Marine Environmental Research 108 (2015) 45e54

Contents lists available at ScienceDirect

Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev

The influence of fire-coral colony size and agonistic behaviour of territorial damselfish on associated coral reef fish communities Isabela Carolina Silva Leal a, Maria Elisabeth de Araújo a, Simone Rabelo da Cunha a, Pedro Henrique Cipresso Pereira b, * a b


ria, Recife, PE 50670-901, Brazil Departamento de Oceanografia, CTG, Universidade Federal de Pernambuco (UFPE), Av. Arquitetura, s/n, Cidade Universita School of Marine and Tropical Biology, James Cook University e JCU, Townsville, QLD 4811, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 December 2014 Received in revised form 11 April 2015 Accepted 26 April 2015 Available online 27 April 2015

Branching hydrocorals from the genus Millepora play an important ecological role in South Atlantic reefs, where branching scleractinian corals are absent. Previous studies have shown a high proportion of reef fish species using branching fire-coral colonies as shelter, breeding, and feeding sites. However, the effects of Millepora spp. colony size and how the agonistic behaviour of a competitive damselfish affect the associated reef fish community are still unknown. The present study examined how fire-coral colony volume and the presence of a highly territorial and aggressive damselfish (Brazilian endemic Stegastes fuscus) affects the reef fish community associated with the fire-coral Millepora alcicornis. M. alcicornis
Reefs off Northeast Brazil. Our colonies were surveyed from September 2012 to April 2013 at Tamandare results show that the abundance and richness of coral associated fish was positively correlated with M. alcicornis coral colony volume. Additionally, behaviour of S. fuscus, the most abundant reef fish species found associated with fire-coral colonies (almost 57% of the fish community), was also influenced by firecoral colony volume. There was a clear trend of increased agonistic behaviour and feeding on coral polyps as colony volume increased. This trend was reversed for the non-occupational swimming category, which decreased as M. alcicornis colony volume increased. Behavioural ontogenetic changes were also detected for S. fuscus individuals. Juveniles mainly showed two distinct behaviours: sheltered on coral branches and feeding on coral polyps. In contrast, adults presented greater equitability among the behavioural categories, mostly non-occupational swimming around coral colonies and agonistic behaviour. Lastly, S. fuscus individuals actively defended fire-coral colonies from intruders. A large number of agonistic interactions occurred against potential food competitors, which were mainly roving herbivores, omnivores, and sessile invertebrate feeders. To our knowledge, the present study provides the first evidence that through habitat competition, the presence of S. fuscus may affect reef fish communities associated with M. alcicornis coral colonies. Our findings also indicate that S. fuscus uses M. alcicornis coral colonies as part of their territory for shelter and foraging. In conclusion, M. alcicornis fire-coral colonies are extremely important habitats for reef fishes and the size and presence of a territorial damselfish are relevant variables for associated reef fish community. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Habitat complexity Coral-fish association Competition Millepora alcicornis Stegastes fuscus South Atlantic reefs

1. Introduction Habitat structural complexity is critically relevant in several marine ecosystems such as macroalgal beds, seagrass beds, mangrove forests, kelp forests, and coral reefs (Syms and Jones, 2000; MacNeil et al., 2009; Pereira et al., 2010; Chaves et al.,

* Corresponding author. E-mail address: [email protected] (P.H.C. Pereira). http://dx.doi.org/10.1016/j.marenvres.2015.04.009 0141-1136/© 2015 Elsevier Ltd. All rights reserved.

2013). In tropical coral reefs, habitat structural complexity plays a key role in the life cycles of a diverse range of organisms, including crustaceans (Brown and Edmunds, 2013) and fish (Johnson et al., 2011; Graham and Nash, 2012). Architecturally complex branching cnidarians (e.g. Acropora spp. and Millepora spp.) have been shown to host a great diversity of organisms (e.g. invertebrates and fishes), providing shelter, food, and reproduction sites (Brooks et al., 2007; Coni et al., 2012; Pereira et al., 2012; Leal et al., 2013). Tropical reef habitat complexity has been shown to be related to fish abundance, diversity, predation rates, and competitive

46

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

interactions (Garpe et al., 2006; Grober-Dunsmore et al., 2007; Cabaitan et al., 2008). Positive relationships between coral complexity and fish abundance suggest that coral complexity is a key factor structuring fish communities (Hixon and Beets, 1993; Graham and Nash, 2012). Coral colony size and volume are also important in determining habitat quality, which can also affect fish abundance and species richness (Lomolino, 2001; Schiemer et al., 2009). Larger colonies are often a preferred resource for reef fish possibly due to increased space between their branches, allowing for the presence of higher numbers of species and providing superior shelter areas (Hobbs and Munday, 2004; Eggleston et al., 1997). Furthermore, higher abundances and availability of branching corals has been shown to be a strong predictor of fish growth, survival, and reproductive success (Jones, 1988; Holbrook et al., 2000; Brooker et al., 2013). In this context, Coni et al. (2012) first suggested that the size of fire-coral colonies from the genus Millepora could affect fish abundance and species richness in the Abrolhos Bank coral reefs (NorthEast Brazil); however this hypothesis has never been fully tested. Reef fish distribution can also be influenced by competition (Sale, 1977; Almany, 2003; Munday, 2001; Geange et al., 2013) and €v and Diehl, 1994; Almany, 2004). predatoreprey interactions (Eklo Stegastes fuscus Cuvier, 1830, a Brazilian endemic damselfish, defends relatively small territories from its competitors, especially
rio et al., 2006; herbivores that invade their territories (Oso Medeiros et al., 2010; Leal et al., 2013). According to Shulman (1984) the aggressive behaviour of Stegastes spp. in the Caribbean, can reduce recruitment of other fish families, such as Acanthuridae and Chaetodontidae. The genus Stegastes is often found in high abundance within Acropora spp. colonies, defending filamentous algal turfs within their territories (Mahoney, 1981; Robertson, 1984; Hata and Kato, 2004; Precht et al., 2010). On Brazilian reefs, S. fuscus has been largely observed associated with hydrocoral colonies of the genus Millepora, competing with other reef fishes for food and habitat (Leal et al., 2013). Furthermore, Pereira et al. (2012) suggested that juveniles of S. fuscus could feed on Millepora spp. with feeding rates of up to 1.35 ± 0.93 bites/min. Hydrocorals from the genus Millepora are found in tropical coral reefs worldwide, and are considered the second most important reef builders, behind the scleractinian corals (Boschma, 1948; Lewis, 1989). The fire-coral Millepora spp. are the only species with conspicuous branching forms founds on Southwestern Atlantic reefs. Millepora alcicornis and Millepora braziliensis can form colonies with a diameter of more than 2 m and are considered one of the most abundant species along the Brazilian coast (Garcia et al., 2008). The habitat occupied by Millepora spp. in Brazilian reefs is comparable to those of Acropora palmata and Acropora cervicornis in the Caribbean. Previous studies have shown that Millepora spp. are important resources for tropical coral reefs and the associated organisms that inhabit these areas. For instance, a wide variety of fish and invertebrates use Millepora for food and shelter (Lewis, 1989, 2006; Garcia et al., 2008; Pereira et al., 2012; Coni et al., 2012; Leal et al., 2013; Brown and Edmunds, 2013). Varying patterns of association and territorial behaviour of S. fuscus in response to M. alcicornis colony volume could have a greater ecological influence on coral reef fish communities than previously thought. The present study examined how coral colony volume and the competitive interactions by a territorial damselfish, S. fuscus, may influence tropical reef fish communities associated with the firecoral M. alcicornis. The aims of this study were: (1) examine whether M. alcicornis coral colony volume influences reef fish species abundance and richness; (2) investigate whether the volume of a M. alcicornis colony affects S. fuscus abundance and behaviour; (3) analyse ontogenetic changes in S. fuscus behaviour

associated with M. alcicornis; (4) examine agonistic interactions between S. fuscus and other trophic guilds associated with the firecoral M. alcicornis. 2. Materials and methods 2.1. Study area The present study was conducted on the coastal reefs in the
(88
440 5400 S and 35
860 1400 W); located municipality of Tamandare 110 km from Recife, the capital of Pernambuco State. The climate has two distinct seasons; the rainy season which occurs from September to April and has average air and sea temperature of 20
C and 22
C, respectively. The dry season occurs from August to April, and has average air and sea temperature of 28
C and 27
C, respectively. The coral reef complex examined in the present study is located within the “APA Costa dos Corais” (Coral Coast MPA),
in which encompasses over 135 km of coastline, from Tamandare the southern state of Pernambuco, to Paripueira in the northern state of Alagoas. This study focused on the “Ilha do Norte” reef (Fig. 1), which is located 300 me600 m from the shoreline, and has an average depth of 2.5 m. The reef is composed by macroalgal species, hermatypic corals (Favia gravida, Montastrea cavernosa, Mussismilia spp., and Porites astreoides) (Lamarck, 1816), and hydrocoral colonies of M. alcicornis and M. braziliensis, (Ferreira and Maida, 2006; Pereira et al., 2014), with M. alcicornis being the more abundant. 2.2. Underwater census and behavioural observations Underwater observations were conducted by snorkel between September 2012 and April 2013. A total of 40 h of direct observation were collected during the present study. A total of 20 M. alcicornis colonies were randomly selected, tagged, and measured (diameter and height) with measuring tape. Volume (m3) of each colony was calculated using the formula for the volume of a cylinder (base area vs. colony height) (Holbrook et al., 2000; Coni et al., 2012). The base area was calculated from the diameter of the colony (longest base axis) and coral colony height was measured from colony substratum to the highest branch. Large Millepora spp. corals are very scarce and intermediate colonies are in low abundance in the sampling area. Therefore, due to the low amplitude homogeneity of fire-coral colony volumes and size distribution skewness in the study area (Leal et al., 2013), colony volume was characterized into differing size categories. The categories created were small (S ¼ 0.03e0.41 m3), medium (M ¼ 0.41e1.4 m3), large (L ¼ 1.4e2.3 m3), or very large (VL ¼ 2.3e7.85 m3). To evaluate the influence of Millepora coral colony volume on fish richness and abundance, the total number of fish and species in each colony was counted. Counting time and distance of the observer to each coral colony were standardized to 5 min and 1.5 m respectively. To maintain a consistent sampling, observations were performed by a single observer under the same tide, same time of day (08:00e14:00) and same time of year. The most abundant species recorded that was associated with fire-coral colonies in previous studies was S. fuscus (Coni et al., 2012; Leal et al., 2013). Therefore, to detect ontogenetic changes in behaviour in relation to body size, individuals of S. fuscus were classified as having a total length (TL) that was either less than 6.0 cm (juveniles) or greater than 6.0 cm (adults). Colouration patterns were also used in order to classify the life phase of each individual (Fig. 2). Additionally, due to Leal et al. (2013) suggesting that S. fuscus could include M. alcicornis colonies within their territories, this study wanted to evaluate if S. fuscus abundance and behaviour is affected by M. alcicornis colony size. To determine if M. alcicornis colony volume

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

47


municipality, Northeast Brazil, indicating the sampled M. alcicornis fire-coral colonies. N ¼ 20. Fig. 1. Map of the study area, “Ilha do Norte” reef in Tamandare

affected the distribution of S. fuscus individuals and their behaviour, M. alcicornis colonies were divided into the categories mentioned above.

2.3. S. fuscus behavioural observations S. fuscus behaviour associated with Millepora coral colonies was observed using focal animal sampling (Altmann, 1974) for five minutes observations. Behaviour categories were quantified as percentages of each five minutes observation intervals. Observations were recorded for each S. fuscus individual associated with the Millepora colony, consequently; each individual was observed and their displayed behaviour recorded continuously during the five minute period. Previous studies have shown that S. fuscus individuals display ontogenetic changes in their diet and behaviour
rio et al., 2006; Medeiros et al., 2010); and (Ferreira et al., 1998. Oso individuals inhabiting fire-coral colonies should display stronger ontogenetic changes due to changes in habitat use, food availability and competition. However, this has never been investigated for individuals associated with fire-coral colonies. To determine S. fuscus ontogenetic changes in behaviour, five categories were established. These categories included: Sheltered on coral branches: where individuals were recorded sheltering in the crevices and hollows of the coral; Feeding on algae: when individuals were observed eating algae on M. alcicornis structures; Feeding on coral polyps: when fishes were observed biting M. alcicornis polyps; Agonistic: when fishes were engaged in aggressive behaviour with other fish that tried to approach or were roving close to M. alcicornis colonies; Non-occupational swimming: recorded when associated fishes were swimming in close proximity to fire-coral colonies.

2.4. S. fuscus agonistic interactions When S. fuscus individuals were observed behaving aggressively toward fish that were within close proximity to M. alcicornis colonies, counts were taken to determine the agonistic interactions between S. fuscus and the trophic guilds associated with Millepora. Individuals that were confronted by S. fuscus’ aggressive behaviour were then grouped into seven trophic categories based on their main dietary preference (Ferreira et al., 2004) to evaluate the possibility of interspecific competition. These categories were: Territorial Herbivores (TERH); Roving Herbivores (ROVH); Mobile invertebrate feeders (MIF); Sessile invertebrate feeders (SIF); Carnivores (CAR); Omnivores (OMN) and Planktivores (PLANK). 2.5. Data analysis Logarithmic regressions were used to determine the relationship between M. alcicornis volume and reef fish abundance and richness. This analysis was chosen due to the non-linear nature of the data, and subsequent saturation of the number of individuals and species richness with increased coral colony volume. Chi-square tests were performed to determine whether the categories of M. alcicornis colony volume affected the abundance of S. fuscus individuals. A one-way analysis of similarity (ANOSIM), a non-parametric tool proposed by Clarke (1993) was used to test for a significant difference between S. fuscus behaviour and M. alcicornis coral colony categories. A BrayeCurtis distance was used to obtain a dissimilarity matrix, and this analysis was performed using Primer-e 6 PERMANOVAþ1.0 software (Ver. 6.1.14). The ANOSIM determined whether S. fuscus behaviour differed statistically between M. alcicornis coral colony size. The ANOSIM test statistic, R, is based on the ratio of the between-group to

48

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

Fig. 2. S. fuscus individuals at juvenile (A) and adult (B) life phase and the associated M. alcicornis fire-coral colonies (C).

within-group similarity ranking and ranges from 0 to 1, with the value indicating the degree of dissimilarity (1 ¼ completely dissimilar; 0 ¼ completely similar). To investigate ontogenetic changes in S. fuscus behaviour between juveniles and adults we also applied a one-way permutational multivariate analysis of variance (PERMANOVA) on a univariate way. S. fuscus data was grouped according to individuals life phase, log transformed (Xþ1) and reassembled in a BrayeCurtis similarity matrix. Unrestricted permutation of raw data was applied as the best technique analysing one factor. A permutational analysis of multivariate dispersions (PERMIDISP) was also applied to analyse whether the multivariate variations are homogeneous (Anderson et al., 2008; Anderson and Walsh, 2013). Additionally, a Principal component analysis (PCA) by life phase and behaviour categories was used to show the relationship between S. fuscus juveniles and adults and their behavioural changes. All the data were standardized and log-transformed prior to multivariate analyses. Primer-e 6 PERMANOVAþ1.0 software (Ver. 6.1.14) was used to conduct PERMANOVA, PERMIDISP and PCA analysis. Ivlev's index of electivity (Ivlev, 1961) was calculated to examine agonistic interactions between S. fuscus and fish categorized in trophic guilds associated with the M. alcicornis coral colonies. The index was calculated using the following equation:

Ei ¼

ri  pi ri þ pi

where electivity for the trophic guild i (Ei) was considered the

percentage of the trophic guild i associated with fire coral colonies (ri) and the percentage of agonistic interactions performed against the trophic guild i (pi). E values vary from 1.0 to þ1.0, with positive values indicating agonistic interaction and negative values indicating no interaction; values equal to 0 are an intermediate relationship in which some individuals interacted agonistically while others did not. Univariate regression trees (URT) using Tree Plus were then used to explore the relationship of coral colony volume and the influence of S. fuscus presence on fish abundance and species richness associated with fire-coral colonies. Coral colony volume was categorized as small, medium, large and very large following the description mentioned above. S. fuscus abundance on fire-coral colonies was then categorized in low (less than 3 individuals per coral colony), medium (3e6 individuals) and high (>6 individuals). URT is well suited for describing patterns in complex ecological datasets because they separate the variables in a series of binary splits and continuous and categorical variables can be compared in the same analyses (De'ath and Fabricius 2000). 3. Results A total of 1051 reef fish from 24 species were observed associated with M. alcicornis fire-coral colonies (Table 1). Among them, S. fuscus was the most abundant species, representing 56.8% of the fish assemblage (n ¼ 598). The most represented family was Pomacentridae, which represented 64% of the fish assemblage. Logarithmic regressions showed that coral colony volume

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

49

Table 1 Fish species and their respective families associated with Millepora alcicornis colonies and corresponding trophic guilds, listed in decreasing order of the total number of individuals and relative frequency of occurrence. TERH ¼ territorial herbivores; OMN ¼ omnivores; ROVH ¼ roving herbivores; MIF ¼ mobile invertebrate feeders; SIF ¼ sessile invertebrate feeders; CARN ¼ carnivores; PLANK ¼ planktivores. Family

Species

Trop. Guilds

Individuals (N)

Freq. occur. (%)

Pomacentridae Holocentridae Pomacentridae Epinephilidae Acanthuridae Holocentridae Labrisomidae Epinephelidae Scaridae Blenniidae Scianidae Chaetodontidae Haemulidae Scaridae Haemulidae Labridae Lutjanidae Pomacentridae Labrisomidae Lutjanidae Labridae Pomacentridae Monacanthidae Muraenidae

Stegastes fuscus (Curvier, 1830) Holocentrus adscensionis (Osbeck, 1765) Abudefduf saxatilis (Linnaeus, 1758) Ephinephelus adscensionis (Osbeck, 1765) Acanthurus coeruleus Bloch and Schneider, 1801 Myripristis jacobus Curvier, 1829 Labrisomus nuchipinnis (Quoy and Gaimard, 1824) Cephalopholis fulva (Linnaeus, 1758) Sparisoma axillare (Steindachner, 1878) Ophioblenius trinitatis (Miranda Ribeiro, 1919) Pareques acuminatus (Bloch and Schneider, 1801) Chaetodon striatus Linnaeus, 1758 Anisotremus virginicus (Linnaeus, 1758) Scarus zelindae Moura, Figueiredo & Sazima, 2001 Haemulon squamipinna (Desmarest, 1823) Halichoeres brasiliensis (Bloch, 1791) Lutjanus spp. (Bloch, 1787) Chromis multilineata (Guichenot, 1853) Labrisomus kalisherae (Jordan, 1904) Ocyurus chysurus (Bloch, 1791) Halichoeres poyei (Steindachner, 1867) Microspathodon chrysurus (Curvier, 1830) Aluterus scriptus (Osbeck, 1765) Muraena pavonina Richardson, 1845

TERH MIF OMN CARN ROVH PLANK CARN CARN ROVH TERH MIF SIF MIF ROVH MIF MIF CARN PLANK CARN CARN ROVH TERH OMN CAR

598 111 73 46 42 34 31 21 17 15 10 9 8 8 7 6 3 3 3 2 1 1 1 1

56.8 10.5 6.9 4.3 4 3.2 3 2 1.6 1.4 0.9 0.8 0.7 0.7 0.6 0.5 0.2 0.2 0.2 0.2 0.09 0.09 0.09 0.09

increased both the abundance (R2 ¼ 0.78; p < 0.05) and richness (R2 ¼ 0.79; p < 0.05) of associated reef fish species (Fig. 3A and B). The volume of fire-coral colonies affected the abundance of

S. fuscus individuals, particularly in colonies classified as very large (VL). The proportion of juveniles was very similar among colonies with small, medium, and large volumes (21%e27%). However, VL colonies included juveniles in 45% of the analysed colonies. The number of S. fuscus juveniles and adults in colonies classified as very large (VL) was significantly different from the other colony class sizes (chi-square ¼ 37.12, df ¼ 4, p ¼ < 0.001). Significant differences between M. alcicornis coral colony size categories and S. fuscus behaviour were observed (R ¼ 0.68, p < 0.001) (Table 2). Specifically, behavioural differences among certain coral colony categories were confirmed by an ANOSIM test for small vs. very large (R ¼ 1.00, p ¼ 0.001), medium vs. large (R ¼ 0.81, p ¼ 0.03), medium vs. very large (R ¼ 0.93, p < 0.02), large vs. very large (R ¼ 0.81, p < 0.03) (Table 2). The frequency of agonistic behaviour and feeding on polyps categories increased with an increase in colony volume (Fig. 4). For instance, agonistic behaviour represented 19% of all behaviours in colonies classified as small (S) and 38% in VL colonies. Feeding on coral polyps represented 2% of all behaviours in colonies classified as (S) and 18% in VL colonies. On the other hand, this trend was reversed for the nonoccupational swimming category, which decreased from 53% in the S colonies to 9% in the VL colony (Fig. 4) (Table 3) S. fuscus individuals associated with M. alcicornis colonies showed significant behavioural ontogenetic changes (PERMANOVA; Pseud F ¼ 48.67; p < 0.01) and a non significant PERMIDISP test (F ¼ 7.18, P (perm) ¼ 0.32) revealed that multivariate variations

Table 2 Results of analyses of similarity (ANOSIM) evaluating variation in the S. fuscus behaviour and M. alcicornis coral colony categories. S ¼ small; M ¼ medium; L ¼ large and VL ¼ very large.

Fig. 3. Logarithmic regressions between M. alcicornis coral colony volume (m2) and the total abundance (A) and richness (B) of associated fish.

M. alcicornis colony size

R statistic

P value

Small vs. medium Small vs. large Small vs. very large Medium vs. large Medium vs. very large Large vs. very large

0.52 0 1.0 0.81 0.93 0.81

0.2 0.1 0.001 0.03 0.02 0.03

50

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

Table 3 Fish species and their respective families that S. fuscus performed agonistic behaviour, listed in decreasing order of the total number of individuals encountered, the percentage of individuals registered, and the corresponding trophic guilds. Species

Family

Individuals (N)

Freq. occur. (%)

Trophic guilds

Stegastes fuscus Abudefduf saxatilis Acanthurus coeruleus Halichoeres brasiliensis Scarus zelindae Sparisoma axillare Lutjanus spp. Chaetodon striatus Holocentrus adscensionis Microspathodon chrysurus Aluterus scriptus Ophioblennius trinitatis Myripristis jacobus Chromis multilineata

Pomacentridae Pomacentridae Acanthuridae Labridae Labridae Labridae Lutjanidae Chaetodontidae Holocentridae Pomacentridae Monacanthidae Blenniidae Holocentridae Pomacentridae

33 29 18 10 6 4 3 2 1 1 1 1 1 1

30 26 16 9 5 4 3 2 0.9 0.9 0.9 0.9 0.9 0.9

TERH OMN ROVH MIF ROVH ROVH CARN SIF MIF TERH OMN TERH PLANK PLANK

Fig. 4. Behaviour patterns of S. fuscus associated with coral colony size category (S ¼ small, M ¼ medium, L ¼ large, and VL ¼ very large).

are homogeneous strengthening the PERMANOVA results. For juveniles, only two behaviours were detected: sheltered on coral branches (55%) and feeding on coral polyps (23%). In contrast, for adults a proportional distribution among behaviour categories was observed: non-occupational swimming (29%), agonistic (26%), and feeding on algae (24%). Principal component analysis (PCA) was also performed and explained 58.2% of the total variability in behavioural changes, with 32.1% of the variability explained by PC1 (Eigenvalue 718) and 26.1% by PC2 (Eigenvalue 256) (Fig. 5). During this analysis an evident separation was observed once the split in life phase behaviour was highly correlated with non-occupational swimming for adults and sheltered on coral branches and feeding on coral polyps for juveniles (Fig. 5). S. fuscus individuals performed 111 agonistic interactions against 13 different tropical reef fish species, including intraspecific interactions. Territorial herbivores (mainly S. fuscus) were the most affected trophic guild. The absolute abundance of interactions within this guild was expected, once it was the most frequently observed trophic guild associated with fire-coral colonies. Nevertheless, based on the Ivlev's electivity index (Fig. 6), the trophic guilds positively interacted by S. fuscus for agonistic interactions, in comparison to their total abundance, were roving herbivores, omnivores and sessile invertebrate feeders. In contrast, other trophic guilds, such as carnivores, territorial herbivores and planktivores were negatively selected and received a proportionally low number of agonistic interactions, in comparison to their total abundance (Fig. 6). Lastly, univariate regression tree (URT) analysing fish abundance and species richness associated with fire-coral colonies featuring

Fig. 5. Principal component analysis (PCA) of ontogenetic changes in the behaviour demonstrated by juveniles and adults individuals of S. fuscus associated with M. alcicornis fire-coral colonies.

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

51

lowest richness (less than 1.22 species) were correlated with high S. fuscus presence (Fig. 7). 4. Discussion

Fig. 6. Ivlev Electivity index of relative abundance and agonistic interactions demonstrated by S. fuscus individuals against different reef fish trophic guilds associated with M. alcicornis fire-coral colonies. See Table 1 for abbreviations.

the best standard error (S.E) resulted in a 4-leaf tree (Fig. 7). Both the abundance and richness of reef fishes associated with Millepora spp. were mostly influenced by coral colony volume. Abundance higher than 44.5 individuals and richness up to 2.26 species were observed associated with very large Millepora spp. colonies. In contrast, the smallest abundance (less than 17.5 individuals) and

The present study shows that M. alcicornis colony volume positively affects abundance and species richness of associated reef fish. Our results emphasize the close and complex association of fish assemblage and fire-corals in the South Atlantic Ocean. The volume of M. alcicornis coral colonies also affected the behaviour of S. fuscus, the most abundant species found associated with firecoral colonies. Agonistic behaviour and feeding on polyps increased relatively to an increase in coral colony volume. This trend was reversed for non-occupational swimming category, which decreased with an increase in colony volume. Behavioural ontogenetic changes were also detected for S. fuscus individuals. Juveniles mainly showed two distinct behaviours: sheltered on coral branches and feeding on coral polyps. In contrast, adults were often observed during non-occupational swimming and performing agonistic behaviour. Lastly, S. fuscus individuals displayed a higher territorial behaviour against potential food competitors that seemed to influence the reef fish community associated with fire-corals. These results indicate that the volume of fire-coral colonies and the presence of a territorial damselfish are drivers affecting the associated reef fish community. The complex structures created by branching cnidarians, such as Millepora spp. and Acropora spp. have shown positive effects on the distribution and abundance of reef fish (Holbrook et al., 2000; McClanahan and Arthur, 2001; Johnson et al., 2011). In general,

Fig. 7. Four-leaf regression trees of reef fish community abundance (A) and species richness (B) associated with Millepora spp. fire-corals. Variables were small, medium and large coral colony volume and low, medium and high abundance of S. fuscus.

52

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

reef fish abundance and species richness varied according to coral colony size and structural complexity (Holbrook et al., 2000; Lomolino, 2001; Schiemer et al., 2009). Coni et al. (2012) briefly suggested that the volume of Millepora spp. colonies could affect both the richness and abundance of fish species associated with Abrolhos Bank in Northeast Brazil. This hypothesis was fully supported by the present study where a positive relationship between Millepora colony volume and fish abundance and species richness was observed. Colony size also seemed to influence species demographic traits and other fish community attributes. It appears to determine growth, survival and reproduction for some coraldwelling fish (Munday, 2001; Brooker et al., 2013), and accounted for variation in group size, juvenile survival and habitat selection in the damselfish Chromis retrofasciata (Noonan et al., 2012). Therefore, it is likely that the survival of some reef fishes such as Labrisomus kalisherae, is strictly dependent on fire-coral colonies once they are normally not recorded out of Millepora spp. (pers. obs.). Environmental variables such as current, water temperature, depth and seasonality may also influence fish abundance and species richness associated with fire-coral colonies. Although we have not specifically analysed these variables it is unlikely they could have an effect on our results. For instance, variations in weather conditions were small during the present study period and seasonality does not affect fishes associated with fire-corals colonies once they are relatively sedentary (Leal et al., 2013). On the other hand, habitat structural variables such as rugosity, substratum composition and refuge metrics are relevant to explain spatial patterns in invertebrate/fish diversity and species-habitat relationships (Alexander et al., 2009; Cameron et al., 2014). The rugosity and substratum composition on Millepora coral colonies are very similar when comparing different coral colonies, so our selected variable (i.e. coral volume), seems to be the most significant factor affecting fish abundance and species richness. S. fuscus individuals represent more than half the tropical reef fish species sampled in Millepora colonies during the present study, which indicates a strong association between S. fuscus and Millepora. S. fuscus was also the most abundant species recorded associated with Millepora spp. colonies in Abrolhos Bank (NE Brazil) and accounted for 43% of the fish community in that area (Coni et al., 2012). Johnson et al. (2011) observed a similar specialized relationship in the South Pacific Ocean between Acropora pulchra and Stegastes planifrons. Additionally, Precht et al. (2010) suggested that a reduction in the availability of S. planifrons preferred habitat, A. cervicornis, could alter the distribution and recruitment of this damselfish in the Caribbean. These findings highlight the importance of Millepora coral colonies as a tropical reef ecosystem engineer, for site attached reef fishes such as S. fuscus, due to the absence of scleractinian branching corals in the Southwest Atlantic Ocean. Behavioural changes demonstrated by S. fuscus individuals were also influenced by Millepora colony volume. A direct relationship was seen between agonistic behaviour and the volume of M. alcicornis, once S. fuscus individuals acquired sufficient time to establish and defend their territory. The concept of economic defensibility is widely used to explain territorial behaviour in tropical reef fishes (Mumby and Wabnitz, 2002), and states that the development of territorial behaviour is only established when the benefits acquired from exclusive access to a particular resource are greater than the cost of defending that resource (Brown, 1964). Therefore, it is likely that S. fuscus benefits from access to larger fire-coral colonies are favourable even with higher defending cost. Larger coral colonies could provide better breeding/nesting grounds and also superior food quality as suggested for coraldwelling gobies (Hobbs and Munday, 2004; Schiemer et al., 2009). Different ontogenetic behavioural changes were also

demonstrated by S. fuscus utilizing M. alcicornis coral colonies during the present study. Juvenile individuals were observed more frequently among coral branches and feeding on corals, which may be due to their smaller body size allowing for access to M. alcicornis branches, which are not easily accessed by larger predators. Juveniles of S. fuscus likely spend more time sheltering from predators in protected locations (Menegatti et al., 2003) a same trend that was observed in the present study. In contrast, individuals greater than 6 cm in length (adults), were mostly observed eating algae within the M. alcicornis structure. Previous studies have shown that highly territorial species (e.g. S. fuscus) may cause an increase in algal growth in their territories (Ceccarelli et al., 2001), and this behaviour can affect the competitive interaction between algae and corals (Vine, 1974; Lobel, 1980; Sammarco et al., 1986). Some species can destroy large volumes of corals to promote algae growth, resulting in reduced coral cover (Potts, 1977; Wellington, 1982). However, an alternative hypothesis assumes that associated reef fish (e.g. coral-dwelling gobies) can increase coral health through nutrient input (Holbrook et al., 2008) and also by removing toxic seaweeds (Dixson and Hay, 2012). In this context, the ecological role of the Brazilian endemic S. fuscus and its consequences for the M. alcicornis coral host require further evaluation. For examples, if S. fuscus individuals reduce algal growth while inhabiting fire-coral colonies then its presence should favour coral performance. However, increased feeding on Millepora tissue by S. fuscus could damage Millepora structure, and act as potential vector of coral disease, as observed for some corallivorous species (Chong-Seng et al., 2011). To further elucidate the impact of S. fuscus on Millepora performance, cage-exclusion experiments could be used to test Millepora performance with and without S. fuscus presence. Adults of S. fuscus exhibited higher territorial behaviour and showed agonistic interactions against other reef fish. The frequency of agonistic interactions usually increased with fish size, as was previously reported for the families Labridae (Jones, 1984), Lutjanidae (Mueller et al., 1994), Haemulidae (Pereira and Ferreira,
rio et al., 2012), and Pomacentridae (Ferreira et al., 1998; Oso 2006; Medeiros et al., 2010; Souza et al., 2011). Agonistic behaviour can be observed between species that are territorial, nonterritorial, or both (Robertson et al., 1976). These interactions are mainly due to competition for food resources, but may also occur with competition for space and habitat segregation (Johnson et al., 2011). During the present study, a large number of agonistic interactions on fire-coral colonies occurred between S. fuscus and potential food competitors, which were mainly roving herbivores, omnivores, and sessile invertebrate feeders. A similar trend was observed by Souza et al. (2011) for Stegastes rocasensis that performed agonistic interactions against mostly herbivores and omnivores. Therefore, food competition based on overlapping diets may be an important factor affecting the tropical reef fish community associated with M. alcicornis colonies. According to previous studies (Ebersole, 1977; Harrington and Losey, 1990; Draud and Itzkowitz, 1995) territorial aggression is escalated based on the degree of ecological overlap among the species entering the territory. For example, individuals of S. rocasensis guarding their egg clutches showed agonistic reactions towards planktivores and omnivores, which could potentially prey on their eggs (Souza and Ilarri, 2014). However, individuals without egg clutches in their territories showed agonistic interactions mainly towards herbivorous fish (Souza and Ilarri, 2014). Pereira et al. (in prep.) discussed recently the efficiency of different methods (direct diver observation vs. remote cameras) for sampling reef fish behaviour associated with coral colonies. During fishecoral interactions, fish behaviour seems to be influenced by diver presence, where fish spent more time sheltered on the coral branches during direct diver observation compared to passively

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

swimming by coral colonies with a remote video. Despite this evidence, we do not believe that diver presence influenced S. fuscus behaviour during the present study, especially due to the highly territoriality and aggressiveness of this species, which does not usually fear human activity. Assuming the hypothesis that space is a limiting resource for coral reef fish (Holbrook and Schmitt, 2002; Forrester et al., 2006; Bonin et al., 2009), fire-coral colonies may be considered habitats for superior competitors in Brazilian waters due to the lack of architecturally complex cnidarian structures (e.g. Acropora spp.). According to previous studies, Millepora spp. coral colonies are an important source of refuge and shelter for S. fuscus; Holocentrus adscensionis, Parablennius marmoreus and Malacoctenus sp. mostly during the juvenile life phase (Coni et al., 2012; Leal et al., 2013). Therefore, the tolerance of fish to the powerful stinging dactylozooids characteristic of Millepora spp. may be species-specific, as well as dependent on life phase. Further research using laboratory experiments are necessary to test competitive hierarchies and tolerance for nematocysts on Brazilian reef fishes. The genus Millepora is the only branching cnidarian found in the Southwest Atlantic Ocean (Lewis, 1989). It is considered one of the more susceptible to bleaching events, in comparison to other hermatypic corals (Marshall and Baird, 2000). According to Bozec et al. (2015) under severe and frequent thermal stress, a model predicted a dramatic loss of coral complexity over the next decades. Consequently, a decrease in abundance and complexity of Millepora coral colonies caused by bleaching events, may decrease the architectural complexity of the tropical reef environment and therefore affect the abundance and species richness of reef fish communities. The present study, which is the first to focus on the ecological aspects of the fire coral M. alcicornis, showed an increase in coral colony volume favoured a positive association (abundance and richness) to tropical reef fish. Moreover, M. alcicornis colonies had high abundances of S. fuscus, which showed ontogenetic changes in their behaviour that affected other tropical reef fish associated to Millepora, especially for food and habitat competitive interactions. M. alcicornis coral colonies are extremely important habitats for reef fish in the South Atlantic Ocean, and decreases in Millepora abundances may dramatically affect natural processes such as recruitment and competition within reef fish communities. Acknowledgements The authors would like to thank Ronaldo Franici-Filho, Allan Souza, Darren Brown and Matthew Jankowski for helping in the manuscript preparation. We also thank Marcus Vinícius Santos for the photograph of S. fuscus adult. Conselho Nacional de Desenvol
gico, Coordenaç~ vimento Científico e Tecnolo ao de Aperfeiçoamento
gicas de Longa de Pessoal de Nível Superior, Pesquisas Ecolo ~o (PELD), Rufford Small Grants (RSG) and Idea Wild for the Duraça financial and logistical support. References Alexander, T.J., Barrett, N., Haddon, M., Edgar, G., 2009. Relationships between mobile macroinvertebrates and reef structure in a temperate marine reserve. Mar. Ecol. Prog. Ser. 389, 31e44. Altmann, J., 1974. Observational study of behavior: sampling methods. Behavior 49, 227e265. Almany, G.R., 2003. Priority effects in coral reef fish communities. Ecology 84, 1920e1935. Almany, G.R., 2004. Does increased habitat complexity reduce predation and competition in coral reef fish assemblage? Oikos 106, 275e284. Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA þ for PRIMER: Guide to Software and Statistical Methods. PRIMER-E, Plymouth. Anderson, M.J., Walsh, D.C.I., 2013. What null hypothesis are you testing? PERMANOVA, ANOSIM and the Mantel test in the face of heterogeneous dispersions. Ecol. Monogr. 83, 557e574.

53

Bonin, M.C., Srinivasan, M., Almany, G.R., Jones, G.P., 2009. Interactive effects of interspecific competition and microhabitat on early post-settlement survival in a coral reef fish. Coral Reefs 28, 265e274. Boschma, H., 1948. The species problem in Millepora. Zool. Verh. (Leiden) 1, 3e115. Bozec, Y.-M., Filip, L.A., Mumby, P.J., 2015. The dynamics of architectural complexity on coral reefs under climate change. Glob. Change Biol. 21, 223e235. Brooker, R.M., Jones, G.P., Munday, P.L., 2013. Prey selectivity affects reproductive success of a corallivorous filefish. Oecologia 172, 409e416. Brooks, A.J., Holbrook, S.J., Schmitt, R.J., 2007. Patterns of microhabitat use by fishes in the patch-forming coral Porites rus. Raffles Bull. Zool. 14, 245e254. Brown, J.L., 1964. The evolution of diversity in avian territorial systems. Wilson Bull. 76, 160e169. Brown, D., Edmunds, P.J., 2013. The hermit crab Calcinus tibicen lives commensally on Millepora spp. in St. John, United States Virgin Islands. Coral Reefs 32, 127e135. ~ o, P.M., 2008. Effects of coral transplantation and Cabaitan, P.C., Gomez, E.D., Alin giant clam restocking on the structure of fish communities on degraded patch reefs. J. Exp. Mar. Biol. Ecol. 357, 85e98. Cameron, M.J., Lucieer, V., Barrett, N.S., Johnson, C.R., Edgar, G.J., 2014. Understanding community-habitat associations of temperate reef fishes using fineresolution bathymetric measures of physical structure. Mar. Ecol. Prog. Ser. 506, 213e229. Ceccarelli, D.M., Jones, G.P., McCook, L.J., 2001. Territorial damselfishes as determinants of the structure of benthic communities on coral reefs. Oceanogr. Mar. Biol. Annu. Rev. 39, 355e389. Chaves, L.T., Pereira, P.H.C., Feitosa, J.L., 2013. Coral reef fish association with macroalgal beds on a tropical reef system in North-eastern Brazil. Mar. Freshw. Res. 64, 1101e1111. Chong-Seng, K.M., Cole, A.J., Pratchett, M.S., Willis, B.L., 2011. Selective feeding by coral reef fishes on coral lesions associated with brown band and black band disease. Coral Reefs 30, 473e481. Clarke, K.R., 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18, 117e143. Coni, E.O.C., Ferreira, M.C., Moura, R.L., Meirelles, P.M., Kaufman, L., FranciniFilho, R.B., 2012. An evaluation of the use of branching fire-corals (Millepora spp.) as refuge by reef fish in the Abrolhos Bank, eastern Brazil. Environ. Biol. Fishes 75, 1e12. De'ath, G., Fabricius, K.E., 2000. Classification and regression trees: a powerful yet simple technique for the analysis of complex ecological data. Ecology 81, 3178e3192. Dixson, D.L., Hay, M.E., 2012. Corals chemically cue mutualistic fishes to remove competing seaweeds. Science 338, 804e807. Draud, M.J., Itzkowitz, M., 1995. Interspecific aggression between juveniles of two Caribbean Damselfish species (Genus Stegastus). Copeia 431e435. Ebersole, J.P., 1977. Adaptive significance of interspecific territoriality in the reef fish Eupomacentrus leucostictus. Ecology 58, 914e920. Eggleston, D.B., Lipcius, R.N., Grover, J.J., 1997. Predation and shelter size effects on coral reef fish and spiny lobster prey. Mar. Ecol. Prog. Ser. 149, 43e59. €v, P., Diehl, S., 1994. Piscivore efficiency and refuging prey: the importance of Eklo predator search mode. Oecologia 98, 344e353. Ferreira, C.E.L., Gonçalves, J.E.A., Coutinho, R., Peret, A.C., 1998. Herbivory by the dusky damselfishStegastes fuscus (Cuvier, 1830) in a tropical rocky shore: effects on the benthic community. J. Exp. Mar. Biol. Ecol. 229, 241e264. Ferreira, C.E.L., Floeter, S.R., Gasparine, J.L., Joyeux, J.C., Ferreira, B.P., 2004. Trophic structure patterns of Brazilian reef fishes: a latitudinal comparison. Jounal Biogeogr. 31, 1093e1106. Ferreira, B.P., Maida, M., 2006. Monitoramento dos recifes de coral do Brasil: sit~o atual e perspectivas. Brasília. MMA, 250 pp. [Serie Biodiversidade 18.]. uaça Forrester, G.E., Evans, B., Steele, M.A., Vance, R.R., 2006. Assessing the magnitude of intra- and interspecific competition in two coral reef fishes. Oecologia 148, 632e640. Garcia, T.M., Matthews-Cascon, H., Franklin-Junior, W., 2008. Macrofauna associated with branching fire-coral Millepora alcicornis (Cnidaria: Hydrozoa). Braz. J. Oceanogr. 57, 153e155. € Garpe, K.C., Yahya, S.A.S., Lindahl, U., Ohman, M.C., 2006. Long-term effects of the 1998 coral bleaching event on reef fish assemblages. Mar. Ecol. Prog. Ser. 315, 237e247. Geange, S.W., Stier, A.C., Shima, J.S., 2013. Competitive hierarchies among three species of juvenile coral reef fishes. Mar. Ecol. Prog. Ser. 472, 239e248. Graham, N.A.J., Nash, K.L., 2012. The importance of structural complexity in coral reef ecosystems. Coral Reefs 32, 315e326. Grober-Dunsmore, R., Frazer, T.K., Lindberg, W.J., Beets, J., 2007. Reef fish and habitat relationships in a Caribbean seascape: the importance of reef context. Coral Reefs 26, 201e216. Hata, H., Kato, M., 2004. Monoculture and mixed-species algal farms on a coral reef are maintained through intensive and extensive management by damselfishes. J. Exp. Mar. Biol. Ecol. 313, 285e296. Harrington, M.E., Losey, G.S., 1990. The importance of species identification and location on interspecific territorial defense by the damselfish, Stegastes fasciolatus. Environ. Biol. Fishes 27, 139e145. Hixon, M.A., Beets, J.P., 1993. Predation, prey refuges and the structure of coral reef fish assemblages. Ecol. Monogr. 63, 77e101. Hobbs, J.P.A., Munday, P.L., 2004. Intraspecific competition controls spatial distribution and social organization of the coral-dwelling goby Gobiodon histrio. Mar. Ecol. Prog. Ser. 278, 253e259.

54

I.C.S. Leal et al. / Marine Environmental Research 108 (2015) 45e54

Holbrook, S.J., Schmitt, R.J., 2002. Competition for shelter space causes density dependent predation mortality in damselfishes. Ecology 83, 2855e2868. Holbrook, S.J., Brooks, A.J., Schmitt, R.J., 2000. Spatial patterns in abundance of a damselfish reflect availability of suitable habitat. Oecologia 122, 109e120. Holbrook, S.J., Schmitt, R.J., Brooks, A.J., 2008. Resistance and resilience of a coral reef fish community to changes in coral cover. Mar. Ecol. Prog. Ser. 371, 263e271. Ivlev, V.S., 1961. Experimental Ecology of the Feeding of Fishes. Yale University Press, San Diego. Johnson, M.K., Holbrook, S.J., Schmitt, R.J., Brooks, A.J., 2011. Fish communities on staghorn coral: effects of habitat characteristics and resident farmerfishes. Environ. Biol. Fishes 91, 429e448. Jones, G.P., 1984. Population ecology of the temperate reef fish Pseudolabrus celidotus Bloch and Schneider (Pisces: Labridae). II. Factors influencing recruitment. J. Exp. Mar. Biol. Ecol. 75, 257e276. Jones, G.P., 1988. Ecology of rocky reef fish of north-eastern Australia. New Zealand: view. Mar. Freshw. Res. 22, 445e462. Leal, I.C.S., Pereira, P.H.C., Araújo, M.E., 2013. Coral reef fish association and behaviour on the fire-coral Millepora spp. in north-east Brazil. J. Mar. Biol. Assoc. U. K. 93 (1703), 1711. Lewis, J.B., 1989. The ecology of Millepora e a review. Coral Reefs 8, 99e107. Lewis, J.B., 2006. Biology and ecology of the hydrocoral Millepora on coral reefs. Adv. Mar. Biol. 50, 1e55. Lobel, P.S., 1980. Herbivory by damselfishes and their role in coral reef community ecology. Bull. Mar. Sci. 30, 273e289. Lomolino, M.V., 2001. Elevational gradients of species diversity: historical and prospective views. Glob. Ecol. Biogeogr. 10, 3e13. MacNeil, M.A., Graham, N.A.J., Polunin, N.V.C., Kulbicki, M., Galzin, R., HarmelinVivien, M., Rushton, S.P., 2009. Hierarchical drivers of reef-fish meta community structure. Ecology 90, 252e264. Mahoney, B.M., 1981. An examination of interspecific territoriality in the dusky damselfish, Eupomacentrus dorsopunicans. Bull. Mar. Sci. 31, 141e146. Marshall, P.A., Baird, A.H., 2000. Bleaching of corals on the Great Barrier Reef: differential susceptibilities among taxa. Coral Reefs 19, 155e163. McClanahan, T.R., Arthur, R., 2001. The effect of marine reserves and habitat on populations of East African coral reef fishes. Ecol. Appl. 11, 559e569. Medeiros, P.R., Souza, A.T., Ilarri, M.I., 2010. Habitat use and behavioural ecology of the juveniles of two sympatric damselfishes (Actinopterygii: Pomacentridae) in the southwestern Atlantic Ocean. J. Fish Biol. 77 (7), 1599e1615. ~es agonísticas e forMenegatti, J.V., Vescovi, D.L., Floeter, S.R., 2003. Interaço rageamento do peixe-donzela, Stegastes fuscus (Perciformes: Pomacentridae). Nat. on line 1, 45e50. Mueller, K.W., Dennis, G.D., Eggleston, D.B., Wicklund, R.I., 1994. Size-specific social interactions and foraging styles in a shallow water population of mutton snapper, Lutjanus analis (Pisces: Lutjanidae), in the central Bahamas. Environ. Biol. Fishes 40, 175e188. Mumby, P.J., Wabnitz, C.C.C., 2002. Spatial patterns of aggression, territory size, and harem size in five sympatric Caribbean parrot fish species. Environ. Biol. Fishes 63, 265e279. Munday, P.L., 2001. Fitness consequences of habitat use and competition among coral-dwelling fishes. Oecologia 128, 585e593. Noonan, S.H.C., Jones, G.P., Pratchett, M.S., 2012. Coral size, health and structural

complexity: effects on the ecology of a coral reef damselfish. Mar. Ecol. Prog. Ser. 456, 127e137.
rio, R.M., Rosa, I.L., Cabral, H., 2006. Territorial defense by the Brazilian damsel Oso Stegastes fuscus (Teleostei:Pomacentridae). J. Fish Biol. 69, 233e242. Pereira, P.H.C., Ferreira, B.P., Rezende, S.M., 2010. Community structure of the ichthyofauna associated with seagrass beds (Halodule wrightii) in Formoso River ^nc. 82, 1e12. estuary e Pernambuco, Brazil. An. Acad. Bras. Cie Pereira, P.H.C., Ferreira, B.P., 2012. Agonistic behavior of Haemulon spp. coral reef fishes (Actinopterygii: Hamulidae) in Northeastern Brazil. Cybium 36, 361e367. Pereira, P.H.C., Leal, I.C.S., Araújo, M.E., Souza, A.T., 2012. Feeding association between reef fishes and the fire-coral Millepora spp. (Cnidaria: Hydrozoa). Mar. Biodivers. Rec. 5, 1e4. Pereira, P.H.C., Moraes, R.L., dos Santos, M.V.B., Lippi, D.L., Feitosa, J.L.L., Pedrosa, M., 2014. The influence of multiple factors upon reef fish abundance and species richness in a tropical coral complex. Ichthyol. Res. 61, 375e384. Pereira, P.H.C., Leal, I.C.S., Araújo, M.E., 2015. Influence of diver presence on behavioural observations of reef fish associated with fire-corals. J. Exp. Mar. Biol. Ecol. (in prep.). Potts, D.C., 1977. Suppression of coral populations by filamentous algae within damselfish territories. J. Exp. Mar. Biol. Ecol. 28, 207e216. Precht, W.F., Aroson, R.B., Moody, R.M., Kaufman, L., 2010. Changing patterns of microhabitat utilization by the threespot damselfish, Stegastes planifrons, on Caribbean reefs. PLoS ONE 5, 10835. Robertson, D.R., 1984. Cohabitation of competing territorial damselfishes on a Caribbean coral reef. Ecology 65, 1124e1135. Robertson, D.R., Sweatman, H.P.A., Fletcher, E.A., Cleland, M.G., 1976. Schooling as a mechanism for circumventing the territoriality of competitors. Ecology 57, 1208e1220. Sale, P.F., 1977. Maintenance of high diversity in coral reef fish communities. Am. Nat. 111, 337e359. Sammarco, P.W., Carleton, J.H., Risk, M.J., 1986. Effects of grazing and damselfish territoriality on bioerosion of dead corals: direct effects. J. Exp. Mar. Biol. Ecol. 98, 1e19. Schiemer, L., Niedermuller, S., Herler, J., 2009. The influence of colony size and coral health on the occupation of coral-associated gobies (Pisces: Gobiidae). Coral Reefs 28, 137e142. Shulman, M.J., 1984. Resource limitation and recruitment patterns in a coral reef fish assemblage. J. Exp. Mar. Biol. Ecol. 74, 85e109. Souza, A.T., Ilarri, M.I., Rosa, I.L., 2011. Habitat use, feeding and territorial behavior of a Brazilian endemic damselfish Stegastes rocasensis (Actinopterygii: Pomacentridae). Environ. Biol. Fishes 91, 133e144. Souza, A.T., Ilarri, M.I., 2014. Behavioral changes of a Brazilian endemic damselfish Stegastes rocasensis when guarding egg clutches. Environ. Biol. Fishes 97, 1295e1303. Syms, C., Jones, G.P., 2000. Disturbance, habitat structure, and the dynamics of a coral-reef fish community. Ecology 81, 2714e2729. Vine, P.J., 1974. Effect of algal grazing and aggressive behaviour of the fishes Pomacentrus lividus and Acanthurus sohal on coral-reef ecology. Mar. Biol. 24, 131e136. Wellington, G.M., 1982. Depth zonation of corals in the Gulf of Panama: control and facilitation by resident reef fishes. Ecol. Monogr. 52, 223e224.