Influence of the Leeuwin Current on Recruitment of Tropical Reef Fishes at Rottnest Island, Western Australia

BULLETIN OF MARINE SCIENCE, 54(1): 245-255, 1994

INFLUENCE OF THE LEEUWIN CURRENT ON RECRUITMENT OF TROPICAL REEF FISHES AT ROTTNEST ISLAND, WESTERN AUSTRALIA

J. B. Hutchins and A. F. Pearce

ABSTRACT The appearance of recently settled juveniles of tropical reef fishes at Rottnest Island (32°S) in autumn is examined in relation to the oceanic circulation offWestem Australia. The arrival of the larvae in March and April coincides with the strengthening of the Leeuwin Current, a southward flow of warm tropical water, at that time. Satellite images are used to indicate that tongues of warm water from the Leeuwin Current can cross the shelf and reach Rottnest Island just before recruitment commences. This is discussed in terms of a conceptual model of transport of the larvae in the current.

Rottnest Island, which is situated at 32°S on the outer continental shelf off Perth, Western Australia (Fig. I), is bathed by the warm waters of the Leeuwin Current, the dominant ocean current off Western Australia. It is a southward flow of relatively low-salinity tropical water which is strongest in the southern autumn, winter and early spring months (Rochford, 1969; Cresswell and Golding, 1980; Godfrey and Ridgway, 1985; Pearce and Griffiths, 1991). Unlike the other southern hemisphere eastern boundary currents (the Benguela and Humboldt Currents) which flow equatorwards, the Leeuwin Current flows polewards, and (despite upwelling-favorable winds) there is no significant upwelling off Western Australia (Pearce, 1991). This current transports the larvae of tropical marine organisms down the west coast of Australia and eastwards into the Great Australian Bight (Maxwell and Cresswell, 1981; Cresswell, 1986; Hutchins, 1991; Morgan and Wells, 1991). It is also responsible for the presence of substantial coral reefs as far south as the Houtman Abrolhos (Fig. 1), a group of offshore islands at 29°S (Wilson and Allen, 1987; Veron and Marsh, 1988; Hatcher, 1991). Satellite images and free-drifting buoys have shown that the Leeuwin Current consists of a complex system of meanders and eddies superimposed on the net southward flow (Cresswell, 1980; Legeckis and Cresswell, 1981; Pearce and Griffiths, 1991). Hutchins (199 I) reported on the dispersal of tropical reef fishes to temperate areas of Western Australia, indicating that the Leeuwin Current was probably responsible for the relatively rich tropical fauna at Rottnest Island. From a total of 360 fish species recorded for the island, 98 were tropical, of which 61 were reef-associated species. By comparison, he noted that reefs along the nearby mainland coast supported few tropical fishes. However, this tropical diversity was not widespread at Rottnest Island, instead being confined to particular areas along the island's western and southern coastlines. Some of the species were present in prominent numbers, whereas many were poorly represented. Furthermore, Hutch- ins considered that most tropical species were not breeding successfully at the island. In marked contrast with observations from more northern reefs of the state, new recruits were rarely found at the island during the summer months, first appearing instead in autumn. He suggested that this late start of tropical recruitment at Rottnest Island was probably related to the strengthening of the Leeuwin Current at that time. This recruitment was shown to continue sporadi- cally throughout the winter, before ceasing in spring at about the same time that

245 246 BULLETIN OF MARINE SCIENCE, VOL. 54, NO. I, 1994

WESTERN AUSTRALIA

o 200

Kilometres

30°

115° E 35°'8 " Figure I. Location chart showing Rottnest Island and the Houtman Abrolhos in relation to the (schematic) Lecuwin Current. The dashed line indicates the approximate position of the continental shelfbreak (taken as the 200 m isobath). HUTCHINS AND PEARCE: RECRUITMENT AT ROTTNEST ISLAND 247

115"30'E t 32·5

ROTTNEST ISLAND

o Parker Point

Kilometres

Figure 2. Map of Rottnest Island showing the localities where tropical fish recruitment was observed.

the Leeuwin Current weakened. Hutchins therefore believed that the populations of most tropical reef fishes at the island were not self-sustaining. Here we present evidence in support of the above view. Observations on the settlement of tropical reef fishes in early autumn are here linked with satellite images showing the movement of warm Leeuwin Current waters across the continental shelf to Rottnest Island. Some general concepts on the transport of planktonic larvae in the Leeuwin Current are then addressed.

METHODS

Fish Monitoring. - The data presented here were obtained during the periods 1979-1980, 1983-1985, and 1990-1992. However, correlation between the times of settlement and satellite imagery has only been attempted for the years 1991 and 1992 because of the lack of suitable images for the earlier periods. Visits were made to Rottnest Island every 3 weeks, allowing for disruptions due to unfavorable weather conditions, but additional trips were included during the expected settlement season. The duration for each trip was generally 1 day, but continuous daily monitoring for periods of up to 4 days were occasionally conducted. Monitoring of settlement occurred mostly at Parker Point on the south side of Rottnest Island (Fig. 2). Surveys around the whole island during the early phases of this investigation had shown that settlement of tropical species was best observed in this area. Other sites at Rottnest Island where evidence of settlement was found are also indicated in Figure 2; these were subjected to less intensive surveying, although they were regularly visited during autumn to compare with the results obtained at Parker Point. All locations are shallow bays, well protected from the prevailing southwesterly swell, but still under the influence of moderate currents that ensure a ready supply of nutrients. Most possess a prominent coral fauna, particularly Parker Point where extensive colonies of Poci!lopora damicornis occur (see Hutchins, 1991 for a more detailed description of this location). On each visit, the snorkel-equipped diver (J.B.H.) searched the area for juvenile fish which had obviously just settled. This was determined on the basis of minute (10-20 mm standard length) post- larval individuals. At Parker Point, each dive (approximately I h in duration) involved a transect that followed the extensive system of Poci!lopora colonies, ending in the rocky shallows at the base of the headland (Parker Point). Some species were first sighted in the shallows where no effort was made to conceal themselves; others were found hiding in crevices in the ree( Certain species were initially 248 BULLETIN OF MARINE SCIENCE, VOL. 54, NO.1, 1994 observed in small schools swimming just above the substrate. The times of arrival of species which hide in the reef were difficult to determine as they usually remained undetected until well after settlement. Members of the wrasse genus Thalassoma, for example, have a particularly long cryptic phase before they venture out into the open. Thus most monitoring effort was directed at those species which remained in the open after settling. Censussing involved counting the number of juvenile individuals of each species sighted, and separating these into two categories, those which were new arrivals and those which had been previously recruited (the former were identified by their more larval characteristics, such as a lack of pigmentation, whereas the latter were larger in size with fully developed juvenile features). Attempts were made to follow the survivorship of these fish but with little success. Individuals were difficult to identify as most species tended to associate with others of their species. Some appeared to continually change their location, although remaining essentially in the same area. Several quite large juveniles would appear suddenly in an area where no prior evidence of settlement had been found. This was usually associated with the loss of individuals from a nearby area where recruitment had occurred, which suggested that a short migration must have taken place. Nevertheless, the prominent adult numbers of some tropical species in the area indicate that the survival of these recruits was significant. Satellite Imagery. -NOAA! AVHRR (Advanced Very High Resolution Radiometer) satellite imagery has been used to infer current patterns in the Leeuwin Current during the recruitment periods in Marchi April of 1991 and 1992. Full resolution (1.\ km) images of the brightness temperature in the AVHRR band 4 have been mapped to a transverse Mercator projection and processed to display most clearly the surface thermal structure on the continental shelf in the vicinity of Rottnest Island (Figs. 3,4). Sealevel and Wind Data. -Coastal sealevels may be used as an indicator of the strength of the Leeuwin Current (Pearce and Phillips, 1988). Monthly mean sealevels for Fremantle (the coastal port for Perth) have been obtained from the Australian National Tidal Facility. Figure 5 illustrates the seasonal cycle, with the rising sealevel (and, by inference, strengthening of the Leeuwin Current, Pattiaratchi and Buchan, 1991) during the early part of the year, including the period of larval recruitment at Rottnest Island in Marchi April. Monthly mean wind data have been obtained from a coastal site near Perth (Fig. 5). The alongshore component has a distinct seasonal pattern, with strongly northward wind-stress in the summer period (October to March) and a weaker southwards tendency in winter. The surface Ekman drift is corre- spondingly directed offshore during the summer period, tending to transport any larvae in the surface layer away from the coast, as is the case with the early stage larvae of the western rock lobster (Phillips, 1981; Pearce and Phillips, 1988). By the time of recruitment of the tropical larvae at Rottnest Island, however, the northward wind-stress and offshore Ekman transport are weakening. The period oflarval fish recruitment at Rottnest Island therefore takes place in a regime of increasing sealevel (strengthening Leeuwin Current) and weakening northward wind-stress (Godfrey and Ridgway, 1985).

RESULTS Table 1 lists the months of first observation at Rottnest Island of the post- pelagic juvenile stage of 20 tropical species of reef fishes. Of these, four species were generally the first to be sighted each year, namely Abudefduf sexfasciatus, Abudefduf vaigiensis, Acanthurus triostegus, and Parupeneus spilurus. All remained in the open after settling and were thus reasonably easy to find. They were therefore identified as indicator species, their appearance representing the commencement of tropical fish recruitment at Rottnest Island. During the years 1991 and 1992, special emphasis was placed on detecting the first arrivals of these four species. Settlement in 1991. - Three surveys were conducted at Rottnest Island in March (Table 2), but no evidence of settlement was found. During April, one new recruit of Acanthurus triostegus was observed at Parker Point on the 8th, followed by another individual on the 16th. A new arrival of either Abudefdufvaigiensis or A. sexfasciatus (these species are difficult to separate before attaining all of their pigmentation) was found at Parker Point on the 16th, and two additional individuals were sighted on the same day in Fish Hook Bay. On the 27th, a total of 30 new recruits of both species of Abudefdufwere observed at Parker Point, as HUTCHINS AND PEARCE: RECRUITMENT AT ROTTNEST ISLAND 249

Figure 3. (Upper left and right) Brightness temperature images of the Leeuwin Current near Rottnest Island in 1991: (a) Orbit N 11/12876,26 March, (b) N11/13003, 4 April. Rottnest Island is the brown horizontal island center-right of the image. The warmest water is in brown/red, and the coolest is in blue; the white band in the upper left is cloud. Each image has been individually enhanced to show the thermal features most clearly, so the color coding does not match between the two pictures. Figure 4. (Lower left and right) Brightness temperature images of the Leeuwin Current near Rottnest Island in 1992: (a) Orbit N 11/18057, 27 March, (b) N 11/18184, 5 April. Rottnest Island is the brown horizontal island center-right of the image. The warmest water is in brown/red, and the coolest in blue; the mottled blue region on the left of (b) is cloud. Each image has been individually enhanced to show the thermal features most clearly, so the color coding does not match between the two pictures.

well as two new individuals of Acanthurus triostegus. Four recruits of Parupeneus spilurus and two additional Abudefdufwere sighted on the 29th at Parker Point. On the same day near Salmon Point, three new arrivals of Abudefdufwere found. A satellite image on the 26th March indicated that the Leeuwin Current was flowing relatively strongly some distance offshore of Rottnest Island, with tongues of water approaching the shelf both north of the island and from the south in an anticlockwise eddy-like pattern (Fig. 3a). By 4 April, the meander had further developed, the nearshore branch having moved onto the shelf and the tropical water now reaching the southern side of the island (Fig. 3b). Examination of wind records indicates that the coastal winds were generally weak « 1a knots) for most of April, although there were strong (> 20 knot) northwest or westerly winds on the 8th and lath, and again on the 21st (due to the proximity of cyclone Fifi). These wind events would have moved surface waters shorewards, however, we cannot tell what role they actually played in the recruitment observed on the 16th and 27th respectively. Settlement in 1992.-Surveys were carried out on five occasions in March (Table 2), with a sighting of four new arrivals of Parupeneus spilurus at Green Island on 250 BULLETIN OF MARINE SCIENCE, VOL. 54, NO. I, 1994

Monthly mean sealevels and winds Perth 100 5

4- .•.•...m 90 ! 8' 3 ..• CJ J:l ....., eo Ql 2 J:l u 0 -~ P- u 1 a ~ 70 0 U CJ l'IJ 0 ~ 60 0 -1 Z

50 2 0 1 2 3 5 6 7 B 9 10 11 12 13- Months ...•...Sealevel ..•.. lrmd component

Figure 5. Monthly mean sealevels (solid line) and northward wind components (dotted) for Perth. the 31st. In April, eight new recruits of Abudefduf were observed on the 11th (four at Parker Point and four near Salmon Point). On the same day, one new arrival of Acanthurus triostegus and four of Parupeneus spilurus were found near Salmon Point. On the 15th, 27 new recruits of Abudefduf were sighted (12 at Parker Point, seven at Fish Hook Bay, and eight at Cathedral Rocks), as well as a school (approximately 20) of new arrivals of Parupeneus spilurus at Fish Hook Bay and six recruits of Acanthurus triostegus (one at Parker Point and five at Cathedral Rocks). At Parker Point on the 20th, the number of Abudefdufrecruits had grown to 27 (11 new recruits), and a newly arrived school of Parupeneus spilurus (approximately 20) was present. Eight new recruits of Abudefduf were seen near Salmon Point on the 21st. Figure 4 shows the structure of the Leeuwin Current off Rottnest Island during the 1992 recruitment period. On 27 March, prior to the arrival of tropical fish larvae, there was an anticlockwise eddy off Rottnest (Fig. 4a), but the warm water had apparently not yet reached the island. By 5 April, although the core of the Leeuwin Current was further offshore, the warm water had flooded across the shelf (Fig. 4b), and boating operations on the 6th indicated that clear oceanic water was present around the whole island. Wind records show that the coastal winds were largely easterly (directed offshore) or southerly, and weak to moderate, in late March and April 1992. It would appear therefore that direct wind-driven surface transport may not have assisted recruitment in 1992.

DISCUSSION It is clear that the Leeuwin Current can provide a transport mechanism for eggs and larvae which originate in warmer northern waters of Western Australia and are subsequently dispersed to more southern areas. Successful recruitment is the end result of a chain of events commencing with the source of eggs and larvae, through the transport processes, and ending with settlement. The following dis- HUTCHINS AND PEARCE: RECRUITMENT AT ROTTNEST ISLAND 251

Table 1. Approximate times of first sightings of selected tropical fishes at Rottnest Island

Family and species 1979 1980 1983 1984 1985 1990 1991 1992 Haemulidae Plectorhi nchus jlavomacuialUs Mar. Apr. Mullidae Parupeneus chrysopleuron Mar. Apr. Parupeneus spilurus Mar. May Jun. Mar. Jun. Apr. Mar. Pempherididae Pempheris species Jul. Jun. Mar. Chaetodon tidae Chaetodon auriga Apr. Chaetodon lunula May Pomacentridae Abudefduf sexfasciatus Apr. Mar. Apr. Apr. Mar. May Apr. Apr. Abudefduf sordidus Abudefdllf vaigiensis Apr. Mar. Apr. Apr. Mar. Apr. Apr. Apr. Plectroglyphidodon leucozonus Apr. Pomacentrlls milleri Mar. Apr. Labridae Anampses geographic us Jun. Jul. Labroides dimidiatus May Stethojulis bandanensis Jul. Jul. May Aug. Stethojulis strigiventer Jun. Aug. Thalassorna lunare Sep. Jun. Aug. Thalassoma lutescens Oct. Thalassoma purpureum May Jul. Sep. Scaridae Scams species Apr. Jun. Mar. Acanthuridae Acanthurus triostegus Mar. Apr. Apr. Mar. May Apr. Apr. Apr.

cussion focuses on a conceptual model of propagules of a tropical species being released from a breeding population, picked up by the Leeuwin Current and then subsequently recruited to Rottnest Island. Source of Larvae. - The most likely source of tropical species which recruit to Rottnest Island is the Houtman Abrolhos where high numbers of individuals can be found (Hutchins, 1991). Breeding takes place there during the spring, summer and autumn months. The eggs of most species are released into the pelagic environment, to be dispersed by local currents, and shortly thereafter to hatch as larvae. Obviously some eggs and larvae are washed away from the islands by advective and/or cross-shelf mixing processes to be entrained into the Leeuwin

Table 2. Results of surveys conducted at Rottnest Island in Marchi April during the years 1991 and 1992 (new sightings only)

1991 1992 March April March April

5 12 29 2 5 8 16 27 29 II 13 14 20 21 31 6 11 15 20 21

Abudefduf sexfasciatus 3 30 5 - - 8 27 11 8 - - - - Abudefduf vaigiensis } Acanthurus triostegus 2 ------I 6 Parupenells spilurus 4 4 - 4 20 20 - 252 BULLETIN OF MARINE SCIENCE, VOL. 54, NO. I, 1994

Longitude E Longitude E 111 112 113 114 115 116 III 112 113 114 115 116 N IU OJ II>

N IU In In

w w 0 0 r r (lJ (lJ rT •....rT •.... rT rT w w C .. C .. Cl. D. CIl CIl

UJ UJ

w W N N

W W W W

w W •.. J> Figure 6. Trajectories of satellite-tracked buoys in the Leeuwin Current south of the Houtman Abrolhos in summer and autumn (February to May) 1976: (left) buoy 413, days 83 to 126; (right) buoy 415, days 101 to ISO. For clarity, only the trajectories east of 112°E are shown. Features A to H are discussed in the text.

Current. Satellite imagery shows that the current frequently flows close to the western margins of the Houtman Abrolhos, with warm water intruding onto the shelf between the island groups (Pearce and Griffiths, 1991). Larval Transport Phase. -Fish larvae should not be considered as passively float- ing particles, for they quickly develop into reasonably competent swimmers (Leis, 1988). Some of those which are dispersed by the local circulation may find their way back to the reefs of the Houtman Abrolhos, possibly using a variety of senses such as smell and hearing (Leis, 1988). However, those picked up by the Leeuwin Current when it is flowing strongly would be carried well to the south into more temperate waters. Many are lost to the system as mortality rates are expected to be high: this can be due to high levels of predation, to increasingly unfavorable environmental conditions, or to dispersion out of the current. In particular, the duration of the transport phase must be shorter than the longest time that the larvae can survive without settling. Although there were no satellite-tracked buoys in the Leeuwin Current during the recruitment periods in 1991 and 1992, the trajectories of two buoys drifting off Western Australia between March and May in 1976 show typical features of the summer/autumn circulation which would transport planktonic larvae present in the near-surface layer. The buoys were drogued with parachutes at 20 m depth (Cresswell and Vaudrey, 1977). Figure 6 shows the segments oftrajectory which have been subset from the full tracks to cover only the area between 28° and 34°S, and l12°E to the coast, and for the period between March and May when recruitment at Rottnest is occurring. HUTCHINS AND PEARCE: RECRUITMENT AT ROTTNEST ISLAND 253

m m N N

o o 1Il ~ 1Il ~

, 1:l" \ JET B Z ~ ~ ...N ------ROTTNEST " 36 em/s 5 ~ BQ...TI.N~S! _

JET 0 ~ ~ 6' March 90 April 120 May .00 60 March 90 April 120 May 15' Days (1976) Days 119761 Figure 7. Time-distance plots corresponding to the buoy trajectories (and current features) in Figure 6. The change of latitude with time represents the mean southward speed of the buoy; estimates of the speed for different segments of each trajectory are shown.

Buoy 413 was released into a clockwise eddy (A in Fig. 6 left), and then jetted southwards along the upper continental slope (B) past Rottnest Island until at almost 34°8 it moved onto the shelf and returned northwards in a countercurrent. At 32.5°8 it was re-entrained into the Leeuwin Current and continued southwards beyond 34"S. Buoy 415 entered the Leeuwin Current off the Houtman Abrolhos (E in Fig. 6 right), was carried strongly southwards in the alongshore jet F, cir- culated twice in the clockwise eddy G and then anticlockwise in eddy H. Estimates of the net southward. current speed between the Houtman Abrolhos (29°8) and Rottnest Island (32°8) can be made from the time-latitude plots in Figure 7. While current speeds in the strongest alongshore jets reached almost 50 cm· S-I or 42 km· day-I (about 1 kn), net southward speeds were between 10 and 20 cm· S-I as a result of the eddies. The time interval between the latitudes of the two island groups varied between about lO and 30 days. It is clear from Figure 6 that, during the recruitment period, both clockwise and anticlockwise eddies abound. Larvae could become trapped for periods of days to weeks in such eddies, and may in fact not survive. However, there are also relatively strong alongshore current jets which could carry larvae down to Rottnest Island within two weeks. This is much shorter than the larval duration of between 39 and 55 days for Thalassoma lunare (Victor, 1986) and 19 to 27 days for Abudefduf vaigiensis (Thresher et a1., 1989). Vertical shear may also playa role, in that currents at different depths are likely to move with different speeds (and possibly different directions). However, we know nothing about the vertical distribution of these larvae. Larval Settlement. - The larvae require appropriate habitat on which to settle. For this, they must leave the current either by swimming or by dispersion processes. Analysis of satellite imagery over a period of years indicates that the core of the Leeuwin Current rarely extends across the shelf as far as Rottnest Island, although warm eddies or tongues of water (possibly associated with instabilities in the flow: Pearce and Griffiths, 1991; Mills et a1., 1992) are frequently seen to approach the island. Figures 3 and 4 suggest that such tongues probably played an important role in recruitment to the reefs along the south coast of Rottnest Island in autumn 1991 and 1992. The presence of Leeuwin Current water at the island will not necessarily ensure that recruitment will commence, as there must obviously be larvae in the water mass at the time; the distribution and density of tropical larvae in the Leeuwin Current are presently unknown. After the larvae have reached the island, settlement will only occur if both habitat and environmental conditions are suitable for surviva1. Many ofthe trop- 254 BULLETIN OF MARINE SCIENCE, VOL. 54. NO. I. 1994 ical species found at Rottnest Island occur in the intertidal zone in tropical areas. Obviously they are able to tolerate the harsher environmental conditions that exist in this zone, and therefore are better suited for survival in the more temperate conditions that exist at Rottnest Island. Of course, these conditions are modified by the presence of the Leeuwin Current near the island in autumn and winter. Although Pearce et al. (1985) have shown that the shallow coastal waters off Perth can cool rapidly in autumn, water temperatures along the western half of the island tend to remain high (mean winter minimum of 18°C) due to effective exchange with the warmer offshore waters of the Leeuwin Current (Hutchins, 1991). In addition, the clear oceanic conditions that prevail at the island during autumn and winter, in contrast to the silt laden waters found along the nearby mainland coast, are also a feature of the Leeuwin Current. Thus the warmer and clearer waters at Rottnest Island have allowed a rich coral fauna to develop, which in tum has provided a suitable habitat for tropical reef fishes. To the south of Rottnest Island, however, such conditions are rare, and successful recruitment of tropical reef species is low.

CONCLUSIONS We have demonstrated that the observed recruitment of tropical fish in reefs at Rottnest Island may be explained in terms of transport in the Leeuwin Current, and have presented satellite images which suggest that tongues ofLeeuwin Current water enveloped the Island at about the time that recruitment was observed in 1991 and 1992. It has been inferred that the larvae of these species originated from the Houtman Abrolhos because of the large number of individuals in breeding condition there, and also because the Leeuwin Current flows close to the western portion of this island group. However, the Houtman Abrolhos itself is probably receiving recruits of tropical species from areas further to the north. Some of these may also reach Rottnest Island. For example, several tropical species which were recorded for Rottnest Island on the basis of a few individuals only also occur in very low numbers at the Houtman Abrolhos. This sugge~ts ~hat both areas received recruits of these species from more northern reefs. For example, the rich coral areas of Ningaloo Reef (approximately 21-22°S) lie adjacent to the path of the Leeuwin Current and possess an extremely diverse fish fauna. All the tropical species recorded for Rottnest Island and the Houtman Abrolhos are in prominent numbers at Ningaloo Reef. Nevertheless, because ofthe much smaller distances involved, the more likely scenario is that the Houtman Abrolhos is the main upstream source of tropical recruits at Rottnest Island. The recruitment of tropical fishes at Rottnest Island will continue to be mon- itored. Access to a larger data base will eventually allow a quantitative analysis of this recruitment to be made, and perhaps show differences that may occur during disturbances such as EI Nino/Southern Oscillation (EN SO) phenomena. Furthermore, studies are planned to gain additional data on the relationships between the tropical faunas at Rottnest Island and the Houtman Abrolhos. In particular, a plankton sampling program across the Leeuwin Current in the region between these two areas is proposed for 1995. The results of these investigations should yield a better understanding of the recruitment process at Rottnest Island and the part played by the Leeuwin Current.

ACKNOWLEDGMENTS

The assistance ofthe Western Australian Satellite Technology Applications Consortium (WASTAC) in supplying the images is gratefully acknowledged. The satellite work is supported by the Western Australian Fisheries Research and Development Trust Fund. G. Cresswell (Commonwealth Scientific HUTCHINSANDPEARCE:RECRUITMENTATROTTNESTISLAND 255 and Industrial Research Organisation) kindly provided the buoy data, and N. Sinclair, C. Stott, J. Keesing, and K. Smith (Western Australian Museum) helped with the fish surveys. Tidal material for Fremantle were supplied by the National Tidal Facility, The Flinders University of South Australia, Copyright reserved. Wind data are from Fremantle Port Authority and Steedman Science and Engi- neering. An early draft of this paper was critically read by L Marsh (Western Australian Museum) and G. Cresswell (CSIRO).

LITERATURE CITED

Cresswell. G. R. 1980. Satellite-tracked buoys in the eastern Indian Ocean. Proc. 4th Int. Symp. Rem. Scns. Environ., Costa Rica, April 1980: 531-541. --. 1986. The role of the Leeuwin Current in the life cycles of several marine creatures. Proc. Int. Conf. Pelagic Biogeography, The Netherlands, May/June 1985; UNESCO Tech. Pap. Mar. Sci. 49: 60-64. -- and T. J. Golding. 1980. Observations of a south-flowing current in the southeastern Indian Ocean. Deep Sea Res. 27: 449-466. -- and D. J. Vaudrey. 1977. Satellite-tracked buoy report I: Western Australian releases 1975 and 1976. CSIRO Aust. Fish. Oceanogr. Rep. 86: 1-49. Godfrey, J. S. and K. R. Ridgway. 1985. The large-scale environment of the poleward-flowing Lecuwin Current, Western Australia: longshore steric height gradients, wind stresses and geo- strophic flow. J. Phys. Oceanogr. 15: 481-495. Hatcher, B. G. 1991. Coral reefs in the Leeuwin Current-an ecological perspective. Proc. Leeuwin Current Symp., J. Roy. Soc. West. Aust. 74: 115-127. Hutchins, J. B. 1991. Dispersal of tropical fishes to temperate seas in the southern hemisphere. Proc. Leeuwin Current Symp., 1. Roy. Soc. West. Aust. 74: 79-84. Legeckis, R. and G. R. Cresswell. 1981. Satellite observations of sea-surface temperature fronts off the coast of wcstern and southern Australia. Deep-Sea Res. 28: 297-306. Leis, J. M. 1988. The odd life of larval fish. Aust. Nat. Hist. 22: 452-455. Maxwell, J. G. H. and G. R. Cresswell. 1981. Dispersal of tropical marine fauna to the Great Australian Bight by the Leeuwin Current. Aust. J. Mar. Freshwat. Res. 32: 493-500. Mills, D. A., N. D'Adamo, A. Wyllie and A. F. Pearce. 1992. Dynamical interplay between barotropic and baroclinic mechanisms and relevance to winter flushing of nutrients from the Perth metro- politan region. Sixth Biennial Conference on the Physics of Estuaries and Coastal Seas, Margaret River, Western Australia, 8-10 Dccember 1992 (extended abstract only). Morgan, G. J. and F. E. Wells. 1991. Zoogeographic provinces of the Humboldt, Benguela and Leeuwin Cllrrent systems. Proc. Leeuwin Current Symp., J. Roy. Soc. West. Aust. 74: 59-69. Pattiaratchi, C. B. and S. J. Buchan. 1991. Implications oflong-term climate change for the Leeuwin Current. PrOc. Leeuwin Current Symp., J. Roy. Soc. West. Aust. 74: 133-140. Pearce, A. F. 1991. Eastern boundary currents of the southern hemisphere. Proc. Leeuwin Current Symp.,1. Roy. Soc. West. Aust. 74: 35-45. --, R. E. Johannes, C. R. Manning, D. W. Rimmer and D. F. Smith. 1985. Hydro]ogyand nutrient data off Marmion, Perth, 1979-1982. CSIRO Aust. Mar. Labs. Rep. ]67: 1-45. -- and R. W. Griffiths. ]991. The mesoscale structure of the Leeuwin Current: a comparison of laboratory models and satellite imagery. 1. Geophys. Res. 96(C9): ]6739-16757. -- and B. F. Phillips. 1988. ENSO events, the Leeuwin Current, and larval recruitment of the western rock lobster. J. Cons. Int. Exp]or. Mer 45: 13-21. Phillips, B. F. 1981. The circulation of the southeastern Indian Ocean and the planktonic life cycle of the western rock lobster. Oceanogr. Mar. BioI. Ann. Rev. 19: 11-39. Rochford, D. 1. 1969. Seasonal interchange of high and low salinity surface waters off south-west Australia. CSIRO Aust. Div. Fish. Oceanogr. Tech. Pap. 29: 1-8. Thresher, R. E., P. L Colin and L J. Bell. 1989. Planktonic duration, distribution and population structure of Western and Central Pacific damselfishes (Pomacentridae). Copeia 1989: 420-434. Veron, 1. E. N. and L. M. Marsh. 1988. Hermatypic corals of Western Australia: records and annotated species list. Rec. West. Aust. Mus. Suppl. no. 29: 1-136. Victor, B. C. 1986. Duration of the planktonic larval stage of one hundred species of Pacific and At]antic wrasses (family Labridae). Mar. BioI. 90: 317-326. Wilson, B. R. and G. R. Allen. 1987. Major components and distribution of marine fauna. Pages 43-68 in G. R. Dyne and D. W. Da]ton, eds. Fauna of Australia, Vol lA, Aust. Govt. Pub!. Service, Canberra.

DATEACCEPTED: June 16,1993.

ADDRESSES:(J.B.H.) Western Australian Museum. Francis Street, Perth, Western Australia 6000; (A.F.P.) CSIRO Marine Laboratories. P.O. Box 20. North Beach. Western Australia 6020.