Download Mechanisms of density-and number

Mar. Freshwater Res., 2002, 53, 175–179
Mechanisms of density- and number-dependent population regulation of
a coral-reef fish
Jeffrey S. Shima
Department of Ecology, Evolution, and Marine Biology, and the Marine Science Institute, University of California,
Santa Barbara, CA 93106, USA. Present address: School of Biological Sciences, Kirk Building, Kelburn Parade,
PO Box 600, Victoria University of Wellington, Wellington, New Zealand
Abstract. Density-dependent post-settlement losses are a common feature of many reef fish populations, and
resent observations suggest that losses may also scale with population size independent of density (i.e.
‘number-dependence’). Despite the potential importance of these two forms of compensatory loss, there have been
relatively few attempts to determine the mechanisms underlying these demographic patterns. A factorial
experiment was conducted to test whether density- and/or number-dependent losses observed for newly settled six
bar wrasse (Thalassoma hardwicke) are attributable to predation or another factor (e.g. migration). Losses of
recently settled fish from reefs within predator exclosures were ∼26% over a 7-day period and apparently
independent of density and number of residents. Losses from reefs that were accessible to predators averaged ∼62%
over 7 days, and were dependent upon both the density and number of resident fish. Behavioural observations
indicate the frequency of agonistic interactions between resident six-bar wrasse scales with the number of fish
independent of density. Overall, results attribute both density-and number-dependent losses to predation, and
suggest that number-dependent behavioural interactions (perhaps important for the social control of sex change)
rather than proximate resource limitation may underlie observed number-dependent mortality.
Extra keywords: density-dependence, number-dependence, predation, recruitment, reef fish.
MF01 3
JePforpueylatSio.nhriemugaltionofacoral-ref sih
Introduction
Local population density may influence rates of growth,
migration, and mortality to reshape patterns of larval
settlement and contribute to the dynamics of marine reef
populations (reviewed in Doherty 1991; Hixon 1991; Jones
1991). Because these density-dependent (i.e. compensatory)
demographic rates potentially regulate local populations
(Murdoch 1994; Turchin 1995), they have served as a focal
point for reef fish ecology (Booth and Brosnan 1995; Caley
et al. 1996).
Although numerous studies have documented
density-dependent mortality in reef fishes (reviewed in
Schmitt and Holbrook 1999; Shima 2001a), there have been
relatively few investigations of the mechanisms producing this
density dependence. Studies that have manipulated predator
access or prey refuges (e.g. Hixon and Beets 1989, 1993;
Steele 1999) coupled with circumstantial and correlative
evidence for post-settlement predation in reef fish systems
(reviewed in Hixon 1991) have contributed to the widely held
view that density-dependent losses arise from predation upon
individuals competing for a limited number of prey refuges.
Only in rare instances has such a mechanism been confirmed
in experimental studies (e.g. Hixon and Carr 1997).
© CSIRO 2002
Sinclair (1988) discusses the importance of separating
demographic consequences of population density from
population size, and more recent work has demonstrated that
compensatory losses of a reef fish may also result from
numerical effects independent of density (Shima 2001a).
Such ‘number-dependence’ is characterized by per capita
loss rates that scale with the absolute number of individuals
per site rather than with numbers per site area (i.e. the ratio
of individuals to resources is inconsequential). For example,
per capita losses in a group of 10 individuals occupying a
10 m2 patch are greater than a group of 2 individuals in a
2 m2 patch, despite the fact that in both cases density is
1 fish m–2. On the assumption that reef area is a reasonable
proxy for resources including refuges from predators, the
existence of number-dependent losses appears to require a
different mechanism than one where prey species compete
for limited refuge space.
The present study employs a factorial experiment to test
whether density- and/or number-dependent losses observed
for the six-bar wrasse (Thalassoma hardwicke) may be attributable to predation or another factor (e.g. migration).
Larval stages of six-bar wrasse develop in pelagic environments for ∼47 days (Victor 1986) and subsequently settle to
10.1071/MF01133
1323-1650/02/020175
176
Jeffrey S. Shima
small patch reefs on Moorea during new moons from January through June (Shima 1999a, 2001a, 2001b). Newly settled individuals experience strong density-dependent losses
within weeks (Shima 2001a) to months (Shima 1999a,
2001b) of settlement, and early post-settlement losses are
also influenced by the number of individuals resident on a
patch reef independent of density effects (‘number-dependence’, sensu Shima 2001a). This study addresses potential
mechanisms of density- and number-dependent losses experienced by newly settled six-bar wrasse.
Materials and methods
Research was conducted within the shallow lagoons of the island of
Moorea, French Polynesia (17°30′S,149°50′W), and focused on
post-settlement survival and behaviour of the six-bar wrasse
(Thalassoma hardwicke).
Density- versus number-dependent mortality and the role of predators
Decoupling effects of ‘density’ and ‘number’.
Shima (2001a) presents the details and rationale behind the factorial
experiment conducted in May 1999 to decouple variation in per capita
loss rates attributable to density (no. fish per reef area) from those
attributable to the number of residents per reef (i.e. independent of
density). This experiment followed survival over 7 days of newly
settled six-bar wrasse that were transplanted to patch reefs (depopulated
of conspecifics prior to experiment) at two typical densities (0.25 or
0.50 fish m–2). Fish densities were achieved by manipulating the
number of fish (either 1 or 2 individuals; most common range for this
species) on reefs of three typical size categories (2, 4, or 8 m2). This
design resulted in replicate fish (n = 50 fish per treatment; see Table 1)
experiencing a group size of 1 at densities of 0.25 or 0.50 fish m–2, or a
group size of 2 at densities of 0.25 or 0.50 fish m–2. Per capita loss rates
were estimated for each patch reef as the fraction of transplanted
individuals that were lost over a 7-day period, and were analysed by
logistic ANOVA to accommodate a trinomial distribution of the data
(i.e. fractional losses were 0%, 50% or 100%). The design incorporated
samples of each treatment equally distributed across 5 sites separated
from each other by ∼1 km. These were treated as randomized blocks in
the analysis, and block effects were subsequently deemed
non-significant.
Effects of predators on losses
At one of these sites, I repeated the factorial manipulation of density
and number of fish on a subset of patch reefs that were enclosed in
Table 1.
cages. Using a random number table, I selected reefs for predator
exclusion from a larger set of reefs that conformed to reef size and
substratum attributes detailed in Shima (2001a). Cages were
custom-designed to exclude potential predators (i.e. large piscivorous
fishes and invertebrates) while still allowing emigration of new settlers
(verified by observing emigration of individuals through mesh when
cages were placed over barren sand). Cages were constructed of 0.635
cm square knotless mesh (Delta Style, Memphis Net & Twine Co.), and
were cylindrical, measuring either 1.83 or 3.05 m diameter, and 3.66 m
in height. Cylinders were weighted with galvanized chain (0.95 cm link
diameter) flush with the reef flat surrounding the base of patch reefs,
and floated at the surface with a contiguous ring of foam buoys to
exclude predator immigration. Prior to transplantation of new settlers,
I attempted to remove all resident piscivores from caged reefs by use of
hand nets and quinaldine anaesthetic. Because I was working with a
limited number of cages (n = 12), densities × number treatments were
conducted over field seasons in 1998 and 1999 (concurrent with
un-caged factorial treatments in 1999 only, Table 1); hence results
necessarily include the assumption that treatment effects were constant
between years. Caged reefs were censused daily for predators; caged
reefs containing predators were excluded from statistical analyses (no.
of excluded reefs, 3 for 1998, 2 for 1999). The role of predators was
evaluated along with density- and number-dependent effects with the
aid of a categorical analysis of variance (CATMOD Procedure, SAS).
This procedure estimates linear parameters based on weighted least
squares (response function = mean), but does not assume the response
(i.e. mortality, in this case 0%, 50%, or 100%) to be normally
distributed. Non-significant interactions were sequentially excluded
from ANOVA models.
Behaviour patterns dependent upon density or number?
I conducted a set of behavioural time-budget observations to explore
patterns of density- and/or number-dependence in behavioural
attributes that may contribute to observed patterns of mortality of newly
settled six-bar wrasse. For seventy focal individuals haphazardly
selected from 70 different patch reefs (reefs were selected randomly
with the aid of a random number table), I recorded the number of
agonistic encounters (i.e. chases by other conspecifics or heterospecific
residents) in a 10-min period. I recorded all behavioural data while on
snorkel, positioned >3 m from the focal individual, and my presence did
not appear to alter the behaviour patterns of residents on the focal patch
reef. All observations were made on focal individuals ≤ 45 mm in total
length (corresponding to fish <4 months post-settlement age; Shima
1999a) between 1100 and 1500 hours (time of peak activity), May–June
1999. Following each 10-min observation, I estimated the number of
resident conspecifics and quantified the size of each reef (to facilitate
estimates of density as number of six-bar wrasse per area of reef) using
Experimental design used to explore effects of predators on density- and
number-dependent mortality
Predator
exposure
Population
size (no. reef–1)
Population
density (no. m–2)
Reef
area (m–2)
Reef
n
Fish
n
Year
Uncaged
Uncaged
Uncaged
Uncaged
Caged
Caged
Caged
Caged
1
1
2
2
1
1
2
2
0.25
0.50
0.25
0.50
0.25
0.50
0.25
0.50
4
2
8
4
4
2
8
4
50A
50A
25A
25A
5
4
5
5
50
50
50
50
5
4
10
10
1999
1999
1999
1999
1998
1998
1999
1999
A
Reefs were evenly divided among five blocks separated by ~1 km; block effects were not significant
(Shima 2001a)
Population regulation of a coral-reef fish
177
methods of Shima (2001a, 2001b). Chase frequencies were analysed by
ANCOVA (GLM Procedure, SAS), with ‘number’ of resident
conspecifics treated as a categorical variable (‘0’, ‘1’, ‘2’, or ‘3 or
more’; only three reefs had values >3), and conspecific ‘density’ as a
covariate. Non-significant interactions were excluded from analyses.
labrids (primarily Pseudocheilinus hexataenia) accounted for
∼18% of the chases, while territorial damselfish (Stegastes
nigricans) were responsible for ∼7%. Only two chases from
potential predators were observed in >700 min of focal
observations (one from a hawkfish Paracirrhites arcatus, the
Results
Density- versus number-dependent mortality and the role of
predators
In the absence of predators, newly settled six-bar wrasse
experienced rates of loss estimated to be ∼26% over a 7-day
period. Losses from reefs exposed to predators were
significantly greater (∼62% over 7 days), and suggest that
36% of the losses in this experiment resulted from predation
while the remaining 26% may be attributable to some
combination of handling mortality, emigration, starvation or
disease. Patterns of losses from caged reefs appear
independent of both density and number (Fig. 1). In contrast,
reefs exposed to predators experienced significant densityand number-dependent losses (Fig. 1; details in Shima
2001a). Although statistical power to detect a significant
interaction between ‘predator-exposure’ and ‘density’ and/or
‘number’ treatments was quite low in this study (‘predator
exposure’ × ‘density’ P = 0.44; ‘predator exposure’ ×
‘number’ P = 0.39; Table 2), overall results present a
compelling pattern suggestive of predator-mediated densityand number-dependent mortality (Fig. 1).
Table 2. Analysis of variance of per capita mortality over 7 days
of newly settled six-bar wrasse as affected by predators (either
caged or uncaged reefs), number of newly settled wrasse per reef
(either 1 or 2), or the density of newly settled wrasse per reef
(either 0.25 or 0.5 fish m–2)
Results were obtained by fitting linear models to functions of
categorical data (response function = ‘mean’) by the CATMOD
Procedure of SAS. Non-significant interactions were sequentially
dropped from the analysis. Non-significant residual χ2 indicates a
significant model fit
Source
df
χ2
P
Intercept
Predator exposure
Number per reef
Density
Residual
1
1
1
1
4
92.93
17.27
11.42
4.15
3.14
<0.0001
<0.0001
0.0007
0.0416
0.5348
12
8
Behaviour patterns dependent on density or number?
6
Resident conspecifics accounted for ∼67% of all chases
incurred by juvenile six-bar wrasse. Heterospecific juvenile
4
1
50
0.75
50
Chases incurred
(no. per 10 min)
2
50
Per capita loss
A.
10
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0.4
0.6
0.8
1
1.2
1.4
12
B.
10
0.5
50
4
10
10
8
5
6
0.25
4
2
0
No. per patch reef
1
2
1
2
0
0
Density
0.25 Fish / m
2
0.50 Fish / m
2
Fig. 1. Mean per capita mortality over 7 days (±1 s.e.) for fish on reefs
caged to exclude predators (open bars) and uncaged reefs (shaded bars).
Fish were stocked to reefs at low or high density, and either as solitary
individuals or in groups of two. Statistics given in Table 2 indicate
significant effects of ‘predator exposure’, ‘density’, and ‘number per
reef’, with no significant interactions between main-effects.
0.2
Density of conspecifics
(no. per m2 reef)
Fig. 2. Chases incurred by focal six-bar wrasse over a 10 min period
as a function of their density (No. six-bar wrasse m–2 reef). Frequencies
of (A) total incurred chases and (B) chases by conspecifics did not scale
significantly with conspecific densities (A: P = 0.21, r2 = 0.02;
B: P = 0.13, r2 = 0.03; see supporting statistics in Table 3).
178
Jeffrey S. Shima
Table 3. Analysis of covariance of chases incurred by focal
individuals over 10 min, as affected by the number of
neighbouring six-bar wrasse per reef (‘0’, ‘1’, ‘2’, or ‘3 or more’)
and the density (no. per reef area) of six-bar wrasse per reef
(treated as a continuous covariate)
Non-significant interactions were sequentially dropped from the
analysis
A. Chases from all species
Source
df
MS
F
P
Number per reef
Density
Error
3
1
65
13.21
0.53
251.81
3.41
0.14
3.87
0.02
0.71
B. Chases from conspecifics
Source
df
MS
F
P
Number per reef
Density
Error
3
1
65
8.53
2.14
238.73
2.32
0.58
3.67
0.08
0.44
4
Chases incurred
(no. per 10 min)
3
2
1
0
0
1
2
3 OR MORE
Number of conspecific neighbors
(per patch reef)
Fig. 3. Mean chases incurred by focal six-bar wrasse over a 10 min
period (±1 s.e.) as a function of number of neighbouring conspecific
fish per reef. Frequencies of total incurred chases (open bars) and
chases by conspecifics (shaded bars) scaled with the number of
conspecific fish per reef, but not with their densities (Table 3, Fig. 2).
other from a triplefin: Family Tripterygiidae). Frequencies
of chases (both the total number of chases, and those from
conspecifics only) were explained by the ‘number’ of
individual six-bar wrasse per patch reef, with no effects
attributable to density (Fig. 2, Table 3). Number of chases
increased with the number of conspecifics resident on a reef
(Fig. 3); if such agonistic interactions increase the likelihood
of predation, these number-dependent behaviour patterns
may contribute to the predator-induced number-dependent
mortality observed in the above experiment.
Discussion
Results of the factorial manipulation of ‘density’, ‘number’,
and ‘predator exposure’ suggest that newly settled six-bar
wrasse populations may be regulated by predator-mediated
density- and number-dependent mortality. Although the
occurrence of both density- and number-dependent losses
has been documented elsewhere for this species (Shima
2001a), the mechanisms potentially responsible for these
compensatory effects have heretofore remained unexplored.
Losses resulting from density-dependent emigration (i.e.
movement from a focal reef to a neighbouring reef) are likely
to have vastly different consequences for population and
community dynamics at larger scales than losses resulting
from mortality events. In this study I constructed cages that
excluded predators from patch reefs but still allowed
migration of newly settled six-bar wrasse between reefs.
Although it is conceivable that cages may have slowed rates
of migration between reefs, the fact that fish were physically
capable of navigating the cage mesh to move from barren
sand to nearby patch reefs, coupled with circumstantial
evidence that this species rarely migrates between patch
reefs in the absence of cages (Shima 1999b), suggests that
any such bias was likely to be quite low. Elevated
survivorship of individuals (apparently independent of
‘density’ and ‘number of fish per reef ’) within cages was
strongly indicative of predation as an important mechanism
of compensatory loss.
Recent studies have confirmed the important role of
predators in the regulation of local reef-fish populations
(Hixon and Carr 1997), and density-dependence is
commonly viewed as arising from interference competition
between individuals for refuges from predators (Hixon 1991;
Caley 1993; Hixon and Beets 1993; Caley and St John 1996).
The conceptual model underlying many of these studies is
one of refuges (e.g. branching corals, seagrass beds,
macroalgae) that vary in quantity and/or quality, such that the
likelihood of being consumed by a predator is dependent
upon the number of fish relative to the number/quality of
refuges. For patterns of mortality that appear to scale with
the number of fish independent of the number of potential
refuges (i.e. ‘number-dependence’), some other mechanism
is likely to be responsible for compensatory mortality (e.g.
Connell 2000).
Behavioural observations of six-bar wrasse suggest that
frequencies of agonistic encounters between conspecifics
scale with the number of individuals on a patch reef more so
than with their density. Because the six-bar wrasse is a
sex-changing species, there exists the potential that this
numerical scaling may reflect interactions directed at
establishment of dominance hierarchies controlling
sex-change rather than those that result from interference
competition for refuges from predators. Although such
interactions prior to sexual maturity have received little
attention, recent work on another sex-changing reef fish
(Amphiprion spp.) suggests that establishment of social
dominance may begin early in the juvenile stage for some
species (P. Buston, personal communication).
Such
Population regulation of a coral-reef fish
179
interactions, if important, may increase the likelihood of
emigration for the ‘loser’, and this may provide a partial
explanation for the non-significant trend of increased losses
on reefs with higher numbers of individuals in the absence of
predators. If such agonistic encounters also attract the
attention of nearby predators (e.g. Elkin and Baker 2000),
and/or if chase-time is inversely proportional to vigilance
(because chased fish may not be mindful of predators),
number-dependent behaviour patterns interacting with
predation may constitute an alternative mechanism that can
account for number-dependent compensatory mortality of
recently settled six-bar wrasse. In such a situation,
behavioural interactions (e.g. social posturing related to sex
change) that are decoupled from proximate limiting
resources (e.g. food availability, refuges from predators) may
indirectly result in patterns of compensatory mortality
capable of regulating local populations.
This work underscores the need to consider indirect
effects of behavioural interactions in addition to more
traditional views of resource limitation as potential
mechanisms contributing to the regulation of local
populations.
Acknowledgments
This work would not have been possible without field
assistance from S. Kleinschmidt. Logistic support was
provided K. Seydel, B. Williamson, and the staff of the
Gump Research Station on Moorea. The contents of this
manuscript benefited from discussions with R. Schmitt, S.
Holbrook, S. Gaines, R. Warner, N. Phillips, A.
Stewart-Oaten, and A. Brooks, and comments on versions of
the manuscript were kindly provided by S. Holbrook and R.
Schmitt. Funding for this work was provided by RTG and
GRT Programs in Spatial Ecology (NSF BIR94–13141 and
NSF GER93–54870, both to W. Murdoch), an NSF award to
R. Schmitt and S. Holbrook (OCE 99–10677), and the
Partnership for the Interdisciplinary Study of Coastal Oceans
(PISCO—Supported by the Packard Foundation). This is
contribution No. 84 of the UC Berkeley Gump Research
Station and No. 42 of PISCO.
References
Booth, D. J., and Brosnan, D. M. (1995). The role of recruitment
dynamics in rocky shore and coral reef fish communities. Advances
in Ecological Research 26, 309–85.
Caley, M. J. (1993). Predation, recruitment, and the dynamics of
communities of coral reef fishes. Marine Biology 117, 33–43.
Caley, M. J., and St John, J. (1996). Refuge availability structures
assemblages of tropical reef fishes. Journal of Animal Ecology 65,
414–28.
Caley, M. J., Carr, M. H., Hixon, M. A., Hughes, T. P., Jones, G. P., and
Menge, B. A. (1996). Recruitment and the local dynamics of open
marine populations. Annual Reviews of Ecology and Systematics 27,
477–500.
Connell, S. D. (2000). Is there safety-in-numbers for prey? Oikos 88,
527–32.
Doherty, P. J. (1991). Spatial and temporal patterns in recruitment. In
‘The Ecology of Fishes on Coral Reefs’. (Ed. P. F. Sale.) pp. 261–93.
(Academic Press: New York.)
Elkin, C. M., and Baker, R. L. (2000). Lack of preference for
low-predation-risk habitats in larval damselflies explained by costs
of intraspecific interactions. Animal Behaviour 60, 511–21.
Hixon, M. A. (1991). Predation as a process structuring coral reef fish
communities. In ‘The Ecology of Fishes on Coral Reefs’. (Ed. P. F.
Sale.) pp. 475–508. (Academic Press: New York.)
Hixon, M. A., and Beets, J. P. (1989). Shelter characteristics and
Caribbean fish assemblages: experiments with artificial reefs.
Bulletin of Marine Science 44, 666–80.
Hixon, M. A., and Beets, J. P. (1993). Predation, prey refuges, and the
structure of coral reef fish assemblages. Ecological Monographs 63,
77–101.
Hixon, M. A., and Carr, M. H. (1997), Synergistic predation, density
dependence, and population regulation in marine fish. Science 277,
946–9.
Jones, G. P. (1991). Postrecruitment processes in the ecology of coral
reef fish populations: a multifactorial perspective. In ‘The Ecology
of Fishes on Coral Reefs’. (Ed. P. F. Sale.) pp. 294–328. (Academic
Press: New York.)
Murdoch, W. W. (1994). Population regulation in theory and practice.
Ecology 75, 271–87.
Schmitt, R. J., and Holbrook, S. J. (1999). Mortality of juvenile
damselfish: implications for assessing processes that determine
abundance. Ecology 80, 35–50.
Shima, J. S. (1999a). Variability in relative importance of determinants
of reef fish recruitment. Ecology Letters 2, 304–10.
Shima, J. S. (1999b). An evaluation of processes that influence variation
in abundance of a coral reef fish. PhD dissertation, University of
California, Santa Barbara.
Shima, J. S. (2001a). Regulation of local populations of a coral reef fish
via joint effects of density- and number-dependent mortality.
Oecologia 126, 58–65.
Shima, J. S. (2001b). Recruitment of a coral reef fish: roles of
settlement, habitat, and postsettlement losses. Ecology 82, 2190–9.
Sinclair, M. (1988). ‘Marine Populations: an Essay on Population
Regulation and Speciation.’ (University of Washington Press:
Seattle.) 252 pp.
Steele, M. A. (1999). Effects of shelter and predators on reef fishes.
Journal of Experimental Marine Biology and Ecology 233, 65–79.
Turchin, P. (1995). Population regulation: old arguments and a new
synthesis. In ‘Population Dynamics: New Approaches and
Synthesis’. (Eds N. Cappuccino and P. W. Price.) pp. 19–40.
(Academic Press: San Diego.)
Victor, B. C. (1986). Duration of the planktonic larval stage of one
hundred species of Pacific and Atlantic wrasses (family Labridae).
Marine Biology 90, 317–26.
Manuscript received 14 May 2001; revised and accepted 21 September
2001
http://www.publish.csiro.au/journals/mfr