Allelopathic effects of Ulva pertusa, Corallina pilulifera and Sargassum thunbergii on the growth of the dinoflagellates Heterosigma akashiwo and Alexandrium tamarense.

The allelopathic effects of fresh tissue, dry powder and aqueous extracts of three macroalgae, Ulva pertusa, Corallina pilulifera and Sargassum thunbergii, on the growth of the dinoflagellates Heterosigma akashiwo and Alexandrium tamarense were evaluated using coexistence culture systems in which concentrations of the three macroalga were varied. The results of the coexistence assay showed that the growth of the two microalgae was strongly inhibited by using fresh tissue, dry powder and aqueous extracts of the three macroalga; the allelochemicals were lethal to H. akashiwo at relatively higher concentrations of the three macroalga. The macroalgae showing the most allelopathic effect on H. akashiwo and A. tamarense using fresh tissue were U. pertusa and S. thunbergii, using dry powder were S. thunbergii and U. pertusa, and using aqueous extracts were U. pertusa and C. pilulifera. We also examined the potential allelopathic effect on the two microalgae of culture filtrate of the three macroalga; culture medium filtrate initially exhibited no inhibitory effects when first added but inhibitory effects became apparent under semi-continuous addition, which suggested that continuous release of small quantities of rapidly degradable allelochemicals from the fresh macroalgal tissue were essential to effectively inhibit the growth of the two microalgae.

Introduction

In recent years there has been a growing awareness of the problems associated with red tides, which may induce mass mortalities of cultured fish or shellfish, and cause damage to aquaculture industries such as n class="Chemical">abalone farminpan>g, prawnpan> culture and caged fish culture (Mackenzie 1991; Qi et al. 1993; Honjo 1994; Horner et al. 1997; Kim 1997). Red tides are the result of massive blooms of harmful micron class="Species">algae (HABs), especially blooms of the dinoflagellates Heterosigma akashiwo (Hada) Hada ex Hara et Chihara and Alexandrium tamerance (Lebour) Balech, which cause heavy damage almost every year in China and other countries. Moreover, about 2,000 cases of human poisoning resulting from algal toxins are reported every year (Zingone and Enevoldsen 2000).

Because of the severe economic and public health problems caused by harmful micron class="Species">algae, many studies about n class="Chemical">HABs have been conducted and have been well reviewed by Hallegraeff et al. (1995). Rensel (1995) reported several techniques to mitigate the effects of harmful algal blooms on finfish aquaculture. Some promising methods have been developed, including the use of clay to sediment red tide organisms (Na et al. 1996; Choi et al. 1998), chemical agents such as copper sulfate (Steidinger 1983), hydrogen peroxide (Ryu et al. 1998), and some biological control in the form of viruses (Garry et al. 1998) or bacteria (Fukami et al. 1992; Imai et al. 1995). Although these methods seem effective in short-term experiments, they may have potentially dangerous environmental consequences; very few investigations of a direct and specific control of marine harmful algal blooms with few environmental side-effects have been reported (Jeong et al. 2000).

In the search for HAB control agents that are efficient and benign to the environment, scientists are showing interest in allelopathic substances released by other aquatic organisms for growth inhibition of HAB species. Macron class="Species">algae and micron class="Species">algae have long been known to have an antagonistic relationship in both natural and experimental aquatic ecosystems (Hasler and Jones 1949). Hogetsu et al. (1960) surmised that macrophytes released allelochemicals to inhibit algal growth. Several bioactive substances have been extracted and purified successfully for practical HAB control and management (Taft and Martin 1986; Kakisawa et al. 1988; Yu et al. 1992; Suzuki et al. 1998; Nakai et al. 1999; Jeong et al. 2000; Lee et al. 2000; Nakai et al. 2000; Jin and Dong 2003).

However, we have little knowledge of non-nutrient interactions between macron class="Species">algae and micron class="Species">algae. Thus, in this study, we selected three macroalgae, Ulva pertusa Kjellman (Chlorophyta), Corallina pilulifera Postels et Ruprecht (Rhodophyta) and Sargassum thunbergii (Mertens et Roth) Kuntze (Phaeophyta). The two harmful microalgae used in this study are common HAB species that have been recorded worldwide. Heterosigma akashiwo and Alexandrium tamarense are well known ichthyologic species through the production of reactive oxygen species or other bioactive metabolites or mucus substances (Kim et al. 1999, 2000; Nagasaki et al. 1999). Macroalgae are common in the coastal waters of China and collecting abundant macroalgae represents a potentially easy, low cost and relatively environmentally benign means of controlling HABs.

Materials and methods

Algae and experimental conditions

The axenic strains of n class="Species">H. akashiwo and n class="Species">A. tamarense were provided by the Microalgal Laboratory of Ocean University of China. U. pertusa, C. pilulifera and S. thunbergii were collected from the Taiping region of Qingdao, China. The algae were immediately carefully washed with distilled water to remove attached organisms. They were then treated with a mixture of antibiotics and grown aseptically. All the macroalgae and microalgae were cultured at room temperature with illumination of 40 μmol photons m−2 s−1, then prepared in culture for 7 days in f/2 medium (Guillard and Ryther 1962) at 20°C, using a light intensity of about 40 μmol photons m−2 s−1 with a 12:12 h light:dark cycle in illuminated incubators. All cultures were shaken twice a day to prevent wall growth. The seawater was obtained from the coast of Qingdao, filtered through 0.22 μm pore size cellulose nitrate membrane filters and autoclaved at 121°C for 20 min.

Microalgae were cultured to exponpan>enpan>tial phase before subsequenpan>t inoculationpan>. The initial cell denpan>sities of pan> class="Species">H. akashiwo and A. tamarense used in the experiments were approximately 1×105 or 8×104 cells mL−1, respectively. The culture conditions in the following experiments were the same unless otherwise stated.

Coexistence assays with fresh macroalga tissue

To compare the effects on the two micron class="Species">algae of different concentrations of fresh tissue from the three macron class="Species">algae, coexistence assays were performed using a mixed culture system of one macroalga and one microalga. Different initial inoculation concentrations (0, 0.625, 1.25, 2.5, 5.0, 10 and 15 g-wet wt L−1) of fresh algal tissues were inoculated into 100 mL conical glass flasks containing 40 mL f/2 culture medium. Exponentially growing H. akashiwo or A. tamarense were added to the fresh tissue seaweed culture in co-culture systems. Controls were prepared by inoculating the microalgae into f/2 culture medium without the addition of a macroalga.

Microalgae samples (1 mL) were takenpan> daily anpan>d fixed with acidic Lugol’s anpan>d counpan>ted with a haemocytometer. We conpan>ducted the experimenpan>t unpan>der stable enpan>vironpan>menpan>tal conpan>ditionpan>s, anpan>d thenpan> 1 mL nutritionpan> solutionpan>, which conpan>tained 20 x f ingredienpan>ts for 40 mL culture medium, was added to each flask to rule out nutrienpan>t depletionpan> anpan>d avoid nutrienpan>t competitionpan> betweenpan> micropan> class="Species">algae and macroalgae. The concentration of NO3-N and PO4-P was also routinely measured throughout the experiment. The pH of the culture medium and the wet weight of the macroalga were also measured.

Macroalga culture medium filtrate assays

Initial addition of macroalga culture medium filtrate

The initially added culture solution was prepared by culturing the three macroalgae over 3 days in f/2 medium at a conpan>cenpan>trationpan> of 80 g-wet wt L−1. This culture solutionpan> was added with nutrienpan>ts equivalenpan>t to the medium anpan>d filtered through anpan> autoclaved membranpan>e filter (Whatmanpan>, 0.22 μm); 40 mL of the filtrate was thenpan> immediately added to a 100 mL flask anpan>d inoculated with pan> class="Species">H. akashiwo or A. tamarense, while microalgae cultured in fresh f/2 medium were used as controls.

Semi-continuous addition of macroalga culture medium filtrate

The same culture solution as above (40 mL) was added to a 100 mL flask and inoculated with H. akashiwo or A. tamarense. Each day 10 mL of the 40 mL culture medium was removed from each flask and 10 mL nutrient re-enriched macroalga culture medium filtrate was added to keep the culture medium volume constant. In the control the filtrate was replaced by fresh f/2 medium.

Boiled macroalga culture medium filtrate assays

The macroalga culture medium was boiled for 20 min at 121°C, filtered through an autoclaved membrane filter (Whatman, 0.22 μm) and re-enriched with f/2 medium. Cells of H. akashiwo or pan> class="Species">A. tamarense were immediately inoculated into the culture medium filtrate of the three macroalgae, while controls were prepared by inoculating microalgae into f/2 medium only.

Coexistence assays

Fresh macroalga tissue (8 g) was ground with some distilled n class="Chemical">water usinpan>g a mortar and centrifuged three times usinpan>g sean class="Chemical">water. About 80 mL supernatant was collected and then diluted with seawater to give different concentrations of aqueous extracts (0, 2, 4, 8, 16, 24 g-wet wt L−1). These extracts were inoculated into flasks containing microalgae. Controls were prepared by inoculating only with microalgae without adding the aqueous extraction.

The method for preparing the aqueous extracts of the macron class="Species">algae was the same as described above. About 50 mL solution was boiled for 20 minpan> at 121°C and then diluted with sean class="Chemical">water to the desired concentration (8 g-wet wt L−1) and inoculated with exponentially growing microalgae. Controls were prepared by inoculating microalgae in the medium without adding the aqueous macroalga extract.

Coexistence assays with dry macroalga powder

For the convenient use of samples, fresh macroalgal tissue was dried completely for 6 days at room temperature and then ground to a powder using a mortar and pestle. Different amounts of dry powder with an initial concentration of 0, 0.3, 0.6, 1.2 and 2.4 g-dry wt L−1 were inoculated into 100 mL flasks, and the dinoflagellates n class="Species">H. akashiwo and n class="Species">A. tamarense were also immediately added. The control contained only microalgae.

Statistics

All experiments were repeated at least three times for each independent assay. Mean values were compared to the controls using student's t-test. P <0.05 was considered significant.

Results

Sufficient amounts of n class="Chemical">N and P were available for n class="Species">algae growth. The daily addition of 1 mL 20xf solution resulted in nutrient concentrations of NO3-N ≥780 μmol L−1 and PO4-P ≥36 μmol L−1, which are well above limiting levels for microalgae (Guillard and Ryther 1962; Sharp et al. 1979). An effect of lighting conditions was also ruled out by measuring the light intensity at the bottom of the culture flask; the results confirmed that inhibition did not occur due to insufficient light (data not shown). The pH measured daily for all culture media was pH 8.5 ± 0.3, which is well suited for the growth of microalgae (Chen and Durbin 1994).

Effects of fresh macroalga tissue

Figure 1 shows the growth curve of the three macroalgae coexistenpan>ce cultures with pan> class="Species">H. akashiwo or A. tamarense. The growth of H. akashiwo was significantly suppressed (P <0.05) by the three macroalgae at the different concentrations tested (Figure 1a–c). Cells of H. akashiwo were all dead at concentrations higher than 2.5 g-wet wt L−1 after 3 days with U. pertusa, while they were all dead within 8 days with the other two macroalgae when concentrations of >5 g-wet wt L−1 were used. The growth of A. tamarense was significantly (P <0.05) suppressed by all three macroalgae; however, no lethal effect was observed (Figure 1d–f).

Figure 1

Growth curves of Heterosigma akashiwo (a–c) or Alexandrium tamarense (d–f) coexistence assays with different initial inoculation concentrations (□ 0; ▪ 0.625; ▴ 1.25; × 2.5; □ 5; • 10; □ 15 g-wet wt L−1) of fresh tissue of Ulva pertusa (a, d), Corallina pilulifera (b, e) or Sargassum thunbergii (c, f) in coexistence assays. Data are the means ± SD from at least three independent assays.

Figure 2 shows the respective dose-response relationships between the three macroalgae anpan>d the two micropan> class="Species">algae. Results were normalized to maximum growth based on previous methods (Nakai et al. 1999). At the concentration used, normal growth of microalgae was inhibited 50% by the macroalga.

Figure 2

Normal maximum growth of H. akashiwo (a) or A. tamarense (b) in the presence of fresh tissue of U. pertusa (▴), C. pilulifera (□) or S. thunbergii (▪). Data are the means ± SD from at least three independent assays.

The growth of H. akashiwo was more stronpan>gly suppressed by fresh tissue of pan> class="Species">U. pertusa than fresh tissue of the other two macroalgae (Figure 2a). The EC50 concentration for the fresh tissue of U. pertusa was 0.5 g-wet wt L−1. The growth of A. tamarense was more strongly suppressed by fresh tissue of S. thunbergii than fresh tissue of the other two macroalgae (Figure 2b). The EC50 concentrations for S. thunbergii, U. pertusa and C. pilulifera were 0.8, 1.7 and 4.0 g-wet wt L−1, respectively (Table 1).

Table 1

Summary inhibitory effects (EC50 values) of fresh tissue (g-wet wt L−1) or dry powder (g-dry wt  L−1) of Ulva pertusa, Corallina pilulifera or Surgassum thunbergii on Heterosigam akashiwo or Alexandrium tamarense

Treatment	Heterosigam akashiwo	A. tamarense
Fresh tissue of U. pertusa	0.5	1.7
Fresh tissue of C. pilulifera	1.8	4.0
Fresh tissue of S. thunbergii	2.5	0.8
Dry powder of U. pertusa	0.5	0.6
Dry powder of C. pilulifera	1.0	1.0
Dry powder of S. thunbergii	0.45	1.3

Summary inhibitory effects (EC50 values) of fresh tissue (g-wet wt L−1) or dry powder (g-dry wt  L−1) of Ulva pertusa, pan> class="Species">Corallina pilulifera or Surgassum thunbergii on Heterosigam akashiwo or Alexandrium tamarense

Effects of macroalga culture medium filtrate

The growth of n class="Species">H. akashiwo was not signpan>ificantly (P > 0.05) inpan>hibited by the culture medium filtrate of n class="Species">U. pertusa or S. thunbergii under the initial or semi-continuous filtrate addition conditions. The initial addition of C. pilulifera culture solution did not inhibit growth of H. akashiwo; however, semi-continuous addition significantly inhibited the growth of H. akashiwo (P < 0.05), and the cells all died (Figure 3a–c).

Figure 3

Growth curves of H. akashiwo (a–c) or A. tamarense (d–f) under initial (▪) or semi-continuous (▴) culture medium filtrate addition of U. pertusa (a, d), C. pilulifera (b, e) or S. thunbergii (c, f) with control (□ initial addition; □ semi-continuous addition). Data are the means ± SD from at least three independent assays.

Growth curves of H. akashiwo (a–c) or pan> class="Species">A. tamarense (d–f) under initial (▪) or semi-continuous (▴) culture medium filtrate addition of U. pertusa (a, d), C. pilulifera (b, e) or S. thunbergii (c, f) with control (□ initial addition; □ semi-continuous addition). Data are the means ± SD from at least three independent assays.

The growth of n class="Species">A. tamarense was signpan>ificantly (P < 0.05) inpan>hibited by the culture medium filtrate of n class="Species">U. pertusa under the initial or semi-continuous filtrate addition conditions, and all cells died at the end. The effect of the culture medium filtrate of S. thunbergii on A. tamarense was the same as for U. pertusa, but no lethal effect was observed. The growth of A. tamarense was not significantly inhibited by the culture medium filtrate of C. pilulifera under either initial or semi-continuous addition (Figure 3d–f).

Effects of boiled macroalga culture medium filtrate

The growth of A. tamarense was signpan>ificanpan>tly (P < 0.05) inhibited by the boiled culture medium filtrate of pan> class="Species">S. thunbergii. However, the inhibitory effect was lower than that of the non-boiled culture medium filtrate and no lethal effect was observed (Figure 4a). The semi-continuous addition of the boiled C. pilulifera culture medium filtrate significantly (P < 0.05) inhibited the growth of A. tamarense but the effect was lower than that observed with non-boiled filtrate. No lethal effect was observed on H. akashiwo (Figure 4b).

Figure 4

Growth curves of A. tamarense (a, c) or H. akashiwo (b) with addition of boiled culture medium filtrate of S. thunbergii (a), C. pilulifera (b) or U. pertusa (c) with control (□ control; ▪ treatment). Data are the means ± SD from at least three independent assays.

Growth curves of A. tamarense (a, c) or pan> class="Species">H. akashiwo (b) with addition of boiled culture medium filtrate of S. thunbergii (a), C. pilulifera (b) or U. pertusa (c) with control (□ control; ▪ treatment). Data are the means ± SD from at least three independent assays.

Within 5 days, the growth of A. tamarense was significantly lowered by the boiled culture medium filtrate of U. pertusa (P < 0.05). However, over the period of the experiment (5–8 days), the growth of A. tamarense recovered to normal levels (Figure 4c).

Effects of different concentrations of aqueous extracts

The growth of H. akashiwo was signpan>ificanpan>tly reduced (P < 0.05) by all tested conpan>cenpan>trationpan>s of aqueous extracts of pan> class="Species">U. pertusa , C. pilulifera and S. thunbergii within the first 4, 2 and 2 days, respectively. Inhibition of growth increased with extract concentration. Aqueous extracts of all three macroalgae at a concentration of 24 g L−1 resulted in the death of H. akashiwo within 2 days. The growth of H. akashiwo was stimulated at the lower concentration treatments (2–8 g L−1). During the 4th and 10th day of the experiment, aqueous extract of U. pertusa at a concentration of 2 g L−1 and extracts of the two other macroalgae at a concentration of 2–4 g L−1, resulted in faster and higher growth than in the control culture of H. akashiwo (Figure 5a–c).

Figure 5

Growth curves of H. akashiwo (a–c) or A. tamarense (d–f) coexistence assays with different initial concentrations (□0; ▪2; ▴4; ×8; □16; •24 g-wet L−1) of aqueous extracts of U. pertusa (a, d), C. pilulifera (b, e) or S. thunbergii (c, f). Data are the means ± SD from at least three independent assays.

The growth of n class="Species">A. tamarense was signpan>ificantly (P < 0.05) inpan>hibited by all tested concentrations of the aqueous extracts of the three macron class="Species">algae within the first 3 days of the experiment. The inhibitory effect of the aqueous extract of C. pilulifera was higher than that of the two other macroalgae during the first 2 days of the experiment. Aqueous extracts of C. pilulifera and S. thunbergii at a concentration of 8 g L−1 or 16 g L−1, respectively, resulted in the death of the A. tamarense culture. However, no lethal effects were observed for the aqueous extract of U. pertusa (Figure 5d–f).

Effects of boiled aqueous extracts

Within the first 2 days, the growth of H. akashiwo was stronpan>gly (P < 0.05) inhibited by the boiled aqueous extracts of pan> class="Species">U. pertusa. The growth of H. akashiwo recovered and was not inhibited significantly (P > 0.05) at the end of the experiment (Figure 6a). Growth of H. akashiwo and A. tamarense was inhibited (P < 0.05) by the boiled aqueous extracts of C. pilulifera, but the inhibitory effect was weaker than that of the non-boiled aqueous extracts (Figure 6b, c). The growth of A. tamarense was inhibited (P < 0.05) by boiled aqueous extracts of S. thunbergii, and the inhibitory effect was greater than that observed with the non-boiled extracts (Figure 6d).

Figure 6

Growth curves of H. akashiwo (a, b) or A. tamarense (c, d) coexistence assays with boiled aqueous extracts of U. pertusa (a), C. pilulifera (b, c) and S. thunbergii (d). Data are the means ± SD from at least three independent assays.

Effects of dry macroalga powder

In the first 6 days, the growth of n class="Species">H. akashiwo was signpan>ificantly (P < 0.05) inpan>hibited by all tested concentrations of dry powder of n class="Species">U. pertusa and S. thunbergii and at the two highest concentrations of dry powder (1.2 and 2.4 g-dry wt L−1) all the cells of H. akashiwo died within 1 day. The inhibition effect of the dry powder of Corallina was lower than for the two other macroalgae; at high concentrations (1.2 g-dry wt L−1), all cells of H. akashiwo died within 5 days, whereas at lower concentrations (0.3–1.2 g-dry wt L−1), the growth of H. akashiwo was not significantly inhibited. In contrast, the growth of H. akashiwo was significantly increased at a dry powder concentration of 0.3 g-dry wt L−1(Figure 7a–c).

Figure 7

Growth curves of H. akashiwo (a–c) or A. tamarense (d–f) coexistence assays with different initial concentrations (□ 0; ▪ 0.3; □ 0.6; × 1.2; • 2.4 g-dry wt L−1) of dry powder of U. pertusa (a, d), C. pilulifera (b, e) or S. thunbergii (c, f). Data are the means ± SD from at least three independent assays.

In the 10-day coexistence assays, the growth of n class="Species">A. tamarense was signpan>ificantly inpan>hibited at all concentrations of dry powder of n class="Species">U. pertusa and S. thunbergii. At the highest concentrations of the dry powder of U. pertusa, the biomass of A. tamarense was lower and no lethal effect was observed. However, all cells of A. tamarense died when 2.4 g-dry wt L−1 of the dry powder of S. thunbergii was added (Figure 7d, f). The growth of A. tamarense was not significantly inhibited by the dry powder of C. pilulifera at concentrations lower than 0.6 g-dry wt L−1. On the contrary, at the end of the experiment, the growth of A. tamarense was significantly heightened in comparison with the control. At the highest concentrations (1.2 and 2.4 g-dry wt L−1) A. tamarense died within 1 day (Figure 7e).

The growth of H. akashiwo was more stronpan>gly inhibited by the dry powder of pan> class="Species">S. thunbergii than by the two other macroalgae (Figure 8a). The EC50 concentrations of the dry powder of S. thunbergii, U. pertusa and C. pilulifera on H. akashiwo were 0.45, 0.5 and 1.0 g-dry wt L−1, respectively. The growth of A. tamarense was more strongly inhibited by the dry powder of U. pertusa than by the two other macroalgae (Figure 8b), and the EC50 concentrations of U. pertusa, C. pilulifera and S. thunbergii, on A. tamarense were 0.6, 1.0 and 1.3 g-dry wt L−1, respectively (Table 1).

Figure 8

Normal maximum growth of H. akashiwo (a) or A. tamarense (b) in the presence of dry powder of U. pertusa (▴), C. pilulifera (□) or S. thunbergii (▪). Data are the means ± SD from at least three independent assays.

Discussion

Allelopathy is a common natural phenomenon in aquatic ecosystems, being seen in many biochemical interactions among n class="Species">higher plants and between n class="Species">higher plants and microorganisms, and including both stimulatory and inhibitory interactions (Molisch 1937). However, it is difficult for researchers to study allelopathic effects among aquatic organisms under natural conditions because factors such as nutrient and light competition, temperature and pH changes can totally mask allelopathic effects (Keating 1977).

We conducted laboratory experiments under stable environmental conditions, precluding nutrient and light competition, in order to investigate the allelopathic effects of three macron class="Species">algae on two micron class="Species">algae. The growth of Heterosigma akashiwo and Alexandrium tamarense was strongly inhibited by the fresh tissues of three macroalgae, with cells of H. akashiwo being completely killed at higher concentrations of macroalgae. Ulva pertusa and Sargassum thunbergii culture medium filtrate did not exhibit an apparent inhibitory effect on the growth of H. akashiwo under the initial or semi-continuous addition conditions; however, they did exhibit significant growth inhibition on A. tamarense.

Schmidt and Hansen (2001) investigated the effects of pH on the immobilization of Heterocapa triquetra cells by n class="Species">Chrysochromulina polylepis and noted that pH had a dramatic effect on n class="Species">H. triquetra. Macroalgae may change the pH of the culture medium during growth, making it unsuitable for microalgal growth. In the coexistence assays, we measured the pH of the culture medium at the beginning and the end of the experiment and there was no evidence that a change in pH of the culture medium played an important role in the growth inhibition of the two microalgae. Hence, the secretion of allelopathic substances by the three macroalgae is the most likely explanation for the observed growth inhibition.

The results from both coexistence and macroalga culture medium filtrate assays indicate that the three macron class="Species">algae release rapidly degradable allelopathic substances, and that a continpan>uous allelochemical secretion from fresh tissue is essential to effectively inpan>hibit the growth of the micron class="Species">algae. Nakai et al. (1999) demonstrated that the growth inhibition of cyanobacteria by the macrophyte Miriophyllum spicatumn required continuous secretion of some unstable, growth-inhibitory allelopathic compounds. They also found that growth of Microcyctis aerugimosa was not apparently inhibited by the initial addition of culture solution of Miriophyllum spicatumn, whereas a quasi-continuous addition did lead to inhibitory effects on growth. Jin and Dong (2003) reported the same phenomenon, indicating that a very small amount of the hypothetical rapidly degradable allelochemicals was continuously released into the culture medium by Ulva pertusa. Van Donk and van de Bund (2002) found that growth reduction of Scenedesmus acutus by Chara aspera occurred only when C. aspera was actually present in the medium during the experiment; when the algae were inoculated into medium in which C. aspera had been grown but removed before the experiment, no effect could be demonstrated. The phenomenon observed in the latter experiment is similar to our observations. Although the investigations mentioned above indicated the continuous secretion of unstable growth inhibitory allelochemicals in freshwater ecosystems, we speculate that this phenomenon may also exist in marine ecosystems. Sanna et al. (2004) reported that the sensitivity of an organism might also depend on the nature of the allelochemicals to which it is exposed, since the same organism may respond differently to filtrates from different algae. In our experiment, we believe that the inhibitory effect of the macroalga culture medium filtrate on microalgae is species-specific.

The inhibitory effects of fresh tissue and dry powder of n class="Species">Ulva pertusa on the growth of n class="Species">Heterosigma akashiwo and Alexandrium tamarense were considerably stronger than those of the two other macroalgae, with the exception of the inhibitory effect of dry powder of Sargassum thunbergii on A. tamarense, which can be inferred from the EC50 (Table 1). Therefore, we think that our experiments indicate that U. pertusa is a broad-spectrum macroalga that exhibits an inhibitory effect on these two HABs, but that S. thunbergii is a differential macroalga. The dry powder of the three macroalga inhibited the growth of the two microalgae considerably more than the fresh tissue of the macroalgae, with the exception of the effect of the dry powder of S. thunbergii on A. tamarense. This was probably because the allelochemicals in the dry powder were added to the culture medium of the microalgae as a large pulse, so that the initial allelochemical concentrations were much higher than those in the culture media of the two microalgae coexisting with the fresh tissue of the macroalga, even though the allelochemical supply pattern was not continuous.

We observed that high concentrations of the dry powder and aqueous extracts of the three macroalga strongly inhibited the growth of the micron class="Species">algae and killed all the cells of n class="Species">H. akashiwo and A. tamarense. At lower concentrations neither macroalga affected growth, or they in fact stimulated microalgae growth. This was probably because, at higher concentrations, the microalgae cells were all killed by the allelopathic substances in a short time interval, while at lower concentrations some microalgal cells were still alive after degradation of a part of the allelochemicals, enabling the growth of cells that survived the allelochemicals to be supplied by the large quantity of nutrients released from dead microalgae cells by physical leaching. Jin and Dong (2003) reported the effect of dry powder of the nonsexual strain of U. pertusa on the growth of H. akashiwo. Their observations were similar to ours.

n class="Species">Heterosigma akashiwo and n class="Species">A. tamarense showed different responses when coexisting with the fresh tissue of the macroalgae. Relatively low concentrations had lethal effects on H. akashiwo, whereas the cells of A. tamarense were not killed completely even at the highest concentration of macroalgae. Guo (1994) states that there is no cell wall around H. akashiwo cells and that the outermost layer of the cell is a naked cell membrane, whereas the A. tamarense cell is covered by a hard wall (Qi and Qian 1994). The difference in the cell surface structures between H. akashiwo and A. tamarense cells may account for the different responses to allelochemicals. Kakisawa et al. (1988) reported that the brown alga Cladosiphon okamuranus produced some allelopathic substances against several HAB algae including H. akashiwo. They discovered that these allelochemicals were active against phytoplankton without cell coverings and inactive to those with rigid cell walls.

Macron class="Species">algae such as n class="Species">U. pertusa, C. pilulifera and S. thunbergii are widely dispersed. Collection and cultivation of abundant macroalgae species represents an easy, economical and environmentally benign means of potential HAB control in confined areas (Jeong et al. 2000). In this study, our results have demonstrated that three macroalgae from the Chlorophyta, Phaeophyta and Rhodophyta can release some allelochemicals that effectively inhibit the growth of the dinoflagellates H. akashiwo and A. tamarensis. Application of allelopathy in the control of microalgae blooms may require identification of the allelochemicals; however, such identification of natural products is often difficult. It is paramount in this case because it would instigate the development of ecologically desirable, highly specific, biological algaecides. These three macroalgae possess several allelopathic substances, and purification of the active substances is now in progress.

Further research is also needed to elucidate the mechanism of selective allelopathic effects against harmful red tide micron class="Species">algae. Although it is possible to control algal growth inpan> an actual ecosystem by the addition of macron class="Species">algae, their overgrowth would have a negative impact (Nakai et al. 1999), thus the algicidal effect of macroalgae extracts, powders or culture medium filtrates should be encouraged for the control of red tides in confined coastal areas.