Potential for collecting tropical marine fish fry by light-attraction

John H. Carleton and Peter J. Doherty[1]

Abstract

Wild fry of marine fishes support a limited range of artisanal grow-out operations in Asia but this does not apply yet to any high value species of coral reef fishes. Pelagic juveniles of many reef species can be attracted with light and AIMS has sampled the supply of these juveniles on the Great Barrier Reef with anchored and drifting light-traps. The results from two projects (a two-year study of abundance at cross-shelf scales and a four-year study of larval supply near two coral reefs) are described with respect to catch composition and catch rates at family level. The cross-shelf study illustrates how different taxa use the available habitat space, while the longer study reveals the temporal (interannual) variations in this fauna. Collectively, the two studies captured almost 700,000 fishes but the food fishes of highest value to the Australian market were rare, probably reflecting the representation of macro-carnivores in natural assemblages. Among these targets, only lethrinids (emperors) were common enough to warrant further interest. On the basis of abundance and catchability, the technique has potential to replace more labour-intensive and destructive methods of collecting some of the more common ornamental fishes from coral reefs, though its real value may be through research into the early life histories of wild stocks leading to improved rearing in artificial environments.

Introduction

With recent evidence that global fish catches are approaching the theoretical limit of sustainable production from the oceans (Pauly and Christensen, 1995), aquaculture and sea ranching offer alternative ways of raising total food production and repairing depleted wild stocks. Both require reliable supplies of seed and, to date, this has been a major bottleneck in mariculture. In contrast to some freshwater and estuarine species (e. g., Lates calcarifer), the larvae of truly marine fishes, including all high value tropical reef fishes (e. g., serranids, labrids, lutjanids), have proven difficult to culture in the numbers required (Lim, 1993; Sorgeloos, 1994).

One way of bypassing the husbandry problems is to capture wild fry after they have passed through the most difficult stages and then grow out juveniles. The best example of such an enterprise is the “bangus” industry of several Asian countries, most notably the Philippines, where artisanal fishers harvest annually several hundred million juvenile milkfish, Chanos chanos, from shallow coastal waters with a variety of push nets and weirs (Kumagai et al., 1980). The same approach is used for culture of groupers that are collected with attractors made of coconut fronds (Mamauag et al., 1995). A third method of harvest that offers some potential is light-attraction (Pascoe, 1990).

Pelagic juveniles of many marine fish larvae are photopositive during the night, when surface residence may assist their dispersal (Wolanski et al., 1997). These juveniles are often taken as by-catch in tuna bait fisheries using lift nets under submerged lights (Rawlinson, 1990). We have capitalised on this behaviour to develop automated light-traps as a research tool to investigate the issues of distribution, abundance and larval supply of a wide range of tropical fishes on the Great Barrier Reef (GBR). Because these passive traps depend on the behaviour of the targets, this method is selective for active swimmers (pelagic juveniles with the full complement of median fins) and is therefore a useful guide to larval supply (Milicich et al., 1992). The maturity of these animals and their live state when captured in traps facilitate identification, allowing definition of the spatial and temporal distribution and abundance of individual species.

In the context of this workshop, we have two purposes for reviewing data on larvae collected by light-traps in two sampling projects (one emphasising spatial pattern across the continental shelf and one emphasising temporal variations in larval supply at fixed locations). The first is to establish the constancy and hence reliability of larval supplies on the GBR. The second is to establish whether harvesting with light would be a viable method of concentrating the wild fry of any species suitable for mariculture. This paper summarises the results of these analyses.

Materials and Methods

All sampling was based on submerged light-traps (Figure 1) similar to those described by Doherty (1987). These traps have no moving parts and depend upon attracting photopositive organisms with slow flashing lights through two baffled chambers into a third where most of the catch remains alive until collected.

Operation of the lights is automated, which can allow arrays of traps to reveal synoptic views of spatial pattern. With daily clearance, anchored traps also can yield extensive time series of data (Milicich 1992).

Our general protocol for fixed (anchored) traps was to show light over three periods during each night (2100-2200, 2400-0100, 0300-0400), partly to conserve battery- life for extended sampling, but also partly to spread tidal and diel effects more evenly among nights of sampling (Doherty 1987). In addition, most sampling was concentrated around new moons between October and February when seasonal and lunar spawning patterns produce the largest catches of coral reef fishes (Doherty 1991).

Cross Shelf Study

In two consecutive years, we anchored light-traps immediately downstream (within 100 m) of four coral reefs located between 50 and 110 km off the coast of Townsville (Figure 2) and sampled them daily for at least a week around each of five consecutive new moon periods. In 1990/91, each reef was sampled by three traps suspended near the surface and one trap suspended at 20 m depth from the middle mooring. In 1991/92, two deep traps were suspended from the outside moorings in addition to the three surface traps (Table 1).

During the same nights, we also sampled open water habitats with free light-traps that were allowed to drift for one-hour periods at 15 stations between the coast and the Coral Sea (Figure 2). Each such release involved an independent pair of shallow and deep traps. In 1990/91, all deep drifting traps were limited to a maximum of 20 m to match those on the reef. In 1991/92, the deep traps were suspended at 5 m above the bottom or to a maximum depth of 100 m.

All sampling was done between 2000-0500H, which with transit time allowed five consecutive stations to be sampled per night. Consequently, one night transect consisted of samples from the GBR Lagoon (stations 1-5), Magnetic Passage (stations 6-10) or Coral Sea (stations 11-15). Whenever possible, we sampled each shelf station three times during a cruise. Lower replication was attempted in the Coral Sea (Table 1) due to the persistent low catches of coastal fish larvae beyond the continental margin.

Figure 2. Map of study sites

Table 1. Sampling program and fishing effort (in light-trap hours) by major habitat in the cross-shelf study

Habitat		Trap	1990/91			1991/92
	Stations	Deployment	Depth (m)	Replicates	Effort	Depth (m)	Replicates	Effort
GBR Lagoon	5	Drifting	0	2	102	0	2	119
			20	2	102	Near bottom	2	120
Magnetic Passage	5	Drifting	0	2	81	0	2	90
			20	2	81	Near bottom	2	88
Coral Sea	5	Drifting	0	2	71	0	2	40
			20	2	72	100	2	39
Reefs	3	Anchored	0	*3	724	0	3	996
			20	1	228	20	2	627

*Only two surface traps were deployed during the first cruise in October 1990.

Larval supply study

Starting in 1990/91, we monitored larval supply over four years around two reefs off Cairns (Arlington, Green) with depth-stratified anchored arrays of traps (Figure 2). All traps were cleared daily when possible (exceptions were discarded) and sampled for at least 10 days around five consecutive new moons matching those in the cross-shelf study.

Although the distribution of effort varied among years (Table 2), we tried to balance effort between depths within reefs and years, with the exception of 1993/94. However, information from three prior seasons of sampling elsewhere confirmed that the pelagic juveniles of benthic reef species are most common near the surface. We then abandoned deep-sampling in the final year in order to achieve a more-balanced effort between reefs and to provide replication at locations within reefs.

Table 2. Sampling program and fishing effort (in light-trap hours) at Arlington and Green Reefs during the larval supply study

	1990/91				1991/92				1992/93				1993/94
	Sites	Depth	Repl	Effort	Sites	Depth	Repl	Effort	Sites	Depth	Repl	Effort	Sites	Depth	Repl	Effort
Arlington	4	1	1	618	5	1	1	888	5	1	1	888	5	1	2	1221
		20	1	558		20	1	663		20	1	774
Green	3	1	1	516	3	1	1	399	3	1	1	507	5	1	2	1212
		20	1	354		20	1	258		20	1	357

Results

Cross Shelf Study

Over two spawning seasons, we collected samples of pelagic juvenile fish with >3,500 hours of light representing >1000 samples from open water and >850 samples from the leeward sides of the four coral reefs (Table 1). All samples represent collection from a site and depth combination; the difference in total hours of light between reef and non-reef habitats is explained by the longer operation of anchored traps (Table 1). To facilitate uniform comparisons, most of the following data have been standardised to catch-per-hour of light.

In total, we collected >130,000 fish consisting of >600 recognisable types belonging to >70 families with many individual distributional patterns. The strongest pattern was related to depth with the majority of families having higher catch rates near the surface (Table 2). A number of exceptions (with higher catch rates in deep traps) were encountered. Of those with sufficient numbers to be convincing, most were pelagic fishes belonging to the families Scombridae, Carangidae, Engraulidae. In each case, this pattern was driven by the abundance of one dominant genus: Rastrelliger, Decapterus, and Stolephorus, respectively. Only a few families of benthic fishes (e. g., Synodontidae, Priacanthidae, Trichiuridae) were more abundant near the bottom. Drifting traps in open water habitats collected all of these deep-dwelling taxa.

After standardisation of effort, pelagic juveniles of coral reef fishes were captured in greatest numbers by the traps anchored near coral reefs. Approximately half of these fishes were a mixture of larvae, juveniles, and adults of two species of small sprat (Spratelloides delicatulus and S. gracilis), that are known to complete their lifecycles in coral reef lagoons (Leis 1994). Although the overall abundance of these fishes remained constant across the two years (Clupeidae, Table 2), their proportional representation in shallow samples decreased from 64% in 1990/91 to 40% in 1991/92. This reduction in dominance was caused by significantly greater catches of reef species during the second year. This increase was coherent across many of the non-clupeid taxa and is illustrated here with the catch rates for the second and third most abundant taxa: the damselfish subfamilies Pomacentrinae and Chrominae (Figure 3). Whereas the former is a diverse assemblage of almost 100 species, making interpretation ambiguous, the Chrominae were dominated by a single species, Chromis atripectoralis. Thus, it appears that many species were affected in similar manner as a result of some widespread influence. Furthermore, the temporal changes were consistent for both groups on three of the four coral reefs (exception Helix Reef) spanning the full width of the reef matrix. After adjustment of the different catch rates among taxa and years, the distribution of abundance among the three reefs (excluding Helix) was remarkably consistent with greatest abundance at Keeper Reef (Figure 4).

Consistent spatial patterns were detected among most of the major taxa collected with drifting traps from open water habitats, including the low catches of shelf species from all stations beyond the continental margin mentioned earlier. We illustrate the range of patterns observed on the shelf with four examples (Figure 5). Lethrinids were most common in surface samples from the GBR Lagoon, whereas Rastrelliger was most abundant near the bottom in the same major habitat. Decapterus was also most common near the bottom but was found in greatest abundance at stations further offshore in the Magnetic Passage. Mackerels, belonging to the genus Scomberomorus, were most common in the GBR Lagoon but without such clear preference for depth. Further dissection of their catches, however, showed a difference in the size of individuals captured near the surface and bottom (Figure 6) indicating a possible ontogenetic change in the use of space.

Figure 3. Catch rates of two subfamilies of reef fish at four coral reefs.

Figure 4. Relative distribution of the two subfamilies among the four reefs. Shading patterns as in the previous Figure 3.

Larval Supply Study

Over four spawning seasons, we collected >500,000 pelagic juvenile fish (Table 4) with >9,000 hours of light (>3,000 samples) from multiple traps anchored close to the two reefs. Because of the lack of “deep” effort in the final year and the much lower catches near the bottom, we focus here only on fishes caught near the surface.

As in the cross-shelf study, temporal trends were strong. Catches at individual traps varied among days, months, and years. We have pooled the first two scales (as representing only patchiness and spawning cycles) and report the interannual variation as the factor most likely to affect population dynamics and any attempts to harvest natural supplies of fish fry.

In contrast to the observations on the four reefs off Townsville, there was no increase between 1990/91 and 1991/92 in the CPUE of the Pomacentrinae or Chrominae on either reef (Tables 5, 6). In the third season, collections from both reefs were dominated (~70%) by an influx of Pomacentrinae but Chrominae did not change in a similar way suggesting that the patterns were driven at the level of species. This pattern of independent changes was repeated among the dominant non-clupeid taxa within and between reefs (Tables 5, 6). The blue sprat (Spratelloides delicatulus) was again captured in great numbers (49% of total catch on Arlington - rank 1; 32% of total from Green - rank 2). Catch rates of this species varied by more than an order of magnitude among years but without coherence among the two adjacent populations.

We attribute these differences to local influences given their closed life cycles.

In contrast to the incoherent temporal patterns among taxa within and between reefs, one spatial effect was obvious besides the previously reported effect of depth. Catch rates for the 10 most abundant species were all greater from traps set near Arlington than Green, though this may reflect their different positioning around the two reefs. While such positional effects can determine the absolute catch rates of individual taxa, the most desirable targets (species with high unit value attractive to commercial fishers) were always rare, which simply may reflect their natural representation in reef communities

Discussion

A comprehensive understanding of the ecology of presettlement tropical fishes has been elusive in the past; partly because of the high costs of sampling this notoriously patchy fauna and also partly because of a failure of suitable methods. Traditional methods of sampling plankton, which rely upon towed nets, may capture large numbers of individuals but their catches are biased towards minute preflexion (no tail formation) stages that are difficult to identify and clearly are not viable to culture. Until the advent of alternative sampling methods, the scarcity of the more mature and robust stages could be attributed to extremely high rates of natural mortality in the ocean, but the deployment of light-traps (Doherty 1987) in the same water has shown that the larger stages are agile swimmers that simply avoid conventional gears (Choat et al., 1993).

Light-traps were developed as tools to study spatial and temporal variations in the supply of pelagic juveniles to coral reef habitats and their catches have been shown to correlate well with the replenishment of benthic populations (Milicich, 1992). This should not be surprising because the efficiency of these passive devices depends upon the active orientation (photopositive behaviour and swimming ability) of the targets. Lack of either capacity means that light-traps will be selective, obviously undersampling fish that are not attracted to light (taxonomic selectivity) and feeble swimmers regardless of their attraction (size-selective bias).

The inherent size-selectivity of light-traps is an advantage for certain studies. Since most of the fish caught are large, agile swimmers with fully-developed fins, much of the catch can be identified to species level, which allows better ecological study. Since we still have a rudimentary knowledge of the early life histories of most tropical fishes, better field studies could assist with attempts to close their lifecycles in hatcheries by defining the dietary and water quality requirements of natural populations. In the short-term, there is potential also to consider light-trapping as an alternative method for collecting wild fry since the majority of fish captured by this technique are alive when collected and robust enough to be candidates for mariculture.

Table 3. Depth-stratified catch rates (and relative abundance) from the cross-shelf study. Percentages of <0.01 after rounding were suppressed to increase readability of the table.

Families	1990/91				1991/92				Total
	Shallow		Deep		Shallow		Deep		Shallow		Deep
Clupeidae	27.59	(63.49)	2.67	(32.57)	24.26	(39.84)	4.48	(31.00)	25.72	(48.29)	3.83	(31.38)
Pomacentrinae	8.71	(20.05)	0.56	(6.79)	21.74	(35.70)	0.44	(3.06)	16.03	(30.11)	0.48	(3.95)
Chrominae	1.2	(2.77)	0.04	(0.43)	6.42	(10.54)	0.2	(1.40)	4.13	(7.76)	0.14	(1.17)
Apogonidae	1.82	(4.20)	0.45	(5.42)	2.46	(4.03)	0.98	(6.81)	2.18	(4.09)	0.79	(6.48)
Blenniidae	0.48	(1.10)	0.07	(0.88)	1.48	(2.44)	0.17	(1.18)	1.04	(1.96)	0.14	(1.11)
Monocanthinae	0.56	(1.28)	0.16	(1.92)	0.83	(1.37)	0.12	(0.83)	0.71	(1.34)	0.13	(1.09)
Atherinidae	0.56	(1.29)			0.64	(1.05)	0.003	(0.02)	0.6	(1.13)	0.002	(0.02)
* Scombridae	0.36	(0.82)	1.37	(16.75)	0.56	(0.92)	1.85	(12.83)	0.47	(0.89)	1.68	(13.77)
Gobiidae	0.14	(0.32)	0.15	(1.87)	0.66	(1.08)	0.29	(1.98)	0.43	(0.80)	0.24	(1.95)
* Carangidae	0.25	(0.58)	0.74	(9.01)	0.25	(0.41)	1.36	(9.42)	0.25	(0.47)	1.14	(9.32)
Holocentridae	0.4	(0.92)	0.08	(0.93)	0.1	(0.16)	0.002	(0.02)	0.23	(0.43)	0.03	(0.23)
Nomeidae	0.24	(0.55)	0.008	(0.10)	0.16	(0.26)	0.008	(0.06)	0.19	(0.36)	0.008	(0.07)
Myctophidae	0.33	(0.76)	0.02	(0.30)	0.04	(0.06)	0.02	(0.12)	0.17	(0.31)	0.02	(0.16)
Engraulidae	0.06	(0.15)	0.41	(4.97)	0.2	(0.32)	2.18	(15.09)	0.14	(0.26)	1.55	(12.67)
Bregmacerotidae	0.05	(0.11)	0.3	(3.660	0.18	(0.29)	0.17	(1.20)	0.12	(0.23)	0.22	(1.78)
Labridae	0.12	(0.28)	0.04	(0.43)	0.13	(0.21)	0.01	(0.09)	0.12	(0.23)	0.02	(0.17)
* Lethrinidae	0.12	(0.29)	0.008	(0.10)	0.12	(0.19)	0.02	(0.13)	0.12	(0.23)	0.01	(0.12)
* Siganidae	0.06	(0.13)	0.06	(0.68)	0.15	(0.24)	0.12	(0.86)	0.11	(0.20)	0.1	(0.82)
Scorpaeninae	0.04	(0.09)	0.004	(0.05)	0.11	(0.19)	0.009	(0.06)	0.08	(0.15)	0.007	(0.06)
Synodontidae	0.04	(0.09)	0.26	(3.20)	0.09	(0.14)	0.46	(3.19)	0.07	(0.12)	0.39	(3.20)
* Mullidae	0.05	(0.11)	0.09	(1.06)	0.05	(0.08)	0.02	(0.15)	0.05	(0.09)	0.04	(0.37)
Syngnathidae	0.06	(0.15)	0.004	(0.05)	0.02	(0.03)	0.01	(0.07)	0.04	(0.07)	0.008	(0.07)
Chaetodontidae	0.01	(0.03)	0.004	(0.05)	0.04	(0.06)	0.006	(0.04)	0.03	(0.05)	0.005	(0.04)
Dactylopteridae	0.02	(0.04)	0.008	(0.10)	0.01	(0.02)	0.002	(0.02)	0.02	(0.03)	0.004	(0.04)
* Nemipteridae	0.003	(0.01)			0.03	(0.05)	0.002	(0.02)	0.02	(0.04)	0.001	(0.01)
Pseudochromidae	0.01	(0.03)	0.02	(0.28)	0.03	(0.05)	0.37	(2.57)	0.02	(0.04)	0.25	(2.02)
Priacanthidae	0.03	(0.06)	0.23	(2.83)	0.01	(0.02)	0.41	(2.84)	0.02	(0.04)	0.35	(2.84)
* Caesionidae	0.04	(0.10)	0.04	(0.48)	0.006	(0.01)	0.09	(0.63)	0.02	(0.04)	0.07	(0.60)
Plesiopidae	0.005	(0.01)	0.03	(0.40)	0.01	(0.02)	0.003	(0.02)	0.01	(0.02)	0.01	(0.11)
Sphyraenidae	0.02	(0.05)			0.003	(0.01)	0.003	(0.02)	0.01	(0.02)	0.002	(0.02)
* Serranidae	0.009	(0.02)	0.006	(0.08)	0.01	(0.02)	0.03	(0.21)	0.01	(0.02)	0.02	(0.18)
Schindleriidae	0.001		0.02	(0.20)	0.02	(0.04)	0.17	(1.18)	0.01	(0.03)	0.12	(0.95)
Pseudoplesiopidae	0.004	(0.01)	0.01	(0.13)	0.01	(0.02)	0.006	(0.04)	0.008	(0.01)	0.007	(0.06)
Canthigasterinae	0.003	(0.01)			0.009	(0.01)			0.006	(0.01)
Microdesmidae			0.002	(0.03)	0.01	(0.02)	0.23	(1.60)	0.005	(0.01)	0.15	(1.22)
Amphiprioninae	0.003	(0.01)			0.006	(0.01)			0.005	(0.01)

Table 3 cont'd.

Families	1990/91				1991/92				Total
	Shallow		Deep		Shallow		Deep		Shallow		Deep
Eleotridae	0.001				0.006	(0.01)	0.002	(0.02)	0.004	(0.01)	0.001	(0.01)
Acanthuridae	0.001		0.01	(0.13)	0.006	(0.01)	0.04	(0.27)	0.004	(0.01)	0.03	(0.23)
Grammistidae	0.007	(0.02)	0.004	(0.05)	0.002		0.003	(0.02)	0.004	(0.01)	0.004	(0.03)
Gerreidae	0.005	(0.01)	0.002	(0.03)	0.002		0.009	(0.06)	0.004	(0.01)	0.007	(0.05)
Scaridae					0.005	(0.01)			0.003	(0.01)
Diodontidae	0.004	(0.01)			0.002				0.003	(0.01)
Hemiramphidae	0.007	(0.02)							0.003	(0.01)
Balistinae	0.001				0.002				0.002
Bothidae	0.001		0.006	(0.08)	0.002		0.002	(0.02)	0.002		0.004	(0.03)
Istiophoridae	0.004	(0.01)	0.002	(0.03)					0.002		0.0007	(0.01)
Teraponidae	0.002	(0.01)			0.002				0.002
Muraenidae	0.001				0.002				0.002
* Lutjanidae	0.003	(0.01)	0.03	(0.40)	0.002		0.006	(0.04)	0.002		0.02	(0.13)
Ostraciidae	0.002	(0.01)			0.002				0.002
Pempheridae			0.01	(0.15)	0.003	(0.01)	0.04	(0.26)	0.002		0.03	(0.23)
Anguillidae	0.001				0.002				0.001
Gobiesocidae	0.003	(0.01)	0.002	(0.03)					0.001		0.0007	(0.01)
Ophichthidae	0.002	(0.01)			0.0008				0.001
Exocoetidae	0.001				0.0008				0.001
Pomacanthidae	0.001		0.02	(0.20)	0.0008		0.005	(0.03)	0.001		0.009	(0.07)
Bythitidae					0.002				0.001
Tetraodontinae	0.003	(0.01)	0.006	(0.08)			0.001	(0.01)	0.001		0.003	(0.02)
Uranoscopidae	0.001				0.0008		0.001	(0.01)	0.001		0.0007	(0.01)
Cirrhitidae					0.0008				0.0005
Congrogadidae					0.0008				0.0005
Trichiuridae			0.24	(2.93)	0.0008		0.06	(0.42)	0.0005		0.12	(1.02)
Ephippidae					0.0008				0.0005
Kyphosidae	0.001								0.0005
Pseudogrammidae							0.001	(0.01)			0.0007	(0.01)
Plotosidae							0.001	(0.01)			0.0007	(0.01)
Leiognathidae			0.008	(0.10)			0.003	(0.02)			0.005	(0.04)
Centriscidae			0.002	(0.03)							0.0007	(0.01)
Congridae							0.001	(0.01)			0.0007	(0.01)
Gempylidae			0.002	(0.03)			0.003	(0.02)			0.003	(0.02)
Mugiloididae							0.001	(0.01)			0.0007	(0.01)
Total fish	41843		3963		75268		12618		117111		16581

* families of maximum interest as sources of food

We have grown-out various fish caught with traps where their identification were ambiguous at the time of collection. For example, the larval supply study described above was done for the purpose of monitoring the replenishment of the most valuable fish taken by commercial and recreational fishers on the Great Barrier Reef - coral trouts belonging to the genus, Plectropomus (Doherty et al. , 1994). During our study, we collected approximately 650 individuals from this species complex but all looked alike when collected due to their intense red pigmentation. During the course of the project, we transferred several hundred live individuals to culture facilities on the mainland, where, after initial problems with water quality, some were grown-up to one-year-old juveniles confirming that the catches were dominated by Plectropomus leopardus. It was this experience that raised our expectations that wild fry collected by light attraction might be used to circumvent the technical problems that have been encountered in closing the life cycles of high value reef fishes in the laboratory. Although our intention has never been to substitute for artificial mass culture, it raises the question of whether the fry of any species attractive to mariculture is so abundant in the wild that it could be harvested in numbers to support a growing industry.

To answer this question, we have summarised the results of two sampling exercises that collected almost 700,000 fish. The first project sampled fish across the continental shelf off Townsville in habitats both near and far from coral reefs. Our reason for this study was to identify the use of habitat space by a wide range of reef and coastal fishes. In the limited number of examples presented here, we have shown that different species may have unique requirements, some of which may also vary during the ontogeny of that species. Such information allows future sampling to be better targeted to places of maximum abundance. The second project sampled fish from near two coral reefs, where our intention was to monitor temporal variations in larval supply in order to examine coupling between the availability of propagules and the replenishment of local populations. From the perspective of this conference, which is concerned with the culture of marine fishes, it provides data on the availability of attractive targets among years.

Figure 5. Catch rates from the three major shelf habitats (GBR Lagoon, Magnetic Passage, coral reefs) of four taxa with contrasting distributions. Shading identifies catch rates from shallow (unfilled) and deep (filled) traps respectively. The horizontal lines identify back-transformed catch rates of one (unbroken) and 10 (broken) fish per hour of effort.

Figure 6. Size-frequency distributions of mackerels in the genus, Scomberomorus, in deep and shallow samples from the Great Barrier Reef Lagoon.

While recognising that the GBR is not necessarily a model for other countries, our results can be summarised briefly here for the purpose of this conference. Our cross-shelf study was completed at considerable cost to the Australian Government because it covered deep and shallow habitats up to 150 km from the coast. While it revealed a lot of new information about the distribution and abundance of the early life history stages of a wide range of coastal fishes, it did not reveal an untapped source of juveniles suitable for cost-effective culture. The possible exception is that of the lethrinids (emperors), which were most abundant in surface waters near the coast. The data that we have shown may not be particularly convincing because we caught fewer than 300 individuals from around 100 hours of light-trapping in the GBR Lagoon though returns were considerably higher from specific locations. In a third season of sampling in this habitat (data not reported here), catch rates of lethrinids were approximately 4 times greater. This example is meant to make two points. First, catch-per-unit effort can be maximised by focussed effort (a fact long exploited by fishermen with their traditional knowledge of the habits of fish) but it will vary among spawning seasons due to a range of environmental influences on larval transport and survival.

The cross-shelf study, sampling both near and far from coral reefs, showed that the most effective place to collect the pelagic juveniles of reef-associated species is at the surface in the lee of coral reefs (Doherty and Carleton in press). We believe that this concentration factor arises from the combination of hydrodynamic forces and the active orientation of presettlement reef fishes (Wolanski et al., 1997). We also believe, on the preliminary data presented here, that the size of the reef has an effect on this concentration factor. The evidence affecting this belief is the poor and erratic larval supply sampled at Helix Reef (the smallest and most isolated reef in our sampling) in the cross-shelf study against the greater catch rates from traps placed downstream of the huge and extensive Arlington Reef. The latter extends across most of the outer shelf off Cairns. Such an effect is consistent with the lower catch rates observed at the neighbouring Green Reef, although without replication, we cannot be certain that the larger reef was not starving its neighbour of recruits by its upstream position. In any case, we suggest that large reefs will be the best places to seek pelagic propagules in useful concentrations, though we caution that this result may pertain only to reefs subject to strong hydrodynamic flows like those of the GBR.

Accepting that reefs with the characteristics of Arlington may be optimal places from which to harvest the fry of valuable reef fishes, we still conclude that such an enterprise would be a risky business. During our four-year study at Arlington, the catches of the high-value coral trouts ranged from 339 to 27, with the least number of juveniles collected during the year of greatest effort (Table 2). Interannual variability in larval supply comes as no surprise to students of fish population dynamics but, at such levels, it poses an unacceptable risk to any commercial venture. While this risk could be spread by not concentrating on a single species, we note that all of the potential target species that is considered most valuable in Australia (mostly fish of high trophic status) were rare. This may reflect nothing other than the natural representation of macrocarnivores in reef fish communities but, separate from the economic considerations, conservation imperatives in Australia would rule such harvesting out of the question because of the extensive by-catch.

Based on abundance alone, it would make more sense to exploit the resource for ornamental reef fishes (Andrews 1990) and this may be a valuable niche market. Though preliminary trials with pomacentrids suggest that their larviculture is not particularly challenging, due to the large size of the hatchlings emerging from demersal eggs. The experience of a few companies that have bred tropical marine fishes for the aquarium trade is that their product is uncompetitive and commands low prices due to the drab colours of cultured juveniles compared with those from wild stocks (Anon., 1992). These factors explain why the aquarium trade in marine fishes remains almost totally reliant on the collection of individuals from the wild despite the technical capacity to culture some species. While the unit values of common species are relatively low, the catch rates that are possible through light-attraction (>30,000 damselfishes in one night, Milicich, 1992) suggest that such harvesting may be more effective and less destructive than present techniques for collecting the same species from reef habitats.

In the case of food fishes, after a vast amount of effort and a substantial number of dead fish, we can look only to two families (Siganidae, Lethrinidae) with any interest for this audience. The rabbitfishes are of little interest at present to the Australian market but the recent successes with their mass- culture in Asia challenges the need for expanded harvest of their wild fry. Until a real shortage of reef fishes places a higher unit value on alternative species, we can point only to the catch rates of emperors (family: Lethrinidae) as a possible cost-effective target for the harvest of wild fry. These fish (attractive to the Australian market) have been harvested in large numbers in several of our sampling projects (often as much as 25% of the catch - see also Milicich and Doherty, 1994), have proven to be resilient in captivity and have shown excellent growth rates.

Acknowledgement

We thank the crews of the RV Lady Basten for logistic support; John Collingwood for trap manufacture; Helen Sturmey, Kim Smith, Chad Lunow for sorting thousands of samples; David Harris, Phillip Light, and volunteers for maintaining the extended sampling in the Cairns study; Tara Anderson for the preliminary analysis of data from the cross-shelf study. This is AIMS Contribution no. 879.

Literature Cited

Andrews, C. 1990. The ornamental fish trade and fish conservation. J. Fish. Biol. 37: 53-59.

Anon., 1992. Marine ornamental fish: world trade. Aqua. Farm News X (1): 1-5.

Carleton, J. H. and P. J. Doherty. 1997. The distribution and abundance of pelagic juvenile fish near Grub Reef, central Great Barrier Reef. In: Lession, H. A. and I. G. MacIntyre. Proc. 8th Int. Coral Reef Symp. 1:1155-1160.

Choat, J. H., P. J. Doherty, B. A. Kerrigan, and J. M. Leis. 1993. A comparison of towed nets, purse seine and light aggregation devices for sampling larvae and pelagic juveniles of coral reef fishes. Fish. Bull. 91:195-209.

Doherty, P. J. 1987. Light-traps: selective but useful devices for quantifying the distributions and abundances of larval fishes. Bull. Mar. Sci. 41:423-431.

Doherty, P. J. 1991. Spatial and temporal patterns in recruitment. pp. 261-293 In: Sale, P. F. (ed.). The ecology of coral reef fishes. NY: Academic Press.

Doherty, P. J., A. J. Fowler, M. A. Samoilys, and D. A. Harris. 1994. Monitoring the replenishment of coral trout (Pisces: Serranidae) populations. Bull. Mar. Sci. 54: 343-355.

Kumagai, S., T. Bagarinao, and A. Unggui. 1980. A study on the milkfish fry fishing gears in Panay Island, Philippines. SEAFDEC Tech. Rep. 6.

Leis, J. M. 1994. Coral Sea atoll lagoons: Closed nurseries for the larvae of a few coral reef fishes. Bull. Mar. Sci. 54:206-227.

Lim, L. C. 1993. Larviculture of the greasy grouper, Epinephelus tauvina F., and the brown-marbled grouper, E. fuscoguttatus F., in Singapore. J. World Aquacult. Soc. 24:262-274.

Mamauag, S., L. Penolio, P. Alino, and G. Russ. 1995. Juvenile groupers in the Philippines: gaining insights into their recruitment patterns. Third Natl. Symp. Mar. Sci., Philippine Assoc. Mar Sci.

Milicich, M. J. 1992. Light-traps: a novel technique for monitoring larval supply and replenishment of coral reef fish populations. Unpubl. Ph.D. Thesis, Griffith University, Australia. 127 pp.

Milicich, M. J., P. J. Doherty. 1994. Larval supply of coral reef fish populations: magnitude and synchrony of replenishment to Lizard Island, Great Barrier Reef. Mar. Ecol. Prog. Ser. 110:121-134.

Milicich, M. J., M. G. Meekan, and P. J. Doherty. 1992. Larval supply: a good predictor of recruitment of three species of reef (Pomacentridae). Mar. Ecol. Prog. Ser. 86: 153-166.

Pascoe, P. L. 1990. Light and the capture of marine animals. Cambridge Uni. Press.

Pauly, D. and V. Christensen. 1995. Primary production required to sustain global fisheries. Nature 374: 255-257.

Rawlinson, N. J. F. 1990. Catch composition of the tuna bait-fishery of Solomon Islands and possible impact on non-target species. ACIAR Proc. 30: 190-211.

Sorgeloos, P. 1994. State of the art in marine fish larviculture. World Aqua. 25: 34-37.

Wolanski, E., P. J. Doherty, and J. H. Carleton. 1997. Directional swimming by fish larvae determines connectivity of the Great Barrier Reef. Naturwissenshaften 84:262-268

---

[1] Australian Institute of Marine Science, PMB No. 3, Townsville MC Q4810, Australia