See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/227526001

The relationship between fish

species richness, abundance and
habitat complexity in a range of
shallow tropical marine...

Article in Journal of Fish Biology · March 2005

DOI: 10.1111/j.0022-1112.2005.00629.x

CITATIONS READS

333 3,115

2 authors:

Brian Gratwicke Martin R. Speight

Smithsonian Institution University of Oxford
54 PUBLICATIONS 1,124 CITATIONS 51 PUBLICATIONS 1,328 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Development of Artificial Reproductive Technologies for Endangered

Amphibian Species in Panama View project

Teaching View project

All content following this page was uploaded by Brian Gratwicke on 01 December 2017.

The user has requested enhancement of the downloaded file.

Journal of Fish Biology (2005) 66, 650–667
doi:10.1111/j.1095-8649.2005.00629.x, available online at http://www.blackwell-synergy.com

The relationship between fish species richness,

abundance and habitat complexity in a range of shallow
tropical marine habitats
B. G R A T W I C K E * AND M. R. S P E I G H T
Department of Zoology, Oxford University, South Parks Road, OX1 3PS, U.K.

(Received 3 May 2004, Accepted 4 November 2004)

A simple habitat assessment score (HAS) was designed to assess habitat complexity across
several different shallow tropical marine habitats including sandy patches, algal beds, seagrass
beds and reefs. It measured rugosity, variety of growth forms, height, refuge size categories,
percentage live cover and percentage hard substratum. Multiple regression models using HAS
variables as predictors accounted for 71 and 22% of the variation in observed species richness
and total fish abundance respectively. The two most important predictors of observed species
richness were rugosity and variety of growth forms, while height was the most important
predictor of total fish abundance. The HAS method worked consistently across a variety of
habitat types and the complexity map closely mirrored the map of observed species richness,
reflecting the patchy habitat mosaic of shallow tropical marine areas. Stations at the mouth of
an enclosed lagoon, however, had a higher number of species than might have been expected
judging from the habitat complexity scores. It is possible that this was linked to the preferential
settling of pelagic fish larvae in this area as tidal water exchanges between the bay and the reef
were funnelled through one small gap. This study highlights the need for fish biodiversity
studies to take habitat complexity into account. # 2005 The Fisheries Society of the British Isles

Key words: British Virgin Islands; environmental impact; fish diversity; habitat structure; rapid
assessment.

INTRODUCTION
Most fishermen know that the best place to find fishes is around a structure, such
as a fallen log in a stream, a shipwreck or even an artificial reef. In fact, the first
artificial reefs were built in Japan as early as the 1700s as fish aggregating
devices (Meier et al., 1989). More recently, attention has focused on the rela-
tionship between fish species richness and habitat complexity at local scales in
both freshwater and marine environments (Risk, 1972; Gorman & Karr, 1978;
Luckhurst & Luckhurst, 1978; Roberts & Ormond, 1987; Gorham & Alevizon,
1989; McClanahan, 1994; Caley & St John, 1996; Beukers & Jones, 1997; Horan
et al., 2000; Ferreira et al., 2001; Harding & Mann, 2001). These studies use a
wide variety of different methods for assessing habitat complexity, making it

*Author to whom correspondence should be addressed at present address: National Fish and Wildlife
Foundation, 1120 Connecticut Avenue, N. W. Suite 900, Washington D.C. 20036, U.S.A. Tel.:
þ1 202 857 0166; fax: þ1 202 857 0162; email: [email protected]

650
# 2005 The Fisheries Society of the British Isles
 FISH AND HABITAT COMPLEXITY 651

difficult to draw any general conclusions. The Oxford English Dictionary

defines complexity as ‘consisting of many different and connected parts’. It
follows therefore that an assessment of habitat complexity is a multivariate
problem and that several different aspects of habitat structure and composition
need to be taken into account when addressing it.
Most authors only examine one or two variables out of at least six recognized
components of habitat complexity. These are (1) topographic complexity or
rugosity of the substratum (Risk, 1972; Luckhurst & Luckhurst, 1978; Roberts
& Ormond, 1987; Jennings et al., 1996; Chabanet et al., 1997; Ohman &
Rajasuriya, 1998), (2) substratum diversity (Roberts & Ormond, 1987; Ohman
& Rajasuriya, 1998), (3) variety of refuge hole sizes (Roberts & Ormond, 1987),
(4) vertical relief or height of substratum architecture (Luckhurst & Luckhurst,
1978; Molles, 1978), (5) percentage live cover, including corals (Carpenter et al.,
1981; Bell & Galzin, 1984; Sano et al., 1984) and seagrasses (Aliaume et al.,
1990; Edgar & Shaw, 1995; Hemminga & Duarte, 2000) and (6) percentage hard
substratum (Cohen et al., 1993; Guidetti, 2000; Khalaf & Kochzius, 2002).
Rugosity has been positively correlated with observed fish species richness
in most reports, but the relationship between fish species richness and other
complexity variables has been less consistent. Live coral cover is a particu-
larly controversial example because some authors provide data to support the
positive relationship between observed species richness and live coral cover
(Carpenter et al., 1981; Bell & Galzin, 1984; Sano et al., 1984; Lewis, 1997)
while others authors report none (Sale & Dybdahl, 1975; Luckhurst &
Luckhurst, 1978; Roberts & Ormond, 1987). A variety of different method-
ologies are used in these studies and many only examine selected fish taxa in
one habitat type. Predictions about the relative importance of different
components of habitat complexity are practically impossible because there
have been no systematic studies across a range of habitats. One reason for
this might be that the measurement of complexity variables is often time-
consuming and tedious to measure. For example, Risk (1972) assessed the
structural complexity of the substratum by draping a finely linked chain over
it and warned that it took 5 h underwater to assess a single 1 m2 quadrat.
Another factor might be connected with the fact that different methods are
used in different habitats. For example, draping a chain over a reef is a
suitable way to measure reef complexity, but it is difficult to apply in a
seagrass bed where blade height and density measurements would be more
appropriate.
There are many different factors currently affecting marine habitats, ranging
from bleaching of corals to mangrove clearing and the death of seagrass beds
due to eutrophication. Fisheries managers clearly need a simple, inexpensive
approach that can be used to quantify habitat complexity in a range of different
habitats that can be used to predict how environmental change might affect fish
assemblages.
The aims of this study were: (1) to devise a simple rapid assessment method
that, unlike most previous methods, was not habitat specific, (2) to determine
which of these variables were the most important predictors of fish species
richness and abundance and (3) to describe the spatial patterns of variation in
habitat complexity, species richness and abundance.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
652 B. GRATWICKE AND M. R. SPEIGHT

METHODS

STUDY AREA
The study area was located on the south shore of Tortola, a 61 km2 island of the
British Virgin Islands (BVI) in the Caribbean. The islands are steep-sided, volcanic
formations with no perennial streams or rivers. The moderate tropical climate and
clear waters of the BVI support a wide range of habitats including mangroves,
seagrasses and coral reefs. Over 530 of the 1500 species of Caribbean fishes have
been recorded in the Virgin Islands (Froese & Pauly, 2003). One hundred and six
stations were sampled in Hodges Creek, Paraquita Bay and Brandywine Bay (Fig. 1).
These bays were chosen because they encompass a wide variety of different habitats.
Mangroves, sand, seagrass, algal beds and coralline boulders form a patchy mosaic
inside the bays that are fringed by spur and groove fore reefs. Stations were selected in
a semi-systematic way by moving at least 30 m from the nearest other sampling point
and dropping a weighted float overboard and marking the station’s position with a
Garmine Etrex GPS receiver, then diving or snorkelling depending on the depth, to
sample the fishes and environmental variables.
Prior to the survey, the observers familiarized themselves with the area and the fishes
by intensively snorkelling and recording all fish encountered in the study area during a 4
week period prior to the survey. Certain ‘problem’ species were apparent and a protocol
was established for each of them. For example, Malacoctenus macropus (Poey) males
were clearly distinguished by the rosy flush on their chins, but in the field, females and
juveniles were difficult to distinguish from female and juvenile Malacoctenus versicolor
(Poey), as a result the two species were ‘lumped’ as M. macropus/versicolor. Some other
fish taxa were also identified as far as possible into the self-explanatory groups presented
in the results section.
64·57° W

Road Town

Tortola Hodges Creek

Study area

18·42° N Paraquita Bay

Brandywine Bay

0 0·3 0·6
km

FIG. 1. Study area, shoreline features and sampling stations in the three bays, Tortola, British Virgin
Islands ( , dock; , station; , mangrove; , building).

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 653

SA MPL I NG
A four-quadrat sample was taken at each station by placing a 25  25 m quadrat on
each of the four corners of an unsampled 25 m square with the station marker in its
centre. After the placing of a quadrat, the fishes were allowed a few minutes to become
accustomed to the presence of the quadrat and the observer. During this time the habitat
complexity was assessed using a habitat assessment score (HAS) sheet printed on the
back of the underwater recording slate (Table I). The quadrat was then observed for
4 min from a distance of c. 1 m, counting and identifying all fishes in the quadrat or
moving through it. A further minute was spent actively searching for fishes hiding in the
structure or the camouflaged ones that were not detected previously. Stations were visited
in randomized order to control for possible time of day effects and were visited twice in
total, the second sampling occasion being 2–4 weeks after the first. Samples were all
recorded during the summer months of July, August and September 2001. The percent-
age cover of the actual habitats was also recorded on the slate. The categories used were:
seagrass, fleshy algae, sand, mangrove, rubble, reef (including corals, zooanthid and
rock) and ‘other’. Stations were assigned a rough habitat category depending on which
substratum variable was dominant. For example a station with 30% reef, 40% seagrass
and 30% sand would have been classified as seagrass. It must be noted here that these
categories are very approximate and have been used only to give an approximate
impression of which dominant substratum type corresponds to the relevant HAS score.
One of the reasons for creating a non habitat-specific complexity index is so that all
habitats can be surveyed and analysed by direct gradient analysis without the user bias
associated with site selection of pre-conceived ‘suitable’ habitats.

TABLE I. Habitat assessment score (HAS) sheet for near-shore coastal habitats, copied
on to the back of an underwater slate. The HAS score for each category was assessed
visually in each quadrat and recorded with the fish data

HAS SCORE

1 2 3 4 5

Rugosity (visual topographic

estimate of the substratum
in each quadrat)
Variety of growth forms <2 3&4 5&6 7&8 9–10
(stalked/lobed/filamentous/
ribbon-like/massive/branching/
cylindrical/tube/fan/plate/
pinnate/encrusting/other)
Height (visual estimate of 0–9 10–19 20–39 40–79 >80
average height of habitat
architecture) (cm)
Refuge size categories (holes or 0–1 2 3 4 5
gaps in habitat architecture or
substratum in the following size
categories: 1–5, 6–15, 16–30,
31–50 and >50cm)
Live cover (total per cent cover of 0–19 20–39 40–59 60–79 80–100
e.g. living corals, mangrove roots,
seagrass, macroalgae and sponges)
Hard substratum (%) 0–19 20–39 40–59 60–79 80–100

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
654 B. GRATWICKE AND M. R. SPEIGHT

HAS SHEET
The substratum profile across the 25 m quadrat was examined to assess the rugosity
score (in this study substratum refers to mud, sand, rubble, boulders, rock, hard coral
skeletons, concrete pillars and other artificial structures). This means that two flat sandy
areas would score 1 regardless of any seagrass, mangroves or soft corals growing on it. If
the substratum was generally flat with few bumps it was rated 1 or 2 while the very
complex substratum-profile of a quadrat filled with hard, branching corals would score 5.
The term ‘growth forms’ was applied very generally to living organisms such as coral
algae and seagrass. The algae Penicillus spp. are examples of a ‘stalked’ growth form, the
coral Montastrea annularis is usually lobed, manatee grass Syringodium filiforme is
filamentous, turtle grass Thalassia testudinum is ribbon-like, brain corals Diploria spp.
are massive, staghorn coral Acropora cervicornis is branched, mangrove roots are usually
cylindrical or branched, tube sponges (e.g. Agelas conifera) are tubular, the seafan
Paramuricea placomus is fan-like, some firecorals Millepora spp. have a plate-like growth
forms while others may be encrusting or branched, and an example of a pinnate growth
form is the alga Caluerpa taxifolia. The aim of this score is to assess the diversity of
structural attributes of the habitat that might provide resources for different fish species.
The average height of the habitat architecture was a subjective visual estimate to
distinguish between taller and shorter habitats. For example in areas with mangrove
roots this would usually be equivalent to the water depth, while seagrass beds were
generally 10–19 cm, but in sparse or closely cropped areas would be 0–9 cm. The average
height of reefs was usually assessed in relation to the lowest point in the quadrat.
The refuge-size categories of Roberts & Ormond (1987) were used in this study. Holes
in reefs were easily measured, and gaps between structures that would provide an avenue
for a fish to escape a predator were estimated visually. For example small gaps of 1–5 cm
between seagrass blades were abundant in seagrass beds but larger gaps were uncommon.
The gaps between mangrove prop roots, however, ranged from small to quite large
providing hiding places for fishes of many different sizes.
Live cover was estimated by looking at the substratum with a ‘bird’s-eye view’ and
judging the percentage area covered by live coral, mangrove roots, seagrass, macroalgae
or any other living substratum (excluding epiphytic algal films). Hard substratum
referred to the percentage of substratum that was not mud, sand or rubble. A total
HAS score was calculated by adding the scores of each of the six complexity variables to
give an approximate impression of the overall habitat complexity for rapid assessment
purposes.

ANALYSIS
Rare species occurring at <10 stations were omitted from the analysis and abundance
data were square root transformed to normalize the data then tested for multicolinearity
before running a canonical correspondence analysis (CCA) (ter Braak, 1986) on the fish
and habitat complexity data using ECOM software (# Pisces Conservation, Lymington,
U.K.). This analysis was used to describe the relationship between complexity variables
and individual fish species. A Monte-Carlo randomization test (1000 permutations) was
used to assess the probability of the observed pattern being due to chance. A guide to the
interpretation of these plots is found in ter Braak (1986).
The CCA analysis indicates species preferences for the different complexity variables,
but does not allow a prediction to be made of where the most species or the most fishes
will be found. A multiple regression analysis with HAS variables as predictors and
observed species richness and total fish abundance as response variables was performed
(SPSS v.11 5#). These data were checked for multicolinearity, outliers, normality and
homogeneity of variance and were square root transformed if necessary to meet the
assumptions of multiple regression (Jongman et al., 1987). It might be argued that
different habitats have inherently different species richness and that any observed differ-
ences are due to habitat type rather than complexity. The hypothesis that different
habitat types had different complexity-richness slopes was tested by treating habitat

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 655

type as a covariable in an analysis of ANCOVA using SPSS. In this case, observed

species richness and total fish abundance were the dependent variables, HAS score the
fixed factor and habitat type was a covariate.
The spatial relationship between total HAS and species richness and abundance were
investigated using MAPINFO GIS software# to produce interpolated maps. Fish dis-
tribution maps were drawn by plotting relative abundance data for all species occurring
at >10 stations on a base map of the area.

RESULTS

CCA ANALYSIS
Rugosity and percentage hard substratum were colinear (r2 ¼ 084) suggesting
that areas with more hard substratum cover are also more topographically
complex. This caused multicolinearity problems so hard substratum was
excluded from further analysis. The CCA analysis of stations showed that reef
stations were the most rugose habitats with a variety of refuge spaces, but the
sand seagrass or algae stations had a high degree of over lap with similar
habitat complexity such as low rugosity, low variety of refuges and higher
percentages of live cover than reef stations (Fig. 2). The CCA plot could be
split into three ‘rough’ groups along the rugosity axis (Fig 3). The first group
included mostly reef-dwelling species that preferred rugose substrata: Abudefduf
saxatilis (L.), Acanthurus bahianus Castelnau, Acanthurus chirurgus (Bloch),

1
Axis 2

–1 Rugosity

–2
Refuge
categories
–3 Live cover
Growth forms Scale
–2 –1 0 1 2

Axis 1

FIG. 2. A CCA plot of habitat complexity variables and stations categorized by the main habitat variable

present: *, reef; , sand; ., algae; ,, seagrass. Eigenvalues: axis 1, 045; axis 2, 012. r species-
environment: axis 1, 092; axis 2, 082. First 2 axes account for 21% of the variance. Monte-Carlo
test of all canonical axes significant (P < 001), 1000 permutations.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
656 B. GRATWICKE AND M. R. SPEIGHT

4
G. saep

3 Bath. sp.

S. fusc
2 Eucin. spp.
C. glauc
P. mac
M. aurolin S. intermed L. syn
C. rost
A. bah M. cil
1 H. rad M. mac
O. atlant C. rub L. cyp
H. ruf H. biv S. test
Axis 2

S. aurofr
H. mac M. triang S. leuc/dien
A. coer A. chir
0 H. nig H. plum H. poey G. cinn
C.stri
S. vet O. saur O. chry
T. bif Steg. ads C. cap
S. ise/taen Scar. juvs S. rad
S. virid H. ruf H. flav M. cur
–1 Rugosity H. puella L. gris
M. tuc
S. plan S. barr
M. chry
Scar. juvs H. sciur
L. apod
A. sax
–2

Refuge Growth forms Scale Cover

–3 categories

–2 –1 0 1 2
Axis 1

FIG. 3. A CCA plot of habitat complexity variables and species. A. sax, A. saxatilis; A. bah, A. bahianus;
A. chir, A. chirurgus; A. coer, A. coeruleus; Bath. sp., Bathygobius sp.; C. rost, C. rostrata; C. rub,
C. ruber; C. cap, C. capistratus; C. stri, C. striatus; C. glauc, C. glaucofraenum; Eucin spp.,
Eucinostomus spp.; G. cinn, G. cinereus; G. saep, G. saepepallens; H. flav, H. flavolineatum;
H. plum, H. plumieri; H. sciur, H. sciurus; H. biv, H. bivittatus; H. poey, H. poeyi; H. mac,
H. maculipinna; H. rad, H. radiatus; H. ruf, H. rufus; H. puella, H. puella; H. nig, H. nigricans;
L. cyp, L. cyprinoides; L. gris, L. griseus; L. apod, L. apodus; L. syn, L. synagius; M. aurolin,
M. aurolineatus; M. cil, M. ciliatus; M. cur, M. curema; M. mac, M. macropus/versicolor; M. triang,
M. triangulatus; M. chry, M. chrysurus; M. tuc, M. tuckeri; O. atlant, O. atlanticus; O. chry,
O. chrysurus; O. saur, O. saurus; P. mac, P. maculatus; S. intermed, S. intermedius; Scar. juvs,
Scarus juveniles; S. ise/taen, S. iseri/taeniopterus (adults); S. vet, S. vetula; S. aurofr, S. aurofrenatum;
S. rad, S. radians; S. rub, S. rubripinne; S. virid, S. viride; S. test, S. testudineus; S. barr, S. barracuda;
Steg. ads, Stegastes adults; S. fusc, S. fuscus; S. leuc/dien, S. leucostictus/diencaeus (juveniles); S. plan,
S. planifrons; S. dors, S. dorsopunicans; T. bif, T. bifasciatum. Eigenvalues: axis1, 045; axis 2, 012.
r species-environment: axis 1, 092; axis 2, 082. First two axes accounted for 21% of the variance.
Monte-Carlo test of all canonical axes were significant (P < 001), 1000 permutations.

Acanthurus coeruleus (Bloch & Schnieder), Chaetodon striatus (L.), Haemulon

maculipinna (Muller & Troschel), Halichoeres radiatus (L.), Holocentrus rufus
(Walbaum), Hypoplectrus nigricans (Poey), Malacoctenus aurolineatus Smith,
Malacoctenus triangulatus Springer, Microspathodon chrysurus (Cuvier),
Ophioblennius atlanticus (Valenciennes), Synodus intermedius (Spix & Agassiz),
Scarus iseri (Bloch) and Scarus taeniopterus Desmarest (S. iseri/taeniopterus;
adults), Scarus vetula (Bloch & Schneider), Sparisoma rubripinne (Valenciennes),
Sparisoma viride (Bonaterre), Stegastes dorsopunicans (Poey) and Thalassoma
bifasciatum (Bloch). The second group included species that might be found in
areas intermediate between bays and reefs with moderate substratum rugosity,

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 657

or some species that use bays as juveniles and move onto reefs as adults. They
were: Acanthurus chirurgus (Bloch), Canthigaster rostrata (Bloch), Caranx ruber
(Bloch), Chaetodon capistratus (L.), Coryphopterus glaucofraenum Gill,
Gobionellus saepepallens Gilbert & Randall, Haemulon flavolineatum (Desmarest),
Haemulon plumieri (Lacapède), Haemulon bivittatus (Bloch), Haemulon poeyi
(Steindachner), Hypoplectrus puella (Cuvier), Lutjanus griseus (L.), Lutjanus
apodus (Walbaum), M. macropus/versicolor, Pseudupeneus maculatus (Bloch),
Scarus juveniles, Sparisoma aurofrenatum (Valenciennes), Stegastes adults,
Stegastes leucostictus (Müller & Traschel) and Stegastes diencaeus (Jordan &
Rutter) (S. leucostictus/diencaeus; juvenile colouration) and Stegastes planifrons
(Cuvier). The third group was associated with bay areas that had low rugosity and
sandy substrata, it included: Bathygobius sp. Eucinostomus spp., Gerres cinereus
(Walbaum), Haemulon sciurus (Shaw), Lophogobius cyprinoides (Pallas), Lutjanus
synagris L., Monacanthus ciliatus (Mitchill), Monacanthus tuckeri Bean, Mugil
curema Valenciennes, Ocyurus chrysurus (Bloch), Oligoplites saurus (Bloch &
Schneider), Sparisoma radians (Valenciennes), Sphoeroides testudineus (L.) and
Sphyraena barracuda (Walbaum) (Fig. 3). These groups were internally split by
correlations with other variables for example in group 1, H. sciurus and
S. barracuda preferred areas with high per cent live cover (probably mangroves,
seagrass and algae) while Bathygobius sp. was found in low rugosity, low live
cover situations (probably sand). Similarly, S. dorsopunicans preferred rugose
substrata, without much cover (e.g. shallow, wave-washed reefs) while A. saxatilis
preferred rugose areas with many refuge size categories and growth forms (e.g.
deeper forereef zones). The purpose of this paper is not to indirectly analyse fish
habitat preferences, but to pick out any effects of complexity variables on fish
assemblages. At least 19 species (c. 35% of the 55 fish species common enough for
inclusion in the analysis) favoured rugose substrata of the forereef (Fig. 3)
accounting for a substantial proportion of species, but indicating that other
complexity variables must strongly influence the remaining species.

R E G R ES S I O N S
The combination of habitat complexity variables accounted for 71% of
the variation in observed species richness and 22% of the variation in total
fish abundance (Table II). Two significant predictor variables, rugosity and
variety of growth forms, were positively correlated with observed species
richness while the only significant predictor of total fish abundance was height
(Table II).
When the overall HAS score (sum of all habitat complexity variables) was
plotted against observed species richness and abundance, it was clear that sandy
stations were generally less complex, and had fewer species than algal and
seagrass beds, which in turn were less complex and species-rich than reefs
(Fig. 4). A similar, but less marked pattern was noted with overall fish abun-
dance (Fig. 4). The ANCOVA indicated that the habitats did not have signifi-
cantly different slopes for species richness (ANCOVA, d.f. ¼ 1, P > 005) or
abundance (ANCOVA, d.f. ¼ 1, P > 005) (Table III). It was therefore complex-
ity, rather than habitat types that were responsible for the observed variations
in number of fish species, and to a lesser extent, fish abundance.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
658 B. GRATWICKE AND M. R. SPEIGHT

TABLE II. The multiple regression of all contributing HAS variables on observed species
richness and total fish abundance. The partial correlations and significances are presented

Observed species richness Total fish abundance

Partial Partial
correlation P correlation P

Rugosity 052 0000*** 014 0141 NS

Variety of growth forms 024 0016* 008 0391 NS
Height 007 0450 NS 021 0032*
Refuge size categories 008 0370 NS 009 0359 NS
Live cover (%) 005 0736 NS 011 0260 NS
Overall regression r2 ¼ 071 0000*** r2 ¼ 022 0000***

NS, not significant; *, significant (P < 005); ***, very highly significant (P < 0001).

SP A T IA L V A R IA T IO N
Habitat complexity accounted for the distinctive spatial patterns of species
richness, with more complex fore-reefs areas having more species than the bays
(Fig. 5). A similar, but less marked, pattern was evident with total fish abun-
dance. Closer examination revealed patches outside bays with low complexity
and species richness, while certain areas inside bays were both complex and
species-rich, illustrating the uneven complexity of the habitats in the study area.
The total fish abundance at two stations in Paraquita Bay was exceptionally
high (Fig. 5). This was connected to an exceptional recruitment of thousands of
juvenile parrotfish Scarus spp. that might ordinarily have been less abundant at
these stations. One final interesting pattern to note is that the number of species
immediately surrounding the mouth of Paraquita Bay, which is entirely cut off
from the ocean except by a small gap is slightly higher than might be expected
judging from the habitat complexity index.

DISCUSSION

HABITAT COMPLEXITY AND SPECIES RICHNESS

It was clear that more complex marine habitats supported a greater number
of fish species than less complex ones. This is unsurprising, as many studies at
local scales have shown a similar effect in a range of animal taxa such as: fishes
(Risk, 1972; Gorman & Karr, 1978; Luckhurst & Luckhurst, 1978; Roberts &
Ormond, 1987; Gorham & Alevizon, 1989; McClanahan, 1994; Caley & St
John, 1996; Beukers & Jones, 1997; Horan et al., 2000; Ferreira et al., 2001;
Harding & Mann, 2001), birds (Mac Arthur & Mac Arthur, 1961; Finch, 1989;
Telleria & Carrascal, 1994), lizards (Pianka, 1966, 1973), mammals (Dueser &
Brown, 1980; August, 1983; Williams et al., 2002), insects (Southwood et al.,
1979; Haslett, 1997; Davidowitz & Rosenzweig, 1998), aquatic invertebrates

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 659

(a)

30
√Total fish abundance

20

10

0
(b)

30
Number of species

20

10

0
0 5 10 15 20 25
HAS

FIG. 4. The relationship between the total habitat assessment score (HAS) and (a) total fish abundance
and (b) number of species (* reef, 
sand, . algae, , seagrass). The curves were fitted by (a)
y ¼ 491 þ 078x (r2 ¼ 017) and (b) y ¼ 497 þ 161x (r2 ¼ 055). The multiple regression models,
however, accounted for 22 and 71% of the variation in abundance and fish species richness
respectively (see Table II).

(Heck & Wetstone, 1977; Gilinsky, 1984), gastropods (Kohn, 1967; Beck, 2000)
and even ostracods (Hull, 1997). A recurring theme in all of these papers,
however, is that the term ‘complexity’ can be assessed in a number of different

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
660 B. GRATWICKE AND M. R. SPEIGHT

TABLE III. The results of ANCOVA testing for the covariate effect of habitat type
(covariate) and HAS (fixed factor) on observed number of fish species and abundance

Type III
sum of Mean
Source squares d.f. square F P

Species
Intercept 2028649 1 2028649 129570 0000
Habitat 9831 1 9831 0628 0433
HAS 4923025 65 75739 4837 0000
Error 579303 37 15657
Total 29712000 104
Abundance
Intercept 1409691 1 1409691 56304 0000
Habitat 1935 1 1935 0077 0783
HAS 2293566 65 35286 1409 0130
Error 926370 37 25037
Total 25458250 104

ways, depending on interpretation, and the measurement of just one or two

complexity variables can be a difficult task in the field.
The rapid assessment approach used in this study was both user-friendly and
accounted for a large proportion (71%) of the variation in fish species richness,
and a lesser, but statistically significant proportion (22%) of the variation in fish
abundance across a range of habitats. With a closer examination of the com-
plexity variables some plausible explanations of the mechanisms responsible can
be suggested.

HAS VARIABLES
Rugosity and percentage hard substratum were closely correlated, probably
because rocks are usually coralline boulders that are the primary hard substrata
in the study area, and they have rugose surfaces. Rocky substrata are stable
with large surface areas on which microscopic algae, macroalgae, live hard
corals, gorgonians, encrusting gorgonians, sponges and other animals can live.
It is biologically important because many fish species graze on the epilithic algae
covering the rocks (Carpenter, 1986; Hixon & Brostoff, 1996; Lawson et al.,
1999; Ferreira et al., 2001). These algae are important nitrogen fixers and are
the primary source of production on reefs (Klumpp & McKinnon, 1989;
Larkum 1999). They cannot grow on sand, which is unstable and moves in
turbulent environments with a grinding action that is not suitable for the settling
and growth of algae, corals or certain invertebrates (Rogers, 1990; Booth, 2001;
Hillebrand & Kahlert, 2002). Rugose substrata may also offer more cover from
predators, or nesting sites that would make them attractive to some species. At
least 35% of the fish species analysed were positively correlated with rugosity,
making this the most significant variable related to the observed fish species
richness and its partial correlation accounted for 52% of the variance.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 661

(a)
N

Hodges Creek

Paraquita Bay
18·42° N

Brandywine Bay
HAS score
22
14
64·59° W

64·58° W

64·57° W
12
10
6
0 0·25 0·5
km

(b)
N

Hodges Creek

Paraquita Bay
18·42° N

Brandywine Bay
Species
richness
33
64·59° W

64·58° W

64·57° W

19
15
12
0 0·25 0·5 3

km

(c)
N

Hodges Creek

Paraquita Bay
18·42° N

Brandywine Bay
Abundance
1500
340
64·59° W

64·57° W
64·58° W

270
170
10
0 0·25 0·5
km

FIG. 5. A map of (a) habitat complexity index, (b) species richness and (c) abundance, interpolated using
MAPINFO software from 106 stations in the sample area ( , dock; , station; , mangrove; ,
building).

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
662 B. GRATWICKE AND M. R. SPEIGHT

Variety of growth forms was the second most important habitat complexity
variable for predicting number of fish species and it accounted for only 24% of
the variance in the overall regression. The variety of growth forms is an indica-
tion of different types of microhabitat and a greater variety of these would offer
a larger number of potential resource axes such as food type, camouflage,
refuge spaces, breeding sites or cleaning stations with a range of different uses
to different fishes. The Hutchinsonian concept of the niche predicts that more
species would be able to coexist in areas with a greater variety of resources, as there
would be more potential niches that could be occupied by habitat specialists
(Hutchinson, 1957). For example, trumpetfish Aulostomus maculatus Valenciennes
are dependent on cylindrical sponges and branching gorgonians for camouflage
to stalk their prey (DeLoach, 1999); juvenile yellowtail damselfish M. chrysurus
are strongly associated with blade-like firecorals (Robertson, 1984) and
bucktooth parrotfish Sparisoma radians (Valenciennes) feed only on ribbon-like
seagrasses (Greenway, 1995; Kirsch et al., 2002). While different methods were
used, at least two other studies indicate positive correlations between fish
species richness and coral species richness or substratum diversity indices
(Roberts & Ormond, 1987; Ohman & Rajasuriya, 1998). Others, however,
found no correlation between coral species richness and fish species richness
(Luckhurst & Luckhurst, 1978) and Risk (1972) found no correlation between
fish species richness and substratum diversity. The assessment of variety of
growth forms is an attractive method because it is a simple, direct measure of
the variety of microhabitats and it does not require expertize in algal and coral
taxonomy.
Height was a significant predictor of fish abundance but it only accounted for
a relatively small proportion of the variance (21%). Height was not a significant
predictor of the observed species richness. Fish aggregating devices take advan-
tage of the well-known tendency for fishes to cluster around prominent habitat
features and tall habitats may aggregate fishes in a similar way (Chou, 1988;
Beets, 1989). Alternatively, in the nutrient-poor tropical marine systems, taller
reefs with higher surface areas would have higher levels of substratum-linked
productivity and might therefore realistically support more fishes by increasing
the availability of food resources.
Increasing the variety of refuge size categories would theoretically enhance
species richness by increasing the refuge availability to a range of species with
different body sizes (Luckhurst & Luckhurst, 1978). Other authors suggest that
any observed increase in species richness at stations with more holes is an
artefact of the main effect increasing fish abundance (Molles, 1978; Caley &
St John, 1996). The methods in this study were chosen based on a protocol
established by Roberts & Ormond (1987) who found that the variety of hole-
categories accounted for a significant proportion of total fish abundance, but
the total number of holes was not significantly correlated with total fish species
richness. In this study, however, other variables were more important predictors
of fish abundance and species richness.
The per cent live cover was not a significant predictor variable for fish species
richness or abundance, it is probably because some habitats with low fish
species diversity, such as seagrass and algal beds, had a very high proportion
of live cover. The CCA analysis reflected the unduly high weighting that

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 663

seagrasses received because species favouring seagrass beds at some stage in

their life history, such as G. cinereus, H. sciurus, L. griseus, M. curema,
O. chrysurus, S. radians and S. barracuda (Nagelkerken et al., 2000, 2002),
were correlated with high percentages of live cover (Fig. 3). As a result, it
seems likely that the ratio of live coral cover on reefs and the percentage
cover of seagrass may be useful fish species diversity predictors within their
respective habitats (Carpenter et al., 1981; Bell & Galzin, 1984; Sano et al.,
1984; Aliaume et al., 1990; Edgar & Shaw, 1995; Hemminga & Duarte, 2000),
but the relationship breaks down when applied between habitat types.

SP A TI A L P A T T E R N S
Reef areas are generally more complex than seagrass and algal beds and these
are more complex than sandy areas. The more complex habitats supported
more fishes, a pattern that has been noted by others (Sedberry & Carter,
1993; Nagelkerken et al., 2000; Mateo & Tobias, 2001). This is the first attempt,
however, to quantify habitat complexity variables and relate them to fish species
richness across a wide range of habitats. The index performs well, accounting
for both within-habitat and between-habitat differences in complexity. Forereef
areas tend to be the most complex with high species richness, and the patchy
nature of the complexity and species richness and abundance maps reflects the
patchy nature of the existing habitat mosaic.
The mouth of Paraquita Bay had more species than might be expected
judging from its complexity score. Nagelkerken et al. (2000) noted a similar
pattern around the mouth of an enclosed bay in Curaçao. This is probably
related to large volumes of water entering the bay through one restricted
opening. As the tide rises and pelagic juveniles of reef fishes enter Paraquita
Bay, they probably settle onto the first suitable substratum they encounter,
resulting in higher species richness and abundance around the mouth of the
bay. One part of Paraquita Bay also had exceptional numbers of schooling
juvenile Scarus spp. Seagrass habitats are known to support very high numbers of
juvenile Scarus spp. (Nagelkerken et al., 2000), but this exceptional density was
probably related to a recent recruitment pulse of pelagic juveniles settling in this
thick, shallow seagrass bed where they could feed and hide from predators.

A P P L I C A TI O N S
This research suggests that more complex habitats tend to have higher fish
species richness and abundance than less complex ones, which is not an unex-
pected discovery. This rapid assessment method, however, was designed in a
very robust manner to compare a range of different habitats. The fact that HAS
variables account for 71% of the variation in observed fish species richness
across habitats is quite impressive considering that it is a ‘quick and dirty’ rapid
assessment method. The method used involved at least two visits to each station
to counter potential bias from weather or time of day. The small quadrat size
means that it can be used under visibility conditions as low as 4 m. Since this
research was carried out in summer, it is not known what seasonal variations
might occur, but judging from other studies these need to be taken into account

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
664 B. GRATWICKE AND M. R. SPEIGHT

(Mateo & Tobias, 2001). User-errors resulting from the innate subjectivity of
this method have not been quantified yet and will need to be determined.
Compiling information from multiple users will probably reduce precision and
this is a common problem with most rapid assessment methods.
While this method implied that complex habitats are best for increasing local
fish species richness, it should be remembered that all habitat types, regardless
of their complexity are important in maintaining regional fish diversity. This is
because some fishes use different habitat types at different stages of their
ontogenetic development (Sedberry & Carter, 1993; Nagelkerken et al., 2000,
2001, 2002; Cocheret de la Moriniere et al., 2002) and others are habitat
specialists (Greenway, 1995; Lieske & Myers, 2001; Kirsch et al., 2002).

We would like to thank the Rhodes Trust for funding B.G. and the British Ecological
Society for funding M.R.S. A. Morris assisted with fieldwork and B. Lettsomme,
N. Eristhee and E. Kadison of the Conservation and Fisheries Department of the BVI
offered valuable support and permits. R. Southwood, P. Johnson, A. Foggo and
N. Brown have made helpful comments on the manuscript.

References
Aliaume, C., Lasserre, G. & Louis, M. (1990). Spatial community organization of fishes
in the Turtlegrass Beds of Grand Cul-De-Sac Marin in Guadeloupe French West
Indies. Revue d’Hydrobiologie Tropicale 23, 231–250.
August, P. V. (1983). The role of habitat complexity and heterogeneity in structuring
tropical mammal communities. Ecology 64, 1495–1507.
Beck, M. W. (2000). Separating the elements of habitat structure: independent effects of
habitat complexity and structural components on rocky intertidal gastropods.
Journal of Experimental Marine Biology and Ecology 249, 29–49.
Beets, J. P. (1989). Experimental evaluation of fish recruitment to combinations of fish
aggregating devices and benthic artificial reefs. Bulletin of Marine Science 44,
973–983.
Bell, J. D. & Galzin, R. (1984). Influence of live coral cover on coral reef fish communities.
Marine Ecology Progress Series 15, 265–274.
Beukers, J. S. & Jones, G. P. (1997). Habitat complexity modifies the impact of piscivores
on a coral reef fish population. Oecologia 114, 50–59.
Booth, J. D. (2001). Habitat preferences and behaviour of newly settled Jasus edwardsii
(Palinuridae). Marine and Freshwater Research 52, 1055–1065.
ter Braak, C. J. F. (1986). Canonical Correspondence Analysis: A new eigenvector
technique for multivariate direct gradient analysis. Ecology 67, 1167–1179.
Caley, M. J. & St John, J. (1996). Refuge availability structures assemblages of tropical
reef fishes. Journal of Animal Ecology 65, 414–428.
Carpenter, K. E., Miclat, R. I., Albaladejo, V. D. & Corpuz, V. T. (1981). The influ-
ence of substrate structure on the local abundance and diversity of Philippine
Reef Fishes. In Proceedings of the 4th International Coral Reef Symposium,
pp. 497–502. Quezon City: Marine Sciences Center, University of the Philip-
pines.
Carpenter, R. C. (1986). Partitioning herbivory and its effects on coral reef algal
communities. Ecological Monographs 56, 345–364.
Chabanet, P., Ralambondrainy, H., Amanieu, M., Faure, G. & Galzin, R. (1997).
Relationship between coral reef substrata and fish. Coral Reefs 16, 93–102.
Chou, L. M. (1988). The commercial and recreational values of artificial reefs. Journal of
the Singapore National Academy of Science 17, 50–52.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 665

Cocheret de la Moriniere, E., Pollux, B. J. A., Nagelkerken, I. & van der Velde, G. (2002).
Post-settlement life cycle migration patterns and habitat preference of coral reef
fish that use seagrass and mangrove habitats as nurseries. Estuarine, Coastal and
Shelf Science 55, 309–321.
Cohen, A. S., Bills, R., Cocquyt, C. Z. & Caljon, A. G. (1993). The impact of sediment
pollution on biodiversity in Lake Tanganyika. Conservation Biology 7, 667–677.
Davidowitz, G. & Rosenzweig, M. L. (1998). The latitudinal gradient of species diversity
among North American grasshoppers (Acrididae) within a single habitat: a test of
the spatial heterogeneity hypothesis (in animal patterns). Journal of Biogeography
25, 553–560.
DeLoach, N. (1999). Reef Fish Behaviour. Jacksonville, FL: New World Publications, Inc.
Dueser, R. D. & Brown, W. C. (1980). Ecological correlates of insular rodent diversity.
Ecology 61, 50–56.
Edgar, G. J. & Shaw, C. (1995). The production and trophic ecology and shallow-water
fish assemblages in southern Australia I. Species richness, size-structure and
production of fishes in Western Port, Victoria. Journal of Experimental Marine
Biology and Ecology 194, 53–81.
Ferreira, C. E. L., Goncalves, J. E. A. & Coutinho, R. (2001). Community structure of
fishes and habitat complexity on a tropical rocky shore. Environmental Biology of
Fishes 61, 353–369.
Finch, D. M. (1989). Habitat use and habitat overlap of riparian birds in three elevational
zones. Ecology 70, 866–880.
Gilinsky, E. (1984). The role of fish predation and spatial heterogeneity in determining
benthic community structure. Ecology 65, 455–468.
Gorham, J. C. & Alevizon, W. S. (1989). Habitat complexity and the abundance of juvenile
fishes residing on small scale artificial reefs. Bulletin of Marine Science 44, 662–665.
Gorman, O. T. & Karr, J. R. (1978). Habitat structure and stream fish communities.
Ecology 59, 507–515.
Greenway, M. (1995). Trophic relationships of macrofauna within a Jamaican seagrass
meadow and the role of the Echinoid Lytechinus variegatus (Lamarck). Bulletin of
Marine Science 56, 719–736.
Guidetti, P. (2000). Differences among fish assemblages associated with nearshore
Posidonia oceanica seagrass beds, rocky-algal reefs and unvegetated sand habitats
in the Adriatic Sea. Estuarine Coastal and Shelf Science 50, 515–529.
Harding, J. M. & Mann, R. (2001). Oyster reefs as fish habitat: Opportunistic use of
restored reefs by transient fishes. Journal of Shellfish Research 20, 951–959.
Haslett, J. R. (1997). Insect communities and the spatial complexity of mountain habitats.
Global Ecology and Biogeography Letters 6, 49–56.
Heck, K. L., Jr. & Wetstone, G. S. (1977). Habitat complexity and invertebrate species richness
and abundance in tropical seagrass meadows. Journal of Biogeography 4, 135–142.
Hemminga, M. A. & Duarte, C. M. (2000). Seagrass Ecology. Cambridge: Cambridge
University Press.
Hillebrand, H. & Kahlert, M. (2002). Effect of grazing and water column nutrient supply on
biomass and nutrient content of sediment microalgae. Aquatic Botany 72, 143–159.
Hixon, M. A. & Brostoff, W. N. (1996). Succession and herbivory: effects of differential
fish grazing on hawaiian coral-reef algae. Ecological Monographs 66, 67–90.
Horan, D. L., Kershner, J. L., Hawkins, C. P. & Crowl, T. A. (2000). Effects of habitat
area and complexity on Colorado River cutthroat trout density in Uinta Mountain
streams. Transactions of the American Fisheries Society 129, 1250–1263.
Hull, S. L. (1997). Seasonal changes in diversity and abundance of ostracods on four
species of intertidal algae with differing structural complexity. Marine Ecology
Progress Series 161, 71–82.
Hutchinson, G. E. (1957). Concluding remarks, Cold Spring Harbor Symposium.
Quantitative Biology 22, 415–427.
Jennings, S., Boulle, D. P. & Polunin, N. V. C. (1996). Habitat correlates of the
distribution and biomass of Seychelles reef fishes. Environmental Biology of Fishes
46, 15–25.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
666 B. GRATWICKE AND M. R. SPEIGHT

Jongman, R. H. G., ter Braak, C. J. F. & van Tongeren, O. F. R. (1987). Data Analysis in
Community and Landscape Ecology. Wageningen: Pudoc.
Khalaf, M. A. & Kochzius, M. (2002). Community structure and biogeography of shore
fishes in the Gulf of Aqaba, Red Sea. Helgoland Marine Research 55, 252–284.
Kirsch, K. D., Valentine, J. F. & Heck, K. L., Jr. (2002). Parrotfish grazing on turtlegrass
Thalassia testudinum: Evidence for the importance of seagrass consumption in food
web dynamics of the Florida Keys National Marine Sanctuary. Marine Ecology
Progress Series 227, 71–85.
Klumpp, D. W. & McKinnon, A. D. (1989). Temporal and spatial patterns in primary
production of a coral-reef epilithic algal community. Journal of Experimental
Marine Biology and Ecology 131, 1–22.
Kohn, A. J. (1967). Environmental complexity and species diversity in the gastropod genus
Conus on Indo-West Pacific reef platforms. American Naturalist 101, 251–259.
Larkum, A. W. D. (1999). The cyanobacteria of coral reefs. Bulletin de l’Institut
Oceanographique Monaco Special Issue 19, 149–167.
Lawson, G. L., Kramer, D. L. & Hunte, W. (1999). Size-related habitat use and schooling
behaviour in two species of surgeonfish (Acanthurus bahianus and A. coeruleus) on
a fringing reef in Barbados, West Indies. Environmental Biology of Fishes 54, 19–33.
Lewis, A. R. (1997). Effects of experimental coral disturbance on the structure of fish
communities on large patch reefs. Marine Ecology Progress Series 161, 37–50.
Lieske, E. & Myers, R. (2001). Collins Pocket Guide: Coral Reef Fishes. London: Collins.
Luckhurst, B. E. & Luckhurst, K. (1978). Analysis of the influence of substrate variables
on coral reef fish communities. Marine Biology 49, 317–323.
Mac Arthur, R. H. & Mac Arthur, J. W. (1961). On bird species diversity. Ecology 42,
594–598.
Mateo, I. & Tobias, W. J. (2001). Distribution of shallow water coral reef fishes on the
northeast coast of St. Croix, USVI. Caribbean Journal of Science 37, 210–226.
McClanahan, T. R. (1994). Kenyan coral reef lagoon fish: effects of fishing, substrate
complexity, and sea urchins. Coral Reefs 13, 231–241.
Meier, M. H., Buckley, R. & Polovina, J. J. (1989). A debate on responsible artificial reef
development. Bulletin of Marine Science 44, 1051–1057.
Molles, M. C. (1978). Fish species diversity on model and natural reef patches:
experimental insular biogeography. Ecological Monographs 48, 289–305.
Nagelkerken, I., Dorenbosch, M., Verberk, W. C. E. P., Cocheret de la Moriniere, E. &
van der Velde, G. (2000). Importance of shallow-water biotopes of a Caribbean
bay for juvenile coral reef fishes: patterns in biotope association and spatial
distribution. Marine Ecology Progress Series 202, 175–192.
Nagelkerken, I., Kleijnen, S., Klop, T., van den Brand, R. A. C. J., Cocheret de la
Moriniere, E. & van der Velde, G. (2001). Dependence of Caribbean reef fishes
on mangroves and seagrass beds as nursery habitats: a comparison of fish faunas
between bays with and without mangroves/seagrass beds. Marine Ecology Progress
Series 214, 225–235.
Nagelkerken, I., Roberts, C. M., van der Velde, G., Dorenbosch, M., Riel, M. C.,
Cocheret de la Moriniere, E. & Nienhuis, P. H. (2002). How important are
mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested
on an island scale. Marine Ecology Progress Series 244, 299–305.
Ohman, M. C. & Rajasuriya, A. (1998). Relationships between habitat structure and fish
communities on coral and sandstone reefs. Environmental Biology of Fishes 53, 19–31.
Pianka, E. R. (1966). Convexity, desert lizards and spatial heterogeneity. Ecology 47,
1055–1059.
Pianka, E. R. (1973). The structure of lizard communities. Annual Review of Ecology and
Systematics 4, 53–74.
Risk, M. J. (1972). Fish diversity on a coral reef in the Virgin Islands. Atoll Research
Bulletin 153, 1–6.
Roberts, C. M. & Ormond, R. F. G. (1987). Habitat complexity and coral reef fish
diversity and abundance on Red Sea fringing reefs. Marine Ecology Progress Series
41, 1–8.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667
 FISH AND HABITAT COMPLEXITY 667

Robertson, D. R. (1984). Cohabitation of competing territorial damselfishes on a

Caribbean coral reef. Ecology 65, 1121–1135.
Rogers, C. S. (1990). Responses of coral reefs and reef organisms to sedimentation.
Marine Ecology Progress Series 62, 185–202.
Sale, P. F. & Dybdahl, R. (1975). Determinants of community structure for coral reef
fishes in an experimental habitat. Ecology 56, 1343–1355.
Sano, M., Shimizu, M. & Nose, Y. (1984). Changes in structure of coral reef fish
communities by destruction of hermatypic corals: observational and experimental
views. Pacific Science 38, 51–79.
Sedberry, G. R. & Carter, J. (1993). The fish community of a shallow tropical lagoon in
Belize, Central America. Estuaries 16, 198–215.
Southwood, T. R. E., Brown, V. K. & Reader, P. M. (1979). The relationships of plant
and insect diversities in succession. Biological Journal of the Linnean Society 12,
327–348.
Telleria, J. L. & Carrascal, L. M. (1994). Weight-density relationships between and within
bird communities: implications of niche space and vegetation structure. American
Naturalist 143, 1083–1092.
Williams, S. E., Marsh, H. & Winter, J. (2002). Spatial scale, species diversity, and habitat
structure: Small mammals in Australian tropical rain forest. Ecology 83,
1317–1329.

Electronic Reference
Froese, R. & Pauly, D. (2003). FishBase. Available at www.fishbase.org.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 66, 650–667

View publication stats