Western Australia has an extensive coastline ranging from the tropical north to the temperate south, with a high diversity of neritic marine fishes. Distribution data of 1855 neritic fish species were used to compare a range of methods for identifying priority areas for their conservation. Species richness and endemism richness hotspots, biogeographic zoning and complementarity analysis were tested for their efficiency at representing the total suite of species. The hotspot approaches demonstrated low efficiency, as the sections of coastline selected were grouped together in isolated geographic locations and relatively few of the total suite of neritic fish species were represented. Biogeographic zoning divided the coast of Western Australia into six regions, and as such, priority areas were selected around the entire coastline, which resulted in fairly high levels of efficiency. However, the complementarity analysis proved to be the most efficient method, as >95% of all neritic fish species could be represented in six, appropriately located, 100km long sections of coastline. Complementarity analysis indicated a total of 26 priority areas for neritic fish conservation, which were spread around the entire coast of Western Australia. However, as current marine conservation measures in Western Australia are focused on the west and northwest coasts, this study highlighted the need for marine conservation efforts to be extended to cover the north and south coasts.

Ages were estimated for the tropical sparid, Acanthopagrus berda, caught in northern KwaZulu-Natal estuaries. Whole otoliths were used in the age determination. Age estimates were validated by marginal increment analysis and oxytetracycline labelling, which indicated that opaque deposition occurs primarily from September to November each year. The reproducibility of age estimates was described by a coefficient of variation of 10%. The von Bertalanffy growth curve was found to best describe the growth of A. berda. The parameters of the von Bertalanffy growth curve indicated that A. berda in northern KwaZulu-Natal is a slow-growing species, capable of reaching at least 16 years of age. Longevity of the species, coupled with sex change, late maturation and estuarine dependency, give cause for concern for the continued sustainable utilization of this species.

The recreational linefisheries in Durban Harbour and the Mgeni Estuary were surveyed using roving creel and access-point surveys during the period January–December 2000. In total, 3 351 shore-anglers and 652 boat-anglers were checked for catch-and-effort information, and 432 shore-anglers were interviewed using a demographic and socio-economic questionnaire. Durban Harbour had much higher angling effort than the Mgeni Estuary. Angling activity was higher on weekends (121 and 23 anglers per count for the harbour and Mgeni Estuary respectively) than during the week (34 and 9 anglers per count respectively). Total effort expended in Durban Harbour and the Mgeni Estuary shore-fisheries during 2000 was estimated to be 54 024 and 11 977 angler-outings respectively. Annual effort for the Durban Harbour boat-fishery for the same period was estimated to be 9 991 angler-outings. The flathead mullet Mugil cephalus was the most commonly harvested species in both Durban Harbour and the Mgeni Estuary shore-fisheries (25.2 and 68.1% respectively), and spotted grunter Pomadasys commersonnii (41.5%) was the most commonly harvested species in the Durban Harbour boat-fishery. Because of their small size, a large proportion (>63%) of the total fish catch at both localities was released. Mean catch per unit effort for the Durban Harbour boat-based fishery (0.11 fish angler−1 h−1 or 0.15 kg angler−1 h−1) was higher than that of the shore-based fishery (0.071 fish angler−1 h−1 or 0.034 kg angler−1 h−1). The questionnaire survey indicated that anglers had high site fidelity, considerable years of fishing experience (mean of 17.2 years for the harbour, 13.6 years for Mgeni) and high compliance with possession of fishing permits (86% harbour, 84% Mgeni). Although anglers generally supported the regulations currently applicable to the linefishery (>71% harbour, >77% Mgeni), the questionnaire results showed that specific knowledge for target species was poor (45–70% harbour, 29–60% Mgeni). The level of fisheries law enforcement (6.4% of harbour outings inspected, 7% Mgeni outings inspected) was poorer than in the previously studied KwaZulu-Natal marine shore-fishery. Economic investment by participants in terms of angling equipment used in the shore-fisheries of the two systems was calculated to be >R10 million, and expenditure in terms of bait, travel and tackle costs was approximately R9 million per year.

A robust, perspex light-trap designed for capture of settlement stage coastal fishes is described. Trials with this light-trap were conducted in the entrance channel to Durban harbour over four nights, each a week apart during September and October 1996. Larvae and settlement stage fishes of 22 teleost families representative of the ichthyofauna of the study area were captured. Clupeidae (72.3%), Tripterygiidae (7.9%), Gobiidae (7.7%), Blenniidae (5.0%) and Sparidae (3.9%) contributed most of the catch. Although variable, in comparison to other international studies, mean catch rate was high at 241 fish per hour. Excluding clupeoids, mean catch rate was 67 fish per hour with Sparidae, a target group for future recruitment studies, contributing a mean catch rate of 14 fish per hour. Construction of the trap was cost-effective and it proved to be robust and reliable over the trial period.

The coast of South Africa is subject to a range of oceanographic conditions and habitats, and this translates into a wide diversity of fishes. Indigenous people knew of, and used, some of these fishes long before the arrival of explorers and naturalists from Europe, who subsequently described many of the species in the scientific literature. The appointment of J.D.F. Gilchrist as the marine biologist to the Cape Colony in 1895 started a tradition of ichthyology in South Africa that was fostered by K. H. Barnard at the South African Museum. The discovery of the coelacanth in 1938 put South African ichthyology firmly on the world map, and J.L.B. Smith produced the first edition of the remarkable volume Sea Fishes of Southern Africa in 1949. Subsequently, in 1986, with the assistance of numerous local and international contributors, M. M. Smith and P. H. Heemstra compiled Smiths' Sea Fishes, which documented >2000 species from the region. With the information in this volume as a basis, the coastal and shelf ichthyofauna of South Africa is examined at a biogeographic level. Diversity and endemism are discussed and complementarity analysis used to address the suitable location of marine protected areas to conserve the ichthyofauna of South Africa.

We describe the larval development of the sparid Pagellus natalensis based on 34 field specimens of 2–19 mm from the western Indian Ocean off South Africa. P. natalensis has unusual larval morphology for a sparid: in particular, extensive head spination including a serrate supraoccipital crest, serrate supraocular, pterotic, and supracleithral ridges, interopercular spines, strong preopercular spination, and pitted frontal bones. Development in the family Sparidae is briefly reviewed, and larval development of P. natalensis is central to this review. We suggest that larval morphology reveals much about sparid relationships, and pose several hypotheses for further testing. These include: Pagellus, Pagrus and Dentex as currently conceived are polyphyletic; Pagellus affinis, bellottii and natalensis and Pagrus pagrus are closely related, as is an unidentified larval sparid from the eastern north Atlantic; spiny larval morphology is derived within the Sparidae and indicates that the currently recognized subfamilies are polyphyletic; Spondyliosoma is monophyletic; Calamus and Spondyliosoma are sister taxa.

The southern African coastline is dominated by strong currents. Along the eastern seaboard, the warm western boundary Agulhas Current sweeps close inshore along the shelf edge before diverging from the coast on the Agulhas Bank and retroflecting back into the Indian Ocean. On the western seaboard, strong jet currents develop in the southern Benguela, associated with the strong thermal gradients induced by upwelling and Agulhas Current intrusions and eddies. There is, in general, northward drift of surface waters in the Benguela Current with strong offshore losses in the vicinity of an exceptionally active upwelling region off LÜderitz. Several potent mechanisms exist for offshore dispersal and loss from the productive shelf waters, such as eddies, filaments, retroflections and offshore Ekman drift, which pose special problems for successful retention of planktonic eggs and larvae from broadcast spawners. Most fish species in southern Africa have evolved highly selective reproductive patterns, which ensure that sufficient progeny are retained or can enter the nursery grounds along the coastline. Four important reproductive habitats, comprising spawning areas, transport mechanisms and nursery grounds, occur between Moçambique and Angola. These are used by a wide variety of pelagic, demersal and inshore-dwelling fish species.

St Lucia is one of the largest estuarine systems in Africa, and attracts thousands of anglers each year. Catch card data from the National Marine Linefish System for the years 1986-1999 were analysed to determine catch composition, catch per unit effort (cpue) and seasonality of catches by recreational anglers. Because not all anglers completed catch cards, estimates of total catch were made using additional data on the number of private boat outings, the number of boat trailers at boat slipways and the number of boats recorded entering the campsite gates during 1992 and 1993. In all, 27 fish families, constituting 55 species, were recorded by recreational anglers. Dusky kob Argyrosomus japonicus, spotted grunter Pomadasys commersonnii, perch Acanthopagrus berda, Natal stumpnose Rhabdosargus sarba, springer Elops machnata and mini-kob Johnius dorsalis were the most prominent species caught in terms of numbers and mass. Catch rates expressed numerically (fish angler(-1) h(-1)) peaked during the summer and early winter. However, in terms of mass, catch rates peaked during late winter and spring, when there were increased landings of large dusky kob. Fluctuations in cpue were linked to salinity and estuary mouth conditions (i.e. mouth closure). Despite annual fluctuations in cpue, regression analysis revealed an overall downward trend for the dominant species (dusky kob and spotted grunter) and a gradual increase for stumpnose, perch and springer, but with the exception of stumpnose, these trends were not significant. Socio-economic aspects of the fishery were also investigated by conducting an independent boat-angler survey. The value of the recreational fishery, in terms of accommodation and direct angler expenditure, was estimated to be in the region of R9 million during 1992. Angler attitudes towards fishing regulations were positive and anglers generally had a good knowledge of the regulations for target species. Based on this assessment, a number of suggestions are made regarding the future management of the recreational fishery at St Lucia.

Age and growth of Rhabdosargus sarba, sampled along the KwaZulu-Natal coast was estimated by counting the number of growth rings in sectioned sagittal otoliths read under transmitted light. The reproducibility of otolith readings was described by the average percentage error index, which was equal to 4.7%. Validation of the annual periodicity of growth zone deposition was achieved through marginal zone analysis and oxytetracycline labeling of captive fish. The age-length data from the sample of 216 fish was best described by the special Von Bertalanffy growth function of the form:L-t = 715 mm FL (1 - e(-0.16/year(t+0.996 years)))Like most sparids, R. sarba is a slow growing, long-lived species. The study revealed a maximum age of 16+ years and, based on previously published data, the age-at-maturity was estimated to be 1.8 years. (C) 2002 Elsevier Science B.V. All rights reserved.

The Kosi estuarine system, located in northern KwaZulu-Natal, is becoming an increasingly popular angling venue. Recreational catch card data from the National Marine Linefish System for the years 1986 to 1999 were analysed to determine total catch, catch composition, catch per unit effort (CPUE), seasonality of catches and annual trends in catch composition. A total of 17 families, comprising 34 species, was reported in catches of recreational anglers. Pomadasys spp., predominantly Pomadasys commersonni was the most prominent taxon caught in terms of both numbers and mass. Seasonal trends in catch rates were observed, with more fish caught in the summer and autumn months. The catch rate fluctuated annually. Fluctuations in CPUE were closely correlated to the number of cards collected, which reflects bias in the data. Despite annual fluctuations in CPUE, regression analysis revealed a slight increase in overall CPUE with time. However, CPUE for Acanthopagrus berda and Lutjanus argentimaculatus decreased steadily since 1986. The decline of these two estuarine-dependent species may indicate that the total fishing levels are not sustainable.

Age and growth of the endemic, reef-dwelling sparid, Polysteganus undulosus were determined by using archived otoliths collected in 1962 and 1963, before the fishery collapsed. Estimates of age ranged up to 20 years, but reproducibility of the age estimates was low because of stacking on the otolith margins in older fish. Validation of annual deposition of the opaque band indicated that active deposition occurred during late winter and spring. No difference in growth rate between the sexes was observed and the growth rate for the combined sexes was described by the following logistic equation: Lt=942mmTL/(1+exp(−0.277year−1(t−5.178years))). The age-at-50% maturity was 8.8 years. A retrospective stock assessment revealed that, although in the early 1960s, the yield-per-recruit was optimal, the spawning biomass of P. undulosus was already at 25% of its unfished level. This study thus suggests that management strategies for slow-growing, long-lived, aggregating reef fishes should go beyond optimising yield alone and pay particular attention to the status of the spawning stock.

The structure of the larval fish assemblages along an ocean-estuarine gradient in the St Lucia region on the northern KwaZulu-Natal coast of South Africa was examined using a combination of univariate, distributional and multivariate techniques. The data was comprised of a full annual set of ichthyoplankton samples taken from three types of environment: nearshore coastal waters, surf zone and within the St Lucia Estuary itself. The mean monthly densities of each species in each environment were used in the species matrix, and the mean monthly values of salinity, temperature and turbidity were used in the physical variables matrix. The mean species diversity and eveness index was significantly higher in the nearshore waters than the surf zone and estuary. The patterns of relative species abundances in each environment (K-dominance curves) showed that the estuarine environment was dominated by a few species in large numbers, the surf zone was intermediate, and the nearshore coast was the most diverse. Classification and multidimensional scaling (MDS) ordination analyses of larval fish densities grouped together into three main clusters based on the three different environments. The species similarity matrix (inverse analysis) clustered into four groups at the 10% similarity level. The MDS analysis of the same matrix showed that the groups separated out more or less according to the type of environment they occur in, and hence the level of estuarine dependence of the various species. Species belonging to each assemblage showed similarities with regards to their reproduction modes and/or preference to a particular physical condition. Some species were restricted to one environment, whilst others were common to two or all three environments. The occurrence of partially estuarine-dependent species in all three environments suggests that ocean-estuarine coupling is an important process for the recruitment success of these species. The ‘best fitting’ physical variable which explained the larval fish community patterns was turbidity on its own, with the relationship of larval densities to the physical variables being species-specific. The present study indicates that a number of factors must play a role in determining the structure and composition of larval fish assemblages occurring in different types of environments along an ocean-estuarine gradient. It is suggested that these distinct assemblages might be considered indicators for the different environments which they inhabit.

Fishes occurring in a lower shore tidal pool near Durban, South Africa, were collected using rotenone at monthly intervals for a year. Eighteen teleost families and 50 lower taxa of fishes were recorded. Cryptic Tripterygiidae, Blenniidae and Gobiidae were particularly abundant while Pomacentridae, Acanthuridae, Serranidae, Sparidae and Labridae were recorded in lower numbers. Resident intertidal fishes recolonized the pool between monthly visits, with tripterygiids and gobies showing evidence for year–round recruitment of juveniles. Juveniles of more transient Indo-Pacific species recruited primarily during the summer months.

Spatial and temporal distribution patterns of scombrid larvae along the east coast of South Africa were investigated from ichthyoplankton collections made during May-June 1990 (winter), October 1990 (spring) and February 1991 (summer). Results were analysed in relation to oceanographic conditions and known spawning localities of tuna and mackerels in the western Indian Ocean. In total, eight species were represented in the samples, with highest diversity in February and lowest numbers in May June. Larvae of the temperate chub mackerel Scomber japonicus were most abundant at shelf stations during October. Larvae of neritic tunas Auxis sp. and Euthynnus affinis occurred in shelf stations off KwaZulu Natal in February and extended southward in a plume along the shelf edge. Larvae of skipjack tuna Katsuwonus pelamis were most abundant in the Agulhas Current during February. Only a few larvae of oceanic tunas Thunnus spp., wahoo Acanthocybium solandri and king mackerel Scomberomorus commerson were collected in the Agulhas Current in the north of the study area during February when there was an intrusion of warm Tropical Surface Water. This indicates that spawning of these species probably does not occur off the east coast of South Africa.

Sardines have been a major contributor to the pelagic fishery off southern Africa for several decades. The South African and Namibian stocks are separate, and seasonal hydrology, in conjunction with age-related behaviour, produces the observed size-distribution patterns. Sardines are relatively fast-growing and spawn repeatedly over the spring and summer months. Eggs and larvae are transported from the Agulhas Bank to the west coast by the north-flowing jet current at the shelf edge. Sardines are omnivorous, microphagist filter-feeders and are important prey for many piscivorous fishes, birds and mammals. There have been major fluctuations in sardine population biomass in southern Africa which have been associated with environment, regime shifts, recruitment and fishing. South African and Namibian sardine catches peaked in the 1960s, declining dramatically thereafter, and the pelagic fishing industry switched to anchovy. During the 1990s, sardine biomass increased but, although South African catches rose, the Namibian stock collapsed again. The annual South African total allowable catch is set through an operational management procedure that takes into account biomass estimates from comprehensive hydroacoustic surveys and the sardine by-catch in the anchovy fishery.

This paper is presented as a guide to assist in identifying larval stages of lanternfish (family Myctophidae) of the Agulhas Current region. The work is based on the larvae collected during three cruises conducted off the eastern coast of South Africa from 29°S to 34°S in 1990 and 1991. The larval distributions obtained from these surveys are summarised briefly. Generally, only literature citations are given for species of the region whose larvae have been described previously. Descriptions are limited to those species for which complete larval development have not been published previously and to those species not known to spawn in the region, to compare larval features with those from regions where the species is known to spawn (i.e., Benthosema fibulatum, B. pterotum, Hygophum proximum, Lampadena luminosa, Myctophum selenops, Notoscopelus caudispinosus and Triphoturus nigrescens). Also, diagnostic characters and illustrations are presented for species which may be difficult to identify in some developmental stages, particularly the early larval stages (i.e., Benthosema suborbitale, Bolinichthys spp., Ceratoscopelus townsendi, Diogenichthys panurgus, Myctophum phengodes, Lobianchia dofleini, L. gemellarii and Scopelopsis multipunctatus).

This paper reports on the composition, abundance and distribution of the larval fish assemblage in the nearshore coastal waters off the St Lucia Estuary mouth, South Africa. Ichthyoplankton samples were collected over a 12 month period from five stations located along a transect up to 2·5km offshore, and from two stations north and south of the estuary mouth, respectively. In all, 6126 fish larvae, representing 89 families and 186 species, were collected. Larvae in the families Myctophidae and Tripterygiidae comprised 21% and 16% of the total catch, respectively. The most abundant species were an unidentified triplefin, Tripterygiid 1 and the lanternfish Benthosema fibulatum, together which contributed nearly 18% of the total catch. Larvae of marine spawners independent of estuaries dominated the catch both in terms of density (90%) and in terms of number of taxa (89%). Some larvae of estuarine-associated species were present, in addition to a few specimens of estuarine resident species. Overall the dominant environmental variable affecting larval densities was temperature, particularly for Trypterygiid 1 where temperature contributed to 9% of the variance model. Densities of fish larvae peaked in November and December 1990 (late spring and early summer) and were lowest from January to June 1991 (summer, autumn an early winter). Different taxa dominated the catch each month with reef- and shelf-associated species accounting for the peak in August and September 1990, oceanic species in November 1990 and a mixture of the two groups in December. Overall larval densities were significantly higher in bottom samples with a trend of increasing densities offshore for reef and shelf taxa. The larvae of reef and shore taxa were predominantly preflexion larvae, whilst the few estuarine spawner species that were collected were mainly postflexion. Ontogenetic patterns related to depth and distance offshore were evident for the dominant species in each estuarine-association category.The present study has shown that temporal and spatial variations in the larval fish assemblage off St Lucia are related to environmental conditions and ontogenetic behavioural patterns of certain species. The origin of many of the larvae in the assemblages off the coast of St Lucia is probably from both local spawning populations in the shelf waters off KwaZulu-Natal and spawning populations farther north in shelf waters off Mozambique. Additional studies with more detailed oceanographic measurements will further our understanding of the physical processes that supply larvae to the St Lucia region.

The age of the western Indian Ocean pelagic scombrid Scomberomorus plurilineatus was determined by counting rings in whole otoliths. Validation using marginal increment analysis was inconclusive and it was assumed that a single opaque band was laid down annually in the otolith of this species. The maximum observed age was six years. S. plurilineatus is fully recruited to the fishery at the age of one year and 50%-maturity is attained at the age of two years. For the age-length data, the von Bertalanffy model was found to be the best fitting growth model. Growth was also estimated using a length-based method, Shepherd's Length Composition Analysis, but this proved to be inappropriate because of the seasonality of the catches of S. plurilineatus in KwaZulu-Natal. Based on a catch curve analysis, fishing mortality was estimated to be 0.28yr−1. Preliminary per-recruit analyses revealed that spawner biomass of S. plurilineatus is currently reduced to 50% of its unfished level and it appears that the species is optimally exploited.

The paper is based on larval distribution data from five lanternfish species outside the spawning grounds of the adult populations. Using hydrographic information collected with the larvae, the effect of surface currents on the observed patterns of distribution is discussed. In the South-East Atlantic, adjacent to the Benguela upwelling system, larvae of the pseudoceanic Lampanyctodes hectoris were found farther offshore (250 - 450 km) than usual. The general flow of the current is parallel to the coast, but upwelling filaments extend offshore. The size distributions of the L. hectoris larvae found, together with the rather low temperature at two of the oceanic stations sampled, indicate that they had been displaced by a filament. In the South-West Indian Ocean, the distribution of larvae of highly oceanic species of lanternfish over the shelf is related to onshore intrusions of the Agulhas Current. The size distributions of Hygophum hygomii and Scopelopsis multipunctatus indicate southward transport of larvae. The presence of larvae of Myctophum selenops and Benthosema pterotum, which have not been documented as adults in the region, indicates not only transport to the area, bur also that spawning could be taking place closer to the area than previously reported.

Ages were estimated for the small sparid Sarpa salpa obtained off the east coast of South Africa, from whole otoliths read under reflected light. Growth zones were composed of opaque and translucent zones and one opaque band was laid down per year. This was validated by marginal increment analysis and by an oxytetracycline labelling experiment using captive fish. The reproducibility of otolith readings was described by an index of average percentage error, which was equal to 3.9%. Von Bertalanffy growth parameters, which adequately described the growth of S. salpa, were determined. These parameters indicated a rapid growth rate for this species with a maximum age of six years being attained. Growth curves for separate sexes could not be determined because of the lack of sex-specific data in the younger age classes.