ATOLL RESEARC ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. AUGUST 1994 COLONIZATION OF FISH LARVAE IN LAGOONS OF RANGIROA (TUAMOTU ARCHIPELAGO) AND MOBREA (SOCIETY ARCHIPELAGO) BY V. DUFOUR The colonization of the lagoon by coral reef fish larvae was compared between two islands of French Polynesia, the atoll of Rangiroa and the high volcanic island of Moorea. In both cases the larval flux corning into the lagoon followed a daily cycle. Larvae were mainly caught at dusk and during the night, and on both islands the colonization was higher during moonless than moonlit periods. The la~val flux did not appear to be dependent on the waterflow in the lagoons. A comparison of larval abundance and taxonomic lists indicates that Scarids and Labrids were dominant in Rangiroa while Gobiidae was the major family on Moorea. This difference could be in part related to the different sampling periods, but other environmental and biological factors could also be important. Most reef fishes have a pelagic larval phase, ending with the colonization of the reef (Leis, 1991). This recruitment of fish larvae on coral reefs is now studied in detail since it has been assumed that events occuring during this period determine the characteristics of reef-fish stocks (Sale, 1980; Richards and Lindernan, 1987; Doherty and Williams, 1988). Although some studies emphasized the importance of the processes during the settlement of fish larvae among coral reefs (Sweatman, 1985, 1988; Victor, 1986), this phenomenon is not clearly understood. For fifteen years, scientists have studied mechanisms of this return to the parental habitat. These studies have been limited mainly to continental reefs (e. g.: reefs of Central America), or patch reefs along continental platfoims (e.g.: Great Barrier Reef of Australia) with very little data available on recruitment of reef fish species in oceanic islands and in atolls. This is a first attempt to compare some features of fish colonization of the lagoons of two geomosphologically different islands located in French Polynesia. Laboratoire d'Ichtyoecologie Tropicale et Mediterraneenne, Ecole Pratique des Hautes Etudes, URA CNRS 1453, Universitk de Perpignan 66860 Perpignan Cedex France et Antenne EPHE-Museum Centre de 1'Environnement BP 1013 Moorea Polyntsie Fran~aise Manuscript received 14 October 1993; revised 2 June 1994 Although the data were not obtained simultaneously in both islands, i t is still useful to compare these two sets of data. It is also woizh considering whether or not the observed differences ase due to the location, the geomorphological features of the islands, or the time lag between sample collecting on the two islands. MATERIAL A STUDY AREA Rangiroa Atoll (figure 1) is one of the largest atolls in the world and the most important of the Tuarnotu Archipelago (Ricard, 1985). It is 70 kin long, 30 kin wide and the peripheral rim is 225 km long. One third of the s i n is above the sea surface and consists of small cays separated by channels. The rim flat ic generally wider in the northern than in the southern part (800 in vs 500 in). The lagoon is biologically very rich compared to the other atolls of Tuamotu and is one of the most important reef fisheries centers of this kchipelago. The lnaximuin estimated depth is 35 111 and a lot of pinnacles ase evenly distributed on its suiface. Two passes, 450 to 550 n~ wide and 14 to 35 m deep ase located in the North coast and lagoon waters are flushed out through these passes during ebb tides (35 cm to 60 cm tide range). Oceanic waters flow into the lagoon through channels over the atoll rim and the two passes duiing flood tides and also when trade winds blow. The fish larvae were collected in a channel, midway between the two passes. Moorea Island (figure 1) is located 25 kin north-west of Tahiti (Galzin and Pointier, 1985). This volcanic island has a triangular shape with a 61 krn coastline and a relief of 1200 m. The island is sui-sounded by a ba-sier reef, which encloses a lagoon, 800 to 1600 m wide. The reef is intersected by several passes. Two bays are located on the northern part of the island. The lagoon is generally shallow (1 to 5 m), but deeper near the passes. The oceanic water enters the lagoon by waves breaking over the outer reef crest, and return to the ocean through the passes. The very weak tides on Moorea (average range 15 cm) do not reverse the cui-sent in the passes. Sampling was canied out on the outer reef crest, 600 m away from the pass. METHODS Samples were collected off the northern coasts of both islands. Fish larvae were collected with an anchored net that filtered the wateiflow coming into the lagoon. The net with rectangular mouth (1 x 0.25 m) was of mesh size 0.5 mm. A General Oceanics flowmeter was fixed in the mouth of the net. On Rangiroa Atoll, the net filtered the water coming from the seaward reef flat to the lagoon. It was located 500 m from the outer reef front. The channel was made of gravel in a shallow area (0.5 m). On Moorea Island, the fish larvae were collected on the outer reef crest. The net was fixed on the reef substrate and filtered the water coming over the crest with the ayl uy uope-[arro3 1ue3yjyu%ys jo ax~asqe ayL .1ue3yjyu%ys lou seM (11 a-[qeL) MOTJ mleM ayl pue xnu Iemq ayl uaaMlaq pa~e-[nqw riel s, -[IepuaX ayL ' (2 am%yj) ysnp pug ly%yu le uayel a m u aemq ysyj ISOW leyl M O ~ S O S ~ ealooK uo apeur sa13iC3 layp ayL foul- studied cycles confirms that larval flux did not seem to be quantitatively dependent of the water flow. The study of larval flux on the two islands reveals that the lava1 flux on Rangiroa reached 3 times the value of 500 larvae per sample, which was obtained only once on Moorea, despite a larger sampling effort. On both islands these larval peaks occured in the early evening. A second peak was found just befose dawn on the second cycle on Rangisoa. The water flow during these larval peaks on Rangiroa was not very high and similas to that found during lasval peaks of Moorea. As a result, these high peaks of larval colonization on Rangil-oa and Moosea do not appear to be ueated by variation in ater flow over the reef of these islands. The comnpiirison o f the average jar-val f h x recorded on the two islands at di t times indicates that this f lux appeass to be nwrc significant on Rangiroa than on ea (Table 111). It was obt ious that a high larval flux from these islands was neves recorded during full moon. PIowever, during moonlight periods of the first lunar quarter, thc lasval abundance on Rangiroa was higher than the abundance on Moorea. The number of larvae and the number of larval types were diffesent between the two islands (Table IV). The total number of larvae fsom Rangiroa was almost half the number of those collected from Moorea during eleven months, although the number of samples was higher. Based on the two studied periods, the average larval flux on Rangiroa reached three times the average larval flux on Moorea. The number of larval types on Moorea was 56 for the three cycles. The number of larval types on Rangiroa during only two nights was 43. Sevei-a1 larval types from Moorea were not found on Rangiroa, while only one larval type from Rangiroa was missing from Moorea. Some of these types were represented by more than 50 larvae. The compa~ison between Rangiroa and all the samples of Moorea indicates that the number of types was twice as less as that found in all the samples of Moorea despite the fact that the number of samples collected was eight times higher and the sample period was much longer in Moorea. Therefore, the number of larval types caught in two nights on Rangiroa was significantly higher that those caught off Moorea. The list of the larval types and their abundance is presented for both Rangiroa and Moorea (Fig. 3). The pie diagrams show the percentages of the main larval types for each island. The abundance of the larvae from Moorea is presented for all the 358 samples made between March 1989 and November 1989 (grey bars) and for the three die1 cycles previously studied (black bars). The most abundant larval type on Rangiroa was the Scaridae forming 52% of the total catch. The two most abundant larval types on Moorea were Gobiidae (Gobiidae type 1 and Gobiidae type 56). The abundance of Gobioid types on Moorea represents 63% of the total catch. Scaiidae were the second most abundant family on Moorea but they represented half the number of Scaridae collected from Rangiroa. On Rangiroa Gobioid types were the second most important group but their number were far below those of the Scaidae. The other significant larval types were found in similar numbers on both islands although periods of sampling were different. This was the case for the Labridae, the Callionymidae and the Schindleriidae. It is apparent that the number of larvae of these families would have been much higher on Rangiroa if the extent of sampling was similar to that carried out off Moorea. The Apogonidae type 2 were more abundant on Rangiroa but the total number of Apogonidae from both areas was not very different. Juvenile fishes were caught in both islands in relatively high number. Different families were gathered in this type (Mullidae, Holocentsidae ...). It is interesting to note that these juveniles were collected at dusk despite the fact that daylight was supposed to assist in a higher avoidance of the net. The Gobiidae type 8 was only collected at Rangiroa. The daily patterns of the reef colonization by reef fish larvae have been demonstrated only recently on coral reefs (Dufour, 1991, 1993). The fish larvae that enter the lagoon were caught only at night and dusk. Their abundance was also found to be higher duiing moonless periods. This pattern has been confirmed by samples over a two yeas period. The data from Rangiroa in this study confirm this finding. Each cycle made at Rangiroa demonstrated that fish larvae were abundant during the moonless nights in the channel of the atoll. The larval abundance could reflect higher larval activity above the reef at night (Hobson et Chess, 1978). However, the fixed nets could not catch larvae that do not move into the lagoon. Hobson and Chess (1978, 1986) have demonstrated that planktonic organisms drifted at night over the reef of Enewetak atoll to enter the lagoon. Their appearence over the reef was related to a vertical migsation at night, followed by a passive drift in a cun-ent flow induced by breaking waves. However, colonization by fish larvae at Rangiroa and at Moorea was only accomplished by individuals ready to settle. The larval flux observations do not include preflexion larvae because these larvae were scarce in samples, although they could have drifted more easily than postflexion larvae. It is known that postflexion larvae are able to swim (Blaxter, 1986; Webb and Weihs, 1986). Moreover, reef fish larvae can avoid the reef area until they are competent for metamorphosis (Kingsford and Choat, 1989). These phenomena imply other mechanisms of colonization in addition to passive drift. The larval flux in the lagoon could thus be viewed as an active process made nightly by competent fish larvae. Night activity correllated to the darker phases of the moon cycle has also been demonstrated for other planktonic organisms over reefs (Aldredge and King, 1980, Tranter and al., 1981). These authors found that this moonless activity was an adaptative advantage against predation. In a similar way, the colonization of fish larvae occurs at night when predation is lower (Hobson, 1973, 1975). Therefore, larval colonization of the lagoons at night could be viewed as an adaptative process against predation, as predation plays a major role during the recruitment of reef fishes (Shulrnan and Ogden, 1987, Victor, 1986, Hixton, 1991). Both the geomorphology of the reef and hydrodynamic characteristics of the waters flushing into the lagoons appear to have no significant control on larval colonization. The difference of the abundance of fish larvae between the two islands can be explained by the difference of the sampling periods. Although it has not been established that fish larvae were more abundant in French Polynesia during February than during April, the summer season was considered to be the reci-uitment season in other coral reef areas (Williams, 1983, Victor, 1987). Thus, the lower abundance i n samples fi-om Moorea could be explained by variations related to seasonal recruitment. The difference in abundance and diversity of fishes during colonization between these two islands could also be related to the size of the lagoon. The quotient of reef periphery to surface of the lagoon is also much lower for Rangiroa than for Moorea. This is because the lagoon of Moorea encloses the volcanic island and does not cover all the surface delimited by the outer reef like an atoll. On Moorea, the quotient of the lagoon suiface to the reef length is around 0.86 km-1 (60 km/70 kin*), on Rangiroa it is 0.11 ki11-1 (230 kn1/2100 km2), but the sand cays over one third of the reef lower this coefficient to 0.074. This last value i s more than 10 times smaller than assume that the density of the larval flux pel- unit of lagoon .sui- related to this coefficient, the number of fish colonizing the lagoon should be propoltionally higher. This assumption could explain the higher rate of colonization for angiroa. This hypothesis cannot be vesified, however. because the larval flux over all the seef ~ i r n has not been determined. The difference between the major la~val types from the two islands could also be explained by other hypothesis. The composition and divessity of adult fishes in both lagoons was probably not the same. It is possible that the number of fish species in the lagoons of atoll is related to the surface area of these atolls (Galzin et al., 1994). Scaridae and Labridae are among the most abundant fishes in atoll lagoons (Bouchon- Navano, 1983, Morize d, 1 WO), while Pomacentridae and Acanthuridae are more abundant i n Moorea lagoon (Galzin, 1987). Although we have no information about theii- density in Rangiroa atoll, the higher abundance of Scaiidae lasvae on Rangiroa was not suiprising. But this highei- abundance could be related to the low number of samples collected in Rangii-oa, and the period when they were collected. It is possible, howeves, that the patteln of settlement of fish larvae on reefs could be relatively unpredictable and chaotic and peaks of larvae have be desci-ibed as randomly distributed at different time scales (Doherty and Williams, 1988). Another explanation could be the reproduction period of Scaridae, which could occur easlier. Larvae of Scaridae, however, were caught on Moorea until the end of June and Scaridae and Labridae were also the most abundant families in samples made in May and June 1988 on Moorea. CONCLUSIONS The study of the lai-val flux over the reef on Rangiroa and Mooi-ea was useful to the understanding of some aspects of the settlement processes of fish larvae in lagoons. This study has confirmed some trends in the die1 and lunar cycles of seef colonization by fish larvae. The difference of lasval abundance between samples on both islands can be related to the time lag between the sampling periods of each island. The sizes of the two lagoons could also play a role in this difference. It was more difficult to understand the taxonomic difference. It could be explained by the difference in size of the two lagoons, or by the peiiod of fish reproduction or even by the density of the different families, but few data were available to confilm these hypotheses. REFERENCES Alldredge A. L., King J. M., 1980. Effects of moonlight on the vertical migration patterns of demessal zooplankton. J. exp. mar. Biol. Ecol., 44: 133- 156. Bouchon-Navasro Y. , 1983. Distribution quantitative des principaux poissons herbivores (Acanthuiidae et Scaridae) de l'atoll de Takapoto (Polynesie fsanpise). J. Soc. Ocean. 39(77): 43-54. Blaxter J. H. S., 1986. Development of sense organs and behaviour of Teleost larvae special reference to feeding and predator avoidance. T ~ ~ I I s . Am. Fish. Soc. 1 1 S(86): 98- 1 14 Doherty P.J.. Williams D.McB.. 1988.-The replenishment of coral reef fish populations. Oceanogr. mar. Biol. Ann. Rev., 26: 487-55 1. Dufous V., 1991. Variation d'abondance des lasves de poissons en milieu rkcifal: effet de la lumiere sur la colonisation 1xvaii-e. C.R. Acad. Sci., Paiis, t.3 13, sesie III: 1 87- 194. Dufour V., Galzin R., 1993. Colonization patterns of reef fish larvae to the lagoon at Moorea Island, French Polynesia. Marine Ecology Progress Serie, 102: 143- 152. Galzin R., 1987. Structure of fish community of French Polynesia coral reefs. 11 Spatial scales. Mar. Ecol. Prog. Ser., 41 : 137- 145. Galzin R.. Pointier J-P., 1985. Moorea Island, Society Archipelago. In: Proc. 5th int. Coral Reef Congress, B. Delesalle, R. Galzin and B. Salvat (eds.), 1: 73- 102. Galzin R., Planes S, Dufour V, Salvat B., in press. Variation in diversity of Coral reef fishes among French Polynesian atolls. Coral Reefs. Hixton , 1991. Predation as a process structuring coral reef fish communities. In : The ecology of fish on coral reef, P. F. Sale (ed.), Academic Press : 475-508. Hobson E. S., 1973. Diel feeding migrations in tropical reef fishes. Helgol. Meer. 24 : 36 1-370. Hobson E. S., 1975. Feeding patterns among tropical reef fishes. Am. Scient., 63 : 382-392. Hobson E. S., Chess J. R., 1978. Trophic relationships among fishes and plankton in the lagoon at Enewetak atoll, Marshall island. Fish. Bull, vol76: 133-153. Hobson E.S., Chess J.R., 1986. Diel movements of resident and transient zooplankters above lagoon reefs at Enewetak atoll, Marshall islands. Pac. Sci., (40): 7-25. Kingsford M. J., Choat J. M., 1989. Horizontal distribution patterns of presettlement reef fish: were they influenced by the proximity of reefs? Mar. Biol., 91: 161-17 1. Leis J. M., 1991 The pelagic stage of reef fishes: thc larval biology of coral reef fishes. In: The ecology of fishes on coral reefs. Sale P. F.. ed., Academic Press, San Diego, p. 183- 230. Leis J . M., Rennis D. S., 1983. The lai-vae of Indo-Pacific coral reef fishes. J . M. Leis (ed.) NSW Univ. press & Univ. of Hawaii press, 269pp. Leis J. M., Trnski T., 1989. The larvae of Indo-OPacific shorefishes. NSW Univ. Press & Univ. of Hawaii press, 37 lpp. Morize E., Galzin R., Hal-melin-Vivien M., Arnaudin H., 1990. Osganisation spatiale euplernent ichtyologique d s le lagon de f'atoll de Tikehau (Polynt%ie franpise). Notes et Doc. ORST Ricard M., 1985. angisoa Atoll , Tuarnoeu Archipelago. h: roc. 5th int. Coral Congsess, B. Delesalle, R. Galzin and B. Salvae (eds.), 1: 159-21 0. . J.. Lindeman K . C., 1987. Recruitment dynamics of reef fishes: planctonic processes, settlement and deinersal ecologies, and fishery analysis. Bull. Mar. Sci., 41 (2): 392-410. Shulman M. J., Ogden J. C., 1987. What control tropical reef fish populations: recruitment or benthic mortality? An example in the Caribbean reef fish: Haemulon flavolineatum. Mar. Ecol. Prog. Ser., 39: 233-242. Sweatman H. P. A., 1985. The influence of adults of some coral reef fishes on larval recruitment. Ecol. Monogr. 55: 469-485. Sweatman H. P. A.. 1988. Field evidence that settling coral reef fish larvae detect resident fish using dissolved chemical cues. J. Exp. as. Biol. Ecol., 124(3): 163- 174. Tranter D. J., Bulleid N. C., Campbell R., Higgins H. W., Rowe F., Tranter H. A., Smith D. F., 1981. Nocturnal movements of phototactic zooplancton in shallow waters. Mar. Biol. 61 : 317-326. Victor B. C., 1986. Larval settlement and juvenile mortality in a recruitment-limited coral reef fishes population. Ecol. Monogr., 56(2): 145-160. Victor B. C., 1987. Growth, dispersal, and identification of planktonic labrid and pomacentrid reef fish larvae in the eastein Pacific Ocean. Mar. Biol., 95: 145-152. Webb P. W., Weihs D., 1986. Functional locomotor morphology of early life histoiy stages of fishes. Trans. Am. Fish. Soc., 115: 115- 127. Williams D.McB., 1983.-Daily, monthly and yearly variability in recruitment of a guild of coral reef fishes. Mar. Ecol. Prog. Ser., 10: 23 1-237. Table 1: Values of the Kendall coefficient comelation rank for the lasval flux and the water flow between the two die1 cycles from Rangiroa (n.s: not significant at 5%, s: significant at 5%). Table II: Values of the Kendall coefficient con-elation rank between the water flow and the lasval f lux (n.s: not significant at 5%) . Rangiroa Coinpa~ison of the larval fluxes Comparison of the Wates flows Tableau 111 : Average values of the water flow and the larval flux for the cycles from Kendall coefficient 0.524 s -0.486 s Rangisoa (R) and Moorea (M), standasd deviation ase in brackets. Tableau IV : Abundance of larvae and larval types from Rangiroa and Moorea number of s a m ~ l e s number of larvae larvae / sample number of types Rangiroa I Moosea (3 cycles) 44 I 34 Moorea (all samples) 358 4 165 94.66 43 1369 40.26 5 6 10050 28.1 7 1 Pacific Ocean Figure 1. French Polynesia (above) with the atoll of Rangiroa, Tuarnotu archipelago (middle), and the high Island of Moorea, Society archipelago (below). 16h 1Sh 2011 22h 1 Oh 2h 4h 6h 8h 100 - 100 Second quarter [ Figure 2. Evolution of the larval flux expressed in number of larvae. sample-' (bars) and the water flow in m-3. sample-' (line) during nycthemeral cycles made on Rangiroa and on Moorea. The black thickness on the categories axis represents the night hours, the white frame on the same axis represents moonlit hours. abridae 4 abrdae 5 Gohioidae 8 Le tocephal i i ~aEr idae 6 Juveniles Gobiidae 1 Gobiidac 2 Figure 3. Percentage of the main larval types collected on Moorea and Rangiroa (above) and diagram of larval abundance (below) for all the samples from Rangiroa and for the three cycles of Moorea (black) and all the samples from Moorea (grey). n.i.: not identified to lower taxonomic level; juv.: juvenile; ad.: adulte fishes are also included in this neotenic family.