Effects of density and food supply on postlarval abalone: behaviour, growth and mortality.

ABSTRACT Understanding density dependent processes in abalone is important for sustainable management of fisheries, and optimum aquaculture production. Density and food were manipulated to test their interacting effects on the growth and survival of hatchery produced postlarval Haliotis rubra. Feeding and dispersive behavior was also examined. Initial densities ranged between the densities used in aquaculture and those reported in the natural environment. Higher initial density increased the mortality rates of postlarvae in two experiments. Simultaneous manipulation of diatom food supply and density showed that the effect of density on mortality was mostly independent of food. Growth was strongly related to food supply, but also significantly reduced by high initial density, and these effects were additive. When diatom density was low, postlarvae did not increase rasping rates so the effect was that food was consumed more slowly. This explains the strong effect of food supply on growth. No aggressive interactions between feeding postlarvae were observed, but they graze intensively, leaving small distinct patches that have been cleared of algal food. Postlarvae encountering patches without food move rapidly and apparently randomly in search of food. The dispersal rate of postlarvae increased with density, presumably because post larvae spent more time searching for food. This searching is likely to be energetically expensive, and would increase at high densities because more cleared patches are created. This may explain why high initial densities of postlarvae had persistent reduced growth even when food was abundant, but does not explain the increased mortality at high density irrespective of food supply. We hypothesize that this effect may be attributable to other factors, such as the increased transmission of a pathogen at high density.

KEY WORDS: recruitment variability, intraspecific competition, density dependence, diatoms, Haliotis rubra

INTRODUCTION

The role of density dependent processes in populations has been an important debate in ecology for decades (Nicholson 1933, Andrewartha & Birch 1970, Stubbs 1977, Rose et al. 2001). Understanding any density dependent processes throughout the life cycle is also vital to sustainable management of exploited populations (Fogarty et al. 1991, McShane & Naylor 1995), because the overall compensatory density dependence in the relation between adult population size and future recruitment is crucial for any sustainable fishery (McGarvey 1996, Rose et al. 2001). Density dependent processes will also determine the optimum production strategies in aquaculture.

Although it is understood that some form of density dependence must operate in all populations (Haldane 1953), recruitment to some fish and invertebrate populations is so highly variable that some have questioned whether any relation exists between adult stocks and recruitment (e.g., McShane 1995, McShane & Reyn 1996). There is general agreement that the early life history stages of these organisms are a likely source of much of the variation in recruitment, due to their small size and relatively large numbers (Hunt & Scheibling 1997). This may make them particularly vulnerable to density dependent processes such as predation (Hurlbut 1991, Shepherd & Daume 1996), competition for food (Marshall & Keough 1994, Schmitt 1996, Elmgren et al. 2001), or space (Paine 1984).

Both pre- and postsettlement stages may contribute to recruitment variability, but the early postsettlement period has not received much attention, due to the difficulty of measuring small settlers whose distribution varies greatly in time and space (Keough & Downes 1982). Measuring recruitment at some arbitrary minimum size may ignore mortality that could have occurred in the period between settlement and observation (Keough & Downes 1982, Connell 1985). In cases where settlement rates are high (Gaines & Roughgarden 1985, Hurlbut 1991, McShane 1991) or mortality rates vary over time (Roegner 1991) postsettlement processes will be more important in determining the degree of recruitment variation than larval supply. Further, the significance of the postsettlement period is likely to be especially important in populations where larval dispersal is relatively local, because the dynamics will correspond to a closed rather than open population model (Roughgarden et al. 1988, Underwood & Fairweather 1989). In closed populations the postsettlement phase will both depend on and, with a time lag, impact on adult population density, because larval influx from other populations will not confuse the relationship.

In aquaculture the rate at which a farm can produce abalone of any given size will depend directly on the survival and growth rates of postsettlement individuals. Thus, understanding any density-dependent processes that affect growth and survival is essential for determining the most effective strategies to maximize yield and minimize turnover rates and costs in aquaculture ventures.

Abalone are ideal for examining density dependent mechanisms in the postsettlement phase, because they are now extensively cultured. The issue is important because effective management arising from an understanding of the processes underlying recruitment variation is vital to the sustainability of these fisheries (Shepherd & Rodda 2001), and there is evidence that many abalone species have localized larval dispersal (Prince et al. 1987, Prince et al. 1988, McShane et al. 1988), so that populations are effectively closed. Studies on several species have indicated that various density dependent processes may act in the early postsettlement stage. These include competition for space (McShane 1991), predation (Shepherd & Daume 1996, Preece et al. 1997), and competition for food (Daume et al. 2001, Heasman et al. 2001). While these studies show evidence for density-dependent growth or mortality, the underlying mechanisms remain obscure.

The lack of a mechanistic understanding of early life history processes can be associated with the inherent difficulty in studying tiny individuals in the field (McShane 1992). Using a suction sampling device McShane (1991) sampled postlarval Haliotis rubra in the crustose coralline red algal habitat, where abalone larvae settle almost exclusively (Shepherd & Turner 1985, Shepherd & Daume 1996, Daume et al. 1999). Over time, the results showed a change in mortality rate that was correlated to the number of postlarvae observed soon after settlement. Shepherd & Daume (1996) obtained similar results from an in situ study of Haliotis laevigata and Haliotis scalaris. The limitation of these approaches is that they are observational rather than experimental so that their conclusions are limited to inferences or suggestions.

Ignorance of the underlying processes behind density dependent mortality and growth extends to the hatchery environment. Because the effects of food quality and postlarval diet on growth and survival have received nmch attention in the recent literature, (e.g., Kawamura et al. 1998a, Kawamura et al. 1998b, Roberts et al. 1999a, Daume et al. 2000), the role of density in this has only recently been recognized (Daume et al. 2001, Heasman et al. 2001). This is surprising considering that improved settlement rates using cues such as the green alga Ulvella lens (Daume et al. 2000, Daume et al. 2001) allow very high densities of postlarvae to be cultured. Even this recent work on postlarval density only describes the outcomes of the interaction (in terms of the effects on growth and survival). The same problem as in field studies is evident: while competition for food is proposed, the mechanisms remain untested.

The objective of this study is to identify the mechanisms behind any density dependent effects on growth and survival rates of postsettlement abalone. Identifying these mechanisms might both allow abalone farms to increase their efficiency, and prove useful in predicting the dynamics of natural populations. Small-scale experiments in a hatchery environment and observations of postlarval behavior in relation to food were used to test for intraspecific competition for food in postlarval H. rubra, and to investigate the underlying mechanism of any density effects. The postlarval phase in abalone has been defined as the period after settlement and metamorphosis (~300 [micro]m shell length) to the formation of the first respiratory pore (e.g., Leighton 1974, Kawamura et al. 1998b). This seems arbitrary in an ecologic context. In this study the end of the postlarval phase is considered to be the shift to the cryptic juvenile phase at between 5- and 8-mm shell length (Shepherd & Turner 1985, Shepherd & Daume 1996), which involves a change in habitat and behavior that will influence the resources they use and sources of mortality. In the hatchery environment postlarvae are weaned off microalgal food and onto artificial food at a similar size (Hahn 1989).

METHODS

Spawning was induced in conditioned brood stock at the hatchery of Ocean Wave Seafoods, Lara, Victoria. Larvae were kept in an aerated, 300 L conical tank with flowing UV-sterilized, 3-[micro]m filtered water. Larvae were used in experiments 5-6 days after fertilization, when they were judged competent to settle, as they began exploring the tank's surface and had developed the third tubule of the cephalic tentacle (Daume et al. 2000). The experiments used two larval batches, each the result of a mix of four to six parents.

A culture of the diatom Navicula sp., obtained from the Department of Fisheries, Western Australia, was used as food in all experiments. Navicula sp. has been shown by Daume et al. (2000) to produce high growth and survival rates in postlarval H. rubra. The non-axenic culture was maintained in f/2 growth medium complemented with silicates (Guillard & Ryther 1962) at ambient temperature and under a 12 h L:D photo cycle. For the f/2 growth medium, seawater was filtered to 1 [micro]m before adding sodium hypochlorite (commercial chlorine, 12.5% W/V) and aerating for 3 h. A neutralizing solution of sodium thiosulfate was then added, followed by a further 3 h of aeration.

Behavior

Ten Petri dishes (85 mm diameter) were inoculated with Navicula sp. and cultured for 3 days, alter which diatom density was measured by counting the number of diatoms within a 300- x 300 [micro]m quadrat at x100 magnification for 10 random fields in each dish. Mean density was 517 [mm.sup.-2] (SE = 24, n = 10). Isolated square patches of food approximately 5 x 5 mm were formed by removing a 5-mm wide grid of diatoms with a small cloth. A single postlarva (mean size 1273 [micro]m, SD 200 [micro]m, n = 12) was placed in the center of one of these patches in each dish, and its position recorded every 4 h for a 24 h period (under a 12 h light to dark cycle).

Effect of Density on Dispersal

Plates with a cover of the green alga Ulvella lens of at least 95% were used to examine the grazing behavior of postlarvae. A good cover of U. lens greatly enhances the settlement rate of H. rubra larvae and provides food to larger postlarvae (Daume et al. 2000, Daume et al. 2001, Daume et al. 2004). Squares (200 x 200 mm) cut from commercial settlement plates (30 x 60 cm) were placed horizontally in a 225 x 225 x 145 mm container filled with 3-[micro]m filtered seawater with an average output of 80 [+ or -] 5 mL.[min.sup.-1] per container. Overhead fluorescent lights provided illumination on a 12 h cycle and temperature remained at 18[degrees]C [+ or -] 1[degrees]C.

At settlement, the larvae were restrained to the center of the plates with a vertical PVC pipe, 45 mm in diameter and 16 cm high. After settlement, the pipe was removed and the larvae were allowed to spread. A grid with concentric squares (Fig. 1) was placed behind the settlement plates and the number of postlarvae within the central square was manipulated by haphazardly removing some of the postlarvae with a soft cloth to obtain high (~100) and low (~20) densities. All post larvae outside this central square were removed with a soft cloth. Five replicates were used for each treatment. The distribution of the postlarvae was then recorded by counting the individuals in each marked square of the grid at 1.5, 6, 12, 24, 48, 96, and 192 h after the removal of the pipe.

Density-light Interaction

Two levels of initial postlarval density (20,000 and 5,000 [m.sup.-2]) were tested and the abundance of diatom food was controlled using 2 levels of illumination (one layer of shade cloth and no shading). Higher illumination maintains a higher diatom productivity and abundance (Kinne 1973), so that light management is a common method in commercial hatcheries to control algal abundance on settlement plates (Daume et al. 2001). The higher initial density was chosen according to the highest densities found in the natural environment (McShane 1991, McShane & Smith 1991, Nash et al. 1995) and the scale of the experimental units. Whereas 20,000 individuals [m.sup.-2] is c.a. 1.6 times greater than the maximum recorded in the wild (McShane & Smith 1991) their sampling technique examined density on the scale of [m.sup.2] and small, denser aggregations may occur on a smaller scale--such as that used here. The lower initial density corresponds to that used in the aquaculture industry (Heasman et al. 2001), and was judged sufficient to ensure that a minimum number of postlarvae survived in each jar after settlement and initial mortality, two processes that are notoriously variable.

Commercial plastic settlement plates (30 x 60 cm) were cut into squares (5 x 5 cm). These plates had a cover of the green alga Ulvella lens of at least 95%. Plates were then randomly assigned to glass jars (c.a. 250 mL volume) filled with 3 [micro]m filtered seawater. The jars were placed 35 cm below a fluorescent neon light on a 12 h L:D cycle. Approximately 300 larvae were added to each jar. There was no flow at this stage. Three days after the larvae were released the plates were examined under a dissecting microscope. The initial treatment densities were established by haphazardly removing postlarvae from the plate until the desired number had been reached. After this manipulation, numbers were recounted to ensure accuracy. A flow-through system of 3 [micro]m filtered seawater was turned on at this point.

Opaque PVC piping was placed around all jars to restrict illumination to that coming from above. In the shaded treatments, 70% shade cloth was fastened over the top of the PVC piping. Six replicates were used for each of the four treatment combinations. Survival and growth were monitored weekly for a period of 6 wk. For sampling, platelets were carefully transferred to a petri dish under the dissecting microscope and were kept immersed in water at all times. The postlarvae were counted, and seven randomly selected postlarvae were measured at x50 magnification with a graticule ([+ or -]20 [micro]m resolution). The jars were wiped clean after each weekly measurement to minimize disturbance. Average flow rate was also recorded for each outlet of the system. Because the counts were obtained repeatedly from plates over time, differences in overall mortality of post larvae between initial density treatments were tested using linear comparisons of means in a repeated measures ANOVA. We chose this method rather than survival analysis because it is more familiar to most ecologists and the interpretation is straightforward.

Food quantity and quality

Food abundance (high and low) and initial postlarval density (20,000 and 3,500 [m.sup.-2]) were manipulated to test their combined effects on growth and survival. In the density-light experiment, food quantity and quality could not be closely monitored because diatoms could not be clearly seen through the dense U. lens cover. This second experiment was designed to directly manipulate and monitor the food levels and was thus carried out in petri dishes inoculated with a diatom film. The petri dishes were randomly placed under a fluorescent light (12 h L:D cycle), and water was exchanged with sterile, 3 [micro]m filtered water every 2 days. The experiment was run for 3 wk.

Small petri dishes were used (20, each 3.4 cm diameter, 9.1 [cm.sup.2] area) and a diatom film of Navicula sp. was grown on them before the start of the experiment. Diatom abundance was controlled in two ways. Half of the petri dishes were inoculated 4 days prior to the others to allow for a denser initial diatom biofilm. This difference was then maintained by the addition of extra diatoms to the dishes (1-2 times per week). Five replicates were used for each of the four treatment combinations.

Post-larvae were grown at low density on large (20 x 20 cm) U. lens plates in a 3 [micro]m flow-through system and then transferred 16 days after settlement to the petri dishes at a mean size of 663 [micro]m (n = 118, SD = 70 [micro]m). Transferral involved brushing them off the plates into a clean petri dish with a soft cloth. They were left for 3-4 h and rinsed several times with 3-[micro]m filtered, sterilized seawater to remove diatoms, fecal material, and the like. They were then delicately moved into the treatment dishes with a soft cloth. Damaged shells were removed and postlarvae that did not land the right way up were inverted with a small jet of water from a pipette.

Survival was checked 1, 3, and 7 days after transferral to the petri dishes and every week thereafter. The initial size of all postlarvae in the low-density treatments and 8 randomly selected postlarvae from the high initial density treatments were measured on the second day at x50 magnification. Diatom abundance was measured at the same time with a compound microscope. Ten fields of view at x100 magnification were randomly selected and the number of diatoms within a 500 x500 [micro]m grid was counted along with an estimate of percent coverage. Growth, survival, and diatom abundances were measured weekly.

RESULTS

Behavior and Patchy Food

Postlarvae generally began grazing less than 20 sec after being transferred to a patch of food. Rasping of the surface with the radula could be observed with an inverted compound microscope and diatoms could be seen being removed by the postlarvae.

Grazing behavior of postlarvae consisted of steady rasps of the surface. After each rasp the postlarvae turned a small angle in one direction. This direction was the same until the abalone had turned approximately 120[degrees], as shown in Figure 2. A small movement forward was made to maintain its position on the edge of the food before grazing continued in the opposite direction. Postlarvae grazed areas thoroughly and in a manner that formed bare patches of space.

[FIGURE 2 OMITTED]

The mean grazing rate of postlarvae was 1177 [+ or -] 113 diatoms [hour.sup.-1] in patches of diatoms (n = 10). There was no significant linear regression of size on grazing rate ([F.sub.1,8] = 1.869, P = 0.209). Observations showed that less food is consumed per unit time when diatom densities were low. Where diatoms were closer together, each rasp of the radula picked up more diatoms than when the diatoms were more widely spaced. When diatom density was minimal (approximately 50 [mm.sup.-2]) the radula would only pick up diatoms every three to four rasps, as opposed to every one to two rasps at higher diatom densities. There was no obvious increase in rasping rate of the radula at lower diatom densities, which suggests that postlarvae do not compensate for a lower diatom density by grazing faster.

Post-larvae would often graze to the edge of patches of food. In doing so, their anterior sensory appendages lost contact with the food source. In the absence of this contact postlarvae would stop grazing and start a series of apparently random turns, often more than the 120[degrees] rotations observed when grazing. Their speed of movement also increased. This pattern of movement ceased when the postlarvae found another patch of food. Three postlarvae were observed from the time they stopped grazing to the time they found a patch and started grazing again. This occurred in a mean time of 105 [+ or -] 34 sec. Seven postlarvae were observed to leave their patch over the 28 h of the experiment. This would commonly occur before the entire patch had been consumed (mean cover remaining = 43 [+ or -] 4%, n = 7). This does not include the postlarvae that may have traveled between patches between observation intervals. Contacts between postlarvae were observed in the other experiments, but there was no aggressive interaction: searching or grazing often continued throughout the contact.

Dispersal

Figure 3 shows the distribution of postlarvae in each concentric region expanding from the center of the plate. Postlarvae started spreading immediately and a change in distribution was observed within 1 1/2 h. The movement of postlarvae out of the central zone appeared faster at higher initial densities. Towards the end of the experiment, when significant numbers of postlarvae were reaching the larger outer areas of the plate, the proportions became more even across the plate at high rather than at low initial density (Fig. 4), although the proportion of postlarvae remained highest in the central area.

[FIGURES 3-4 OMITTED]

The Effect of Shading and Density

Figure 5 shows the change in size and the trend in log numbers surviving for the postlarvae in each of the four treatments over time. Although variation in survival within treatments was high, there was a significant interaction between time and density (Table 1), and the linear component of this interaction was significant ([F.sub.1,20] = 8.498, P = 0.009). Thus overall mortality was significantly affected by the density treatment. There was no significant time by light interaction, which indicates that mortality is independent of the shading treatment. There was no significant difference in flow between treatments (Light: [F.sub.1,20] = 0.692; Density: [F.sub.1,20] = 0.126; interaction [F.sub.1,20] = 1.203; P > 0.20 in all cases).

[FIGURE 5 OMITTED]

Analyzing the size data for the whole experiment duration would have greatly reduced the replication level in the ANOVA with repeated measures, due to the mortality of all postlarvae in some of the experimental units. Therefore only the data from the first 2 wk of the experiment were analyzed this way. In weeks 0-2 postlarvae became significantly larger in the unshaded treatments (Table 1, Fig. 5B). There was also a trend towards smaller sizes in the high initial density treatments (Fig. 5B). This trend became significant after 5 wk (Density: [F.sub.1,9] = 9.80, P < 0.05; Light: [F.sub.1,9] = 2.92, P = 0.12, Interaction: [F.sub.1,9] = 0.407, P = 0.54) despite the loss of replication.

Food Quantity and Quality

Food abundance varied between treatments, as shown by the analysis of diatom densities (Table 2). Interestingly, while there was no significant three-way interaction with time, there was a significant interaction between density and food level. From Figure 6C, this seems to stem from the fact that diatom abundance differed more between densities at the high food level, than between densities at the low food level. Diatom abundance increased over time in the high food level treatments but was fairly constant in the low food treatments, which explains the time by food interaction. The smaller time by density effect is due mainly to the slower increase of diatom abundance in the high food treatment where density was high. Presumably this merely reflects the faster consumption of diatoms in this treatment.

[FIGURE 6 OMITTED]

The effect of food level and initial density on postlarval survival is shown in Figure 6A. For the first 14 days there was a steady decline of postlarval numbers in both high-density treatments, compared with a low and stable mortality rate in the lowdensity treatments. These trends seem to be independent of food availability. The repeated measures ANOVA (Table 2) reflects this pattern, as the food by time interaction and the three way interaction were not significant, whereas the time by postlarval density interaction was highly significant. From day 14-21 however, postlarval numbers also sharply decreased in the low-density treatment with excess food (Fig. 6A).

The growth of postlarvae differed significantly between density treatments and between food treatments (as shown by the time by density, and time by food tests in Table 2). Postlarvae in the low density and high food treatment were the biggest, whereas those at high initial density, but low food, were the smallest (Fig. 6B). There was no significant interaction between food and initial density (or three factor interaction with time), indicating that food level and initial density have additive effects on postlarval growth.

DISCUSSION

Behavior

There are several aspects of the observed grazing behavior of the postlarvae that seem to be important in explaining our results and results from other experiments on abalone postlarvae by Daume et al. (2004). First, postlarvae within a patch of food were efficient grazers, removing the majority of diatoms present, so that they created small patches cleared of food. Second, food consumption in a patch appears to vary directly with diatom density. Third, postlarvae that encountered areas without food began what appeared to be a random search pattern, with rapid movements.

Rapid movement in abalone and other gastropods has been shown to be a significant energetic cost (Donovan et al. 1999). This may be due to the high levels of anaerobic respiration involved, or the costs associated with mucus production for movement (Denny 1980), so that increased movement will have a significant impact on the balance between nutritional intake and energetic output (Carefoot 1987). This balance may be specially delicate in postlarvae, as compared with adults, because of the greater energetic costs associated with a large surface area to volume ratio as well as rapid morphologic development (Roberts et al. 1999a, Roberts et al. 1999b). As substantial time is spent in grazing even a small patch of algae, the time required to search for a patch may only be a small proportion of the total time spent consuming it. Nevertheless, in areas where patches of food are scarce, the sum of these erratic movements must be a significant energetic cost for the postlarvae, and impact on health and performances.

Postlarvae did not search for denser patches of diatoms. This is consistent with observations of other marine grazers and indicates an opportunistic approach to foraging (Hawkins & Hartnoll 1983). Random patterns of movement were also observed in the dispersal of postlarvae from the center of horizontal plates. If dispersal was based on the availability of food, then one might expect that postlarvae would move out in a noticeable "wave" as food became depleted in the center of the plate. The spread across the plate, with a higher proportion at the center even after 192 h, suggests instead a more random movement that was independent of food density. Dispersal was faster at higher initial densities of postlarvae. This pattern can be explained by our behavioral observations. Postlarvae at high density presumably more often encountered patches grazed by other postlarvae, which would lead to more time spent moving and thus a faster dispersal.

Similarly, postlarvae in high-density treatments would encounter many grazed patches, leading to relatively costly rapid movements. Where diatoms occur in patches of high and low density, postlarval nutrition may be even more variable, because some would spend long periods grazing areas with low food density. This food situation is likely to arise when an area has been grazed some time previously. Diatoms would have started to recolonize the grazed patches, but would not be at high density. It seems obvious that the more pregrazed areas encountered by the postlarvae, the lower the energy intake in proportion to the costs to the postlarvae.

Behavior may explain the distribution of postlarvae within the hatchery environment. Postlarvae are often grown on vertical plates covered in microalgae (Hahn 1989). Our observations of blacklip abalone on these plates a few weeks after settlement indicate a higher density on the top of the plates compared with the bottom. The top parts of the plates receive more light and are therefore likely to have denser food. Towards the bottom, grazed patches would not be quickly recolonized by diatoms, so that postlarvae may spend less time grazing in these areas. Their rapid searching would lead to dispersal from the lower parts of plates, whereas those at the top would spend more time grazing and thus move slowly. We cannot discount taxes in response to light or gravity, but in the natural environment postlarvae are found on crustose coralline algae that often have very rugose or uneven surfaces, and they occur on the tops or sides of rocks (McShane 1991, Shepherd & Turner 1985, Shepherd & Daume 1996). Thus, light and gravity may not be reliable cues to find food in natural situations.

Effect of Density on Postlarval Growth

As larvae from different batches may differ in quality (Martinez & Searcy-Bernal 1998, Daume et al. 2000), direct comparisons between batches must be examined with caution. Nevertheless, the mean size reached by postlarvae was negatively affected at high initial densities in both experiments. There are several mechanisms by which density could cause changes in postlarval size distributions. If size-dependant mortality was occurring, the death of smaller (or larger) individuals would skew the size distribution of the remaining postlarval population. This would give the impression of different growth rates between treatments. Because size-frequency distributions in the present experiments seemed normal, their small scale and the small number of individuals measured in each population prohibit definitive conclusions. Normal size distributions are usually observed in well-fed populations in the hatcheries however (Huchette 2003) suggesting that size-dependent mortality is unlikely, and the size differences reflect different growth rates.

This difference in growth rates continued to the end of the experiment despite the fact that differences in density between treatments were reduced by higher mortality in the high initial density treatments, so that densities were similar by the end of the experiment. A negative relationship between growth and density suggests there is density-dependent intraspecific competition for space or food (Stimson 1970, Branch 1975, Hughes 1986, Jarayabhand & Newkirk 1989, Parsons & Dadswell 1992, Foster & Stiven 1996). Direct competition for food has been shown for several marine gastropods (e.g., Stimson 1970). These are often limpets defending patches of algae near their home scars (Chapman & Underwood 1992). Direct competition is unlikely in abalone postlarvae due to their free-ranging foraging behavior, and no observations of aggressive confrontations were made throughout this study. The competitive interactions are therefore likely to be indirect.

Exploitative (indirect) competition for food has been shown in numerous benthic marine invertebrates, particularly herbivorous gastropods (Creese & Underwood 1982, Marshall & Keough 1994). It usually occurs when the abundance of a resource is depleted by consumers, and as a result this resource becomes limiting (Schmitt 1996). Food was a limiting factor because it significantly affected growth in both experiments. But even when food was clearly in excess in one of these experiments, the effect of density on growth appeared to have the same negative effect. It seems that postlarval food consumption was affected by the grazing activity of conspecifics despite an abundance of food.

Grazing would have increased the patchiness of food. Even when the densities were reduced by mortality in the high initial density treatments, the food would have been very patchy due to intensive previous grazing. Postlarvae entering recently grazed bare patches would have begun costly searching movements, whereas those encountering patches with a low diatom density would have consumed food at a slower rate. In both cases their growth would be reduced, and this form of exploitative competition, which essentially reduces access to food, would increase with density, and have persistent effects.

Effect of Density on Postlarval Survival

Density dependent mortality of postlarvae was found in both experiments, and has also been found in large-scale hatchery experiments, where food limitation at high density has been suggested as the underlying cause (Heasman et al. 2001, Daume et al. 2001, Daume et al. 2004). While food limitation may cause density dependent growth, our results suggest a different process causes the density dependent mortality.

Because alternative processes such as density-dependent predation (Shepherd & Daume 1996) or competition for space (McShane 1991) are absent in the hatchery, the density dependent mortality might be due to exploitation competition leading to either starvation, or a reduction in overall health that makes the postlarvae more susceptible to other agents of mortality. Density dependent mortality of this nature has been described for several species (Creese & Underwood 1982, Marshall & Keough 1994). One would expect such mortality to be delayed depending on the energetic reserves of the postlarvae at the start of the experiments, yet mortality in the high-density treatments actually occurs soon after the start of the experiments. It seems highly unlikely that starvation would take effect so quickly, and before differences in growth start to appear. In the first 2-3 wk after settlement, growth is minimally affected by the diatom composition because of the remaining yolk reserves (Daume et al. 2000, Roberts & Lapworth 2000) and the uptake of dissolved organic matter (Manahan & Jaeckle 1992, Shilling et al. 1996), so that effects on mortality are unlikely until after this time.

Density dependent mortality would be expected to be less intense, or take longer to come into effect where the food is more abundant, because the food intake of the abalone would be substantially higher. In these experiments, mortality at high initial postlarval densities occurred at the same rate, irrespective of food level. In a large scale experiment in the hatchery (Daume et al. 2004) mortality was shown to be higher overall in the excess food treatment. This suggests that mortality is unrelated to starvation, and that some other mechanisms were at work.

Contact between individuals reduces growth in juvenile abalone, perhaps because of elevated levels of stress (Huchette et al. 2003a, Huchette et al. 2003b). At higher densities this would occur more often and might be a cause of the density-dependent mortality observed. Contacts between postlarvae however, did not seem to affect their behavior and are unlikely to cause the sharp declines in survival recorded.

Pathogens such as protists, viruses, or bacteria are well known agents of mortality in aquaculture (Bower 1987, Shepherd & Breen 1992). Although we have no direct evidence for any pathogen, higher rearing densities imply an increased contact between individuals and a higher probability that one or more individuals carry a pathogen. Increased contact between individuals or with mucus trails thus increased transmission of a pathogen at high densities, and may be the cause of the mortality that is observed here. Bower (1987) notes that smaller abalone (<1.5 mm) were far more susceptible to a pathogenic protist, so that these postlarvae may be highly vulnerable.

These density-dependent processes may simply be an additional pressure that increases the postlarvae's susceptibility to other stresses. These stresses could include aspects such as oxygen or carbon dioxide supersaturation or depletion in the diffusive boundary layer above the substratum (Searcy-Bernal 1996). This effect might account for the greater mortality that occurred after 2 wk in the excess food treatments of the second experiment. The diatoms were very abundant in these treatments by this time, so that the diurnal oxygen production, or nocturnal consumption, of the algae may have contributed to the mortality in the postlarvae. Because the contribution of the biomass of postlarvae to oxygen concentrations is not likely to be significant in comparison to diatoms, and diatoms were reduced by high larval densities, the boundary layer effect alone cannot explain the density-dependent mortality.

We tentatively conclude that the effect of density on mortality may be due to a pathogen, or a complex interaction between the factors discussed above.

Implications

When considering commercial systems these results must be interpreted with care, due to the small scale at which these experiments were conducted. Whereas small containers such as petri dishes and jars may not closely match conditions in hatcheries, their use has been widespread in abalone aquaculture research (Kawamura et al. 1998a, Roberts et al. 1999a, Roberts et al. 1999b). The fact that they can be easily manipulated makes them ideal for experiments aiming at identifying what is happening at a process level. Here, the existence of density dependent growth and mortality has been demonstrated in larger scale experiments (Heasman et al. 2001, Daume et al. 2001, Daume et al. 2004). Using small containers has allowed us to confirm that intraspecific competition for food is the cause of reduced growth at high density in postlarval H. rubra. Identifying that density-dependent mortality is not due to a shortage of food is also of enormous value.

From an abalone farm perspective, the efficiency of production is paramount. This often requires a balance to be found between the net biomass yield, the time taken to grow the abalone to market size and the cost of infrastructure and labor. Whereas density was shown to have a significant impact on the growth of postlarvae, food level was more important. Careful monitoring and the regular addition of supplemental diatoms would allow farms to grow postlarvae in high densities while maintaining reasonably high growth rates. The causes of density-dependent mortality have not been adequately identified, so that it is hard to say what management strategies would be most effective, but our results should encourage more work in this area. Commercial scale trials (Heasman et al. 2001, Daume et al. 2004) indicate that although survival rates are poorer, the number alive after 2 mo is still greater at high initial densities. Because production of large numbers of larvae is becoming easier, settlement of larvae at high densities may be the most efficient mode of production.

Although natural conditions are also different from those used here, it seems possible that the same density dependent mechanisms would operate on postlarvae in small patches of habitat, and these may play a role in regulating natural populations. This study highlights the importance of a process level understanding of the mechanisms underlying density-dependent growth and survival in abalone and opens up a range of new questions on these processes. Possible pathogens of postlarvae and the role of the boundary layer in the early life stages deserve further research.

TABLE 1.

Repeated measures ANOVA testing for the effects of density and
shading on the log numbers surviving (week 0-5) and size (week
0-2) of post-larvae on plates over time. Probabilities were adjusted
using the Greenhouse-Geiser statistic.

                          Survival (Log Counts)

Source of Variation      Df       MS          F

Between plates
  Density                 1      1.00        0.19
  Light                   1      1.09        0.21
  Density x Light         1      9.38        1.80
  Residual               20      5.22
Within plates
  Time                    5     23.14       35.01 *
  Time x Density          5      2.34        3.55 *
  Time x Light            5      0.88        1.34
  Time x Density
    x Light               5      0.83        1.25
  Residual              100      0.66

                               Mean Size

Source of Variation      df       MS          F

Between plates
  Density                 1     29172        3.79
  Light                   1     65550        8.51 *
  Density x Light         1      5602        0.73
  Residual               13      7701
Within plates
  Time                    2    285272      100.95 *
  Time x Density          2      8022        2.84
  Time x Light            2     12336        4.37 *
  Time x Density
    x Light               2      1441        0.51
  Residual               26      2825

* P < 0.05.

TABLE 2.

Repeated measures ANOVA of diatom density (day 0-21), logged
post-larval counts (day 0-21) and mean size (day 0-14), testing
for the effects of density and food levels. Probabilities were
adjusted using the Greenhouse-Geiser statistic.

                               Diatom Density

Source of Variation        df       MS          F

Between Petri Dishes
  Density                   1       1239      5.4 *
  Food                      1     144746    637.0 *
  Density x Food            1       1471      6.5 *
  Residual                 16        227
Within Petri Dishes
  Time                      3       9243     62.4 *
  Time x Density            3        605      4.1 *
  Time x Food               3       8455     57.1 *
  Time x Density x Food     3        219      1.5
  Residual                 48        148

                             Survival (Log Nos)

Source of Variation        df       MS          F

Between Petri Dishes
  Density                   1      8.29      37.0 *
  Food                      1      0.01       0.1
  Density x Food            1      0.58       2.6
  Residual                 16      0.22
Within Petri Dishes
  Time                      3      7.23      69.8 *
  Time x Density            3      2.71      26.2 *
  Time x Food               3      0.31       3.0
  Time x Density x Food     3      0.33       3.2
  Residual                 48      0.10

                                  Mean Size

Source of Variation        df       MS          F

Between Petri Dishes
  Density                   1       9412      3.0
  Food                      1      29068      9.2 *
  Density x Food            1        204      0.1
  Residual                 16       3166
Within Petri Dishes
  Time                      2     115088    118.8 *
  Time x Density            2       3621      3.7 *
  Time x Food               2      13658     14.1 *
  Time x Density x Food     2        736      0.8
  Residual                 32        968

* P < 0.05.

LITERATURE CITED

Andrewartha, H. G. & L. C. Birch. 1970. The distribution and abundance of animals. Chicago, University of Chicago.

Bower, S. M. 1987. Pathogenicity and host specificity of Labyrinthuloides haliotidis (Protozoa: Labyrinthomorpha), a parasite of juvenile abalone. Can. J. Zool. 65:2008-2012.

Branch, G. M. 1975. Mechanisms reducing intraspecific competition in Patella spp.: migration, differentiation and territorial behaviour. J. Anim. Ecol. 44:575 600.

Carefoot, T. H. 1987. Diet and its effect on oxygen uptake in the sea hare Aplysia. J. Exp. Mar. Biol. Ecol. 114:275-288.

Chapman, M. G. & A. J. Underwood. 1992. Experimental designs for analyses of movements by molluscs. In: J. Grahame, P. J. Mill & D. G. Reid, editors. Proceedings of the 3rd International Symposium on Littorinid Biology. London: The Malacological Society of London. pp. 169 180.

Connell, J. H. 1985. The consequences of variation in initial settlement vs. post-settlement mortality in rocky intertidal communities. J. Exp. Mar. Biol. Ecol. 93:1145.

Creese, R. G. & A. J. Underwood. 1982. Analysis of inter- and intraspecific competition amongst limpets with different methods of feeding. Oecologia (Berlin) 53:337-346.

Daume, S., S. Brand-Gardner & W. J. Woelkerling. 1999. Settlement of abalone larvae (Haliotis laevigata Donovan) in response to nongeniculate coralline red algae (Corallinales, Rhodophyta). J. Exp. Mar. Biol. Ecol. 234:125-143.

Daume, S., S. M. H. Huchette, S. Ryan & R. W. Day. 2004. Nursery culture of Halitotis rubra: The effect of cultured algae and larval density on settlement and juvenile production. Aquaculture 236:221-239.

Daume, S., A. Krsinich, S. Farrell & M. Gervis. 2000. Settlement, early growth and survival of Halitotis rubra in response to different algal species. J. App. Phycol. 12:479-488.

Daume, S., S. Ryan, S. M. H. Huchette & R. W. Day. 2001. Settlement and growth of abalone on cultured algae in commercial abalone nurseries: a preliminary report. In: A. E. Fleming, editor. 8th Annual abalone aquaculture subprogram workshop. Canberra: Fisheries Research and Development Corporation. pp. 139-148.

Denny, M. 1980. Locomotion: the cost of gastropod crawling. Science 208:1288-1290.

Donovan, D., J. Baldwin & T. Carefoot. 1999. The contribution of anaerobic energy to gastropod crawling and a re-estimation of minimum cost of transport in the abalone, Haliotis kamtschatkana (Jonas). J. Exp. Mar. Biol. Ecol. 235:273-284.

Elmgren, R., G. Ejdung & S. Ankar. 2001. Intraspecific food competition in the deposit-feeding benthic amphipod Monoporeia affinis--a laboratory study. Mar. Ecol. Prog. Ser. 210:185-193.

Fogarty, M. J., M. P. Sissenwine & E. B. Cohen. 1991. Recruitment variability and the dynamics of exploited marine populations. Trends Ecol. Evol. 6:241-246.

Foster, B. A. & A. E. Stiven. 1996. Experimental effects of density and food on growth and mortality of the Southern Appalachian land gastropod, Mesodon normalis (Pilsbry). Am. Midl. Nat. 136:300-314.

Gaines, S. & J. Roughgarden. 1985. Larval settlement rate: a leading determinant of structure in an ecological community of the marine intertidal zone. Proc. Nat. Acad. Sci. USA. 82:3707-3711.

Guillard, R. R. L. & J. H. Ryther. 1962. Studies of marine planktonic diatoms. Can. J. Microbiol. 8:229-239.

Hahn, K. O. 1989. Handbook of culture of abalone and other marine gastropods. Boca Raton, Florida, CRC press.

Haldane, J. B. S. 1953. Animal populations and their regulation. New Biology 15:9-24.

Hawkins, S. J. & R. G. Hartnoll. 1983. Grazing of intertidal algae by marine invertebrates. Oceanogr. Mar. Biol. Ann. Rev. 21:195-282.

Heasman, M. P., C. P. Savva, C. P. Brand, R. C. Chick, D. G. Worthington & P. C. Gibson. 2001. Effect of seeding density on subsequent growth and survival of Haliotis rubra larvae set onto conventional diatom coated plates--a preliminary assessment. In: A. E. Fleming, editor. 8th Annual abalone aquaculture subprogram workshop. Canberra: Fisheries Research and Development Corporation. pp. 139-148.

Huchette, S. M. H. 2003. Density dependence in the early life stages of abalone. PbD thesis, University of Melbourne.

Huchette, S. M. H., C. S. Koh & R. W. Day. 2003a. Growth of juvenile blacklip abalone (Halitotis rubra) in aquaculture tanks: effects of density and ammonia. Aquaculture 219:457-470.

Huchette, S. M. H., C. S. Koh & R. W. Day. 2003b. The effects of density on the behaviour and growth of juvenile blacklip abalone (Halitotis rubra). Aquac. Int. 71:411-428.

Hughes, R. N. 1986. A functional biology of marine gastropods. London: Croom Helm.

Hunt, H. L. & R. E. Scheibling. 1997. Role of early post-settlement mortality in recruitment of marine invertebrates. Mar. Ecol. Prog. Ser. 155: 267-301.

Hurlbut, C. J. 1991. Community recruitment: settlement and juvenile survival of seven co-occurring species of sessile marine invertebrates. Mar. Biol. 109:507-515.

Jarayabhand, P. & G. F. Newkirk. 1989. Effects of intraspecific competition on growth of the European oyster, Ostrea edulis Linnaeus, 1750. J. Shellfish Res. 8:359-365.

Kawamura, T., R. D. Roberts & C. M. Nicholson. 1998a. Factors affecting the food value of diatom strains fur post-larval abalone Haliotis iris. Aquaculture 160:81-88.

Kawamura, T., R. D. Roberts & H. Takami. 1998b. A review of the feeding and growth of postlarval abalone. J. Shellfish Res. 17:615-625.

Keough, M. J. & B. J. Downes. 1982. Recruitment of marine invertebrate: the role of active larval choices and early mortality. Oecologica 55: 348 352.

Kinne, O. 1973. Marine Ecology, Vol. 1. London: Wiley.

Leighton, D. L. 1974. The influence of temperature on larval and juvenile growth in three species of southern California abalones. Fish. Bull. 72:1137-1145.

Manahan, D. T. & W. B. Jaeckle. 1992. Implications of dissolved organic material in seawater for the energetics of abalone larvae Haliotis rufescens: a review. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford: Fishing News Books. pp. 95-106.

Marshall, P. A. & M. J. Keough. 1994. Asymmetry in intraspecific competition in the limpet Cellana tramoserica (Sowerby). J. Exp. Mar. Biol. Ecol. 177:121-138.

Martinez, P. D. R. & R. Searcy-Bernal. 1998. Grazing rates of red abalone (Haliotis rufescens) postlarvae feeding on the benthic diatom Navicula incerta. J. Shellfish Res. 17:627-630.

McGarvey, R. 1996. Is long-term persistence of harvested populations evidence of density dependence? In: R. B. Floyd, A. W. Sheppard & P. J. De Barro, editors. Frontiers of population ecology. Melbourne: CSIRO Publishing. pp. 65-77.

McShane, P. E. 1991. Density-dependent mortality of recruits of the abalone Halitotis rubra (Mollusca: Gastropoda). Mar. Biol. 110:385-389.

McShane, P. E. 1992. The early life history of abalone. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford: Fishing News Books. pp. 120-138.

McShane, P. E. 1995. Recruitment variation in abalone: Its importance to fisheries management. Mar. Freshw. Res. 46:555-570.

McShane, P. E., K. P. Black & M. G. Smith. 1988. Recruitment processes in Halitotis rubra (Mollusca: Gastropoda) and regional hydrodynamics in southeastern Australia imply localized dispersal of larvae. J. Exp. Mar. Biol. Ecol. 124:175-203.

McShane, P. E. & J. R. Naylor. 1995. Density independent growth of Haliotis iris Martyn (Mollusca: Gastropoda). J. Exp. Mar. Biol. Ecol. 190: 51-60.

McShane, P. E. & N. Reyn. 1996. Variation in spawning and recruitment of Haliotis iris (Mollusca: Gastropoda). N. Z. J. Mar. Freshw. Res. 30: 325-332.

McShane, P. E. & M. G. Smith. 1991. Recruitment variation in sympatric populations of Halitotis rubra (Mollusca: Gastropoda) in southeast Australian waters. Mar. Ecol. Prog. Ser. 73:2-3.

Nash, W. J., J. C. Anderson, J. Bridley, D. Wideley, S. Dickson & B. Hislop. 1995. Post-larval recruitment of blacklip abalone (Haliotis rubra) on artificial collectors in southern Tasmania. Mar. Freshw. Res. 46:531-538.

Nicholson, A. J. 1933. The balance of animal populations. J. Anim. Ecol. 2 (Supplement): 131-178.

Paine, R. T. 1984. Ecological determinism in the competition for space. Ecology 65:1339-1348.

Parsons, G. J. & M. J. Dadswell. 1992. Effect of stocking density on growth, production and survival of the giant scallop, Placopecten magellanicus, held in intermediate culture in Passamaquoddy Bay, New Brunswick. Aquaculture 103:291-309.

Preece, P. A., S. A. Shepherd, S. M. Clarke & J. K. Keesing. 1997. Abalone stock enhancement by larval seeding: effect of larval density on settlement and survival. Moll. Res. 18:265-273.

Prince, J. D., T. L. Sellers, W. B. Ford & S. R. Talbot. 1987. Experimental evidence for limited dispersal of haliotid larvae (genus Haliods; Mollusca: Gastropoda). J. Exp. Mar. Biol. Ecol. 106:243-263.

Prince, J. D., T. L. Sellers, W. B. Ford & S. R. Talbot. 1988. Confirmation of a relationship between the localized abundance of breeding stock and recruitment for Halitotis rubra Leach (Mollusca: Gastropoda). J. Exp. Mar. Biol. Ecol. 122:91-104.

Roberts, R. & C. Lapworth. 2000. Starvation tolerance of post-larval abalone (Haliotis iris). J. Shellfish Res. 19:529.

Roberts, R.D., T. Kawamura & C. M. Nicholson. 1999a. Growth and survival of postlarval abalone (Haliotis iris) in relation to development and diatom diet. J. Shellfish Res. 18:243-250.

Roberts, R. D., T. Kawamura & H. Takami. 1999b. Morphological changes in the radula of abalone (Haliotis iris) during post-larval development. J. Shellfish Res. 18:637-644.

Roegner, G. C. 1991. Temporal analysis of the relationship between settlers and early recruits of the oyster Crassostrea virginica (Gmelin). J. Exp. Mar. Biol. Ecol. 151:57-69.

Rose, K. A., J. H. Cowan, K. O. Winemiller, R. A. Myers & R. Hilborn. 2001. Compensatory density dependence in fish populations: importance, controversy, understanding and prognosis. Fish and Fisheries 2:293-327.

Roughgarden, J. S., S. Gaines & H. Possingham. 1988. Recruitment dynamics in complex life cycles. Science 241:1460-1466.

Schmitt, R. J. 1996. Exploitation competition in mobile grazers: trade-offs in use of a limited resource. Ecology 77:408-425.

Searcy-Bernal, R. 1996. Boundary layers and abalone postlarval culture: Preliminary studies. Aquaculture 140:129-137.

Shepherd, S. A. & P. A. Breen. 1992. Mortality in abalone: its estimation, variability, and causes. In: S.A. Shepherd, M.J. Tegner & S.A. Guzman del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford, UK: Fishing News Books. pp. 276-304.

Shepherd, S. A. & S. Daume. 1996. Ecology and survival of juvenile abalone in a crustose coralline habitat in South Australia. In: Y. Watanabe, Y. Yamashita & Y. Oozeki, editors. Survival strategies in early life stages of marine resources. Yokohama, Japan: Balkema. pp. 297-313.

Shepherd, S. A. & K. R. Rodda. 2001. Sustainability demands vigilance: evidence for serial decline of the South Australian greenlip abalone fishery and a review of management. J. Shellfish Res. 20:829-841.

Shepherd, S. A. & J. A. Turner. 1985. Studies on southern Australian abalone (genus Haliotis). VI. Habitat preference, abundance and predators of juveniles. J. Exp. Mar. Biol. Ecol. 93:285-298.

Shilling, F. M., H. O. Guldberg & D. T. Manahan. 1996. Sources of energy for increased metabolic demand during metamorphosis of the abalone Haliotis rufescens (Mollusca). Biol. Bull. 191:402-412.

Stimson, J. 1970. Territorial behaviour of the owl limpet Lottia gigantea. Ecology 51:113-118.

Stubbs, M. 1977. Density dependence in the life cycles of animals and its importance in k- and r-strategies. J. Anim. Ecol. 46:677-688.

Underwood, A. J. & P. G. Fairweather. 1989. Supply-side ecology and benthic marine assemblages. Trends Ecol. Evol. 4:16-20.

ROB DAY, * PATRICK GILMOUR, AND SYLVAIN HUCHETTE

Zoology Department, The University of Melbourne, Parkville, Victoria 3010, Australia

* Corresponding author. E-mail: [e-mail].