You are on page 1of 18

Hydrobiologia (2006) 563:45–60  Springer 2006

DOI 10.1007/s10750-005-1428-2

Primary Research Paper

Different surrounding landscapes may result in different fish assemblages

in East African seagrass beds

M. Dorenbosch, M.G.G. Grol, I. Nagelkerken* & G. van der Velde

Department of Animal Ecology and Ecophysiology, Institute for Water and Wetland Research, Faculty of Science,
Radboud University Nijmegen, Toernooiveld 1, 6525 ED, Nijmegen, The Netherlands
(*Author for correspondence: E-mail:

Received 3 June 2005; in revised form 13 October 2005; accepted 28 October 2005; published online 11 April 2006

Key words: coral reef fishes, seagrass beds, mangroves, nursery, connectivity, migration

Few studies have considered how seagrass fish assemblages are influenced by surrounding habitats. This
information is needed for a better understanding of the connectivity between tropical coastal ecosystems.
To study the effects of surrounding habitats on the composition, diversity and densities of coral reef fish
species on seagrass beds, underwater visual census surveys were carried out in two seagrass habitat types at
various locations along the coast of Zanzibar (Tanzania) in the western Indian Ocean. Fish assemblages of
seagrass beds in a marine embayment with large areas of mangroves (bay seagrasses) situated 9 km away
from coral reefs were compared with those of seagrass beds situated on the continental shelf adjacent to
coral reefs (reef seagrasses). No differences in total fish density, total species richness or total juvenile fish
density and species richness were observed between the two seagrass habitat types. However, at species
level, nine species showed significantly higher densities in bay seagrasses, while eight other species showed
significantly higher densities in reef seagrasses. Another four species were exclusively observed in bay
seagrasses. Since seagrass complexity could not be related to these differences, it is suggested that the
arrangement of seagrass beds in the surrounding landscape (i.e. the arrangement on the continental shelf
adjacent to the coral reef, or the arrangement in an embayment with mangroves situated away from reefs)
has a possible effect on the occurrence of various reef-associated fish species on seagrass beds. Fish
migration from or to the seagrass beds and recruitment and settlement patterns of larvae possibly explain
these observations. Juvenile fish densities were similar in the two types of seagrass habitats indicating that
seagrass beds adjacent to coral reefs also function as important juvenile habitats, even though they may be
subject to higher levels of predation. On the contrary, the density and species richness of adult fish was
significantly higher on reef seagrasses than on bay seagrasses, indicating that proximity to the coral reef
increases density of adult fish on reef seagrasses, and/or that ontogenetic shifts to the reef may reduce adult
density on bay seagrasses.

Introduction linked to adjacent habitats through fish migration.

Various studies have suggested ontogenetic
Seagrass beds are considered important habitats migrations of reef fishes from seagrass nurseries
for reef fish assemblages throughout the tropical towards adjacent coral reefs (McFarland et al.,
seas and may function as nursery, shelter and/or 1985; Kanashiro, 1998; Nagelkerken et al., 2000a;
feeding habitats (Pollard, 1984; Parrish, 1989; Beck Cocheret de la Morinière et al., 2002; Nakamura
et al., 2001). Seagrass habitats can be directly & Sano, 2004b), while other studies described the

occurrence of daily twilight or tidal migrations distance from the coral reef habitat. No in-depth
between seagrass beds and adjacent habitats studies have been published that compared fish
(Ogden & Buckman, 1973; Ogden & Ehrlich, 1977; assemblages on seagrass beds adjacent to coral
Burke, 1995; Nagelkerken et al., 2000b; Doren- reefs with those on seagrass beds located in shel-
bosch et al., 2004a). Because seagrass beds could tered marine bays away from coral reefs. It can be
function as an extra source of new adults for hypothesised that in situations where the distance
adjacent habitats (through its assumed nursery between coral reefs and seagrass beds is large,
function) or increase the feeding or sheltering regular migrations by large reef-associated pisci-
opportunities for fish from adjacent habitats, it is vores to seagrass beds are unlikely. As a result,
likely that the presence of seagrass beds positively predation pressure by reef-associated predators on
influences the fish assemblages of adjacent habitats. juvenile fish present on seagrass beds at a large
This has been confirmed by a few studies which distance from reefs may be considerably smaller
showed that the presence of seagrass and mangrove than that on seagrass beds adjacent to a coral reef.
nurseries influence fish assemblages on coral reefs This might result in increased survival of juvenile
(Nagelkerken et al., 2000a, 2002; Halpern, 2004; fish that recruit on seagrass beds situated away
Mumby et al., 2004; Dorenbosch et al., 2004b). from coral reefs (reviewed in Parrish, 1989).
Vice versa, adjacent habitats may also influence Several studies only showed weak interlinkages
the fish assemblage of seagrass beds themselves. between fish assemblages of mangroves and coral
Weinstein & Heck (1979), Baelde (1990) and reefs in the Indo-Pacific region (Quin & Kojis,
Kochzius (1997) showed that the presence of a 1987; Thollot & Kulbicki, 1988; Kimani et al.,
coral reef influenced the structure of the fish 1996; Huxham et al., 2004). It may therefore be
assemblages on adjacent seagrass beds, mainly as a possible that interlinkages between fish assem-
result of migration of fishes from the reef towards blages on seagrass beds adjacent to these man-
the seagrass habitat. Shulman (1985) and Sweat- groves (i.e. seagrass beds situated in estuaries or
man & Robertson (1994) suggested that the pres- marine bays) and coral reefs are also weak. Just as
ence of a coral reef resulted in high predation in the case for mangroves, fish assemblages on
pressure by reef-associated piscivores on juvenile these seagrass beds may show a different species
fish present on the seagrass zone adjacent to the composition and/or increased juvenile fish densi-
reef. The presence of a coral reef is therefore likely ties because of the absence of large predators.
to influence the structure of the fish assemblage of To our knowledge there are no studies within
adjacent seagrass beds. the Indian Ocean that investigated fish assemblages
Apart from these post-settlement processes, on seagrass beds in relation to the surrounding
also pre-settlement processes in fish larvae may be landscape of shallow water habitats such as coral
an important factor structuring fish assemblages in reefs or mangroves. On the East African island of
a landscape of tropical shallow water habitats. Zanzibar, seagrass beds can be distinguished that
Various environmental factors may influence are located on the continental shelf directly adja-
recruitment of fish larvae on reef habitats and cent to coral reefs, and seagrass beds that are
larvae may actively select their settlement habitat located in a marine embayment adjacent to large
(Sweatman, 1988; Jenkins et al., 1997; Atema mangrove areas. This provides an opportunity to
et al., 2002; Kingsford et al., 2002; Simpson et al., investigate the effect of the setting of these seagrass
2004; Lecchini et al., 2005; Nanami et al., 2005). beds with respect to the surrounding landscape on
Subsequently, spatial combinations of shallow the fish assemblages present on these seagrass beds.
water habitat types may result in unique environ- Due to these different surrounding landscapes, two
mental conditions that may direct recruitment different fish assemblages are expected on each type
processes of fish larvae and therefore influence the of seagrass bed. Reduced predation pressure and
structure of the fish assemblage. higher food availability in seagrass beds situated in
The few studies that focussed on interlinkages the embayment could result in higher total densities
between seagrass beds and other habitats have of juvenile fish. Subsequently, migration of fishes
investigated a seagrass – reef gradient where from the coral reef and differences in recruitment
seagrass beds are adjacent to or located at a short patterns due to different environmental variables as

a result of the setting in the landscape (e.g. in an Materials and methods

embayment with mangroves or on the continental
shelf adjacent to the reef), may result in different Study area
species assemblages.
Based on the above hypothesis and outlines, The present study was carried out off the island of
the following questions were asked in the present Zanzibar along the Tanzanian coast (East Africa) in
study: (1) is the density of juvenile reef fish on the western Indian Ocean (Fig. 1). Two types of
seagrass beds situated in the embayment with seagrass habitats were distinguished: (1) shallow
mangrove areas far away from the coral reef seagrass beds situated in a marine embayment far
greater than the density of juvenile fish on seagrass from coral reefs but adjacent to large areas of
beds situated on the continental shelf adjacent to mangroves and mud flats (three sites were surveyed,
coral reefs, and (2) is the species composition, hereafter called bay seagrasses), and (2) shallow
species richness and density of reef fishes on seagrass beds situated on the shallow continental
seagrass beds situated in the embayment different shelf directly adjacent to coral reefs and located far
from that on seagrass beds adjacent to coral reefs? away from mangroves and mud flats (four sites were

Figure 1. Overview of the island of Zanzibar (latitude 610¢ S, longitude 3910¢ E) Reef seagrasses: (1) Muyuni, (2) Chapwani, (3)
Changu, (4) Chumbe; Bay seagrasses: (5) Marumbi, (6) Tunduni, (7) Chwaka; Coral reefs: (1) Sensations, (2) Nungwi, (3) Leon’s wall,
(4) Mnemba atoll, (5) Kiwengwa, (6) Pongwe, (7) Uroa, (8) Jambiani, (9) Kizimkazi, (10) Adriana bank, (11) Boribu, (12) Tambare,
(13) Pwaku, (14) Chumbe, (15) Murogo, (16) Changu, (17) Timbatu bay, (18) Timbatu island, (19) Hanga.

Table 1. Mean fish densities (500 m)2) of species associated with bay seagrasses (species with significantly higher densities on
bay seagrasses than on reef seagrasses (tested with a t-test) or species observed only on bay seagrasses), and their mean densities
on 19 coral reef sites in a spatial gradient around Zanzibar

Reef Bay Western reefs

seagrasses seagrasses

Mean SD Mean SD p-value Adriana Pwaku Boribu Tambare Chumbe Morogo

density density Bank

Species associated with Bay seagrasses (n=9)

Cheilinus fasciatus 0.3 3.6 6.3 18.4 0.020 0.6 15.5 1.0 1.4 0.2 24.5
C. trilobatus 7.4 15.5 16.0 31.0 0.014 1.6 4.6 8.2 12.7 0.1 2.8
Chlorurus strongylocephalus 0.2 1.8 7.5 25.3 0.005 1.8 7.6 4.8 11.5 0.3
Hipposcarus harid 3.1 16.5 62.0 121.0 <0.001 2.1 0.3 0.8 0.0 0.8
Lethrinus harak 16.4 33.3 56.2 175.8 0.029 2.1 2.1 25.7 27.6 0.2 9.2
Lutjanus fulviflamma 7.9 31.9 71.1 145.7 <0.001 11.3 12.2 14.5 92.5 0.1 1.0
Parupeneus barberinus 27.6 49.6 47.5 70.7 0.019 4.9 2.6 11.4 7.2 0.1 5.8
P. rubescens 5.0 22.7 17.6 37.7 0.004 1.2 0.0 0.1
Scarus ghobban 13.6 34.5 212.9 345.8 <0.001 1.0 0.7 1.1 1.5 0.0 1.3
Species observed only on Bay seagrasses (n=4)
Cheilinus undulatus 11.1 22.9 <0.001 0.2 0.0
Lutjanus argentimaculatus 7.5 28.1 0.009 0.0
L. ehrenbergi 3.8 20.1 0.061
Parupeneus indicus 1.8 10.7 0.095 0.2 0.1 0.1 0.4

surveyed, hereafter called reef seagrasses). The Field sampling and study design
mean distance between the bay seagrasses and the
nearest coral reef was 9.2 km (minimum distance Reef fish species were selected on the basis of our
6.5 km), while the mean distance between the reef ability to identify them in underwater visual census
seagrasses and mangroves and mud flats was 8.5 km surveys, as well as on their abundance and com-
(minimum distance 3.8 km). The three bay seagrass mercial value for fisheries (Jiddawi & Stanley,
sites could be considered as independent from each 1997). All species belonging to the families
other since the minimum distance was 2.1 km and Gerreidae, Haemulidae, Lethrinidae, Lutjanidae,
each seagrass sites was distinguished from other Mullidae, Nemipteridae, Scaridae and Siganidae
seagrass site by large algae flats and tidal channels. were included in the present study, as well as two
All reef seagrasses bordered shallow (water depth species of Chaetodontidae (Chaetodon melannotus
<10 m) stony coral reefs and were in open con- and C. auriga), four species of large-sized Labri-
nection with the sea. In contrast, bay seagrasses dae (Cheilinus fasciatus, C. trilobatus, C. undula-
were located in a shallow marine embayment and tus and Cheilio inermis), and one species of
therefore sheltered from the open sea. Both types of Zanclidae (Zanclus cornutus). Large groups of
seagrass beds were dominated by Thalassodendron Scaridae that could not be identified under water
ciliatum (cover >60%), sometimes intermingled were regularly observed. Identification of individ-
with Enhalus acoroides, Thalassia hemprichii and uals of several of these groups that were caught
Cymodocea rotundata. Although the tidal range in with a net revealed that they consisted predomi-
the study area was large (3.3 m), seagrass beds never nantly of juvenile Chlorurus sordidus intermingled
fell dry at low tide (minimum water depth 2.2 m). with juvenile Calotomus spinidens and Leptosca-
Water depth during neap high tide was approxi- rus vaigiensis. In all further analyses, these Scari-
mately 4 m. dae were pooled as Scarus spp.

North western reefs Northern reefs Eastern reefs Southern


Changu Timbatu Hanga Timbatu Sensation Nungwi Leon’s Mnemba Kiwengwa Pongwe Uroa Jambiani Kizimkazi
island bay wall (atoll reef)

0.2 1.5 0.0 0.0 1.1 0.2 0.1 0.2 0.0 0.1
0.1 4.1 0.0 0.1 0.2 5.2 6.1 13.7 2.3 18.9 0.2 3.2 6.8
0.2 6.8 0.0 0.0 0.1 3.1 0.3 0.2 2.0 0.0 0.2 3.0
0.1 24.3 0.1 0.4 0.8 6.4 0.9 1.2 9.1 0.0 4.8 10.7
0.0 14.7 0.0 0.1 6.5 1.4 3.1 1.4 0.0 1.9
0.1 35.6 3.8 0.6 3.7 27.7 34.4 254.8 1.4 0.7 1.8 16.4 25.2
0.0 6.5 0.3 7.6 5.0 3.9 4.4 3.5 1.0 0.1 3.0 2.8
0.2 0.0 0.9 0.4 2.0 1.1 0.3 0.0 0.6
0.0 0.5 0.0 0.1 2.8 2.8 7.6 0.2 0.3 0.7 0.3 2.2 4.3

0.4 1.7 1.1 0.3 0.0 0.8

0.0 0.6 0.1 0.0
0.1 0.0

The selected reef fish assemblage was sampled conducted between 8:00 and 13:00 h at high neap
by means of underwater visual census with a tide when water movement was minimal. Species
stationary point-count method using SCUBA identification and quantification were first thor-
(Watson & Quinn, 1997; Polunin & Roberts, oughly practiced simultaneously by the two
2004). Because underwater visibility ranged observers until their results were comparable.
between 5 and 10 m, square quadrats of 55 m Estimation of size classes was trained by repeat-
were surveyed. Data were collected by two edly estimating the sizes of 50 pieces of electrical
observers who independently surveyed the quad- wire of known lengths, placed at random under
rats. A single rope with a length of 5 m was used water and representing all fish size classes.
as a visual reference for quadrat size. After Training was continued until differences in size
placing the line, the observer waited 3 min to estimation were minimal (i.e. a deviation of no
minimise fish disturbance. All target species more than ±2.5 cm from the actual length for
within or passing through the quadrat were objects <20 cm). Practicing quantification and
counted over a period of 10 min. During the first size estimation was repeated regularly during the
7 min, the observer was situated on the edge of visual census period. Habitats were surveyed in
the quadrat, while during the last 3 min, the ob- the beginning of southern hemisphere summer in
server moved through the quadrat to search for the period October–December 2004, around the
small juvenile fish hiding behind or under sea- start of the northeast monsoon, ca. three months
grass leaves or other structures. Care was taken before the beginning of the long rainy season. In
not to count individuals or groups of fish more total, we surveyed 121 replicate quadrats distrib-
than once that regularly moved in and out of the uted over four sites on reef seagrasses and 99
quadrat. Fish were classified into 2.5 cm size replicate quadrats distributed over three sites on
classes using an underwater slate. Surveys were bay seagrasses (Fig. 1).

Table 2. Mean fish densities (500 m)2) of species associated with reef seagrasses (species with significantly higher densities on reef
seagrasses than on bay seagrasses, t-test), and species without a clear habitat preference (species without significant different
fish densities between reef and bay seagrasses, t-test), and their mean densities on 19 coral reef sites in a spatial gradient around Zanzibar

Reef seagrasses Bay seagrasses Western reefs

Mean SD Mean SD p-value Adriana Pwaku Boribu Tambare Chumbe Morogo

density density Bank

Species associated with reef seagrasses (n=8)

Calotomus carolinus 11.9 25.9 3.8 19.7 0.009 3.7 0.5 0.1
Cheilio inermis 76.9 130.8 30.3 41.5 <0.001 2.2 0.1 2.1 4.0 0.9
Chlorurus sordidus 157.4 311.3 44.0 101.5 <0.001 25.0 22.6 15.0 20.6 1.2 34.4
Lethrinus mahsena 19.2 36.2 2.4 10.4 <0.001 0.8 0.6 0.1
L. variegatus 51.7 201.5 0.6 3.4 0.006 0.4
Lutjanus gibbus 10.7 46.0 2.0 11.2 0.046 0.1
Scarus. frenatus 2.8 11.0 0.2 2.0 0.012 1.1 0.4 3.2 0.3 3.1
Scarus spp. 239.3 602.4 28.7 112.5 <0.001
Species without a clear habitat preference (n=18)
Calotomus spinidens 7.6 36.1 3.0 14.1 NS
Chaetodon auriga 0.5 3.1 2.6 11.3 NS 2.6 2.2 0.8 1.4 0.1 1.0
C. melanotus 0.3 2.6 1.6 6.8 NS 8.0 6.6 3.1 2.4 10.9
Gerres oyena 4.3 47.3 0.6 3.4 NS
Leptoscarus vaigiensis 100.7 185.3 85.7 172.5 NS
Lethrinus lentjan 60.3 153.4 58.8 93.1 NS 4.5 0.6 16.7 17.8 0.1 0.9
Mulloidichthys flavolineatus 3.8 38.2 1.4 10.0 NS 0.9 0.1
Parupeneus macronema 44.1 69.0 29.9 55.4 NS 27.9 5.9 30.1 35.0 0.1 12.9
P. pleurostigma 2.5 17.2 0.6 6.0 NS 1.0 0.3 0.7 0.1
Plectorhinchus flavomaculatus 1.7 6.6 1.2 6.3 NS 0.5 1.4 0.0 0.2
P. gaterinus 3.1 10.3 6.7 17.4 NS 14.0 1.7 22.8 16.5 0.6 1.2
Scarus niger 2.1 18.4 0.8 4.9 NS 6.4 17.4 2.2 7.3 0.5 13.5
S. psittacus 4.1 16.5 18.8 122.9 NS 12.3 2.0 4.9 1.7 0.1 0.3
S. russelli 1.3 6.7 0.6 6.0 NS 6.7 4.7 8.9 0.0 0.6
S. scaber 3.6 18.6 1.6 8.9 NS 1.8 0.3 0.1 0.8 0.0 3.6
Scolopsis ghanam 49.3 147.9 27.9 57.6 NS 26.3 2.0 23.2 27.2 0.2 31.0
Siganus stellatus 6.0 47.2 0.8 6.3 NS 0.6 1.2 1.2 0.8 0.0 2.3
S. sutor 46.6 74.7 55.2 89.2 NS 1.3 0.4 1.8 0.1 0.1

NS: not significant.

Bay seagrasses were situated on the eastern possible effect, the distribution of 39 species were
side of Zanzibar while reef seagrasses were situ- compared among 19 different coral reefs located
ated on the western and north-eastern side of the around the entire coast of Zanzibar island (see
island. These differences in the spatial arrange- Fig. 1 for the locations of these coral reefs and
ment of the seagrass habitats could have an effect Tables 1 and 2 for species names). Since adults of
on the present fish assemblages, i.e., differences in most of the fish species that were investigated on
ocean current patterns in the waters around the seagrass beds also occur on the coral reef
Zanzibar island may result in an unbalanced (either as a result of ontogenetic migration or
recruitment of fish species on habitats off the local recruitment), a spatial effect due to differ-
coast of the island. This could result in presence ences in ocean currents that would influence the
or absence of species in either one of the seagrass recruitment of species on the two seagrass habitat
habitat types. To analyze the influence of this types, would likewise also be visible for the dis-

North western reefs Northern reefs Eastern reefs Southern reef

Changu Timbatu Hanga Timbatu Sensation Nungwi Leon’s Mnemba Kiwengwa Pongwe Uroa Jambiani Kizimkazi
island bay wall (atoll reef)

0.0 0.0 0.0 4.6 0.6 3.5 1.0 8.8 0.0 3.0 9.2
0.1 2.6 0.0 1.3 3.5 3.1 0.5 3.5 3.3 0.1 3.9 3.3
0.7 15.8 0.2 0.4 1.5 18.8 20.3 41.6 25.5 29.6 0.3 45.8 13.1
0.1 0.8 0.0 0.4 2.8 0.8 1.1 3.0 39.1 0.1 0.4 0.5
0.0 0.3 0.3
0.0 0.5 0.2 0.1 0.7 5.6 29.6 3.0 0.0 2.4 69.1
0.1 1.1 0.0 3.1 0.6 0.0 0.1

0.1 0.0
0.0 1.2 0.0 0.1 3.5 1.4 1.4 5.9 1.2 0.1 3.7 1.1
0.0 4.7 0.1 0.2 0.2 1.5 1.8 4.8 1.0 6.2 0.0 6.7 0.8
0.1 0.3 0.0
0.0 6.3 1.5 0.0 8.2 9.5 12.1 3.5 1.3 0.1 11.2 33.9
0.0 6.3 0.0 8.2 11.7 57.0 2.3 1.0 0.0 8.0
0.1 19.4 0.2 0.4 21.5 78.8 34.8 51.4 28.3 23.8 0.4 54.1 37.7
0.0 0.0 0.0 16.3 6.6 2.1 8.8 0.9 0.0 2.2 8.1
0.3 0.0 0.0 2.4 2.9 0.0 0.6 0.8
0.0 0.9 0.0 0.6 10.9 4.8 1.2 0.7 2.6 10.3
0.1 5.4 0.0 0.1 0.3 4.7 0.9 0.2 0.0 1.5
0.0 11.4 0.1 0.2 2.8 20.7 4.3 14.8 8.5 4.2 0.1 15.6 6.6
0.0 5.0 0.0 0.0 1.1 1.0 1.9 2.8 1.3 0.0 1.5 1.0
0.1 0.0 0.0 0.2 2.0 1.1 0.2 0.0 0.1
0.0 12.6 1.7 0.4 67.3 36.9 15.0 6.9 3.8 3.3 0.2 5.8 10.7
0.0 1.1 0.0 0.4 1.3 0.2 1.0 0.0 0.6 2.1
0.0 5.3 0.0 0.1 12.0 11.9 133.9 8.3 4.2 0.5 0.7 4.2

tribution of these species on the coral reefs. Because seagrass complexity may influence fish
Distribution of the species on the reef locations densities (Sogard et al., 1987; Hyndes et al., 2003),
were sampled in quadrats of 88 m that were seagrass height and seagrass shoot density were
surveyed using the similar under water visual measured at each site where quadrats were sur-
census technique that was used to survey the veyed. Mean seagrass height of a site was quantified
distribution of fish species on the seagrass beds. by measuring the length of both shoots and leaves of
Quadrats were placed randomly on the reef and randomly selected seagrass plants on seagrass beds
included all depth zonations that could be dis- where quadrats were surveyed. The number of
tinguished (until a maximum depth of 20 m). The plants that was measured per site ranged between 60
number of surveyed quadrats per coral reef are and 110 for reef seagrasses and between 60 and 80
listed in Table 3. for bay seagrasses. The mean shoot density of a site

Table 3. Total number (#) of observed species on reef seagrasses, bay seagrasses and the 19 coral reef sites in a spatial gradient around Zanzibar (all surveyed quadrats pooled),
and the # of surveyed quadrats. Total # of species is based on the 39 most common species (total mean fish density on seagrass beds >0.05 individuals 100 m)2, all surveyed
quadrats pooled)

Reef Bay Western reefs

seagrasses seagrasses
Adriana Pwaku Boribu Tambare Chumbe Morogo Changu

# Of observed species 35 39 28 22 30 26 29 26 34
# Of surveyed quadrats 99 121 49 77 72 62 34 110 117

North western reefs Northern reefs Eastern reefs Southern reef

Timbatu island Hanga Timbatu Sensation Nungwi Leon’s Mnemba Kiwengwa Pongwe Uroa Jambiani Kizimkazi
bay wall (atoll reef)

# Of observed species 31 22 30 19 29 27 29 29 20 35 28 27
# Of surveyed quadrats 52 15 86 36 71 40 78 45 24 105 42 30

was quantified by counting the number of shoots in sulted in violation of assumptions for performing a
25 by 25 cm quadrats, which were placed randomly nested ANOVA. Therefore, mean density (pooling
on the seagrass beds where quadrats were surveyed. juveniles and adults per quadrat) of each of the
The number of measured quadrats per site ranged most common 39 species (density >0.05 individ-
between 16 and 32 for reef seagrasses and between uals 100 m)2) were compared using independent
28 and 32 for bay seagrasses. sample t-tests assuming either equal variances or
unequal variances (as determined with a Levene’s
Data analysis test). For the same reason independent sample t-
tests were used to compare mean seagrass height
The abundance of each observed species per reef and seagrass density between the two seagrass
location was expressed as mean total density per habitats.
100 m2 (all size classes pooled per quadrat). To study the relationship between seagrass
Euclidean cluster analysis (using group average height or seagrass density and densities of indi-
linkage) and Shannon–Wiener diversity indices vidual species, simple linear regressions were per-
were used to compare the distribution of species formed. Linear regressions were also performed
along the various reef locations. Cluster and between total fish density or total species richness
diversity analysis of coral reef sites was performed and seagrass height or seagrass density. Fish den-
for 38 of the 39 species that occurred in mean sity or species richness (either for the whole fish
densities >0.05 individuals 100 m)2 on the seagrass assemblage or for species separately) per seagrass
beds (pooling all seagrass quadrats). One of the site was used as the dependent variable and mean
species (Scarus spp.) was excluded for these two seagrass height or density as the regressor.
analyses because it was not observed on the coral To study the spatial distribution of the 39
reefs. Cluster analyses and Shannon–Wiener species that occurred in mean densities >0.05
diversity calculations were performed using individuals 100 m)2 on the seagrass habitats
the computer programme Biodiversity Pro 2 (pooling all seagrass quadrats) at the various sites,
(McAleece, 1997). Principal Component Analysis (PCA) was used on
Fish were recorded as juveniles when they were log-transformed fish densities. PCA was carried
smaller than one third of the maximum species’ out with the Canoco 4.0 program (ter Braak &
length. Nagelkerken & van der Velde (2002) used Smilauer, 1998). Scaling was focussed on inter-
this rule to distinguish juveniles from adults and species distances, and species scores were divided
confirmed that for species for which maturation by the standard deviation. Nested ANOVA’s,
size was known this rule was appropriate to dis- t-tests and linear regressions were performed using
tinguish juveniles from adults. For species with a SPSS 12.
maximum length >90 cm, specimens were recorded
as juveniles when <30 cm. Maximum lengths of
species were obtained from FishBase World Wide Results
Web (Froese & Pauly, 2003).
To compare mean fish densities and mean Of the selected fish families, 49 species were
species richness of juveniles, adults and the total observed in the quadrats on the seagrass beds, 39 of
fish assemblage (i.e. pooling juveniles and adults which occurred in mean densities >0.05 individuals
per quadrat) between the two types of seagrass 100 m)2 (Tables 1 and 2). The distribution of 38 of
habitats, nested one-way ANOVA were used these 39 species (Scarus spp. was not observed on
based on log-transformed values. Seagrass habitat reefs) on the various coral reefs, did not show a
type (i.e. bay seagrasses versus reef seagrasses) was clear spatial pattern (Fig. 2, Tables 1 and 2). Only
used as a fixed factor, seagrass site (i.e. four reef the distribution of fishes at Mnemba reef (an off-
seagrass sites, three bay seagrass sites) was used as shore atoll reef surrounded by water with a depth
a random factor and was nested within the factor >100 m) was clearly different from that of the other
seagrass habitat type. 18 coral reefs. Among these 18 coral reefs, there
At species level, many quadrats consisted of was no difference in the distribution of fish species
zero counts (i.e. absence of fish species) which re- in eastern versus western reefs. Two species were

observed only at one reef location (Lutjanus eh- grasses and reef seagrasses. However, reef sea-
renbergi and Gerres oyena), one species at only two grasses showed both a significantly higher total
reef locations (Calotomus spinidens) and one spe- adult density and species richness of adults than
cies at only three reef locations (Leptoscarus vai- bay seagrasses (fish density p=0.022, species
giensis). All other species occurred on a wide richness p=0.007, nested ANOVA, Fig. 3a, b).
variety of coral reefs (Tables 1 and 2). Seagrass sites nested within seagrass habitat
Total fish density and species richness were types showed a significant effect for both juvenile
not significantly different between the two sea- and adult fish density and species richness (both
grass habitat types (Fig. 3a, b), although nesting juvenile fish density and species richness
showed a significant effect for total fish density p<0.001, adult fish density p=0.001, adult spe-
(p<0.001) and species richness (p=0.005, nested cies richness p=0.007).
ANOVA). There were no significant differences Mean seagrass height (±SD) was significantly
observed either in total juvenile density or greater (p<0.001; t-test) for bay seagrasses
species richness of juveniles between bay sea- (55.4±22.7 cm) than for reef seagrasses

Figure 2. Euclidean cluster analysis and Shannon–Wiener diversity of 38 fish species on the 19 coral reefs (see also Fig. 1 and Tables 1
and 2).

250 between fish density and seagrass density were

(a) Mean fish densities only significant for Chaetodon auriga (p=0.040;
Mean fish density + SE (100 m-2)

Reef seagrass sites I b=)0.01; Chlorurus sordidus

Bay seagrass sites (p=0.047; b=1.10; R2=0.58) and Parupeneus
barberinus (p=0.028; b=)0.18; R2=0.65). Linear
A regressions between all other species, total fish
density or species richness and seagrass height or
density were not significant.
100 PCA of the 39 most common species showed
1 that of all reef seagrass sites, most species clustered
on the Chapwani site, whereas of all bay seagrass
50 sites most species clustered on the Chwaka site
(Fig. 4). A species-level analysis of these 39 most
common species showed that 17 species showed
significantly different densities between the two
Juveniles Adults Total
seagrass habitats, while Cheilinus undulatus,
40 Lutjanus argentimaculatus, L. ehrenbergi and
(b) Mean species richness
Mean species richness + SE (100 m-2)

Parupeneus indicus occurred in bay seagrasses

35 Reef seagrass sites
I only (Table 1, Fig. 4). Nine of the 17 species had
A Bay seagrass sites
30 their highest densities in bay seagrasses, whereas
eight species had their highest densities in reef
25 seagrasses. Most species that were associated with
A bay seagrasses (i.e., species with significantly
higher densities in bay seagrasses than in reef
seagrasses, or species only observed in bay seag-
rasses; Table 1) were predominantly observed at
10 2 the Chwaka and Tunduni sites. The Marumbi site
that of all bay seagrass sites was located closest to
5 the reef (6.5 km), was characterised by the highest
density of only one bay seagrass associated species
Juveniles Adults Total (Lethrinus harak, Fig. 4). Four species that were
associated with reef seagrasses (i.e., species with
Figure 3. Mean fish density (a) and mean species richness (b) of
the 49 observed species on reef seagrasses and bay seagrasses significantly higher densities in reef seagrasses than
(bay), of juveniles, adults and the total fish assemblage (pooling in bay seagrasses; Table 2) were observed in
juveniles and adults per quadrat). Different letters and numbers highest densities at the Chapwani site (Lethrinus
indicate significant differences in fish density or species richness variegatus, Lutjanus gibbus, Scarus frenatus and
between the two seagrass habitats (nested ANOVA).
Scarus spp.), whereas the Changu and Muyuni
sites were each characterised by the highest den-
(33.2±10.1 cm), but mean seagrass density was sities of two of the species associated with reef
significantly higher (p<0.001; t-test) for reef seagrasses (Changu: Calotomus carolinus and
seagrasses (896.6±316.0 plants m)2) than for bay Chlorurus sordidus; Muyuni: Cheilio inermis and
seagrasses (510.1±202.9 plants m)2). Linear Lethrinus mahsena). The remaining 18 species did
regressions at species level between fish density and not show a clear preference for either of the two
seagrass height for the species that were observed habitat types.
both in reef seagrasses and bay seagrasses, showed
that significant relationships existed only for
Chaetodon auriga (p=0.018; b=0.02; R2=0.71), Discussion
Lutjanus fulviflamma (p=0.030; b=0.70;
R2=0.64) and Scarus ghobban (p=0.005; Total fish density and total species richness of the
b=0.33; R2=0.82), whereas linear regressions pooled fish assemblage did not differ between the

Chapwani 1.0
Let len
Sca fre Sco gha
Ger oye
Sca spp Sig ste

Cal spi Lut gib Par bar

Let var Lut ful

Let har Sca psi
Sig sut Sca gho
Ple gat
Lut ehr/
Par rub
Lut arg
Cal car
Changu Par ind
Marumbi Hip har
Cha mel
-1.0 Sca rus 1.0
Chl sor Cha aur
Lep vai Chumbe Mul fla
Let mah Sca sca
Che ine Che und
Sca nig
Ple fla Chl str
Che fas
Par mac
Par ple

Muyuni Che tri Tunduni

Species: Sites:
Significant higher density on bay seagrasses Bay seagrass sites
Significant higher density on reef seagrasses Reef seagrass sites
Only observed on bay seagrasses
No significant difference in density -1.0
Figure. 4 Principal Component Analysis (PCA) of the 39 most common fish species (mean density >0.05 individuals 100 m)2)
observed on quadrats in seagrass beds at seven sites. Reef seagrass sites were situated adjacent to coral reefs, whereas bay seagrass sites
were situated in a marine bay close to mud flats with mangroves but far from coral reefs. The eigenvalues of the first and second
ordination axis were 0.38 and 0.26 respectively, while the cumulative percentage variance of species data for both axes was 64%.
Species abbreviations: Cal spi: Calotomus spinidens; Cal car: C. carolinus; Cha aur: Chaetodon auriga; Cha mel: C. melanotus; Che fas:
Cheilinus fasciatus; Che tri: C. trilobatus; Che und: C. undulatus; Che ine: Cheilio inermis; Ger oye: Gerres oyena; Chl sor: Chlorurus
sordidus; Chl str: C. strongylocephalus;Hip har: Hipposcarus harid; Lep vai: Leptoscarus vaigiensis; Let har: Lethrinus harak; Let len: L.
lentjan; Let mah: L. mahsena; Let var: L. variegatus; Lut arg: Lutjanus argentimaculatus; Lut ehr: L. ehrenbergi; Lut ful: L. fulviflamma;
Lut gib: L. gibbus; Mul fla: Mulloidichthys flavolineatus; Par bar: Parupeneus barberinus; Par ind: P. indicus; Par mac: P. macronema;
Par ple: P. pleurostigma; Par rub: P. rubescens; Ple fla: Plectorhinchus flavomaculatus; Ple gat: P. gaterinus; Sca fre: Scarus frenatus; Sca
gho: S. ghobban; Sca nig: S. niger; Sca psi: S. psittacus; Sca rus: S. russelli; Sca sca: S. scaber; Sca spp.: S. spp.; Sco gha: Scolopsis
ghanam; Sig ste: Siganus stellatus; Sig sut: S. sutor.

two seagrass habitats, suggesting that the standing species richness of juvenile fish were equal in both
stocks of reef fish in the two habitats were similar seagrass habitats. This suggests that both habitats
and no large differences exist between the carrying function as important juvenile fish habitats and
capacities of the two habitats. Comparable to total that no large differences exist between the function
fish density and species richness, total densities and of the two seagrass habitats as a juvenile fish

habitat. The hypothesis that seagrass beds in the better explain differences in the two fish assem-
marine embayment situated far from coral reefs blages.
would show higher juvenile fish densities than The differences in species composition and
seagrass beds adjacent to coral reefs due to a species densities between the two seagrass habitats
higher predation pressure in the latter (based on may be partly explained by different recruitment
Shulman, 1985; Sweatman & Robertson, 1994) can and settlement patterns of juvenile reef fish species.
therefore not be confirmed by the present study. A The location of seagrass beds (e.g. in a sheltered
possible explanation for this observation could be marine embayment or on the continental shelf
a similar predation pressure on juvenile fish in exposed to the open sea) in combination with the
both seagrass habitats, possibly because piscivores type of adjacent habitat (e.g. mangroves and mud
are also present on seagrass beds in marine bays flats or coral reefs) may result in differences in
(Sheaves, 2001; Nakamura & Sano, 2004a; Baker environmental or biological variables that cha-
& Sheaves, 2005). racterise these seagrass beds, such as salinity,
Contrary to juvenile fish density and species water temperature, water depth, food availability,
richness, both adult fish density and species rich- shelter opportunities or predation risk (reviewed in
ness was significantly higher on the reef seagrasses Blaber, 2000). Since several studies have shown
than on the bay seagrasses. These differences may that coral reef fish larvae actively select their set-
possibly be explained by a combination of two tlement habitat (Atema et al., 2002; Fisher &
processes. Firstly, ontogenetic migrations of adults Bellwood, 2003; Simpson et al., 2004; Lecchini
towards the nearest coral reefs may result in et al., 2005), partly influenced by some of the
decreased densities on bay seagrasses (Parrish, variables listed above, differences in some of these
1989; Beck et al., 2001). Since the coral reef is variables possibly result in different recruitment
located far away, these adults most likely use the and settlement patterns of juvenile reef fish species.
coral reef as a permanent habitat. On the contrary In the Caribbean, settlement of some juvenile coral
to bay seagrasses, reef seagrasses are located reef fish has been suggested to be influenced by the
directly adjacent to the coral reef. Here, adult fish habitat types present, e.g., coral reefs, mangroves
may use both the seagrass beds and the coral reef or seagrass beds (Yáñez-Arancibia et al., 1993;
as an adult habitat and adult fish densities on the Risk, 1997; Nagelkerken et al., 2001; Cocheret
reef seagrasses are not reduced as a result of de la Morinière et al., 2002). In the present study,
ontogenetic emigration. Secondly, additional the species that showed highest densities in bay
migrations of adult fish of species that are reef- seagrasses were barely observed on the bay seag-
associated from the coral reefs towards directly rasses at the Marumbi site. In comparison with the
adjacent seagrass beds may result in increased two other bay seagrass sites, Marumbi is situated
densities on reef seagrasses (Weinstein & Heck, farthest from mangroves and mud flats, and rela-
1979; Kochzius, 1997). tively close to the entrance of the embayment (and
Although total fish densities were comparable, the adjacent coral reef). This provides further
fish species composition and densities of individual support for the earlier suggestion that some fish
species were clearly different between the two species actively select seagrass beds that are char-
habitat types. Some fish species clearly occur in acterised by variables influenced by the presence of
seagrass beds adjacent to the coral reef (reef mangroves and mud flats, and/or the setting in a
seagrasses), whereas other species occur in seagrass marine embayment.
beds adjacent to a combination of mud flats and For some of the observed species their associ-
mangroves situated in a sheltered embayment (bay ation with either bay seagrasses or reef seagrasses
seagrasses). Seagrass complexity appeared to be of may be related to their ecology. Dorenbosch et al.
minor importance in explaining these differences, (2005) showed a clear effect of the presence of bays
since relationships between seagrass complexity with seagrass beds and mangroves on fish densities
and fish densities were minimal for the majority of on adjacent coral reefs for Cheilinus fasciatus,
species, while also seagrass species composition C. undulatus, Chlorurus strongylocephalus, Lutj-
was similar between the two habitats. Therefore, anus argentimaculatus, L. fulviflamma, Parupen-
another hypothesis can be suggested that may eus barberinus, P. rubescens and Scarus ghobban

suggesting these species are associated with bay the other two bay seagrass sites (located more dee-
habitats in their juvenile stage. Likewise, these ply into the bay). Although seasonal differences may
species were associated with seagrass beds located also have an effect on the structure of the fish
in the embayment with mangroves and not with assemblage on seagrass beds (Laroche et al., 1997),
seagrass beds located on the continental shelf these influences are generally highest during the cold
adjacent to coral reefs in the present study. Fur- season or during periods with heavy rains. The data
thermore, juveniles of Lutjanus argentimaculatus in the present study were collected outside the cold
and L. fulviflamma have been reported from or rainy seasons, and data were collected simulta-
mangroves (Wakwabi & Mees, 1999; Kiso & neously during the same time period. Seasonally
Mahyam, 2003) and may prefer mangroves and/or influences were therefore most likely to be low in this
areas close to mangroves as a juvenile habitat case, and if present, comparable for all sites.
which may possibly explain the presence of both In conclusion, the present study shows that two
species on bay seagrasses close to mangroves and types of seagrass beds that differ in their setting in
no observations of these species on reef seagrasses the surrounding landscape (i.e., in a sheltered
far away from mangroves. Lugendo et al. (2005) embayment with mangroves and mudflats away
showed the presence of juveniles of both species in from the coral reef or on the open continental shelf
the mangroves in the embayment in the present directly adjacent to the coral reef), have a similar
study, located at short distance from the bay density of juvenile fish suggesting that both types of
seagrasses. Chlorurus sordidus and Scarus frena- seagrass beds have an important function as a
tus are strongly associated with coral reefs and are juvenile fish habitat. Despite these similar densities
known to use the coral reef as a juvenile habitat of juvenile fish, the structure of the fish assemblage
(Crook, 1999; Gust et al., 2002). Subsequently, on seagrass beds in an embayment with mangroves
their strong association with seagrass beds located and mud flats is different from that of seagrass beds
directly adjacent to coral reefs in the present study located next to coral reefs. The hypothesis that the
may be a reflection of this general preference for setting of the seagrass beds with respect to the sur-
coral reefs. rounding landscape affect fish migration and
An alternative explanation for the different fish recruitment patterns, may partly explain these lat-
assemblages in the two seagrass habitats could be ter observations. Regarding the conservation of
the influence of a spatial or temporal effect such as coral reef fish stocks in relation to the design of
different ocean current patterns between the east- marine protected areas, this study suggests that it is
ern side of Zanzibar island (bay seagrasses) and important to cover various types of seagrass beds
the western and northern side of the island (reef that stretch from the coral reef edge until mangrove-
seagrasses). In this case the occurrence of species dominated embayments farther from the reef.
on the two seagrass habitat types might rather
depend on a different fish recruitment associated
with these ocean current patterns than on the Acknowledgements
arrangement of the habitat types. Several argu-
ments can be proposed, however, to discard this This study was funded by Nuffic through the
alternative explanation. Firstly, the data showed ENVIRONS-MHO programme, by PADI’s
that 35 of the 39 observed fish species occurred on Aware project and by the Quo Vadis fund of
both seagrass habitat types (either in significant Radboud University Nijmegen (The Netherlands).
different or equal densities), suggesting that almost We would like to thank the management and staff
all of the species were able to reach seagrass beds of the Institute of Marine Sciences in Zanzibar for
on various sides of Zanzibar island. Secondly, 34 allowing us to use their facilities and for their
of the 39 species, also occurred on coral reef support. We also wish to thank Simgeni Juma
locations throughout Zanzibar island, without Simgeni and Abdalla Makame Mweleza for their
showing a separation in density between the east- help in the field. W. Verberk provided useful
ern and western reefs. Finally, as noted earlier, the comments to earlier versions of the manuscript.
fish assemblage at Marumbi (located near This is publication No. 360 of the Centre for
the mouth of the bay) was different from that of Wetland Ecology.

References Huxham, M., E. N. Kimani & J. Augley, 2004. Mangrove fish:

a comparison of assemblage structure between forested and
Atema, J., M. J. Kingsford & G. Gerlach, 2002. Larval reef fish cleared habitats. Estuarine, Coastal and Shelf Science 60:
could use odour for detection, retention and orientation to 637–647.
reefs. Marine Ecology Progress Series 241: 151–160. Hyndes, G. A., A. J. Kendrick, L. D. MacArthur & E. Stewart,
Baelde, P., 1990. Differences in the structure of fish assemblages 2003. Differences in the species- and size-composition of fish
in Thalassia testudinum beds in Guadeloupe, French West assemblages in three distinct seagrass habitats with differ-
Indies, and their ecological significance. Marine Biology 105: ing plant and meadow structure. Marine Biology 142:
163–173. 1195–1206.
Baker, R. & M. Sheaves, 2005. Redefining the piscivore Jenkins, G. P., K. P. Black, M. J. Wheatley & D. N. Hatton,
assemblage of shallow estuarine nursery habitats. Marine 1997. Temporal and spatial variability in recruitment of a
Ecology Progress Series 291: 197–213. temperate, seagrass-associated fish is largely determined by
Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. physical processes in the pre- and post-settlement phases.
Eggleston, B. M. Gillanders, B. Halpern, C. G. Hays, Marine Ecology Progress Series 148: 23–35.
K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan & M. Jiddawi, N. S. & R. D. Stanley, 1997. Fisheries Stock Assess-
P. Weinstein, 2001. The identification, conservation, and ment in the Traditional Fishery Sector: The Information
management of estuarine and marine nurseries for fish and Needs. Canadian International Development Agency, Zan-
invertebrates. BioScience 51: 633–641. zibar, Tanzania.
Blaber, S. J. M., 2000. Tropical Estuarine Fishes: Ecol- Kanashiro, K., 1998. Settlement and migration of early stage
ogy, Exploitation and Conservation. Blackwell Science, spangled emperor, Lethrinus nebulosus (Pisces: Lethrinidae),
Oxford. in the coastal waters off Okinawa Island, Japan. Nippon
Burke, N. C., 1995. Nocturnal foraging habitats of French and Suisan Gakkaishi 64: 618–625.
bluestriped grunts, Haemulon flavolineatum and H. sciurus, Kimani, E. N., G. K. Mwatha, E. O. Wakwabi, J. M. Ntiba &
at Tobacco Caye, Belize. Environmental Biology of Fishes B. K. Okoth, 1996. Fishes of a shallow tropical mangrove
42: 365–374. estuary, Gazi, Kenya. Marine and Freshwater Research 47:
Cocheret de la Morinière, E., B. J. A. Pollux, I. Nagelkerken & 857–868.
G. van der Velde, 2002. Post-settlement life cycle migration Kingsford, M. J., J. M. Leis, A. Shanks, K. C. Lindeman,
patterns and habitat preferences of coral reef fish that use S. G. Morgan & J. Pineda, 2002. Sensory environments,
seagrass and mangrove habitats as nurseries. Estuarine, larval abilities and local self-recruitment. Bulletin of Marine
Coastal and Shelf Science 55: 309–321. Science 70: 309–340.
Crook, A. C., 1999. Quantitative evidence for assortative Kiso, K. & M. I. Mahyam, 2003. Distribution and feeding
schooling in a coral reef fish. Marine Ecology Progress Series habits of juvenile and young John’s snapper Lutjanus johnii
176: 17–23. in the Matang mangrove estuary, west coast of Peninsular
Dorenbosch, M., M. G. G. Grol, M. J. A. Christianen, Malaysia. Fisheries Science 69: 563–568.
I. Nagelkerken & G. van der Velde, 2005. Indo-Pacific Kochzius, M., 1997. Interrelation of ichthyofauna from a sea-
seagrass beds and mangroves contribute to fish density and grass meadow and coral reef in the Philippines. Proceedings
diversity on adjacent coral reefs. Marine Ecology Progress of the 5th Indo-Pacific Fisheries Conference, Nouméa 1997,
Series 302: 63–67. 517–535.
Dorenbosch, M., M. C. van Riel, I. Nagelkerken & G. van der Laroche, J., E. Baran & N. B. Rasoanandrasana, 1997. Tem-
Velde, 2004b. The relationship of reef fish densities to the poral patterns in a fish assemblage of a semiarid mangrove
proximity of mangrove and seagrass nurseries. Estuarine, zone in Madagascar. Journal of Fish Biology 51: 3–20.
Coastal and Shelf Science 60: 37–48. Lecchini, D., J. Shima, B. Banaigs & R. Galzin, 2005.
Dorenbosch, M., M. C. Verweij, I. Nagelkerken, G. van Larval sensory abilities and mechanisms of habitat
der Velde & N. S. Jiddawi, 2004a. Homing and daytime selection of a coral reef fish during settlement. Oecologia
tidal movements of juvenile snappers (Lutjanidae) be- 143: 326–334.
tween shallow-water nursery habitats in Zanzibar, wes- Lugendo, B. R., A. Pronker, I. Cornelissen, A. de Groene,
tern Indian Ocean. Environmental Biology of Fishes 70: I. Nagelkerken, M. Dorenbosch, G. van der Velde &
203–209. Y. Mgaya, 2005. Habitat utilisation by juveniles of com-
Fisher, R. & D. R. Bellwood, 2003. Undisturbed swimming mercially important fish species in a marine embayment in
behaviour and nocturnal activity of coral reef fish larvae. Zanzibar, Tanzania. Aquatic Living Resources 18: 149–158.
Marine Ecology Progress Series 263: 177–188. McFarland, W. N., E. B. Brothers, J. C. Ogden, M. J. Shulman,
Froese, R. & D. Pauly, 2003. FishBase. World Wide Web E. L. Birmingham & N. M. Kotchian-Prentiss, 1985.
electronic publication ( 24-6-03. Recruitment patterns in young French grunts, Haemulon
Gust, N., J. H. Choat & J. L. Ackerman, 2002. Demographic flavolineatum (Family Haemulidae), at St Croix, Virgin
plasticity in tropical reef fishes. Marine Biology 140: Islands. Fisheries Bulletin 83: 413–426.
1039–1051. McAleece, N., 1997. Biodiversity Pro 2. World Wide Web elec-
Halpern, B. S., 2004. Are mangroves a limiting resource for tronic publication by the National History Museum London
two coral reef fishes? Marine Ecology Progress Series 272: & The Scottisch Association of Marine Science (http://
93–98. zip).

Mumby, P. J., A. J. Edwards, J. E. Ernesto Arias-Gonzalez, Pollard, D. A., 1984. A review of ecological studies on seagrass-
P. G. Lindeman, A. Gall, M. I. Gorczynska, A. R. Harborne, fish communities, with particular reference to recent studies
C. L. Pescod, H. Renken, C. C. C. Wabnitz & G. Llewellyn, in Australia. Aquatic Botany 18: 3–42.
2004. Mangroves enhance the biomass of coral reef fish Polunin, N. V. C. & C. M. Roberts, 2004. Greater biomass and
communities in the Caribbean. Nature 427: 533–536. value of target coral-reef fishes in two small Caribbean marine
Nagelkerken, I., M. Dorenbosch, W. C. E. P. Verberk, E. reserves. Marine Ecology Progress Series 100: 167–176.
Cocheret de la Morinière & G. van der Velde, 2000a. Quin, T. P. & B. L. Kojis, 1987. The influence of diel cycle, tidal
Importance of shallow-water biotopes of a Caribbean bay direction and trawl alignment on beam trawl catches in an
for juvenile coral reef fishes: patterns in biotope association, equatorial estuary. Environmental Biology of Fishes 19:
assemblage structure and spatial distribution. Marine Ecol- 297–308.
ogy Progress Series 202: 175–193. Risk, A., 1997. Effects of habitat on the settlement and post
Nagelkerken, I., M. Dorenbosch, W. C. E. P. Verberk, E. settlement success of the ocean surgeonfish Acanthurus
Cocheret de la Morinière & G. van der Velde, 2000b. Day– bahianus. Marine Ecology Progress Series 161: 51–59.
night shifts of fishes between shallow-water biotopes of a Sheaves, M., 2001. Are there really few piscivorous fishes in
Caribbean bay, with emphasis on the nocturnal feeding of shallow estuarine habitats?. Marine Ecology Progress Series
Haemulidae and Lutjanidae. Marine Ecology Progress Series 222: 279–290.
194: 55–64. Shulman, M. J., 1985. Recruitment of coral reef fishes: effects of
Nagelkerken, I., S. Kleijnen, T. Klop, R. A. C. J. van den Brand, distribution of predators and shelter. Ecology 66: 1056–1066.
E. Cocheret de la Morinière & G. van der Velde, 2001. Simpson, S. D., M. G. Meekan, R. D. McCauley & A. Jeffs,
Dependence of Caribbean reef fishes on mangroves and 2004. Attraction of settlement-stage coral reef fishes to reef
seagrass beds as nursery biotopes: a comparison of fish fau- noise. Marine Ecology Progress Series 276: 263–268.
nas between bays with and without mangroves/seagrass beds. Sogard, S. M., G. V. N. Powell & J. G. Holmquist, 1987.
Marine Ecology Progress Series 214: 225–235. Epibenthic fish communities on Florida Bay banks: relations
Nagelkerken, I., C. M. Roberts, G. van der Velde, M. Dore- with physical parameters and seagrass cover. Marine Ecol-
nbosch, M. C. van Riel, E. Cocheret de la Morinière & P. H. ogy Progress Series 40: 25–39.
Nienhuis, 2002. How important are mangroves and seagrass Sweatman, H., 1988. Field evidence that settling coral reef fish
beds for coral-reef fish? The nursery hypothesis tested on an larvae detect resident fishes using dissolved chemical cues.
island scale. Marine Ecology Progress Series 244: 299–305. Journal of Experimental Marine Biology and Ecology 124:
Nagelkerken, I. & G. van der Velde, 2002. Do non-estuarine 163–174.
mangroves harbour higher densities of juvenile fish than Sweatman, H. & D. R. Robertson, 1994. Grazing halos and
adjacent shallow-water and coral reef habitats in Curaçao predation on juvenile Caribbean surgeonfishes. Marine
(Netherlands Antilles)? Marine Ecology Progress Series 245: Ecology Progress Series 111: 1–6.
191–204. ter Braak, C. J. F. & P. Smilauer, 1998. Reference Manual and
Nakamura, Y. & M. Sano, 2004b. Comparison between User’s Guide to Canoco for Windows: Software for
community structures of fishes in Enhalus acoroides- and Canonical Community Ordination (version 4). Microcom-
Thalassia hemprichii-dominated seagrass beds on fringing puter power, Ithaca, New York.
coral reefs in the Ryuku Islands, Japan. Ichthyological Re- Thollot, P. & M. Kulbicki, 1988. Overlap between the fishfauna
search 51: 38–45. inventories of coral reefs, soft bottoms and mangroves in
Nakamura, Y. & M. Sano, 2004a. Is there really lower preda- Saint–Vincent Bay (New Caledonia). Proceedings of the 6th
tion risk for juvenile fishes in a seagrass bed compared with International Coral Reef Symposium Australia 2: 613–618.
an adjacent coral area?. Bulletin of Marine Science 74: Wakwabi, E. O. & J. Mees, 1999. The epibenthos of the back-
477–482. waters of a tropical mangrove creek (Tudor creek, Mombasa,
Nanami, A., M. Nishihira, T. Suzuki & H. Yokochi, 2005. Kenya). Netherlands Journal of Zoology 49: 189–206.
Species-specific habitat distribution of coral reef fish assem- Watson, R. A. & J. T. Quinn, 1997. Performance of transect
blages in relation to habitat characteristics in an Okinawan and point count underwater visual census methods. Eco-
coral reef. Environmental Biology of Fishes 72: 55–65. logical Modelling 104: 103–112.
Ogden, J. C. & N. S. Buckman, 1973. Movements, foraging Weinstein, M. P. & K. L. Heck, 1979. Ichthyofauna of seagrass
groups, and diurnal migrations of the striped parrotfish meadows along the Caribbean coast of Panama and the Gulf
Scarus croicensis Bloch (Scaridae). Ecology 54: 589–596. of Mexico: composition, structure and community ecology.
Ogden, J. C. & P. R. Ehrlich, 1977. The behaviour of hetero- Marine Biology 50: 97–107.
typic resting schools of juvenile grunts (Pomadasyidae). Yáñez-Arancibia, A., A. L. Lara-Dominguez & J. W. Day, 1993.
Marine Biology 42: 273–280. Interactions between mangrove and seagrass habitats medi-
Parrish, J. D., 1989. Fish communities of interacting shallow- ated by estuarine nekton assemblages – coupling of primary
water habitats in tropical oceanic regions. Marine Ecology and secondary production. Hydrobiologia 264: 1–12.
Progress Series 58: 143–160.