Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Coral Reefs (1995) 14:225-235 Coral Reefs 9 Springer-Verlag1995 ~ The effects of fishing on the diversity, biomass and trophic structure of Seychelles' reef fish communities S. Jennings 1, E. M. Grandcourt 2, N. V. C. Polunin 1 1Departmentof Marine Sciencesand Coastal Management,The University,Newcastle-upon-TyneNE 1 7RU, UK 2SeychellesFishingAuthority,PO Box 449, FishingPort, Mah6, Seychelles Accepted: 10 January 1995 Abstract. A fishery independent underwater visual census technique was used to assess the effects of fishing on the diversity, biomass and trophic structure of the diurnally active non-cryptic reef-associated fish communities of the Seychelles. One hundred and thirty four species associated with three significantly different types of reef habitat were censused at one unfished ground and in six fishing grounds subject to different fishing intensities. There was an inverse relationship between fishing intensity and the biomass of several species targeted by the fishery. The diversity of families containing target species (lutjanidae, lethrinidae) was significantly higher at unfished and lightly fished sites as was the total biomass of the fish community and the biomass of piscivorous, piscivorous/invertebrate feeding and herbivorous trophic groups. However, there was no indication that the biomass of non-target species increased in response to the removal of their predators by fishing. The findings of this study are significant for fishery managers because they suggest that the intensive differential cropping of top predators will not necessarily lead to increases in the biomass and productivity of their prey. Introduction Fishing is the most widespread human exploitative activity on coral reefs, and fisheries of socio-economic significance have developed in all areas where humans can access fishing grounds (e.g. Munro and Williams 1985; Russ 1991). Fishing activities have direct and indirect effects on reef fish populations and their ecosystems (Russ 1991). Such effects are significant because they may lead to changes in the size and composition to yield from the fishery and alter subsequent fish production processes (Jennings and Lock 1996; Jennings and Polunin 1996). Correspondence to. S. Jennings Given current concerns for the sustainability of reef fisheries and a burgeoning literature suggesting approaches for their management (Pauly 1988; Medley et al. 1993) there are remarkably few studies of fishing effects (Munro et al. 1987; Russ 1991; Jennings and Lock 1996). Existing knowledge of fishing effects has been gained using two principal approaches: temporal comparisons of fish community structure in an area subject to changing rates of exploitation (e.g. Russ 1985; Koslow et al. 1988; Alcala and Russ 1990) and spatial comparison between the fish communities in two or more areas subject to different levels of fishing intensity (e.g. Samoilys 1988; Polunin and Roberts 1993). The former approach has typically been used to demonstrate short-term fishing effects following the opening or closing of reserves, but these may differ from the long term effects in established fisheries. The latter approach has generally been applied to compare fished and unfished areas or has considered a relatively small proportion of the fish community (e.g. Grigg 1994). Fishing effects have rarely been examined across quantifiable gradients of fishing intensity using fishery independent assessments of community structure. Decreases in the size, abundance and biomass of those piscivorous or carnivorous species which are favoured for consumption or sale are expected to be the most readily detectable direct effects of fishing (Russ and Alcala 1989). These effects have been recorded in reef fisheries as changes in catch-per-unit-effort (e.g. Ralston and Polovina 1982; Koslow et al. 1988) or by independent assessment techniques such as underwater visual census (Bohnsack 1982; Samoilys 1988; Russ and Alcala 1989; Polunin and Roberts 1993; Grigg 1994). The removal of piscivorous fishes may have repercussions throughout the ecosystem. Piscivorous fishes are probably the most significant consumers offish biomass (Grigg et al. 1984) and in most reef systems will consume considerably more fish biomass than is removed by fishing (Jennings and Lock 1996). Theorists have predicted that the removal of predatory species should result in the proliferation of their prey (Beddington and May 1982; Beddington 1984) but 226 evidence for such effects in reef ecosystems r e m a i n s equivocal (Russ 1985, 1991; H i x o n 1991; Caley 1993). I n the p r e s e n t study we use a fishery i n d e p e n d e n t assessm e n t t e c h n i q u e to e x a m i n e the effects Of fishing o n the diversity, b i o m a s s a n d t r o p h i c structure o f the fish comm u n i t i e s f r o m three of the m a i n types o f fished reef h a b i t a t in the Seychelles. W i t h i n each h a b i t a t type, fish comm u n i t i e s were investigated at o n e unfished g r o u n d a n d in six fishing g r o u n d s subject to different fishing intensities. W e a i m to describe the direct effects o f fishing o n target species a n d d e t e r m i n e w h e t h e r changes i n the c o m p o s i t i o n o f their c o m m u n i t i e s are associated with changes in the 9diversity or b i o m a s s of species a n d t r o p h i c g r o u p s which are n o t targeted by the fishery. 55~ 55~ i lin NE Cous I ~ U? NI 55~ 55~ Methods Study sites The study was conducted on reefs surrounding the granitic islands of Mah~ and Praslin (Fig. 1). These islands rise from the SeychellesBank, an extensive submarine platform with mean depths of 44 to 65 m. Fringing reefs with a carbonate framework have developed around the islands and, in deeper water, corals grow directly on the granite substrate. Corals are scarce at depths in excess of 20 m (see review by Stoddart 1984). The majority of the Seychellois people live on Mah~ and Praslin and reef-associated fishes are targeted by the artisanal fishery (Mees 1990a). Fishing is not permitted in a number of areas which have been designated as Marine Parks or Special Reserves (Anon 1990a; Jennings et al. 1996). Seven inshore grounds subject to different levels of fishing intensity were selected for study (Fig. 1). Five of these grounds (Mah~ NW, Mah6 W, Mah~ E, Praslin SW and Praslin NE) were based on the strata used by the Seychelles Fishing Authority for their assessment of the artisanal fishery (Anon 1993a). Sainte Anne Marine Park (Ste. Anne) was established by the Government of Seychelles in 1973 (Anon 1973) and Cousin Island Special Reserve (Cousin) was established by the International Council for Bird Preservation (now Birdlife International) in 1968 (Diamond 1975). Poaching takes place at Ste. Anne and a few local families have been granted fishing rights (Anon 1993b). Cousin is considered to be an unfished site (Jennings et al. 1996). Fig. 1. Location of the seven grounds subject to different fishing intensities. Continuous lines bordering Ste. Anne and Cousin indicate boundaries of the reserves. Broken lines indicate fringing reefs. Habitat types: 9 = 1, well-developed fringing reef with a carbonate framework. O -- 2, Coral growth on a granitic substrate in exposed locations. 9 = 3, small rock and coral patch reefs on a predominantly sandy substrate Assessment offish community structure Quantitative estimates of the size and abundance of fish in the seven grounds were made using an underwater visual census (uvc) technique. This was an appropriate technique for censusing the fish communities because speartishing is banned in Seychelles and the fish were not expected to avoid divers. Moreover, recreational diving and the feeding of fish, which can result in fish gathering around divers, do not take place at the Cousin reserve. One hundred and thirty four non-cryptic diurnally active reef associated species which could be positively identified and censused effectively were selected for study (Table 1). Fish were counted and their lengths were estimated at three sites within each fishing ground. The sites were selected to include each of the three main types of fished reef habitat, i.e. (1) well-developed fringing reef with a carbonate framework, (2) coral growth on a granitic substrate in exposed locations and (3) small (typically 3-7 m circumference) rock and coral patch reefs on a predominantly sandy substrate. Counts were conducted by the same observer (SJ) from 26 April to 12 June 1994. Sixteen replicate counts were completed at each site during daylight hours at depths of 3.0 to 13.0 m below chart datum. Replicates were placed randomly with the proviso that the boundaries of adjacent counts were separated by a minimum of 15 m (to avoid bias due to push-pull effects in preceding counts; Samoilys 1992). The uvc point count technique was based on that developed by Samoilys and Carlos (1992). The sizes and abundance of target species within a census area of 7 m radius were determined by counting each individualand making an estimate of its length (to + 1 cm). Counts of the species in each area were conducted sequentially, the most active species being recorded first. When a count for one species was complete, all further movements of that species in or out of the census area were disregarded because counts of incoming fish would bias biomass estimates (Samoilys 1992). When the active, non-territorial species had been counted, a 7 m line was laid to confirm estimates of the position of the census boundary. Smaller territorial species such as Stegastes and Plectroglyphidodonwere then recorded. Only those fish estimated to be >__8 cm were included. The accuracy of length estimation was maintained by practising length estimation with objects of known size (60 lengths of 2 cm diameter plastic pipe cut to lengths of 5-65 cm and threaded onto a line in random order) during the census 227 Table 1. Species for which size and abundance estimates were obtained, the trophic groups to which they were assigned (co corallivore, hb herbivore, ip invertebrate feeder and piscivore, iv invertebrate feeder, om omnivore, pi piscivore,pk planktivore, dt detritivore), their maximum lengths (cm; from Smith and Heemstra 1986; Randall 1992 and personal observation) and their role in the artisanal fishery Trophic group Aeanthuridae Acanthurus leucosternon Bennett, 1832 A. lineatus (Linnaeus, 1758) A. nigrofuscus (Forsskgd, 1775) A. triostegus (Linnaeus, 1758) A. tenneti Gtinther, 1861 Ctenochaetus binotatus Randall, 1955 C. striatus (Quoy and Gaimard, 1825) C. strigosus (Bennett, 1828) Naso lituratus (Forster, 1801) Paracanthurus hepatus (Linnaeus, 1766) Zebrasoma scopas (Cuvier, 1829) Z. desjardini (Bennett, 1835) Balistidae Sufflamen chrysopterus (Bloch and Schneider, 1801) Chaetodontidae Chaetodon auriga Forsskfil, 1775 C guttatisimus Bennett, 1823 C. kleinii Bloch, 1790 C lineolatus Cuvier, 1831 C. lunula Lacep6de, 1803 C. madagaskariensis Ahl, 1923 C. melannotus Bloch and Schneider, 1831 C. meyeri Bloch and Schneider, 1831 C. trifaseialis Quoy and Gaimard, t824 C trifasciatus Park, 1797 C. xanthocephalus Bennett, 1832 C zanzibariensis Playfair, 1867 Haemulidae Diagramma pictum (Thunberg, 1792) Plectorhinchus gibbosus (Lacep6de, 1802) P. orientaIis (Bloch, 1793) P. schotaf (Forsskfil, 1775) Labridae Anampses meleagrides Valenciennes, 1840 Bodianus axillaris (Bennett, 1831) Cheilinus digrammus (Lacep6de, 1801) C. fasciatus (Bloch, 1791) C. trilobatus (Lacep~de, 1801) Corisformosa (Bennett, 1834) Epibulis insidiator (Pallas, 1770) Gomphosus caeruleus Lacep6de, 1801 Halichoeres cosmetus Randall and Smith, 1982 H. hortulanus (Lacep6de, 1801) H. marginatus Riippell, 1839 H. scapularis (Bennett, 1831) Hemigymnusfasciatus (Bloch, 1792) H. melapterus (Bloch, 1791) Labrichthys unilineatus (Guichenot, 1847) Labroides biocoIor (Fowler and Bean, 1828) L. dimidiatus Valenciennes, 1839 Macropharyngodon bipartus Smith, 1957 No vaculichthys taeniourus (Lacep6de, 1801) hb hb hb hb hb dt dt dt hb pk om om Length in cm and role 23 d 38 ~ 21 d 27 ~ 31 ~ 22 d 26 ~ 19 d 45 d 31 ~ 19 ~ 40 d iv 20 d iv 20 ~ 11 d 12 d 30 ~ 20 ~ 14 d 15 ~ 20 d 17 d 14d ROd 15c iv co CO iv iv CO CO co co iv iv iv iv iv iv iv iv pi iv iv iv iP iv iv 1V 1V 1V ap xp 1v 1V 1V 1V 1V 90 b 70" 55 a 40" 21 d 20 d 30 ~ 36 c 40 b 60 ~ 35 d 28 d 12 d 27 d 17 d 20 d 45 r 50 ~ 17 d 14d 12a 12d 25 d Table 1. (Continued) Stethojulis albovittata (Bonnaterre, 1788) Thalassoma hardwicke (Bennett, 1828) T. herbracium (Lacep6de, 1801) T. lunare (Linnaeus, 1758) Lethrinidae Lethrinus concyliatus (Smith, 1959) L. enigmatus undescribed L. harak (Forsskgd, 1775) L. lentjan (Lacep+de, 1802) L. mahsena (Forssk~d, 1775) L. mahsenoides Valenciennes, 1830 L. nebulosus (Forsskgd, 1775) L. obsoletus (ForsskN, 1775) L. olivaceus Valenciennes, 1830 L. rubrioperculatus Sato, 1978 Monotaxis grandoculis (Forsskgd, 1775) Lutjanidae Aprion virescens Valenciennes, 1830 Lutjanus argentimaculatus (Forsskfd, 1775) L. bohar (ForsskN, 1775) L. fulviflamma (Forsskgd, 1775) L. gibbus (ForsskN, 1775) L. kasmira (ForsskS,1, 1775) L. monostigma (Cuvier, 1828) L. rivulatus (Cuvier, 1828) L. russelli (Bleeker, 1849) Macolor niger (Forsskfil, 1775) Monacanthidae Cantherines pardalis (Rfippell, 1837) Oxymonocanthus longirostris (Bloch and Schneider, 1801) Mullidae Mulloidesftavolineatus (Lacepbde, 1801) P. barberinus (Lacep~de, 1801) P. bifasciatus (Lacep6de, 1801) P. ciliatus (LacepSde, 1801) P. cyclostomus (Lacep6de, 1801) P. rubescens (Lacep6de, 1801) P. macronema (Lacep~de, 1801) Trophic group Length in cm and role iv 12 d 18 d 23 ~ 27 c iv iv iv lV 76 a 40 a 60 a 40 a 38 a 25 a 87 a 40 a 100 a 50 a 56" iol px pl 11o 11o 11o lp lp lp pk 100 a 100 a 95 ~ 30 a 50 a 30 b 45 a 60 b 40 b 70 b iv 21 c 10d lp lp lp lp lp lp lp lp lp lp CO ip ip ip ip iv 40 c 50 ~ 35 c 40 ~ 50 ~ 42 c 32 c iv 28 c Pomaeanthidae Apolemichthys trimaculatus (Lacep~de, 1831) i v Centropyge multispinis (Playfair, 1867) iv Pomacanthus imperator (Bloch, 1787) iv P. semicireulatus (Cuvier, 1831) iv 21 d 12d 40 c 38 d Nemipteridae Scolopsisfrenatus (Cuvier, 1830) Pomacentridae Ambyglyphidodon leucogaster (Bleeker, 1847) Chromis atripectoralis Welander and Schultz, 1951 C. ternatensis (Bleeker, 1856) C. weberiFowler and Bean, 1928 Dascyllus trimaculatus (Riippell, 1828) Neoglyphidodon melas (Cuvier, 1830) Plectroglyphidodon dickii (Lienard, 1839) P. johnstionus Fowler and Ball, 1924 P. lacrymatus Quoy and Gaimard, 1824 iv ip pk 14a pk 14d pk pk pk iv 11 d 14d 14d 186 11 d 10 d 12 d om om hb 228 Table 1. (Continued) Description o f habitat characteristics Trophic group Length in cm and role hb hb 12d 12 d hb hb hb hb 54c 80 b 75 b 35 c hb hb hb hb hb hb hb hb hb hb hb hb hb hb hb 70 b 476 476 75 b 70 b 27 ~ 356 306 75 b 37 b 356 386 40 b 33 ~ ip pi 61 ~ 50 b Cephalopholis argus pi 50 ~ (Bloch and Schneider, 1801) C. leopardus (Lacep6de, 1801) C. miniata ((Forssk~d, 1775) C. urodeta (Bloch and Schneider, 1801) Epinephelus caeruleopunctatus (Bloch, 1790) E. fasciatus (Forsskgd, 1775) E. hexagonatus (Bloch and Schneider, 1801) E. rnerra Bloch, 1793 E. spilotoceps Schultz, 1953 E. tukula Morgans, 1959 ip pi ip pi ip ip ip ip ip 206 41 ~ 28 b 60 a 35 a 25 a 28 b 356 200 a Siganus argenteus hb 37 a (Quoy and Gaimard, 1825) S. puelloides Woodland and Randall, 1979 S. stellatus Forsskfd, 1775 S. sutor (Valenciennes, 1835) hb hb om 356 30 b 45" iv 22 d Pomacentrus trilineatus Cuvier, 1830 Stegastesjasciolatus (Ogilby, 1889) Searidae Calotomus carolinus (Valenciennes, 1840) Cetoscarus bicolor (Rtippell, 1829) Hipposcarus harid ((Forsskgd, 1775) Leptoscarus vaigiensis (Quoy and Gaimard, 1824) Scarus atrilunula Randall and Bruce, 1923 S. caudofasciatus (Gunther, 1862) S. falcipinnis Playfair, 1867 S. frenatus Lacep6de, 1802 S. ghobban Forsskgd, 1775 S. gibbus Riippell, 1829 S. globiceps Valenciennes, 1840 S. niger Forsskgd, 1775 S. psittacus Forssk~l, 1775 S. rubroviolaceus (Bleeker, 1849) S. scaber Valenciennes, 1840 S. sordidus Forssk~l, 1775 Scarus sp S. tricolor Bleeker, 1847 S. viridifucatus (Smith, 1956) 30 b Serranidae Aethaloperca rogaa (ForsskN, 1775) Anyperodon leucogramma (Valenciennes, 1828) Siganidae Habitat was described both within the perimeter of each replicate count area and within the overall boundaries of the study site. When a count was complete the percentage cover (based on plan view) of sand, rubble, rock, massive coral, soft and branching coral was estimated and topographic complexity of the substrate was described using the 6 point scale of Polunin and Roberts (1993). An exposure index was calculated for each site by determining the 15~ sectors (first sector beginning at true north) from which the site was fully exposed to an unobstructed wave fetch of 3 km or more, and summing the product of the mean annual windspeed on Mah6 (km h ~) and duration (hours) in each of these sectors (Seychelles Meteorological Office, unpublished data). The general seabed habitat at each site was described in terms of composition and profile. The number of counts in which the dominant underlying substrate consisted of carbonate framework, sand or granite was expressed as a proportion of the total number of replicate counts within each site, and the minimum and maximum depth within each count area was recorded to provide a crude index of site rugosity. D a t a analysis Estimates of fish length were converted to mass using published length-weight relationships (from Kulbicki et al. 1993 and unpublished sources). When the weight-length relationship for a given species was not available, then the relationship for a species from the same genus and with similar morphology was used. For 19 species this was not possible and the relationships were selected for fish with similar morphology. Fish were assigned to trophic groups (Table 1) on the basis of information provided by Hiatt and Strasburg (1960), Vivien (1973), Sano et al. (1984), Randall (,1983, 1992), Blaber et al. (1990), Randall et al. (1990) and observations during the census period. An examination of the relationships between habitats at the sites selected for study was made by subjecting the means of all replicatespecific habitat indices and the values of site specific habitat indices to an agglomerative hierachical clustering procedure, using the average linkage method (Sokal and Michener 1958). All data were standardized to mean = 0 and SD = 1 before clustering. The significance of differences between clusters was assessed using a test of dissimilarity (Carr 1991). The significance of differences in the biomass of species or trophic groups between habitats and grounds was assessed using two-way crossed A N O V A with loge +1 transformed data. When interactions between habitats and grounds were not significant (P>0.05) and when differences between habitats were not significant (P > 0.05) biomass data were presented as pooled means. Otherwise, the data were re-analyzed within-habitats using one-way ANOVA. Tukey's method (P = 0.05; Day and Quinn 1989) was used to determine the significance of differences in mean biomass between specific habitats or grounds. One-way and two-way crossed ANOVA were also used to compare the number of species recorded per count in the different grounds or habitats. Zanelidae Zanclus cornutus (Linnaeus, 1758) a Primary target species caught by most of the fishing techniques in use b Secondary target species caught using a limited number of techniques ~ Bycatch sometimes retained if caught d Species which are not retained if caught; from Seychelles Fishing Authority (unpublished data) period and assessed using methods of Bell et al. (1985) and Polunin and Roberts (1993). The time required to complete a census was not standardized since this is dependent on the number and diversity of fish in the census area. Estimation o f fishing intensity The Seychellois fish from small outboard, powered open boats and a small proportion ofunpowered pirogues (Anon 1993a). Larger boats, such as 'whalers' and 'schooners', rarely fish around inshore reefs although they do land catches on Mah6 and Praslin (Mees 1990b). The Seychelles Fishing Authority conducts a regular census of the number of small boats in operation and their landing sites (Anon 1991, 1992, 1993a). Data from the 1992 censuses (Anon 1993a) were used to estimate the monthly mean number of boats fishing with lines or traps (the methods used on inshore reefs) on a sea-area specific basis within each fishing ground. This provides a reasonable index of fishing intensity because the majority of small boats fish in areas relatively 229 close to their landing sites and a similar proportion of the fleet use specific gear types in the different grounds (Anon 1993a). There are no records of the number of boats fishing or poaching in Ste. Anne, so the effective number of boats fishing was calculated from knowledge of the number of traps and lines licensed to local families (Seychelles Division of Environment, unpublished) and records of illegally positioned traps confiscated by wardens of the National Parks Service (Anon 1993b). It was assumed (1) that 50% of the traps set by poachers were confiscated, (2) that the ratio of trap to line fishing effort was 1:1 and (3) that a single boat with fishing power equivalent to those in other grounds would deploy four traps day-~. Areas of reef habitat (of the three types where fish counts were conducted) within each fishing ground were determined on the basis of sea-surface area using digitized navigational charts (UK hydrographic office charts 722 and 742) or maps (UK ordnance survey series Y851) and records of substrate type collected during diving o o o f- qS: -c," o o o o o I I I I I I I I 7 6 5 4 3 2 1 0 dissimilarity Fig. 2. A dendrogram which shows the grouping of habitat types formed by hierachical classification analysis. For symbols see Fig. 1 surveys. Areas were measured from the coast to the 20 m isobath for grounds based on Seychelles Fishing Authority strata and to the reserve boundary at Ste. Anne and Cousin. Results C l u s t e r i n g o f sites o n the basis o f their h a b i t a t c h a r a c t e r istics (Fig. 2) p r o d u c e d t h r e e significantly different g r o u p s ( P < 0 . 0 5 ) . T h e g r o u p s c o r r e s p o n d e d With the c a r b o n a t e b a s e d c o r a l r e e f (code 1), r o c k b a s e d c o r a l reef (2) a n d p a t c h reef (3) h a b i t a t t y p e s selected for study, w i t h the e x c e p t i o n o f the c a r b o n a t e b a s e d r e e f site in M a h 6 W (Fig. 1) w h i c h c l u s t e r e d w i t h the p a t c h reefs o n s a n d y s u b s t r a t e . C h a r a c t e r i s t i c s o f the different h a b i t a t t y p e s are s u m m a r i z e d in T a b l e 2. T h e d i s t r i b u t i o n o f fishing intensity indices ( T a b l e 3) in r e l a t i o n to the g e o g r a p h i c l o c a t i o n o f the g r o u n d s (Fig. 1) i n d i c a t e s t h a t g r o u n d s w i t h similar fishing i n t e n s i t y were n o t g r o u p e d o n the s a m e islands o r c o a s t s o f islands. A l l l u t j a n i d , lethrinid, c h a e t o d o n t i d , m u l l i d a n d scarid species e n c o u n t e r e d d u r i n g the survey were i n c l u d e d in the census a n d their w i t h i n - f a m i l y species richness is s h o w n in Fig. 3. T h e n u m b e r o f l e t h r i n i d species r e c o r d e d p e r c o u n t was significantly h i g h e r at the unfished g r o u n d in h a b i t a t s 1 ( P < 0.005) a n d 2 ( P < 0.001) a n d the least intensively fished g r o u n d in h a b i t a t 3 ( P < 0.005). C o r r e l a t i o n s bet w e e n the n u m b e r o f l e t h r i n i d species p e r c o u n t a n d fishing i n t e n s i t y were significant in h a b i t a t 1 (r = - 0 . 7 7 , P < 0.05) b u t n o t in h a b i t a t s 2 o r 3 ( P > 0.05). T h e n u m b e r o f l u t j a n i d species r e c o r d e d p e r c o u n t was h i g h e r in the u n f i s h e d o r least intensively fished g r o u n d s o f all h a b i t a t t y p e s ( P < 0 . 0 5 ) a n d c o r r e l a t i o n s b e t w e e n species n u m b e r a n d fishing i n t e n s i t y were significant at h a b i t a t s 2 (r = - 0 . 8 8 , P < 0.01) a n d 3 (r = - 0 . 8 8 , P < 0.05) b u t n o t 1 (P>0.05). Differences in the m e a n n u m b e r of c h a e t o d o n t i d , m u l l i d a n d s c a r i d species were significantly different in all h a b i t a t s ( P < 0.05), w i t h the e x c e p t i o n o f m u l l i d s in h a b i t a t 3 ( P > 0.05). H o w e v e r , t h e r e were n o significant c o r r e l a t i o n s b e t w e e n fishing i n t e n s i t y a n d the Table2. Characteristics ofthe habitat types selected for study. Vahies are means + SD(n = 112) unless indicated otherwise. Habitat codes follow Fig. 1 Coral cover (%) Sand cover (%) Rubble cover (%) Rock cover (%) Mean depth(m) Mean (max-min depth) (m) Habitat complexitya Exposurea,c Calcareous substrateb,c Sand substrateb'~ Rock substrateb.~ Habitat 1 Habitat 2 Habitat 3 34.1 7.1 32.7 13.4 5.8 2.2 0.62 0.18 0.86 0.05 0.09 24.6 4.1 25.0 40.3 7.1 4.1 0.64 0.55 0.09 0.13 0.79 19.22 13.4 22.1 30.5 6.8 2.8 0.65 0.38 0.44 0.48 0.08 _+ 1t.ll + 9.70 + 20.40 + 12.50 +_ 0.61 + 0.35 + 0.039 + 0.236 + 0.152 + 0.076 + 0.134 a Expressed in relative terms where maximum possible value = 1 u Expressed as the proportion of replicate counts in which a given substrate was dominant ~ _+ 10.91 + 6.28 + 25.25 +_ 28.73 _+ 1.55 + 0.70 + 0.069 + 0.156 + 0.087 + 0.114 + 0.157 _+ 7.10 + 12.44 + 13.15 + 10.36 +_ 1.23 + 0.57 + 0.053 + 0.137 _+ 0.051 + 0.059 + 0.059 230 Table 3. Fishing intensity indices for seven Seychelles' fishing grounds Ground Cousin Ste. Anne Mah6 N W Mah6 W Mah6 E Praslin SW Praslin NE Boats (mean number fishing month ~) Reef area 0 3.7 25.9 27.0 45.2 31.2 20.2 0.4 2.1 2.7 4.6 9.0 4.7 5.7 (km 2) - Intensity index (boats km-2) 0 3.5 19.9 11.7 10.0 13.3 7.0 Lutjanidae 4 Lethrinidae 2 o 0 I I Chaetodontidae o ID 1 o 9 OoO I E I I Mullidae E 9 9 0 A9 I lk O I I I (P< 0.05). Scaridae -*o! 0 e 0 ; o t 10 between grounds (P< 0.05). The biomass of A. virescens and L. obsoletus was significantly higher in the two least heavily fished sites and that ofL. bohar significantly higher in the unfished site (P < 0.05). Similarly, in two of three habitat types the biomass of C. argus and S. argenteus was significantly higher in the unfished site (P < 0.05). For the ten,species considered to be Of secondary importance as target species (Fig. 4b) the mean biomass at different grounds differed significantly only for L. harak in habitat 2 (P < 0.05) L. fulviflamma in habitat 3 (P < 0.05), S. ghobban (pooled data, P < 0.05) and S. frenatus (pooled data, P < 0.01). In addition, there was a significant negative correlation between the biomass of S. frenatus and fishing intensity (r = -0.86, P < 0.01). Differences between the biomass of the ten non-target and by-catch species in different grounds (Fig. 4c) were frequently significant. However, there were no significant correlations between fishing intensity and mean biomass of these ten species (P<0.05) and no trends indicating consistently higher biomass at specific sites. Hence the significance of decreases in the biomass of individual target species was generally attributable to differences between the unfished and/or most lightly fished site and all others. The total biomass of all fish censused (Fig. 5) was significantly higher in the unfished and least heavily fished site than at all others in all habitats (P<0.05). Total biomass in habitats 1 and 3 was significantly correlated with fishing intensity (r=-0.84, P<0.05; r=-0.75, P < 0.05) but not in habitat 2 (P < 0.05). Differences between mean biomass in different grounds were significant (P < 0.05) for herbivores and invertebrate (Fig. 5; feeders in all habitats (with the exception of invertebrate feeders in habitat 2). However, the only significant correlations between biomass and fishing intensity were for herbivores and invertebrate feeders in habitat 3 (r =-0.81, P < 0.05; r =-0.74, P< 0.05). Piscivore and invertebrate feeder/piscivore biomass (Fig. 6) was significantly correlated with fishing intensity (r = -0.76, P < 0.05; r=-0.85, P<0.01) and the biomass was significantly higher at the unfished and least intensively fished sites 2'0 fishing intensity (boats km-2) Fig. 3. The relationship between the numbers of Species recorded and fishing intensity within five families of reef fishes. Symbols indicate habitat types and follow Fig. 1 mean number of chaetodontid, mullid or scarid species recorded per count (except for chaetodontids in habitat 3 (r---0.82, P<0.05) and no indication of consistently higher numbers of these species at specific sites. Twenty-five species each contributed more than 1% of the total recorded biomass, and their biomass at the different grounds is shown in Fig. 4. The mean biomass of primary target species (Fig. 4a) differed significantly The total biomass of corallivorous, detritivorous, omnivorous and invertebrate feeding fish which grow to a maximum size of less than 20 cm (Fig. 7), and which are unlikely to be directly affected by fishing due to their small size and feeding strategy, differed significantly between grounds in all habitat types (P<0.05), but in no case was biomass significantly correlated with fishing intensity (P< 0.05). Discussion Evidence for fishing effects Clearly, differences in the diversity, biomass or trophic structure of reef fish communities may be attributed to a number of factors and may not be persistent in space or time. However, there is evidence to suggest that fishing effects were responsible for some of the significant trends we observed. In particular, trends were examined across a 231 A. virescens . eel 0 9 C. argus '~ L. obsoletus 2 ~0 0 H. harid ~0 ~176 L, harak L, bohar 2 e L. fulviflamma S. argenteus Coo Scarus frenatus S. ghobban 1t~ 9 E ; o 1 .I=~ "i~ "o S. gibbus ~,o ; ~ S. niger 9 t 0 9 350 ~o S. scaber 9 l9 9 Ill 0 ~O o9 9 0 " 9 9 S. puelloides 04 I 9 9 9 9o I 0 9 9 9 S. sordidus 0 9 OO O 9 ~0 I0 9 ~176176 e~ A. leucostemum C, trilobatus 9 0 L , e~ e ' 9 C. atripectoralis . ~ ~" ' ~ q 0 H. fasciatus ~0 ," P.barberinus 0 20 o- -'o 0 9 io 20 i o ;~ ~' all groups 8oI 0 I I 40 ,r ~o 0 20 0 fishing intensity (boats km-2) I ~40m ~' 0 I 0 $ 9 ,,t herbivores I !20 I i I { 0 I ~ 0 I 20 i ]t ~ ", 0 ~ IA~ "o~ 20 grounds and habitats or differences between habitats were not significant (two-way crossed ANOVA, P> 0.05). In other cases symbols indicate habitat types and follow Fig. 1. Values of zero are not plotted invertebratefeeders 2of I 9 : Scolopsis frenatus ~'o 20 r { i oo P. lacrymatus Fig. 4a-e. The relationship between biomass and fishing intensity for a primary target species; b secondary target species and e species which are retained as bycatch or not fished. Mean values for replicates pooled across habitats are presented when interactions between 80 C.strigosus 51 o0o. P. ciliatus 'I: o C. striatus 4i: 0 I I I I I I 10 I i 0 i 4oI t 00 20 p 5 r 10 lt5 i 0 20 0 10 t 5 i 10 i 15 ~i 0 20 0 1tO 115 20 fishing intensity (boats km-2) Fig. 5. Relationships between mean biomass (+95% CL, n = 16) of trophic groups and fishing intensity (habitats: 1. upper, 2. centre, 3. lower). Interactions between grounds and habitats, and differences between habitats, were significant (two-way crossed ANOVA, P < 0.05). Symbols indicate habitat types and follow Fig. 1 range o f seven g r o u n d s subject to different fishing intensities rather than using the paired (in space or time) c o m p a r i s o n s which typified m a n y previous studies. Whilst h a b i t a t characteristics such as coral cover, reef size, reef height, t o p o g r a p h i c c o m p l e x i t y o f the substrate, current n o w and exposure h a v e been s h o w n to influence fish a b u n d a n c e (review by Williams 1991), the analysis o f h a b i t a t d a t a f r o m the different g r o u n d s suggested that similarities within h a b i t a t types between fishing g r o u n d s are considerably greater t h a n those between h a b i t a t types 232 piscivores 20I lO ,{ E 0 }{}, v invertebrate feeders/piscivores 0 o I I I I 5 10 15 20 fishing intensity (boats kin-2) Fig. 6. Relationships betweenmean biomass (+95% CL, n = 48) of piscivores and invertebrate feeders/piscivoresand fishing intensity. Mean values for replicates pooled across habitats are presented because interactions between grounds and habitats, and differences between habitats, were not significant (two-way crossed ANOVA, P>0.05) 0 2 ~ oE ~5 0 I I I I r i i , 1; 1'5 20 } I 5 I fishing intensity (boats km -2) Fig. 7. Relationships between mean biornass (+95% CL, n = 16) of small (maximum length < 20 cm) corallivorous, detritivorous, omnivorous and invertebrate feeding fishes and fishing intensity in three habitats (habitats: 1. upper, 2. centre, 3. lower). Interactions between grounds and habitats, and differences between habitats, were significant (two-way crossed ANOVA, P<0.05). Symbols indicate habitat types and follow Fig. 1 and within grounds. Furthermore, the distribution offishing intensity indices in relation to geographic location of grounds indicated that grounds with similar fishing intensity were not grouped on the same islands or coasts of islands and, therefore, geographic trends in fish distri- bution were unlikely to lead to differences between fish populations which would be falsely attributed to fishing. Existing descriptions of the movements and migrations of many reef fish species (review by Parrish 1989; Jennings and Lock 1996) suggest that only a relatively small proportion of fish would move between relatively large fishing grounds and thereby affect biomass estimates. However, given the proximity of individual grounds, the current flows on the Seychelles Bank (Seychelles Division of Environment, unpublished) and the duration of pelagic egg and larval phases in many reef fish (review Victor 1991) it is reasonable to assume that fish populations within individual grounds are not predominantly self-recruiting. As a result the observed fishing effects are expected to indicate the responses of the resident post-recruitment fish community &fishing. Diversity and biomass The findings of the present study corroborate existing evidence that fishing causes detectable decreases in the diversity and biomass of target species (review by Russ 1991). For the species which showed the strongest responses to fishing (predominantly scarids, lethrinids and lutjanids) differences in the response to fishing in different habitats were rare and inconsistent. This suggests that the responses to sustainable fishing intensities were no more marked on true carbonate reefs than in sandy areas with low coral cover or on reefs where corals form a veneer on the rock substrate. There was no indication that target species were replaced by non-target species with similar functional roles in the censused community. If such effects were typical responses to fishing, it is likely that they would have occurred because the relative differences in fishing intensities between grounds have been sustained for at least seven years (Anon 1991, 1992, 1993a). The results suggest that increases in effort within sustainably fished areas would have a smaller impact on biomass than starting to fish in an unfished system. Jennings and Polunin (1995a) noted similar effects when using catch-per-unit-effort as an index of fishing effects in six Fijian reef fisheries subject to different levels of exploitation. The two factors which determine yield at a given biomass will be the rate of recruitment to the fishery and the growth of these recruits. On the basis of information available it is impossible to determine their relative significance. Biomass may not change across a range of fishing intensities because the fishers are consistently cropping the majority of those fish which recruit to the fishery. In addition, an increased rate of biomass regeneration (fish production) may compensate, in part, for the removal of fish by the fishery. This possibility is compatible with the widely reported decreases in size that have been attributed to fishing (reviews Russ 1991; Jennings and Lock 1996) since the form of fish growth trajectories is such that, within a population of given biomass, decreases in mean size are expected to lead to increases in mean production. An additional consequence of there being a greater proportion of the biomass in smaller fish of fewer dominant age classes may be that fluctuations in recruitment will 233 have greater relative effects on the total stock biomass and thus lead to greater instability (variance) in population size. This may make fishing effects harder to detect in heavily fished areas. Intensity of exploitation The results of the present study suggest that, despite decreases in total biomass, processes of recruitment and fish production are sufficient to allow yields to be maintained at existing fishing levels. The total catch of reef associated species landed by the small boat fishery off Mah6, Praslin and La Digue was 759 t in 1992 (Anon 1993a) and the area (based on sea surface area) to 20 m depth around these islands is approximately 170 km 2. The seabed in this area is often sand, rock, seagrass or rubble bottom and may not be fished as intensively as reef areas, but it nevertheless forms a habitat crudely comparable with those measured in other studies of area specific reef fish yields (Russ 1991; Jennings and Polunin 1995b). The mean yield from this area would be 4.5 t km 2 U x, which suggests that the fishing grounds we studied were fished with a range of fishing intensities from unfished to close to the limits of sustainability (Russ 1991). It should be emphasized, however, that the yields from heavily fished areas may only be maintained because larvae can recruit from adjacent and less heavily fished areas. Thus the collective effects of fishing must be considered and the study does not necessarily indicate that it is safe to increase effort to a theoretical maximum throughout the islands. Predator-preyrelationships Whilst piscivory is one of the major processes of energy transfer within the reef ecosystem (Grigg et al. 1984; Parrish et al. 1985, 1986; Norris and Parrish 1988; Hixon 1991) the present study provides no evidence for the proliferation of prey species following a significant reduction in the biomass of some of their predators. Thus small (<20 cm maximum size) corallivores, detritivores, omnivores, and invertebrate feeders are neither target nor retained bycatch species and yet there were no signs of increased biomass in grounds where predator populations had been reduced. Furthermore, if any responses were not detected due to the variability in the data, they would have been too small to compensate for the reduction in the biomass of target species, and thus total community biomass decreased significantly in response to fishing. Other studies of predation effects have also shown no marked responses in prey populations but have been hampered by the immigration of new piscivores to the study area and experimental artifacts or other extraneous factors which affect prey abundance (e.g. Lassig 1982; Stimson et al. 1982; Thresher 1983; Doherty and Sale 1985; Sphigel and Fishelson 1991). However, in the present study there were significant differences between the biomass of the censused piscivorous and invertebrate feeding/piscivorous species at the different grounds, although it is not clear how fishing has affected the density of cryptic piscivores such as muraenids or roving piscivores such as carangids and some elasmobranchs. These species may play an important functional role in the ecosystem. Sudekum et al. (1990) calculated the fish consumption of two carangid species on Hawaiian reef and they consumed more fish per unit area than are characteristically removed by man from the most heavily fished reef areas (Russ 1991). The lack of significant prey release may be a function of the complexity of trophic interactions within the reef ecosystem. Hixon (1991) suggested that diffuse predation may mask predator-prey relationships because the overall effect of all piscivores on the prey population is substantial and yet the impact of each species is minor. Bohnsack (1982) demonstrated significant decreases in the biomass of a few piscivorous species in a fished area and also found no convincing evidence for widespread changes in prey populations. Caley (1993) on the other hand removed all reef species likely to consume fish (with the exception of roving species) and found evidence for increases in the abundance of some prey species. The findings of the present study are important for fishery managers because they suggest that the intensive differential cropping of top predators may not necessarily lead to increases in the biomass or productivity of their prey. In enclosed lake fisheries there are examples of prey release following piscivore removal (Marten 1979 a,b). However, these systems are not open and lack the complexity of trophic linkages between fish species which characterize reef ecosystems and the diversity of piscivorous species is considerably lower (Jones 1982). Munro and Williams (1985) and Grigg et al. (1984) suggested that actively fishing piscivorous species may lead to increases in catches of prey species. It is unlikely that such a strategy would provide sufficient additional yield to be effective unless a very wide range of predators could be cropped, using a method akin to the complex age, size and situation related feeding strategies of the piscivorous fishes. A fishing strategy of this type is likely to be impractical. It may be more beneficial to crop prey species on the basis that they form a major proportion of total community biomass and may be faster growing and therefore more productive than species at higher trophic levels. Acknowledgements. We wish to thank: Philippe Michaud, David Boull6, Joel Nageon de Lestang and other staff of Seychelles Fishing Authority for considerable technical and logistical assistance: JeanClaude Michel, Suzanne Marshall and staffof the Seychelles Division of Environment for permission to work in Ste. Anne Marine Park, logistical assistance and the provision of unpublished data; Robbie Bresson and Jim Stevenson of BirdLife International for their assistance with the work at Cousin; Chris Mees of the Renewable Resources Assessment Group for useful advice; and the referees for some valid comments which have been incorporated in the text. The study was supported by the British Overseas Development Administration through its Fishery Management Science Programme. References Alcala AC, Russ GR (1990) A direct test of the effects of protective management on the abundance and yield of tropical marine resources. J Cons Int I'Explor Mer 46:40-47 Anon (1973) National Parks and Nature Conservancy Act. Ste. Anne Marine Park Regulations. Government of Seychelles SI 58/1973 234 Anon (1990) Environmental Management Plan of the Seychelles 1990-2000. Department of Environment Government of Seychelles Mah6 Seychelles (no continuous pagination) Anon (1991) Seychelles artisanal fisheries statistics for 1990. Seychelles Fishing Authority Mah6 Seychelles Anon (1992) Seychelles artisanal fisheries statistics for 1991. Seychelles Fishing Authority Mah6 Seychelles Anon (1993a) Seychelles artisanal fisheries statistics for 1992. Seychelles Fishing Authority Mah6 Seychelles Anon (1993b) 1992 Annual Report: Conservation and National Parks. Ministry of Environment,Economic Planningand External Relations, Government of Seychelles, Mah6, Seychelles Beddington JR (i 984) The response of multispecies systems to perturbations. In: RM May (ed) Exploitation of marine communities. Springer, Berlin, pp 209-225 Beddington JR, May RM (1982) The harvesting of interacting species in a natural ecosystem. Sci Am 247:42-49 Bell JD, Craik GJS, Pollard DA, Russell BC (l 985) Estimating length frequency distributions of large reef fish underwater. Coral Reefs 4:41-44 Blaber SJM, Milton DA, Rawlinson NJF (1990) Diets of lagoon fishes of the Solomon Islands: predators of tuna baitfish and trophic effects of baitfishing on the subsistence fishery. Fish Res 5:263-286 Bohnsack JA (1982) Effects of piscivorous predator removal on coral reef fish community structure. In: Caillet GM, Simenstad CA (eds) Gutshop '81: Fish food habits and studies, Washington Sea Grant Publication, University of Washington Seattle, pp 258-267 Caley MJ (1993) Predation, recruitment and the dynamics of communities of coral-reef fishes. Mar Biol 117:33 43 Carr MR (1991) User guide to Plymouth routines in multivariate ecological research. Plymouth Marine Laboratory, Plymouth, UK Diamond AW (1975) Cousin Island Nature Reserve Management Plan 1975-1979. International Council for Bird Preservation, Oxford UK Day RW, Quinn GP (1989) Comparisons of treatments after an analysis of variance in ecology. Ecol Monogr 59:433 463 Doherty PJ, Sale PF (1985) Predation on juvenile coral reef fishes: an exclusion experiment. Coral Reefs 4:225~34 Grigg RW (1994) Effects of Sewage discharge, fishing pressure and habitat complexity on coral ecosystems and reef fishes in Hawaii. Mar Ecol Prog Ser 103:25-34 Grigg RW, Polovina JJ, Atkinson MJ (1984) Model of a coral reef ecosystem III. Resource limitation, community regulation, fisheries yield and resource management. Coral Reefs 3:23-27 Hiatt RW, Strasburg DW (1960) Ecological relationships of the fish fauna on coral reefs of the Marshall Islands. Ecol Mono 30:65-127 Hixon MA (1991) Predation as a process structuring coral reef fish communities. In: Sale PF (ed) The ecology of fishes on coral reefs. Academic Press, San Diego, pp 475-508 Jennings S, Lock JM (1996) Population and ecosystem effects of fishing In: NVC Polunin, CM Roberts (eds) Tropical reef fisheries. Chapman and Hall, (in press) Jennings S, Polunin NVC (1995a) Relationships between catch and effort in multispecies reef fisheries subject to different levels of exploitation. Fish Manag Ecol 2:8%101 Jennings S, Polunin NVC (1995b) Comparative size and composition of yield from six Fijian reef fisheries. Journal of Fish Biology 46:28 46 Jennings S, Polunin NVC (1996) Impacts of fishing on tropical reef ecosystems. Ambio (in press) Jenning S, Marshall SS, Polunin NVC (1996) Seychelles' marine protected areas: comparative structure and status of reef fish communities. Biol Cons (in press) Jones R (1982) Ecosystems, food chains and fish yield In: Pauly D, Murphy GI (eds) Theory and management of tropical fisheries. ICLARM Con Proc 9:195-237 Koslow JA, Hanley F, Wicklaud R (1988) Effects of fishing on reef fish communities at Pedro Bank and Port Royal Cays, Jamaica. Mar Ecol-Prog Ser 43:201-212 Kulbicki M, Mou Tham G, Thollot P, Wantiez L (1993) Length- weight relationships of fish from the lagoon of New Caledonia. Naga 2+3:26 30 Lassig BR (1982) The minor role of large transient fishes in structuring small-scale coral patch reef assemblages. PhD Thesis, Macquarie University, Australia Marten GG (1979a) Predator removal: effects on fisheries yields in Lake Victoria (East Africa). Science 203:646-648 Marten GG (1979b) Impact of fishing on the inshore fishery of Lake Victoria (East Africa). J Fish Res Board Can 36:891-900 Medley P, Gaudian G, Wells SM (1993) Coral reef fisheries stock assessment. Rev Fish Biol fish 3:242~85 Mees CC (1990a) The fishermen of Seychelles: results of a socioeconomic study of the Seychelles fishing community. Seychelles Fishing Authority Mah6 Seychelles Mees CC (1990b) Seychelles schooner fishery: an analysis of data collected during the period January 1985 to July 1990 Seychelles Fishing Authority Mah~ Seychelles Munro JL, Williams DMcB (1985) Assessment and management of coral reef fisheries: biological, environmental and socioeconomic aspects. Proc Fifth Int Coral Reef Symp 4:545-581 Munro JL, Parrish JD, Talbot FH (1987) The biological effects of intensive fishing upon reef fish communities. In: Salvat B (ed) Human impacts on coral reefs: facts and recommendations. Antenne Museum E.P.H.E., French Polynesia, pp 41-49 Norris JE, Parrish JD (1988) Predator prey relationships among fishes in pristine coral reef communities. Proc Sixth Int Coral Reef Symp 2:107 113 Parrish JD (1989) Fish communities of interacting shallow-water regions in tropical oceanic regions. Mar Ecol Prog Ser 58:143160 Parrish JD, Callahan MW, Norris JE (1985) Fish trophic relationships that structure reef communities. Proc Fifth Int Coral Reef Syrup 4:73 78 Parrish JD, Norris JE, Callahan MW, Magarifug EJ, Schroeder RE (1986) Piscivory in a coral reef community. In: Caillet GM, Simenstad CA (eds) Gutshop '81: Fish food habits and studies, Washington Sea Grant Publication, University of Washington, Seattle pp 73-78 Pauly D (1988) Some definitions of overfishing relevant to coastal zone management in Southeast Asia. Trop Coastal Area Manag 3:14-15 Polunin NVC, Roberts CM (1993) Greater biomass and value of target coral-reef fishes in two small Caribbean marine reserves. Mar Ecol Prog Ser 100:167 176 Ralston S, Polovina JJ (1982) A multispecies analysis of the commercial deep-sea handline fishery in Hawaii. Fish Bull 80:435448 Randall JE (1983) Red Sea reef fishes. Immel Publishing, London Randall JE (1992) Fishes of Maldives. Immel Publishing, London Randall JE, Allen RA, Steene RC (1990) Fishes of the Great Barrier Reef and Coral Sea. Crawford House Bathurst Russ GR (1985) Effects of protective management on coral reef fishes in the central Philippines. Proc Sixth Int Coral Reef Symp 4:219-224 Russ GR (1991) Coral reef fisheries: effects and yields. In: Sale PF (ed) The ecology of fishes on coral reefs. Academic Press, San Diego pp 601-635 Russ GR, Alcala AC (1989) Effects of intense fishing pressure on an assemblage of coral reef fishes. Marine Ecology Progress Series 56:13-27 Samoilys M (1988) Abundance and species richness of coral reef fish on the Kenyan coast: the effects of protective management and fishing. Proc Sixth Int Coral Reef Symp 2:261-266 Samoilys M (1992) Review of the underwater visual census method developed by the QDPI/ACIAR project: visual assessment of reef fish stocks. Conf Workshop Seri QC92006, Department of Primary Industries, Brisbane, Australia Samoilys M, Carlos GM (1992) Development of an underwater visual census method for assessing shallow water reef fish stocks in south-west pacific. Queensland Department of Primary Industries Cairns, Australia 235 Sano M, Shimizu M, Nose Y (1984) Food habits of teleostean reef fishes in Okinawa Island, southern Japan. University of Tokyo, Univ Mus Bull 25:1-128 Smith MM, Heemstra PC (1986) Smiths' sea fishes. Springer, Berlin Sokal RR, Michener CD (1958) A statistical method for evaluating systematic relationships. University of Kansas Sci Bull 38:1409-1438 Sphigel M, Fishetson L (1991) Experimental removal of piscivorous groupers of the genus Cephalopholis(Serranidae) from coral habitats in the Gulf of Aqaba (Red Sea). Env Biol Fish 31:131 138 Sudekum AE, Parrish JD, Radtke RL, Ralston S (1991) Life history and ecology of large jacks in undisturbed, shallow oceanic communities. Fish Bull 89:493-513 Stimson J, Blum S, Brock RE (1982) An experimental study of the influence of muraenid eels on reef fish sizes and abundance. Hawaii Sea Grant Q 4, 4:1-6 Stoddart DR (1984) Coral reefs of the Seychelles and adjacent regions. In: Stoddart DR (ed) Biogeography and ecology of the Seychelles Islands. DR W Junk, The Hague, pp 63 81 Thresher RE (1983) Enivornmental correlates of the distribution of planktivorous fishes in the One Tree Reef Lagoon. Mar Ecol Prog Ser 10:137-145 Victor BC (1991) Settlement strategies and biogeography of reef fishes. In: Sale PF (ed) The ecology of fishes on coral reefs. Academic Press, San Diego, pp 231 260 Vivien ML (1973) Contribution /t la connaissance de l'~thologie alimentaire de l'icthyofauna du platier interne des r6cifs coralliens de Tul6ar (Madagascar). Tethys Supple 5:221-308 Williams D McB (1991) Patterns and processes in the distribution of coral reef fishes. In: Sale PF (ed) The ecology of fishes on coral reefs. Academic Press, San Diego pp 437-474