Early Life History Traits and Recruitment Patterns of Caribbean Wrasses (Labridae)

Abstract

Despite the fact that recruitment can significantly influence the population dynamics of benthic marine populations, relatively little is known about the biological and physical processes controlling recruitment. We selected eight closely related coral reef fishes (wrasses in the family Labridae) to examine the temporal and spatial patterns of juvenile recruitment to the Caribbean island of Barbados. We used a comparative approach to study the relationships among patterns of recruitment, early life history traits, and aspects of the physical environment. For 10 wk during each of three peak recruitment (spring) seasons (1990-1992), we used a biweekly census of recently settled juveniles (8-25 mm standard length, SL) to measure the abundance of six congeners, Halichoeres bivittatus, H. radiatus, H. poeyi, H. garnoti, H. pictus, and H. maculipinna, and two confamilial labrids, Thalassoma bifasciatum and Bodianus rufus. Analysis of the otoliths of a sample of collected specimens provided estimates of larval durations, postsettlement ages, sizes at settlement, and juvenile growth rates, enabling back-calculation of settlement day for all collected juveniles. We compared temporal patterns of recruitment among species, and spatial patterns of recruitment for the most common species. Temporal patterns of recruitment were consistent among seasons for most of the labrids examined, although the magnitude of recruitment was less predictable (particularly for H. poeyi, H. maculipinna, and B. rufus). The eight labrids could be divided into two groups based on their early life history traits and within-season temporal patterns of recruitment. Halichoeres bivittatus, H. radiatus, H. poeyi, H. garnoti, and H. pictus had larval durations that were relatively short and invariant (means of 23-27 d), and all settled at fairly large sizes (9-12 mm SL) during the new moon and first maximum amplitude tide. In contrast, T. bifasciatum, B. rufus, and H. maculipinna had larval durations that were longer or more variable, and all three were able to delay metamorphosis. These three species settled at relatively smaller sizes (8-10 mm SL) during the third-quarter moon and second minimum amplitude tide. We compared temporal patterns of T. bifasciatum recruitment between Barbados and Caribbean Panama in an attempt to identify further the proximate environmental cues operating during settlement. Contrasting patterns of T. bifasciatum recruitment between the two geographical locations probably result from differences in the relative timing of the lunar and tidal amplitude cycles. Recruitment of labrids to Barbados occurred along the entire west coast of the island. Although some labrids had rather specific habitat requirements (e.g., B. rufus associated exclusively with large seaward-facing coral heads such as Montastrea spp.), most species were ubiquitous along the west coast. Species-specific juvenile densities did not often vary significantly among sites following major recruitment events, although overall densities were generally lower at a central site. Lower recruitment to that site likely results from reduced rates of larval supply due to prevailing offshore tidal flows. Thus, temporal and spatial patterns of labrid recruitment to Barbados appear to be more predictable than previously thought for reef fishes. In particular, variation in the tidal amplitude cycle may influence both the timing of settlement and, to a lesser degree, the spatial scale of larval supply. Finally, the interaction of larval biology with such physical processes is evident in the correlation between temporal patterns of recruitment and early life history traits. The functional nature of this relationship clearly warrants further study.

Full text

177
Ecological Monographs,
67(2), 1997, pp. 177–202
q
1997 by the Ecological Society of America
EARLY LIFE HISTORY TRAITS AND RECRUITMENT PATTERNS OF
CARIBBEAN WRASSES (LABRIDAE)
S
U
S
PONAUGLE AND
R
OBERT
K. C
OWEN
Marine Sciences Research Center, State University of New York, Stony Brook, NY 11794-5000 USA
Abstract.
Despite the fact that recruitment can significantly influence the population
dynamics of benthic marine populations, relatively little is known about the biological and
physical processes controlling recruitment. We selected eight closely related coralreef fishes
(wrasses in the family Labridae) to examine the temporal and spatial patterns of juvenile
recruitment to the Caribbean island of Barbados. We used a comparative approach to study
the relationships among patterns of recruitment, early life history traits, and aspects of the
physical environment. For 10 wk during each of three peak recruitment (spring) seasons
(1990–1992), we used a biweekly census of recently settled juveniles (8–25 mm standard
length, SL) to measure the abundance of six congeners,
Halichoeres bivittatus, H. radiatus,
H. poeyi, H. garnoti, H. pictus,
and
H. maculipinna,
and two confamilial labrids,
Thalassoma
bifasciatum
and
Bodianus rufus.
Analysis of the otoliths of a sample of collected specimens
provided estimates of larval durations, postsettlement ages, sizes at settlement, and juvenile
growth rates, enabling back-calculation of settlement day for all collected juveniles. We
compared temporal patterns of recruitment among species, and spatial patterns of recruit-
ment for the most common species.
Temporal patterns of recruitment were consistent among seasons for most of the labrids
examined, although the magnitude of recruitment was less predictable (particularly for
H.
poeyi, H. maculipinna,
and
B. rufus
). The eight labrids could be divided into two groups
based on their early life history traits and within-season temporal patterns of recruitment.
Halichoeres bivittatus, H. radiatus, H. poeyi, H. garnoti,
and
H. pictus
had larval durations
that were relatively short and invariant (means of 23–27 d), and all settled at fairly large
sizes (9–12 mm SL) during the new moon and first maximum amplitude tide. In contrast,
T.
bifasciatum, B. rufus,
and
H. maculipinna
had larval durations that were longer or more
variable, and all three were able to delay metamorphosis. These three species settled at
relatively smaller sizes (8–10 mm SL) during the third-quarter moon and second minimum
amplitude tide. We compared temporal patterns of
T. bifasciatum
recruitment between Bar-
bados and Caribbean Panama in an attempt to identify further the proximate environmental
cues operating during settlement. Contrasting patterns of
T. bifasciatum
recruitment between
the two geographical locations probably result from differences in the relative timing of the
lunar and tidal amplitude cycles. Recruitment of labrids to Barbados occurred along the entire
west coast of the island. Although some labrids had rather specific habitat requirements (e.g.,
B. rufus
associated exclusively with large seaward-facing coral heads such as
Montastrea
spp.), most species were ubiquitous along the west coast. Species-specific juvenile densities
did not often vary significantly among sites following major recruitment events, although
overall densities were generally lower at a central site. Lower recruitment to that site likely
results from reduced rates of larval supply due to prevailing offshore tidal flows.
Thus, temporal and spatial patterns of labrid recruitment to Barbados appear to be more
predictable than previously thought for reef fishes. In particular, variation in the tidal
amplitude cycle may influence both the timing of settlement and, to a lesser degree, the
spatial scale of larval supply. Finally, the interaction of larval biology with such physical
processes is evident in the correlation between temporal patterns of recruitment and early
life history traits. The functional nature of this relationship clearly warrants further study.
Key words: Barbados;
Bodianus;
Caribbean; coral reef fishes; early life history;
Halichoeres;
larval
duration; lunar synchrony; metamorphosis, delay of; tidal synchrony; otoliths;
Thalassoma;
tides.
I
NTRODUCTION
In recent years, much attention has been directed
toward understanding the role of recruitment in struc-
turing marine populations (e.g., reviewed in Doherty
Manuscript received 2 February 1996; revised 24 July
1996; accepted 26 July 1996.
1991, Jones 1991, Olafsson et al. 1994, Booth and
Brosnan 1995). Because most benthic marine organ-
isms have complex life cycles in which larvae and
adults occupy different environments, the demography
of these open populations can be influenced by pro-
cesses occurring in the benthos after recruitment or by
processes occurring in the plankton prior to or during

178
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
T
ABLE
1. Potential mechanisms creating variation in recruitment.
Scale of variation Physical Biological
a) Temporal
Seasonal (interannual) Fluctuations in large-scale cur-
rents, storm frequency Reproductive output
Within season
Episodic Winds, fronts, eddies, storms,
upwelling, relaxation events Reproductive timing
Biweekly/monthly Lunar and tidal amplitude cy-
cle Reproductive timing, length of
larval life, delay of meta-
morphosis, vertical migra-
tion, directional swimming
Daily Diel and tidal cycle Vertical migration, directional
swimming
b) Spatial†
Geographic Large-scale current flow
Differences in tidal amplitude
cycles
Reproductive timing
Reproductive source, length of
larval life, behavioral re-
sponses to current flow cues
Island-wide (among
sites) Local current flow Habitat selection
Microhabitat (within
sites) Boundary-layer flows Microhabitat selection (includ-
ing gregarious settlement)
† Geographic spatial scale is measured over hundreds to thousands of kilometers, island-wide
over tens of kilometers, and microhabitat
,
1 km.
recruitment. ‘‘Recruitment’’ is generally defined as the
settlement (transition from a planktonic larvae into a
benthically oriented juvenile) and survival of new ju-
veniles to the time of sampling (Richards and Linde-
man 1987, Forrester 1990). Although this definition
encompasses early postsettlement mortality, when ju-
veniles are sampled very shortly after settlement, ‘‘re-
cruitment’’ becomes functionally equivalent to ‘‘set-
tlement.’’
During the past several decades, demographic par-
adigms for the regulation of open marine populations
have shifted in emphasis between the importance of
benthic vs. planktonic processes. The current most
widely celebrated view recognizes the importance of
both benthic and planktonic processes in structuring
marine populations, and suggests that at any point in
space and time, benthic populations may be controlled
to a greater or lesser degree by density-independent
processes influencing the supply and settlement of lar-
vae, or density-dependent processes occurring after re-
cruitment. This concept has been particularly empha-
sized in the literature on coral reef fishes (Victor 1986
a
,
Warner and Hughes 1988, Forrester 1990, Jones 1990,
1991, Doherty 1991, Hixon and Beets 1993). However,
although recruitment dynamics clearly can have an im-
portant role in the regulation of benthic marine pop-
ulations (e.g., for coral reef fishes, Williams 1980, Do-
herty 1981, 1983, Victor 1983
a,
1986
a,
Doherty and
Fowler 1994; as well as a diversity of other species:
Cowen 1985, Gaines et al. 1985, Choat et al. 1988,
Keough 1988, Roughgarden et al. 1988, Raimondi
1990, Gaines and Bertness 1992, Peterson and Sum-
merson 1992, Cowen and Bodkin 1993, Botsford et al.
1994, Eggleston and Armstrong 1995), relatively little
is known about the physical and biological processes
controlling recruitment patterns.
Variation in recruitment on a variety of temporal and
spatial scales (reviewed in Doherty 1991, Olafsson et
al. 1994) may be influenced by the interaction of phys-
ical transport mechanisms and active behavior by
adults or larvae (Table 1). Physical mechanisms such
as large-scale shifts in prevailing currents can create
temporal variation in recruitment among seasons (Cow-
en 1985, Choat et al. 1988, Farrell et al. 1991, Gaines
and Bertness 1992), while within-season recruitment
events may be associated with episodic events such as
wind stress (Milicich 1994), storms (Reed et al. 1988),
wind-induced onshore current flow (Shenker et al.
1993, Thorrold et al. 1994
b
), fronts or surface slicks
(Kingsford and Choat 1986, Wolanski and Hamner
1988, Kingsford et al. 1991), upwelling (Roughgarden
et al. 1991) or relaxation events (Wing et al. 1995), or
the impingement of mesoscale eddies (Sale 1970, Lobel
and Robinson 1986, Boehlert et al. 1992, Lee et al.
1992). On a monthly scale, recruitment may vary with
the lunar cycle (reviewed in Doherty 1991, Robertson
1992, Booth and Beretta 1994, Sponaugle and Cowen
1994, 1996
a, b
) or the tidal amplitude cycle as a result
of variable transport during spring or neap tides, in-
cluding transport by tidally induced internal waves or
surface slicks (reviewed in Shanks 1995). Active larval
behavior such as vertical migration among water mass-
es (reviewed in Leis 1991, Young 1995) or directional
swimming (e.g., Hare and Cowen,
in press,
Leis et al.
1996) probably interacts with these physical mecha-
nisms to control the timing of recruitment.
Aspects of the early life history of benthic marine
organisms may also influence the timing of recruitment

May 1997 179
RECRUITMENT PATTERNS OF WRASSES
(Table 1). For some reef fishes, particularly those with
rather invariant larval durations, variation in the mag-
nitude and timing of adult reproduction determines re-
cruitment variability both among and within seasons
(Robertson et al. 1988, Meekan et al. 1993, but see
Robertson et al. 1993). Early life history attributes such
as the length of larval life have obvious impacts on the
degree to which larvae are subjected to transport by
physical processes (Scheltema 1977, Jackson and
Strathman 1981, Cowen 1991, Jenkins and May 1994).
The capacity to delay metamorphosis may decouple the
temporal relationship between spawning and recruit-
ment for some fishes (Victor 1986
a,
Sponaugle and
Cowen 1994), and enable greater recruitment synchro-
nization to environmental cues (Sponaugle and Cowen
1994). Such developmental flexibility, including vari-
ability in larval growth rates (e.g., McCormick 1994)
could potentially modify the supply of settlers.
Variation in the spatial patterns of recruitment is
probably similarly influenced by physical and behav-
ioral processes (Table 1). Large-scale currentscan pro-
duce concurrent recruitment events at wide spatial
scales (Cowen 1985, Victor 1986
a,
Doherty 1987,
Choat et al. 1988, Pitcher 1988, Farrell et al. 1991),
while local differences in tidal currents have the po-
tential to create contrasting patterns of recruitment
among distant locations. Because some environmental
cycles differ among locations (e.g., the tidal amplitude
cycle), environmentally cued spawning or recruitment
may differ among sites. Among-site differences in the
source of larvae also could lead to variable spatial re-
cruitment. For example, a downstream site may receive
a relatively continuous stream of larvae spawned from
several upstream sites, whereas an upstream site may
receive larvae in pulses correlated with the spawning
of that population. In addition, because the duration of
the larval period may vary among different populations
(Thresher and Brothers 1989, Thresher et. al. 1989,
Wellington and Victor 1989, 1992), the influence of
spawning patterns and physical mechanisms on the sup-
ply of larvae may vary among sites. Smaller scale (e.g.,
within island or region) spatial patterns of recruitment
may be controlled by the specifics of local flows; how-
ever, if the distribution of various habitats, conspecif-
ics, or predators varies spatially, habitat selection by
larvae may contribute to spatial variation in recruitment
(e.g., Sweatman 1983, Sale et al. 1984, Eckert 1985,
Shulman 1985, Booth 1992). Larvae of sessile inver-
tebrates may actively select microhabitats based on a
suite of chemical cues (e.g., Crisp 1974, Pawlik 1992),
although physical processes occurring in the boundary
layer (e.g., small-scale eddies) also likely passively
modify recruitment patterns (e.g., Butman 1987, Eck-
man et al. 1994).
Clearly, there are many physical and behavioral pro-
cesses that may function to create the high degree of
recruitment variability apparent in benthic marine pop-
ulations. In general, however, relatively few have been
shown to directly influence the temporal and spatial
patterns of recruitment. Because the proximate causes
of recruitment variability have been difficult to identify
for organisms such as coral reef fishes, there has been
little consensus among the theories of ultimate causes
of pattern (particularly temporal) in spawning and re-
cruitment (e.g., Johannes 1978, Thresher 1984, Rob-
ertson et al. 1990, Robertson 1991). This is due in part
to the lack of extensive data on a variety of species
(especially closely related species) from diverse loca-
tions. Most studies to date have focused on one or two
species (but see Robertson 1992) recruiting to a single
location. In addition, for many reef fishes studied, lar-
val durations are rather invariant, and temporal patterns
of recruitment may simply reflect spawning patterns
(Robertson et al. 1988, Meekan et al. 1993). Thus, prior
to exploring the ultimate causes of recruitment vari-
ability, it is necessary to identify the proximate causes
of temporal and spatial patterns of recruitment.
In this study, we selected a group of closely related
coral reef fishes to examine the proximate causes of
recruitment variability. We examined variability in the
temporal and spatial patterns of recruitment for eight
species of labrids recruiting to the upstream Caribbean
island of Barbados: the slippery dick,
Halichoeres bi-
vittatus
(Bloch); the puddingwife,
H. radiatus
(Lin-
naeus); the blackear wrasse,
H. poeyi
(Steindachner);
the yellowhead wrasse,
H. garnoti
(Valenciennes); the
painted wrasse,
H. pictus
(Poey); the clown wrasse,
H.
maculipinna
(Muller and Troschel); the bluehead
wrasse,
Thalassoma bifasciatum
(Bloch); and the Span-
ish hogfish,
Bodianus rufus
(Linnaeus). While there
may be some seasonality to the spawning of at least
one species at Barbados (
T. bifasciatum
; Hunt von
Herbing and Hunte 1991), there does not appear to be
any lunar synchrony within-season; all eight species
are pelagic spawners that spawn daily (Warner and
Robertson 1978, Robertson 1981). Because particular
patterns of recruitment may be associated with various
larval behaviors or early life history characteristics, we
conducted a comparative analysis to determine whether
certain larval traits (i.e., length of larval life, size at
settlement, growth rate) are correlated with specific
temporal patterns of recruitment. We also examined
within-island variability in spatial patterns of recruit-
ment. We compared these patterns to various environ-
mental characteristics to determine the degree to which
recruitment patterns are influenced by active habitat
selection or by physical processes controlling larval
supply. Finally, to identify which proximate environ-
mental cues might be important in the timing of re-
cruitment, we compared the recruitment patterns of one
wrasse (
T. bifasciatum
) to those reported from another
geographic location. Environmental cycles (e.g., tidal
amplitude cycle) are often different, or offset, among
various geographical locations, while lunar cycles are
constant; thus, comparisons of temporal recruitment
patterns among geographical locations may reveal

180
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
F
IG
. 1. Map of Barbados, West Indies, with the nearshore
census sites (solid circles) located inside the bank reef
(dashed line).
which of these two proximate cues are important in the
monthly timing of recruitment.
M
ETHODS
Site description
Barbados is the easternmost island in the Lesser An-
tilles chain of islands located in the eastern Caribbean.
Lying within the belt of northeast trade winds, in the
path of the wind-driven Equatorial Current, Barbados
is upstream of the other Caribbean islands. Recent data
collected offshore suggest that large-scale flow around
the island may be topographically steered: as the north-
westerly-flowing current impinges upon the southeast-
ern coast of Barbados, it diverges around the island,
continuing north and recirculating around a deep ridge
well north of the island before continuing downstream
(Cowen and Castro 1994). This flow may retain larvae
spawned locally, contributing to the maintenance (on
ecological time scales) of the high diversity of fishes
(Cowen and Castro 1994).
We selected census sites to span most of the western
coast of Barbados (Fig. 1). All sites were located on
the nearshore reefs, which are distributed between 100
and 400 m offshore and
ø
600 m inshore of the bank
reef (which parallels the west coast). These nearshore
reefs consist of high relief spurs of dead coral matrix,
outcropping seaward from the continuous reef crest and
separated from other spurs by sand or sand–rubble
grooves. Reefs are concentrated at small coastal head-
lands, and extend toward shore, with no regions of
seagrass. Sites were 2–2.5 km apart, and each encom-
passed
ø
4000 m
2
of coral and rubble habitat in water
depths of 2.5–6.5 m.
Recruitment census
To determine the temporal and spatial patterns of
labrid recruitment to Barbados we censused and col-
lected new recruits biweekly (every other week) during
the spring of three consecutive years (1990–1992).
Spring (March–June) corresponds to the period of gen-
erally high larval abundance (Munro et al. 1973, Pow-
les 1975) and peak reef fish recruitment in the Carib-
bean (Luckhurst and Luckhurst 1977), although the re-
cruitment of
Thalassoma bifasciatum
to Barbados may
be higher during summer months (Hunt von Herbing
and Hunte 1991). During each census, we counted and
collected all recently settled juveniles (8–25 mm stan-
dard length, SL) from six randomly selected replicate
5
3
1 m quadrats; we counted schools (
.
10 fish) in
six 5
3
5 m quadrats, and collected a subsample (10%)
of each school for analysis. For otolith aging purposes
we made additional, qualitative collections of several
rarer species (
Halichoeres pictus
and
Bodianus rufus
).
We used the anesthetic Quinaldine and hand nets during
all counts and collections. We preserved collected spec-
imens in 70% ethanol prior to measuring SL, thus our
reported SL may be somewhat smaller than the actual
sizes at collection, due to shrinkage.
We attempted to measure larval supply directly by
the nightly deployment of larval light traps. Although
the traps proved successful at collecting the late-stage
larvae of a diversity of reef fishes, the abundance of
labrids was insufficient for analysis (Sponaugle and
Cowen 1996
a
). However, we were able to use these
late-stage larvae to estimate size at settlement. Note
that these larvae were similarly preserved in 70% eth-
anol prior to length measurements.
Otolith analysis
To obtain a measure of larval duration and construct
an age-to-length relationship for back-calculating re-
cruitment date, we removed and examined the otoliths
from a subsample of juveniles collected during each
season. The otoliths of many bony fishes contain con-
centric marks that are deposited daily, allowing a pre-
cise estimate of age at capture (e.g., Brothers et al.
1976, Victor 1982). For reef fishes settling to the ben-
thos, an additional, conspicuous settlement mark is of-
ten visible, permitting an estimate of larval duration as
well as postsettlement age (Victor 1982). Daily incre-
ment deposition has previously been validated for two
of the wrasses in this study (Victor 1982); it is rea-
sonable to extrapolate from these for the remaining
species. During this analysis, we randomly selected
individuals from each 1.0 mm length category (Cam-
pana and Jones 1992), and extracted the otoliths using
standard techniques (Brothers 1987). Storage of the
sagittae and lapilli in medium viscosity immersion oil
on microscope slides for 30 d facilitated interpretation.
We examined the sagittae with a computer-aidedimage
enhancer (Optical Pattern Recognition System, Bio-
sonics Incorporated, Seattle, Washington) that was at-
tached to a Zeiss compound microscope. We viewed
the otoliths under transmitted light at 250
3
(oil im-
mersion) magnification, using an adjustable polarizing

May 1997 181
RECRUITMENT PATTERNS OF WRASSES
T
ABLE
2. Age-on-length reduced major axes (RMA) re-
gression slopes (
6
1
SE
) and intercepts (
6
1
SE
) used to
back-calculate settlement date for eight species of labrids
recruiting to Barbados.
Species Year Slope Intercept
r
†
H. bivittatus
1990 2.61
6
0.16
2
27.67
6
2.82 0.81
1991 2.56
6
0.21
2
26.98
6
3.31 0.77
1992 2.24
6
0.16
2
20.85
6
2.74 0.88
H. radiatus
1990 2.62
6
0.20
2
28.28
6
3.17 0.88
1991 2.65
6
0.60
2
27.35
6
8.64 0.86
1992 1.99
6
0.21
2
16.26
6
3.57 0.85
H. poeyi
1990 2.54
6
0.21
2
30.12
6
3.62 0.87
H. garnoti
1990 1.83
6
0.21
2
19.90
6
3.88 0.71
1991 2.10
6
0.29
2
24.73
6
5.19 0.81
1992 1.99
6
0.19
2
20.87
6
3.24 0.79
H. pictus
1990 1.38
6
0.08
2
12.11
6
1.33 0.96
1992 1.59
6
0.10
2
15.10
6
1.73 0.89
H. maculipinna
1992 2.46
6
0.25
2
22.44
6
3.70 0.88
T. bifasciatum
1990 2.43
6
0.16
2
19.07
6
2.40 0.90
1991 2.92
6
0.03
2
25.85
6
4.23 0.70
1992 2.67
6
0.14
2
23.49
6
2.12 0.93
B. rufus
1991 1.60
6
0.16
2
14.41
6
2.32 0.85
1992 2.03
6
0.12
2
20.22
6
1.62 0.90
† Pearson product-moment correlation coefficient.
filter placed between the light source and the first stage.
We made at least two independent sagital counts of
each sample. For all labrid species except
Bodianus
rufus,
a wide discontinuous band was clearly evident
between the relatively narrow larval increments and
the wider juvenile increments. We interpreted the inner
edge of this band to indicate settlement (see Victor
1982), corresponding to the time when settling larvae
enter and remain buried under the sand for several days
(Victor 1983
b
; S. Sponaugle,
unpublished data
); like-
wise, the outer edge of the band corresponded to the
emergence of juveniles onto the reef. Where increments
were unclear in this region, we used interpolation based
on surrounding increment widths to estimate the num-
ber of days fish were buried. For
B. rufus,
no wide band
was evident; instead, a sharp transition (interpreted as
settlement) existed between the narrow larval incre-
ments and the relatively wider juvenile increments.For
all species, we added 2 d to the count of presettlement
increments to account for the time to hatching (Fritzche
1978, Victor 1982).
Habitat characteristics and resident
mature fish populations
Because spatial patterns of recruitment could be in-
fluenced by habitat selection by settling larvae or re-
cently settled juveniles, or by differential survival of
recruits in different habitats, we selected census sites
to minimize site-specific habitat differences. To mea-
sure site similarity, we quantified 10 variables at each
site: the percentage and diversity of live coral substrate;
the percentage of coral rock substrate, coral rubble sub-
strate, sand substrate, algal turf cover, algal turf–sand
cover (algal turf visible through heavy sand coating),
and sand cover (only sand visible); mean depth; and
rugosity. We used a standard point-contact method
(Greig-Smith 1964) to record the substrate and cover
under, and the water depth above, 20 points along six
randomly placed 5-m transects at each site. We con-
sidered rugosity to be the degree of variability (stan-
dard deviation) in depth among the points.
Spatial patterns of recruitment may also be influ-
enced by conspecific populations (e.g., Sweatman
1983, Booth 1992), as larvae settle into existing pop-
ulations, or are excluded by (or select not to settle near)
predators or dense resident populations (e.g., Shulman
1985). Therefore, we surveyed mature fishes during
1991 and 1992 by counting the number of mature fishes
within six randomly placed 5
3
5 m quadrats at each
site. For the purposes of this census, we defined ‘‘ma-
turity’’ by size alone (SL
.
3 cm). We did not include
small, cryptic species such as gobies, blennies, and
apogonids in the survey.
S
TATISTICAL
A
NALYSIS
Early life history characteristics
We compared otolith-derived measurements of lar val
duration (non-normally distributed) among species and
among years using a nonparametric Kruskal-Wallis
test, followed by a nonparametric multiple comparison
analysis (Dunn 1964, Zar 1984). We used reduced ma-
jor axes (RMA) procedures to adjust for the inherent
variability in the independent variable (length) of the
age-on-length linear regressions (Ricker 1973, Laws
and Archie 1981). Length-on-age regressions provided
an estimate of growth rate (slope) and size at settlement
(intercept), and the jackknife method produced esti-
mates of standard error for each statistic (Sokal and
Rohlf 1981). We compared growth rates and sizes at
settlement among species and years using the
T
9
meth-
od for unplanned comparisons (Sokal and Rohlf 1981).
Temporal and spatial patterns of recruitment
Because all age-on-length regressions had generally
high Pearson’s product-moment correlation coefficients
(Table 2), we used separate regressions for each species
from each year to estimate the postsettlement age of
new recruits of that year. In some instances, where few
fish were collected during a particular sampling season,
we used the age-to-length regression from another year
to estimate ages (i.e., for
Halichoeres poeyi
in 1992,
H. pictus
in 1991, and
H. maculipinna
in 1990 and
1991). For each collection, we weighted the size (age)
distribution of collected fish by abundance to provide
a more accurate estimate of the relative magnitude of
each event. Where species (i.e.,
H. bivittatus
and
Thal-
assoma bifasciatum
) appeared in the 5-m
2
quadrats (as
individuals) and in the 25-m
2
quadrats (in schools), we
tabulated the data from each quadrat type separately.
We tested periodicity in recruitment over the lunar
cycle for each species during each year with Rayleigh
tests (Batschelet 1981, Zar 1984). Used in the analysis
of biological rhythms when cycle periods are prede-
termined (e.g., lunar or tidal cycle), this test compares

182
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
the temporal distribution of recruitment to that of uni-
form recruitment over the given environmental cycle.
Where the distribution of recruitment is non-uniform,
circular statistics provide an estimate of peak time
(mean vector angle) and dispersion (mean angular de-
viation) of recruitment (Batschelet 1981). To obtain a
general temporal pattern of recruitment, we also pooled
the data for each species across all years and repeated
the tests.
The biweekly juvenile census also provided data on
spatial patterns of recruitment. We calculated themean
density of juveniles at each site for the most common
species from each census following a large recruitment
event. We tested these densities for site-related differ-
ences using the nonparametric Kruskal-Wallis test (due
to non-normality of the data), followed where indicated
by a nonparametric Tukey-type multiple comparison
analysis (Zar 1984). To obtain an overall view of spatial
variability in recruitment, we also calculated the mean
rank of each site for all these events. We obtained rough
estimates of mean juvenile mortality rates through
counts made during the second census after a large
recruitment event. Although such calculations may un-
derestimate true mortality due to recruitment between
sampling periods, we selected windows to minimize
these effects. Initial densities were calculated 1 wk after
each recruitment peak, and subsequent densities were
calculated before the next peak. For each species and
each large recruitment event we calculated daily mor-
tality rates (Ricker 1975). For the less common species,
we obtained a qualitative measure of site-specific dif-
ferences in abundance by summing the number of ju-
veniles censused at each site over the three seasons.
Habitat characteristics and resident mature
fish populations
We transformed (arcsine square root) habitat mea-
surements where appropriate (Sokal and Rohlf 1981),
and then compared these transformed proportions
among sites using standard ANOVA and multiple com-
parison procedures. In five instances (percentage of live
coral, diversity of live coral, percentage of sand sub-
strate, percentage of algal turf–sand cover, and per-
centage of sand cover), transformed variables remained
non-normally distributed, so we used nonparametric
Kruskal-Wallis tests, followed by nonparametric Tu-
key-type multiple comparisons (Zar 1984). To provide
a relative measure of site similarity, we used principal
components analysis (PCA) techniques to group the 10
habitat variables (Tatsuoka 1971). We then regressed
juvenile densities (for both common and rarer species)
as well as mature fish densities against three separate
PCA factors to determine whether habitat descriptors
influenced the spatial patterns of recruitment.
To measure the impact of mature fishes on juvenile
recruitment patterns, we grouped the counts of resident
mature fishes from each site and year into several cat-
egories: mean diversity of fishes, mean number of all
fishes, and mean number of labrids per quadrat. Due
to low abundances, we did not analyze these data at
the level of individual species. We transformed the data
where necessary to meet assumptions of normality, and
tested among sites and between years with two-way
ANOVA and multiple comparison procedures (Zar
1984). We regressed the densities of the most common
juveniles against mature fish densities to identify any
relationship between mature fishes and new recruits.
We also regressed the densities of mature fishes against
the three habitat PCA factors to determine whether ma-
ture fishes were associated with particular habitat fea-
tures.
R
ESULTS
Early life history characteristics and temporal
patterns of recruitment to Barbados
The eight species of labrids in this study could be
broadly divided into two groups based on several life
history traits and their temporal patterns of recruitment.
Most of the
Halichoeres
species (
H. bivittatus, H. ra-
diatus, H. poeyi, H. garnoti,
and
H. pictus
) had rela-
tively short, invariant larval durations and exhibited no
evidence of delayed metamorphosis. Although recruit-
ment rates varied interannually for several of these spe-
cies, monthly settlement occurred consistently during
the new moon (maximum amplitude tides). The other
three labrids,
H. maculipinna, T. bifasciatum,
and
B.
rufus,
had relatively longer or more variable larval du-
rations, and all three could delay metamorphosis. Set-
tlement by these species occurred during the third-quar-
ter moon and minimum amplitude tides.
New-moon recruits:
Halichoeres bivittatus, H.
radiatus, H. poeyi, H. garnoti,
and
H. pictus
Larval durations were similar over all years for the
five species of
Halichoeres
(Dunn method:
Q
,
3.17,
P
.
0.05), and were generally short and relatively in-
variant (Table 3, Fig. 2). Estimated size at settlement
was also similar among most of the five
Halichoeres
species (
T
9
method: difference [
d
]
5
0.06
2
3.85, min-
imum significant difference [msd]
5
2.25
2
4.95; Table
4), as were early juvenile growth rates (
T
9
method:
d
5
0
2
0.35, msd
5
0.1
2
0.4). However,
H. pictus
tended
to settle at a slightly smaller size (
T
9
method:
d
5
2.63
2
3.96, msd
5
2.40
2
2.95) and exhibit higher
growth rates than some of the other wrasses (
T
9
method:
d
5
0.22
2
0.36, msd
5
0.20). The size of the smallest
H. pictus
collected (11.9 mm, 5–6 d old, postsettle-
ment) tended to be smaller than the smallest
H. poeyi
(14.0 mm, 7–8 d),
H. garnoti
(13.0 mm, 6 d), and
H.
radiatus
(12.0 mm, 8 d), but not
H. bivittatus
(11.0–
11.4 mm, 5–8 d).
Interannual variability in the recruitment of the five
Halichoeres
species differed among species (Fig. 3).
Halichoeres bivittatus
was consistently one of the most
abundant labrids, and recruited regularly each year.

May 1997 183
RECRUITMENT PATTERNS OF WRASSES
T
ABLE
3. Larval durations† (mean
6
1
SE
) and coefficients
of variation (
CV
) for eight labrids collected from Barbados
during three spring seasons (1990–1992).
Species Year
Larval
duration
(d) Range
CV
n
H. bivittatus
1990 23.5
6
0.3 20–27 7.4 38
1991 26.8
6
0.6 21–32 10.8 25
1992 24.5
6
0.3 21–29 6.3 31
H. radiatus
1990 25.3
6
0.2 23–27 4.7 24
1991 23.9
6
0.4 22–28 6.2 18
1992 23.4
6
0.4 20–28 7.9 24
H. poeyi
1990 25.9
6
0.4 22–29 6.1 17
H. garnoti
1990 26.0
6
0.5 20–31 8.7 25
1991 24.8
6
0.7 21–30 11.3 17
1992 23.8
6
0.5 21–28 8.7 19
H. pictus
1990 26.7
6
0.5 23–30 7.7 18
1992 26.2
6
0.4 23–31 7.3 25
H. maculipinna
1992 27.0
6
0.7 23–37 13.4 25
T. bifasciatum
1990 49.0
6
1.1 38–64 12.2 32
1991 53.2
6
2.5 41–94 24.3 26
1992 49.2
6
0.8 41–66 10.0 37
B. rufus
1991 40.0
6
1.0 32–52 12.2 25
1992 34.3
6
0.3 31–41 4.4 22
† We added2dtoeach count of pre-settlement increments
to account for time to hatching.
Halichoeres radiatus
was less abundant than
H. bivitta-
tus,
but new recruits also were consistently present on
the reef each year. The least predictable of all the la-
brids,
H. poeyi,
was quite rare on the reefs, recruiting
in a strong pulse only in 1990. Both
H. garnoti
and
H.
pictus
were typically less abundant on the reefs than
either
H. bivittatus
or
H. radiatus,
but their temporal
patterns of recruitment were consistent among years.
Halichoeres garnoti
tended to recruit in rather small
events spread over several days during all three years,
while most of the recruitment of
H. pictus
occurred
during a single pulse of short duration each year (Fig.
3).Although there was interannual variability in the size
of recruitment events for many species, the timing of
these events was generally consistent among years for
each species. For these five
Halichoeres
species,
monthly patterns of recruitment were similar: pulses
occurred during the time of the new moon. Because
the lunar and tidal amplitude cycles are closely coupled
in Barbados, particularly during our sampling season,
pulses of recruitment occurred during maximum am-
plitude tides associated with the new moon (Table 5,
Fig. 4).
Third-quarter moon recruits:
Halichoeres
maculipinna, Thalassoma bifasciatum,
and
Bodianus rufus
Larval durations of
Halichoeres maculipinna, Thal-
assoma bifasciatum,
and
Bodianus rufus
tended to be
both longer and more variable than larval durations of
the first five
Halichoeres
species (new-moon group;
Table 3). While mean larval durations of
T. bifasciatum
were significantly longer than durations of all species
of
Halichoeres
(Dunn method:
Q
5
4.77
2
9.46,
P
,
0.05), mean larval durations of
H. maculipinna
tended
to be but were not significantly longer than the other
Halichoeres
species (Dunn method:
Q
,
3.17,
P
.
0.05; Table 3). Mean larval durations of
B. rufus
were
also consistently longer than
Halichoeres
species, but
due to high variability, this tendency was not always
significant (Dunn method:
Q
5
4.01
2
6.12,
P
,
0.05).
Variability in the larval durations of
H. maculipinna,
T. bifasciatum,
and
B. rufus
(1991) was generally quite
high (Table 3). Of all of the labrids,
T. bifasciatum
had
the highest variation in larval durations (1991
CV
5
24.3). Both the high mean larval duration and the high-
er variability in durations that year were due to the
presence of 4 fish (out of 26 dissected) whose otoliths
had numerous narrow increments immediately prior to
settlement, suggesting a delay of metamorphosis (Vic-
tor 1986
b
, Cowen 1991). Among all the other labrids
dissected, only
H. maculipinna
and
B. rufus
exhibited
evidence of delay of metamorphosis (6 out of 25
H.
maculipinna
dissected in 1992, and 1 out of 25
B. rufus
dissected in 1991, resulting in high
CV
s for each in
those years; Fig. 2). These are relatively conservative
estimates of the proportion of fish that delayed meta-
morphosis because only those with additional, narrow-
er increments on the outer edge of the presettlement
region of their otoliths (interpreted as reflecting re-
duced growth, Victor 1986
b
, Cowen 1991) were con-
sidered to be delayers. Other individuals with longer
larval durations may also be delayers without exhib-
iting these narrower increments (Sponaugle and Cowen
1994).
The sagittae of all of the labrids except
B. rufus
were
all rather similar in appearance, with relatively narrow
presettlement increments separated from the wider
postsettlement increments by a wide discontinuous
zone. This zone has previously been interpreted as the
period following settlement when the settled larvae are
buried under the sand undergoing metamorphosis (Vic-
tor 1983
b
). Because this zone was not evident in the
otoliths of
B. rufus,
this wrasse presumably settles di-
rectly to the reef, without burrowing in the sand. In
fact, several very small specimens of
B. rufus
in various
stages of larval to juvenile coloration were collected
in the water column near large coral heads, suggesting
that metamorphosis is gradual and occurs while these
fishes remain above the substrate.
Mean estimated sizes at settlement for
H. maculi-
pinna, T. bifasciatum,
and
B. rufus
were all at the small-
er end of the size range (Fig. 5), although
H. maculi-
pinna
and
B. rufus
were not significantly different from
each other or the other
Halichoeres
species (
T
9
method:
d
5
0.05
2
3.89, msd
5
1.70
2
4.95; Table 4).
Thalas-
soma bifasciatum
tended to be smaller at settlement
than most of the other labrids and was significantly
smaller than
H. bivittatus
(1990, 1991),
H. radiatus
(1990), and
H. poeyi
(
T
9
method:
d
5
2.67
2
4.00, msd
5
2.45
2
2.95). The size of the smallest fish collected
of all three species was also smaller than any of the

184
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
F
IG
. 2. Frequency of larval durations for eight species of labrids collected at Barbados during three seasons.
other
Halichoeres
species:
H. maculipinna
(10.8–11.0
mm, 3–6 d old, postsettlement),
T. bifasciatum
(10.0
mm, 4–10 d old), and
B. rufus
(9.7 mm, 2 d old). In
addition, late-stage larvae of
H. maculipinna
(mean SL
5
11.4 mm,
SD
5
0.73 mm, range
5
10.1–12.9 mm,
n
5
18) and
T. bifasciatum
(mean SL
5
10.8 mm,
SD
5
0.18 mm, range
5
10.6–11.0 mm,
n
5
4) were small-
er than
H. bivittatus
larvae (mean SL
5
12.1 mm,
SD
5
0.71 mm, range
5
10.8–13.7 mm,
n
5
33). Note
that late-stage larvae were often larger than estimated
settlement sizes and the size of the smallest juveniles.
This is probably due to shrinkage during metamorpho-
sis (see
Discussion
). In general, labrids recruiting dur-
ing the third-quarter moon tended to be smaller at set-
tlement than several other
Halichoeres
species. Fur-
thermore, and in contrast to the new-moon recruits,
third-quarter-moon recruits exhibited relatively greater
variation in larval duration than in estimated size at
settlement (Fig. 5). Following settlement, growth rates
for newly settled juvenile
H. maculipinna, T. bifascia-
tum,
and
B. rufus
were similar to one another and to
growth rates of the other
Halichoeres
species (
T
9
meth-

May 1997 185
RECRUITMENT PATTERNS OF WRASSES
T
ABLE
4. Estimated juvenile growth rates and size at set-
tlement (mean
6
1
SE
) from reduced major axes (RMA)
length-on-age regressions for eight labrids.
Species Year
Juvenile
growth rate
(mm/d)
Size at
settlement
(mm)
H. bivittatus
1990 0.38
6
0.02 10.60
6
0.45
1991 0.39
6
0.03 10.53
6
0.46
1992 0.45
6
0.03 9.30
6
0.60
H. radiatus
1990 0.38
6
0.03 10.79
6
0.42
1991 0.38
6
0.08 10.31
6
0.87
1992 0.50
6
0.05 8.15
6
0.99
H. poeyi
1990 0.39
6
0.03 11.86
6
0.49
H. garnoti
1990 0.55
6
0.06 10.87
6
0.87
1991 0.48
6
0.07 11.75
6
0.90
1992 0.50
6
0.05 10.47
6
0.66
H. pictus
1990 0.73
6
0.04 8.79
6
0.48
1992 0.63
6
0.04 9.52
6
0.53
H. maculipinna
1992 0.41
6
0.04 9.13
6
0.58
T. bifasciatum
1990 0.41
6
0.03 7.86
6
0.50
1991 0.34
6
0.04 8.84
6
0.59
1992 0.37
6
0.02 8.79
6
0.34
B. rufus
1991 0.62
6
0.06 8.99
6
0.63
1992 0.49
6
0.03 9.96
6
0.30
od:
d
5
0
2
0.35, msd
5
0.1
2
0.4; Table 4), except for
H. pictus.
Interannual temporal patterns of recruitment varied
among
H. maculipinna, T. bifasciatum,
and
B. rufus
(Fig. 3). By far one of the most abundant fishes on
Barbados reefs,
T. bifasciatum
recruited consistently in
pulses of a similar magnitude to
H. bivittatus.
In con-
trast,
Halichoeres maculipinna
and
B. rufus
juveniles
were much less abundant and exhibited greater inter-
annual variability in patterns of recruitment. After poor
recruitment seasons in 1990, both species were more
abundant in 1991 and 1992. During these years,
H.
maculipinna
typically recruited in pulses spread over
a number of days, while recruitment of
B. rufus
oc-
curred as relatively short events.
For all three species,
T. bifasciatum, B. rufus,
and
to a lesser degree
H. maculipinna,
monthly recruitment
pulses occurred during the third-quarter moon and min-
imum amplitude tides (Table 5, Fig. 4). This pattern
was consistent among years for both
T. bifasciatum
and
B. rufus
(Table 5, Fig. 3). However, in 1992, recruit-
ment of
H. maculipinna
was not as tightly coupled to
the third-quarter moon, but was spread over a broader
period of time (Fig. 3).
Mortality rates
For the three most abundant labrids,
H. bivittatus,
H. radiatus,
and
T. bifasciatum,
initial juvenile den-
sities
ø
1 wk following a major recruitment peak ranged
from 0.6 to 3.8 fishes per quadrat (5 m
2
; Table 6). Over
the next 2 wk, juveniles exhibited daily mortality rates
of 1.4 to 13.1%, based on a loss of 0.5–1.9 juveniles
per quadrat (Table 6). There wasno correlation between
these daily mortality rates and initial recruitment den-
sities, thus no evidence of density-dependent mortality
during this time period.
Spatial patterns of recruitment to Barbados
Examination of species-specific juvenile densities
over the eight sites along the west coast of Barbados
revealed a consistent pattern (Fig. 6). Although this
trend was significant for only two individual recruit-
ment events (Fig. 6a), in general, recruitment of the
three most common wrasses (
H. bivittatus, H. radiatus,
and
T. bifasciatum
) tended to be lower in the central
region of Barbados (site 6) than at either the northern
or southern ends (Fig. 6b). When the other, less abun-
dant labrids were considered together (over all census
seasons) for qualitative purposes, a similar pattern was
revealed, with most juveniles censused at either end of
the island, particularly the northern end, rather than in
the center (Fig. 7). Thus, although there was relatively
little significance in the spatial pattern of recruitment,
there was a consistent tendency for juvenile densities
to be reduced at site 6.
Spatial variation in environmental characteristics
To examine whether the observed spatial patterns of
recruitment were influenced by habitat selection, we
analyzed two environmental components: habitat (sub-
strate and cover) characteristics and densities of resi-
dent mature fishes. Of the 10 habitat variables consid-
ered, 6 varied significantly among sites (Table 7). Be-
cause juvenile densities were lower at site 6, we will
focus our discussion of particular habitat variables on
identifying differences between that site and the others.
Overall, site 6 had the highest percentage of live coral,
but it was only significantly higher than sites 7 and 3
(Tukey-type:
q
5
5.9,
P
,
0.005). Coral diversity was
also higher at site 6 than at sites 7, 3, or 8 (Tukey-
type:
q
5
4.97
2
6.83,
P
,
0.05
2
0.001). Not surpris-
ingly, site 6 had the lowest percentage of sand cover,
but this was only significantly lower than site 3 (Tukey-
type:
q
5
5.81,
P
,
0.05). Although not the deepest
site, site 6 was significantly deeper than sites 7, 2, and
8 (Tukey:
q
5
4.57–4.95,
P
,
0.05).
Principal components analysis (PCA) of the same
habitat variables revealed that sites 3, 7, and 8 could
be distinguished from the other sites based on PCA
factor 1; factor 2 did not seem to contribute to grouping
sites 3, 7, and 8, but may help separate site 1 from the
others (Fig. 8). There was less of a distinction among
sites based on PCA factor 3 (Fig. 8). PCA factors 1, 2
and 3 explained 36.7%, 28.0%, and 18.6% of the vari-
ance, respectively (Table 8).
In general, there was little significant influence of
habitat on the spatial patterns of recruitment for most
of the labrids, although in a few cases specific habitat
requirements may have influenced the distribution of
juveniles. The regressions of juvenile densities (means
following each major event for the common species,
and totals for each of the rarer species) against each
PCA factor indicated that there were no significant re-
lationships between juvenile densities and any of the
habitat variables (
r
2
5
0.001
2
0.47,
P
.
0.05), except

186
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
→
Fig. 3. Recruitment by eight species of labrids collected at Barbados during each ofthree spring recruitment seasons(1990–
1992). Lunar phase is indicated by solid (new moon) and open (full moon) circles. Note that
y
-axes vary for certain species
and years.
for
H. bivittatus
(1992
a
) and
B. rufus
(total). Site-spe-
cific densities of
H. bivittatus
following the first major
recruitment event in 1992 were significantly influenced
by PCA factor 1 (
r
2
5
0.70,
P
5
0.01; Fig. 9), indicating
a negative correlation with the percentage and diversity
of live coral, and a positive relationship with the per-
centage of rubble substrate, algal turf and sand cover,
and sand cover alone. Juvenile densities of
B. rufus
over all years were significantly influenced by PCA
factor 3 (
r
2
5
0.65,
P
5
0.016; Fig. 9), indicating a
positive relationship with mean depth and the per-
centage of algal turf cover.
Although there were few correlations between ju-
venile densities and among-site habitat characteristics,
on a smaller scale, several within-habitat differences
were noted in the distribution of each species. Palest
in coloration of all the labrid species considered in this
study,
H. bivittatus
juveniles were most commonly
found in small groups (but also in large schools fol-
lowing strong recruitment pulses) swimming relatively
low in the water column over sand–rubble substrates.
Juveniles of
H. radiatus
and
H. poeyi
were almost al-
ways solitary and occurred near the sand–coral inter-
face, under large coral overhangs.
H. garnoti
juveniles
were fairly widespread, occurring most often as solitary
individuals, but also in small, mixed schools near the
sand–coral interface.
H. pictus
juveniles were always
found in small schools (some mixed), relatively high
in the water column near areas of high vertical relief.
Newly settled juveniles of
H. maculipinna
and
T. bi-
fasciatum
were similar in their distribution (and rough-
ly, in their coloration), occurring as solitary individuals
very low in coral crevices. Over time,
T. bifasciatum
juveniles swam progressively higher in the water col-
umn, forming increasingly larger schools. Older
H. ma-
culipinna
juveniles tended to remain more solitary.
B.
rufus
juveniles were found exclusively around large
coral heads (such as
Montastrea cavernosa
), located at
the offshore end of coral spurs. Solitary or small groups
remained fairly high in the water column around the
coral structure.
There were only a few site-specific differences in the
abundance of resident mature fishes, and these differ-
ences did not seem to significantly influence the den-
sities of juveniles. Two-way ANOVAs for each variable
(mean number and diversity of all fishes, and mean
number of labrids censused) indicated a significant in-
fluence of site (
F
5
2.94–3.90,
P
5
0.001–0.009, df
5
7, 80), but not year (
F
5
0.001–0.107,
P
5
0.74–
0.98, df
5
1, 80), with no significant interaction (
F
5
0.90–1.02,
P
5
0.43–0.51, df
5
7, 80). In general, the
mean number and diversity of all mature fishes was
similar across all sites (ANOVA:
F
5
1.28–1.93,
P
5
0.09–0.29, df
5
7, 40; Fig. 10). The high number of
mature fishes at site 2 in 1992 was due to the presence
of several very large schools of grunts (Haemulidae).
In 1991, the mean number of mature labrids was sig-
nificantly lower at site 6 than three other sites (Tukey:
q
5
4.57–4.95,
P
,
0.05). However, there were no
significant relationships between mature fish counts
and any habitat variables (
r
2
5
0.001–0.47,
P
.
0.05),
or between juvenile densities of
H. bivittatus
and
T.
bifasciatum
and mature fish densities for a given season
(
r
2
5
0.001–0.086,
P
.
0.05).
D
ISCUSSION
Temporal recruitment patterns and early
life history characteristics
For most of the wrasses in this study, recruitment to
Barbados occurs rather predictably during the spring.
While the magnitude of recruitment events is somewhat
variable, the timing of these events is generally con-
sistent within seasons and among years. Eight species
of labrids exhibit two distinct temporal patterns of re-
cruitment, and these appear to be correlated with sev-
eral early life history traits. The recruitment of five
species (
Halichoeres bivittatus, H. radiatus, H. poeyi,
H. garnoti,
and
H. pictus
) was closely coupled to the
new moon and maximum amplitude tides, while re-
cruitment pulses of three other labrids (
H. maculipinna,
Thalassoma bifasciatum,
and
Bodianus rufus
) were
synchronized to the minimum amplitude tides during
the third-quarter moon. Because these recruitment pat-
terns were back-calculated from biweekly collections
of juveniles, it is possible that high daily mortality rates
of new settlers have obscured real settlement patterns.
This effect might be particularly suspect if we obtained
a semilunar (biweekly) recruitment pattern. However,
we obtained two distinct lunar recruitment patterns for
species collected during the same biweekly census.
Furthermore, data from a concurrent study on another
reef fish (
Stegastes partitus
) confirm that similar back-
calculated lunar recruitment patterns exactly matched
patterns of larval supply (Sponaugle and Cowen
1996
b
).
Several early life history traits were correlated with
these temporal recruitment patterns. Duration of the
larval period tended to be longer and more variable for
labrids recruiting during the third-quarter moon. This
high variability in larval duration reflects a capacity to
delay metamorphosis. While this capacity has been re-
ported previously for
T. bifasciatum
(Victor 1986
b
),
and a temperate labrid (Cowen 1991), this is the first

May 1997 187
RECRUITMENT PATTERNS OF WRASSES

188
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
T
ABLE
5. Rayleigh test statistics for lunar periodicity in settlement of eight labrids for each
year that juveniles were collected and for all years summed together. Because the tidal
amplitude cycle is closely coupled to the lunar cycle throughout the study period, the timing
of settlement relative to the tidal amplitude cycle is similar.
Species Year
nZ
†
Lunar
settlement
day‡
s
§
H. bivittatus
1990 314 111.1*** 2.5 4.2
1991 100 21.2*** 1.1 4.9
1992 242 11.1*** 0.0 5.9
All years 656 118.2*** 1.8 5.0
H. radiatus
1990 104 38.6*** 29.3 4.2
1991 40 11.2*** 29.1 4.6
1992 72 7.3*** 29.3 5.5
All years 216 53.5*** 29.3 4.7
H. poeyi
1990 23 14.3*** 0.6 3.1
1992 3 1.4
NS
All years 26 15.6*** 0.7 3.2
H. garnoti
1990 29 8.2*** 0.3 4.6
1991 19 5.9** 2.0 4.4
1992 17 2.9
NS
All years 65 16.4*** 0.9 4.8
H. pictus
1990 82 46.3*** 4.1 3.3
1991 10 9.9*** 19.5 0.5
1992 18 15.2*** 28.7 1.9
All years 110 34.8*** 2.9 4.4
H. maculipinna
1990 1 1.0
NS
1991 21 8.8*** 21.5 3.9
1992 52 1.7
NS
All years 74 5.2** 19.5 5.7
T. bifasciatum
1990 207 25.3*** 25.3 5.4
1991 111 29.7*** 20.1 4.6
1992 478 51.8*** 18.1 5.4
All years 796 68.2*** 20.1 5.6
B. rufus
1991 30 27.5*** 21.8 1.4
1992 18 12.9*** 25.7 2.6
All years 48 34.3*** 23.1 2.6
†
Z
5
Rayleigh test statistic; **
P
,
0.01, ***
P
,
0.001,
NS
,
P
.
0.05.
‡ Lunar settlement day was calculated from mean vector angle.
§
s
5
mean angular deviation (days; Batschelet 1981, Zar 1984); day 1
5
new moon.
report of delayed metamorphosis by
H. maculipinna
and
B. rufus.
For all species, mean larval durations
were comparable to those reported previously for la-
brids from Panama (
Halichoeres
spp. 24.1–25.9 d,
T.
bifasciatum
49.3 d,
B. rufus
41.6 d; Victor 1986
c
) and
Bermuda (
Halichoeres
spp. 22–30.5 d,
T. bifasciatum
55 d,
B. rufus
32 d; Schultz and Cowen 1994).
Labrids recruiting during the third-quarter moon also
tended to be smaller at settlement than the other
Hal-
ichoeres
species. The only previously reported sizes at
settlement are for
T. bifasciatum
and
H. poeyi
(each
ø
12 mm, Victor 1991;
T. bifasciatum
mean length of
11.5 mm, Robertson 1992). Although our estimated
settlement sizes are similar for
H. poeyi, T. bifasciatum
appears to settle to Barbados at much smaller sizes
(7.9–8.8 mm). Our estimated sizes at settlement are
also smaller than the sizes of three labrids collected in
light traps during the same period. Based on obser-
vations of
H. maculipinna
and several other labroid
species, we hypothesize that during metamorphosis,
some shrinkage occurs and juveniles emerge at a small-
er standard length than the presettlement larvae (S.
Sponaugle and R. K. Cowen,
unpublished data
). How-
ever, it is also possible that growth is simply curvilinear
during metamorphosis, so that the linear model used
to estimate size at settlement results in a smaller es-
timate of size.
Considering that
T. bifasciatum
spends a signifi-
cantly longer time in the plankton prior to settlement,
yet settles at a smaller size than many of the other
labrids, and assuming that size at hatching is similar
for all species, mean larval growth rates must be sub-
stantially less for
T. bifasciatum
than for the other la-
brids. Similarly,
B. rufus
probably has an intermediate
larval growth rate, with several
Halichoeres
species
exhibiting the highest larval growth rates. Following
settlement, however, most juvenile labrids grew at sim-
ilar rates (Table 4), which were slightly to substantially
higher than growth rates previously published for
H.
bivittatus, H. poeyi,
and
T. bifasciatum
(
ø
0.30 mm/d;
Victor 1991). Clearly, there is a decoupling of growth
between the larval and juvenile phases of the life cycle.

May 1997 189
RECRUITMENT PATTERNS OF WRASSES
F
IG
. 4. Distribution of recruitment over the lunar cycle for eight species of labrids collected at Barbados over three spring
recruitment seasons combined (1990–1992; total of 10 lunar cycles). Lunar phase is indicated by solid (new moon) and open
(full moon) circles.
Processes operating in the pelagic environment appear
to select for a range of larval growth rates, whereas
reef-based constraints appear to select for more equiv-
alent rates of growth. The relationship between larval
life history strategies, recruitment patterns, and juve-
nile life history traits is probably complex, but its elu-
cidation is critical to a complete understanding of the
population dynamics of organisms with complex life
cycles.
Thus, the labrids in this study with relatively short,
invariant larval durations (
H. bivittatus, H. radiatus,
H. poeyi, H. garnoti,
and
H. pictus
) recruited to Bar-
bados during the new moon and maximum amplitude
tides, while those with generally longer or more vari-

190
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
F
IG
. 5. Mean estimated settlement size vs. mean larval
duration for eight labrids with relatively invariant larval du-
rations (
Halichoeres poeyi, H. garnoti, H. bivittatus, H. ra-
diatus,
and
H. pictus
; shaded circles) and for those with more
variable larval durations (
H. maculipinna, T. bifasciatum,
and
B. rufus
; solid circles). Error bars indicate
6
1
SD
. Note that
error bars are not available for
H. poeyi
and
H. maculipinna,
since data from only one year were used to estimate size at
settlement.
T
ABLE
6. Mean densities and mortality estimates (
6
1
SD
) following large recruitment pulses
for three common species of labrids recruiting to Barbados during three spring seasons.
Means were calculated for each site separately; means for all sites are presented.
Year
Initial
density
(no. fish/
quadrat)†
No. days
after
peak‡
Final
density
(no. fish/
quadrat)†
No. days
after
peak‡
Mean
decrease
(no. fish/
quadrat)
Daily
mortality
(%)§
H. bivittatus
1990 3.8
6
2.2 7 1.9
6
1.0 20 1.9
6
2.5 5.3
1991 0.6
6
0.6 6 0.1
6
0.2 23 0.5
6
0.6 11.3
1992a 2.1
6
2.0 7 0.8
6
0.9 22 1.2
6
1.4 5.9
1992b 1.2
6
1.0 7 0.1
6
0.2 23 1.1
6
0.9 13.1
H. radiatus
1990 1.3
6
0.7 10 0.7
6
0.4 23 0.7
6
0.6 5.0
T. bifasciatum
1990 3.0
6
1.1 17
\
1.6
6
0.5 38 1.4
6
1.5 3.0
1992 3.7
6
1.7 8 2.9
6
1.5 25 0.8
6
1.3 1.4
† Initial densities were calculated from the first census following the event; final densities
were calculated from the second census following the event (number censused/5-m
2
quadrat).
‡ Number of days between peak in recruitment event and midpoint of census period.
§ Calculated from Ricker (1975).
\
Although time after peak is long for
T. bifasciatum
in 1990, this particular event was quite
broad and extended up to the day prior to the census (see Fig. 3).
able larval durations, with a demonstrated capacity to
delay metamorphosis (
T. bifasciatum, H. maculipinna,
and
B. rufus
), recruited during the third-quarter moon
and minimum amplitude tides. In addition, several of
these later species (
T. bifasciatum
and
B. rufus
) may
have slower larval growth rates, and may tend to be
relatively small at settlement. Consideration of growth
rates and settlement sizes places
H. pictus
slightly out-
side the two groups because after a rather short, in-
variant period in the plankton,
H. pictus
appears to
settle at a smaller size, and then exhibit faster juvenile
growth rates than the others. Interestingly, the temporal
pattern of settlement exhibited by
H. pictus
is also
slightly offset from (3–4 d after) the new moon.
At this time, the significance of the correlation be-
tween early life history traits and recruitment patterns
is unclear. Why should a longer larval duration, a ca-
pacity to delay metamorphosis, and a possibly smaller
size at settlement lead to recruitment timed to the third-
quarter moon and minimum amplitude tides? Prior to
examining this issue, we need to explore the causes
and potential advantages of synchronized recruitment.
The comparison of recruitment patterns among these
closely related labrids may clarify or eliminate several
of the alternatives.
Synchronized recruitment may be a result of either
active (behavioral) or passive (physical) processes act-
ing on the early egg and larval stages, the late-stage
larvae, or on the new recruits. Several investigators
have proposed that recruitment synchrony results from
synchronous spawning, which occurs to enhance trans-
port, survival, or growth of the eggs and larvae (Jo-
hannes 1978, Thresher 1984, Gladstone and Westoby
1988; see review in Robertson 1991), or survival of
the adults (Robertson et al. 1990). Temporal coupling
between spawning and recruitment is more likely for
species with short, invariant larval durations, such as
pomacentrids (e.g., Robertson et al. 1988, Meekan et
al. 1993) and is less likely for the labrids in this study
because spawning is asynchronous (daily, year-round;
Warner and Robertson 1978, Robertson 1981), and lar-
val durations of several species are relatively long and
variable. At present, there are no reported data on
monthly variation in labrid fecundity; until such data
are available, we will assume that cyclic patterns of
fecundity do not contribute significantly to patterns of
reproduction. While the frequency of spawning may
vary seasonally (Hunt von Herbing and Hunte 1991),
and fertilization success may be reduced in rough seas

May 1997 191
RECRUITMENT PATTERNS OF WRASSES
F
IG
. 6. Site-specific juvenile densities for three common labrids recruiting to Barbados. (a) Mean densities (
6
1
SE
)
following each major recruitment event each year. Significant differences among sites are indicated in the inset (Tukey-type:
q
5
4.33
2
5.37,
P
,
0.05). (b) Mean rank of each site for all three species and events. See Fig. 1 for site locations.
(Petersen et al. 1992), these factors likely contribute
to seasonal rather than monthly cycles of reproduction.
For daily spawners, periodic, synchronous patterns
of recruitment suggest that the survival of late-stage
larvae and juveniles may be temporally predictable. For
such species, a variable larval duration may enable
greater synchronization of recruitment (Sponaugle and
Cowen 1994). In a concurrent study, we demonstrated

192
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
F
IG
. 7. Site-specific number of juveniles collected during
three seasons of biweekly censuses. For qualitative purposes,
the total number of juveniles censused was tabulated for the
five rarer labrids. Solitary juveniles were censused in 5-m
2
quadrats (a), and schooling species were censused in 25-m
2
quadrats (b).
that temporal patterns of recruitment to Barbados re-
flect variability in the supply of late-stage larvae as
measured by light traps (Sponaugle and Cowen 1996
b
).
Thus, the timing of recruitment is not simply the result
of differential survival of continuously settling larvae,
and probably reflects either predictable processes act-
ing on the pelagic stage, or cued behavior to enhance
the growth and survival of new recruits (e.g., Kingsford
1980). For example, pulses of late-stage larvae may
form as a result of differential growth and survival rates
cued by variable patterns of food availability, mortality,
or transport by currents. Similarly, settlement may be
behaviorally timed to coincide with cycles of prey
availability on the reef, periods of reduced predation
on juveniles, or particular flow conditions.
At present, we have no data on food limitation of
tropical larvae or new recruits, thus our discussion will
focus on the potential effects of predation and transport
on the timing of settlement. Recently, piscivorous reef
fishes have been shown to reduce the survivorship of
recently settled
H. pictus
and
Chromis cyanea
(Po-
macentridae), but not
T. bifasciatum
(Carr and Hixon
1995). Settling larvae may be even more vulnerable to
predation; settlement at night, particularly during the
darkest nights, has often been suggested to be the result
of selection to reduce predation by reef organisms (e.g.,
Hobson and Chess 1978, Johannes 1978). However,
although five of the labrids in our study settle during
the new moon, three labrids settle during the third-
quarter moon following the full moon, hardly the
darkest (nor the brightest) time of the month. Thus,
while predation may ultimately drive the timing of re-
cruitment for some species, it is obviously not the sole
selective pressure.
Monthly patterns in the supply of late-stage larvae
may be due to variable transport during particular times
of the tidal cycle. For example, transport of late-stage
larvae to reefs may be enhanced when strong onshore
flood tides occur at night. Alternatively, late-stage lar-
vae may be able to swim onshore more successfully
during other times of the tidal cycle (e.g., when currents
are weakest). It is also possible that detection of a
suitable settlement site may only occur during partic-
ular flow regimes (e.g., offshore flow over the reef).
Because labrids exhibit two distinct patterns in the
timing of their recruitment (i.e., during very different
stages of the tidal cycle), it is likely that these species
vary with respect to their behaviors during approach
and settlement. For example, swimming ability is prob-
ably linked to a certain degree with size prior to set-
tlement. If labrids actively swim to reefs during set-
tlement, smaller species such as
T. bifasciatum, H. ma-
culipinna,
and to a lesser degree
B. rufus
may be rel-
atively slower or weaker swimmers compared to the
larger species, and may be more successful at swim-
ming during periods of low current flow (i.e., minimum
amplitude tides during the quarter moons). This selec-
tive pressure may be balanced by the pressure of vi-
sually oriented predators. Because predation is prob-
ably reduced on dark, moonless nights (i.e., new
moons), recruitment may be most successful during the
time of the third-quarter moon (minimum amplitude
tides and nights becoming darker). Relatively larger
species such as
H. bivittatus, H. radiatus, H. poeyi,
and
H. garnoti
may be more successful swimming against
strong currents, and therefore pulses of recruitment can
be more closely coupled to the darkest nights around
the new moon. Furthermore, as Carr and Hixon (1995)
demonstrate (but in contrast to what size at settlement
would suggest), species such as
H. pictus
may be more
susceptible to predation than species such as
T. bifas-
ciatum,
thus settlement during the darkest nights may
be more important for those species.
Finally, other differences in larval behavior may in-
teract with transport processes and predation pressures
to influence the timing of settlement. For example, the
depth at which larvae approach reefs may vary among
species (Leis et al. 1996), resulting in differences in
tidally mediated transport, and possibly in larval vul-

May 1997 193
RECRUITMENT PATTERNS OF WRASSES
T
ABLE
7. Habitat characteristics of eight sites along the west coast of Barbados. Site 1 is the
northernmost site; site 8 is southernmost (see Fig. 1). Where significant differencesoccurred
among sites, values for each site are presented, with solid underlines connecting sites with
no significant differences. There were no significant differences among sites for mean per-
centage coral rock substrate, rubble substrate, sand substrate, or algal turf cover.
Mean percentage live coral substrate (PLC)
Site 7 3 854126
PLC 4.2 4.2 10.0 15.0 18.0 19.0 24.0 26.0
Mean diversity live coral (DLC)
Site 7 3 852461
DLC 2.5 4.2 4.2 6.7 8.4 13.4 13.4 15.0
Mean percentage algal turf and sand cover (ATS)
Site 4 5 236781
ATS 2.5 8.3 9.2 11.7 11.7 31.7 34.7 48.3
Mean percentage sand cover (SC)
Site 6 5 817243
SC 1.7 4.2 7.5 8.3 10.8 10.8 15.0 24.2
Mean depth (m)
Site 7 2 854631
Depth 2.97 3.26 3.42 4.28 4.31 4.49 4.95 5.26
Mean rugosity
Site 1 8 734625
Rugosity 0.52 0.64 0.66 1.21 1.34 1.40 1.57 1.80
nerability to predators. Differences in larval swimming
ability, position in the water column, and susceptibility
to predation may account for the two patterns of re-
cruitment by the labrids in this study; however, it re-
mains unclear how other early life history traits, such
as the capacity to delay metamorphosis, influence re-
cruitment. The capacity to extend larval life until a
suitable settlement site is found may be particularly
important for species with long early developmental
(i.e., precompetent) periods (Jackson and Strathman
1981, Cowen 1991). In addition, delay of metamor-
phosis would probably be most advantageous where
the successful settlement to the reef is less temporally
predictable. For example, depending on local flow con-
ditions, weaker swimmers may not always be able to
enter the reef environment. Such a constraint may se-
lect for the ability to successfully delay settlement and
metamorphosis. In light of this, it is interesting to note
that a diversity of other reef fishes and crustaceans
recruit to Barbados during the third-quarter moon and
neap tides (Sponaugle and Cowen 1996
a
). In general,
the relationship between a number of early life history
traits and patterns of subsequent recruitment warrants
further investigation (Cowen and Sponaugle,
in press
).
Proximate environmental cues
Proximate cues operating during settlement have
typically been difficult to distinguish, largely because
environmental cycles (i.e., lunar and tidal amplitude
cycles) are often temporally coupled at a single geo-
graphic site. Because the lunar cycle is the same ev-
erywhere, but the tidal amplitude cycle can shift among
geographical locations, intraspecific comparisons of re-
cruitment patterns among distant locations may poten-
tially isolate the relevant proximate cues operating dur-
ing settlement. Due to the ‘‘open’’ nature of coral reef
fish populations and the dynamics of large-scale current
flow in the Caribbean, most fish populations are not
likely to be genetically isolated (Mitton et al. 1989,
Lacson 1992, Shulman and Bermingham 1995). There-
fore, larvae at all locations should use the same cues
or suite of cues to time settlement. Geographic differ-
ences in the timing of recruitment should only result
from offset cycles of these cues. However, to date, few
such comparisons have been made.
Among labrids, patterns of recruitment that are syn-
chronized to the third-quarter moon and minimum am-
plitude tides have not previously been reported. Syn-
chronous recruitment cued to different phases of the
lunar cycle has been reported for a number of other
species from diverse locations. Lunar cycles of re-
cruitment are exhibited by several pomacentrids re-
cruiting to Pacific reefs (Kingsford 1980, Ochi 1985)
as well to as reefs off St. Thomas, U.S. Virgin Islands
(Booth and Beretta 1994), and Panama (Robertson et
al. 1988, Robertson 1992). In fact, many species re-
cruiting to Caribbean Panama have lunar-cyclic re-
cruitment patterns, with pulses occurring around the
new moon (in addition to a labrid and several poma-
centrids, several acanthurids, and a chaetodontid; Rob-
ertson 1992). Synchrony to the third-quarter moon is
less frequently encountered, and outside of Barbados,
these patterns are usually semilunar (a haemulid at St.
Croix, McFarland et al. 1985; and a pomacentrid re-

194
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
F
IG
. 8. Principal component analysis (PCA) factors de-
rived from habitat variables measured at each of the eight
census sites (indicated by numbers). See Table 8 for factor
loadings.
F
IG
. 9. (a) Densities of
Halichoeres bivittatus
juveniles
after a major recruitment pulse in 1990, and (b) total number
of
Bodianus rufus
juveniles collected over three seasons, plot-
ted against principal components analysis (PCA) factors 1
and 3, respectively.
T
ABLE
8. Principal components factor loadings for habitat variables.
Variable Factor 1 Factor 2 Factor 3
Mean percentage live coral
2
0.961 0.007
2
0.192
Mean diversity live coral
2
0.838 0.348 0.071
Mean percentage rubble 0.686
2
0.531
2
0.387
Mean percentage algal turf–sand 0.684 0.663
2
0.113
Mean percentage sand cover 0.629
2
0.128 0.396
Mean rugosity
2
0.531
2
0.655 0.484
Mean percentage sand substrate
2
0.118
2
0.687 0.329
Mean percentage coral rock 0.372 0.655 0.486
Mean depth
2
0.297 0.628 0.684
Mean percentage algal turf 0.429
2
0.459 0.669
cruiting to Panama, Robertson 1992). Recruitment syn-
chrony to the first-quarter moon has been reported for
a holocentrid recruiting to Panama (Robertson 1992),
and two pomacentrids recruiting to St. Thomas around
the time of the full moon (Booth and Beretta 1994).
Interestingly, Thorrold et al. (1994
a
) report that labrid
larvae appear in Bahamian channel nets with semilunar
frequency, but it is unclear how this periodicity relates
to recruitment patterns.
In an attempt to identify the most important proxi-
mate cues operating during settlement, we compared
temporal patterns of
T. bifasciatum
recruitment to Bar-
bados with previously published records from San Blas,

May 1997 195
RECRUITMENT PATTERNS OF WRASSES
F
IG
. 10. Mean number and diversity of mature fishes (
6
1
SE
) and mean number of mature labrids (
6
1
SE
) censused at
each of eight sites at Barbados during two sampling seasons (left column, 1991; right column, 1992). Significant differences
among sites are indicated in the inset.
Panama (Victor 1986
a
). Although both locations ex-
perience the same lunar phase, there are important geo-
graphical differences in the expression of the tidal am-
plitude cycle. Tides in both locations are mixed, but
the semi-diurnal signal is much greater at Barbados
relative to the more diurnal signal at Panama. More
importantly, the relationship between the synodic and
the declinational cycle differs between the two loca-
tions: tides at Barbados are less influenced by the dec-
linational cycle, thus the tidal amplitude cycle (29.4 d)
is in phase with the lunar cycle during much of the
year, including the study period. The lunar declination
cycle has greater influence in Panama, where the tidal
amplitude cycle is 27.3 d (see Barnwell 1976 for further
discussion). As a result, the tidal amplitude cycle in
Panama is only periodically in phase with the lunar
cycle (Fig. 11).
To determine whether
T. bifasciatum
recruitment is
synchronized to the lunar or tidal amplitude cycle, we
first compressed each data set into a single lunar cycle
(as before) and tested with Rayleigh tests (Batschelet
1981, Zar 1984); in a separate analysis, we repeated
the tests on each data set reduced into a single tidal
amplitude cycle (day 1 being the first day of maximum
amplitude tides). Because sampling in Panama was
conducted over the entire year, we further subdivided
these data into months where maximum amplitude tides
occurred at night, months when these tides occurred

196
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
F
IG
. 11. Annual tidal amplitude record for Barbados (up-
per curves) and Panama (lower curves), obtained from the
British Admiralty and from U.S. Department of Commerce
(NOAA) tidal charts, respectively. Lines indicate the 3-d
moving average of maximum amplitude tides occurring at
night (solid lines) and during the day (dashed lines). New
moons are indicated by solid circles.
T
ABLE
9. Rayleigh test statistics for tidal amplitude and lunar periodicity in recruitment of
Thalassoma bifasciatum
to Barbados (1990–1992) and Caribbean Panama (1981–1982). See
Table 5 for the test statistics for lunar periodicity of recruitment to Barbados.
Time period
nZ
†
Settlement
day
(tidal or
lunar)‡
s
Barbados—tidal
1990 218 23.0*** 25.7 5.4
1991 111 31.0*** 21.3 4.6
1992 482 48.7*** 20.2 5.5
All years 811 77.2*** 21.7 5.5
Panama—tidal§
Night maximum 102 11.2*** 12.5 5.0
Day maximum 257 13.7*** 5.3 5.4
Transition 344 88.7*** 12.5 4.3
Night and transition 446 97.4*** 12.5 4.5
Day and transition 601 53.3*** 11.1 5.2
Total 703 63.6*** 11.3 5.1
Panama—lunar§
Night maximum 69 2.1
NS
Day maximum 310 6.1** 13.4 6.2
Transition 307 92.6*** 3.5 4.5
Night and transition 376 86.9*** 3.5 4.8
Day and transition 617 37.1*** 4.7 5.8
Total 686 38.8*** 4.6 5.8
†
Z
5
Rayleigh test statistic; **
P
,
0.01, ***
P
,
0.001,
NS
,
P
.
0.05.
‡ Settlement day was calculated from mean vector angle;
s
5
mean angular deviation (days;
Batschelet 1981, Zar 1984); day 1 of tidal cycle
5
first day of maximum amplitude tide; day
1 of lunar cycle
5
new moon.
§ Data for
T. bifasciatum
recruiting to Panama were reanalyzed from Victor (1986
a
), and
split into months where maximum amplitude tides occurred at night, during the day, or both
(transition months).
during the day, and transition months when they oc-
curred during both daylight and nighttime (e.g., March,
April, September, and October; Fig. 11).
The timing of
T. bifasciatum
recruitment differs be-
tween Barbados and Panama. Reanalysis of Victor’s
(1986
a
) recruitment data over a lunar cycle indicated
that in general,
T. bifasciatum
recruited to Panama in
pulses occurring
ø
4–5 d after the new moon (Table 9,
Fig. 12). Based on the analysis of recruitment data over
a tidal amplitude cycle, the timing of
T. bifasciatum
recruitment also varied between the two locations. Be-
cause the tidal amplitude cycle in Barbados was in
phase with the lunar cycle during our study, recruitment
of
T. bifasciatum
was similarly synchronized to the
tidal amplitude cycle, occurring during minimum am-
plitude tides 21–22 d after the first day of maximum
amplitude tides (Table 9, Fig. 12). In Panama,however,
most
T. bifasciatum
recruitment occurred 11–12 d after
maximum amplitude tides. Overall recruitment of
T.
bifasciatum
to Panama was highly correlated with both
the tidal amplitude and lunar cycle, although in general,
synchronization to the tidal amplitude cycle may be
tighter than to the lunar cycle.
Because the moon is in the same phase at both geo-
graphical locations yet recruitment patterns differ,tides
must play a role either solely, or more likely, in concert
with the lunar cycle. A number of crabs possess spe-
cies-specific hierarchies of rhythms that enable the
tracking of several environmental cycles at the time of
larval release (Morgan and Christy 1994, Morgan
1995). Phase shifts of these cycles at different locations
result in variable temporal patterns of reproduction
(Morgan and Christy 1994). Similar responses by reef
fishes to the lunar and tidal amplitude cycles may create
complex variation in recruitment patterns at distant
geographical locations.

May 1997 197
RECRUITMENT PATTERNS OF WRASSES
F
IG
. 12. Distribution of recruitment over (a) the lunar and (b) the tidal cycles for
Thalassoma bifasciatum
at Panama
(from reanalysis of data from Victor 1986
a
), and over (c) the lunar and (d) the tidal cycle for
T. bifasciatum
at Barbados.
Lunar phase is indicated by solid (new moon) and open (full moon) circles. Day 1 of each tidal cycle corresponds to the
first day of maximum amplitude tides.
Seasonal shifts in the relation of the tidal amplitude
cycle to the lunar cycle may contribute to seasonal
variation in recruitment patterns at Panama. Most re-
cruitment of
T. bifasciatum
occurred during months
when maximum amplitude tides occurred during night
and day (transition months) and months when maxi-
mum tidal amplitudes occurred only during the day
(Table 9). The timing of recruitment events during the
months with daytime maximum tidal amplitudes dif-
fered from events during other times of the year: pulses
were less distinct and occurred just prior to the first
minimum amplitude tide, often near the full moon. In
fact, Robertson (1992) suggests that recruitment of
T.
bifasciatum
is semilunar cyclic for up to one-third of
the months during which significant recruitment oc-
curs. This pattern alone points to the importance of
tides in the timing of recruitment pulses.
Although data from Barbados indicate a distinct lu-
nar pattern of recruitment for
T. bifasciatum,
it is pos-
sible that these patterns shift seasonally in Barbados
as well. Such a shift may have led Hunt von Herbing
and Hunte (1991) to conclude that recruitment of
T.
bifasciatum
to Barbados is not lunar-cyclic, although
it is also possible that the lack of a clear signal may
have been due to their sampling of much larger (SL
.
30 mm) fishes. Even if patterns of recruitment to Bar-
bados shift seasonally, recruitment patterns in Panama
were most distinct during the transition months, and it
is during these months that recruitment patterns of
T.
bifasciatum
are in clear contrast to those in Barbados.
Nonetheless, Robertson’s (1992) point that long time-
series (at least year-round) measurements of recruit-
ment are necessary to obtain an accurate perspective
on synchrony is valid. In fact, by comparing intra-an-
nual shifts in the timing of recruitment to seasonal
shifts in cues such as the tides, further insight may be
gained into identifying the proximate cues involved.
In addition to the relationship between the lunar and
tidal amplitude cycles, the influence of the tidal cycle
on the nearshore environment may differ between Bar-
bados and Panama. The tidal range is greater in Bar-
bados (1.2 m) than in Panama (0.6 m) and may result
in stronger tidal flows. Interaction of these flows with
the prevailing currents may translate into rather strong
alongshore (north–south) currents. Perhaps due to this
stronger tidal influence, recruitment during minimum
amplitude tides is necessary for fishes that are weaker
swimmers. In Panama, lower amplitude tides may have
less overall effect on nearshore flow, and recruitment
during maximum amplitude tides may be more feasible.
However, the influence of tidal amplitude cycle may
shift seasonally as well. Strong trade winds during the
dry season (mid-December through mid-April) gen-
erate heavy wave action and northward-flowing near-
shore currents, in contrast to the wet season, where
winds and currents are typically lighter (Glynn 1972,
Robertson 1990). Such a seasonal change in flow con-
ditions may contribute to the shift in recruitment timing
between the months with daytime maximum amplitude
tides (during most of the dry season) and other months.

198
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
Seasonal shifts in nearshore currents and in the relative
phasing of the tidal amplitude cycle may also influence
the length and timing of the recruitment season, al-
though these seasons are also likely to be defined by
seasonal reproduction and the interaction of larval, ju-
venile, and adult biology with environmental cycles
(Robertson 1991).
Variability in the synchrony of other behaviors sug-
gests that tides are important in other aspects of reef
fish biology. In particular, the timing of spawning shifts
among sites where the relative influence of tides varies.
Several labrids spawn daily during ebb tides in Florida
(Thresher 1979) and in Barbados (Hunt von Herbing
and Hunte 1991), but not in Panama (Robertson 1981).
Similarly, at one site at Enewetak Atoll where tidal
flows are minimal, several labroids spawn at a fixed
time each day, while at a nearby site where tidal flows
are stronger, the timing of reproduction shifts daily
while tracking the tidal cycle (Colin and Bell 1991).
Our data suggest that tides can influence the timing of
recruitment for reef fishes, resulting in geographical
variation in recruitment synchrony among locations
where the relative strength or timing of tidal currents
varies.
Despite the fact that most benthic marine populations
are open and genetically mixed, behavioral and other
life history traits may vary significantly among pop-
ulations from widely separated geographic locations.
This variation is inherently introduced by local differ-
ences in the expression of proximate environmental
cues. Careful examination of these traits in the context
of variable environmental cues is clearly necessary
both in the identification of critical proximate cues as
well as in understanding ultimate constraints to adap-
tation. Further comparisons of behavioral life histories
among distant locations may reveal the mechanisms by
which open populations adapt to local environmental
conditions. While behavioral plasticity is one solution,
adaptation to common temporally persistent physical
or environmental processes (such as the tidal amplitude
cycle) may be more likely. Less predictable, episodic
features may lend a certain degree of noise to the sys-
tem, but are unlikely to underlie behavioral traits.
Spatial pattern of labrid recruitment to Barbados
Recruitment of labrids to Barbados occurred on an
island-wide scale: pulses of new recruits arrived con-
currently at all sites along the west coast of the island.
The overall pattern of lower juvenile densities at a cen-
tral site along the west coast does not appear to be
determined by large-scale habitat selection or by the
presence of resident fishes. Estimates of early juvenile
mortality for the most abundant labrids indirectly sug-
gest that spatial patterns of recruitment were probably
not greatly influenced by differential mortality rates of
new recruits. For the most common labrids, juvenile
densities 1 wk after recruitment were lower than the
estimated carrying capacity for mature
T. bifasciatum
on several Barbados reefs (0.8–1.2 fishes/m
2
; Hunt von
Herbing and Hunte 1991). In addition, there was no
correlation between daily mortality rates and initial ju-
venile densities, unlike those reported for older
T. bi-
fasciatum
in Barbados (0–60% monthly mortality, pos-
itively correlated with fish density; Hunt von Herbing
and Hunte 1991). While reported mortality rates are
typically highest during the first few days following
settlement and taper off with time (Victor 1986
a
, Sale
and Ferrell 1988), it is possible that juvenile
T. bifas-
ciatum
density in Barbados only impacts survivorship
at later ages.
Because early juvenile mortality is apparently not
density-dependent, and overall labrid recruitment ap-
pears to be unrelated to habitat features and densities
of mature fishes, spatial patterns of recruitment prob-
ably reflect variation in the supply of larvae. Lower
overall densities of labrids at the central site (site 6)
are likely the result of differential larval supply due to
variable nearshore tidal currents. During a concurrent
study on nearshore currents and patterns of larval sup-
ply at Barbados, we demonstrated that the east–west
component of nightly transport by tidal currents is di-
rected offshore at the central site, in contrast to gen-
erally onshore transport at the northern and southern
ends of the west coast (Sponaugle and Cowen 1996
a
).
Light traps deployed during this time period collected
a lower abundance and diversity of larval reef fishes
than traps from a northern and southern site (Sponaugle
and Cowen 1996
a
). Thus, offshore flow near the central
region of the west coast may result in a lower rate of
larval supply to these reefs.
Coupling of behavioral and physical processes can
occur on a variety of temporal and spatial scales to
influence patterns of larval supply and recruitment. The
intersection of these mechanisms must be understood
in order to fully define population boundaries, includ-
ing the identification of sources and sinks of larvae.
Due to consistent flow patterns, certain areas may al-
ways receive lower numbers of larvae. Additionally,
these same flow conditions may result in spatial vari-
ation in the reproductive contribution of adults. Be-
havioral responses to these physical features lend sig-
nificant complexity to the system and call into question
the degree to which particular populations are truly
open. Successful tracking of fish or invertebrate larvae
in the field has proved quite difficult (e.g., Stoner 1990;
G. P. Jones and M. J. Milicich,
unpublished data
), but
remains essential to fully understanding the population
dynamics of most benthic marine organisms.
Conclusions
The results of this study demonstrate that both phys-
ical and biological processes are important to the pop-
ulation replenishment of coral reef fishes. Spatial vari-
ation in physical phenomena such as tidal currents can
lead to consistent spatial differences in the supply of
larvae and subsequent distribution of juveniles. The

May 1997 199
RECRUITMENT PATTERNS OF WRASSES
tidal amplitude cycle also appears to be more important
than previously thought in the specific timing of re-
cruitment. Temporal patterns of recruitment are prob-
ably the result of a balance between tidal amplitude
and lunar cues at the time of settlement. However, the
correlation between particular early life history traits
and temporal patterns of recruitment indicates that bi-
ology is also clearly involved. Physiological and be-
havioral differences during larval life differ among
eight closely related fishes, leading to distinct temporal
patterns of recruitment.
The transition between a pelagic and benthic exis-
tence involves a radical change in both physical en-
vironment and the physiological traits necessary for
survival in each environment (Thorson 1950). Because
most marine organisms have complex life histories,
stage-specific adaptation to environmental constraints
is probably a common phenomenon. Due to environ-
mental complexity, the range of solutions to these con-
straints probably varies enormously between the pe-
lagic and benthic realms. A potentially intricate shifting
of life history strategies in response to all of these
constraints must occur rapidly following settlement.
Given the importance of this transition, it is surprising
how very little we currently know about any of the
processes influencing the stages on either side of the
settlement event. Although workers in temperate ma-
rine systems have traditionally focused great attention
on early life history stages in attempt to better under-
stand stock-recruitment issues (e.g., Sissenwine 1984,
Rothchild 1986), until recently, very few studies have
examined the relationship between events occurring
during larval life and subsequent patterns of recruit-
ment in tropical reef fishes (reviewed in Cowen and
Sponaugle,
in press
). In the present study, we attempted
to bridge this gap and to examine whether events oc-
curring during one phase of the life cycle are related
to patterns evident in a subsequent phase. Our results
demonstrate that aspects of larval life, as reflected in
several early life history traits, are related to recruit-
ment dynamics. We suggest that further work into the
specific linkage between the pelagic larval stage and
the juvenile phase has the potential to reveal the pro-
cesses operating in each environment to control pop-
ulation dynamics. Furthermore, comparative studies
among and within diverse groups of organisms that
experience a similar phase transition between environ-
ments may help isolate general physical processes im-
portant to all marine organisms with complex life cy-
cles.
A
CKNOWLEDGMENTS
We thank S. Morgan for valuable discussions during all
phases of the work. The comments of R. Cerrato, D. Conover,
L. Crowder, J. Hare, K. Lwiza, S. Morgan, R. Warner, and
two anonymous reviewers substantially improved earlierver-
sions of the manuscript. Additionally, we thank R. Cerrato
and J. Hare for statistical advice. We are gratefulto K. Miles,
J. Hare, and P. Allard for jumping in (literally) as expert
substitute dive buddies when the need arose. Field facilities
were made available by the Bellairs Research Institute of
McGill University and W. Hunte. This research was supported
by National Science Foundation Grant No. OCE-8911120 to
R. K. Cowen and M. J. Bowman. This represents contribution
number 1035 from the Marine Sciences Research Center.
L
ITERATURE
C
ITED
Barnwell, F. H. 1976. Variation in the form of the tide and
some problems it poses for biological timing systems.
Pages 161–187
in
P. J. DeCoursey, editor. Biological
rhythms in the marine environment. University of South
Carolina Press, Columbia, South Carolina, USA.
Batschelet, E. 1981. Circular statistics in biology. Academic
Press, New York, New York, USA.
Boehlert, G. W., W. Watson, and L. C. Sun. 1992. Horizontal
and vertical distributions of larval fishes around an isolated
oceanic island in the tropical Pacific. Deep-Sea Research
39:439–466.
Booth, D. J. 1992. Larval settlement and preferences by
domino damselfish
Dascyllus albisella
Gill. Journal of Ex-
perimental Marine Biology and Ecology 155:85–104.
Booth, D. J., and G. A. Beretta. 1994. Seasonal recruitment,
habitat associations and survival of pomacentrid reef fish
in the US Virgin Islands. Coral Reefs 13:81–89.
Booth, D. J., and D. M. Brosnan. 1995. The role of recruit-
ment dynamics in rocky shore and coral reef fish com-
munities. Advances in Ecological Research 26:309–385.
Botsford, L. W., C. L. Moloney, A. Hastings, J. L. Largier,
T. M. Powell, K. Higgins, and J. F. Quinn. 1994. The
influence of spatially and temporally varying oceanograph-
ic conditions on meroplanktonic metapopulations. Deep-
Sea Research II 41:107–145.
Brothers, E. B. 1987. Methodological approaches to the ex-
amination of otoliths in aging studies. Pages 319–330
in
R. C. Summerfeldt and G. E. Hall, editors. Age and growth
of fish. Iowa State University Press, Des Moines, Iowa,
USA.
Brothers, E. B., C. P. Mathews, and R. Lasker. 1976. Daily
growth increments in otoliths from larval and adult fishes.
(U.S.) Fisheries Bulletin 74:1–8.
Butman, C. A. 1987. Larval settlement of soft-sediment in-
vertebrates: the spatial scales of pattern explained by active
habitat selection and the emerging role of hydrodynamical
processes. Oceanography and Marine Biology Annual Re-
view 25:113–165.
Campana, S. E., and C. M. Jones. 1992. Analysis of otolith
microstructure data. Pages 73–98
in
D. K. Stevenson and
S. E. Campana, editors. Otolith microstructure examination
and analysis. Canadian Special Publication on Fisheries and
Aquatic Science 117.
Carr, M. H., and M. A. Hixon. 1995. Predation effects on
early post-settlement survivorship of coral-reef fishes. Ma-
rine Ecology Progress Series 124:31–42.
Choat, J. H., A. M. Ayling, and D. R. Schiel. 1988. Temporal
and spatial variation in an island fish fauna. Journal of
Experimental Marine Biology and Ecology 121:91–111.
Colin, P. L., and L. J. Bell. 1991. Aspects of the spawning
of labrid and scarid fishes (Pisces: Labroidei) at Enewetak
Atoll, Marshall Islands with notes on other families. En-
vironmental Biology of Fishes 31:229–260.
Cowen, R. K. 1985. Large scale pattern of recruitment by
the labrid,
Semicossyphus pulcher
: causes and implications.
Journal of Marine Research 43:719–742.
. 1991. Variation in the planktonic larval duration of
the temperate wrasse
Semicossyphus pulcher.
Marine Ecol-
ogy Progress Series 69:9–15.
Cowen, R. K., and J. L. Bodkin. 1993. Annual and spatial
variation of the kelp forest fish assemblage at San Nicolas
Island, California. Pages 463–474
in
F. G. Hochberg, editor.
Third California Islands Symposium: Recent Advances in

200
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
Research on the California Islands. Santa Barbara Museum
of Natural History, Santa Barbara, California, USA.
Cowen, R. K., and L. R. Castro. 1994. Relation of coral reef
fish larval distributions to island scale circulation around
Barbados, West Indies. Bulletin of Marine Science 54:228–
244.
Cowen, R. K., and S. Sponaugle.
In press.
Relationships be-
tween early life history traits and recruitment among coral
reef fishes.
In
R. C. Chambers and E. Trippel,editors. Early
life history and recruitment in fish populations. Chapman
and Hall, London, UK.
Crisp, D. J. 1974. Factors influencing the settlement of ma-
rine invertebrate larvae. Pages 177–265
in
P. T. Grant and
A. M. Makie, editors. Chemoreception in marine organ-
isms. Academic Press, New York, New York, USA.
Doherty, P. J. 1981. Coral reef fishes: recruitment-limited
assemblages? Proceedings of the Fourth International Coral
Reef Symposium 2:465–470.
. 1983. Tropical territorial damselfishes: is density
limited by aggression or recruitment? Ecology64:176–190.
. 1987. The replenishment of populations of coral reef
fishes, recruitment surveys, and the problems of variability
manifest on multiple scales. Bulletin of Marine Science 41:
411–422.
. 1991. Spatial and temporal patterns in recruitment.
Pages 261–293
in
P. F. Sale, editor. The ecology of fishes
on coral reefs. Academic Press, San Diego, California,
USA.
Doherty, P., and T. Fowler 1994. An empirical test of re-
cruitment limitation in a coral reef fish. Science 263:935–
939.
Dunn, O. J. 1964. Multiple contrasts using rank sums. Tech-
nometrics 6:241–252.
Eckert, G. J. 1985. Settlement of coral reef fishes to different
natural substrata and at different depths. Proceedings of the
Fifth International Coral Reef Symposium 5:385–390.
Eckman, J. E., F. E. Werner, and T. F. Gross. 1994. Modeling
some effects of behavior on larval settlement in a turbulent
boundary layer. Deep-Sea Research 41:185–208.
Eggleston, D. B., and D. A. Armstrong. 1995. Pre- and post-
settlement determinants of estuarine Dungeness crab re-
cruitment. Ecological Monographs 65:193–216.
Emery, A. R. 1972. Eddy formation from an oceanic island:
ecological effects. Caribbean Journal of Science 12:121–
128.
Farrell, T. M., D. Bracher, and J. Roughgarden. 1991. Cross-
shelf transport causes recruitment to intertidal populations
in central California. Limnology and Oceanography 36:
279–288.
Forrester, G. E. 1990. Factors influencing the juvenile de-
mography of a coral reef fish population. Ecology 71:1666–
1681.
Fritzche, R. A. 1978. Development of fishes of the Mid-
Atlantic Bight. Volume 5. U.S. Department of Interior,
Washington, D.C., USA.
Gaines, S. D., and M. D. Bertness. 1992. Dispersal of ju-
veniles and variable recruitment in sessile marine species.
Nature 360:579–580.
Gaines, S., S. Brown, and J. Roughgarden. 1985. Spatial
variation in larval concentrations as a cause of spatial vari-
ation in settlement for the barnacle,
Balanus gladula.
Oec-
ologia 67:267–272.
Gladstone, W., and M. Westoby. 1988. Growth and repro-
duction in
Canthigaster valentini
(Pisces, Tetradontidae):
a comparison of a toxic reef fish with other reef fishes.
Environmental Biology of Fishes 21:207–221.
Glynn, P. W. 1972. Observations on the ecology of the Ca-
ribbean and Pacific coasts of Panama. Bulletin of the Bi-
ological Society of Washington 2:13–20.
Greig-Smith, P. 1964. Quantitative plant ecology. Butter-
worths, London, UK.
Hare, J. A., and R. K. Cowen.
In press.
Transport mechanisms
of larval and pelagic juvenile bluefish (
Pomatomus salta-
trix
) from South Atlantic Bight spawning grounds to Mid-
dle Atlantic Bight nursery habitats. Limnology and Ocean-
ography.
Hixon, M. A., and J. P. Beets. 1993. Predation, prey refuges,
and the structure of coral-reef fish assemblages. Ecological
Monographs 63:77–101.
Hobson, E. S., and J. R. Chess. 1978. Trophic relationships
among fishes and plankton at Enewetak Atoll, Marshall
Islands. Fishery Bulletin 76:133–153.
Hunt von Herbing, I., and W. Hunte. 1991. Spawning and
recruitment of the bluehead wrasse
Thalassoma bifasciatum
in Barbados, West Indies. Marine Ecology Progress Series
72:49–58.
Jackson, G. A., and R. R. Strathman. 1981. Larval mortality
from offshore mixing as a link between precompetent and
competent period of development. American Naturalist
118:16–26.
Jenkins, G. P., and H. M. A. May. 1994. Variation in settle-
ment and larval duration of King George Whiting,
Silla-
ginodes punctata
(Sillaginidae), in Swan Bay, Victoria,
Australia. Bulletin of Marine Science 54:281–296
Johannes, R. E. 1978. Reproductive strategies of coastal ma-
rine fishes in the tropics. Environmental Biology of Fishes
3:65–84.
Jones, G. P. 1990. The importance of recruitment to the dy-
namics of a coral reef fish population. Ecology 71:1691–
1698.
. 1991. Postrecruitment processes in the ecology of
coral reef fish populations: a multifactorial perspective.
Pages 294–328
in
P. F. Sale, editor. The ecology of fishes
on coral reefs. Academic Press, San Diego, California,
USA.
Keough, M. J. 1988. Benthic populations: is recruitment lim-
iting or just fashionable? Proceedings of the Sixth Inter-
national Coral Reef Symposium 1:141–148.
Kingsford, M. J. 1980. Interrelationships between spawning
and recruitment of
Chromis dispilis
(Pisces: Pomacentri-
dae). Thesis. University of Auckland, Auckland, New Zea-
land.
Kingsford, M. J., and J. H. Choat. 1986. Influence of surface
slicks on the distribution and onshore movements of small
fish. Marine Biology 91:161–171.
Kingsford, M. J., E. Wolanski, and J. H. Choat. 1991. In-
fluence of tidally induced fronts and Langmuir circulations
on distribution and movements of pre-settlement fishes
around a coral reef. Marine Biology 109:167–180
Lacson, J. M. 1992. Minimal genetic variation among sam-
ples of six species of coral reef fishes collected at La Par-
guera, Puerto Rico, and Discovery Bay, Jamaica. Marine
Biology 112:327–331.
Laws, E. A., and J. W. Archie. 1981. Appropriate use of
regression analysis in marine biology. Marine Biology 65:
13–16.
Lee, T. N., C. Rooth, E. Williams, M. McGowan, A. F.
Szmant, and M. E. Clarke. 1992. Influence of Florida Cur-
rent, gyres and wind-driven circulation on transport of lar-
vae and recruitment in the Florida Keys coral reefs. Con-
tinental Shelf Research 12:971–1002.
Leis, J. M. 1991. The pelagic stage of reef fishes: the larval
biology of coral reef fishes. Pages 183–230
in
P. F. Sale,
editor. The ecology of fishes on coral reefs. Academic
Press, San Diego, California, USA.
Leis, J. M., H. P. A. Sweatman, and S. E. Reader. 1996. What
are the pelagic stages of coral reef fishes doing out in blue
water: direct obser vations of larval behavioural capabilities
in the field. Marine and Freshwater Research 47:401–412.

May 1997 201
RECRUITMENT PATTERNS OF WRASSES
Lobel, P. S., and A. R. Robinson. 1986. Transport and en-
trapment of fish larvae by ocean mesoscale eddies and cur-
rents in Hawaiian waters. Deep-Sea Research 33:483–500.
Luckhurst, B. E., and K. Luckhurst. 1977. Recruitment pat-
terns of coral reef fishes on the fringing reefs of Curac¸ao,
Netherlands Antilles. Canadian Journal of Zoology 55:
681–689.
McCormick, M. I. 1994. Variability in age and size at set-
tlement of the tropical goatfish
Upeneus tragula
(Mullidae)
in the northern Great Barrier Reef lagoon. Marine Ecology
Progress Series 103:1–15.
McFarland, W. N., E B. Brothers, L. C. Ogden, M. J.Shulman,
E. L. Bermingham, and N. M. Kotchian-Prentiss. 1985.
Recruitment patterns in young french grunts,
Haemulon
flavolineatum
(Family Haemulidae), at St. Croix, Virgin
Islands. Fishery Bulletin 83:413–426.
Meekan, M. G., M. J. Milicich, and P. J. Doherty. 1993.
Larval production drives temporal patterns of larval supply
and recruitment of a coral reef damselfish. Marine Ecology
Progress Series 93:217–225.
Milicich, M. J. 1994. Dynamic coupling of reef fish replen-
ishment and oceanographic processes. Marine Ecology
Progress Series 110:135–144.
Mitton, J. B., C. J. Berg, Jr., and K. S. Orr. 1989. Population
structure, larval dispersal, and gene flow in the queen
conch,
Strombus gigas,
of the Caribbean. Biological Bul-
letin 177:356–362.
Morgan, S. G. 1995. The timing of larval release. Pages 157–
191
in
L. McEdward, editor. The ecology of marine in-
vertebrate larvae. CRC, New York, New York, USA.
Morgan, S. G., and J. H. Christy. 1994. Plasticity,constraint,
and optimality in reproductive timing. Ecology 75:2185–
2203.
Munro, J. L., V. C. Gaunt, R. Thompson, and P. H. Reeson.
1973. The spawning seasons of Caribbean reef fishes. Jour-
nal of Fish Biology 5:69–84.
Ochi, H. 1985. Temporal patterns of breeding and larval
settlement in a temperate population of the tropical ane-
monefish,
Amphiprion clarkii.
Japanese Journal of Ichthy-
ology 32:248–257.
Olafsson, E. B, C. H. Petersen, and W. G. Ambrose. 1994.
Does recruitment limitation structure populations and com-
munities of macro-invertebrates in marine soft sediments:
the relative significance of pre- and post-settlement pro-
cesses. Oceanography and Marine Biology Annual Review
32:65–109.
Pawlik, J. R. 1992. Chemical ecology of the settlement of
benthic marine invertebrates. Oceanography and Marine
Biology Annual Review 30:273–335.
Petersen, C. W., R. R. Warner, S. Cohen, H. C. Hess, and A.
T. Sewell. 1992. Variable pelagic fertilization success:im-
plications for mate choice and spatial patterns of mating.
Ecology 73:391–401.
Peterson, C. H., and H. C. Summerson. 1992. Basin-scale
coherence of population dynamics of an exploited marine
invertebrate, the bay scallop: implications of recruitment
limitation. Marine Ecology Progress Series 90:257–272.
Pitcher, C. R. 1988. Spatial variation in the temporal pattern
of recruitment of a coral reef damselfish. Proceedings of
the Sixth International Coral Reef Symposium 2:811–816.
Powles, H. 1975. Abundance, seasonality, distribution and
aspects of the ecology of some larval fishes off Barbados.
Dissertation. McGill University, Montreal, Canada.
Raimondi, P. T. 1990. Patterns, mechanisms, consequences
of variability in settlement and recruitment of an intertidal
barnacle. Ecological Monographs 60:283–309.
Reed, D. C., D. R. Laur, and A. W. Ebling. 1988. Variation
in algal dispersal and recruitment: the importance of epi-
sodic events. Ecological Monographs 58:321–335.
Richards, W. J., and K. C. Lindeman. 1987. Recruitment
dynamics of reef fishes: planktonic processes, settlement
and demersal ecologies, and fishery analysis. Bulletin of
Marine Science 41:392–410.
Ricker, W. E. 1973. Linear regressions in fisheries research.
Journal of the Fisheries Research Board of Canada 30:409–
434. . 1975. Computation and interpretation of biological
statistics of fish populations. Bulletin of the Department of
the Environment, Fisheries and Marine Science 191.
Robertson, D. R. 1981. The social and mating systems of
two labrid fishes,
Halichoeres maculipinna
and
H. garnoti,
off the Caribbean coast of Panama. Marine Biology 64:
327–340.
. 1990. Differences in the seasonalities of spawning
and recruitment of some small neotropical reef fishes. Jour-
nal of Experimental Marine Biology and Ecology 144:49–
62. . 1991. The role of adult biology in the timing of
spawning of tropical reef fishes. Pages 356–386
in
P. F.
Sale, editor. The ecology of fishes on coral reefs.Academic
Press, San Diego, California, USA.
. 1992. Patterns of lunar settlement and early recruit-
ment in Caribbean reef fishes at Panama. Marine Biology
114:527–537.
Robertson, D. R., D. G. Green, and B. C. Victor. 1988. Tem-
poral coupling of production and recruitment of larvae of
a Caribbean reef fish. Ecology 69:370–381.
Robertson, D. R., C. W. Petersen, and J. D. Brawn. 1990.
Lunar reproductive cycles of benthic-brooding reef fishes:
reflections of larval biology or adult biology? Ecological
Monographs 60:311–329.
Robertson, D. R., U. M. Schober, and J. D. Brawn. 1993.
Comparative variation in spawning output and juvenile re-
cruitment of some Caribbean reef fishes. Marine Ecology
Progress Series 94:105–113.
Rothchild, B. J. 1986. Dynamics of marine fish populations.
Harvard University Press, Cambridge, Massachusetts,
USA.
Roughgarden, J., S. Gaines, and H. Possingham. 1988. Re-
cruitment dynamics in complex life cycles. Science 241:
1460–1466.
Roughgarden, J., J. T. Pennington, D. Stoner, S. Alexander,
and K. Miller. 1991. Collisions of upwelling fronts with
the intertidal zone: the cause of recruitment pulses in bar-
nacle populations of central California. Acta Oecologica
12:35–51.
Sale, P. F. 1970. Distribution of larval Acanthuridae off Ha-
waii. Copeia 1970:765–766.
Sale, P. F., W. A. Douglas, and P. J. Doherty. 1984. Choice
of microhabitats by coral reef fishes at settlement. Coral
Reefs 3:91–99.
Sale, P. F., and D. J. Ferrell. 1988. Early survivorship of
juvenile coral reef fishes. Coral Reefs 7:117–124.
Scheltema, R. S. 1977. Dispersal of marine invertebrate or-
ganisms: paleobiogeographic and biostrategraphic impli-
cations. Pages 73–108
in
E. G. Kauffman and J. E. Hazel,
editors. Concepts and methods in biostratigraphy.Dowden,
Hutchinson and Ross, Stroudsburg, Pennsylvania, USA.
Schultz, E. T., and R. K. Cowen. 1994. Recruitment of coral
reef fishes to Bermuda: local retention or long-distance
transport? Marine Ecology Progress Series 109:15–28.
Shanks, A. L. 1995. Mechanisms of cross-shelf dispersal for
larval invertebrates and fish. Pages 323–367
in
L. Mc-
Edward, editor. The ecology of marine invertebrate larvae.
CRC, New York, New York, USA.
Shenker, J. M., E. D. Maddox, E. Wishinski, A. Pearl, S.
Thorrold, and N. Smith. 1993. Onshore transport of set-
tlement-stage Nassau grouper
Epinephelus striatus
and oth-
er fishes in Exuma Sound, Bahamas. Marine Ecology Prog-
ress Series 98:31–43.

202
SU SPONAUGLE AND ROBERT K. COWEN
Ecological Monographs
Vol. 67, No. 2
Shulman, M. J. 1985. Recruitmentof coral reef fishes: effects
of distribution of predators and shelter. Ecology 66:1056–
1066.
Shulman, M. J., and E. Bermingham. 1995. Early life his-
tories, ocean currents, and the population genetics of Ca-
ribbean reef fishes. Evolution 49:897–910.
Sissenwine, M. P. 1984. Why do fish populations vary? Pages
59–94
in
R. M. May, editor. Exploitation of marine com-
munities. Springer-Verlag, New York, New York, USA.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W. H. Freeman,
San Francisco, California, USA.
Sponaugle, S., and R. K. Cowen. 1994. Larval durations and
recruitment patterns of two Caribbean gobies (Gobiidae):
contrasting early life histories in demersal spawners. Ma-
rine Biology 120:133–143.
Sponaugle, S., and R. K. Cowen. 1996
a.
Nearshore patterns
of larval supply to Barbados, West Indies. Marine Ecology
Progress Series 133:13–28.
Sponaugle, S., and R. K. Cowen. 1996
b.
Larval supply and
patterns of recruitment for two Caribbean fishes,
Stegastes
partitus
and
Acanthurus bahianus.
Marine and Freshwater
Research 47:433–447.
Stoner, D. S. 1990. Recruitment of a tropical colonial as-
cidian: relative importance of pre-settlement vs. post-set-
tlement processes. Ecology 71:1682–1690.
Sweatman, H. P. A. 1983. Influence ofconspecifics on choice
of settlement sites by larvae of two pomacentrid fishes
(
Dascyllus aruanus
and
D. reticulatus
) on coral reefs. Ma-
rine Biology 75:225–229.
Tatsuoka, M. M. 1971. Multivariate analysis: techniques for
educational and psychological research. John Wiley and
Sons, New York, New York, USA.
Thorrold, S. R., J. M. Shenker, E. D. Maddox, R. Mojica,
and E. Wishinski. 1994
a.
Larval supply of shorefishes to
nursery habitats around Lee Stocking Island, Bahamas. II.
Lunar and oceanographic influences. Marine Biology 118:
567–578.
Thorrold, S. R., J. M. Shenker, R. Mojica, Jr., E. D. Maddox,
and E. Wishinski. 1994
b.
Temporal patterns in the larval
supply of summer-recruiting reef fishes to Lee Stocking
Island, Bahamas. Marine Ecology Progress Series 112:75–
86.
Thorson, G. 1950. Reproductive and larval ecology of ma-
rine bottom invertebrates. Biological Review 25:1–45.
Thresher, R. E. 1979. Social behavior and ecology of two
sympatric wrasses (Labridae:
Halichoeres
spp.) off the
coast of Florida. Marine Biology 53:161–172.
. 1984. Reproduction in reef fishes. T. F. H. Publish-
ers, Neptune City, New Jersey, USA.
Thresher, R. E., and E. B. Brothers. 1989. Evidence ofintra-
and inter-oceanic regional differences in the early life his-
tory of reef-associated fishes. Marine Ecology Progress Se-
ries 57:187–205.
Thresher, R. E., P. L. Colin, and L. J. Bell. 1989. Planktonic
duration, distribution and population structure of western
and central Pacific damselfishes (Pomacentridae). Copeia
1989:420–434.
Victor, B. C. 1982. Daily otolith increments and recruitment
in two coral-reef wrasses,
Thalassoma bifasciatum
and
Hal-
ichoeres bivittatus.
Marine Biology 71:203–208.
. 1983
a.
Recruitment and population dynamics of a
coral reef fish. Science 219:419–420.
. 1983
b.
Settlement and larval metamorphosis pro-
duce distinct marks on the otoliths of the slippery dick,
Halichoeres bivittatus.
Pages 47–51
in
M. L. Reaka, editor.
The ecology of deep and shallow coral reefs. Underwater
Research Symposium Series, Volume 1. National Oceanic
and Atmospheric Administration, Rockville, Maryland,
USA.
. 1986
a.
Larval settlement and juvenile mortality in
a recruitment-limited coral reef fish population. Ecological
Monographs 56:145–160.
. 1986
b.
Delayed metamorphosis with reduced larval
growth in a coral reef fish (
Thalassoma bifasciatum
). Ca-
nadian Journal of Fisheries and Aquatic Science 43:1208–
1213.
. 1986
c.
Duration of the planktonic larval stage of
one hundred species of Pacific and Atlantic wrasses (family
Labridae). Marine Biology 90:317–326.
. 1991. Settlement strategies and biogeography of reef
fishes. Pages 231–260
in
P. F. Sale, editor. The ecology of
fishes on coral reefs. Academic Press, San Diego, Califor-
nia, USA.
Warner, R. R., and T. P. Hughes. 1988. The population dy-
namics of reef fishes. Proceedings of the Sixth International
Coral Reef Symposium 1:149–155.
Warner, R. R., and D. R. Robertson. 1978. Sexual patterns
in the labroid fishes of the western Caribbean. I. The
wrasses (Labridae). Smithsonian Contributions to Zoology
254:1–27.
Wellington, G. M., and B. C. Victor. 1989. Planktonic larval
duration of one hundred species of Pacific and Atlantic
damselfishes (Pomacentridae). Marine Biology 101:557–
567.
Wellington, G. M., and B. C. Victor. 1992. Regional differ-
ences in duration of the planktonic larval stage of reef fishes
in the eastern Pacific Ocean. Marine Biology 113:491–498.
Williams, D. McB. 1980. Dynamics of the pomacentrid com-
munity on small patch reefs in One Tree Lagoon (Great
Barrier Reef). Bulletin of Marine Science 30:159–170.
Wing, S. R., L. W. Botsford, J. L. Largier, and L. E. Morgan.
1995. Spatial structure of relaxation events and crab set-
tlement in the northern California upwelling system. Ma-
rine Ecology Progress Series 128:199–211.
Wolanski, E., and W. M. Hamner. 1988. Topographically
controlled fronts in the ocean and their biological influence.
Science 241:177–181.
Young, C. M. 1995. Behavior and locomotion during the
dispersal phase of larval life. Pages 249–277
in
L. Mc-
Edward, editor. The ecology of marine invertebrate larvae.
CRC, New York, New York, USA.
Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, En-
glewood Cliffs, New Jersey, USA.