winners
often reached a higher maximal rate of energy
expenditure (Table IV). Frequencies for all
tactical transitions and those constituting deescalations to energetically less costly tactics did
not differ significantly between overall winners
and losers (Table IV). Experimental context
(AsymR and SymNR) did not affect these
differences between opponents (Table IV).
Consistent behavioural differences between
overall winners and losers early in the match
ultimately resulted in greater oxygen consumptions, or energy expenditures, by winners in both
experiments (Table V). Pooling across experiments, winners expended on average 2.3 ìl O2/g
more than losers in the first two bouts of combat
(range= "7.3–17.2), a more than 12% increase
over the mean net oxygen consumption of match
losers. This difference between opponents tended
Table V. Calculated total net oxygen consumptions
(ìl/g) of overall winners and losers over the initial two
bouts of each match, for both asymmetry/resource
(AsymR) and symmetry/no resource (SymNR) matches
Mean&
Maximum
AsymR (N=20)
Winner
Loser
25.806&5.426
22.369&5.109
70.780
67.230
SymNR (N=10)
Winner
Loser
10.936&4.426
10.848&5.363
44.010
36.680
Repeated measures ANOVA: overall status F1,28 =7.31,
P=0.01; experiment F1,28 =2.58, P>0.10; experiment *
overall status F1,28 =3.38, P=0.07.
to be even larger in matches involving burrow
competition (Table V).
Asymmetries in energy expenditure and mass
were not correlated (r=0.03, N=30, P=0.9), suggesting the independent influence of each on fight
outcome. The heavier combatant or the one
expending greater energy was the overall winner
in 21 and 22 of 30 matches, respectively; in only
three matches did neither win the match. In the
most escalated bouts, or those with wrestling,
fight outcome tended to be more closely associated with relative energy expenditure than with
relative mass (energy: 66.7% of 18 matches; mass:
50.0% of 22 matches).
Bout initiation
In all three experiments, initiators were more
Hack: Assessment in cricket contests
85***
1.0
Proportion of bouts won
by the initiator
Not escalated
Escalated
0.9
0.8
20**
29**
452***
0.7
110*
0.6
739
an identical change in its opponent, only one
combatant’s fighting success can be considered an
independent datum; overall match losers were
arbitrarily chosen to represent each opponent pair
in the analysis. On average, initiation increased
the percentage of bouts individuals won by 15%
and 35% in AsymR and AsymNR matches,
respectively (Wilcoxon signed-ranks test on the
percentage of bouts won as an initiator versus as a
responder, AsymR: T=3, N=7 comparisons with
sufficient data, P<0.07; AsymNR: T=26.5, N=31,
P<0.001; SymNR: insufficient data to test).
123
0.5
(P = 0.01)
AsymR
(P < 0.0001)
AsymNR
(P = 0.12)
SymNR
Experiment
Figure 1. The proportion of fighting bouts won by
initiators within each experiment: AsymR=males asymmetric in mass and age/contested resource present;
AsymNR=males asymmetric in mass and age/no contested resource present; SymNR=males symmetric in
mass and age/no contested resource present. The
number of bouts within each analysis category is given
above each column along with the statistical significance
of the initiator’s advantage (log-likelihood ratio test:
*P<0.05, **P<0.01, ***P<0.001). The probability values beneath paired analysis categories are from Fisher’s
exact tests of the decrease in initiator advantage with
escalation. Note that the proportion won axis begins at
0.5, or the expected value if initiation has no effect on
fighting success.
likely to win a fighting bout than responders,
particularly if the bout did not escalate beyond the
antennal contact defining the start of each bout
(Fig. 1). Excluding non-escalated bouts from the
analysis eliminated the initiator’s advantage in
AsymR bouts and reduced it substantially in
AsymNR and SymNR bouts (Fig. 1). In a large
percentage of bouts won by the initiator, the
initiator had also won the preceding bout
(AsymR: 84.3%, AsymNR: 66.2%, SymNR:
93.2%). Thus, much of the initiator’s apparent
advantage can be attributed to factors determining
its prior fighting success.
To control for other intrinsic determinants of
success and test for an advantage specific to
initiation I compared the fighting success of combatants as initiators with their success as responders. Since the proportion of fights won by each
combatant in a pair must sum to 1.00, and change
in one’s success as a result of initiation mirrors
Burrow residency
Of the 374 bouts observed in AsymR matches,
48.1% occurred within, on top of, or less than one
body length from a burrow and directly determined which individual resided in it. The current
occupant of the burrow when the fight began won
87.8% of these bouts (G1 =115.85, P<0.0001),
suggesting a strong advantage of residency to
fighting success. However, to examine the effect of
residency independently of other fighting success
determinants I asked whether an individual’s success as a burrow resident differed relative to its
success in bouts fought during the same match in
the open arena, away from a burrow. Residency
generally increased individual fighting success,
relative to that in the open arena, by 13–14%
(Wilcoxon signed-ranks test on the percentage
of bouts won in each context: T=14, N=17
comparisons with sufficient data, P<0.01).
Consistency of individual fighting success
The overall winner in AsymR and SymNR
matches generally won every bout fought or
lost only single, isolated bouts (mean % bouts
won/match&: AsymR=88.3&2.2%, N=43;
SymNR=91.3&3.8%, N=10). AsymNR match
winners were more likely to lose several consecutive bouts, although they still won 73.2% of bouts
per match on average (=2.5%, N=38). Thus,
experimental context caused the overall winner’s
consistency of fighting success to vary (ANOVA
on arcsine square-root transformed % bouts won/
match: F2,88 =12.77, P<0.0001). Neither asymmetry in mass nor behavioural asymmetry between opponents correlated with the percentage of
bouts won by the overall winner, in any context.
Fighting success over consecutive matches with
740
Animal Behaviour, 53, 4
the same opponent, either 24 or 96 h apart, also
remained consistent for most pairs (AsymNR).
For 78.4% of 37 pairs, the same individual won
a majority of bouts in consecutive matches
(G1 =12.66, P<0.001, one pair did not survive the
between-match interval). Asymmetry in mass
between opponents did not affect the probability
of the same individual winning both matches, nor
did the duration of the between-match interval.
However, overall winners that won only a small
majority of bouts in their first match were more
likely to lose the subsequent match than winners
that won a large majority of bouts in their first
match (median % bouts won=58.2% and 80.0%
for reversal and no reversal in relative success,
respectively; Mann–Whitney U=178.5, N1 =8,
N2 =29, P=0.02). Thus, uncertainty in a pair’s
dominance relationship within a match was also
reflected in uncertainty of dominance across
consecutive matches.
A male’s fighting success against a novel opponent of similar mass and age, encountered
within 1 or 30 min of a prior match, varied
depending on its previous fighting success. In all
10 SymNR matches, the overall loser from the
initial match also lost the subsequent match. In
contrast, the overall winner of the initial match
won only four of 10 matches against a new
opponent. Recently winning and losing a series of
fighting bouts had clearly different effects on
subsequent fighting success (Fisher’s exact
P=0.04). Overall winners tended to lose against a
new opponent if they fought immediately after the
initial match (80%), whereas they lost less often if
given a 30-min interval between matches (40%).
Discussion of Fighting Success
The resource-holding potential of male house
crickets in this study derived from both intrinsic
male attributes and extrinsic, contextual factors.
Relative mass, or body size, generally correlated
with fighting success, suggesting its primary contribution to fighting ability. This is not a surprising result given the importance of size to fighting
ability in other species (e.g. Sigurjónsdóttir &
Parker 1981; Leimar et al. 1991), including
crickets (Dixon & Cade 1986; Simmons 1986;
Souroukis & Cade 1993). More interesting are my
observations that relative size was not the only
intrinsic determinant of male fighting success,
nor was it a consistently strong determinant in
all experimental contexts. Clearly dominant
individuals emerged in matches with no or small
asymmetry in mass (<0.10), indicating a substantial difference between opponents in an alternative
component of RHP. The relative contribution
of body size to fighting ability also varied with
the presence and absence of burrows (see also
Simmons 1986), suggesting the greater influence
of alternative RHP components in some contexts
(e.g. Barlow et al. 1986). Relative size may be the
primary determinant of success when attacking or
defending the confined space of a burrow, but not
when fighting an opponent in an open area where
all sides are vulnerable to attack.
Consistent behavioural differences between
winners and losers reflect the influence of a second
determinant of success in house cricket contests.
This determinant probably constitutes an intrinsic
component of RHP since differences between
combatants in motivation, or resource value, were
controlled for in the experimental design. The
greater energy expenditures by winners, resulting
from their more frequent use of costly tactics,
suggest that winners may have had larger energy
reserves (e.g. Marden & Waage 1990; Marden
& Rollins 1994) or a better ability to recover
expended energy (e.g. Vehrencamp et al. 1989).
This inference relies on the accuracy with which
calculated total energy expenditures, based on
measured costs per tactic and tactic frequencies,
represent the true energy expenditures of combatants. Although variance in male energy
expenditure within a tactic is generally smaller
than variance in expenditure across tactics (Hack,
in press), it is possible that losers consistently
expended more energy per repetition of each tactic
than winners. This may have resulted in equivalent
energy expenditures by winners and losers. However, winners still outperformed losers for this
expenditure, and appeared to win fights as a result.
Measurements of energy reserves before and after
fighting are required to test these hypotheses.
The more escalated behaviour of bout winners
may alternatively reflect a difference between
opponents in aggressiveness or tendency to attack
(e.g. Barlow et al. 1986). A combatant that escalates to tactics entailing increasingly hard physical
contact may be demonstrating a greater ability to
inflict injury on its opponent. However, the rarity
of serious injury in even the most escalated house
cricket contests argues against this hypothesis.
Despite the care taken to exclude from analysis
Hack: Assessment in cricket contests
any acts made by the winner as its opponent
retreated, behavioural asymmetry between opponents may still have resulted from, rather than
caused, differences in fighting success. After
repeated assessments of an opponent’s RHP, the
eventual bout winner may realize its advantage
and alter its strategy from assessment to forcing
its opponent into retreat. In doing so, it may
shorten the contest and reduce its time costs from
fighting, or it may gain a benefit in subsequent
competition with the same opponent (i.e. punishment: Clutton-Brock & Parker 1995). Statusdependent differences in the frequencies of some
tactics during combat clearly occur in this species
(Hack 1994) and other crickets (Adamo & Hoy
1995) once a dominance relationship is formed.
However, most of these tactics do not entail
physical contact (e.g. stridulate, shake) and are
therefore unlikely to force an opponent’s retreat
through injury. Preliminary analysis of the initial
bouts’ act sequences in this study indicates that
stridulation, shaking and antennae lashing
accounted for most of the asymmetry in behaviour between winners and losers, but more
extensive analysis is needed to investigate the
coercion and punishment hypotheses further.
Even if either were true, differences between combatants in agonistic behaviour would still reflect
asymmetry in an intrinsic component of RHP
unrelated to body size. Whether male crickets
actually assess asymmetries in tactical behaviour
will be addressed in the next section.
The consistency of relative fighting success over
intervals of several days implies that fighting
success primarily derived from intrinsic, temporally stable characters. However, two extrinsic
factors, burrow residency and bout initiation,
affected the outcome of individual fighting bouts
and represent secondary, contextual components
of RHP. The greater fighting success of residents
presumably derived from a strategic or positional
advantage in fighting ability (e.g. Parker et al.
1974), rather than an asymmetry in resource
value, since a male’s fighting success improved
within seconds of its occupying a burrow. The
small inherent advantage in fighting success
gained by initiators may have resulted from a
small benefit for winning relative to a large cost
for engaging an opponent in combat. In this case,
initiation could be an arbitrary asymmetry for
contest settlement (Hammerstein 1981). Alternatively, initiation may correlate weakly with fight-
741
ing ability or motivation, but the costs of testing
its reliability as a signal of these qualities is not
always justified by the expected benefit (Dawkins
& Guilford 1991). For example, immediately after
losing an escalated fight, an individual may retreat
more readily from an opponent that initiates
combat since the opponent is likely to be the
recent victor with an already known, relatively
superior RHP.
Recent fighting experience constitutes an
additional component of RHP for house crickets.
Recent losers appeared less likely than expected to
defeat a new opponent of similar mass. This effect
of losing presumably stemmed from a psychophysiological mechanism since losers were not
visibly injured or incapacitated by the prior
combat. Similarly, four of 10 initial matches
escalated to wrestling, implying that losers were
not generally in poor condition or unable to fight
at the beginning of the experiment. Decreased
fighting ability as a result of prior losses probably accounts for the large majority of bouts,
irrespective of mass or energy expenditure asymmetries, won by the overall match winner. Thus,
experience effects may reinforce and stabilize
dominance relationships in house crickets and
other species (Landau 1951). One implication of
an experiential component of RHP is that the
initial winner of a series of interactions will continue to dominate its opponent, even if it has the
lower intrinsic fighting ability and simply won the
initial bout by luck. However, the non-arbitrary
outcome of most matches in this study argues
against this and suggests that initial bouts were
settled through the accurate assessment of other
RHP components, such as relative body size or
endurance.
ASSESSMENT STRATEGIES
The core prediction of models that allow repeated
assessment is that decreased asymmetry between
opponents in RHP, assuming equal motivation to
win, results in more costly contests (Parker &
Rubenstein 1981; Enquist & Leimar 1983). This is
because closely matched opponents require more
information about relative RHP to distinguish
their asymmetry accurately. Thus, tests for assessment must focus on changes in the costs of
fighting as opponent asymmetry in RHP varies
(e.g. Sigurjónsdóttir & Parker 1981; Leimar et al.
1991).
Cumulative energetic cost of
combat per match (µl oxygen/g)
180
160
140
120
100
80
60
40
20
0
–0.1
0.0
0.1
0.2
0.3
Mass asymmetry
0.4
0.5
Figure 2. The cumulative net oxygen consumed during
combat by both opponents per match as a function of
the asymmetry between them in body mass (winner’s"
loser’s mass/mean mass). Data are from asymmetry/
resource matches (N=18). Two matches with strongly
negative asymmetry values (< "0.10, i.e. the lighter
opponent wins) were excluded from the analysis since
asymmetries other than mass probably determined fighting success in these cases. Small negative asymmetries
(> "0.10) have been included in the analysis since
relative mass did not correlate with success in the range
"0.10 to 0.10. Asymmetries in this range would presumably require similar costs to assess regardless of
which opponent won. Regression equation: Y= "168.2
X+78.7.
Results
To test for the use of assessment strategies by
male A. domesticus, I first calculated a pair’s
cumulative net oxygen consumption, or combined
energy expenditure per match (see Methods).
Total energy expenditure is likely to reflect accurately the fitness costs of fighting since it integrates
three important cost currencies: energy, time and
injury risk (i.e. force of movement, Table II). I
then regressed cumulative oxgyen consumption
against (1) relative mass difference, and (2) relative difference in initial oxygen consumption (first
two bouts of each match). Asymmetry in oxygen
consumption simultaneously accounted for all the
observed differences between opponents in tactic
use (see Table IV).
In AsymR matches, cumulative net oxygen consumption during combat decreased significantly
with increasing asymmetry in mass (Fig. 2:
r2 =0.27, F1,16 =5.85, P=0.03). Similarly, increasing asymmetry in opponents’ oxygen consumptions during the first two bouts of each match was
associated with a decrease in the cumulative
Cumulative energetic cost of combat per match
(µl oxygen/g)
Animal Behaviour, 53, 4
742
200
180
160
140
120
100
80
60
40
20
0
60
AsymR
0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25
SymNR
50
40
30
20
10
0
0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25
Asymmetry in rate of energy expenditure
Figure 3. The cumulative net oxygen consumed during
combat by both opponents per match as a function of
the asymmetry between them in their initial rates of
oxygen consumption (winner’s"loser’s rate/mean rate,
first two bouts of each match only). Results are presented for both asymmetry/resource (AsymR) and
symmetry/no resource (SymNR) matches. Negative
asymmetry values were excluded from the analysis (N=5
and 2 for AsymR and SymNR, respectively) since alternative asymmetries presumably determined fighting success in these cases. (Inclusion of these data do not change
the direction or decrease the statistical significance of the
observed relationships.) For the analysis of AsymR
matches, the discontinuous variation in asymmetry in
energy expense and heteroscedasticity in cumulative
costs precluded a linear regression analysis (see Results).
Regression equation for SymNR: Y= "21.4X+41.0.
energetic cost of fighting per match (Fig. 3). In
other words, the more one combatant’s initial rate
of energy expenditure exceeded his opponent’s,
the lower was the overall energy expenditure of
both combatants over a complete match. This
result was obtained in both AsymR (linear regression precluded by heteroscedasticity; median O2
consumption for asymmetry <0.75 was 81.33 ìl/g
versus 5.88 ìl/g for asymmetry >0.75, Mann–
Whitney U=39, N1 =11, N2 =4, P<0.05) and
SymNR (r2 =0.61, F1,6 =9.20, P=0.02) matches.
Hack: Assessment in cricket contests
Discussion of Assessment Strategies
The preceding section of this paper demonstrated a correlation between fighting success and
several potential components of RHP. The results
presented in this section indicate that asymmetries
in two of these components, mass and energy
expenditure, or their close correlates, are also
assessed by combatants to resolve conflict in a
cost-effective manner. Moreover, the high costs of
140
(a)
120
100
Total energetic cost per bout (µl oxygen/g)
Energy expenditure asymmetry was smaller
in matches that escalated to wrestling, or the
maximum cost level, than in matches that did
not (Mann–Whitney U=180, N1 =18, N2 =12,
P<0.01). In contrast, small asymmetry in mass
was not more likely to lead to wrestling (U=311,
N1 =23, N2 =27, P>0.75). Asymmetry in energy
expenditure also varied inversely with the total
duration of wrestling per match (Spearman rank
correlation: rS = "0.62, N=30, P<0.001).
If asymmetries in both mass and rate of energy
expenditure represent independently assessed
components of fighting ability, assessment should
be more costly when each favours a different
opponent to win. For matches involving both
types of asymmetry, cumulative oxygen consumptions per match were greater when each asymmetry favoured a different opponent than when
both asymmetries favoured the eventual winner
(t13 =2.68, P=0.01). In an analogous conflict of
asymmetries, bouts over burrows resulting in the
eviction of the prior resident had a median cumulative energetic cost over 17 times greater than
that for matches resulting in the prior resident’s
retention of the burrow (Fig. 4a: Mann–Whitney
U=48, N1 =5, N2 =11, P<0.05). This indicates
that fighting has high costs when residency
provides an advantage to an otherwise inferior
opponent.
Comparison of fighting bouts over burrows
with those occurring in the open arena reveals
whether opponents assessed resource value during
fights. Assuming that winning control of a burrow
provides a greater fitness benefit to males than
defeating an opponent when no tangible resource
is contested, fights over burrows should be more
costly than those in the open arena. Indeed, the
median cumulative energetic cost of bouts over
burrows was nearly twice that of bouts in the
open arena (Fig. 4b: Mann–Whitney U=3734.5,
N1 =71, N2 =86, P<0.05).
743
80
60
40
20
0
35
Resident stays Resident evicted
Bout outcome
(b)
30
25
20
15
10
5
0
Burrow
Open arena
Bout location
Figure 4. The energetic cost of fighting for bouts (a) that
did (N=5) or did not (N=11) result in the eviction of the
prior resident and (b) that determined burrow ownership
(N=71) compared with bouts that occurred in the open
arena (N=86). For (a) only the first two bouts of each
match were included in the analysis in order to control
for the effects of recent fighting experience on subsequent behaviour and fighting success. Energy costs per
bout are net values of oxygen consumption summed
across opponents. Horizontal lines delineate the 10th,
25th, 50th (median), 75th and 90th percentiles in each
distribution.
fighting when these asymmetries contradict, or
favour different opponents to win, indicates that
each independently influences male fighting
strategy. Burrow occupation and the expected
benefit of winning may also be assessed by fighting
crickets and affect fighting strategy. Thus, the
decision to continue fighting or retreat is based on
a complex integration of several types of information, as suggested by theory (Parker 1974;
Hammerstein & Parker 1982; Enquist et al. 1990).
The tactics performed by crickets during combat
presumably provide this information, yet how are
different asymmetries assessed?
744
Animal Behaviour, 53, 4
Crickets may assess relative body size through
visual, tactile or acoustic inspection of an opponent. The high correlation of body length and
head size with mass, or overall size, may allow
accurate visual or tactile assessment without
entailing hard physical contact. Males often faced
each other at close range to butt their heads
together briefly and spread their mandibles, perhaps in a comparison of head size or the force
behind each butt. The acoustic structure (frequency and temporal characteristics) of advertisement stridulation provides body size information
in a related cricket (Simmons 1988), suggesting
the possible encoding of size information in the
structurally similar agonistic call of A. domesticus.
Thus, several tactics may function to communicate body size differences directly. Tactics of this
type, or ‘assessment signals’ (sensu Maynard
Smith & Harper 1988) provide reliable information without necessarily incurring costs, since
their performance simply reveals an unbluffable
quality such as body size. The fact that many of
these tactics do impose energetic costs on assessors indicates that even direct assessment is a
costly process, and is therefore likely to involve
the trade-off between accuracy and cost predicted
by theory (Parker & Rubenstein 1981; Enquist &
Leimar 1983).
An important result of this study is that male
house crickets also assess relative RHP with conventional signals (sensu Maynard Smith & Harper
1988; Guilford & Dawkins 1995), or by using
tactics that are informative because of the costs
they incur. Males appear to monitor which tactics
an opponent uses and how frequently it uses each,
which collectively determine its rate of energy
expenditure. The eventual winner’s more energetically costly behaviour may reliably signal its
greater capacity to incur the energetic costs of
further, more escalated combat (Grafen 1990; see
also General Discussion). In this study, most
fights culminated in wrestling and, since wrestling
is an obligately bilateral behaviour, its duration
was identical for each opponent. Therefore, asymmetry in tactic use, or energy expenditure, must
have developed prior to escalation to wrestling.
Opponents that did not perceive sufficient asymmetry in their relative energy expenditures on
tactics such as antennae lashing, shaking and
stridulation lashing may have escalated to
wrestling to determine the better fighter. Three
results support the hypothesis that pre-wrestling
behaviour reliably indicated a male’s ability to
outlast its opponent in the most costly phase of
combat: (1) more frequent escalation to wrestling
when energy expenditure asymmetry was small;
(2) a negative correlation between wrestling duration and energy expenditure asymmetry; and (3)
the superior fighting success of the male with the
greater energy expenditure in bouts that escalated
to wrestling.
The argument presented above could also be
applied to the injury risk a combatant willingly
assumes since house cricket tactics varied in the
degree of physical contact, and possible injury
risk, they entailed. An individual’s escalation to a
more risky tactic may reliably signal a greater
capacity to withstand its opponent’s use of the
same tactic. The tactic may or may not impose a
direct cost to be reliable, but it must commit the
signaller to the more costly (i.e. risky) stage of
combat if its opponent similarly escalates (Enquist
1985). Although differences in injury risk may
account for the observed asymmetry in escalation
frequency and repertoire size, they are less likely
to account for asymmetry in the repetition rate of
specific tactics. Also, in over 3000 fighting bouts,
injuries never directly resulted from combat, and
damage possibly attributable to fighting was typically minor (loss of parts of the antennae, cerci
and legs). Nevertheless, relative injury risk may
account for differences in the use of some tactics,
particularly those involving different degrees of
physical contact but incurring similar energetic
costs. The general correlation across tactics
between the degree of physical contact entailed
and energy expenditure (Table II) suggests that,
even if injury risk varies significantly between
A. domesticus tactics, conclusions drawn from
broad patterns of energy expenditure will apply
regardless of which cost currency is the most
relevant.
GENERAL DISCUSSION
In this study I have tried to evaluate strategic
behaviour in contests by taking into account the
economics of alternative agonistic tactics. This
approach, advocated by recent theory (Parker &
Rubenstein 1981; Enquist et al. 1990), is necessary
to gain an understanding of how individuals
acquire and use information to resolve conflict.
The relatively high energetic costs of several
Hack: Assessment in cricket contests
cricket tactics have the consequence that rates of
cost accrual are not just high during the final,
most escalated phase of combat (e.g. wrestling).
Assessment for male house crickets is a costly
process, as predicted by sequential assessment
models (Parker & Rubenstein 1981; Enquist et al.
1990), rather than a cost-free phase prior to choosing a fighting strategy (Hammerstein & Parker
1982). An important result to emerge from this
study is that a combatant’s rate of cost accrual, or
the tactical sequence it performs, functions as a
signal. This indicates that the accrual of energetic
costs is more than a by-product of the assessment
process but a parameter controlled by combatants
to reveal their relative competitive abilities.
Although my results (but see Hack 1994) do not
directly test the sequential assessment model
(Enquist et al. 1990), they provide support for one
of its qualities that differs from previous contest
models (Hammerstein & Parker 1982). The intrinsic fighting ability of male A. domesticus appears
to derive from two components, body size and
endurance, and each is independently assessed.
Multiple components of fighting ability suggest
the need for multiple agonistic tactics, each
specialized to reveal a particular component
(Enquist et al. 1990).
The observation that winners of house cricket
contests performed certain tactics more frequently
than losers is a result consistent with the mechanism of sequential assessment but one that suggests
greater complexity to real contest behaviour
than is considered by the model. In house cricket
fights, only two of 12 agonistic tactics require
both opponents to perform the same act simultaneously. Thus, males often perform different
tactics during combat, and consequently accrue
different costs and, presumably, acquire different
information. Since the sequential assessment
model assumes opponents perform the same tactic
per assessment round, it is not clear whether it can
provide an evolutionarily stable strategy when
opponents differ substantially in their tactic use,
as observed in this study. This is a general problem since, for example, contest winners in other
species often perform certain tactics more frequently than their opponents (Riechert 1978;
Turner & Huntingford 1986; Franck & Ribowski
1989), demonstrating a similar independence in
their choice of tactics.
Constraining opponents to perform the same
tactic per assessment round also precludes two
745
hypothesized functions for agonistic tactics. First,
a tactic may provide information to the actor by
causing its opponent to respond in a particular
manner. For example, a rapid lunge may force
an opponent to step back revealing its agility or
speed. Different propensities to probe an opponent may account for some of the observed difference in tactic use between cricket opponents, but
it is not clear why winners may have consistently
probed at higher frequencies than losers. Second,
an individual cannot communicate through its
choice of tactic (Enquist 1985) if it is obligated to
perform the same tactic as its opponent (Grafen &
Johnstone 1993). This study presents evidence
that winners signalled a greater capacity to incur
the high energetic costs of wrestling by more
frequently choosing to perform other tactics of
moderate energetic cost. Signals of this type represent ‘strategic handicaps’ (sensu Zahavi 1977;
Grafen 1990) since high-quality individuals (e.g.
superior fighters) can better afford to produce
them than low-quality individuals. Unfortunately,
there are difficulties in applying the theory of
strategic handicap signalling to contests. The
model has a formal weakness in its inability to
model contexts where communication is symmetrical, or when signallers also act as receivers
(Grafen 1990), as is clearly the case during
contests.
Honest, but costly, indicators of condition (e.g.
strategic handicaps) also generate an apparent
contradiction with assessment theory. By definition, a handicap is a more costly signal given by
a higher quality individual (Grafen 1990). For
example, to signal greater energy reserves, an
individual should expend more energy than its
opponent, as suggested by my results. However,
assessment models are based on the idea that high
RHP combatants accrue costs at a slower rate
during combat than their opponents (Parker &
Rubenstein 1981; Hammerstein & Parker 1982;
Enquist & Leimar 1983). Although this appears
true for injury costs (Austad 1983; Hammerstein
& Riechert 1988), energetic costs may often violate this assumption. A solution to this apparent
contradiction is an accelerating decline in fitness
with decreasing energy reserves; the same energy
expenditure produces a larger decrease in fitness
for individuals with low reserves than for individuals with high reserves. For a small excess energy
expenditure by a high RHP individual, its ultimate
fitness cost would not generally exceed that of its
746
Animal Behaviour, 53, 4
opponent, meeting the basic assumption underlying assessment theory and the requirement of
costly, honest signalling. In house cricket matches,
the winner’s excess energy expenditure was usually
small (12%) relative to the total expended per
opponent. Fitness may generally decline at a
faster rate with decreasing energy reserves
(Wilkinson 1984; Marden & Rollins 1994),
although whether this applies to male house
crickets is not known.
As an alternative to strategic handicap signalling, the winner’s greater rate of energy expenditure may directly reveal its superior energy
reserves. Individuals with lower reserves may
simply be incapable of performing energetically
costly acts at the same rate. Although this hypothesis may account for differences in the performance of a particular tactic used by both opponents
(e.g. roaring in red deer, Cervus elaphus: CluttonBrock & Albon 1979), it seems less able to explain
differences in the tactic choices of opponents (e.g.
escalation to a higher cost level). For example, a
male house cricket able to perform two repetitions
of shaking could have instead performed a single
repetition of the more costly stridulation lash for
an equivalent total energy expenditure. It would
not be incapable of escalating because of low
energy reserves.
In conclusion, assessment is fundamental to the
settlement of contests in A. domesticus. Males
appear to gain useful information about relative
body size directly from the performance quality of
certain tactics, while the choice to perform particularly costly tactics reveals information about
endurance, or a male’s willingness to incur further
energetic costs. Thus, assessment for male house
crickets is not simply a passive process of receiving
information, but depends on the sequence of
tactics each opponent performs. Future studies of
assessment should consider the possible signalling
strategies combatants use to reveal and acquire
information during combat.
ACKNOWLEDGMENTS
I thank J. Bradbury, V. Crawford, J. Kohn, T.
Price, S. Vehrencamp and K. Williams for their
advice in the design of this study and their comments on an early version of the manuscript. K.
Congdon helped in the data collection. G. Parker,
P. Stockley, N. Wedell and M. Gage provided
many helpful comments in the writing of the
manuscript. This research was funded in part
by a National Science Foundation Predoctoral
Fellowship.
REFERENCES
Adamo, S. A. & Hoy, R. R. 1995. Agonistic behaviour
in male and female field crickets, Gryllus bimaculatus,
and how behavioural context influences its expression.
Anim. Behav., 49, 1491–1501.
Alexander, R. D. 1961. Aggressiveness, territoriality,
and sexual behaviour in field crickets (Orthoptera:
Gryllidae). Behaviour, 17, 130–223.
Andersson, M. 1980. Why are there so many threat
displays? J. theor. Biol., 86, 773–781.
Austad, S. N. 1983. A game theoretical interpreation of
male combat in the bowl and doily spider (Frontinella
pyramitela). Anim. Behav., 31, 59–73.
Barlow, G. W., Rogers, W. & Fraley, N. 1986. Do
Midas cichlids win through prowess or daring? It
depends. Behav. Ecol. Sociobiol., 19, 1–8.
Burk, T. U. 1983. Male aggression and female choice in
a field cricket (Teleogryllus oceanicus): the importance
of courtship song. In: Orthopteran Mating Systems:
Sexual Selection in a Diverse Group of Insects (Ed. by
D. T. Gwynne & G. K. Morris), pp. 97–119. Boulder,
Colorado: Westview Press.
Caryl, P. G. 1979. Communication by agonistic displays: what can games theory contribute to ethology?
Behaviour, 68, 136–169.
Chase, I. D., Bartolomeo, C. & Dugatkin, L. A. 1994.
Aggressive interactions and inter-contest interval:
how long do winners keep winning? Anim. Behav., 48,
393–400.
Clutton-Brock, T. H. & Albon, S. D. 1979. The roaring
of red deer and the evolution of honest advertisement.
Behaviour, 69, 145–169.
Clutton-Brock, T. H. & Parker, G. A. 1995. Punishment
in animal societies. Nature, Lond., 373, 209–216.
Dawkins, M. S. & Guilford, T. 1991. The corruption of
honest signalling. Anim. Behav., 41, 865–874.
Dixon, K. A. & Cade, W. H. 1986. Some factors
influencing male–male aggression in the field cricket
Gryllus integer (time of day, age, weight and sexual
maturity). Anim. Behav., 34, 340–346.
Enquist, M. 1985. Communication during aggressive
interactions with particular reference to variation in
choice of behaviour. Anim. Behav., 33, 1152–1161.
Enquist, M. & Leimar, O. 1983. Evolution of fighting
behaviour: decision rules and assessment of relative
strength. J. theor. Biol., 102, 387–410.
Enquist, M. & Leimar, O. 1987. Evolution of fighting
behaviour: the effect of variation in resource value.
J. theor. Biol., 127, 187–205.
Enquist, M., Leimar, O., Ljungberg, 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.
Francis, R. C. 1988. On the relationship between aggression and social dominance. Ethology, 78, 223–237.
Hack: Assessment in cricket contests
Franck, D. & Ribowski, A. 1989. Escalating fights for
rank-order position between male swordtails (Xiphophorus helleri): effects of prior rank-order experience
and information transfer. Behav. Ecol. Sociobiol., 24,
133–143.
Garland, T. J., Hankins, E. & Huey, R. B. 1990.
Locomotor capacity and social dominance in male
lizards. Func. Ecol., 4, 243–250.
Grafen, A. 1990. Biological signals as handicaps.
J. theor. Biol., 144, 517–546.
Grafen, A. & Johnstone, R. A. 1993. Why we need ESS
signalling theory. Proc. R. Soc. Lond. Ser. B, 340,
245–250.
Guilford, T. & Dawkins, M. S. 1995. What are conventional signals? Anim. Behav., 49, 1689–1695.
Hack, M. A. 1994. The choreography and energetics
of house cricket (Acheta domestica Linn.) contests:
patterns of assessment. Ph.D. thesis, University of
California, San Diego.
Hack, M. A. In press. The energetic costs of fighting in
the house cricket, Acheta domesticus L. Behav. Ecol.
Hammerstein, P. 1981. The role of asymmetries in
animal contests. Anim. Behav., 29, 193–205.
Hammerstein, P. & Parker, G. A. 1982. The asymmetric
war of attrition. J. theor. Biol., 96, 647–682.
Hammerstein, P. & Riechert, S. E. 1988. Payoffs and
strategies in territorial contests: ESS analyses of two
ecotypes of the spider Agelenopsis aperta. Evol. Ecol.,
2, 115–138.
Huntingford, F. & Turner, A. 1987. Animal Conflict.
New York: Chapman & Hall.
Landau, H. G. 1951. On dominance relations and the
structure of animal societies: II. Some effects of
possible social factors. Bull. Math. Biophys., 13, 245–
262.
Leimar, O. & Enquist, M. 1984. Effects of asymmetries
in owner–intruder conflicts. J. theor. Biol., 111, 475–
491.
Leimar, O., Austad, S. & Enquist, M. 1991. A test of the
sequential assessment game: fighting in the bowl and
doily spider. Evolution, 45, 862–874.
Loher, W. & Dambach, M. 1989. Reproductive behavior. In: Cricket Behavior and Neurobiology (Ed. by F.
Huber, T. E. Moore & W. Loher), pp. 83–113. Ithaca,
New York: Cornell University Press.
Marden, J. H. & Rollins, R. A. 1994. Assessment of
energy reserves by damselflies engaged in aerial contests for mating territories. Anim. Behav., 48, 1023–
1030.
747
Marden, J. H. & Waage, J. K. 1990. Escalated damselfly
territorial contests are energetic wars of attrition.
Anim. Behav., 39, 954–959.
Maynard Smith, J. 1982. Evolution and the Theory of
Games. New York: Cambridge University Press.
Maynard Smith, J. & Harper, D. G. C. 1988. The
evolution of aggression: can selection generate variability? Phil. Trans. R. Soc. Lond. Ser. B, 319, 557–570.
Maynard Smith, J. & Parker, G. A. 1976. The logic of
asymmetric contests. Anim. Behav., 24, 159–175.
Parker, G. A. 1974. Assessment strategy and the evolution of fighting behaviour. J. theor. Biol., 47, 223–243.
Parker, G. A. & Rubenstein, D. I. 1981. Role assessment, reserve strategy, and acquisition of information
in asymmetric animal conflicts. Anim. Behav., 29,
221–240.
Parker, G. A., Hayhurst, G. R. G. & Bradley, J. S. 1974.
Attack and defence strategies in reproductive interactions of Locusta migratoria and their adaptive significance. Z. Tierpsychol., 34, 1–24.
Riechert, S. E. 1978. Games spiders play: behavioural
variability in territorial disputes. Behav. Ecol. Sociobiol., 3, 135–162.
Sigurjónsdóttir, H. & Parker, G. A. 1981. Dung fly
struggles: evidence for assessment strategy. Behav.
Ecol. Sociobiol., 8, 219–230.
Simmons, L. W. 1986. Inter-male competition and
mating success in the field cricket, Gryllus bimaculatus
(De Geer). Anim. Behav., 34, 567–579.
Simmons, L. W. 1988. Male size, mating potential and
lifetime reproductive success in the field cricket, Gryllus bimaculatus (De Geer). Anim. Behav., 36, 372–379.
Souroukis, K. & Cade, W. H. 1993. Reproductive
competition and selection on male traits at varying sex
ratios in the field cricket, Gryllus pennsylvanicus.
Behaviour, 126, 45–62.
Turner, G. F. & Huntingford, F. A. 1986. A problem for
game theory analysis: assessment and intention in
male mouthbrooder contests. Anim. Behav., 34, 961–
970.
Vehrencamp, S. L., Bradbury, J. W. & Gibson, R. M.
1989. The energetic cost of display in male sage
grouse. Anim. Behav., 38, 885–896.
Wilkinson, G. 1984. Reciprocal food sharing in vampire
bats. Nature, Lond., 308, 181–184.
Zahavi, A. 1977. Reliability in communication systems
and the evolution of altruism. In: Evolutionary Ecology (Ed. by B. Stonehouse & C. Perrins), pp. 253–259.
New York: Macmillan Press.