Environ Biol Fish (2006) 76:71–80

DOI 10.1007/s10641-006-9009-4

ORIGINAL PAPER

The influence of habitat complexity on reef fish communities

in the northeastern Gulf of Mexico
Mark E. Lingo Æ Stephen T. Szedlmayer

Received: 28 January 2005 / Accepted: 13 March 2006 / Published online: 11 May 2006

Springer Science+Business Media B.V. 2006

Abstract Several previous studies have attempted richness were significantly (P < 0.05) higher on more
to correlate habitat complexity and reef fish species complex habitats (H¢=1.7, S=11–12) compared to less
diversity. These studies have mostly examined natural complex habitats (H¢=0.8–1.0, S=4–9). In 2002, pat-
reef systems, but results differed. To examine this terns among diversity, richness and reef complexity
relation, we built 1 m2 habitats with 20 replicates of were less apparent with only the least complex habi-
five complexity levels from July to August 2001 in the tats shell and cage showing significantly lower values.
northeastern Gulf of Mexico (n=100). In June and July In both years, multidimensional scaling grouped by
2002, we built new habitats using the 2001 design, but complexity levels with cage and shell habitats show-
also added a sixth complexity level (n=120). In order ing the clearest separation from other habitat types.
of increasing complexity these included: cage, shell, Also, with few exceptions (only 8%) analysis of
cage-shell, block-shell, cage-block-shell, and shell- similarities showed significant (P < 0.05) differences
block-pyramid habitats. Most fish in both years were in fish communities across complexity levels. Al-
juveniles and included species common to reef though community composition varied between years,
structures in the northeastern Gulf of Mexico. In 2001, this study provided evidence to support the hypothesis
we identified 26 fish species, and the dominant species that habitat complexity increased reef fish species
was red snapper, Lutjanus campechanus (41%), fol- diversity.
lowed by rock sea bass, Centropristis philadelphica
(23%), and sand perch, Diplectrum spp. (14%). In Keywords Habitat complexity Æ Multidimensional
2002 we identified 36 species, and the dominant scaling Æ Species diversity
species was tomtate, Haemulon aurolineatum (36%),
followed by Diplectrum spp. (19%), and L. camp-
echanus (13%). In 2001, species diversity and Introduction

The influence of habitat complexity on species

M. E. Lingo
diversity has been studied extensively by community
Texas Parks and Wildlife Department, 95 Fish Hatchery
Road, Brownsville, TX 78520, USA ecologists. They have sought correlations among
various components of habitat complexity and spe-
S. T. Szedlmayer (&) cies diversity with most work based on terrestrial
Marine Fish Laboratory, Department of Fisheries and
systems. These studies have suggested a positive
Allied Aquacultures, Auburn University, 8300 State Hwy.
104, Fairhope, AL 36532, USA correlation between habitat complexity and species
e-mail: sszedlma@acesag.auburn.edu diversity. Rozensweig and Winakur (1969) found that

123
72 Environ Biol Fish (2006) 76:71–80

rodent species diversity could be accounted for by treatments, Talbot et al. (1978) used four replicates
habitat complexity. Similar results were found for of four treatments, and Fujita et al. (1996) used one
birds (MacArthur and MacArthur 1961; Roth 1976), replicate of three treatments.
insects (Murdoch et al. 1972), and lizards (Pianka Bonhsack and Sutherland (1985) stressed the need
1967). for controlled experimentation in marine ecological
In recent years, increased concern over habitat and studies and suggested that artificial habitats could
species loss in marine systems has prompted inves- provide an ideal tool for testing habitat complexity
tigations into the relation between habitat and species hypotheses. They are easy to construct, manipulate,
diversity in marine systems. Habitat complexity was survey, and allow for controlled experimentation with
an important factor in settlement and habitat prefer- replication (Gratwicke and Speight 2005a). Equally
ences for sponges, corals, and gastropods (Russ 1980; important for the practical aspects of ecological
Carleton and Sammarco 1987; Beck 2000), but re- experimental design is that marine communities settle
sults differed when fish diversity and habitat com- quickly, often within days of structure placement
plexity were compared. For example, several studies (Shulman 1984).
have shown a positive correlation between substrate In this study, we substantially increased replication
complexity and fish diversity (Risk 1972; Roberts and (n=20 per treatment) by using artificial habitats.
Ormond 1987; McClanahan 1994; McCormick 1994; Specifically, we tested the null hypothesis that in-
Szedlmayer and Able 1996; Gratwicke and Speight creased habitat complexity does not affect reef fish
2005a, b), while one study showed a positive corre- species diversity in the northeastern Gulf of Mexico.
lation between habitat complexity and biomass, but
not with species diversity (Carpenter et al. 1981). In
contrast, other studies have shown no correlation Materials and methods
(Bourget et al. 1994; Gray 1994; Caballero and
Schmitter-Soto 2001) or conflicting results between The study site was 21 km southeast of Dauphin Is-
study sites (Luckhurst and Luckhurst 1978; Ohman land, Alabama (Fig. 1). The area lacked natural reef
and Rajasuriya 1998).
Increased habitat complexity may reduce preda-
tion pressures by providing refuge and escape routes
from predators. Rooker et al. (1998), suggested that
habitat complexity was linked to survival of newly
settled red drum, Sciaenops ocellatus, and Beukers
and Jones (1997) found that habitat complexity
modified predator impact on lemon damsel, Poma-
centrus moluccensis. Similar results were found for
fish communities associated with seagrass meadows
(Hindell et al. 2000). Complex habitats may also al-
low coexistence of competing species through
reduction of territorial competition (Sale 1980).
Most previous studies investigating the correlation
between habitat complexity and marine fish com-
munity structure have been descriptive and conducted
on natural reef systems (McCormick 1994; Beukers
and Jones 1997; Ohman and Rajasuriya 1998) with
few controlled experimental studies (Talbot et al.
1978; Hixon and Beets 1989; Fujita et al. 1996).
While these studies have contributed significantly to
understanding reef fish communities, a persistent
difficulty has been low replication. For example, Fig. 1 Location of artificial habitats sites in the northeastern
Hixon and Beets (1989) used two replicates of four Gulf of Mexico

123
Environ Biol Fish (2006) 76:71–80 73

structure, was uniform in depth (20 m), and domi- (a)

nated by sand-mud substrates (Schroeder et al. 1995).
From 24 July to 28 August 2001, we built 1 m2
habitats with 20 replicates of five complexity levels
(n=100). The following year we built new habitats 25cm
from 12 June to 31 July using the 2001 design, but
1m
also added 20 replicates of a sixth complexity level
(n=120). These 220 artificial habitats included: 40 1m
cage (1·1·0.25 m) only habitats made of galvanized
wire with 5·10 cm openings (cage), 40 oyster shell (b)
habitats (1·1·0.1 m, shell), 40 cage on shell habitats
(cage-shell), 40 block on shell habitats (shell with
four 20·20·40 cm concrete blocks, block-shell), 40
cage on block on shell habitats (cage-block-shell), m
and 20 habitats that each had 8 concrete blocks
stacked in a pyramid shape on top of shell (pyramid-
block-shell, Fig. 2). Complexity levels were based on
the surface area that each different component added
to each m2 plot. Block surface area was calculated
1m
from simple linear measures (each block = 5129 cm2)
while shell surface areas were measured using an (c)
image analyses system (total shell surface area for
each habitat = 128,940 cm2, Image-Pro Plus1). Also,
the wire cage added substantially to the habitat 812 mm

structure, but only added 162-cm2 surface area. Thus,

habitats were ranked in order of increasing
complexity as: cage, shell, cage-shell, block-shell,
cage-block-shell, and pyramid-block-shell. We built 1m
habitats in 200 m transects with each habitat 20 m Fig. 2 Examples of three complexity levels and materials
apart, with alternating complexity levels. Two SCU- used: (a) cage, (b) block-shell, and (c) pyramid-block-shell
BA divers constructed habitats by placing a 1 m2 habitats. Other habitat types were combinations of these
components. All habitat types contained shell except cage
PVC square on the bottom and spreading oyster shells
and concrete blocks within the square.
Fish populations were visually surveyed by SCU- Wiener diversity index [H0 =R-pi ln(pi)], species
BA divers trained in reef fish species identification. richness (S), and evenness (J). For each year analysis
Each habitat was surveyed completely and fish were of variance (ANOVA) was used to compare com-
identified to species. Species were also video taped munity measures among habitat types, and Tukey’s
for later verification. Transient schools of blue run- multiple comparison test was used to show specific
ner, Caranx crysos, crevalle jack, Caranx hippos, differences. Abundance of the common species (>1%
round scad, Decapterus punctatus, nurse shark, of the total) were compared among habitat types for
Ginglymostoma cirratum, greater amberjack, Seriola each year using ANOVA, and Student-Neuman–
dumerili, S. ocellatus, and gulf flounder, Paralichthys Keuls multiple comparison test (Zar 1984). To show
albigutta, were not included in the reef fish analysis. seasonal trends for 2001, mean abundances were
Mean abundances for each fish species were compared across habitat types within each of five
calculated for each habitat, and fish communities time periods (16–22 August, 29–30 August, 18–21
were compared with community measures: Shannon– September, 24 October–8 November, 19 November),
with ANOVA and Student-Neuman–Keuls multiple
comparison test (Zar 1984). Surveys outside these
1
MediaCybernetics, Silver Spring, MD 20910 USA time periods were not used in seasonal comparisons.

123
74 Environ Biol Fish (2006) 76:71–80

Reefs were only sampled once within each time Orthopristis chrysoptera (5%, Table 1). These five
period, but not all reefs were surveyed within each species accounted for 88% of the fish surveyed. Most
time period. However, nearly equal replicates of each of the common species (>1% of the total abundance,
reef type were sampled within each time period. 10 of 12) preferred the more complex habitats. Five
To compare the whole fish community among species L. campechanus, O. chrysoptera, lane snap-
habitat types we calculated Bray–Curtis similarity per, Lutjanus synagris, gray triggerfish, Balistes
coefficients among all habitats based on the root capriscus, and cubbyu, Pareques umbrosus, showed
transformed abundance of each species for each year. significantly increased abundance with increased
We then used non-parametric Multidimensional complexity from cage or shell, to cage-shell, block-
Scaling (MDS) to plot the distances in two- shell, and cage-block-shell (Table 1). Five other
dimensional space (Field et al. 1982). One-way species, C. philadelphica, S. caprinus, bank sea bass,
analysis of similarities (ANOSIM) permutation tests Centropristis ocyurus, pygmy filefish, Monacanthus
were used for both global and pair-wise testing for setifer, and pinfish, Lagodon rhomboides, also
significant differences among habitat types (Clarke showed significantly higher abundances on more
and Green 1988). All statistical differences were complex habitats, but appeared to separate on only
considered significant at P < 0.05. two levels, high complexity (cage-block-shell, block-
shell, cage-shell) compared to low complexity (shell
or cage). In 2001, only two species appeared to prefer
Results less complex habitats, Diplectrum spp. and pygmy
sea bass, Serraniculus pumilio, which showed sig-
All species identified during 2001 and 2002 surveys nificantly greater abundance on shell habitats com-
were reef fish common to the northeastern Gulf of pared to other habitats.
Mexico (Hoese and Moore 1977; McEachran and After reef deployment (13–55 days) we completed
Fechhelm 1998, 2006). We completed 368 individual 119 (one reef was missed) individual visual surveys
visual surveys from August to November 2001, and during August 2002, after which habitats were de-
identified 26 fish species. Red snapper, Lutjanus stroyed by tropical storms. We identified 35 fish
campechanus, was the dominant species (41%), fol- species, and tomtate, Haemulon aurolineatum, was
lowed by rock sea bass, Centropristis philadelphica the dominant species (36%), followed by Diplectrum
(23%), sand perch, Diplectrum spp. (14%), longspine spp. (19%), L. campechanus (13%), P. umbrosus
porgy, Stenotomus caprinus (6%), and pigfish, (5%), and O. chrysoptera (4%, Table 2). These five

Table 1 Mean abundance (mean) of fish per 1 m2 and percent abundance (%) for species >1% of the total abundance on artificial
habitats in the northeastern Gulf of Mexico for 2001
Species Mean Percent (%) Habitat type
CBS BS CS S C

L. campechanus 16.2 41 28.5a 21.9b 17.2c 6.9d 4.7d

C. philadelphica 8.9 23 15.3a 10.9a 10.8a 5.0b 0.1c
Diplectrum spp. 5.6 14 4.8b 5.4b 7.4b 10.3a 1.1c
S. caprinus 2.5 6 3.7a 3.8a 2.6ab 1.5bc 0.1c
O. chrysoptera 1.9 5 6.0a 2.8b 0.8c 0.4c 0.1c
C. ocyurus 1.0 3 2.1a 1.6ab 1.2b 0.4c 0.0c
L. synagris 0.8 2 2.0a 1.1b 0.5c 0.2c 0.2c
B. capriscus 0.6 2 1.7a 0.9b 0.5c 0.2cd 0.0d
P. umbrosus 0.4 1 1.1a 0.6b 0.1c 0.1c 0.0c
M. setifer 0.3 1 0.6a 0.5a 0.4ab 0.2b 0.1b
S. pumilio 0.3 1 0.1b 0.1b 0.1b 1.4a 0.0b
L. rhomboides 0.2 1 0.4a 0.2ab 0.2ab 0.0b 0.1b
Habitat types: cage-block-shell = CBS, block-shell = BS, cage-shell = CS, shell = S, and cage = C. Significant differences (P < 0.05)
in abundance for individual species among habitat types are shown by different letters

123
Environ Biol Fish (2006) 76:71–80 75

Table 2 Mean abundance (mean) of fish per 1 m2 and percent abundance (%) for species >1% of the total abundance on artificial
habitats in the northeastern Gulf of Mexico for 2002
Species Mean Percent (%) Habitat type
PBS CBS BS CS S C

H. aurolineatum 25.68 36 42.2a 53.1a 27.3ab 33.5ab 0.7b 0.0b

Diplectrum spp. 13.93 19 13.2 11.4 10.6 13.1 19.3 16.3
L. campechanus 9.57 13 10.7abc 14.4a 12.7ab 8.4bc 7.7c 3.8d
P. umbrosus 3.82 5 3.3bcd 8.7a 6.0b 3.7cb 1.3cd 0.2d
O. chrysoptera 3.02 4 7.0a 8.2a 2.1b 1.1b 0.0b 0.3b
S. caprinus 2.46 3 1.1 1.5 5.7 3.5 2.4 0.5
C. philadelphica 2.44 3 1.5b 3.4ab 2.9ab 4.9a 1.2b 1.0b
B. capriscus 2.04 3 4.1a 4.6a 2.5ab 1.0b 0.2b 0.1b
S. pumilio 1.68 2 1.3b 1.5b 2.2ab 2.0b 3.1a 0.1c
Halichoeres spp. 1.44 2 1.8ab 1.2ab 1.5ab 1.5ab 2.8a 0.0b
M. setifer 1.40 2 1.2 2.1 1.5 2.1 0.7 0.7
C. ocyurus 1.17 2 0.6 1.2 1.8 1.2 2.0 0.1
P. variabilis 0.96 1 1.7a 1.3ab 1.2ab 1.0ab 0.6cb 0.0c
L. synagris 0.81 1 2.6a 0.6b 0.4b 0.4b 0.9b 0.1b
Habitat types: pyramid-block-shell = PBS, cage-block-shell = CBS, block-shell = BS, cage-shell = CS, shell = S, and cage = C.
Significant differences (P < 0.05) in abundance for individual species among habitat types are shown by different letters

species accounted for 77% of the fish surveyed. The

additional level of complexity (pyramid) added only
two new species to the 2002 species richness, blue
runner, Caranx crysos, and cobia, Rachycentron 2.0 (a)
a
Diversity (H`)

canadum. Also, after excluding the fishes observed 1.6

a
b
on pyramids, dominant species showed little change, b

H. aurolineatum (33%), Diplectrum spp. (21%), 1.2

L. campechanus (14%), P. umbrosus (5%), and c
0.8
O. chrysoptera (3%). Similar patterns of fish abun-
dance compared to habitat complexity were observed
14 (b)
in 2002, but patterns were less distinct (not significant a
12
Richness (s)

a
across all levels). Complexity levels tended to over-
10
lap, but still showed the general trend of increased b
8 b
abundance with increased complexity. Surprisingly,
only one species L. synagris, showed significantly 6

higher abundance on the pyramid-block-shell habitats 4 c

(the most complex habitat). (c)

Significant differences in the community measures a a
Evenness (J)

0.70 a a
of species diversity and richness were detected a

among habitat types. In 2001, we detected a signifi-

cant decrease in diversity and richness as complexity 0.65

decreased from the more complex block-shell and

cage-block-shell habitats (H¢=1.7, S=11.1–11.9) to 0.60
intermediate complexity shell and cage-shell habitats Cage Shell Caged shell Block Caged block
Habitat type
(H¢=1.3–1.4, S=7.5–8.5) to the least complex cage
habitat (H¢=0.7, S=3.6). No significant differences Fig. 3 Mean fish community measures by habitat type for
were detected for evenness among habitat types 2001: (a) Shannon–Wiener diversity index + SE, (b) richness +
SE, and (c) evenness + SE. Significant differences (P < 0.05) in
(Fig. 3). Global and pair-wise ANOSIM tests community measures among habitat types are shown by
detected significant differences among all habitat different letters

123
76 Environ Biol Fish (2006) 76:71–80

Fig. 4 MDS plot, based on

Bray–Curtis similarity 1
coefficients among all
habitats from root
transformed abundance in
2001
0

-1 Cage
Shell
Cage-shell
Block-shell
Cage-block-shell

-2

-3 -2 -1 0 1

types in 2001, and visual analysis of the MDS ordi- habitat type, with cage and shell showing separate
nation plot of Bray–Curtis similarity coefficients groups compared to the higher complexity levels
showed a clear separation by habitat type (Fig. 4, (Fig. 8). The higher complexity levels (pyramid-
Table 3). block-shell, cage-block-shell, block-shell, and cage-
Reef fish quickly recruited to the artificial reefs. At shell) tended to show more overlap compared to the
the end of August 2001, approximately 1 month after 2001 MDS plots. However, global and pair-wise
reef deployment, mean abundance peaked for most ANOSIM tests again detected significant (P < 0.05)
habitats (Fig. 5). Only cage habitats showed a later differences among most (86%) habitat type pair-wise
peak in mid-September. A general trend of greater comparisons in 2002 (Table 3).
abundance on higher complexity reefs was observed
over all sample periods although significance varied.
An exception was a consistent overlap in mean
abundance between block and cage-shell reefs Table 3 One-way global and pair-wise analysis of similarities
(Fig. 5). After the fish abundance peaked in the sec- test among reef types
ond sample period (29–30 August 2001) there was a
Pair-wise comparisons R statistic 2001 R statistic 2002
subsequent decline in abundance on most reef types
into the fall. Species richness and diversity showed Block, Caged block 0.155 0.015 ns
later peaks in mid-September compared to abundance Block, Cage 0.934 0.685
Block, Shell 0.579 0.335
(Table 4). Patterns of individual species abundance
Block, Caged Shell 0.180 0.011 ns
showed similar peaks near the end of August for the Caged block, Cage 0.970 0.782
two dominant species L. campechanus and C. phila- Caged block, Shell 0.763 0.581
delphica (Fig. 6). Caged block, Caged shell 0.446 0.082
Cage, Shell 0.804 0.387
Significant differences in species diversity and
Cage, Caged shell 0.863 0.559
richness were again detected among habitats in 2002. Shell, Caged shell 0.304 0.236
Cage habitats had significantly lower diversity Block, Pyramid 0.124
(H¢=0.7) compared to all other habitats (H¢=1.5–1.7). Caged block, Pyramid 0.085
Cage, Pyramid 0.735
Significantly lower richness values were detected for
Caged shell, Pyramid 0.153
cage (S=3.2) and shell (S=7.5) compared to all other
habitats (S=9.3–10.5). Only cage (J=0.6) and shell Global test among habitat types for 2001, Global R=0.545, and
for 2002, Global R=0.357. Both Global R’s and the following
(J=0.8) habitats differed in evenness (Fig. 7). In
pair-wise tests showed significant differences (P < 0.05) be-
2002, the MDS analysis again showed separation by tween all pairs except those labeled ns

123
Environ Biol Fish (2006) 76:71–80 77

Fig. 5 Total mean number 120

– SE per m2 for all fishes by
habitat type and sample a
period from artificial reefs 100
in the northeastern Gulf of Block-cage
Mexico. Significant
a

Mean number per m2

differences (P < 0.05) in
80
mean abundance among a
habitat types are shown by a
a a
different letters within each a
time period 60
Shell-cage a
ab a Block
b a
40 b
b
Shell
b b b
20 cb
c
c Cage b cb
c c c
0
g g p v v
Au Au Se No No
22 30 21 t-8 19
16 - 29 - 18- 4O
c
2
Sample period

Table 4 Species richness and diversity (H¢) by habitat type and sample period in the northeastern Gulf of Mexico for 2001
Sample period n Species richness habitat type Diversity=H¢ habitat type
CBS BS CS Shell Cage CBS BS CS Shell Cage

16–22 Aug. 49 6.2 6.0 5.4 4.6* 0* 1.3 1.4 1.4 1.3 –
29–30 Aug. 48 6.8* 5.4 4.8 4.5 1.2* 1.2 1.2 1.1 1.2 0.4*
18–21 Sep. 50 7.6 8.0 6.1 5.0* 3.3* 1.5 1.5 1.4 1.3 0.8*
24 Oct.–8 Nov. 44 7.1* 4.7 4.9 4.0 1.9* 1.5* 1.1 1.2 1.1 0.4*
19 Nov. 29 6.8 5.4 4.0* 4.4 1.7* 1.3 1.3 1.0 1.3 0.4*
Habitat types: cage-block-shell = CBS, block-shell = BS, cage-shell = CS. n = number of individual reef surveys. *Significant
differences (P < 0.05)

Discussion artificial habitats. Apparent preferential selection of

habitat type was noted for several species. For
The habitat complexity and reef fish diversity pat- example: L. campechanus, O. chrysoptera, L. syna-
terns in the present study were similar to patterns gris, B. capriscus, and P. umbrosus, all showed sig-
observed in several other marine studies (Risk 1972; nificantly increased abundance at each level of
Roberts and Ormand 1987; Etter and Grassle 1992; complexity (Table 1). An interesting observation was
McCormick 1994; McClanahan 1994; Beck 2000), as the increased abundance of S. caprinus with in-
well as the classical studies in terrestrial systems creased complexity on two levels, high complexity
(MacArthur and MacArthur 1961; Pianka 1967; (cage-block-shell, block-shell, cage-shell) vs. low
Rozensweig and Winakur 1969; Murdoch et al. 1972; complexity (shell or cage). This species is usually
Roth 1976). Colonization of artificial habitats by considered an open water habitat species, yet still
juvenile reef fish occurred rapidly. The ten most showed significant attraction to more complex habitats.
abundant species by habitat type were observed Shulman (1984) suggested that reef fish prefer
within 3–4 weeks of habitat placement. Talbot et al. habitats with hole sizes near their own body size and
(1978) and Shulman (1984) also observed the present study supported this relation. Larger reef
rapid recruitment of juveniles to newly constructed fish, L. campechanus, O. chrysoptera, L. synagris,

123
78 Environ Biol Fish (2006) 76:71–80

Fig. 6 Mean number per m2 for

dominant fish species by sample L.campechanus
50 Diplectrum spp.
period from artificial reefs in the
northeastern Gulf of Mexico S.caprinus
C.philadelphica
40 O.chrysoptera

Mean number per m2

C.ocyurus
L.synagris
B.capriscus
30

20

10

0
Au
g
Se
p
Se
p ov No
v
22 -6 21 7N 9
16 - g 18 - Oc
t- 8-1
Au 23
29
Sample period

B. capriscus, and P. umbrosus, were significantly

more abundant on habitats that offered larger hole (a) a a
a
a
1.6 a
Diversity (H`)

sizes (i.e., blocks). Also, newly settled H. auroline-

atum, and L. campechanus were found on all habitat 1.2
types where oyster shell rubble provided hole sizes
b
that matched their body sizes. We suggest that the 0.8
main function of increased complexity was increased
12 (b)
predation refuge, for example mesh size of cages was a a
Richness (S)

a
10 a
small enough to exclude large predators, but large
8 b
enough to allow easy access to habitat by small fish
common to shelf habitat in the northeastern Gulf of 6
Mexico. Previous studies have predicted similar 4 c
predation effects on juvenile reef fish (Jones 1991;
Carr and Hixon 1995; Connell 1997). (c)
a
During this study there were several major storm 0.8
Evenness (J)

ba ba
ba
events that destroyed habitats. In 2001, a series of 0.7
ba
b
winter storms destroyed habitats after four surveys
were completed. Then in 2002, tropical storm Hanna 0.6
destroyed habitats after one complete survey. Previ-
0.5
ous studies have concluded that full recruitment may C S CS BS CBS PBS
take several months (Bonhsack and Talbot 1980), and Habitat type
we suggest that the less distinct patterns in 2002 Fig. 7 Fish community measures by habitat type for 2002:
compared to 2001, were due to less time allowed for (a) Shannon–Wiener diversity index + SE, (b) richness + SE,
reef fish recruitment. A difficult pattern to explain is and (c) evenness + SE. Cage = C, shell = S, cage-shell = CS,
the increased number of species (35) observed in block-shell = BS, cage-block-shell = CBS, and pyramid-
block-shell = PBS habitats. Significant differences (P < 0.05)
2002 compared to 2001 (26 species), even though we in community measures among habitat types are shown by
had fewer samples and less time for fish recruitment different letters

123
Environ Biol Fish (2006) 76:71–80 79

Fig. 8 MDS plot, based on

Bray–Curtis similarity 2 Cage
Shell
coefficients among all habitats
Cage-shell
from root transformed Block-shell
abundance in 2002 Cage-block-shell
1 Pyramid-block-shell

-1

-1 0 1 2

in 2002 compared to 2001. It is well recognized that complexity is not clear. Future research should at-
fish recruitment is highly variable from year to year tempt to separate these components of prey avail-
which may account for the different species in these ability from complexity and shelter effects.
2 years, but the fact that diversity and complexity
patterns were persistent between these two different Acknowledgements We thank M. Boles, C. Furman,
years further validates this relation. A. Chapin, A. Ouzts, A. Piko, R. Schroepfer, and B. Wild-
Artificial habitats provided an ideal tool for com- berger for field assistance. This project was funded by Marine
Resources Division, Alabama Department of Conservation and
parisons of community structure across different
Natural Resources. This is a contribution of the Alabama
levels of habitat complexity. They allowed for rep- Agricultural Experiment Station, Auburn University.
lication, were easily manipulated and of known age
and type. Previous studies have incorporated various
levels of complexity, but few replicates (Talbot et al. References
1978; Hixon and Beets 1989). We used 20 replicates
each of five complexity levels in 2001, and added 20 Beck MW (2000) Separating the elements of habitat structure:
independent effects of habitat complexity and structural
replicates of a sixth complexity level in 2002. Greater components on rocky intertidal gastropods. J Exp Mar
replication allowed increased power in testing the Biol Ecol 249:29–49
effects of habitat complexity on fish species compo- Beukers JS, Jones GP (1997) Habitat complexity modifies the
sition. Habitat complexity appears to play an impor- impact of piscivores on a coral reef fish population.
Ecology 114:50–59
tant role in reef fish community structure. In 2001 Bonhsack JA, Talbot FH (1980) Species-packing by reef fishes
there was a clear relation between habitat complexity on Australian and Caribbean reefs: an experimental ap-
and fish diversity. We also detected a significant proach. Bull Mar Sci 30:710–723
complexity effect on fish diversity in 2002 even Bonhsack JA, Sutherland DL (1985) Artificial reef research: a
review with recommendations for future priorities. Bull
though we were only able to complete one survey. Mar Sci 37:11–39
One possible caveat with the present study was that Bourget E, Degauss J, Daigle G (1994) Scales of substratum
as habitat complexity increases, food resources may heterogeneity, structural complexity, and the early estab-
also increase. In contrast, Russ (1980) showed that lishment of a marine epibenthic community. J Exp Mar
Biol Ecol 181:31–51
increased complexity reduced fish feeding efficiency Caballero JA, Schmitter-Soto JJ (2001) Diversity of fishes in
on marine epifaunal communities, which suggest small coral patches of the Mexican Caribbean. Bull Mar
that increased prey shown with increased complexity Sci 68:337–342
may be negated by decreased foraging ability on the Carleton JH, Sammarco PW (1987) Effects of substratum
irregularity on success of coral settlement: quantification
part of predatory fishes. Thus, the relations of fish by comparative geomorphological techniques. Bull Mar
diversity to increased food resources and habitat Sci 40:85–98

123
80 Environ Biol Fish (2006) 76:71–80

Carpenter KE, Miclat RI, Albaladejo VD, Corpuz VT (1981) McClanahan TR (1994) Kenyan coral reef lagoon fish: effects
The influence of substrate structure on the local abun- of fishing, substrate complexity, and sea urchins. Coral
dance and diversity of Philippine reef fishes. The reef and Reefs 13:231–241
man. Proc Fourth Int Corel Reef Symp 2:497–502 McCormick MI (1994) Comparison of field methods for
Carr MH, Hixon MA (1995) Predation effects on early post- measuring surface topography and their associations with
settlement survivorship of coral-reef fishes. Mar Ecol Prog a tropical reef fish assemblage. Mar Ecol Prog Ser
Ser 124:31–42 112:87–96
Clarke KR, Green RH (1988) Statistical design and analysis for McEachran JD, Fechhelm JD (1998) Fishes of the Gulf of
a ‘biological effects’ study. Mar Ecol Prog Ser 46:213– Mexico. Volume 1: Myxiniformes to Gasterosteiformes.
226 University of Texas Press, Austin pp 1112
Connell SD (1997) The relationship between large predatory McEachran JD, Fechhelm JD (2006) Fishes of the Gulf of
fish and recruitment and mortality of juvenile coral reef- Mexico. Volume 2: Scorpaeniformes to Tetraodontifor-
fish on artificial reefs. J Exp Mar Biol Ecol 209:261–278 mes. University of Texas Press, Austin pp 1008
Etter RJ, Grassle JF (1992) Patterns of species diversity in the Murdoch WW, Evans FC, Peterson CH (1972) Diversity and
deep sea as a function of sediment particle size diversity. patterns in plants and insects. Ecology 53:819–829
Nature 360:576–578 Ohman MC, Rajasuriya A (1998) Relationships between hab-
Field JG, Clarke KR, Warwick RM (1982) A practical strategy itat structure and fish communities on coral and sandstone
for analyzing multispecies distribution patterns. Mar Ecol reefs. Environ Biol Fishes 53:19–31
Prog Ser 8:37–52 Pianka ER (1967) On lizard species diversity: North American
Fujita T, Kitagawa D, Okuyama Y, Jin Y, Ishito Y, Inada T flatland deserts. Ecology 48:333–351
(1996) Comparison of fish assemblages among an artifi- Risk MJ (1972) Fish diversity on a coral reef in the Virgin
cial reef, a natural reef and a sandy-mud bottom site on Islands. Atoll Res Bull 193:1–6
the shelf off Iwate, northern Japan. Environ Biol Fishes Roberts CM, Ormond RFG (1987) Habitat complexity and
46:351–364 coral reef fish diversity and abundance on Red Sea
Gratwicke B, Speight MR 2005a. Effects of habitat complexity fringing reefs. Mar Ecol Prog Ser 41:1–8
on Caribbean marine fish assemblages. Mar Ecol Prog Ser Rooker JR, Holt GJ, Holt SA (1998) Vulnerability of newly
292:301–310 settled red drum (Sciaenops ocellatus) to predatory fish: is
Gratwicke B, Speight MR 2005b. The relationship between fish early-life survival enhanced by seagrass meadows? Mar
species richness, abundance and habitat complexity in a Biol 131:145–151
range of shallow tropical marine habitats. J Fish Biol Roth RR (1976) Spatial heterogeneity and bird species diver-
66:650–667 sity. Ecology 57:773–782
Gray JS (1994) Is deep-sea species diversity really so high? Rozensweig ML, Winakur J (1969) Population ecology of
Species diversity of the Norwegian continental shelf. Mar desert rodent communities: habitats and environmental
Ecol Prog Ser 112:205–209 complexity. Ecology 50:558–572
Hindell JS, Jenkins GP, Keough MJ (2000) Evaluating the Russ GR (1980) Effects of predation by fishes, competition and
impact of predation by fish on the assemblage structure of structural complexity of the substratum on the establish-
fishes associated with seagrass (Heterozoztera tasmanica) ment of a marine epifaunal community. J Exp Mar Biol
(Martens ex Ascherson) den Hartog, and unvegetated sand Ecol 42:55–69
habitats. J Exp Mar Biol Ecol 255:153–174 Sale PF (1980) The ecology of fishes on coral reefs. Ocean-
Hixon HA, Beets JP (1989) Shelter characteristics and Carib- ography and Marine Biology. Annu Rev 18:367–421
bean fish assemblages: experiments with artificial reefs. Schroeder WW, Shultz AW, Pilkey OH (1995) Late quaternary
Bull Mar Sci 44:666–680 oyster shells and sea-level history, inner shelf, northeast
Hoese HD, Moore RH (1977) Fishes of the Gulf of Mexico, Gulf of Mexico. J Coastal Res 11:664–674
Texas, Louisiana, and adjacent water. Texas A&M Uni- Shulman MJ (1984) Resource limitation and recruitment pat-
versity Press, College Station pp 327 terns in a coral reef fish assemblage. J Exp Mar Biol Ecol
Jones GP (1991) Postrecruitment processes in the ecology of 74:85–109
coral reef fish populations: a multifactorial perspective. Szedlmayer ST, Able KW (1996) Patterns of seasonal avail-
In: Sale PF (eds) The ecology of fishes on coral reefs. ability and habitat use by fishes and decapod crustaceans
Academic press, San Diego pp 294–328 in a southern New Jersey estuary. Estuaries 19:697–709
Luckhurst BE, Luckhurst K (1978) Analysis of the influence of Talbot FH, Russel BC, Anderson GRV (1978) Coral reef fish
substrate variables on coral reef fish communities. Mar communities: unstable, high-diversity systems? Ecol
Biol 49:317–323 Monogr 48:425–440
MacArthur RH, MacArthur J (1961) On bird species diversity. Zar JH (1984) Biostatistical analysis, 2nd edn, Prentice-Hall,
Ecology 42:594–598 Englewood Cliffs, NJ pp 718

123