Discrimination of French grunts (Haemulon flavolineatum, Desmarest, 1823) from mangrove and coral reef habitats using otolith microchemistry

Using French Grunts (Haemulon flavolineatum) held captive within mangrove and reef sites, we determined (a) whether otolith microchemical differences existed between mangroves and reefs separated at a biologically relevant spatial scale (0.25-7.1 km), (b) whether patterns in elemental concentrations were consistent across years, and (c) whether it was possible to identify whether a given fish occupied a mangrove as a juvenile.Three sites were established at Great Exuma, Bahamas (two reefs and one mangrove, May 2001) and at Turneffe Atoll, Belize (one reef and two mangroves, August 2001 and 2002). Using concentrations of Sr and Ba, discriminant function analysis (DFA) indicated unique spatial microchemical signatures of fish from each of the three Bahamas sites allowing an average correct classification of 77%. Using concentrations of Sr, Ba, Sn and Pb (2001), and in addition Li, Mg, Cu and Rb (2002), DFA of the three Belize sites indicated an average correct classification of 68% and 85% in 2001 and 2002, respectively. To assess temporal variability in otolith microchemistry, we compared microchemical signatures of Belize fish from 2001 to those from 2002. On average, 42% of fish from 2002 were correctly classified to their captive sites using chemical information from 2001, thus suggesting considerable temporal variability in otolith microchemistry. Finally, to identify whether a given fish occupied a mangrove during its juvenile stage, we ablated the juvenile portion of the otolith taken from reefs in Belize 2002. Results of this analysis indicated that 36% of 39 individuals taken from the reef had a signature more representative of one of the mangrove sites. Although otolith microchemistry varied temporally and our analysis was restricted to the grouping of individuals to only one of three sites, mangroves appeared to contribute to reef populations.     

The importance of chemical environmental cues for juvenile Lethrinus nebulosus Forsskål (Lethrinidae, Teleostei) when settling into their first benthic habitat

A commercially important coral-reef fish, the spangled emperor Lethrinus nebulosus, settles into seagrass beds at the end of its pelagic larval phase, but the mechanism for locating these beds is unknown. To investigate this mechanism we first used a wide-choice, ex situ setup to examine the ability of captivity-reared naïve L. nebulosus settlers to select their first benthic habitat by reference to chemical cues. Second, we examined the morphology and ultrastructure of the nasal olfactory organ in settling L. nebulosus juveniles. We obtained the first evidence of a tropical seagrass-settling coral reef fish that can use chemical environmental cues in selecting its first benthic habitat at ranges up to at least 2 m. The L. nebulosus settlers exhibited a well developed pair of nasal olfactory organs, positioned in nares on the dorsal side of the head. These organs were elliptical radial rosettes, one in each of the olfactory chambers, and each comprised 12 lamellae, six on each side of a midline raphe, which were totally covered with sensory and non-sensory cilia, except for the margins. This type of cilia distribution is thought to indicate an acute sense of smell. The olfactory epithelium contained mature and immature ciliated receptor neurons bearing three to five cilia, and a second type of receptor neuron bearing six to eight microvilli.     

Modelling growth and primary production of the marine mangrove (Rhizophora apiculata BL): A dynamic approach

A mathematical model was developed to simulate the growth dynamics of monospecific stands of the marine mangroves (Rhizophora apiculata), based on physiological and morphological processes. Six variables to describe the tree organs and three morphological variables to illustrate the growth pattern were selected and incorporated in the model. The growth pattern was described by using a series of mathematical relationships.Net tree growth was derived from the combined effect of photosynthesis, respiration, and mortality. Net primary production was allocated to the tree organs based on their allometric ratios. Coefficients reported in literature were used to simulate the life history and demographics of R. apiculata. Initial diameter at breast height (DBH), daily global radiation and mean air temperature are some of the inputs that were used to calculate the biomass and demographic parameters, using daily time steps.The model was capable of simulating the growth dynamics of R. apiculata stands at lower latitudes, with reasonable accuracy (R² ≥ 0.80). However, the predictions at higher latitudes exhibited unsatisfactory model fits. This discrepancy in observations at higher latitudes is probably due to other factors like rainfall distribution, soil structure, which were not incorporated in the model structure.     

Contrasting patterns of reef utilization and recruitment of coral trout (Plectropomus leopardus) and snapper (Lutjanus carponotatus) at One Tree Island, southern Great Barrier Reef

Patterns of abundance, age structure and recruitment of coral trout (Plectropomus leopardus) and snapper (Lutjanus carponotatus) were described in different environments, which varied in benthic cover, in a 12-yr study at One Tree Island. It was hypothesized that both taxa would show strong preferences to different environments and benthic cover and that patterns would be consistent through time. Plectropomus leopardus were abundant on the reef slope and seaward edge of the lagoon, where live coral cover was high, and recruitment was generally low, in all environments. The population was sustained by a trickle of recruits, and total abundance varied little after 10 to 25 yr of protection in a no-take area, suggesting P. leopardus had reached an environment-related carrying capacity. Protogynous P. leopardus recruited to shallow environments at sites with 20% or more hard live coral and age data indicated the abundance of fish on the reef slope was from redistribution. Most recruits of gonochoristic L. carponotatus (<150 mm Standard length, SL) were found in the lagoonal environments, and adults were rare on the reef slope. Abundance of recruit L. carponotatus and P. leopardus did not correlate with percent cover of live and soft coral within environments. Recruits of L. carponotatus were usually rare in all lagoonal environments, but in 2003, many recruits (80 to 120 mm SL) were found in lagoonal environments with low and high hard live coral cover. A substantial proportion of the population (age max 18 yr) was from strong recruitment events. In 2003 and 2004, total abundance of L. carponotatus was supported by 1 year class 51.7 and 41% respectively. The utilization of environments and types of substrata varied among taxa and in some cases among life-history stages. There was also temporal variation in the importance of some environments (e.g. Lagoon Centre).     

Asexual propagation in the coral reef macroalga Halimeda (Chlorophyta, Bryopsidales): production, dispersal and attachment of small fragments

Siphonous, green macroalgae of the genus Halimeda are ubiquitous and ecologically important in tropical and subtropical marine environments. It has been hypothesized that the abundance of Halimeda on coral reefs is in part due to the ability of this genus to propagate asexually via vegetative fragmentation. However, vegetative fragmentation has only been documented for H. discoidea in a laboratory setting. To test the hypothesis that vegetative fragmentation contributes to field populations of Halimeda, we examined three aspects of fragmentation by H. tuna (Ellis and Solander) Lamouroux, H. opuntia (Linneaus) Lamouroux and H. goreaui Taylor on Conch Reef in the Florida Keys: (1) short-term (8 days) and long-term (14 weeks) fragment survival and rhizoid production in the laboratory and field (7 and 21 m), (2) size of the fragment pool and (3) influences of herbivory and water motion on production and dispersal of fragments. Although morphologically similar to H. discoidea, only a small percentage of H. tuna fragments survived. Fragments of H. opuntia and H. goreaui were more robust, and survival and rhizoid production were positively correlated with size in short-term trials. In 14-week field trials, one-third or fewer fragments of any species survived at 7 m, potentially because fragments were covered by large amounts of sediment. Survivors included some buried, seemingly dead individuals that turned green when exposed to light, highlighting the remarkable ability of this genus to survive disturbances. There was much less sediment accumulation at 21 m, where more fragments survived. Most (93%) eight-segment fragments of H. opuntia produced attachment rhizoids by the end of the 14-week trial. Overall, a range of 4.7-9.4 fragments of Halimeda m⁻² day⁻¹ were found on Conch Reef; most fragments were generated by H. goreaui. Fish bite marks were evident on 75-85% of the individuals of H. tuna and the number of bites per thallus ranged from 1 to 23. Herbivorous reef fish commonly fed on all three species of Halimeda. Some fish consumed the biomass, while others rejected most bites. For example, 83% of bites were rejected by the blue-striped grunt. Dispersal distances for rejected bites ranged from 0 to 31 m. Water motion was also responsible for fragment dispersal; experimentally produced fragments moved up to 48 cm day⁻¹. Results presented here suggest that asexual propagation of fragments of Halimeda is an important component of the life-history of this genus and vegetative fragmentation contributes to the abundance of this genus on coral reefs.     

The effect of territorial damselfish (family Pomacentridae) on the space use and behaviour of the coral reef fish, Halichoeres bivittatus (Bloch, 1791) (family Labridae)

Most demersal species of damselfish (family Pomacentridae) are territorial herbivores that aggressively chase other fishes away from their nests. This study investigates whether the aggressive territorial damselfish, Stegastes leucostictus, modifies behaviours and home range area use in a less aggressive, non-territorial species, the slippery dick wrasse, Halichoeres bivittatus (family Labridae). Damselfishes and wrasses are ubiquitous and abundant members of coral reef fish assemblages around the world; hence, this study has broader implications beyond the two Caribbean species chosen for this study. A manipulative field experiment consisted of transplanting one or three S. leucostictus into artificial shelters positioned within adult or adjacent to juvenile H. bivittatus home ranges. The introduction of damselfish had little effect on the size of home range areas of juvenile or adult wrasses, but had a significant effect on the location of their home ranges. The damselfish also affected adult microhabitat use but did not affect use by juveniles. In addition, there was an increase in damselfish-wrasse interactions that resulted in a significant reduction in the amount of time adult wrasses could spend foraging or in a foraging assemblage. It was concluded that aggressive interactions with territorial damselfish affect individual H. bivittatus space use and may consequently affect the spatial distribution of H. bivittatus populations.     

Reproductive fitness,population size and genetic variation in Muscari tenuiflorum (Hyacinthaceae): The role of temporal variation

In plant populations a positive correlation between population size, genetic variation and fitness components is often found, due to increased pollen limitation or reduced genetic variation and inbreeding depression in smaller populations. However, components of fitness also depend on environmental factors which can vary strongly between years. The dry grassland species Muscari tenuiflorum experiences long term habitat isolation and small population sizes. We analyzed seed production of M. tenuiflorum in four years and its dependence on population size and genetic variation. Genetic diversity within populations was high (AFLP: H_e = 0.245; allozymes: H_e = 0.348). An analysis of molecular variance revealed considerable population differentiation (AFLP: 26%; allozyme: 17%). An overall pattern of isolation by distance was found, which, however was not present at distances below 20 km, indicating stronger effects of genetic drift. Genetic diversity was positively correlated to population size. Self pollination reduced seed set by 24%, indicating inbreeding depression. Reproductive fitness was not correlated to genetic diversity and a positive correlation with population size was present in two of four study years. The absence of a general pattern stresses the importance for multi-year studies. Overall, the results show that despite long term habitat isolation M. tenuiflorum maintains seed production in many years independent of population size. The long term persistence of populations is thus expected to depend less on intrinsic genetic or demographic properties affecting seed production but on successful plant establishment and persistence, which latter are based on conservation and protection of suitable habitats.     

Microsatellite multiplexes for high‐throughput genotyping of French grunts (Haemulon flavolineatum,Pisces: Haemulidae) and their utility in other grunt species

Grunts of the genus Haemulon are abundant in the western Atlantic and are an important component of regional fisheries. Management efforts for these and other Caribbean reef fish require an understanding of population structure and the extent of dispersal among populations. We characterized 11 polymorphic microsatellite loci for French grunts (Haemulon flavolineatum) and grouped them into three PCR multiplexes to facilitate high‐throughput genotyping. These loci will be useful for population genetic studies of French grunts. Cross‐species amplifications suggest these loci are also suitable for population genetic studies of H. aurolineatum, H. plumieri, and H. sciurus.     

Recruitment of the infaunal bivalve Soletellina alba (Lamarck, 1818) (Bivalvia: Psammobiidae) in response to different sediment types and water depths within the intermittently open Hopkins River estuary

Studies examining recruitment processes for soft-sediment macroinvertebrate fauna in intermittent estuaries are rare and most studies of active habitat selection have been tested in the laboratory rather than the field. The present field study examined whether recruitment of the infaunal bivalve Soletellina alba was influenced by water depth and sediment particle size in the intermittent Hopkins River estuary, southern Australia. The number of recruits in sediment trays differed between water depths, but active habitat selection was not evident across treatments of varying sediment particle size. The use of sediments with varying particle sizes also provided an opportunity to identify potential discontinuities in body-size distributions of recruits associated with varying habitat architecture. The length (mm) of recruits was converted to the same scale used to express sediment particle size (i.e. phi units: ? = − log₂ of sediment particle size). The size of recruits differed across water depths, but did not differ across treatments with fine (? = 3) versus coarse (? = 1) sediment, and no relationships were apparent between bivalve size and sediments consisting of varying particle size. These patterns of recruitment do not correspond with the distribution of adult S. alba within the Hopkins River estuary. Previous sampling has shown that abundances of juvenile and adult S. alba are variable across time, site and water depth, but are often greater at the deeper water depth (1.05 m below the Australian Height Datum). However, recruitment during the present study was greatest at the shallower water depth (0.05 m below AHD), and the apparent absence of active habitat selection suggests that the distribution of adults is unlikely to be attributable to differences in recruitment associated with sediments of varying particle size.     

Effect of nutrient availability, grazer assemblage and seagrass source population on the interaction between Thalassia testudinum (turtle grass) and its algal epiphytes

Seagrass leaves are often densely covered by epiphytic algae which can suppress seagrass productivity and has been implicated in declines of seagrass meadows worldwide. The net effect of epiphytes on seagrass growth and morphology depends on the independent and interactive effects of a variety of factors, including nutrient availability and the intensity of grazing on epiphytes. Here I report the results of a mesocosm experiment designed to test the effects of nutrient addition and within-functional group variation (grazer species composition and the source population of seagrass) on the strength of the interactions among grazers, epiphytes, and turtle grass (Thalassia testudinum). Turtle grass ramets from two sites in the northern Gulf of Mexico were cleared of epiphytes and transplanted into common-garden mesocosms. Replicate ramets were grown in a split-split plot design with two levels of dissolved nutrients and four different grazer species combinations (Tozeuma carolinense alone, Pagurus maclaughlinae alone, both species together, and no grazers present). As expected, grazers had a significant negative effect on epiphyte biomass/leaf area and a significant positive effect on turtle grass growth in the mesocosms. The two species were more similar in their direct effects on epiphyte biomass than in their indirect effects on turtle grass growth; this may reflect differences in epiphyte community composition under different grazer treatments. The effect of nutrient addition on turtle grass growth depended critically on the intensity of grazing: in the presence of grazers, turtle grass tended to produce a greater biomass of new leaf tissue in the tanks with nutrients added than in the control tanks. However, when grazers were absent, the direction of the effect was reversed, and plants with nutrients added grew less than the control plants. The two source populations of turtle grass differed significantly in epiphyte biomass/leaf area accrued in the mesocosms as well as in the strength of the effect of grazers on turtle grass growth. This suggests that population differentiation in seagrass interactions with epiphytes, as well as spatial and temporal variation in resources and grazer community composition, can greatly effect the role of epiphytes in limiting seagrass productivity.     

Interaction between Neoaplectana carpocapsae (Nematoda: Steinernematidae) and Apanteles militaris (Hymenoptera: Bracondiae), a parasitoid of the armyworm, Pseudaletia unipuncta

The DD-136 strain of Neoaplectana carpocapsae adversely affected the development of immature stages of Apanteles militaris, a gregarious internal parasitoid of the armyworm, Pseudaletia unipuncta. The adverse effect of the nematode-bacterium complex was indirect, i.e., the infection by the nematode killed the host before A. militaris could complete its development. However, if armyworms containing 10- or 11-day-old A. militaris were exposed to dauer juveniles in Petri dishes, 48.1 and 94.4%, respectively, produced normal cocoons. If hosts containing 9- or 10-day-old A. militaris were fed dauer juveniles, 42.6 and 73.4% of the armyworms, respectively, produced A. militaris which formed normal cocoons. Cocoon-spinning A. militaris larvae were infected by the nematode. After cocoon formation was completed, the dauer juveniles could not penetrate the cocoon and infect the pupa. However, pupae in cocoons which had been deliberately cut open at one end became infected. A. militaris adults were infected when exposed to dauer juveniles in Petri dishes. After 3 days of exposure to dauer juveniles, 25.0, 44.2, and 7.0% of the adults in three trials were alive, whereas 100, 100, and 96.7% of the control adults were alive. Examination of dead adults in the nematode treatment showed that 67.6% contained nematodes. N. carpocapsae developed and reproduced in unparasitized armyworms, in armyworms containing 9-day-old A. militaris, and in those from which A. militaris had emerged. Production of dauer juveniles was significantly higher in unparasitized armyworms and in armyworms containing 9-day-old A. militaris than in those from which A. militaris had emerged.     

Effect of temperature and salinity on the energetics of juvenile gray snapper (Lutjanus griseus): implications for nursery habitat value

Gray snapper (Lutjanus griseus) encounter a wide range of temperatures and salinities in nearshore and estuarine juvenile habitats. The energetic response of juvenile gray snapper to temperature and salinity was measured in laboratory experiments to determine the influence of these physicochemical factors on the potential value of different juvenile nurseries. Maximum consumption and growth rates of juvenile (25-50 mm SL) gray snapper were determined in 12-day trials at 20 temperature/salinity combinations representing conditions in juvenile habitats. Ad libitum feeding level of individual fish was measured daily. Maximum weight specific feeding rate increased significantly with temperature and salinity; however, the effect of salinity was much less than that of temperature. Linear growth rate and specific growth rate both increased with temperature, and salinity did not have a significant effect on either. Gross growth efficiency (K₁, growth×consumption⁻¹*100) increased with temperature and was significantly lower at high salinities, indicating increased energetic costs. The higher K₁ at lower salinities has several implications for juvenile gray snapper in low salinity habitats: (1) they would need less food to achieve the same somatic growth as juveniles in high salinity habitats; (2) they would have higher growth at limited ration levels as compared to high salinity habitats; and (3) they would have less impact on prey populations than higher salinity habitats assuming similar gray snapper densities.     

Genetic differences between two spider crabs Pisoides bidentatus (A. Milne-Edwards, 1873) and Pugettia quadridens (de Haan, 1839) (Decapoda: Brachyura: Majoidea) from the Sea of Japan

Biochemical genetics was applied to elucidate the taxonomic position of the spider crab Pisoides bidentatus. Two morphologically similar spider crabs, P. bidentatus and Pugettia quadridens, were compared with four sympatric brachyuran species, Eriocheir japonicus, Hemigrapsus sanguineus, Hemigrapsus penicillatus (Varunidae), and Cancer amphioetus (Cancridae), using 20 allozyme loci. High genetic identity-value between P. bidentatus and P. quadridens (I = 0.758 ± 0.092) suggests that they should evidently be placed to one genus. This conclusion is supported by the similar identity-value between two species of the genus Hemigrapsus (I = 0.821 ± 0.078) and the low genetic similarity (I = 0.125 ± 0.05) between different genera, Hemigrapsus and Eriocheir, within the family Varunidae.     

Variation in diapause potential and strength in eggs of the Australian plague locust, Chortoicetes terminifera (Walker) (Orthoptera: Acrididae)

Samples of eggs of Chortoicetes terminifera were incubated under 3 temperature regimes which would allow non-diapause eggs to develop about 50% and so take them beyond the diapause stage. Even so, many more eggs entered diapause when reared at 20°C for 3 weeks than at 32°C for 1 week. By collecting and incubating eggs at intervals after laying in autumn, diapause potential or strength of eggs at different stages of development was estimated. For eggs in pre-diapause, the proportion in diapause after rearing was used to estimate diapause “potential”; for those in diapause, the proportion was used to estimate diapause “strength”. At laying, eggs varied greatly in their potential to enter diapause. However, those with lower potential at laying often increased their potential during the pre-diapause stage so that by the time diapause was reached diapause strength varied but over a lesser range. In all eggbeds, diapause strength decreased by 7–9 weeks after laying; little or none remained by mid-winter.These variations in diapause potential and diapause strength seem to reflect how much the temperature threshold for development during diapause is increased above that for non-diapause. Both diapause potential and strength may reflect the value of some factor whose level at laying is determined by the environment experienced by the parents but which changes subsequently.     

Antigenic cross-reactivity between Schistosoma mansoni and pollen allergens from the birch tree (Betula verrucosa) and Timothy grass (Phleum pratense): involvement of shared glycan epitopes and implications for the hygiene hypothesis

Previous studies have shown that schistosome infection can protect against allergic symptoms, but the underlying mechanisms are still not fully understood. Here we have shown that rabbit IgG antibodies raised against Schistosoma mansoni soluble egg antigens (SmSEA) are cross-reactive with a wide array of molecules in Timothy grass pollen (TGP) and birch tree pollen (BTP). Five of the cross-reactive pollen molecules (two from TGP and three from BTP) were selected randomly and identified by tandem mass spectrometric (TMS) analysis to be, respectively, the TGP allergens Phl p 1 and Phl p 5b, and BTP glutathione S-transferase (GST), and the BTP allergens Bet v 1 and Bet v 6.0102. Rabbit anti-SmSEA IgG antibodies that cross-reacted with each of the five allergens were found to be reactive with three major S. mansoni egg antigens, IPSE/alpha-1, omega-1 and kappa-5. Pairwise alignment of the amino acid sequences of each of the five TMS-identified pollen allergens with each of the three egg antigens revealed a low level of amino acid sequence identity. Further experiments indicated that the schistosome antigen/allergen cross-reactivity was mostly due to similar glycans present in helminths and plants, but not in mammals: so called cross-reactive carbohydrate determinants (CCDs). Previously, CCDs have been implicated in the cross-reactivity between many plants and invertebrates. Furthermore, pollen-induced anti-CCD IgGs have been found in sera of patients undergoing allergen-specific immunotherapy (SIT) and implicated in the treatment of the allergy. Thus, our finding provides not only possible explanations for the allergy-protective effect of helminth/schistosome infections as explained by the hygiene hypothesis, but also a potential starting point for improved SIT.     

Expansion of Myriophyllum spicatum (Eurasian water milfoil) into Lake Nasser,Egypt: Invasive capacity and habitat stability

The invasion of Myriophyllum spicatum into Lake Nasser, and its impact on submerged macrophyte communities are quantitatively documented. Samples of macrophytes, water and hydrosoil were collected from 17 sites, in October and November 2002. The average dry weight standing crop of each species per grapnel haul was determined at each depth zone (sampling site). Twenty-one environmental variables were measured (12 water and 9 hydrosoil variables). Canonical correspondence analysis (CCA) was used to determine species–environment relationships. Comparing these relationships of the present study with those detected in 1988–1990 indicated significant changes in water and hydrosoil characteristics. These changes are also implicated in the submerged macrophyte communities. M. spicatum has replaced the originally dominant submerged macrophyte Najas marina subsp. armata. The study indicated that the invasion of M. spicatum depends not only on its attributes, but also on the physico-chemical characteristics of Lake Nasser.     

Adult movement of the native holly leafminer,Phytomyza ilicicola Loew (Diptera: Agromyzidae): consequences for host choice within and between habitats

Summary Phytomyza ilicicola is a specialist herbivore, the larvae of which mine, and the adults of which feed and oviposit on the developing leaves of American Holly, Ilex opaca. Adult Phytomyza have been shown to discriminate among host plants on the basis of several host and habitat factors. Such discrimination may explain, in part, the observed density variation of larval Phytomyza among individual trees and between forest and suburban habitats. In order for discrimination to influence density, adult Phytomyza must be sufficiently vagile to move among habitats and to encounter many hosts. We experimentally monitored the movement of adults to ask if this were so. In addition we asked whether inter-host movements were altered by resource concentration, and whether vagility offered an escape from parasitism. To answer these questions, three isolated holly trees to serve as colonization sources were selected in early spring, prior to adult emergence. Potted target trees were then placed singly and in groups of four at 3 compass directions around and at 3 different distances (10, 25, and 50 m) from each of the source trees. Six weeks after exposure to adult Phytomyza, the number of feeding punctures and first instar larvae were counted. Clones were then overwintered in cold frames and in early spring, just prior to adult emergence, all pupae were examined for parasitism by the dominant parasitoid Opius striatriventris. Feeding and parasitism were similar at all 3 distances from the source trees, indicating that Opius was at least as vagile as Phytomyza. However, first instar larvae decreased with distance. There was no evidence that resource concentration affected adult movement. Our results suggest that Phytomyza are sufficiently vagile to choose among hosts within and between habitats. However, movement does not translate into equally high oviposition at all distances from a source, nor does it provide an escape from parasitism by Opius.     

Variability in the numbers of post-settlement King George whiting (Sillaginidae: Sillaginodes punctata, Cuvier) in relation to predation, habitat complexity and artificial cage structure

The importance of predation by fish in altering abundances of juvenile King George whiting (Sillaginodes punctata) was examined at multiple locations in Port Phillip Bay, Australia, by manipulating the numbers of piscivorous fish in unvegetated sand and seagrass habitats using cages. Additional information regarding the local abundances of, and habitat use by, the most common piscivorous fish, Western Australian salmon (Arripidae: Arripis truttacea, Cuvier), was gathered using netting surveys and underwater video. Regardless of habitat, abundances of S. punctata were similar in partial cages and uncaged areas. In unvegetated sand, S. punctata were more abundant inside cages than partial cages or uncaged areas. In seagrass, there was no difference in the numbers of S. punctata between caging treatments. Patterns in abundances of S. punctata between cage treatments in each habitat were consistent between sites, but the relative difference in the abundances of S. punctata between habitats was site specific. Abundances of A. truttacea varied significantly between sites, and they consumed a variety of epibenthic fishes including atherinids, clupeids, gobiids, syngnathids and pleuronectids. At one site in Port Phillip Bay (Blairgowrie), A. truttacea occurred more commonly in patches of unvegetated sand than seagrass. Over unvegetated sand, abundances of A. truttacea varied little between partial cages and uncaged areas. The numbers of S. punctata varied between caging treatments and habitats in a manner that was consistent with a model whereby seagrass interferes with foraging by predatory fish and provides juvenile fish with a refuge from predation. The almost total absence of A. truttacea in seagrass habitats and the lack of S. punctata in their diets implies, however, that patterns in S. punctata in seagrass/unvegetated sand mosaics are driven by processes other than direct predation.