Factors affecting settlement of abalone (Haliotis iris) larvae on benthic diatom films.

ABSTRACT Diatoms are widely used in abalone hatcheries to induce larval settlement (-attachment and metamorphosis) but there are few data on the factors that influence their effectiveness as settlement cues. Of 16 diatom strains tested in Experiment 1, half induced [greater than or equal to] 80% attachment and >50% metamorphosis within 4 days. Settlement success did not correlate significantly with diatom abundance or adhesive strength (P > 0.05). Several diatom strains interfered with settlement. Examples included (1) smothering by highly mobile diatoms (Nitzschia longissima); (2) shells becoming stuck to sticky secretions (Navicula britannica, Achnanthes longipes); and (3) unstable diatom cells preventing pedal attachment (Licmophora sp.). In Experiment 2, only 1 of 15 diatom strains induced over 70% metamorphosis, and three others induced 22-36% metamorphosis, within 4 days. Older cultures of a strain induced higher attachment and metamorphosis than younger cultures of the same strain (P < 0.001), but some old cultures still gave poor settlement. For the young diatom cultures, larval attachment correlated positively with diatom percent cover (r = 0.89, P < 0.05), and metamorphosis with the growth phase of the diatom film (r = 0.91, P < 0.05). Experiments 3-5 examined the role of bacteria in settlement induction by diatom films. In Experiment 3, films of 8 diatom species from Experiment 1 were regrown from ~6 individually isolated and rinsed cells, likely altering the associated bacterial flora. Metamorphosis was 179 fold lower on average than in Experiment 1 (P < 0.0001), but attachment and diatom density were not significantly different (P = 0.87 and P = 0.75 respectively). In Experiment 4, Nitzschia ovalis grown and assayed with antibiotics had lower metamorphosis after 2 days than the same strain grown and assayed without antibiotics (6 vs 64%, P = 0.004), whereas attachment after 2 days did not differ (94 versus 92%, P = 0.49). In Experiment 5 bacteria from a Nitzschia ovalis culture induced as much attachment and metamorphosis as the intact diatom film, whereas cell-free supernatant was much less effective. The presence of antibiotics in settlement assays reduced attachment and metamorphosis by biofilms, but not by coralline algae (Phymatolithon repandum) or GABA. In Experiment 5 metamorphosis on diatoms and bacteria occurred gradually over 2 wk, whereas on coralline algae and GABA it occurred within a few days. This study suggests that many diatoms cue rapid larval attachment, but few induce consistently strong metamorphosis within 4 days in laboratory conditions. The bacteria present in diatom films affect the settlement-inducing activity of the film in at least some cases. The effectiveness of diatom films was generally higher for more mature films but overall, physical factors explained little of the variation in the activity of diatoms.

KEY WORDS: benthic diatom, bacteria, abalone, larval settlement, attachment, metamorphosis, biofilm, Haliotis

INTRODUCTION

Diatom films have been used for decades to induce settlement (=attachment and metamorphosis) of abalone larvae in commercial hatcheries (e.g., Seki 1980). Despite the development of alternative settlement inducers such as [gamma]-aminobutyric acid (GABA, Morse et al. 1979a) abalone mucus trails (Seki & Kan-no 1981) and Ulvella lens (Takahashi & Koganezawa 1988, Daume et al. 2004), the use of diatom films remains widespread.

Early abalone culturists exerted no control over the composition of the algal film, so filamentous diatoms sometimes proliferated and were found unsuitable for settlement (Seki 1980). Pregrazing of plates by juvenile abalone was used to select for tightly attached, prostrate diatoms and other algae (Seki 1980) and to coat the settlement plates with mucus trails (Seki & Kan-no 1981). Pregrazed plates are still used in Japanese abalone hatcheries, but have not been widely adopted in other countries despite experimental evidence of their effectiveness with many abalone species (e.g., Searcy-Bernal et al. 1992, Bryan & Qian 1998, Daume et al. 2000, Stott et al. 2004). Abalone culturists in many countries rely on an ungrazed film dominated by prostrate diatoms to trigger abalone settlement (Hahn 1989, Shallow Seafarming Research Institute 1990, Tong et al. 1992, Gallardo & Buen 2003, Sawatpeera et al. 2004). Selection for small, prostrate diatoms can be achieved by filtration of incoming seawater or mixed diatom slurries, or by culture of isolated strains (e,g., Hahn 1989, Tong & Moss 1992, Krsinich et al. 2000).

Despite the widespread use of diatoms in abalone hatcheries, metamorphosis rates on diatom films are often low (Roberts 2001a, Gallardo & Buen 2003, Sawatpeera et al. 2004, Daume 2006) and research has struggled to identify strains that give consistently good metamorphosis. Most studies have tested only a few diatom strains at a time and there has been little critical analysis of the factors that determine the effectiveness of diatom films. Settlement is generally higher on denser diatom films (Kawamura & Kikuchi 1992, Daume et al. 1999) and filamentous diatoms can cause poor settlement (Seki 1980, Yanagihashi et al. 1986) by sticking to larvae (Kawamura & Kikuchi 1992). Kawamura and Kikuchi (1992) considered that the diatoms most suitable for settlement were noncolonial, prostrate strains, and particularly those that were tightly attached to the substratum. Gordon et al. (2004) assessed the effectiveness of 17 strains of prostrate diatom for larval attachment, and began to investigate the physico-chemical properties of attachment inducers. However, larval attachment is often decoupled from metamorphosis in abalone larvae (e.g., Roberts & Nicholson 1997). Cues for metamorphosis are much more limited than cues for larval attachment, and metamorphosis is the critical step in the transition to postlarval life (Roberts 2001a).

Whereas previous studies have documented wide variation in abalone settlement rates on a range of diatom species they have not determined whether those differences are stable through time. None of the work done to date has been with axenic diatoms, so the bacterial component of the "diatom film" could play a significant role. Changes in the composition of the bacterial population over time would go unnoticed during diatom culture. There are few data on the effect of bacteria on abalone settlement. Haliotis rubra larvae showed zero metamorphosis within 4 days on natural biofilms (Hugget et al. 2005) and Haliotis rufescens did not settle on 33 marine bacterial isolates (Morse 1992). Roberts (2001b) found gradual induction of metamorphosis in Haliotis iris and Haliotis virginea by bacterial biofilms. Bacteria could play a more potent role in combination with another source of cues (e.g., Johnson & Sutton 1994).

The potential gains that ideal diatom strains could offer commercial hatcheries in terms of reliability and efficiency justify further critical investigation. The present paper compares Haliotis iris larval settlement on a wide range of benthic diatom strains, primarily prostrate strains that have proven most suitable for abalone settlement. The experiments address four major questions:

(1) Does the rate of larval settlement vary among diatom strains within an experiment?

(2) Are those patterns repeatable between experiments?

(3) What features of a diatom film correlate positively or negatively with settlement success?

(4) Does the bacterial component of a diatom film influence settlement success?

METHODS

Isolation and Culturing of Diatoms

Diatoms were isolated by drawing individual cells into a fine glass capillary and transferring them to Jorgensen's (1962) medium, supplemented with 0.05 [micro]g/L vitamin [B.sub.12]. Each isolated cell was placed in an individual well of a tissue culture plate (Falcon 3047) and grown at 17 [+ or -] 0.5[degrees]C under fluorescent lights at 50-70 [micro]E [m.sup.-2] [s.sup.-1] on a 12:12 light: dark cycle. When the target strain had established, a sample was transferred to a 75 ml polystyrene pottle (Labserv) to establish a parent culture that was subcultured each 10 or 20 days. For settlement bioassays, an inoculum of parent culture was seeded into tissue culture plates (Falcon 3043) containing modified Jorgensen's medium. Plates were incubated in conditions as described above until taken for bioassays. Immediately prior to addition of larvae, the growth medium was gently removed with a transfer pipette (Samco) and replaced with 0.2 [micro]m-filtered natural seawater (FSW). Diatom cultures were uni-algal but not axenic. Details and taxonomic authorities of the strains used are presented in Table 1.

Measurement of Diatom Film Characteristics

The density of diatoms and bacteria on the floor of the assay wells was quantified using an inverted microscope with video attachment. Counts for each treatment were averaged from 3-5 fields of view in each of 6 replicates. Diatom densities were converted to percent cover by multiplying the mean density by the mean cell size, and assuming cells were elliptical. The thin needle-like extensions on Cylindrotheca closterium and Nitzschia longissima were excluded from the area of the ellipse. Fluorescence staining techniques were not used for bacterial counts because the samples had to be used for bioassay after the counts. The direct counting method used here will include nonviable bacteria. Counts by the method used in this paper were 1.3-1.8 fold higher than acridine orange counts of live cells (Turley 1993) for two actively growing cultures, and ~5.8 fold higher for two stationary phase cultures (R. Roberts, unpubl. data).

The strength of adhesion of diatoms to the substratum was ranked on a semiquantitative scale from 1 = weakest to 5 = strongest. The ranking was based on the extent to which diatom cells were dislodged by larval swimming and water jets from a pipette. For Experiment 2, motility was rated on a semiquantitative scale from 1 = least motile, to 5 = most motile based on the proportion of cells moving. Growth phase of the diatom film at the start of the bioassays was estimated for Experiment 2 by plotting the density of each strain 0, 2 and 4 days prior to bioassay. The age of the biofilm was defined as the number of days from the time the diatom cultures were established (by subculturing from a parent culture) until the start of the bioassay.

Spawning, Larval Culture, and Bioassays

Abalone (Haliotis iris Gmelin) were spawned by the hydrogen peroxide technique (Morse et al. 1977) and larvae reared in flowing seawater as described previously (Roberts & Nicholson 1997) until taken for bioassays. Larval development was quantified by counting the number of transverse rows of fully chitinised teeth on the radula (Tong et al. 1992). For settlement bioassays, ~50-100 competent larvae were transferred to each well of a tissue culture plate (Falcon 3043) containing 3.4 ml of FSW. Assays conducted in the presence of antibiotics (to ensure good survival) contained 150 [micro]g/ml each of penicillin G sodium (Biochemie or Sigma) and streptomycin sulphate (Sigma) as a single dose at the start of the assay. Plates were incubated in the conditions described above. Experiments included negative controls (FSW), and positive controls of 1-2 [micro]M GABA and crustose coralline algae (CCA), all with antibiotics. The CCA were small pebbles (~1 cm diam.) fully encrusted with Phymatolithon repandum (Foslie) Wilks and Woelkerling, which gives consistently high metamorphosis of Haliotis iris (Roberts et al. 2004). The metamorphosis response to GABA is more variable between batches of Haliotis iris larvae (Roberts & Nicholson 1997) so gives some measure of larval receptiveness to metamorphosis cues. The proportion of larvae that had attached and metamorphosed was quantified as described previously (Roberts & Nicholson 1997) after various incubation periods. Settlement counts of CCA treatments were "destructive," but other treatments could be counted repeatedly.

Experiment 1: Settlement on 16 Diatom Strains

This experiment compared settlement response to 16 diatom strains, belonging to 12 diatom species (Table 1). Experimental diatom films were established by subculturing parent cultures as described above. Larvae had 8 rows of radula teeth at the start of bioassays, and all bioassays contained antibiotics.

Experiment 2: Settlement on 15 Diatom Strains

This experiment examined the correlations between settlement and a suite of biofilm characteristics described above. The experiment used 15 diatom treatments (Table 1), including both "old" and "young" cultures of 5 strains. The experiment focused on strains predicted to give either high or low settlement, based on existing data. Larvae had 7 rows of radula teeth at the start of bioassays and all contained antibiotics.

Experiment 3: Settlement on 8 Reisolated Diatom Strains

Experiment 3 retested the settlement response to 8 of the diatom strains used in Experiment l, but using "re-isolated" cultures (Table 1). Reisolated cultures were established by individually picking several cells from a stock culture, ejecting them into and repicking them from 2 consecutive drops of FSW, and then transferring them to sterile tissue culture plates (Falcon 3043) with filter-sterilized growth medium. Experiment 3 included a mixed species diatom film labeled as "hatchery film". This was established by holding transparent plastic discs (22 mm diameter) for 1 wk in an aerated tank receiving 1 L per minute of sand-filtered seawater with an irradiance of 15 [micro]E [m.sup.-2] [s.sup.-1] on a 12:12 light: dark cycle. Six randomly chosen discs were bioassayed by placing them on the floor of tissue culture plate wells. Larvae had 10 rows of radula teeth at the start of bioassays, and all treatments contained antibiotics. Follow up experiments to examine the lack of metamorphosis were started 5 days later with larvae from the same batch (~14 rows of radula teeth) and again contained antibiotics.

Experiment 4: Effect of Antibiotics on Settlement on Nitzschia ovalis

Twelve replicate cultures of Nitzschia ovalis were established simultaneously from a single parent culture, and 6 were grown with antibiotics (as above) and 6 without. For bioassays, the diatom growth medium was replaced with FSW, and antibiotics were added only to those diatom cultures reared with antibiotics. The density of bacteria and diatoms was determined immediately prior to commencement of the bioassay. Larvae had 11 rows of radula teeth at the start of the bioassays.

Experiment 5: Role of Bacteria in Settlement Induction by Nitzschia ovalis Films

Experiment 4 suggested that bacteria were important in settlement on Nitzschia ovalis but it did not separate the effect of antibiotics on biofilm development from their direct effect on the settlement response. Experiment 5 addressed this issue by comparing settlement on Nitzschia ovalis grown to late exponential phase over 15 days with and without antibiotics (as above) before each was assayed with and without antibiotics (Treatments 1-4). Experiment 5 also examined whether settlement was induced by bacteria from the biofilm alone (Treatments 5-6) and by diatom culture supernatant (Treatment 7). These treatments, plus positive and negative controls are summarized in Table 2.

Films of mixed bacteria were established by harvesting and filtering 12-day-old (late exponential phase) cultures of Nitzschia ovalis. Six spare cultures of Nitzschia ovalis were harvested by scraping with a plastic transfer pipette, then mixed and drawn into a sterile syringe. Approximately 1 mL of the suspension was ejected into each of 12 randomized wells through a sterile 0.45- [micro]m filter to exclude diatoms but allow the passage of bacteria* Each well received an additional 2 mL of growth medium and plates were incubated for 3 days before settlement bioassay. Six wells were assayed with antibiotics (Treatment 5) and six without (Treatment 6).

The biofilm supernatant treatment (Treatment 7) represented molecules released from the diatom film to the culture medium. Supernatant was drawn from above six 15-day-old (late exponential phase) cultures of Nitzschia ovalis grown without antibiotics, without damaging the film. The supernatant was mixed then drawn into a sterile syringe. Approximately 2 mL of the supernatant was ejected into each of 6 randomized wells through a sterile 0.22 [micro]m filter to exclude diatom and bacterial cells. Each well received an additional 1 mL of growth medium and was assayed with antibiotics to minimize bacterial degradation or production of settlement inducing chemicals.

Controls (Treatments 8-11) included FSW assayed with and without antibiotics, plus two settlement inducers known to be effective in the presence of antibiotics (GABA, and the CCA Phymatolithon repandum). Larvae (7 rows of radula teeth) reared in flowing seawater were soaked in antibiotics for 2 h then triple rinsed with FSW before being transferred to bioassay plates, with or without antibiotics. This pretreatment provided excellent survival in bioassays without antibiotics. Growth medium in Treatments 1-6 was replaced with FSW immediately prior to addition of larvae. Bioassay plates were incubated at 16 [+ or -] 1 [degrees]C with very low fluorescent light (2-4 [micro]E [m.sup.-2] [s.sup.-1]) to minimize fluctuations in water chemistry caused by biofilm metabolism (Roberts et al. in press). Settlement counts were made after 7, 11, and 15 days of incubation except as explained in Figure 5 (see later).

Data Analysis

Correlations between settlement and biofilm characteristics were examined using Pearson's correlation coefficients. The Dunn-Sidak method was used to hold the experiment wise error rate at [alpha] = 0.05 (Sokal & Rohlf 1981). Percent cover of diatoms was calculated by multiplying the mean diatom cell size by the mean diatom density. There was error associated with both of these means, so the standard deviation of diatom cover was calculated by the propagation of errors formula:

[[sigma].sub.y] = x [square root of ([[sigma].sub.p]/p).sup.2] + ([[sigma].sub.q]/q).sup.2]

where x = mean percent cover of diatoms, p = mean diatom cell area m [cm.sup.2], q = mean diatom density in cells [cm.sup.-2], and [[sigma].sub.y], [[sigma].sub.p] and [[sigma].sub.q] are the standard deviations of percent cover, diatom cell area and diatom density respectively (Skoog 1985).

Larval attachment and metamorphosis on different diatom strains in Experiments 1 and 2 were not compared statistically as this would have involved hundreds of pairwise comparisons and correspondingly low corrected critical values. Mean values with 95% confidence intervals are presented for each treatment and for positive and negative controls. Data from Experiment 3 were compared with data from the same diatom strains in Experiment 1 using paired sample t-tests. The tests compared the relative difference between the Experiments by normalizing Experiment 1 data to a value of 1, then expressing the corresponding Experiment 3 value as a proportion of that value. This eliminated the variation introduced by differences in settlement rate among diatom species so that the two experiments could be more efficiently compared. The same approach was taken to comparison of "young" versus "old" diatom cultures in Experiment 2. Settlement data from Experiment 4 were compared with one-way ANOVA or nonparametric Mann-Whitney U-test. Homogeneity of variances was assessed by Levene test (Kirby 1993), and normality from probability plots. Experiment 5 data were analyzed by one-way ANOVA's with TukeyKramer HSD pairwise comparisons. Separate analyses were conducted for attachment and metamorphosis after 7, 11 and 15 days (6 tests). Homogeneity of variances and normality were assessed as above. Data on larval attachment in the coralline algae treatment after 7 days was excluded from analyses as all values were 100%, so the variance was zero and not rectifiable by transformation.

RESULTS

Experiment 1: Comparison of Settlement on 16 Strains of Benthic Diatom

Of the 16 diatom strains tested, half induced [greater than or equal to] 80% attachment and >50% metamorphosis within 4 days (Fig. 1). The most rapid response was elicited by Cocconeis scutellum-Jp, which induced 76% of larvae to metamorphose within 2 days. The poorest response was elicited by Nitzschia longissima, which was the most loosely attached and highly mobile strain. Larval attachment and metamorphosis across all strains did not correlate strongly with either diatom adhesive strength or diatom density (Table 3, Fig. 1). The 2 [micro]M GABA treatment induced 95% of larvae to metamorphose within 4 days, indicating that the larvae were highly receptive to settlement cues.

[FIGURE 1 OMITTED]

There was evidence of physical interference with settlement by some diatom strains. Nitzschia longissima-A was so motile that attached larvae became coated with diatoms and subsequently died without metamorphosing. Most larvae were incapacitated by Day 4 of the bioassay, so settlement counts were waived. This smothering effect was minor for the less dense culture of the same strain (Nitzschia longissima-B). Achnanthes longipes-A had long polysaccharide stalks that led to larvae becoming stuck and unable to attach by their foot. Four-day counts were consequently waived for this strain. A small proportion of larvae in Navicula britannica became stuck by the side of their shells to the floor of the assay chamber, presumably because of sticky secretions in the biofilm. Physical interference was also seen for a Licmophora sp. Cells of tiffs strain normally grow on a branched stalk, but in culture the cells proliferated loose on the floor of their container. These large and unstable cells stuck to the foot of the larvae, preventing normal attachment and leading to death within several days. No settlement counts were made for this diatom.

Some larvae in the Cocconeis scutellum-R treatment showed abnormal velum loss by day 4 of the bioassay. The abnormal larvae were unattached on the floor of the assay chambers with just a few velar cilia still intact. These larvae were incapable of swimming and did not develop postlarval shell or feeding activity. The settlement status of these larvae was unclear, so the 4-day counts were excluded for this diatom.

Experiment 2: Settlement on 15 Diatom Strains

All of the 15 diatom cultures tested had attained densities capable of inducing larval settlement by the start of the assays (Table 4). "Young" and "old" cultures of the same strain differed in their characteristics. Old cultures had a more mature growth phase, higher diatom and bacterial densities, and often higher diatom adhesive strength and lower diatom motility (Table 4).

Few of the 15 diatom cultures tested induced a high percentage metamorphosis. Metamorphosis after 4 days averaged 74% for tin old culture of Cocconeis scutellum 81. and 22% to 36% for 3 other strains. All remaining strains induced less than 6% metamorphosis within 4 days (Fig. 2). Five diatom strains induced >50% attachment.

[FIGURE 2 OMITTED]

When all 15 diatom treatments were considered, larval attachment and metamorphosis did not correlate significantly with any of the measured features of the diatom film (Table 3). When only the young cultures were considered, larval attachment correlated positively with growth phase of the diatom film, and metamorphosis correlated positively with percent cover (Table 3). The correlation between attachment and diatom cover was also strong, but marginally nonsignificant when critical values were adjusted for multiple tests (Table 3).

The greater activity of older films was most evident when young and old films of the same strain were compared, with the old cultures inducing significantly higher attachment and metamorphosis (t = 95, P << 0.0001, d.f. = 4). However, an old film did not guarantee strong settlement, with three of the six old cultures producing <6% metamorphosis after 4 days (Fig. 2). Diatom density and diatom cover were ~5 fold higher in old cultures than in the corresponding young cultures at the start of assays, but only 2 fold higher after 4 days of bioassay incubation.

Experiment 3: Settlement on 8 Reisolated Diatom Strains

When 8 of the 16 strains used in Experiment 1 were reisolated, and reassayed 2 wk later with a subsequent batch of larvae, markedly different results were obtained. In contrast to the first experiment, none of the 8 strains induced more than a few percent metamorphosis within 4 days (Fig. 3). Metamorphosis was 179 fold lower on average than in Experiment 1 (t = 352, P < 0.0001, d.f. = 7), but attachment and diatom density were not significantly different (P = 0.87 and P = 0.75 respectively).

[FIGURE 3 OMITTED]

Larvae in Experiment 3 were older than those from Experiment 1 (10 versus 8 rows of teeth in radula) and several lines of evidence confirmed their competence to metamorphose. (1) About 99% of larvae metamorphosed in response to the CCA positive control in both Experiments 1 and 3. (2) Larvae attached and metamorphosed in response to GABA in both experiments, although the mean percentage metamorphosis after 4 days in Experiment 3 was only about half that in Experiment 1 (Fig. 3). (3) A multispecies biofilm grown in flowing seawater induced 3-fold more metamorphosis than the most active of the pure strains (Fig. 3). This multispecies film was dominated by Bacillaria paxillifer, a tube-dwelling Navicula sp. and ?Nitzschia longissima. (4) When larvae were transferred from Experiment 3 assays to the highly inductive culture of Cocconeis scutellum-Jp from Experiment 1, they responded well (44 [+ or -] 9% metamorphosis after 2 days), although not as strongly as in Experiment 1 (76 [+ or -] 2%).

Experiment 4: Nitzschia ovalis Grown and Assayed With/Without Antibiotics

At the start of the bioassay, the Nitzschia ovalis cultures grown with antibiotics had diatom densities 1.7 fold higher, but bacterial densities ~200 fold lower than the no-antibiotic cultures (Fig. 4). After 2-4 days of bioassay, larval metamorphosis was ~5-10 fold lower (P < 0.01) on diatoms that were grown and assayed in the presence of antibiotics, whereas percent attachment was comparable between the treatments (Fig. 4).

[FIGURE 4 OMITTED]

Experiment 5: Role of Bacteria in Settlement Induction by Nitzschia ovalis

The majority of larvae metamorphosed within 7 days in the positive control treatments of GABA and CCA (Fig. 5). The only other treatments that reached such high levels of metamorphosis were the bacterial film and diatom film grown and assayed without antibiotics (Fig. 5 treatments labeled "Diat. -/-" and "Bact. -/-"). In these two treatments, the bacterial film was as effective in inducing attachment and metamorphosis as the diatom film from which it was derived. Metamorphosis in these two treatments was slow to develop, reaching ~25% after 7 days and ~90% after 15 days. The activity of the bacterial film was reduced to near the level of the negative control ("FSW/+" in Fig. 5) when assayed in the presence of antibiotics ("Bact. -/+" in Fig. 5). Settlement induction by diatom films also tended to be lower for diatoms grown and/or assayed in the presence of antibiotics, although these differences were not all statistically significant (Fig. 5). FSW assayed without antibiotics developed moderate settlement only between 11 and 15 days, with the later onset presumably caused by the lack of bacterial film at the start of the assay. The supernatant from the diatom film (Super /+ in Fig. 5) was drawn from a diatom culture without antibiotics, then assayed with antibiotics. It induced significantly lower attachment and metamorphosis than the equivalent diatom film ("Diat. +/-"in Fig. 5), but significantly higher attachment than the corresponding negative control ("FSW /+" in Fig. 5) on at least some days.

Diatom densities were comparable for films reared with and without antibiotics at the start of the assays, and densities fell slightly by Day 11 in some treatments because of grazing by metamorphosed postlarvae (Table 5). No diatoms were detected in nondiatom treatments on Day 0 or Day 11. At Day 0, bacterial levels in diatom films reared with antibiotics were 1000 fold lower than in diatom or bacterial films reared without antibiotics. Treatments without a biofilm and assayed with antibiotics reached bacterial densities of 2.4-5.7 x [10.sup.4] [cm.sup.-2] by Day 11, about 17-40 fold lower than the FSW without antibiotics (Table 5). Bacteria and diatoms on CCA were not quantified.

[FIGURE 5 OMITTED]

DISCUSSION

Physical Factors Affecting Settlement Induction by Diatom Strains

The present study extended previous work by attempting to better describe factors that make diatom films favorable for abalone settlement, concentrating on prostrate growth forms that are considered most suitable (Kawamura & Kikuchi 1992). In Experiments 1 and 2, settlement ranged from very low to very high depending on the diatom strain used. Physical interference with settlement was observed for several diatom strains, but the other physical characteristics of diatom films that we measured were poor predictors of settlement success. For prostrate diatoms, there is often a positive relationship between diatom density and abalone larval settlement. This relationship is generally apparent for a given diatom strain within an experiment (Kawamura & Kikuchi 1992, Daume et al. 1999, Fig. 2 of this study) but does not necessarily apply across diatom species or across experiments (this study). Furthermore, it is not clear that settlement success relates directly to diatom abundance. Other factors that often covary with diatom abundance include the age of the film, its growth phase, the mobility and adhesive strength of the diatoms, and the density of bacteria in the film. All of these factors were quantified in Experiment 2. The only significant correlations (both positive) were for larval attachment with growth phase of the diatom film, and metamorphosis with percent cover, and these correlations were only evident when old cultures were omitted. In Experiment 2, older cultures produced higher attachment and markedly higher metamorphosis than the younger cultures of the same strain, but neither a dense film nor a mature growth phase guaranteed good settlement. Marine invertebrate larval settlement commonly increases with biofilm age (e.g., Moss & Tong 1992, Hadfield et al. 1994; Wieczorec & Todd 1997, Roberts 2001b, Jin & Qian 2005). Daume et al. (2004) found higher metamorphosis of Haliotis rubra on older films of Ulvella lens, which may be because of the developmental stage of the Ulvella plants (Daume 2006) or the associated biofilm. Overall, physical characteristics explained relatively little of the variation seen in settlement success amongst diatom strains in this study.

Role of Bacteria in Settlement Induction by Diatom Films

Experiments 3-5 provided strong evidence that bacteria can play an important role in induction of Haliotis iris metamorphosis, but are not critical for larval attachment. The reisolated cultures in Experiment 3 showed a severe reduction in metamorphosis relative to the same cultures in Experiment 1, even though the larvae metamorphosed normally on proven diatom films or CCA. The virtual elimination of metamorphosis was not accompanied by a strong reduction in larval attachment. The reisolation of cells to establish the cultures for Experiment 3 most likely altered the bacterial composition of the film, or affected the development of the biofilm in other ways that reduced its metamorphosis-inducing activity. A similar effect on settlement response was seen in Experiment 4, where the presence of antibiotics inhibited metamorphosis but not attachment on Nitzschia ovalis. This evidence of a role for bacteria in triggering metamorphosis on diatom films was further supported by Experiment 5, where bacteria from a Nitzschia ovalis film induced settlement equivalent to the diatom film when assayed without antibiotics. The presence of antibiotics reduced metamorphosis on biofilms, but not on the positive controls of CCA and GABA.

The presence of antibiotics during bioassays did not prevent strong metamorphosis by many diatom biofilms (e.g., Experiment 1), GABA or CCA. Hence the antibiotics do not directly interfere with metamorphic response, but appear to prevent the microbial production of metamorphosis inducers in certain circumstances. If sufficient metamorphosis inducers are present at the start of the bioassay, then the suppression of bacterial activity by antibiotics need not prevent strong and rapid metamorphosis induction by diatom films (e.g., Fig. 1). If bacteria are required for metamorphosis induction by these highly effective diatom films, then the bacteria appear to have performed their role prior to the antibiotics being introduced, or are unaffected by the antibiotics. For relatively ineffective diatom films, bacteria can induce a gradual increase in metamorphosis over 1-2 wk (as in Experiment 5). A gradual increase in percent metamorphosis of abalone over 1-2 wk in response to bacteria has been described previously for Haliotis virginea, Haliotis iris (Roberts 2001b) and to a lesser extent Haliotis rufescens (Morse et al. 1979b). The gradual increase in metamorphosis may reflect changes in the number or nature of bacteria present (Roberts 2001b, Jin & Qian 2005, Lau et al. 2005) or chance encounters between individual larvae and pockets of sufficient cue in a biofilm where inducing chemicals are being produced and degraded by microbes (Kaspar & Mountfort 1995). Gradual metamorphosis induction often follows an extended period of larval attachment so requires much longer observation periods than are usually used in abalone settlement assays. Such bacterial induction will be muted by the presence of antibiotics in assays.

It is unclear whether abalone settlement inducers from bacteria arise independent of the diatom component of the biofilm, or rely on it. Settlement cues can combine synergistically to enhance or accelerate abalone larval settlement (Trapido-Rosenthal & Morse 1986, Roberts & Nicholson 1997, Seki et al. 1997, Roberts 2001a). So whereas bacteria alone induced abalone metamorphosis only slowly, their effect could be enhanced by combination with cues from diatoms. Alternatively, the bacteria may produce a cue from chemicals provided by diatoms, as suggested for starfish inducers produced by bacteria on CCA (Johnson & Sutton 1994). In other marine invertebrates, there is evidence that diatoms (e.g., Lain et al. 2005, Harder et al. 2002) and bacteria (e.g., Zhao et al. 2004; Lau et al. 2005) can be responsible for the settlement inducing activity of biofilms.

The supernatant from a Nitzschia ovalis culture in Experiment 5 had activity greater than the corresponding negative control, but less than the corresponding diatom film. This suggests that inducers from the diatom film were present in the cell-free supernatant, but in lower concentrations, or less potent forms, than with the film itself. A less likely explanation is that inducers were produced by the low bacterial densities that developed in the supernatant treatment (Table 5) using dissolved constituents in the supernatant.

The wide variation in metamorphosis induction by a given diatom strain in the present study (Fig. 3 and Nitzschia ovalis in Figs. 1-5) could have various causes. Growing conditions and growth rate can affect the settlement-inducing activity of diatom films by altering the quality and quantity of extracellular polymers (Lam et al. 2005). Whereas growth conditions were standardized in the present study, variation in diatom characteristics could arise with routine subculturing or reisolation. Some of the variation seen in the present study appeared to be caused by changes in the accompanying bacterial populations. These populations are generally not controlled in the maintenance of microalgal cultures for abalone farming, and they are likely to change substantially in the artificial conditions of laboratory culture. Whether the metamorphosis inducing activity of a diatom film that loses key bacteria would be reestablished in an aquaculture settlement tank (and how quickly) remains to be seen. Some of the variation between experiments will result from interbatch variation in the receptiveness of larvae to settlement cues (e.g., Baloun & Morse 1984. Trapido-Rosenthal & Morse 1986). The larval response to GABA suggested that Experiment 3 larvae were less responsive than those in Experiment 1, but the difference was insufficient to explain the marked decline in metamorphosis between these experiments.

Separation of Cues for Attachment and Metamorphosis

In Experiments 3 and 4, larval attachment was virtually unchanged by treatments that reduced metamorphosis by 1-2 orders of magnitude. Further, several diatom strains induced moderate to high percentage attachment without correspondingly high metamorphosis, a pattern that has been described previously (Kawamura & Kikuchi 1992, Roberts & Nicholson 1997, Sawatpeera et al. 2004). Hence, in some circumstances, abalone metamorphosis requires a cue separate from, or additional to, the cue(s) for larval attachment (Roberts 2001a). Furthermore, the cues for metamorphosis appear to be more specific than those triggering attachment. This presents a problem for abalone farmers because attached larvae disappear from the water column, but have not undergone the critical step of metamorphosis.

Diatoms Favorable and Unfavorable for Settlement

Cocconeis spp. dominate pregrazed plates (Akashige et al. 1981, Suzuki et al. 1987, Seki 1997), often give good metamorphosis of abalone (Kawamura & Kikuchi 1992, Roberts 2001a) and can be a key species for postlarval nutrition (Kawamura et al. 1998). In the present study, four strains of Cocconeis scutellum were tested. One gave the highest metamorphosis in Experiment 1, and another gave the highest metamorphosis in Experiment 3. However, the other two strains of Cocconeis scutellum, and the strain of Cocconeis pseudomarginata, performed no better than other diatoms. The strains that performed well were cultures where the cells remained large and rapidly growing. In the poor strains, cells had become very small in culture, or the strain was growing very slowly. This warns against persisting with selected diatom cultures that have lost vigor or changed substantially in culture.

Nitzschia longissima and the physically similar Cylindrotheca closterium (= Nitzschia closterium (Ehrenberg) Wm. Smith) are good candidates for early postlarval feeding in abalone farms (Kawamura et al. 1998). Their ability to induce metamorphosis is therefore of particular interest. Published work with abalone settlement on these species reports that response may be poor (Ohgai et al. 1991, Kawamura & Kikuchi 1992) or good (Kawamura & Kikuchi 1992, Kawamura & Takami 1995, Daume et al. 1999). In Experiment 2 of the present study, old cultures of these two species gave moderate metamorphosis rates that were second only to old cultures of Cocconeis scutellum--81. Young cultures in exponential growth phase, however, were consistently poor for attachment and metamorphosis. In Experiment 1, a dense and extremely mobile culture of Nitzschia longissima smothered larvae, causing complete mortality. Thus it seems likely that the variable performance of these species probably relates to physical and/ or chemical changes associated with the growth phase of the film.

Implications for Abalone Culture

Diatom films appear to offer little certainty as settlement cues for abalone larvae. Most strains readily induce attachment, but few strains induce consistently high metamorphosis rates. Filamentous diatoms or excessively mobile solitary diatoms can prevent successful settlement by physical interference. Thus, when abalone farmers detect the disappearance of larvae from the water column, they will be uncertain what proportion has undergone the critical step of metamorphosis, rather than become stuck on the biofilm, or attached without metamorphosing. Older films are more likely to induce a high percentage metamorphosis, but excessive development of the diatom biofilm can lead to high mortality of postlarvae. This dilemma may be managed by prolonging the diatom culture period (via very low light or a small diatom inoculum), or by preconditioning tanks with filtered seawater to establish a bacterial film. If the diatom film does become very dense, provision of bare areas by mechanical disturbance may be beneficial. There is evidence that abalone preferentially occupy bare areas after receiving metamorphosis cues from nearby biofilm (Roberts 2001b).

Diatom strains that perform well for a particular hatchery have potential to greatly aid consistent juvenile production, but no diatom species has yet proven to be effective in all tests. Diatom maintenance is straight forward through routine subculturing in standard microalgal media. However, diatoms can change in prolonged culture, so the efficacy of the diatom film needs to be confirmed periodically. Certain bacteria in the culture could be as important as the diatoms themselves. Given that the bacterial assemblage will not be deliberately controlled (unless key bacteria are identified and cultured separately) this adds uncertainty to reliance on laboratory diatom cultures. Diatom cultures will be colonized by a wide range of bacteria when exposed to nonsterile hatchery seawater. This step may or may not restore activity that had been lost with bacterial changes during laboratory culture. It could also disrupt the activity of the bacteria that were present in the diatom culture.

Given the challenges of obtaining consistently high metamorphosis on diatom films, alternative cues should be considered. Ulvella lens has been used since the 1980s in Japan (Takahashi & Koganezawa 1988), and is now producing consistently high metamorphosis and postlarval survival in Australian abalone hatcheries (Daume et al. 2004; Krsinich et al. 2006). Ulvella lens is not ingested effectively by abalone <3 mm shell length (Daume 2006) so early postlarval feeding is supported by wild diatom growth, or by inoculating tanks with cultured diatoms having favorable nutritional properties. Colonization of settlement surfaces is readily achieved by inducing spore release of Ulvella on seed plates (a system more efficient than large scale diatom culture), and Ulvella seed plates are easily maintained in flowing seawater (Krsinich et al. 2000). Overall, Ulvella lens appears to offer a favorable alternative to conventional diatom-based methods.

Factors Controlling Settlement Inducing Activity of Diatoms

This study documented several examples where diatoms exerted a negative effect on settlement by physical interference. However, physical characteristics that guaranteed successful settlement were not identified. We demonstrated that bacterial activity can greatly influence the activity of a diatom film, and that bacteria alone can induce gradual metamorphosis. The nature and abundance of chemical cues is probably the primary factor underlying the effect of bacteria and the variation in settlement among diatom strains that we observed. Cues for abalone larval attachment were present in many biofilms, whereas cues for metamorphosis were more restricted and greatly reduced by reisolation of diatom strains or antibiotic treatment. Given the wide range of abalone settlement responses to biofilms, a suite of inducers with varying potency seems more likely than a single chemical. Progress in isolating and identifying the chemical nature of inducers for abalone (Suenaga et al. 2004) and other marine invertebrates (e.g., Lam et al. 2005) will provide analytical targets to fuel more detailed investigation.

ACKNOWLEDGMENTS

The authors thank Henry Kaspar, Shirley Plant, and Stephen Webb for review of drafts and Sabine Daume for identification of the coralline alga. This research was supported by the New Zealand Foundation for Research Science and Technology, and the Japan Fisheries Agency.

LITERATURE CITED

Akashige, S., T. Seki, H. Kan-No & T. Nomura. 1981. Effects of [gamma]-aminobutyric acid and certain neurotransmitters on the settlement and the metamorphosis of the larvae of Haliotis discus hannai Ino (Gastropoda). Bull. Tohoku Reg. Fish. Res. Lab 43:37-45.

Baloun, A. J. & D. E. Morse. 1984. Ionic control of settlement and metamorphosis in larval Haliotis rufescens (Gastropoda). Biol. Bull. 167:124-138.

Bryan, P. J. & P. Y. Qian. 1998. Induction of larval attachment and metamorphosis in the abalone Haliotis diversicolor (Reeve). J. Exp. Mar. Biol. Ecol. 223:39-51.

Daume, S. 2006. The roles of bacteria and micro and macro algae in abalone aquaculture: A review. J. Shellfish Res. 25:151-157.

Daume, S., S. Brand-Gardner & Win. J. Woelkerling. 1999. Preferential settlement of abalone larvae: diatom films vs. non-geniculate coralline red algae. Aquaculture 174:243-254.

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

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

Gallardo, W. G. & S. M. A. Buen. 2003. Evaluation of mucus, Navicula, and mixed diatoms as larval settlement inducers for the tropical abalone Haliotis asinina. Aquaculture 221:357-364.

Gordon, N., M. Shpigel, S. Harpaz, J. J. Lee & A. Neori. 2004. The settlement of abalone (Haliotis discus hannai Ino) larvae on culture layers of different diatoms. J. Shellfish Res. 23:561-568.

Hadfield, M. G., C. C. Unabia, C. M. Smith & T. M. Michael. 1994. Settlement preferences of the ubiquitous fouler Hydroides elegans. In: M. F. Thompson, R. Nagabhushanam, R. Sarojini, & M. Fingerman, editors. Recent developments in biofouling control Rotterdam: A. A. Balkema. pp. 65-74.

Hahn, K. O. 1989. Handbook of culture of abalone and other marine dastropods. Boca Raton: CRC Press. pp. 71-98.

Harder, T., S. C. K. Lau, H.-U. Dahms & P.-Y. Qian. 2002. Isolation of bacterial metabolites as natural inducers for larval settlement in the marine polychaete Hydroides elegans (Haswell). J. Chem. Ecol. 28: 2029-2043.

Hugger, M. J., R. de Nys, J. E. Williamson, M. Heasman & P. Steinberg. 2005. Settlement of larval blacklip abalone, Haliotis rubra, in response to green and red macroalgae. Mar. Biol. 147:1155-1163.

Jin, T. & P.-Y. Qian. 2005. Amino acid exposure modulates the activity of biofilms for larval settlement of Hydroides elegans by altering bacterial community components. Mar. Ecol. Prog. Ser. 297:169-179.

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.

Jorgensen, E. G. 1962. Antibiotic substances from cells and solution of unicellular algae with special reference to some chlorophyll derivatives. Physiol. Plant. 15:530-545.

Kaspar, H. F. & D. O. Mountfort. 1995. Microbial production and degradation of [gamma]-aminobutyric acid (GABA) in the abalone larval settlement habitat. FEMS Microbiol. Ecol. 17:205-212.

Kawamura, T. 1994. Taxonomy and ecology of marine benthic diatoms. Mar. Fouling 10:7-25.

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

Kawamura, T. & H. Takami. 1995. Analysis of feeding and growth rate of newly metamorphosed abalone Haliotis discus hannai fed on four species of benthic diatom. Fish. Sci. 61:357-358.

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

Kirby, K. N. 1993. Advanced data analysis with SYSTAT. New York: Van Nostrand Reinhold.

Krsinich, A., S. Daume, S. Farrell, M. Gervis & P. Thompson. 2000. Towards intensification of an abalone (Haliotis rubra) nursery operation via inoculation with a benthic diatom Navicula sp. and seeding with the macroalga Ulvella lens. pp. 86-99. Proceedings of the 7th Annual Abalone Aquaculture Workshop. FRDC, Canberra, Australia.

Krsinich, A., R. Roberts, L. McPherson & S. Rodis. 2006. Achieving consistently high recruitment in selective breeding of abalone, Haliotis spp. Abstracts of the Skretting Australasian Aquaculture Conference. August. FRDC, Canberra, Australia.

Lam, C., T. Harder & P.-Y. Qian. 2005. Induction of larval settlement in the polychaete Hydroides elegans by extracellular polymers from benthic diatoms. Mar. Ecol. Prog. Ser. 289:145-154.

Lau, S. C. K., V. Thiyagarajan, S. C. K. Cheung & P.-Y. Qian. 2005. Roles of bacterial community composition in biofilms as a mediator for larval settlement of three marine invertebrates. AME 38:41-51.

Morse, D. E. 1992. Molecular mechanisms controlling metamorphosis and recruitment in abalone larvae. 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. 107-119.

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, H. Duncan & L. Jensen. 1979a. [gamma]-amino-butyric acid, a neurotransmitter, induces planktonic abalone larvae to settle and begin metamorphosis. Science 204:407-410.

Morse, D. E., N. Hooker, L. Jensen & H. Duncan. 1979b. 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. Worm Maricult. Soc. 10:81-91.

Moss, G. A. & L. J. Tong. 1992. Effect of stage of larval development on the settlement of the abalone Haliotis iris. N. Z. J. Mar. Freshwater Res. 26:69-73.

Ohgai, M., M. Wakano & S. Nagai. 1991. Effect of attached microalgae on the settlement of juveniles in abalone, Haliotis discus hamai Ino. Suisanzoshoku 39:263-266.

Roberts, R. D. 2001a. A review of settlement cues for larval abalone (Haliotis spp.). J. Shellfish Res. 20:571-586.

Roberts, R. D. 2001b. Factors affecting larval settlement and post-larval performance of New Zealand abalone, Haliotis spp. PhD Thesis. University of Otago. New Zealand. 204 pp.

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.

Roberts, R. D., H. F. Kaspar & R. J. Barker. 2004. Settlement of abalone (Haliotis iris) larvae in response to five species of coralline algae. J. Shellfish Res. 23:975-987.

Roberts, R. D., N. P. Revsbech & L. R. Damgaard. Water chemistry experienced by post-larval abalone in boundary layers overlying diatom films. J. Shellfish Res. (In press).

Sawatpeera, S., M. Kruatrachue, P. Sonchaeng, S. Upatham & T. Rojanasarampkit. 2004. Settlement and early growth of abalone larvae, Haliotis asinina Linnaeus, in response to the presence of diatoms. Veliger 47:91-99.

Searcy-Bernal, R., A. E. Salas-Garza, R. A. Flores-Aguilar & P. R. Hinojosa-Rivera. 1992. Simultaneous comparison of methods for settlement and metamorphosis induction in the red abalone (Haliotis rufescens). Aquaculture 105:241-250.

Seki, T. 1980. An advanced biological engineering system for abalone seed production. Proceedings of the International Symposium on Coastal Pacific Marine Life. Washington State University. Bellingham.

Seki, T. 1997. Biological studies on the seed production of the northern Japanese abalone, Haliotis discus hannai Ino. Bull. Tohoku Natl. Fish. Res. Inst. 59:1-71.

Seki, T. & H. Kan-no. 1981. Induced settlement of the Japanese abalone Haliotis discus hannai veliger by the mucous trails of the juvenile and adult abalones. Bull. Tohoku Reg. Fish. Res. Lab. 43:29-36.

Seki, T., K. Taniguchi & K. Kurata. 1997. The metamorphosis inducing role of dibromomethane on the Japanese abalone, Haliotis discus hannai Ino. Abstracts of the Third International Abalone Symposium. 26-31 October, 1997. Monterey, California, USA. 60 pp.

Shallow Seafarming Research Institute. 1990. Training manual on artificial breeding of abalone (Haliotis discus hannai) in Korea DPR. Shallow Seafarming Research Institute. Kosong. Korea.

Skoog, D. 1985. Principles of Instrumental Analysis. 3rd ed. New York: Saunders College Publishing.

Sokal, R. R. & F. J. Rohlf. 1981. Biometry: The principles and practice of statistics in biological research, 2nd ed. New York: WH Freeman and Company.

Stott, A. E., T. Takeuchi & Y. Koike. 2004. Testing various substances that have been bound to plastic plates with agar to induce larval settlement and metamorphosis of abalone Haliotis discus discus (Reeve). Aquaculture 231:547-557.

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, Hydrolithon samoense. Fish. Sci. 70:342-344.

Suzuki, H., T. Ioriya, T. Seki & Y. Aruga. 1987. Changes of algal community on the plastic plates used for rearing the abalone Haliotis discus hannai. Nippon Suisan Gakkaishi 53:2163-2167.

Takahashi. K. & A. Koganezawa. 1988. Mass culture of Ulvella lens as a feed for abalone, Haliotis discus hannai. NOAA Tech. Rep. NMFS 70:29-36.

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.

Tong. L. J., G. A. Moss, P. Redfearn & J. Illingworth. 1992. A manual of techniques for culturing paua, Haliotis iris, through to the early juvenile stage. New Zealand Fisheries Tech. Report No. 31.

Trapido-Rosenthal, H. G. & D. E. Morse. 1985. L-[alpha], [omega]-Diamino acids facilitate GABA induction of larval metamorphosis in a gastropod mollusc (Haliotis rufescens). J. Comp. Physiol. 155:403-414.

Trapido-Rosenthal, H. G. & D. E. Morse. 1986. Regulation of receptormediated settlement and metamorphosis in larvae of a gastropod mollusc (Haliotis ruflescens). Bull. Mar. Sci. 39:383-392.

Turley, C. M. 1993. Direct estimates of bacterial numbers in seawater samples without incurring cell loss due to sample storage. In: P. F. Kemp, B. F. Sherr, E. B. Sherr & J. J. Cole, editors. Handbook of methods in aquatic microbial ecology. Boca Raton. CRC Press: pp. 143-147.

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

Yanagihashi, S., M. Takao & K. Ohiwa. 1986. On seed production of abalone in Aichi Prefectural Sea-Farming Centre. In: Seed production system of abalone Haliotis discus discus in Aichi Prefectural Sea-Farming Centre. Aichi Prefectural Sea-Farming Centre. pp. 2-44.

Zhao, B., S. Zhang & P.-Y. Qian. 2004. Larval settlement of the silveror goldlip pearl oyster Pinctada maxima (Jameson) in response to natural biofilms and chemical cues. Aquaculture 220:1-4.

RODNEY D. ROBERTS, (1) * TOMOHIKO KAWAMURA (2) AND CHRISTINE M. HANDLEY (3) (1) Cawthron Institute, Private Bag 2, Nelson, New Zealand; (2) Ocean Research Institute, The University of Tokyo, 1-15-1 Minamidai, Nakano, Tokyo 164-8639, Japan

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

TABLE 1.
Diatoms used in Experiments 1-3. Strains marked with
* Jp were isolated from coastal waters in Shiogama,
Tohoku, Japan. All other strains were isolated from
coastal waters in Nelson, New Zealand. Diatom growth
forms are from Kawamura (1994). In Experiment 2,
"Young" and "Old" refer to the age of cultures
when bioassayed.

Strain and                       Collection     Growth    Exp. 1
Taxonomic Authority               Reference      Form

Achnanthes longipes Agardh            A            E      [check]
Achnanthes longipes Agardh          * Jp           E      [check]
Achnanthes punctifera Hustedt        25            B
Entontoneis sp.                       E            A      [check]
Cocconeis pseudomarginata
  Gregory                             S            B      [check]
Cocconeis scutellum Ehrenberg       * Jp           B      [check]
Cocconeis scutellum Ehrenberg         R            B      [check]
Cocconeis scutellum Ehrenberg        50            B
Cocconeis scutellum Ehrenberg        81            B
Cylidrotheca closterium
(Ehrenberg) Reimann and Lewin        17            A
Cylindrotheca closterium
(Ehrenberg) Reimann and Lewin        55            A
Navicula britannica
  Hustedt et Aleem                    N            A      [check]
Navicula ramosissima
  (Agardh) Cleve                      G            A      [check]
Navictda ramosissima
  (Agardh) Cleve                      7            A
Navicuht ?ramosissima                 C            G      [check]
Navicula cf. ramosissima              L            A      [check]
Navicula sp.                        * Jp           A      [check]
Nitzschia longissima var.
  reversa Grunow (A)                  K            A      [check]
Nitzschia longissima var.
  reversa Grunow (B)                  K            A      [check]
Nitzschia ovalis Arnot                J            A      [check]
Nitzschia sp.                         F            A      [check]
Pleurosigina sp.                    * Jp           A      [check]
Stauroneis consiricta
  Ehrenberg                           H            A      [check]

Strain and                         Exp. 2       Exp. 3
Taxonomic Authority

Achnanthes longipes Agardh
Achnanthes longipes Agardh
Achnanthes punctifera Hustedt       Young
                                    + Old
Entontoneis sp.                                 [check]
Cocconeis pseudomarginata
  Gregory                            Old
Cocconeis scutellum Ehrenberg
Cocconeis scutellum Ehrenberg
Cocconeis scutellum Ehrenberg    Young + Old
Cocconeis scutellum Ehrenberg    Young + Old
Cylidrotheca closterium
(Ehrenberg) Reimann and Lewin    Young + Old
Cylindrotheca closterium
(Ehrenberg) Reimann and Lewin    Young + Old
Navicula britannica
  Hustedt et Aleem
Navicula ramosissima
  (Agardh) Cleve                                [check]
Navictda ramosissima
  (Agardh) Cleve                    Young
Navicuht ?ramosissima
Navicula cf. ramosissima                        [check]
Navicula sp.                                    [check]
Nitzschia longissima var.
  reversa Grunow (A)            Young and Old   [check]
Nitzschia longissima var.
  reversa Grunow (B)
Nitzschia ovalis Arnot              Young       [check]
Nitzschia sp.                       Young       [check]
Pleurosigina sp.
Stauroneis consiricta
  Ehrenberg                                     [check]

TABLE 2.
Treatments used in Experiment 5 to assess the role
of bacteria and dissolved inducers from Nitzschia
ovalis films. N/A = not applicable.

                      Antibiotics   Antibiotics
                        Present       Present
Treatment               During      Settlement
Number and              Biofilm       Assay?
Description            Culture?

1. Nitzschia
     oxalis film          Yes           Yes
2. N. oxalis film         Yes           No
3. N. oxalis film         No            Yes
4. N. oxalis film         No            No
5. Bacteria from          No            Yes
     N. ovalis film
6. Bacteria from          No            No
     N. oxalis film
7. Supernatant from       N/A           Yes
     N. oxalis film
8. FSW                    N/A           Yes
9. FSW                    N/A           No
10. CCA                   N/A           Yes
11. GABA                  N/A           Yes

TABLE 3.
Correlations of larval attachment or metamorphosis with
characteristics of the diatom films. All characteristics
were quantified immediately before assays, except bacterial
density in Experiment 3, which was measured on the fourth
day of the bioassay. The tabulated values are Pearson's
correlation coefficients. Corrected critical values are
adjusted to hold the experiment-wise error rate at 0.05
(see Methods) with significant results (P < 0.05) in
bold type.

                                        All Diatom Treatments

Experiment 1                         2 day %         4 day %
Correlation matrix:                attachment     metamorphosis

Diatom adhesion                       0.48            0.34
Diatom percent cover                 -0.42           -0.16
Uncorrected critical value            0.48            0.51
Corrected critical value              0.73            0.76
Number of pairs                      17              14

                                        All diatom treatments

                                     4 day %         4 day %
                                   attachment     metamorphosis

Biofilm age                           0.07            0.25
Biofilm growth phase                  0.50            0.42
Diatom adhesion                      -0.21            0.27
Diatom motility                       0.17           -0.25
Diatom percent cover                  0.42            0.26
Bacterial density                     0.15            0.48
Uncorrected critical value            0.51            0.51
Corrected critical value              0.74            0.74
Number of pairs                      15              15

                                        Young treatments only

                                     4 day %         4 day %
                                   attachment     metamorphosis

Biofilm age                          -0.01           0.03
Biofilm growth phase                  0.89#          0.47
Diatom adhesion                      -0.49           0.01
Diatom motility                       0.45          -0.22
Diatom percent cover                  0.83           0.91#
Bacterial density                    -0.28          -0.13
Uncorrected critical value            0.66           0.66
Corrected critical value              0.88           0.88
Number of pairs                       7              7

TABLE 4.
Characteristics of the diatom films used in Experiment 2.
All characteristics were quantified immediately before
assays, except bacterial density, which was measured on
the fourth day of the bioassay. The old film of Achnanthes
punctifera had more than a single layer of diatom cells,
hence diatom cover exceeded 100%.

                                                  Diatom
                                       Biofilm   Adhesive
Diatom Strain                            Age     Strength

Achnanthes punctifera Old                44         4
Achnanthes punctifera Young               5         4
Cocconeis pseudomarginata Old            73         5
Cocconeis sculcllum--81 Old              36         5
Cocconeis scutellum--81 Young             6         5
Cocconeis scutellum--50 Old              73         5
Cocconeis scutellum--50 Young             5         5
Ci'linclrmheca closterium--17 Old        44         3
C'rlindrotheca clostcrium--17 Young       5         2
Crlinch-otheca closicrium--55 Old        44         3
C'ylinclrolhecu clostcrhun--55 Young      5         3
Naricula ramosissima--7 Young             5         2
Nitzschia longissima Young                5         2
Nitzschia ovalis Young                    5         3
Nitzschia sp.--F Young                    5         1

                                                 Biofilm
                                       Diatom     Growth
Diatom Strain                          Motility   Phase

Achnanthes punctifera Old                 1         2
Achnanthes punctifera Young               1         2
Cocconeis pseudomarginata Old             1         5
Cocconeis sculcllum--81 Old               1         4
Cocconeis scutellum--81 Young             1         2
Cocconeis scutellum--50 Old               1         4
Cocconeis scutellum--50 Young             2         1
Ci'linclrmheca closterium--17 Old         1         4
C'rlindrotheca clostcrium--17 Young       4         2
Crlinch-otheca closicrium--55 Old         3         4
C'ylinclrolhecu clostcrhun--55 Young      3         2
Naricula ramosissima--7 Young             5         3
Nitzschia longissima Young                3         2
Nitzschia ovalis Young                    2         3
Nitzschia sp.--F Young                    5         3

                                                 Diatom
                                                 Cover
                                                 [cm.sup.2]
                                                 [+ or -]
Diatom Strain                                    SE)

Achnanthes punctifera Old              2.0 [+ or -] 0.1 X [10.sup.6]
Achnanthes punctifera Young            1.6 [+ or -] 0.2 X [10.sup.5]
Cocconeis pseudomarginata Old          2.5 [+ or -] 0.8 X [10.sup.4]
Cocconeis sculcllum--81 Old            2.4 [+ or -] 0.4 X [10.sup.5]
Cocconeis scutellum--81 Young          1.0 [+ or -] 0.3 X [10.sup.5]
Cocconeis scutellum--50 Old            4.9 [+ or -] 0.5 X [10.sup.5]
Cocconeis scutellum--50 Young          3.9 [+ or -] 0.3 X [10.sup.4]
Ci'linclrmheca closterium--17 Old      2.2 [+ or -] 0.6 X [10.sup.5]
C'rlindrotheca clostcrium--17 Young    6.2 [+ or -] 0.5 X [10.sup.4]
Crlinch-otheca closicrium--55 Old      9.1 [+ or -] 1.9 X [10.sup.5]
C'ylinclrolhecu clostcrhun--55 Young   1.3 [+ or -] 0.1 X [10.sup.5]
Naricula ramosissima--7 Young          4.3 [+ or -] 0.7 X [10.sup.5]
Nitzschia longissima Young             7.2 [+ or -] 0.8 X [10.sup.4]
Nitzschia ovalis Young                 9.8 [+ or -] 0.8 X [10.sup.5]
Nitzschia sp.--F Young                 4.9 [+ or -] 0.7 X [10.sup.5]

                                       Diatom
                                        Cover
                                          (%
                                       [+ or -]
Diatom Strain                             SE)

Achnanthes punctifera Old              122 [+ or -] 8.7
Achnanthes punctifera Young            8.2 [+ or -] 1.2
Cocconeis pseudomarginata Old          15 [+ or -] 0.28
Cocconeis sculcllum--81 Old            43 [+ or -] 7.6
Cocconeis scutellum--81 Young          15 [+ or -] 3.7
Cocconeis scutellum--50 Old            39 [+ or -] 4.2
Cocconeis scutellum--50 Young          2.1 [+ or -] 0.17
Ci'linclrmheca closterium--17 Old      22 [+ or -] 6.2
C'rlindrotheca clostcrium--17 Young    7.0 [+ or -] 0.58
Crlinch-otheca closicrium--55 Old      60 [+ or -] 13
C'ylinclrolhecu clostcrhun--55 Young   11 [+ or -] 0.95
Naricula ramosissima--7 Young          30 [+ or -] 4.9
Nitzschia longissima Young             5.8 [+ or -] 0.65
Nitzschia ovalis Young                 64 [+ or -] 5.7
Nitzschia sp.--F Young                 17 [+ or -] 2.8

                                              Bacterial
                                              Density
                                              (cells
                                              [cm.sup.2]
                                              [+ or -]
Diatom Strain                                 SE)

Achnanthes punctifera Old              6.7 [+ or -] 1.2 X [10.sup.5]
Achnanthes punctifera Young            1.6 [+ or -] 0.4 X [10.sup.5]
Cocconeis pseudomarginata Old          9.2 [+ or -] 0.8 X [10.sup.6]
Cocconeis sculcllum--81 Old            7.8 [+ or -] 0.1 X [10.sup.6]
Cocconeis scutellum--81 Young          2.4 [+ or -] 0.1 X [10.sup.6]
Cocconeis scutellum--50 Old            1.0 [+ or -] 0.1 X [10.sup.7]
Cocconeis scutellum--50 Young          1.8 [+ or -] 0.2 X [10.sup.6]
Ci'linclrmheca closterium--17 Old      7.6 [+ or -] 0.8 X [10.sup.6]
C'rlindrotheca clostcrium--17 Young    I.1 [+ or -] 0.1 X [10.sup.6]
Crlinch-otheca closicrium--55 Old      6.8 [+ or -] 0.8 X [10.sup.6]
C'ylinclrolhecu clostcrhun--55 Young   2.1 [+ or -] 0.4 X [10.sup.5]
Naricula ramosissima--7 Young          2.9 [+ or -] 0.3 X [10.sup.5]
Nitzschia longissima Young             3.0 [+ or -] 0.5 X [10.sup.6]
Nitzschia ovalis Young                 2.7 [+ or -] 0.4 X [10.sup.5]
Nitzschia sp.--F Young                 2.0 [+ or -] 0.4 X [10.sup.5]

TABLE 5.
Densities of diatoms and bacteria in settlement
bioassays for Experiment 5, on Days 0 and 11 of
the assay. Day zero counts apply only to
established diatom and bacterial films, as other
treatments were established in sterile containers
at that time. Data are mean + 95% confidence intervals.

                         Day 0 of Assay

Treatment                   Diatoms
                           [cm.sup.-2]

Diatom +/+       1.7 [+ or -] 0.15 x [10.sup.6]
Diatom +/-       1.9 [+ or -] 0.30 x [10.sup.6]
Diatom -/+        2.6 [+ or -] 2.8 x [10.sup.6]
Diatom -/-        1.9 [+ or -] 1.5 x [10.sup.6]
Bacteria -/+                    0
Bacteria -/-                    0

                         Day 11 of Assay

Diatom +/+       2.2 [+ or -] 0.47 x [10.sup.6]
Diatom +/-       2.2 [+ or -] 0.51 x [10.sup.6]
Diatom -/+        7.5 [+ or -] 5.8 x [10.sup.5]
Diatom -/-        6.9 [+ or -] 4.5 x [10.sup.5]
Bacteria -/+                    0
Bacteria -/-                    0
Supernatant /+                  0
FSW /-                          0
FSW /+                          0
Gaba /+                         0

                         Day 0 of Assay

Treatment                   Bacteria
                           [cm.sup.-2]

Diatom +/+        3.0 [+ or -] 3.9 x [10.sup.3]
Diatom +/-        1.5 [+ or -] 2.4 x [10.sup.3]
Diatom -/+       2.9 [+ or -] 0.71 x [10.sup.6]
Diatom -/-       2.8 [+ or -] 0.67 x [10.sup.6]
Bacteria -/+     2.0 [+ or -] 0.39 x [10.sup.6]
Bacteria -/-     2.1 [+ or -] 0.42 x [10.sup.6]

                         Day 11 of Assay

Diatom +/+       2.1 [+ or -] 0.62 x [10.sup.4]
Diatom +/-        7.6 [+ or -] 3.1 x [10.sup.5]
Diatom -/+       2.0 [+ or -] 0.76 x [10.sup.6]
Diatom -/-        4.6 [+ or -] 1.1 x [10.sup.6]
Bacteria -/+      9.1 [+ or -] 4.0 x [10.sup.5]
Bacteria -/-      2.6 [+ or -] 1.0 x [10.sup.6]
Supernatant /+    2.4 [+ or -] 1.2 x [10.sup.4]
FSW /-           1.4 [+ or -] 0.45 x [10.sup.6]
FSW /+            5.7 [+ or -] 2.9 x [10.sup.4]
Gaba /+           3.0 [+ or -] 1.7 x [10.sup.4]