Revista de Biología Tropical ISSN: 0034-7744 [email protected] Universidad de Costa Rica Costa Rica

Castellanos-Galindo, Gustavo Adolfo; Giraldo, Alan; Zapata, Fernando A. Tidepool fish assemblages of Gorgona Island, Colombian Pacific coast: a local and regional comparison Revista de Biología Tropical, vol. 62, núm. 1, febrero, 2014, pp. 373-390 Universidad de Costa Rica San Pedro de Montes de Oca, Costa Rica

Available in: http://www.redalyc.org/articulo.oa?id=44932442028

How to cite Complete issue Scientific Information System More information about this article Network of Scientific Journals from Latin America, the Caribbean, Spain and Portugal Journal's homepage in redalyc.org Non-profit academic project, developed under the open access initiative Tidepool fish assemblages of Gorgona Island, Colombian Pacific coast: a local and regional comparison

Gustavo Adolfo Castellanos-Galindo1,2,3*, Alan Giraldo2 & Fernando A. Zapata4 1. Leibniz Center for Tropical Marine Ecology (ZMT). Fahrenheitstr. 6, 28359 Bremen, Germany; [email protected] 2. Grupo de Investigación en Ecología , Departamento de Biología, Universidad del Valle. A.A. 25360, Cali, Colombia; [email protected]; [email protected] 3. Centre of Excellence in Marine Sciences (CEMarin), Cra. 2 No. 11-68, Rodadero, Santa Marta, Colombia. 4. Grupo de Investigación en Ecología de Arrecifes Coralinos, Departamento de Biología, Universidad del Valle. A.A. 25360, Cali, Colombia; [email protected]

Recibido 18-X-2013. Corregido 20-Xi-2013. Aceptado 19-Xii-2013.

Abstract: The ecology of tidepool fishes has been extensively investigated worldwide over the past 40 years. As in many other ecological fields, studies in temperate zones outnumber studies carried out in tropical areas. Here, we document the short-term spatio-temporal variation of a tidepool fish assemblage from the Tropical Eastern Pacific (Gorgona Island, Colombia), compare our data with previous surveys made in 1993 and provide the first latitudinal comparison of tidepool fish assemblages in the Eastern Pacific coast (43°N to 36°S). During April-July 2006, monthly samplings were carried in 18 tidepools located at three sites this Island. A relatively -rich tidepool fish assemblage (53 species) was found at Gorgona Island. Labrisomidae, Gobiidae and Gobiesocidae were the dominant resident components of this assemblage, whereas Pomacentridae, Muraenidae and Labridae constituted the transient component. Although the same dominant species were observed in sur- veys from 1993 and 2006, suggesting a high persistence of this assemblage over time, clear differences in the relative abundance of some species were also detected. These differences were partly the result of using dif- ferent sampling methodologies (rotenone vs clove oil). Comparisons with other studies performed in tropical and temperate areas of the Eastern Pacific coast indicate a clear latitudinal pattern in the composition of these assemblages that correspond to well-documented biogeographic subdivisions of the marine fauna in this region. Further investigation of the ecological role that tidepool fishes play in rocky intertidal and subtidal areas of Gorgona Island and the whole Tropical Eastern Pacific region will benefit our understanding of the functioning of these important coastal ecosystems. Rev. Biol. Trop. 62 (Suppl. 1): 373-390. Epub 2014 February 01.

Key words: tidepool icthyofauna, community structure, zoogeography, Colombia, Eastern Pacific Ocean.

Rocky shorelines constitute a predomi- A subset of coastal fishes can be specially nant coastal seascape of the world, harbouring adapted to remain in the intertidal zone even at often very diverse biological communities. low tide in tidepools (Gibson, 2003). Such dif- Organisms living in these zones are subject to ference in behaviour highlights the important environmental changes induced by tidal and role that fish can play in transporting energy day-night cycles (Little & Kitching, 1996; within intertidal and between subtidal habitats. Palmer, 2000). To cope with these chang- Tidepools are common features within ing, often stressing, conditions, sessile and intertidal rocky shores, providing the oppor- mobile organisms respond differently. Fishes, tunity to a large number of organisms to for example, can stay in intertidal zones when escape from harsh conditions on emerged favourable conditions are present (i.e. at high rocks (Metaxas & Scheibling, 1993; Mar- tide) returning to subtidal habitats at low tide. tins, Hawkins, Thompson & Jenkins, 2007).

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 373 Tidepools are often considered as nursery This study aims to (1) document the short- grounds for fish and are also used as small iso- term spatio-temporal patterns of tidepool fish lated systems where organisms and their envi- assemblages found at Gorgona Island in the ronmental conditions can be easily manipulated Colombian Pacific Ocean, (2) compare our to advance the ecological understanding of results with data collected in the 1990’s (Gar- these marine systems (Nielsen, 2001). cía & Zapata, unpublished data) at the same Patterns of spatio-temporal variation in localities investigated here; and (3) compare tidepool fish assemblages have been studied the intertidal fish assemblage composition for the last 40 years worldwide (see reviews and organization found at Gorgona Island to in Gibson & Yoshiyama, 1999; Almada & that found at other localities in a wider area Faria, 2004). These studies have been concen- (Colombia and Eastern Pacific Ocean). By trated in temperate areas, while information doing this, a regional synthesis of current from tropical seas is still scarce (Horn, Martin knowledge on Eastern Pacific tidepool fish & Chotkowski, 1999). Several authors (e.g. assemblages is provided. Thomson & Lehner, 1976; Gibson, 1988; 1992; 1993; Gibson & Yoshiyama, 1999) have classi- MATERIALS AND METHODS fied tidepool fishes into three main categories which vary in terminology among studies: (1) Study area: In the Tropical Eastern Pacif- Resident or primary residents, are commonly ic (TEP), the most isolated tropical marine small and cryptic species that spend all their biogeographic region of the world (Robert- life cycle within tidepools; (2) Transient or son, Grove & McCosker, 2004), rocky shore secondary residents, are those that spend only seascapes are predominant besides two large parts of their life cycle within tidepools and sandy and muddy coastline portions along then migrate to subtidal habitats as adults; and the south eastern Gulf of California (350km) (3) Accidental or casual visitors are typical and the expanse from southern Mexico to El subtidal species that can enter intertidal zones Salvador (1000km; Mora & Robertson, 2005; to forage and occasionally get trapped in tide- Robertson & Cramer, 2009). Other near-shore pools at ebb tide. Although this classification is ecosystems such as coral reefs are uncommon, useful to distinguish major components of these occupying <25km2 (Glynn & Ault, 2000). This assemblages, the lack of homogeneity in the holds true for the Colombian Pacific coast, criteria used to distinguish between categories with ca. 636 km of its shoreline dominated by prevents detailed comparisons among studies. rocky shores of different origin (López-Victo- Although the Tropical Eastern Pacific ria et al., 2004). Rocky shores of volcanic ori- regional fauna is not as diverse as the one in gin are located in the northern part of the coast the Indo-West Pacific region, it presents a and in Gorgona and Malpelo Islands, whereas high level of endemism due to its isolation tertiary sedimentary rocky shores are located in from other biogeographical regions (Glynn & the central (Buenaventura and Bahía Málaga) Ault, 2000; Robertson & Allen, 2002). The and southern parts (Istmo de Pichidó and Gallo near-shore fish fauna in the region is relatively Island) of the Colombian coast. well known mainly on coral reefs. In recent Gorgona is a continental island located years, the great importance that rocky reefs in the southern part of the Colombian Pacific may have for reef fish communities within the coast (2º58’10” N - 78º11’05” W). Marine region has been pointed out (Robertson, 1998, habitats on the island include sandy beaches, Dominici-Arosemena & Wolff, 2006; Benfield, coral reefs and rocky shores. In 2006, we Baxter, Guzman & Mair, 2008). In this context, sampled three different rocky intertidal areas tidepool fishes on rocky shores of the region on the Island: Ventana, Barra and Camaronera. represent a subset of reef fish communities that The first two sites were located in the south has not been properly assessed. and southeast portions of the Island and are

374 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 separated by a <50m-long sandy beach. These random measures to the nearest 0.5 cm in dif- two sites presented ca. 200m of a exposed ferent sections of the pools. With these data, rocky platform at low tide The third site, an estimation of fish density (fish m-2 and Camaronera, is located on the southwestern fish m-3) was subsequently obtained. side of the Island, separated from the other Fish collections in 1993 were made using two sites by a 1km-long sandy beach. The the ichthyocide rotenone (García & Zapata, topographic characteristics of this site vary unpublished data), which is regarded as a very considerable compared to those of Ventana and effective sampling technique (Ackerman & Barra. Camaronera is characterized by gravel Bellwood, 2002; Robertson & Smith-Vaniz, and sandy substrates. Subtidal habitats near 2008). Collections were made monthly from the three sites include mixed sandy and rocky January to July (except February) at the same bottoms surrounded by small pocilloporid coral sites where sampling was carried out in 2006 patches (Zapata, 2001a). (i.e. Ventana, Barra and Camaronera). An average of seven tidepools per site for each Sampling: Six tidepools located between month were sampled with rotenone during low the mid and low intertidal zone at each of the tide. All fish were removed from the selected three sites described above were randomly tidepools and subsequently preserved in 10% selected during a preliminary field trip in formaldehyde. In the laboratory fishes were March 2006. Repeated samplings of these transferred to 70% ethanol solution and identi- pools were made during four months (April- fied using appropriate taxonomic keys for the July 2006) using a clove-oil solution (eugenol). area (e.g., Allen & Robertson, 1994). This anaesthetic is used commonly for collect- The classification of species according to ing coral reef and tidepool fishes (Griffiths, its occurrence in time was based on Gibson 2000; Ackerman & Bellwood, 2002; Robertson & Yoshiyama (1999). Resident species were & Smith-Vaniz, 2008). The solution was pre- those with an exclusive presence in tidepools. pared in a 1:8 clove oil to ethanol proportion Transient species were those present only as and sprayed directly into the tidepools, par- juveniles in tidepools and also seen commonly ticularly in small crevices and under boulders in subtidal areas either as juveniles or adults. where fish may hide. The quantity of anaes- Accidental visitor species were those inhab- thetic employed was variable depending on iting subtidal habitats that occasionally got tidepool size. The reaction on fish species was trapped in tidepools. also variable with the most vulnerable species showing erratic movements after ca. 5 minutes Data Analyses: Data taken in 2006 were of clove oil application. Using small dip nets all used to test for differences in tidepool fish com- fish within tidepools were captured searching position among sampling sites and months. For under rocks and crevices when necessary. All this purpose a 2-way permutational analysis fishes were then identified to species with field of variance (PERMANOVA, Anderson, 2001) taxonomic guides adapted from Robertson & was performed. The analysis was conducted Allen (2002), counted and afterwards returned treating Site (three levels) as a random factor to adjacent tidepools where they rapidly recov- and Month (four levels) as a fixed factor. Sam- ered from the anaesthetic effect. ples among months were assumed to be inde- Prior to fish collections, length, width and pendent due to the demonstrated rapid recovery depth of tidepools were measured to calcu- of tidepool fish assemblages after perturbation late sampled areas and volumes. Length was (Willis & Roberts, 1996). The PERMDIST considered as the distance between the most routine (an analog of the univariate Levene’s distant edges of a pool and the width as the test) was used to check if results from the PER- distance perpendicular to the midpoint length MANOVA test were an artifact of difference in axis. Depth was calculated by averaging 10-15 dispersion among groups (Anderson, 2006). To

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 375 explore the multivariate patterns shown in the Classification analyses were then performed PERMANOVA test, constrained and uncon- using the Bray-Curtis similarity measure on strained ordinations were used (non-metric presence/absence data (Sørensen similarity multidimensional scaling or nMDS and canoni- measure). PERMANOVA+for PRIMER soft- cal analysis of principal coordinates or CAP). ware was used to conduct all analyses (Ander- Species responsible for differences along the son et al., 2008). CAP axes were determined as those species’ abundances having Spearman correlations >0.4 RESULTS with any of the canonical discriminant axes (Anderson & Willis, 2003). All analyses were Small-scale spatio-temporal patterns. performed on square-root transformed density The tidepool ichthyofauna of Gorgona Island -2 data (fish m ) based on the Bray-Curtis simi- during April-July 2006 comprised 38 species larity measure. grouped in 18 families. Three taxa, Malacoc- Due to the different sampling procedures tenus sudensis (Labrisomidae), Bathygobius employed in 1993 (rotenone, García & Zapata, ramosus (Gobiidae) and Stegastes acapulcoen- unpublished data) and 2006 (clove oil), data sis (Pomacentridae), accounted for 67% of the from both periods could not be directly com- total abundance. The most speciose families pared. However, in an attempt to evaluate: were Muraenidae (six species) and Pomacen- (1) the possible effect of the sampling method tridae (five species). The abundance of Labri- and, (2) the persistence of these assemblages somidae, Gobiidae, Pomacentridae, Mugilidae over time, we used an unconstrained ordina- and Gobiesocidae accounted for >85% of the tion technique (Principal coordinates analysis- total abundance (Table 1). Twenty one percent PCO) on species relative abundance data for of the species were classified as residents, 47% the three sites that were sampled in 1993 and as transients, and 32% as accidental visitors. 2006. Data were left untransformed and the The relative abundance of these same groups, Bray-Curtis similarity measure was employed however, was greater for residents (65%) than to generate the similarity matrix upon which for transient species (34%) with accidental the PCO was performed. PCO is very flex- visitors representing a small fraction of the ible as it is the non-metric multi-dimensional assemblage (1%). scaling (nMDS) ordination technique, because The PERMAVOVA test showed significant it can be used on any resemblance matrix. differences in fish assemblages among sites and However, PCO differs from nMDS in that PCO among sampling dates (Table 2). The variation projects points in space onto axes according in multivariate dispersion, however, was sig- to the chosen resemblance measure, whereas nificant for the site factor (PERMDISP, F=7.11, nMDS is focused in preserving the rank order p=0.022), indicating that caution should be of similarities in a defined number of dimen- taken when interpreting the PERMANOVA sions (Anderson, Gorley & Clarke, 2008). results. Non-metric multi-dimensional scale Tidepool fish assemblages along the East- (nMDS) and CAP ordinations showed differ- ern Pacific Ocean were compared based on 17 ent results with no clear aggregation patterns sampling locations spanning the West coast of among samples in the nMDS (high stress value USA at 43°N to the southern part of Chile at =0.23) and a relatively well-defined grouping 36°S. The data were obtained from 10 studies of samples along the CAP axis 1 in the CAP and one unpublished data set from Colom- ordination (Fig. 1). In the CAP ordination, sam- bia (Castellanos-Galindo, periods). All these ples from Camaronera were generally located studies had as a main aim the sampling of to the left of the CAP axis 1, whereas samples the entire tidepool ichthyofauna at each loca- from Ventana and Barra were distributed to the tion. Data from these studies were organized right (Fig. 1). As expected, the leave-one-out and analyzed at the species and family level. allocation success from the CAP analysis was

376 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 ------✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 93 - - - 8.33 8.33 4.17 4.17 4.17 4.17 95.83 95.83 50.00 33.33 20.83 16.67 54.17 29.17 16.67 12.50 16.67 29.17 20.83 33.33 12.50 12.50 12.50 F (%) Barra - - - Density 2.38±0.35 1.47±0.27 0.49±0.15 0.38±0.26 0.07±0.04 0.03±0.01 0.13±0.04 0.05±0.02 0.06±0.04 0.03±0.02 0.03±0.01 0.07±0.03 0.01±0.01 0.02±0.01 0.15±0.08 0.05±0.02 0.03±0.02 0.02±0.01 0.04±0.03 0.01±0.01 0.01±0.01 0.01±0.01 0.01±0.01 - 100 9.09 9.09 4.55 4.55 4.55 9.09 4.55 77.27 54.55 54.55 54.55 36.36 59.09 27.27 22.73 27.27 18.18 22.73 13.64 13.64 22.73 18.18 27.27 13.64 F (%) Ventana - Density 0.11±0.07 2.90±0.64 0.82±0.26 0.52±0.22 0.60±0.22 0.65±0.17 0.28±0.10 0.21±0.08 0.22±0.12 0.12±0.06 0.12±0.05 0.03±0.01 0.14±0.07 0.05±0.03 0.02±0.01 0.02±0.01 0.02±0.02 0.08±0.04 0.03±0.01 0.10±0.10 0.03±0.02 0.02±0.02 0.06±0.06 0.01±0.01 0.02±0.02 - - - - - 100 5.26 5.26 5.26 94.74 84.21 26.32 78.95 57.89 47.37 26.32 21.05 52.63 57.89 10.53 47.37 21.05 26.32 26.32 10.53 15.79 F (%) - - - - - Camaronera Density 0.11±0.03 0.11±0.05 2.27±0.43 1.99±0.22 0.92±0.32 0.13±0.13 0.08±0.03 0.71±0.19 0.13±0.04 0.17±0.06 0.08±0.03 0.05±0.03 0.25±0.08 0.17±0.05 0.03±0.02 0.09±0.05 0.05±0.03 0.03±0.03 0.05±0.04 0.01±0.01 0.01±0.01 TABLE 1 TABLE T T T T T T T T T T T T T T T T R R R R R R RS AV AV AV AV Common name Glossy blenny Panamic frillfin Panamic Acapulco major Snouted mullet Night sergeant Panamic clingfish Panamic Starry grouper Tailspot cardinalfish Tailspot Banded wrasse Mexican blenny Myers’ clingfish Myers’ moray Needle-tooth Panamic sergeant major sergeant Panamic Tinsel squirrelfish Tinsel Cortez rainbow wrasse Cortez Pacific spotted scorpionfish Pacific Large-banded blenny Large-banded Freckled moray Undulated moray soapfish Blackfin Rock clingfish moray Panamic mullet Flathead White mullet White Panama triplefin Panama Sandtop goby to decreasing total abundance (2006 sampling). (-) refers to species absent at that site or sampling year site or sampling (-) refers to species absent at that (2006 sampling). abundance total to decreasing (Blenniidae) (Pomacentridae) Mean density (±standard error) and frequency of occurrence (F%) of tidepool fishes collected in 2006 at three sites in Gorgona Island, sites in Gorgona in 2006 at three fishes collected (F%) of tidepool Mean density (±standard error) and frequency of occurrence (Gobiidae) (Pomacentridae) Species ) in the last column of the table (93). Species are sorted according of the table ( ✓ ) in the last column are indicated unpublished data) sites (García & Zapata, at the same three (Gobiidae) (Tripterygiidae) Colombia, Tropical Eastern Pacific. Residential status (RS): residents (R), transients (T) and accidental visitors (AV). Species present in collections made in 1993 made Species present in collections visitors (AV). status (RS): residents (R), transients (T) and accidental Residential Eastern Pacific. Tropical Colombia, Malacoctenus sudensis (Labrisomidae) Malacoctenus Bathygobius ramosus Bathygobius Stegastes acapulcoensis (Mugilidae) Chaenomugil proboscideus Abudefduf concolor (Pomacentridae) Abudefduf Gobiesox adustus (Gobiesocidae) Epinephelus labriformis (Serranidae) Epinephelus Apogon dovii (Apogonidae) Halichoeres notospilus (Labridae) Halichoeres Paraclinus mexicanum (Labrisomidae) Tomicodon petersi (Gobiesocidae) Tomicodon (Muraenidae) macrocephalus Uropterygius Abudefduf troschelii Abudefduf Sargocentron suborbitalis (Holocentridae) Sargocentron Thalassoma lucasanum (Labridae) Scorpaena mystes (Scorpaenidae) Ophioblennius steindachneri Echidna nocturna (Muraenidae) Echidna Gymnothorax undulatus (Muraenidae) nigripinnis (Serranidae) Rypticus (Gobiesocidae) rhodospilus Arcos Gymnothorax panamensis (Muraenidae) Mugil cephalus (Muraenidae) (Mugilidae) Mugil curema Axoclinus lucillae Axoclinus Gobulus hancocki

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 377 - - - - - ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 93 ------8.33 4.17 4.17 4.17 4.17 4.17 F (%) 29 1021 Barra 5.61±1.51 ------Density 0.02±0.01 0.01±0.01 0.01±0.01 0.01±0.01 0.002±0.002 0.005±0.005 ------4.55 4.55 4.55 4.55 4.55 F (%) 30 968 Ventana 4.74±0.85 ------Density 0.01±0.01 0.07±0.07 0.02±0.02 0.003±0.003 0.003±0.003 ------5.26 5.26 10.53 10.53 F (%) ) 25 770 7.53±0.81 ------Camaronera Density 0.01±0.01 0.02±0.01 0.03±0.02 0.03±0.03 T T R R RS AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV AV TABLE 1 ( Continued TABLE Common name Mottled soapfish Mottled Snowflake moray Frogfish Jewel moray Beaubrummel Labrisomind blenny Labrisomind Flagtail wormfish Flagtail Threebanded butterflyfish Threebanded Sand stargazer Estuarine frillfin Estuarine snake eel Tiger Viviparous brotula Viviparous Toadfish Dow’s toadfish Dow’s Roughjaw frogfish Blackstripe clingfish Blackstripe Cortez angelfish Cortez Island major Thickscale silverside Thickscale Barred flagtail Snapper Spottail grunt Spottail Scissortail damselfish Scissortail Coral hawkfish Wormfish Bigmouth sanddab Bigmouth Spotted sharpnosed puffer ) -2 () (Microdesmidae) (Atherinopsidae) (Paralichthyidae) (Muraenidae) Species (Pomacentridae) (Pomacentridae) sp. (Batrachoididae) Rypticus bicolor (Serranidae) Rypticus Echidna nebulosa (Muraenidae) Echidna Antennarius sp. (Antennariidae) Muraena lentiginosa Stegastes flavilatus Labrisomidae sp (Labrisomidae) Labrisomidae Clarkichthys bilineatus Clarkichthys Chaetodon humeralis (Chaetodontidae) Dactyloscopidae sp. (Dactyloscopidae) Dactyloscopidae (Gobiidae) andrei Bathygobius maculosus (Ophichthidae) Myrichthys Ogilbia sp. (Bythitidae) Porichthys Daector dowi (Batrachoididae) Fowlerichthys avalonis (Antennariidae) Fowlerichthys Tomicodon myersi (Gobiesocidae) Tomicodon Pomacanthus zonipectus (Pomacanthidae) (Pomacentridae) Stegastes cf. arcifrons Atherinella pachylepis Atherinella Kuhlia mugil (Kuhliidae) Lutjanus sp. (Lutjanidae) Haemulon maculicauda (Haemulidae) Chromis atrilobata Chromis oxycephalus Cirrhitichthys sp. (Microdesmidae) Cerdale Citharichthys gilberti Citharichthys Canthigaster punctatissima (Tetraodontidae) Number of species number of individual Total Overall fish density (Ind m

378 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 Stress: 0.23 higher at Camaronera and lower at Ventana, implying that assemblages from Ventana were harder to predict than the ones from Camaro- nera or Barra (Table 3). This is also evident in the PERMDISP results, in which the largest dispersion of assemblages was observed in Ventana samples (average Bray-Curtis dis- tance-to-centroid =41%). Ten species were found to be strongly cor- related with CAP axis 1, thus differentiating assemblages from the three sites. Abudefduf concolor (r=0.41) and Chaenomugil probosci- deus (r=0.52) were species strongly correlated with CAP axis 1 and assemblages at Ventana. -2 0.2 The mean density (ind m ±SD) of these spe- cies at Ventana was the highest amongst sites (A. concolor =0.65±0.17 and C. proboscideus =0.60±0.22). Similarly, species with strong 0.1 Gob. adu correlation with CAP axis 1 (Camaronera) Abu.con Cha.prob

Hal.not Sar.sub were Gobiesox adustus (r=-0.63), Bathygobi- Apo.dov us ramosus (r=-0.68), Tomicodon petersi (r=- Tom.petSte.aca = 0.37 2 0 Uro.mac 0.55), Uropterygius macrocephalus (r=-0.43) Bat.ram and Sargocentron suborbitalis. (r=-0.27). All CAP2 δ these species presented the highest densities at -0.1 Camaronera (Table 1). In 1993, 9 395 fish were collected during the six-month study period. Forty-one species -0.2 grouped in 26 families comprised the tide- -0.2 -0.1 0 0.1 0.2 pool fish assemblage. Five species accounted CAP1δ2 = 0.41 for ca. 65% of the total fish abundance (M. Camaronera Ventana Barra sudensis-26.7%, S. acapulcoensis-13.5%, A. concolor-9.7, B. ramosus-8.2% and Abudefduf Fig. 1. nMDS and CAP ordinations of tidepool fish troschelii-7.6%). This indicates a higher even- assemblages data (ind m-2) taken in 2006 at three sites in Gorgona Island, Colombia, Tropical Eastern Pacific. ness in the assemblage from 1993 compared Abbreviations in CAP plot correspond to the first three to the one sampled in 2006, when the same letters of generic and species names (see Table 1). relative abundance was reached by only three

TABLE 2 Results of a two-way mixed model PERMANOVA testing for the effects of site (random factor) and month (fixed factor) on Gorgona Island’s tidepool fish assemblages surveyed during April-July, 2006

Source df SS MS F p Site 2 12656 6328 3.8539 0.001 Month 3 5165.3 1721.8 2.5223 0.016 Site x Month 6 3936.9 656.14 0.39961 1 Residual 53 87024 1642 Total 64 109620

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 379 TABLE 3 Results of canonical analysis of principal coordinates (CAP) testing the effect of Site (three levels) on Gorgona Island tidepool fish assemblages

Allocation success (%) Data m %Var d2 P Camaronera Barra Ventana Total Site 5 72.43 78.94 75.00 59.09 70.77 0.412 0.0002

%Var=percentage of the total variation explained by the first m principal coordinate axes; Allocation success=percentage of points correctly allocated into each group; d2=square canonical correlation. species. Despite this difference, four of the five 40 most important species during both sampling 1993 periods were the same. The dominant families V 2006 Mal.sud Thal.luc in 1993, in decreasing order, were Labrisomi- B B B V 20 V Par.mex V Hal.not dae, Pomacentridae, Gobiidae, Labridae and Ath.pacV Apo.dov C Ste.aca Mugilidae. These five families accounted for B Uro.mac C Arc.rho C 80% of the total abundance. Finally, the rela- Tom.pet B Por.sp. 0 Ryp.nig C C V tive abundance of the residential categories in V C C C Bat.ram V Sar.sub 1993 was different compared to sampling in C B V Abu.tros Abu.con V 2006. Relative abundance of transient spe- B B Cha.pro -20 cies reached ca. 60%, whereas residents did B C PCO2 (21% of total variation) (21% of total PCO2 not reach 40% of the total abundance. This B was due to the considerably fewer number of individuals captured of the resident spe- -40 cies B. ramosus in 1993. In contrast, three -40 -20 0 20 40 pomacentrids (S. acapulcoensis, A. concolor PCO1 (33.3% of total variation) and A. troschelii) and one labrid (Halichoeres Fig. 2. PCO ordination comparing tidepool fish assemblages notospilus), all transient species, were consis- from Gorgona Island in 1993 and 2006. Abbreviations tently more abundant during sampling in 1993 correspond to the first three letters of generic and species compared to 2006. names (see Table 1). Capital letters indicate sampling Twenty-seven species were shared between localities: (V) Ventana, (C) Camaronera and (B) Barra. both sampling periods. Fifteen species were found only in 1993 when sampling with rote- nigripinnis (r=0.72), H. notospilus (r=0.60), none, whereas 11 species were found only in Porichtys sp. (r=0.55), A. troschelii (r=0.52), S. 2006 with clove oil sampling (Table 1). The suborbitalis (r=0.51) and Apogon dovii (r=0.49, PCO ordination showed separation between all correlated with samples from 1993). both sampling periods (1993 and 2006), with most samples collected in 2006 distributed to Zoogeographical analysis: Based on all the left axis of the PCO axis 1 (Fig. 2). Simi- the available data, the tidepool ichthyofauna larly, most samples collected in 1993, indepen- in the Eastern Pacific Ocean (43°N-36°S) dent of site, were grouped to the right of the contained at least 214 species grouped in PCO axis 1. The species responsible for differ- 52 families. represented 70% of ences among sampling times (correlations >0.4 the families and 63% of species present in with PCO axis 1) were B. ramosus (r=-0.88), T. this region. Scorpaeniformes and Anguilli- petersi (r=-0.67), Arcos rodosphilus (r=-0.50), formes were the next two most species-rich M. sudensis (r=-0.49) and U. macrocephaulus fish orders (13% and 6%, respectively). With- (r=-0.44, all correlated with samples from in the Perciformes, the suborder Blennioidei 2006); and S. acapulcoensis (r=0.71), Rypticus had the highest species representation (44),

380 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 with some of these species being especially This group includes studies carried out in the abundant and dominant in assemblages from southernmost part of the west coast of the Gulf the Panamic province (sensu, Robertson & of California (23°-26°N; Ruiz-Campos et al. Cramer, 2009) and also in assemblages of (2010)), a study carried out inside the Gulf Chile. The suborder Labroidei (28 species) of California (Cortez Province; Thomson & was also especially dominant in the tropical Lehner (1976)) and studies carried out in the portion of the region with some of the species Panamic province (Costa Rica and Colombia; contributing importantly to the transient resi- Weaver (1970), Castellanos-Galindo, Giraldo dential category. Pomacentridae and Labridae & Rubio (2005) and this study). A second were the most important labroid families in group was defined by assemblages of the the tropics, whereas Embiotocidae was found northern part of the Eastern Pacific coast (35°- from 26° to 38°N. In the northern portion of 43°N; Yoshiyama, Sassaman & Lea (1986)), the Eastern Pacific (>26°N), Scorpaeniformes which are part of the California and Oregon became the dominant group, being Cottidae a provinces (Fig. 3). Finally, the tidepool fish very diverse family in tidepools of the area. In assemblage from the warm-temperate Peru- terms of abundance, assemblages in this area Chilean Province is clearly distinguished from were almost exclusively dominated by Cottidae the rest (Fig. 3). with some species of Stichaeidae or Clinidae having a relative importance in some areas. In the southern part of the region (<20°S), DISCUSSION apart from the Blennioidei, a Percoidei family (Kyphosidae) became very important in terms Species richness, composition and resi- of abundance within these assemblages. At dential status: The tidepool fish assemblage the southernmost part of this area (36°S), the found at Gorgona Island was relatively diverse suborder Gobiesocoidei was very abundant compared to assemblages at other localities in together with the Blennioidei (Table 4). the Colombian Pacific coast or the rest of the When comparing tidepool fish assemblag- Tropical Eastern Pacific Ocean. These results es along the latitudinal gradient three consistent combined with information from other locali- groups were observed. Assemblages with tropi- ties in the Colombian Pacific indicate a domi- cal affinities were grouped coinciding with the nance of Perciform families represented mainly Tropical Eastern Pacific region boundaries. by the sub-orders Labroidei, Blennioidei and

TABLE 4 Principal families and number of species found in different tidepool fish assemblages of the Eastern Pacific region. Total number of species is given in parenthesis.

Family/Latitude 43°N 42°N 40°N 38°N 35°N 30-32°N 26-29°N 23-26°N 31°N 10°N 9°N 5-6°N 3°N 2°N 20°S 32-33°S 36°S Labrisomidae (20) - - - - - 2 2 2 4 3 6 4 1 3 2 3 3 Cottidae (15) 9 7 8 15 6 4 1 ------Pomacentridae (13) - - - - - 1 - 4 2 4 5 4 3 6 2 - - Gobiesocidae (16) 1 1 - 1 1 2 - 3 2 2 3 3 1 4 2 2 2 Muraenidae (11) ------1 1 - 3 4 7 1 6 - - - Gobiidae (10) ------1 4 3 3 2 1 3 - 1 - Blenniidae (9) - - - - - 2 4 3 2 1 2 3 - 1 2 2 2 Labridae (8) ------1 2 3 4 2 2 1 2 - - - Haemulidae (8) ------5 - 1 - - 1 1 - - Scorpaenidae (7) - - - 4 - - - 1 1 1 - 1 - 1 - - - Clinidae (7) - - - 2 1 3 3 ------1 2 Kyphosidae (6) - - - - 1 2 2 1 3 - - - - - 2 2 0

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 381 Tamarindo, Costa Rica - 9° N (1)

Playas del Coco, Costa Rica - 10° N (1)

Gorgona Island, Colombia - 2° 58’ N (2)

Utría - Cabo Marzo, Colombia - 5° - 6° N (3)

Southern Western Baja California - 23°- 26° N (4)

Isla Palma, Colombia - 3° 54’ N (5)

Northern Gulf of California, USA - 31° N (6)

Brookings, USA - 42° N (7)

Cape Arago, USA - 43° N (7)

Cape Mendocino, USA - 40° N (7)

Point Arena, USA - 38° N (7)

Piedras Blancas, USA - 35° N (7)

Northern Western Baja California - 30°-32° N (4)

Central Western Baja California - 26°-29° N (4)

Southern coast of Chile - 36° S (8)

Central coast of Chile - 32°-33° S (9)

Northern coast of Chile - 20° S (10)

0 20 40 60 80 100 Similarity (%)

Fig. 3. Cluster dendrograms indicating tidepool fish assemblage affinities along the Eastern Pacific coast from 43°N to 36°S based on species presence/absence data. Localities and sources of information given in parenthesis are as follows: (1) Tamarindo and Playas del Coco, Costa Rica, Weaver (1970); (2) Gorgona Island, Colombia; this study; (3) Northern Colombian Pacific coast, Castellanos-Galindo (unpublished data); (4) Western coast of Baja California, Ruiz-Campos et al. (2010); (5) Bahía Málaga, Colombia, Castellanos-Galindo et al. (2005); (6) Northern Gulf of California, Thomson & Lehner (1976); (7) Western coast of USA, Yoshiyama et al. (1986); (8) Central-South coast of Chile, Quijada & Caceres (2000); (9) central coast of Chile, Muñoz & Ojeda (1997); and (10) Northern coast of Chile, Berrios & Vargas (2000). Bars indicate biogeographical regions and provinces as defined by Briggs & Bowen (2012): (black) Tropical Eastern Pacific region (includes Cortez and Panamian provinces); (white) Eastern North Pacific region (Oregon province); (diagonal lines) Warm-temperate California province and (horizontal lines) Warm-temperate Peru-Chilean Province.

382 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 Gobioidei in this part of the TEP. At Gorgona (estuarine, coral reefs). Playas del Coco, adja- Island, a considerably high number of species cent to coral reefs, had a higher species rich- was found, with M. sudensis (Labrisomidae), S. ness (28) than the other localities. acapulcoensis (Pomacentridae) and B. ramosus When comparing the composition of resi- (Gobiidae) being the most abundant species. dents there is concordance at the family level Resident species at Gorgona belonged to the between studies in Costa Rica and Colombia; same families found at another locality of the however, Gorgona presents a higher number of Colombian Pacific with an estuarine environ- species within each of these families. Differ- ment (Bahía Málaga; Castellanos-Galindo et ences can also be observed in the dominance al., 2005), but the total number of species and (relative abundance) of fish species. Whereas the number of species per family was higher in Bahía Málaga, Colombia, one single spe- in Gorgona than at Bahía Málaga (14 species). cies (Bathygobius ramosus) comprises >50% Additional examination of tidepool fish assem- of the total tidepool fish assemblage, approxi- blages at two other localities in the northern mately this same percentage is shared by the Colombian Pacific (Ensenada de Utría and three principal species in Gorgona Island (M. Cabo Marzo), where rocky shores are the pre- sudensis, S. acapulcoensis, B. ramosus). None- dominant seascape, show that these assemblag- theless, when comparing the five most abun- es are similarly diverse to those encountered dant species at each locality, the same three at Gorgona Island (ca. 40 species; Castella- species dominate. nos-Galindo, unpublished data). Labrisomi- In terms of relative abundance, the residen- dae, Gobiidae, Pomacentridae, Gobiesocidae, tial categories differed between Gorgona Island Muraenidae and Labridae dominate the assem- and Bahía Málaga, even though the number of blage composition at these three localities. resident species in the latter locality is higher. The great differences in species richness Griffiths (2003) found a very similar pattern in between Gorgona Island and Bahía Málaga temperate Australia as in Bahía Málaga, where can be explained by a combination of factors, permanent resident species dominate over which include the different sampling effort transients. Conversely, Thomson and Lehner (time, number of pools and number of sites) and differences in the topographic character- (1976) found a pattern in the Gulf of California istics of the tidepools sampled. However, the more similar to the one observed at Gorgona most important difference may be the pres- Island, where the percentage of transient spe- ence of a richer subtidal fish community in the cies was higher. Mahon & Mahon (1994) in the marine environment of Gorgona Island com- Caribbean Sea, documented a dominance of pared to the poorer estuarine subtidal fish com- true residents over partial ones (transients spe- munity in Bahía Málaga. The presence of larger cies sensu Gibson & Yoshiyama (1999)); how- areas of rocky and coral reefs in Gorgona in ever, many juvenile Anguilliform species were comparison to Bahía Málaga, may also explain classified as true residents by these authors, the higher number of transient and acciden- whereas Gibson & Yoshiyama (1999) con- tal visitors found in tidepools of Gorgona. A sider them as transient species. If anguilliforms comparison with results of visual censuses of are pooled with the rest of partial residents, coral reef fishes at Gorgona indicates that at both categories (residents and transients) reach least 20 coral reef fish species are commonly almost the same percentages. Therefore, it is found in tidepools (Zapata & Morales, 1997; reasonable to expect that tidepool fish assem- Zapata, 2001b). A very similar result was found blages will be more likely to contain a higher by Weaver (1970) in the Pacific coast of Costa number of transient species when adjacent Rica where the assemblage structure of the subtidal ichthyofaunas are diverse (i.e., near tidepool ichthyofauna differed between three coral reefs). Conversely, when tidepools occur localities with different subtidal characteristics near species-poor subtidal habitats, resident

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 383 species will be more likely to dominate tide- differences between sampling months, these pool fish assemblages. were driven by the high variance associated to The use of intertidal habitats by fishes at each of the samples. The leave-one-out alloca- high tide has rarely been examined in tropical tion success procedure for the Month factor areas. However, Castellanos-Galindo, Krumme was low (40%). This indicates difficulties in & Willis (2010) examined this aspect in rocky differentiating samples coming from a specific shores of the northern Colombian Pacific, and month mainly because of large multivariate found that 73 fish species used these intertidal dispersion within months. These results are habitats at high tide. Common species can be in contradiction with most of the studies car- identified between those observed at high ried out in temperate regions, but not with the tide and those remaining in tidepools at low studies carried out in the tropics. The absence tide, especially for members of the families of strong seasonal changes in the Colom- Pomacentridae and Labridae. However, a large bian Pacific compared to temperate areas may proportion of cryptic resident species found in explain these contrasting results. The relative tidepools, were not recorded by Castellanos- stability of environmental conditions in these Galindo et al. (2010), due to the methodology coastal biotopes may allow these assemblages employed in that study (i.e., underwater visual to maintain their structure and function through censuses). Combining both sources of informa- short temporal scales. It will be necessary to tion (tidepool sampling and intertidal visual examine if variation in community structure censuses at high tide) it could be estimated occurs during the rainy season in Gorgona that ca. 100 fish species use intertidal habitats Island (May-December) to extend our conclu- in Colombian Pacific rocky shores. This high- sions to larger temporal scales. lights the importance that these habitats have Finally, studies at greater temporal scales for shore fishes and also suggest the existence (years or decades), have taken place mainly in of important links between subtidal and inter- localities of the West coast of the USA. Simi- tidal areas that are regulated by tides and the larities in assemblage structure are very consis- availability of intertidal prey sources for fishes. tent there over the sampling periods indicating that these assemblages are persistent over large Patterns of temporal variation: Tempo- temporal scales (Grossman, 1982; Yoshiyama, ral variation in tidepool fish assemblages has 1981; Yoshiyama et al., 1986, see review in been evaluated at several scales from day/night Almada & Faria, 2004). The results obtained to periods of 10 years. In a month or year- from a comparison of samplings in 1993 and round scale, variation in community structure 2006 in Gorgona Island support the results has been detected in the majority of studies obtained in temperate areas. The dominant spe- (review in Gibson & Yoshiyama, 1999, but see cies were the same between sampling periods. also Davis, 2000a; Griffiths, 2003; Arakaki & The number of shared species (27 out of 53) Tokeshi, 2006). However, other surveys, which between both periods can reflect that some include some in tropical regions, have found degree of unpredictability in the composition that such variability is not present (temperate: of transient and accidental visitors exist, but Prochazka, 1996; tropical: Chang, Lee, Lee & that the core members of the fish assemblage Chen, 1973; Mok & Wen, 1985; Castellanos- are persistent over time. The variation in the Galindo et al., 2005). Temporal variations often proportions found is very likely due to the have been attributed to the strong changing sea- use of different sampling techniques. In 1993, sonal conditions observed year-round in tem- rotenone was used. This substance is highly perate latitudes, where most of these studies effective, but has been shown to have differing have been carried out. No short-term variation effects on fish species. Weaver (1970) men- was observed during the four months of this tioned that B. ramosus (Gobiidae) was highly study. Although the statistical tests suggested resistant to rotenone when poisoning tidepools

384 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 of Pacific Costa Rica. This may be the reason considerably higher at Camaronera than at the why the relative abundance of this species other two sites. At Camaronera, shallower tide- changed between 1993 and 2006 sampling at pools composed of gravel and sand might offer Gorgona Island (8% and 22%, respectively). more microhabitats to (benthic) fishes than Likewise, Weaver (1970) identified species of tidepools at Ventana and Barra where tidepools Pomacentridae as highly susceptible to rote- had relatively low microhabitat complexity. In none. The relative abundance of the species of contrast, schooling species and those inhabit- this family was considerably higher in 1993 ing the water column (e.g. C. proboscideus, A. than in 2006 at Gorgona Island. troschelii) were more characteristic of Ventana Another important conclusion derived than of Camaronera. Further studies should from the low variability in the assemblage directly test the effect of complexity on fish structure over time is that this assemblage is assemblage composition at these sites. able to recover in a short period (<1 month in this case), after severe perturbations such Latitudinal patterns in the Eastern as those caused by sampling (i.e. defauna- Pacific. The taxonomic composition of tide- tion). It has been suggested previously that pool fish assemblages found in the Eastern re-colonization may occur from individuals of Pacific is in agreement with results of previous adjacent pools migrating to defaunated pools analyses by Chotkowski, Buth & Prochazka within few days (Willis & Roberts, 1996). (1999) and Prochazka, Chotkowski & Buth. Finally, phenomena occurring at large temporal (1999). The number of tidepool fish species scales, such as the El Niño-Southern Oscilla- found in the Eastern Pacific corresponds to ca. tion (ENSO), may affect community structure 30% the total number of species documented (Davis, 2000a). This is an aspect that has been by Chotkowski et al. (1999). This percentage is poorly examined in tidepool fish assemblages. likely to be an over-estimation considering that The TEP offers great potential to improve our there is not an appropriate documentation of understanding of the effects of ENSO on inter- tidepool fish faunas in several tropical diverse tidal rocky shore communities because this areas of the Indo-West Pacific region. region is greatly affected by this phenomenon. The relatively small area that is available for colonization (especially in rocky shore Patterns of spatial variation: Differences areas) is the main restriction that fishes that in habitat complexity often lead to changes reside in intertidal areas may face. This con- in assemblage structure (Mahon & Mahon, straint is probably responsible for the lim- 1994; Griffiths, Davis & West, 2006; Macieira ited range of body plans, sizes and taxonomic & Joyeux, 2011). This may be the case for groups that are observed in tidepool fishes the inter-site differences observed at Gorgona (Chotkowski et al., 1999). The Blennioidei and Island. Griffiths et al. (2006) showed that the Labroidei suborders accounted for 21% and availability of rocks and macro-algae sig- 12% of the total number of 702 species found nificantly affected the species composition in the revision made by Chotkowski et al. and abundance of fishes visiting tidepools (1999). In the present comparison of tidepool in Australia. Davis (2000b) reached a simi- fishes in the Tropical Eastern Pacific, these lar conclusion after she manipulated tidepool same suborders accounted for 29% and 19% complexity in California. At Gorgona Island, of the total number of tidepool fish species the different topographic characteristics of the (Blennioidei and Labroidei). The Neotropics sites where tidepools were sampled may be are a major center of diversification for mem- partially responsible for the differences in bers of the Blennioidei with a high degree of assemblage composition. Mean fish density endemism (Hastings, 2009). Three of the six (especially of benthic species; e.g. G. adustus, families that comprise this suborder (Labri- B. ramosus, U. macrocephalus, T. petersi) was somidae, Chaenopsidae and Dactyloscopidae)

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 385 are endemic or nearly endemic to the New the species dominating these assemblages World. Most species within this suborder belonged to Blenniidae (Scartichthys viridis), inhabit shallow coastal waters and have small Tripterygiidae (Tripterygion chilensis) and body sizes. These two characteristics are likely Kyphosidae (Girella laevifrons). The presence responsible for the successful invasion and of these families in intertidal habitats of the residency of this group in rocky intertidal areas Tropical Eastern Pacific is marginal and often (i.e. tidepools) in the Neotropics (especially the composed of transients and accidental visitors Tropical Eastern Pacific). Contrary to the Blen- (i.e. Ophioblennius steindachneri and Axocli- nioidei, the Labroidei are transient components nus lucillae). Although dispersed, tidepool fish of the tidepool ichthyofauna of the Tropical studies in the Tropical Eastern Pacific region Eastern Pacific region. Labridae and Poma- (and the Colombian Pacific coast) have been centridae are species-rich families in rocky slowly increasing, filling in a gap acknowl- reef areas, where access to intertidal habitats edged by previous authors (Prochazka et al., (including tidepools) offers additional food 1999). Further investigation at discrete regions sources and/or refuge from predation to young of Central America and the southern limit of or small individuals of these species. the TEP (i.e., Ecuador and Peru), will help to At regional scales, few attempts have corroborate the patterns drawn in this initial been made to document tidepool fish assem- zoogeographic comparison. blage differences, but recently Boyle & Horn In summary, tidepool fish assemblages (2006) made comparisons at different levels found at Gorgona Island are likely among the of central Chile and California tidepool fish most species-rich of this type of assemblages in assemblages in the Eastern Pacific finding no the TEP region and the whole Eastern Pacific common species and only two shared fami- coast. The observed inter-site differences in lies (Gobiesocidae and Clinidae). The present assemblage structure may be directly related regional comparison of tidepool ichthyofaunas to differences in the topographic structure of between the temperate and tropical regions of tidepools found at each of the study sites. Rote- the Eastern Pacific Ocean shows taxonomic none and clove oil are both effective methods differences even at the family level in line with the observations of Prochazka et al. (1999). to survey tidepool fish assemblages, how- In the northern hemisphere at 37ºN, Yoshi- ever, caution should be taken when comparing yama (1981) found a dominance of species results due to differences in the susceptibility (more than 50% of all species) by members of fish species to each of these methods. These of the families Cottidae (Scorpaeniformes) differences may be reflected in changes in the and Stichaeidae (Perciformes); these families proportion of the dominant species. The region- are totally absent from tropical southern lati- al patterns observed here reflect commonly tudes. The most speciose families of the tropi- defined biogeographical patterns observed for cal regions are Pomacentridae, Muraenidae, the shore fish fauna of the Eastern Pacific. A Gobiidae and Labrisomidae. It is noteworthy distinct species composition was observed for that anguilliform species (mainly Muraenidae) Chilean and northern USA west coast locali- are present only in the tropical portion of the ties, whereas similarities could be observed Eastern Pacific (Table 4). This was also noted for localities in the TEP including the Cortez by Mahon & Mahon (1994) in the Caribbean province. Tidepool fish assemblages in some where they found a similar number of species areas of the TEP are as species-rich as in the from this order. same habitats in the Caribbean region. The later In the southern part of the Eastern Pacific region is considered to have greater fish species at 32-33ºS, Muñoz & Ojeda (1997) identified richness in certain shallow water habitats (e.g. a predominance of species of the families coral reefs; Floeter et al., 2008). Therefore, Labrisomidae and Gobiesocidae. Nevertheless, other factors operating at local scales may be

386 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 important in determining the number of fish últimos 40 años, especialmente en zonas templadas. En species in these habitats. este trabajo, se documenta la variación espacio-temporal The dynamics of tidepool fish assemblages en corta escala de un ensamblaje de peces de charcos intermareales en el Pacífico Oriental Tropical (Isla Gor- at high tide and the possible connections to the gona, Colombia), y se comparan nuestros resultados, con subtidal environments are practically unknown los resultados obtenidos durante muestreos realizados en in the region and should be addressed in la misma localidad en 1993. Además se realiza la primera the future to fully understand the importance comparación latitudinal de estos ensamblajes para la costa of these particular assemblages in the whole oriental del Océano Pacífico (43°N a 36°S). Durante abril a junio de 2006, se realizaron muestreos mensuales en 18 coastal seascape continuum. Finally, tidepool charcos intermareales distribuidos en tres localidades de fish assemblages at Gorgona Island and in the Isla Gorgona, estableciendo un ensamblaje de peces de whole TEP constitute excellent systems for charcos intermareales con una alta riqueza de especies testing the effects of climatic variability intrin- (53 especies). Labrisomidae, Gobiidae and Gobiesoci- sic to this region, providing understanding dae fueron los componentes residentes dominantes en el that is urgently needed in an era of accelerated ensamblaje, mientras que Pomacentridae, Muraenidae and Labridae fueron los componentes transitorios. Aunque climatic changes. el registro de las mismas especies dominantes en 1993 y 2006 sustenta la idea de una alta persistencia temporal de este ensamblaje, se detectaron diferencias significativas en ACKNOWLEDGMENTS la abundancia relativa de las especies. Estas diferencias podrían ser el resultado de las diferentes técnicas de cap- Sampling at Gorgona Island in 2006 was tura utilizadas durante 1993 y 2006 (rotenona vs aceite de possible thanks to the logistic and financial clavo). Al comparar los estudios sobre peces intermareales support of Animal Ecology Research Group at realizados en la zona tropical y zona templada de la costa Universidad del Valle, Scientific Research Sta- oriental del Océano Pacífico, se identificó un claro patrón tion Henry von Prahl at National Natural Park latitudinal en la composición de los ensamblajes, que es Gorgona, Colciencias-Univalle Research Proj- consecuente con la ampliamente documentada subdivisión biogeográfica de la fauna marina de esta región. Para ect RC-243-2004/1106-09-17092 to A. Giral- comprender el funcionamiento de los ecosistemas costeros do and F.A. Zapata, and Colciencias “Young del Pacífico Oriental tropical, es necesario que futuros Researcher Program” grant to G. Castellanos- esfuerzos de investigación se encaminen a conocer el papel Galindo (p-20005-0028). Tide Pool sampling ecológico de los peces intermareales en esta región. was granted by UAESPNN under research per- Palabras clave: ictiofauna intermareal, estructura de mit DTSO-gr-003/2006 to A. Giraldo. Informa- comunidad, zoogeografía, Colombia, Océano Pacífico tion on tidepool fish assemblages of Gorgona Oriental. in 1993 was extracted from samples collected by H.F. García for his unpublished B.Sc. the- sis supervised by F.A. Zapata and deposited REFERENCES at Universidad del Valle. We are indebt to E. Ackerman, J. L. & Bellwood, D. R. (2002). Comparative Escarria, E. Velasco, D. Ramirez and J. Benitez efficiency of clove oil vs. rotenone for sampling tro- for their dedication during sampling activities. pical reef fish assemblages. Journal of Fish Biology, P. Hastings, J. McCosker and J. Van Tassell 60: 893-901. provided useful literature. English proofread- Allen, G. R. & Robertson, D. R. (1994). Fishes of the ing by P. Tuda and comments on early versions tropical eastern Pacific. University of Hawaii Press, of this manuscript by U. Krumme, M. Wolff Honolulu, Hawaii. and I. Freytag and two anonymous reviewers Almada, V. C. & Faria, C. (2004). Temporal variation of are deeply appreciated. rocky intertidal resident fish assemblages-patterns and possible mechanisms with a note on sampling protocols. Reviews in Fish Biology and Fisheries, RESUMEN 14: 239-250. Anderson, M. J. (2001). A new method for non-parametric A nivel mundial, la ecología de peces de charcos multivariate analysis of variance. Australian Journal intermareales ha sido ampliamente investigada durante los of Ecology, 26: 32-46.

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 387 Anderson, M. J. (2006). Distance-based tests for homo- Seasonal and El Niño Southern Oscillation. Limnolo- geneity of multivariate dispersions. Biometrics, 62: gy and Oceanography, 45: 1368-1379. 245-253. Davis, J. L. D. (2000b). Spatial and seasonal patterns of Anderson, M. J. & Willis, T. J. (2003). Canonical analysis habitat partitioning in a guild of southern California of principal coordinates: a useful method of constrai- tidepool fishes. Marine Ecology Progress Series, ned ordination for ecology. Ecology, 84: 511-525. 196: 253-268. Anderson, M. J., Gorley, R. N. & Clarke, K. R. (2008). Dominici-Arosemena A. & Wolff, M. (2006). Reef fish PERMANOVA+for PRIMER: Guide to software and community structure in the Tropical eastern Paci- statistical methods. PRIMER-E:Plymouth, United fic (Panamá): living on a relatively stable rocky Kingdom. reef environment. Helgoland Marine Research, 60: Arakaki, S. & Tokeshi, M. (2006). Short-term dynamics of 287-305. tidepool fish community: diel and seasonal variation. Floeter S. R., Rocha, L. A., Robertson, D. R., Joyeux, J. C., Environmental Biology of Fishes, 76: 221-235. Smith-Vaniz, W. F., Wirtz, P., Edwards, A. J., Barrei- Benfield, S., Baxter, L., Guzman, H. M. & Mair, J. M. ros, J. P., Ferreira, C. E. L., Gasparini, J. L., Brito, (2008). A comparison of coral reef and coral com- A., Falcón, J. M., Bowen, B., & Bernardi, G. (2008). munity fish assemblages in Pacific Panama and envi- Atlantic reef fish biogeography and evolution. Jour- ronmental factors governing their structure. Journal nal of Biogeography, 35: 22-47. of the Marine Biological Association of the United Gibson, R. N. (1988). Patterns of movement in intertidal Kingdom, 88: 1331-1341. fishes. In G. Chelazzi & M. Vannini (Eds.), Beha- vioural Adaptation to Intertidal Life (pp. 55-63). Berrios, V. L. & Vargas, M. E. (2000). Estructura del Plenum Press, New York, USA. ensamble de peces intermareales de la costa rocosa del norte de Chile. Revista de Biología Marina y Gibson, R. N. (1992). Tidally synchronized bahaviuor in Oceanografía, 35:73-81. marine fishes. In M. A. Ali (Ed.), Rhythms in Fishes (pp. 55-63). Plenum Press, New York. USA. Boyle, K. S. & Horn, M. H. (2006). Comparison of fee- ding guild structure and ecomorphology of intertidal Gibson, R. N. (1993). Intertidal teleosts: life in a fluctua- fish assemblages from central California and central ting environment,. In T. J. Pitcher (Ed.), Behaviour Chile. Marine Ecology Progress Series, 319: 65-84. of Teleost Fishes (pp. 513-536). Chapman & Hall, London, United Kingdom. Briggs, J. C. & Bowen, B. W. (2012). A realignment of marine biogeographic provinces with particular refe- Gibson, R. N. (2003). Go with the flow: tidal migration in rence to fish distribution. Journal of Biogeography, marine . Hydrobiologia 503: 153-161. 39: 12-30. Gibson, R. N. & Yoshiyama, R. M. (1999). Intertidal fish Castellanos-Galindo, G. A., Giraldo, A. & Rubio, E. communities. In M. H. Horn, K. L. M. Martin & A. (2005). Community structure of an assemblage M. A. Chotkowski (Eds.), Intertidal Fishes: Life in of tidepool fishes on a Tropical Eastern Pacific two worlds (pp. 264-296) Academic Press, London, rocky shore, Colombia. Journal of Fish Biology, 67: United Kingdom. 392-408. Glynn, P. W. & Ault, J. S. (2000). A biogeographic analysis Castellanos-Galindo, G. A., Krumme, U. & Willis, T. J. and review of the far eastern Pacific coral reef region. (2010). Tidal influences on fish distributions on tropi- Coral Reefs, 19: 1-23. cal eastern Pacific rocky shores (Colombia). Marine Griffiths, S. P. (2000). The use of clove oil as an anaesthetic Ecology Progress Series, 416: 241-254. and method for sampling intertidal rockpool fishes. Chang, K. H., Lee, S. C., Lee, J. C. & Chen, C. P. (1973). Journal of Fish Biology, 57: 1453-1464. Ecological study on some intertidal fishes of Taiwan. Griffiths, S. P. (2003). Rockpool ichthyofaunas of tempe- Bulletin Institute of Zoology Academia Sinica, 12: rate Australia: species composition, residency and 45-50. biogeographic patterns. Estuarine, Coastal and Shelf Chotkowski, M. A., Buth, D. G. & Prochazka, K. (1999). Science, 58: 173-186. Systematics of Intertidal fishes. In M. H. Horn, K. Griffiths, S. P., Davis, A. R. & West, R. J. (2006). Role of L. M. Martin & M. A. Chotkowski (Eds.), Intertidal habitat complexity in structuring temperate rockpool Fishes: Life in two worlds (pp. 297-331). Academic ichthyofaunas. Marine Ecology Progress Series, 313: Press, London, United Kingdom. 227-239. Davis, J. L. D. (2000a). Changes in a tidepool fish assem- Grossman, G. D. (1982). Dynamics and organization of a blage on two scales of environmental variation: rocky intertidal fish assemblage: the persistence and

388 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 resilience of taxocene structure. American Naturalist, Palmer, J. D. (2000). The clocks controlling the tide- 119: 611-637. associated rhythms of intertidal animals. BioEssays, Hastings, P. A. (2009). Biogeography of New World blen- 22: 32-37. nies. In R. Patzner, E. Gonçalves, P. Hastings & B. Prochazka, K. (1996). Seasonal patterns in a tempera- Kapoor (Eds.), The Biology of Blennies (pp. 95-118). te intertidal fish community on the west coast of Science Publishers, Enfield, New Hampshire, United South Africa. Environmental Biology of Fishes, 45: Kingdom. 133-140. Horn, M. H., Martin, K. L. M. & Chotkowski, M. A. Prochazka, K., Chotkowski, M. A. & Buth, D. G. (1999). (1999). Introduction. In M. H. Horn, K. L. M. Martin Biogeography of intertidal Fishes. In M. H. Horn, K. & M. A. Chotkowski (Eds.), Intertidal Fishes: Life L. M. Martin & M. A. Chotkowski (Eds.), Intertidal in two worlds (pp.1-6). Academic Press, London, Fishes: Life in two worlds (pp. 332-355). Academic United Kingdom. Press, London, United Kingdom. Little, C. & Kitching, J. A. (1996). The biology of rocky Quijada, P. A. & Caceres, C. W. (2000). Patrones de abun- shores. Oxford University Press, Oxford, United dancia, composición trófica y distribución especial Kingdom. del ensamble de peces intermareales de la zona López-Victoria, M., Cantera, J. R., Díaz, J. M., Rozo, D. centro-sur de Chile. Revista Chilena de Historia M., Posada, B. O. & Osorno, A. (2004). Estado de Natural, 73: 739-747. los litorales rocosos en Colombia: acantilados y pla- Robertson, D. R. (1998). Do coral-reef fish faunas have yas rocosas. In INVEMAR (Ed.), Informe del estado a distinctive taxonomic structure? Coral Reefs, de los ambientes marinos y costeros en Colombia: 17:179-186. año 2003 (pp. l71-182). INVEMAR, Santa Marta, Robertson, D. R. & Allen, G. (2002). Shorefishes of the Colombia. Tropical Eastern Pacific: an information system. Macieira, R. M. & Joyeux, J. C. (2011). Distribution pat- CD-ROM. Smithsonian Tropical Research Institute, terns of tidepool fishes on a tropical flat reef. Fishery Balboa, Panamá. Bulletin, 109: 305-315. Robertson, D. R., Grove, J. S. & McCosker, J. E. (2004). Mahon, R. & Mahon, S. D. (1994). Structure and resilience Tropical Transpacific shorefishes. Pacific Science, of a tidepool fish assemblage at Barbados. Environ- 58: 507-565. mental Biology of Fishes, 41: 171-190. Robertson, D. R. & Smith-Vaniz, W. F. (2008). Rotenone: Martins, G. M., Hawkins, S. J., Thompson, R. C. & Jen- an essential but demonized tool for assessing marine kins, S. R. (2007). Community structure and functio- fish diversity. BioScience, 58: 165-170. ning in intertidal rock pools: effects of pool size and Robertson, D. R. & Cramer, K. L. (2009). Shore fishes and shore height at different successional stages. Marine biogeographic subdivisions of the tropical eastern Ecology Progress Series, 329: 43-57. Pacific. Marine Ecology Progress Series, 380: 1-17. Metaxas, A. & Scheibling, R. E. (1993). Community struc- Ruiz-Campos, G., González-Guzmán, S., Ramírez-Valdéz, ture and organization of tidepools. Marine Ecology A., González-Acosta, A. F., Castro-Aguirre, J. L. & Progress Series, 98: 187-198. De La Cruz-Agüero, J. (2010). Composition, density Mok, H. & Wen, P. (1985). Intertidal fish community and biogeographic affinities of the rocky intertidal ecology on Lu Tao Island (Green Island), Taiwan. fishes on the western coast of the Baja California Journal of the Taiwan Museum, 38: 81-118. peninsula, Mexico. California Cooperative Oceanic Mora, C. & Robertson, D. R. (2005). Factors shaping the Fisheries Investigations, 51: 210-220. range-size frequency distribution of the endemic fish Thomson, D. A. & Lehner, C. E. (1976). Resilience of a fauna of the Tropical Eastern Pacific. Journal of Bio- rocky intertidal fish community in a physically uns- geography, 32: 277-286. table environment. Journal of Experimental Marine Muñoz, A. A. & Ojeda, F. P. (1997). Feeding guild structure Biology and Ecology, 22: 1-29. of a rocky intertidal fish assemblage in Central Chile. Weaver, P. L. (1970). Species diversity and ecology of Environmental Biology of Fishes, 49: 471-479. tidepool fishes in three Pacific coastal areas of Costa Nielsen, K. J. (2001). Bottom-up and Top-down forces in Rica. Revista de Biología Tropical, 17: 165-185. tide pools: test of a food chain model in an intertidal Willis, T. J. & Roberts, C. D. (1996). Recolonisation community. Ecological Monographs, 71: 187-217. and recruitment of fishes to intertidal rockpools at

Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014 389 Wellington, New Zealand. Environmental Biology of Island. Proceedings of the 8th International Coral Fishes, 47: 329-343. Reef Symposium, 1: 1029-1034. Yoshiyama, R. M. (1981). Distribution and abundance pat- Zapata, F. A. (2001a). Formaciones coralinas de Isla Gor- terns of rocky intertidal fishes in Central California. gona. In L. M. Barrios & M. Lopez-Victoria (Eds.), Environmental Biology of Fishes, 6: 315-332. Gorgona Marina: contribución al conocimiento de una isla única (pp. 27-50). INVEMAR, Serie Publi- Yoshiyama, R. M., Sassaman, C. & Lea, R. N. (1986). caciones Especiales No. 7, Santa Marta, Colombia. Rocky intertidal fish communities of California: tem- Zapata, F. A. (2001b). Peces Marinos de Gorgona: Genenra- poral and spatial variation. Environmental Biology of lidades. In L. M. Barrios & M. Lopez-Victoria (Eds.), Fishes, 17: 23-40. Gorgona Marina: contribución al conocimiento de Zapata, F. A. & Morales, Y. (1997). Spatial and temporal una isla única (pp. 107-109). INVEMAR, Serie Publi- patterns of fish diversity in a coral reef at Gorgona caciones Especiales No. 7, Santa Marta, Colombia.

390 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 62 (Suppl. 1): 373-390, February 2014