A red tide refers to the periodic reddish discolouration of seawater due to intense accumulations (blooms) of minute planktonic organisms called dinoflagellates, which contain pigments such as peridinin, b-carotene and xanthophyll (besides chlorophyll). Typically 10^-20 x 10⁶ cells per litre, or even more, may be involved.

These dinoflagellates (Pyrrophyta, Dinoflagellata) are unique organisms in a number of respects: The nucleus contains chromosomes that are always visible, even under the light microscope; no true histones are present and, during nuclear division there is no true mitotic spindle; in photosynthetic species, peridinin in a chlorophyll protein is responsible for light harvesting; and they can synthesize a variety of highly toxic molecules.

The marine dinoflagellates are a very successful and old group of eukaryotic microorganisms adapted to a range of pelagic and benthic habitats throughout the world's oceans. Many species are cosmopolitan. Free living species may be photosynthetic or heterotrophic or both. Over 2,000 species have been determined world wide, of which more than 50% are photosynthetic forms. There are also some parasitic and symbiotic forms. Most dinoflagellates are important as food for fish, shrimps and shellfish.

The colour of a bloom can in fact range from yellowish-brown to chocolate, orange, rusty brown to maroon depending on the causative organism, its concentration and physiological state. The organisms involved, other than dinoflagellates, being diatoms, purple bacteria, other flagellates and ciliates. Blooms resulting from these latter groups are often referred to as "brown tides" to differentiate them from the dinoflagellate based red tide.

Red tides are natural phenomena, which used to be considered as unusual events, but which are now common in many coastal areas. They have been reported from the Atlantic and Pacific coasts of North America, from Japan, Australia, China and various other parts of Asia, and Europe (for example, Halstead, 1965; Rounsefell & Nelson, 1966; MacLean, 1979, 1984; Tangen, 1979; Qi et al, 1993). Taylor (1987) claims that the term "red tide" originated from the sporadic reddish brown discolouration of sea water along the Atlantic and Pacific coasts of North America.

Often red tides have no obvious harmful effects, although large concentrations can lead to fish kills as a result of deoxygenation of the water. However, 20 or so species of marine dinoflagellates produce toxins (Steidinger & Baden, 1984) which can kill fish and shellfish or, via contaminated fish and shellfish, affect human health. Red tides have also been reported to cause minor irritations to the eyes, skin and respiratory tracts of swimmers (Steidinger & Joyee, 1973; Hemmert, 1975; Machado, 1979). The toxic species belong to widely divergent genera and there is no obvious explanation as to why these few species produce such potent neurotoxins and haemolytic agents, while over a thousand species do not produce such secondary metabolites. The only thing they have in common is that they are all photosynthetic species and synthesize long-chain, unsaturated fatty acids as food reserves; but then so do most free-living dinoflagellates.

It is still not certain exactly what it is that causes red tides. In order for one to begin, an inoculum must be present, and yet the majority of red tide organisms do not appear in the water column throughout the year. Many of them occur as cysts in the bottom sediments (various Gonyaulax, Gymnodinium and Prorocentrum species) and, on excystment, enter the water column; others are carried by ocean currents from one region to another (for example, Alexandrium catenella is suspected to be brought to Hong Kong in the Kuroshio current (Ho & Hodgkiss, 1993b); and there is evidence that yet others are transported great distances in ship ballast waters (Alexandrium catenella and A. tamarense, Hallegraeff & Bolch, 1991).

The rate at which the dinoflagellates then grow depends upon a whole range of factors including light, temperature, salinity, nutrient supply and grazing. Despite this apparent complexity, various generalizations can be made: In temperate and high latitudes red tides are predominantly summer phenomena; in the tropics they often follow heavy rain; and, in many instances, they have been related to the input of nutrients.

Some scientists believe that long-term climatic shifts may be responsible for the apparent increase in both frequency and force of red tides worldwide (see Hallegraeff, 1993, for the proposed link between the greenhouse effect and blooms); others believe that human caused changes, such as the flow of sewage and other nutrients into estuaries, encourage red tides even if the pollution is not actually the stimulus for their inception (Hodgkiss & Yim, 1995). Undoubtedly various factors contribute to red tide formation - nutrient upwelling and nutrient input, temperature changes, wind, rainfall, salinity, and light have all been identified as possibilities.

Many scientists now agree that the term "bloom" is not entirely appropriate to the description of red tides. This is because, firstly, phytoplankton blooms are predictable and occur with regularity in particular seasons, while large concentrations of dinoflagellates do not appear every year and, even when they occur can show distinct spatial patchiness. Indeed, this characteristic of spatial patchiness is probably the single feature which most easily distinguishes red tides from blooms. Secondly, blooms consist of many different species, whereas red tides are markedly single species events and 90-95% of all the phytoplankters in a red tide belong to a single species.

Not all dinoflagellates colour the sea red. Thus, the non-photosynthetic forms which lack peridinin commonly produce "red tide" which is not red. Perhaps the best example is Noctiluca miliaris, the best known characteristic of which is not red tides but bioluminescence.

The first red tide to be documented was in Exodus Chapter 20, where it is described as a plague in Egypt (Taylor, 1987). The earliest scientific record, however, appears to have been by Carter (1858), who reviewed a number of ancient records such as those of Charles Darwin in 1835 and an outbreak in Ireland in 1694. A host of publications on red tides is now available in the literature, including the proceedings of a number of international conferences on red tides or toxic marine phytoplankton (Lo Cicero, 1975; Taylor & Seliger, 1979; Anderson et al, 1985; Okaichi et al, 1989; Graneli et al, 1990; Smayda & Schimizu, 1993).

Red tides have only caught the attention of researchers outside temperature regions over the last two decades and little attention was paid to them in the developing subtropical countries until the late 1970s and early 1980s (Lam & He, 1989a). Ho & Hodgkiss (1991) reviewed red tide occurrences in subtropical waters from 1828 to 1989 and showed the exponential growth in their numbers during the 1950s to 1980s. In fact, red tides have now been reported in more than 19 countries in the subtropical region, with the majority occurring in the Western Pacific Ocean (Ho & Hodgkiss, 1991). These subtropical red tides show monthly variation with little difference between the Northern and Southern Hemisphere patterns. About 70% of the occurrences (Ho & Hodgkiss, 1991) were recorded between February and May, so that in the Northern Hemisphere maximum occurrence was in spring, whereas in the Southern Hemisphere it was in autumn. The September/November peak is less significant in subtropical waters than in temperate waters (Taylor, 1987), probably related to the prolonged summer and relatively insignificant temperature difference in the vertical column resulting in a more even monthly distribution of the algae.

The major dinoflagellate genera involved in these subtropical red tides included Alexandrium, Ceratium, Exuviella, Gymnodinium, Gonyaulax, Noctiluca, Peridinium, Prorocentrum, Pyrodinium and Scrippsiella (Ho & Hodgkiss, 1991). These authors also noted that Southeast Asia has been severely affected by toxic red tides in recent years. The major paralytic shellfish poisoning (PSP) causative organisms were Alexandrium catenella Balech, A. cohorticula Balech and A. tamarensis Balech. Dinophysis caudata Sarille-Kent and Gymnodinium breve Ehrenberg were responsible for neurotoxic shellfish poisonings and Gambierdiscus toxicus (Adachi and Fukuyo) was the most common ciguatoxin secreting form.

Red tide was first reported in China by Fei (1952), but in recent years there has been an increase in the number of these events (Qi et al, 1989). On the basis of incomplete data, it is impossible to say how many such red tides have occurred, but records are available of 169 between 1980 and 1990 (Qi et al, 1993) and of 87 as at 1987, not including the 171 records for Hong Kong up to that time (Tseng et al, 1993). The latter authors recorded 91 phytoplankton bloom forming species, of which 11 were toxic species; Lin (1987) recorded 81 species in the neritic waters of Guangdong and Hainan Island; Lin (1985) recorded 101 species from the North East South China Sea; Qi et al (1993) recorded 60 red tide causative species; and Lin & Zhou (1993) presented SEM photographs of 61 red tide dinoflagellates from the South China Sea. The most common species recorded were Alexandrium tamarensis Balech, Ceratium furca Ehrenberg, Gonyaulux polyedra Stein, Noctiluca scintillans Ehrenberg, and Prorocentrum minimum Schiller.

In Hong Kong, the first possible reference to red tides was in the 1819 edition of the "Gazetteer" where, in Chapter 13 referring to Natural Disasters' in 1629, oysters were described as having blood in them so that people did not dare eat them (Ng 1983). The first scientific reference to a red tide was given by Morton & Twentyman (1971), who refered to an extensive bloom of Noctiluca scintillans widely affecting the beaches of Hong Kong in June 1971. Increasing numbers were recorded during the 1970s and 1980s (Holmes & Lam, 1985; Hodgkiss & Chan, 1987; Wong & Wu, 1987; Lam & He, 1989a, b; Wong, 1989a; Ho & Hodgkiss, 1993a; Hodgkiss, 1993; Ho & Hodgkiss, 1995; Hodgkiss, 1995). The first massive fish kills related to toxic red tides were recorded by Lam & Yip (1989). The causative organism was Gonyaulax polygramma. Toxification of shellfish by Protogonyaulax catenella was recorded in Junk Bay by Wong (1989b) but Chan (personal communication) has shown the presence of PSP neurotoxins in Hong Kong market shellfish since the late 1980s. Protogonyaulax catenella was the causative organism and the toxin reached 13800 mu kg^-1 in March 1989. An anonymous author has recorded 397 cases of ciguatera toxin poisonings as a result of Gambierdiscus toxicus contaminated fish between 1988 and 1992 (Anon, 1994).

Little experimental research has been carried out on the red tide dinoflagellates of the South China Sea. Exceptions are the work of Ho & Hodgkiss (1993a), who used bottle bioassay methods to determine the factors limiting dinoflagellate growth and that of Lin et al (1993) on enclosure experiments with red tide organisms. As a result of studies on the Tolo Harbour and Tai Tam Bay phytoplankton (Chan & Hodgkiss, 1987; Hodgkiss & Chan, 1983; 1986; 1987; Chan, Chin & Hodgkiss, 1991; Ho & Hodgkiss, 1991, 1993a, b; Hodgkiss, 1993; Chiu, Hodgkiss & Chan, 1994; Ho & Hodgkiss, 1995; Hodgkiss, 1995), the linkage between dinoflagellate increases in Hong Kong waters and nutrient enrichment, particularly the atomic N : P ratios in seawater, has been clearly established (Hodgkiss & Ho, 1997).

Similarly, very little research has been carried out concerning the control of red tides. Chemical addition or zooplankton predation have been suggested by some researchers (Oda, 1935; Kutt and Martin, 1974; Saifullah, 1979). Steidinger (1983) reviews a number of such proposals. Later, Taylor (1987) referred to the possible use of biological control using dinoflagellate parasites. In China, Yu et al (1993) reported that it was very effective to precipitate blooming red tides by means of inorganic coagulants, surfactant or polymers. Recently, Ho et al (1998) also reported that it was effective to make use of ozone to kill red tide organisms and to increase dissolved oxygen in seawater. However, these methods have met with little success due to high cost, difficulty of application and worries over chemical and ecological impacts to the marine environment.

Scientists have tried to develop predictive models for red tide by computation of dinoflagellate population size, water mass dimension and turbulence, meteorological conditions and specific nutrient requirements of the causative organisms. Wong et al (1997) developed a computational model for monitoring water quality and red tide in Tolo Harbour with the use of multiquadric method. The two-dimensional depth-integrated water quality model and the eutrophicated model commonly applied in marine environment were integrated by this method to simulate the spatial and temporal variations of nutrients in Tolo Harbour. As revealed, the multiquadric scheme was effective in elimination of the unexpected negative predictions of pollutants in the boundary regions and exhibits satisfactory results in simulating the occurrence of red tides. However, with regards to making accurate prediction of red tide, similar to other computational models, the model by Wong et al was generally constrained by the lack of information on the autecology of causative organisms, insufficient knowledge on the locality characteristics and the limitations on the labour-intensive monitoring programmes.

Some oceanographic researchers found that polar orbiting satellites were able to track the blooms of red tide. For example, in the Chesapeake Bay of USA a cooperative programme between National Oceanic and Atmospheric Administration (NOAA) and the Woods Hole Oceanographic Institute have shown that sea surface temperature (SST) imagery to be a useful tool in studies of red tides and the onset of PSP in the southern Gulf of Maine (NOAA Coastwatch web site, May 1998). Remotely sensed SST is claimed to have promising value as a tool to provide early warning of the conditions conductive to algal blooms but has yet been explored in Hong Kong and the South China coast.

Hodgkiss (1995) advocated reductions in nutrient loading via sewage discharges, etc., as an immediate step that could be taken to reduce the frequency and intensity of red tides. He based this suggestion on the results from Tolo Harbour which clearly showed an increase in red tides associated with increased organic loads from 1971 to 1988 and a decrease in red tides associated with decreasing organic loads from 1989 to 1993. While there is no better solution in reducing the occurrence of red tide, pollution control, particularly the controls on discharge of inorganic and phosphates, seems to be the measure of the most urgency to be conducted in the coming years.