Settlement of abalone (Haliotis iris) larvae in response to five species of coralline algae.

ABSTRACT Abalone recruits are found predominantly on crustose coralline algae (CCA) and recruitment seems to vary among coralline species or growth-forms. Spatial patterns in the recruitment of marine invertebrates can be formed by processes acting before, during, or after settlement. This study used laboratory experiments to examine the settlement of Haliotis iris larvae on five species of CCA. When the CCA species were presented individually to larvae, all induced >88% of larvae to attach within 1 day, and >80% to metamorphose within 3 days. When given a choice of the five CCA, larvae settled most on Pneophyllum coronatum and Hydrolithon rupestris, and settled least on Mesophyllum printzianum. Phymatolithon repandum, and Lithophyllum carpophylli gave intermediate results. The speed of metamorphosis in the no-choice experiments mirrored the species preferences in the choice experiments, with metamorphosis occurring most rapidly on Pneophyllum coronatum and Hydrolithon rupestris. The two preferred species were thin crusts, whereas the two least preferred species formed thick, morphologically complex crusts. In a choice experiment, the proportion of larvae choosing a CCA specimen showed no correlation with bacterial density (r = -0.25, P = 0.52, n = 25), but a weak positive correlation with diatom density (r = 0.52, P = 0.02, n = 25). However, diatom densities were low on all CCA species, and are unlikely to be a primary cause of settlement preference. It is hypothesized that at least some of the variation in recruitment among CCA species or growth forms observed in the wild is determined by selective settlement.

KEY WORDS: abalone, biofilm, coralline algae, larval settlement, metamorphosis, recruitment.

INTRODUCTION

Patterns of recruitment in marine larvae may be caused by processes occurring before, during, or after settlement (Connell 1985). The relative importance of these stages in determining recruitment varies among species and locations (e.g., Keough & Downes 1982, Connell 1985, Gaines & Roughgarden 1985, Woodin 1986, Rowley 1989, McShane 1992, Stoner et al. 1996, Walters & Wethey 1996, Mundy & Babcock 2000).

Abalone recruits are found almost exclusively on crustose coralline algae (CCA) (Shepherd & Turner 1985, McShane & Smith 1988, Day & Branch 2000). Based on this selection of a particular habitat it was initially suggested that differential postsettlement mortality is unlikely to be important in determining abalone recruitment (McShane 1991, McShane 1992). However, the broad grouping "crustose coralline algae" includes a wide range of species and growth forms (Woelkerling et al. 1993). Recent research on abalone ecology has recognized this complexity and demonstrated that the species and growth form of CCA can influence the number of abalone recruits (McShane 1996, Shepherd & Daume 1996, Daume et al. 1999, Day & Branch 2000). In South Australia, "lumpy" and "uneven encrusting" growth forms had more recruits of Haliotis laevigata and Haliotis scalaris than other coralline growth forms (Shepherd & Daume 1996) and in South Africa, "knobbly" corallines contained the highest densities of Haliotis midae recruits (Day & Branch 2000).

Such patterns could be caused by selective settlement or differential postsettlement survival. Work to date has focused on the latter. It has been suggested that morphologically complex CCA growth forms may be more suitable than flat crusts because they provide greater protection of postlarvae from wave shear (postulated by McShane & Naylor 1995 and McShane 1996, but not supported by experimental evidence of Naylor & McShane 1997a) inadvertent ingestion (Fletcher 1987, McShane 1991), predatory fish (McShane 1996) or because they protect pockets of microbial food from access by larger grazers (Shepherd & Daume 1996, Day & Branch 2000). On the other hand, topographically complex surfaces may increase the thickness of the diffusive boundary layer by impairing water movement near the substrate, resulting in more extreme fluctuations in water chemistry in the postlarval environment (Kaspar 1992, Day & Branch 2000). Morphologically complex CCA growth forms generally grow as thick crusts. These harbor predatory worms that can decimate populations of newly settled abalone (Morse et al. 1979, Naylor & McShane 1997a, Naylor & McShane 1997b).

Coralline algae are among the most effective larval settlement cues for all species of abalone tested to date (Roberts 2001a). Other natural surfaces such as biofilms can induce larval metamorphosis, but few approach the speed and consistency of CCA (Roberts 2001a, Roberts 2001b). Little work has been done on the extent to which settling abalone larvae discriminate among coralline species or growth forms.

Discrimination among substrata is well demonstrated for settling marine invertebrate larvae (reviewed by Crisp 1974, Pawlik 1992, Rodriguez et al. 1993). Some studies have demonstrated settlement preference in multiple choice experiments (e.g., Kirchman et al. 1982, Miron et al. 1996, Stoner et al. 1996, Wieczorek & Todd 1997, Snelgrove et al. 1998, Daume et al. 1999). These choice experiments test the relative preference of larvae for a particular substrate when faced with a range of options, whereas experiments that present cues individually test the absolute inducing ability of a substrate for larvae that are given no other option. The former is more relevant to the situation faced by a larva in nature.

The few data available for abalone larvae provide conflicting evidence on selectivity among types of CCA. Morse et al. (1980) and Morse and Morse (1984) reported equally strong settlement in response to various species of CCA, but the species were presented individually rather than as multiple choices. In contrast, larvae of Haliotis laevigata metamorphosed in greater numbers on the thick, lumpy crusts of Sporolithon durum than on the encrusting Hydrolithon rupestre (= Hydrolithon rupestris) or the warty Mesophyllum engelhartii (Daume et al. 1999). Larvae also metamorphosed more quickly on Sporolithon with a lumpy growth form than the same species with an encrusting growth form. Both patterns were consistent in choice and no-choice experiments.

Data comparing the performance of types of CCA for the settlement of invertebrates other than abalone are also sparse. The scleractinian coral Agaricia agaricites humilis metamorphosed strongly in response to only 3 of 6 CCA species presented individually (Morse et al. 1988). The soft coral Alcyonium siderium metamorphosed in response to both of two CCA species (Sebens 1983).

More data are required to assess the importance of settlement preferences in determining distributions of recruits among types of CCA. The ecological justification for such studies is compelling. CCA are abundant from tropical to polar waters (Johansen 1981, Steneck 1986) occupying more hard substratum within the photic zone than any other organism (Steneck & Paine 1986). CCA trigger the settlement of numerous species of marine invertebrate larvae (Pawlik 1992, Johnson et al. 1991) and difficulties with the taxonomy and identification of CCA have obscured the species specificity of many interactions involving CCA (Steneck & Paine 1986).

The present study compares abalone larval settlement on 5 CCA species with diverse growth forms, in choice and no-choice experiments. It was found that all CCA species induced metamorphosis if presented individually, but that larvae exhibited preferences when given a choice.

METHODS

Larval Rearing and Bioassays

Ripe adult Haliotis iris Gmelin 1791 were spawned by the hydrogen peroxide technique (Morse et al. 1977) and the larvae reared in flowing seawater as described previously (Roberts & Nicholson 1997) until taken for assays. Competent larvae were added to the wells of tissue culture plates (Falcon 3043 or 3047) containing CCA in 2 mL (no choice experiments) or 3.3 mL (choice experiments) of 0.2 [micro]m-filtered seawater with 150 [micro]g/mL of both penicillin G sodium (Sigma) and streptomycin sulphate (Sigma). Assay plates were wrapped in foil and incubated at 17 [+ or -] 0.5[degrees]C. To quantify settlement, the whole of each CCA fragment and its well were examined microscopically (Roberts & Nicholson 1997) after ~1 and 2.5 to 3 days of incubation. Counts were converted to percentage of larvae showing: (a) "shell growth" (velum shed and peristomal growth visible); (b) "metamorphosis" (velum shed, with or without shell growth); and (c) "attached" (metamorphosed or attached by foot). "Settlement" is used as a general term encompassing larval attachment and metamorphosis (Roberts 2001a). Negative controls of filtered seawater (with clay in Experiment 3) showed 0% to 0.3% metamorphosis within 3 days (data not presented). Experimental details are presented in Table 1.

Collection and Preparation of CCA

Details and taxonomic authorities of the five CCA species used are given in Table 2. Fragments of ~0.15 to 0.3 [cm.sup.2] "floor area" (defined later) were prepared by cutting the plastic on which they were growing (for the fragile Pneophyllum coronatum and Hydrolithon rupestris) or breaking off fragments of the plant with a knife (Lithophyllum carpophylli and Mesophyllum printzianum) or a hammer and chisel (Phymatolithon repandum). Each assayed fragment was from a separate plant, and each plant of Lithophyllure carpophylli was from a separate host plant of Carpophyllum maschalocarpum.

In Experiments 1 and 3 the fragments of CCA were held in filtered seawater for <30 min before being moved into bioassay chambers. In Experiment 2, fragments were soaked in ~1L of static 1-[micro]m filtered seawater for 4 days at 17[degrees]C before bioassay, with the intention of removing dissolved components leaking from broken edges of the CCA. This holding period without grazers would also alter the biofilm on the CCA (not quantified). In Experiment 3, the floor of each bioassay chamber was filled with a flattened plug of "colour clay" (Trident Toy Co. Ltd, Hutton, Avon, England) to -2.5 mm depth and the CCA fragments embedded in the clay so that their upper (live) surface was flush with the surface of the clay. The clay was molded to fit closely the edges of the samples. The intention of this treatment was to make all CCA species equally accessible to crawling larvae (see also Discussion).

For choice experiments, the five CCA fragments in a replicate were randomly placed in a circle on a radius two-thirds that of the tissue culture plate well (Falcon 3043), with the center of the well unoccupied. For no-choice experiments the CCA fragment was placed in the center of the tissue culture plate well (Falcon 3047) with species randomly arranged in the plates.

Phymatolithon repandum and Mesophyllum printzianum have uniporate gametangial conceptacles and multiporate tetrasporangial conceptacles, making these life cycle stages easy to distinguish. A previous experiment showed that both life cycle stages induced equally strong larval attachment and metamorphosis within 60 h (data not shown), so life cycle stage of CCA was excluded from further consideration.

For Choice Experiment 1, fragment size was estimated by measuring length and width of fragments with an eye-piece micrometer in a binocular microscope. In Experiment 3, the size of CCA fragments was determined by image analysis (Optimas 6) of videorecords. Fragment size was measured as the area of the CCA when viewed from above (= "floor area"), taking no account of the greater surface area of the topographically complex corallines (notably Mesophyllum printzianum). The floor area measurement was regarded as the most appropriate measure of "target size" (see Discussion).

Scanning Electron Microscope Analysis of Biofilms

Samples of coralline algae were prepared for scanning electron microscope (SEM) analysis by 2 methods. A simple preparation procedure (termed "SEM method 1") was designed to minimize loss of the microbial population on the CCA surface. A more complex procedure (termed "SEM method 2") was followed to preserve the form of microbes and other fine structures on the surface of the corallines.

SEM method 1: for each replicate in Experiment 3, a sample from the same plant was preserved in 2% glutardialdehyde (Merck) then air dried in a laminar flow cabinet. Dried samples were attached to aluminum stubs with double-sided carbon tape and silver paint then sputter-coated with gold/palladium before viewing with a Cambridge $360 Stereoscan SEM. On each sample, bacteria were counted in 12 x 5000 magnification fields of view (each 24 [micro]m x 17 [micro]m) and diatoms in 12 x 1000 fields of view (each 121 [micro]m x 82 [micro]m). Samples were scanned systematically, and nonoverlapping fields of view were haphazardly located along the scan path. Additional samples of each CCA species, collected from the same locations, were fixed in 4% formalin in seawater, air dried, sputter coated, and viewed with a Hitachi 5-2250N SEM. Diatoms were counted in 12 x 1000 fields of view (each 127 x 98 [micro]m).

SEM method 2: Samples of each species of CCA from the same sources were placed directly into foil-wrapped containers with 2.5% glutardialdehyde in the field. A day later samples were transferred to 1% osmium tetraoxide for 24 h, then desalinated with a graded series of 0, 25, 50, 75, and 100% distilled water in filtered seawater (15 min per step), dehydrated with a graded series of 10, 30, 50, 70, 90, 95, and 100% ethanol in distilled water (15 min per step), and critical point dried (Hayat 1989, Johnson et al. 1991). Samples were coated and scanned as earlier.

Statistical Analyses

Correlations were tested among planned comparisons using the Dunn-Sidak method to maintain the experiment-wise Type 1 error rate at 0.05 (SPSS Inc. 1999). Diatom and bacterial densities used in correlations related only to the surface of the CCA fragments, ignoring the surrounding surfaces. In data from the choice experiments, the ANOVA assumption of independent errors was violated--knowing that a larva settled on CCA species A also tells us that it did not settle on the other species, so the errors are correlated. Settlement preferences across the three replicated choice experiments were analyzed by fitting a log-linear model to the 3-way classification of CCA species (5 levels), experiments (3 levels), and replicates within experiments (6 levels). Only those animals that settled on CCA were used in the analysis, and it was assumed that the effects of day and replicate on settlement preference were additive rather than interactive (i.e., the 3 way interaction term was omitted from the model). The analysis used a quasi-likelihood approximation to an over-dispersed log-linear model, with random experiment and replicate effects (Link 1999). This approach treats the random effects as fixed effects, and assumes that larvae behave independently. This leads to an intentionally over-dispersed model, where the over-dispersion is caused by experiment and replicate effects and/or by nonindependent behavior among larvae. There is a corresponding intentional underestimation of the sampling variance and test statistics, which are then adjusted using the quasi-likelihood adjustment factor (= ratio of goodness of fit statistic to its degrees of freedom) (Link 1999). Lack of a significant replicate effect after adjustment for over-dispersion (see Results) allowed the data to be reduced to a 2-way classification (5 CCA species, 3 experiments). To obtain selection probabilities for each CCA species, the data from the 3 experiments were combined by a weighted average that took into account the variance within each experiment (Burnham et al. 1987). The reported 95% confidence intervals include variance components from the estimation of the selection probabilities within experiments, and the between-experiment variation.

RESULTS

Experiments With a Choice of Coralline Species

In Choice Experiment 1, Hydrolithon rupestris and Pneophyllure coronatum attracted more than twice as many settling larvae as the other three CCA species (Fig. 1A). Few larvae (0.7% of total) settled on Mesophyllum printzianum. Almost all of the larvae had metamorphosed, and most had begun to grow postlarval shell, within 22 h. Nearly half of the larvae were found on the floor or walls of the assay chamber, rather than on the CCA themselves (Fig. 1A). Fragment size was almost identical in four of the CCA species (averages of 0.194 to 0.197 [+ or -] 0.02 [cm.sup.2]) whereas the most preferred species for settlement, Hydrolithon rupestris, had slightly smaller fragments (0.15 [+ or -] 0.03 [cm.sup.2]).

[FIGURE 1 OMITTED]

In Choice Experiment 2 the order of preference for the CCA species was almost the same as in Choice Experiment 1 except that Hydrolithon rupestris was less preferred, slipping to second choice and comparable to Phymatolithon repandum (Fig. 1B). All larvae on CCA had metamorphosed, but of the 24% not on CCA, 8% had attached but not metamorphosed, and 7% had not yet attached (Fig. 1B). The proportion of larvae not on CCA was about half of that in Choice Experiment 1 (Fig. 1A).

In Choice Experiment 3, the order of preference for CCA species was again similar with Pneophyllum coronatum and Hvdrolithon rupestris clearly preferred (Fig. 1C, D). Phymatolithon repandum was less preferred in this experiment than previously. Larvae were slower to metamorphose in this experiment, with 46% of larvae attached (mostly to CCA fragments) but not metamorphosed after 23 h. However, the majority of abalone found on the two preferred species, Pneophylhnn coronatum and Hwlrolithon rupestris, had metamorphosed within 23 h. The size of CCA fragments varied little among the CCA species (Fig. 2C), and showed no correlation with the proportion of settlers (r = 0.02, P = 0.999, n = 30).

[FIGURE 2 OMITTED]

Analysis of the data from the three choice experiments demonstrated a strong difference in settlement probabilities among CCA species ([F.sub.12, 40] = 17.55, P < 0.0001). Pneophyllum coronatum and Hydrolithon rupestris attracted significantly more settlers than the other CCA species (Fig. 3). The goodness-of-fit statistic from the model omitting the 3-way interaction term provided evidence of over-dispersion ([[chi square].sub.40] = 174.27, p < 0.001) indicating that either substrate settlement probabilities varied among experiments and/or replicates, or that larvae were not behaving independently. After accounting for over-dispersion, there was little evidence of variation between replicates ([F.sub.20,40] = 1.192, p = 0.31) but strong evidence of variation in settlement probabilities between days ([F.sub.8,40] = 3.081, p = 0.008).

[FIGURE 3 OMITTED]

Biofilms on Coralline Species

The proportion of larvae choosing a CCA specimen in Choice Experiment 3 showed a weak positive correlation (r = 0.52, P = 0.02, n = 25) with the density of diatoms on that plant (Fig. 2A). In contrast, the proportion of settling larvae showed a nonsignificant negative correlation (r = -0.25, P = 0.52, n = 25) with the density of bacteria (Fig. 2B). Bacterial and diatom densities varied greatly between plants within species and between quadrats within plants. Mean diatom densities were low, ranging from zero to only 3 x [10.sup.4] [cm.sup.-2] (Fig. 2A, Table 3). Bacterial densities ranged from 0.2 to 8 x [10.sup.6], [cm.sup.-2] (Fig. 2B).

On Mesophyllum printzianum there were clearly demarcated areas with either few bacteria and little surface polysaccharide, or many bacteria among copious polysaccharide (Fig. 4A to D). Diatoms were seen only as one patch on one of the five Mesophyllum printzianum plants examined. Mesophyllum printzianum previously collected from the same location had dense patches of diatoms in grazing refugia beside protruberances or in sunken conceptacle pore-plates (Fig. 4 E, F). However, searching with SEM confirmed the absence of such patches of diatoms on the plants used in this Choice Experiment 3.

[FIGURE 4 OMITTED]

Lithophyllum carpophylli from Experiment 3 also showed biofilm patchiness on a scale of hundreds of micrometres. Bacteria were much more abundant in the cavities formed by cells that had lost their exterior wall (Fig. 5A). Few bacteria were seen on areas of thallus where the external wall of the cells was flush with the thallus surface (no cell cavities present). However, on another occasion this did not hold true (Fig. 5B). No diatoms were seen on the plants of Lithophyllum carpophylli used in Choice Experiment 3, either in the 60 diatom quadrats (Fig. 2A), or in visual scanning the plants with SEM. On another occasion, variable densities of diatoms were found on this CCA species (Table 3).

[FIGURE 5 OMITTED]

Bacteria were several times less abundant on Phymatolithon repandum than on the other coralline species (Figs. 2B, 5C). The surface of this species typically had extensive fine debris that resulted from disintegration of the calcified lateral cell walls (Fig. 5C). Although this may obscure some bacteria from view, it was obvious that the bacterial density was low. Diatoms were found on only two of the five plants inspected, and the average diatom density was low (Fig. 2A). On other occasions, high diatom densities were seen on Phymatolithon repandum, but these appeared to be removed by epithallial sloughing (Fig. 5D). Extensive grazing marks were often visible on this species (Fig. 5E) and microbial densities were low in areas scarred by grazing.

Pneophyllum coronatum and Hydrolithon rupestris had variable densities of diatoms (Figs. 2A, B, Table 3) and bacteria (Fig. 5F, H). On Pneophyllum coronatum there were fewer microbes on young tissue close to the growing edge of the plant (Fig. 5G).

The diatom assemblage on most plants was dominated by Cocconeis spp., Amphora spp. and other naviculoid diatoms (Table 3). On one occasion Tabularia sp. was abundant in grazing refugia on Mesophyllum printzianum.

No-choice E.rperiments With Coralline Species Assayed Individually

When the five species of CCA were presented individually to larvae, all induced a high proportion of larvae to attach and metamorphose. There were some differences in the timing of responses among CCA species and among experiments (Figs. 6, 7). Those species that induced metamorphosis most rapidly (Pneophyllum coronatum and Hydrolithon rupestris) (Figs. 6, 7) were the species preferred in the choice experiments (Figs. 1 and 3).

[FIGURES 6-7 OMITTED]

In the first experiment, the mean percent metamorphosis within 20 h was [greater than or equal to] 95% on Pneophyllum coronatum and Hydrolithon rupestris, but only 57% on Mesophyllum printzianum and 68% on Phymatolithon repandum (Fig. 6A). By 60 h, most of the remaining larvae had metamorphosed (Fig. 6B). The same pattern was evident in No-choice Experiment 3, where metamorphosis occurred more slowly (Fig. 7A, B).

Movement off CCA

The animals that were not on CCA after 20-24 h were mostly unmetamorphosed (Fig. 8). The exception was in Choice Experiment 1, where a large proportion of larvae were found metamorphosed, and not on CCA (Fig. 1A). By 2.5 to 3 days there was an increase in the total number of larvae not on CCA and a predominance of animals with postlarval shell growth, which had metamorphosed on CCA then migrated off (Fig. 8).

[FIGURE 8 OMITTED]

DISCUSSION

Strong Settlement on all CCA Species Tested

When presented individually, all five species of coralline induced over 80% of Haliotis iris larvae to metamorphose within 3 days. There was variation in the speed of the response among CCA species, and between the two experiments (Figs. 5, 6) but all CCA species were effective inducers of metamorphosis. When presented with a choice of 5 CCA species, Haliotis iris larvae preferentially settled on certain species (Fig. 2), and the preferred species were those that induced metamorphosis most quickly in the no-choice experiments.

The strong settlement response of Haliotis iris larvae to all CCA tested contrasts with the major differences among CCA species reported for Haliotis laevigata (Daume et al. 1999) and a scleractinian coral (Morse et al. 1988), but is consistent with the response of Haliotis rufescens (Morse et al. 1980, Morse & Morse 1984) and a soft coral (Sebens 1983) to multiple CCA species.

Rejection Rates of CCA Underlie Settlement Preferences

Given that the chance of larval encounter is equivalent for the various CCA species (Fig. 2C), the existence of settlement preferences implies that larvae were more likely to reject certain species after contact. This preference was exercised even though all species were strong inducers of metamorphosis when presented alone (Figs. 6, 7). The more rapid metamorphosis seen on Pneophyllum coronatum and Hydrolithon rupestris in the no-choice experiments is consistent with the greater likelihood of larvae "accepting" these substrata when contacted by larvae.

Factors Controlling Acceptance or Rejection of CCA

There are various possible causes for the preferential settlement on certain CCA species. Some potential causes relate to factors that covary with CCA species, rather than being a function of the species itself.

The settlement preferences we observed could result from differences in the settlement-inducing chemicals of the different CCA species (McShane 1996). The specific chemicals involved, the quantities produced, or their availability at the surface of the plant could all vary within and among CCA species, but there is no direct evidence to support or refute this. The recent identification of an abalone metamorphosis-inducing molecule from a CCA (Suenaga et al. 2004) provides an analytical target to begin exploring this question.

The biofilm on CCA is a potential determinant of settlement inducing activity (Johnson et al. 1991, Johnson & Sutton 1994). In the present study, the proportion of larvae choosing a CCA specimen showed no correlation with bacterial density, but a weak positive correlation with diatom density. The density of diatoms on CCA was generally low, with even the highest densities barely reaching levels known to induce high levels of abalone larval attachment (Kawamura & Kikuchi 1992, Roberts 2001a). However, diatoms can combine synergistically with other cues (Roberts & Nicholson 1997, Roberts 2001a) and the densities required to be effective in such a synergism might be much lower.

There was a relationship between the morphology of the corallines and the settlement preferences in choice experiments. The two most preferred species (Pneophyllum coronatum and Hydrolithon rupestris) were both thin encrusting corallines, the intermediate choice (Phymatolithon repandum) was a slightly thicker encrusting to warty growth form, whereas the two least preferred species (Lithophyllum carpophylli and Mesophyllum printzianum) were thick crusts (Table 2). It is interesting to consider whether these choices might make any ecological sense. Thick crusts of Mesophyllum printzianum occur on stable rock substrata (Table 2, Naylor & McShane 1997a) whereas the thin encrusting Pneophyllure coronatum and Hydrolithon rupestris seem to be opportunistic species (author's personal obs.), such as found on mobile rocks in shallow water (McShane 1996, Naylor & McShane 1997a). Experimental evidence shows that the thick-crusted Mesophyllum printzianum hosts polychaete predators that kill many recently settled abalone (Naylor & McShane 1997a, Naylor & McShane 1997b) and that its surface topography offers no demonstrable protection from dislodgement of Haliotis iris postlarvae by wave shear (Naylor & McShane 1997b). Juvenile Haliotis iris (10-70 mm shell length) are found during daylight hours under rocks that have undersides free of bulky encrusting organisms (McShane et al. 1995, author's personal obs.). Such rocks are somewhat mobile, occur in shallow water and are commonly colonized by thin crusts of CCA (McShane 1996, Naylor & McShane 1997a) similar to Pneophyllum coronatum and Hydrolithon rupestris. McShane (1996) remarked on the low frequency of Haliotis iris recruits on thick growth forms of CCA and suggested that lower concentrations of settlement inducers or the presence of predatory polychaetes were possible causes. Thus the preference for thin crusts that we observed in Haliotis iris larvae has potential benefits for both postlarvae and cryptic juveniles, though the suggestion of any link is merely speculative.

There has been some discussion about the effect that CCA growth form has on the biofilm present. Morphologically complex growth forms might protect areas from access by larger grazers, leaving remnants of biofilm as food for abalone recruits (Shepherd & Daume 1996, Day & Branch 2000) or as cues for larval settlement. This supposition is not supported by the observation that the lumpy coralline Sporolithon durum had no more biofilm than encrusting or warty forms (Shepherd & Daume 1996, Daume et al. 1999). However, the "lumps" on Sporolithon durum are several mm diameter, so would not restrict access for smaller grazers, and the sparse biofilm on this coralline probably results from its extensive epithallial sloughing (Shepherd & Daume 1996). In the present study, the effect of grazing refuges was clearly visible on some specimens of the fruticose Mesophyllum printzianum, where pockets of diatoms proliferated in the narrow gaps between protruberances or the sunken pore plates of conceptacles (Fig. 4E, F). Sloughing was also evident in some species (Fig. 5D) and contributed to the extreme variability in biofilm composition. This variability suggests that differences in biofilm characteristics among CCA species are unlikely to explain consistent patterns in abalone settlement or postsettlement survival.

The CCA species used in the present study were from different locations, or different microhabitats within those locations (Table 2). Thus environmental effects such as growing conditions or grazing history could affect settlement preferences. Grazing could affect the inducing activity of CCA species by causing cell damage, allowing access to intracellular inducer molecules. CCA tissue can be broken by the grazing of abalone (Shepherd & Cannon 1988, Shepherd & Daume 1996) and various other invertebrate grazers (Steneck & Watling 1982). In order for grazing damage to explain settlement preferences for certain CCA, the grazing would need to differentially affect the CCA species. In the present study grazing damage was evident on Phymatolithon repandum (Fig. 5E) and Pneophyllum coronatum but not on the favored Hydrolithon rupestris. CCA cell damage from grazing need not result in higher larval settlement. Daume et al. (1999) found that Haliotis laevigata larvae settling on the CCA Sporolithon durum lined up along the edges of cuts in the thallus, but the number of larvae that chose damaged plants was not significantly higher than that choosing undamaged plants. Johnson et al. (1991) found no difference in the number of Acanthaster planci larvae settling on damaged and undamaged Lithothamnium pseudosorum.

Minimal Effect From Dissolved Inducers in Assays

In the present study, the corallines were broken into fragments for the experiments, so the specimens were not intact. Thus active chemicals may have leaked from the plants into the assay water. Water-soluble extracts of Phymatolithon repandum induce larval attachment, and sometimes metamorphosis, in Haliotis iris (Roberts & Nicholson 1997, R. Roberts unpubl. data). The concentration required to induce 50% larval attachment is generally between 100 and 200 [micro]g dry extract per mL of water (R. Roberts unpubl. data). It is unlikely that this concentration would be produced by leakage from damaged edges in the settlement assays, but some effect from soluble inducers cannot be ruled out. Water soluble inducers would cause larvae to cease swimming and begin crawling within several minutes (R. Roberts unpubl. data). Crawling larvae may find access easiest for those corallines supported on flat plastic sheet (Pneophyllum coronatum and Hydrolithon rupestris), and this differential accessibility could confound results. In Choice Experiment 2, the CCA fragments were soaked in FSW for 4 days before assayed to reduce any effect from leaking cell contents. Settlement preferences were not appreciably altered (Fig. 1A, B). Experiment 3 tested the accessibility issue by pressing the coralline samples into clay so that the live surface of each specimen was flush with the surrounding substratum and all plants were equally accessible to crawling (or sinking) larvae. Larvae still displayed the same preference for Pneophyllum coronatum and Hydrolithon rupestris over the remaining species.

Timing of Observations in Settlement Choice Experiments

To detect preferential settlement, counts had to be made early enough to precede the movement of postlarvae off CCA, but late enough to let larvae respond to the CCA. Counts made after 1 day were informative despite the variation in speed of metamorphosis between larval batches. In all cases there was some movement of metamorphosed animals off CCA between 1 and 3 days after settlement (Fig. 8). Haliotis iris remain on the upper surfaces of CCA-coated boulders for a few months after settlement (McShane & Naylor 1995) so their movement off CCA in the present study would not be in search of shelter or darkness. It is likely to be caused by food depletion, as the experiments used high densities of larvae (Table 1) on small fragments of CCA (e.g., Fig. 2C). Newly metamorphosed postlarvae are capable of eating small, loosely attached particles including diatoms and bacteria (Seki & Kan-no 1981, Norman-Boudreau et al. 1986, Roberts et al. 1999a, b), and 6-wk old postlarvae can ingest the cuticle and epithallial cells of CCA (Garland et al. 1985). Copious fecal pellets (devoid of diatoms and CCA cells) are visible around CCA-coated pebbles within 2 days of settlement induction (R. Roberts, pers. obs.). This extracellular food source would probably be quickly depleted at high postlarval densities.

How Should Target Size be Measured?

Competent abalone larvae periodically cease swimming, sink to the bottom, pull themselves onto their foot. They may crawl some distance while testing the surface for the presence of settlement cues (Seki & Kan-no 1981). The larvae require surface contact to respond to intact corallines, rather than showing taxis toward the source of a water-borne cue (Morse et al. 1980). The appropriate "target size" for larvae sinking through the water column is the "floor area" covered by the CCA fragment, as measured here. For crawling larvae, the distance around the perimeter of the fragment is most relevant, but the floor area measurement is a reasonable approximation for objects that are not long and thin.

CONCLUSION

Although the five species of CCA tested were all effective inducers of Haliotis iris larval attachment and metamorphosis, larvae did exhibit a preference for certain species when given a choice. This selection of certain substrata most likely results from different probabilities of acceptance or rejection following contact with a substratum. This study emphasizes the strong effect of selective settlement even among a suite of highly effective settlement substrata. From these laboratory results it is hypothesized that recruitment of Haliotis iris in the wild would vary among CCA species or growth forms, and that this variation is at least partly determined at the time of larval settlement.

TABLE 1.

Details of experiments conducted during this study. Start date is the
day, on which larvae were added to CCA. Larval stage was determined by
the number of fully chitinized transverse rows of teeth on the larval
radula (Tong and Moss 1992).

                                      Larval Stage
Experiment          Larval Batch;       (Rows of       Number of
Number              and Start Date    Radula Teeth)    Replicates

Choice Exp. 1         1; 18/9/98           8-9             6
Choice Exp. 2         1; 22/9/98          12-13            6
Choice Exp. 3         3; 21/4/99           7-8             6
No-choice Exp. 1      1; 18/9/98           8-9             5
No-choice Exp. 3      3; 21/4/99           7-8             6

                         No. Larvae
Experiment                Per Well
Number              Mean [+ or -] SE (n)    Purpose

Choice Exp. 1       192 [+ or -] 30 (6)     Determine relative
                                              settlement of larvae
                                              given a choice of five
                                              CCA species.
Choice Exp. 2       211 [+ or -] 27 (6)     Repeat Choice Exp. 1 but
                                              with CCA fragments soaked
                                              for 4 days to reduce
                                              effects from inducers
                                              leaking from damaged
                                              plants.
Choice Exp. 3       92 [+ or -] 8 (6)       Repeat of Choice Exp. 1 but
                                              with a separate batch of
                                              larvae and with CCA
                                              fragments embedded in
                                              clay to standardize their
                                              accessibility to crawling
                                              larvae.
No-choice Exp. 1    58 [+ or -] 5 (30)      Quantify settlement of
                                              larvae in response to the
                                              5 CCA species presented
                                              individually for
                                              comparison with Choice
                                              Exp. 1.
No-choice Exp. 3    43 [+ or -] 5 (30)      As previous for comparison
                                              with Choice Exp. 3.

TABLE 2.

Morphologic and ecological features of the crustose coralline algae
(CCA) used in this study.

     CCA Species                  Source                Morphology

Pneophyllum coronatum    On plastic plates in       Encrusting
  (Rosanoff) Penrose       GACL hatchery
  in Chamberlain
Hydrolithon rupestris    On plastic plates in       Encrusting
  (Foslie) Penrose         GACL hatchery
Phymatolithon repandum   On stones from shallow     Encrusting to warty
  (Foslie) Wilks and       subtidal at Cable
  Woelkerling              Bay (173[degrees]24'E,
                           41[degrees]11'S)
Lithophyllum             On stipes of sublittoral   Foliose
  carpophylli              fringe Carpophyllum
  (Heydrich)               maschalocarpum
  Heydrich                 (Phaeophyceae) on
                           shore by GACL
Mesophyllium             On stable boulders in      Warty to fruticose
  printzianum              sublittoral fringe on
  Woelkerling and          shore by GACL
  Harvey

                           Crust
     CCA Species         Thickness             Likely Grazers

Pneophyllum coronatum     -0.2 mm     Juvenile Haliotis iris and small
  (Rosanoff) Penrose                    crustacea
  in Chamberlain
Hydrolithon rupestris     -0.2 mm     Juvenile Haliotis iris and small
  (Foslie) Penrose                      crustacea
Phymatolithon repandum    -0.5 mm     Mixture of grazing species but
  (Foslie) Wilks and                    not abalone
  Woelkerling
Lithophyllum               2-4 mm     Small gastropods and crustaceans
  carpophylli                           but not abalone
  (Heydrich)
  Heydrich
Mesophyllium               2-5 mm     Mixture of grazers including
  printzianum                           Haliotis iris
  Woelkerling and
  Harvey

GACL = Glenhaven Aquaculture Centre Ltd (174[degress]21.2'E,
41[degrees]11.4'S). Morphology terminology follows Woelkerling et al.
(1993).

TABLE 3.

Composition of diatom assemblages on the CCA specimens used in Choice
Experiment 3. Data are from 12 random quadrats on each of
5 plants viewed by SEM (see Methods).

                              Percentage of Total Diatom Cells
                                       Seen on Plant:

                                Cocconeis          Amphora
       CCA Species                 spp.              spp.

Experiment 3 plants
  Pneophyllum coronatum       81 [+ or -] 9     13 [+ or -] 9
  Hydrolithon rupestris       70 [+ or -] 24    30 [+ or -] 24
  Phymatolithon repandum      50 [+ or -] 50    25 [+ or -] 25
  Lithophyllum carpophylli          --                --
  Mesophyllum printzianum           0                 0
Additional plants from wine location on other dates
  Pneophyllum coronatum       51 [+ or -] 11    15 [+ or -] 8
  Hydrolithon rupestris       25 [+ or -] 22    54 [+ or -] 19
  Phymatolithon repandum      38 [+ or -] 10     2 [+ or -] 2
  Lithophyllum carpophylli    26 [+ or -] 11     0 [+ or -] 0
  Mesophyllum printzianum     37 [+ or -] 16     1 [+ or -] 1

                              Percentage of Total Diatom Cells
                                       Seen on Plant:

                                Tabularia         Achnanthes
       CCA Species                 spp.              spp.

Experiment 3 plants
  Pneophyllum coronatum
  Hydrolithon rupestris
  Phymatolithon repandum
  Lithophyllum carpophylli
  Mesophyllum printzianum
Additional plants from wine location on other dates
  Pneophyllum coronatum        1 [+ or -] 1      0 [+ or -] 0
  Hydrolithon rupestris        0 [+ or -] 0      0 [+ or -] 0
  Phymatolithon repandum       0 [+ or -] 0      1 [+ or -] 1
  Lithophyllum carpophylli    17 [+ or -] 13     0 [+ or -] 0
  Mesophyllum printzianum     57 [+ or -] 14     3 [+ or -] 3

                              Percentage of
                              Total Diatom
                              Cells Seen on
                                  Plant:

                                  Other         Number of Plants
       CCA Species             Naviculoids        With Diatoms

Experiment 3 plants
  Pneophyllum coronatum        6 [+ or -] 4            5
  Hydrolithon rupestris        0 [+ or -] 0            4
  Phymatolithon repandum      25 [+ or -] 25           2
  Lithophyllum carpophylli          --                 0
  Mesophyllum printzianum          100                 1
Additional plants from wine location on other dates
  Pneophyllum coronatum       33 [+ or -] 12           5
  Hydrolithon rupestris       21 [+ or -] 11           4
  Phymatolithon repandum      59 [+ or -] 10           4
  Lithophyllum carpophylli    58 [+ or -] 23           3
  Mesophyllum printzianum      3 [+ or -] 2            4

                                              Density (Cells
                              Total Number      [cm.sup.2]
       CCA Species             of Diatoms     Mean [+ or -]

Experiment 3 plants
  Pneophyllum coronatum           173         3.1 [+ or -] 1.4 x
                                                [10.sup.4]
  Hydrolithon rupestris            40         6.8 [+ or -] 3.8 x
                                                [10.sup.3]
  Phymatolithon repandum           14         2.4 [+ or -] 2.0 x
                                                [10.sup.3]
  Lithophyllum carpophylli          0         0.0 [+ or -] 0.0
  Mesophyllum printzianum           8         1.4 [+ or -] 1.4 x
                                                [10.sup.3]
Additional plants from wine location on other dates
  Pneophyllum coronatum            61         7.0 [+ or -] 3.4 x
                                                [10.sup.3]
  Hydrolithon rupestris             9         1.2 [+ or -] 0.5 x
                                                [10.sup.3]
  Phymatolithon repandum           97         1.6 [+ or -] 0.4 x
                                                [10.sup.4]
  Lithophyllum carpophylli        201         3.4 [+ or -] 3.3 x
                                                [10.sup.4]

  Mesophyllum printzianum         102         1.7 [+ or -] 0.3 x
                                                [10.sup.4]

Data are mean percentage [+ or -] SE. n = numher of plants with
diatoms.

ACKNOWLEDGMENTS

The authors thank Sabine Daume and Bill Woelkerling for tuition and expertise on corallines identification, and for helpful discussions; Christine Handley for technical assistance; Mark Gould for guidance on scanning electron microscopy and Mike Barker and Philip Mladenov for review of drafts. This research was supported by contracts CAW 801 and CAWX0004 with the New Zealand Foundation for Research Science and Technology and by a University of Otago Alliance Group Scholarship.

LITERATURE CITED

Burnham, K.P., D.R. Anderson, G.C. White, C. Brownie & K.H. Pollock. 1987. Design and analysis methods for fish survival experiments based on release-recapture. American Fisheries Society Monograph 5.

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.

Crisp, D. J. 1974. Factors influencing the settlement of marine invertebrate larvae. In: P.T. Grant & A. M. Mackie, editors. Chemoreception in Marine Organisms. New York: Academic Press.

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.

Day, L. & G.M. Branch. 2000. Relationships between recruits of the abalone Haliotis midae, encrusting corallines and the sea urchin Parechinus angulosus. S. Afr. J. Mar. Sci. 22:137-144.

Fletcher, W. J. 1987. Interactions among subtidal Australian sea urchins, gastropods, and algae: effects of experimental removals. Ecol. Monogr. 57:89-109.

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

Garland, C. D., S. L. Cooke, J. F. Grant & T. A. McMeekin. 1985. Ingestion of the bacteria on and the cuticle of crustose (non-articulated) coralline algae by post-larval and juvenile abalone (Haliotis tuber Leach) from Tasmanian waters. J. Exp. Mar. Biol. Ecol. 91:137-149.

Hayat, M. A. 1989. Principles and techniques of electron microscopy: biological applications 3rd ed. London: MacMillan Press.

Johansen, H.W. 1981. Coralline algae: a first synthesis. Florida: CRC Press. 239 pp.

Johnson, C. R. & D. C. Sutton. 1994. Bacteria on the surface of crustose coralline algae induce metamorphosis of the crown-of-thorns starfish Acanthaster planci. Mar. Biol. 120:305 310.

Johnson, C.R., D.G. Muir & A. L. Reysenbach. 1991. Characteristic bacteria associated with surfaces of coralline algae: a hypothesis for bacteria induction of marine invertebrate larvae. Mar. Ecol. Prog. Ser. 74:281-294.

Kaspar, H. F. 1992. Oxygen conditions on surfaces of coralline red algae. Mar. Ecol. Prog. Ser. 81:97-100.

Kawamura T. & S. Kikuchi. 1992. Effects of benthic diatoms on settlement and metamorphosis of abalone larvae. Suisanzoshoku 40:403-409.

Keough, M. J. & B. J. Downes. 1982. Recruitment of marine invertebrates: the role of active larval choices and early mortality. Oecologia. 54: 348-352.

Kirchman, D., S. Graham, D. Reish & R. Mitchell. 1982. Bacteria induce settlement and metamorphosis of Janua (Dexiospira) brasiliensis Grube (Polychaeta: Spirorbidae). J. Exp. Mar. Biol. Ecol. 56:153-163.

Link, W. A. 1999. Modeling pattern in collections of parameters. J. Wildlife Manage. 63:1017-1027.

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

McShane, P.E. 1992. Early life history of abalone: 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. 120-138.

McShane, P. E. 1996. Recruitment processes in abalone (Haliotis spp.). In: Y. Watanabe, Y. Yamashita & Y. Oozeki, editors. Survival strategies in early life stages of marine resources. Rotterdam: A. A. Balkema. pp. 315-324.

McShane, P. E. & J. R. Naylor. i995. Depth can affect post-settlement survival of Haliotis iris (Mollusca: Gastropoda). J. Exp. Mar. Biol. Ecol. 187:1-12.

McShane, P. E. & M. G. Smith. 1988. Measuring abundance of juvenile abalone Huliotis ruber Leach (Gastropoda Huliotidae): comparison of a novel method with two other methods. Aust. J. Mar. Freshwat. Res. 39:331-336.

McShane, P., S. Mercer & R. Naylor. 1995. Good settlement of paua in Paua 7. Seafood New Zealand 3:20-21.

Miron, G., E. Bourget & P. Archambault. 1996. Scale of observation and distribution of adult conspecifics: their influence in assessing passive and active settlement mechanisms in the barnacle Balanus crenatus (Brugiere). J. Exp. Mar. Biol. Ecol. 201:137-158.

Morse, A. N. C. & D. E. Morse. 1984. Recruitment and metamorphosis of Haliotis larvae induced by molecules uniquely available at the surfaces of crustose red algae. J. Exp. Mur. Biol. Ecol. 75:191-215.

Morse, D. E., H. Duncan, N. Hooker & A. Morse. 1977. Hydrogen peroxide induces spawning in mollusks, with activation of prostaglandin endoperoxide synthetase. Science 196:298-300.

Morse, D. E., N. Hooker, L. Jensen & H. Duncan. 1979. Induction of larval abalone settling and metamorphosis by [gamma]-aminobutyric acid and its congeners from crustose red algae: II: applications to cultivation seed-production and bioassays, principle causes of mortality and interference. Proc. World Mariculture Soc. 10:81-91.

Morse, D.E., M. Tegner, H. Duncan, N. Hooker, G. Trevelyan & A. Cameron. 1980. Induction of settling and metamorphosis of planktonic molluscan (Haliotis) larvae: Part III: signaling by metabolites of intact algae is dependent on contact. In: D. Muller-Schwarz & R. M. Silverstein, editors. Chemical signaling in vertebrate and aquatic animals. New York: Plenum Press. pp. 67-86.

Morse, D. E., N. Hooker, A. N. C. Morse & R. A. Jensen. 1988. Control of larval metamorphosis and recruitment in sympatric agariciid corals. J. Erp. Mar. Biol. Ecol. 116:193-217.

Mundy, C. & R. Babcock. 2000. Are vertical distribution patterns of scleractinian corals maintained by pre- or post-settlement processes? A case study of three contrasting species. Mar. Ecol. Prog. Ser. 198:109-119.

Naylor, J. R. & P. E. McShane. 1997a. Post-settlement survival of abalone (Haliotis iris, H. australis) in turbulent flows. Moll. Res. 18:227-232.

Naylor, J. R. & P. E. McShane. 1997b. Predation by polychaete worms on larval and post-settlement abalone Haliotis iris (Mollusca: Gastropoda). J. Exp. Mur. Biol. Ecol. 214:283-290.

Norman-Boudreau, K., D. Bums, C. A. Cooke & A. Austin. 1986. A simple technique for detection of feeding in newly metamorphosed abalone. Aquaculture 51:313-317.

Pawlik, J. R. 1992. Chemical ecology of the settlement of benthic marine invertebrates. Oceanogr. Mar. Biol. Annu. Rev. 30:273-335.

Roberts, R. D. 2001a. A review of settlement cues for larval abalone (Hallotis spp.). J. Shellfish Res. 20:571-587.

Roberts, R. D. 2001b. Factors affecting larval settlement and post-larval performance of New Zealand abalone, Haliotis spp. PhD Thesis. University of Otago.

Roberts, R. D., T. Kawamura & C. M. Nicholson, 1999a. Growth and survival of post-larval abalone (Haliotis iris) in relation to their 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.

Roberts, R. D. & C. M. Nicholson. 1997. Variable response from abalone larvae (Haliotis iris, H. virginea) to a range of settlement cues. Moll. Res. 18:131-141.

Rodriguez, S.R., F.P. Ojeda & N.C. Inestrosa. 1993. Settlement of benthic marine invertebrates. Mar. Ecol. Prog. Ser. 97:193-207.

Rowley. R. J. 1989. Settlement and recruitment of sea urchins (Stroglyocentrotus spp.) in a sea-urchin barren ground and a kelp bed: are populations regulated by settlement or post-settlement processes? Mar. Biol. 100:485-494.

Sebens. K. P. 1983. Settlement and metamorphosis of a temperate soft-coral larva (Alcyonium siderium Verrill): induction by crustose algae. Biol. Bull 165:286-304.

Seki, T. & H. Kan-no. 1981. Observations on the settlement and metamorphosis of the veliger of the Japanese abalone, Haliotis discus hannai Ino, Haliotidae, Gastropoda. Bull. Tohoku Reg. Fish. Res. Lab. 42:31-39.

Shepherd, S. A. & J. Cannon. 1988. Studies on southern Australian abalone (genus Haliotis) X. Food and feeding of juveniles. J. Malacol. Soc. of Aust. 9:21-26.

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. Rotterdam: A. A. Balkema. pp. 297-313.

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

Snelgrove, P. V. R., J. P. Grassle & C. A. Butman. 1998. Sediment choice by settling larvae of the bivalve, Spisula solidissima (Dillwyn), in flow and still water. J. Exp. Mar. Biol. Ecol. 231:171-190.

SPSS Inc. 1999. SYSTAT 9 Statistics I. SPSS Inc., Chicago.

Steneck, R.S. 1986. The ecology of coralline algal crusts: convergent patterns and adaptive strategies. Ann. Rev. Ecol. Syst. 17:273-303.

Steneck, R. S. & R. T. Paine. 1986. Ecological and taxonomic studies of shallow-water encrusting Corallinaceae (Rhodophyta) of the boreal northeastern Pacific. Phycologia 25:221-240.

Steneck, R. S. & L. Watling. 1982. Feeding capabilities and limitation of herbivorous molluses: a functional group approach. Mar. Biol. 68:299-319.

Stoner, A. W., M. Ray, R. A. Glazer & K. J. McCarthy. 1996. Metamorphic responses to natural substrata in a gastropod larva: decisions related to postlarval growth and habitat preference. J. Exp. Mar. Biol. Ecol. 205:229-243.

Suenaga, K., H. Hori, H. Ishida, H. Nukaya, R. D. Roberts & K. Tsuji. 2004. Inducing substance for abalone larval metamorphosis from the crustose coralline alga, Hvdrolithon samoense. Fisheries Sci. 70:342-344.

Tong, L. J. & G.A. Moss. 1992. The New Zealand culture system for 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. 583-591.

Walters, L. J. & D. S. Wethey. 1996. Settlement and early post-settlement survival of sessile marine invertebrates on topographically complex surfaces: the importance of refuge dimensions and adult morphology. Mar. Ecol. Prog. Ser. 137:161-171.

Wieczorek, S. K. & C. D. Todd. 1997. Inhibition and facilitation of bryozoan and ascidian settlement by natural multi-species biofilms: effects of fihn age and the roles of active and passive larval attachment. Mar. Biol. 128:463-473.

Woelkerling, W. J., L. M. Irvine & A. S. Harvey. 1993. Growth-forms in non-geniculate coralline red algae (Corallinales, Rhodophyta). Aust. Syst. Bot. 6:277-293.

Woodin, S. A. 1986. Settlement of infauna: larval choice? Bull. Mar. Sci. 39:401-407.

RODNEY D. ROBERTS, (1, 2), * HEINRICH F. KASPAR (1) AND RICHARD J. BARKER (3)

(1) Cawthron Institute, Private Bag 2, Nelson, New Zealand; (2) Department of Marine Science, University of Otago, P.O. Box 56, Dunedin, New Zealand; (3) Department of Mathematics and Statistics, University of Otago, P.O. Box 56, Dunedin, New Zealand

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