Guidelines for Canadian Recreational Water Quality

3. Microbiological Characteristics

Recreational waters may be contaminated by direct human contact and by waterborne pollutants from external sources (e.g., sewage, storm water and agricultural runoff). Many epidemiological studies have identified gastrointestinal and upper respiratory illnesses in bathers that were a result of such contamination. The indicator organisms recommended in these guidelines are surrogates for the presence of pathogenic organisms that may cause gastrointestinal illnesses.

The ideal indicator of fecal contamination of recreational waters would be one of the enteric pathogens, such as Salmonella or Norwalk virus, most frequently responsible for waterborne diseases. However, because these are usually present at low levels and are irregularly distributed, even during disease outbreaks, they are difficult to isolate and quantify. Moreover, the absence of one pathogen does not necessarily ensure that other enteric pathogens are also absent. In addition, testing for every possible waterborne disease-causing microorganism would be prohibitively expensive in terms of both time and money. For these reasons, it is common practice to monitor the other more plentiful but non-pathogenic bacteria present in human and animal feces. The presence of elevated numbers of these bacteria in the aquatic environment is indicative of fecal contamination and the possible presence of enteric pathogens.

The best indicators of the presence of enteric pathogens in fecal pollution sources should have the following properties (National Academy of Sciences 1977; Cabelli et al. 1983; Elliot and Colwell 1985):

• present in fecal-contaminated waters when enteric pathogens are present but in greater numbers
• incapable of growth in the aquatic environment but capable of surviving longer than pathogens
• equally or more resistant to disinfection than pathogens
• easily and accurately enumerated
• applicable to all types of natural recreational waters (e.g., fresh, estuarine, and marine)
• absent from non-polluted waters and exclusively associated with animal and human fecal wastes
• density of indicator should be directly correlated with the degree of fecal contamination
• density of indicator should be quantitatively related to swimming-associated illnesses.

In the past, the most widely used indicator of recreational water quality was total coliforms. However, because this group does not conform to most of the above characteristics, it is now considered unsuitable. For example, many of the genera in this group, such as Klebsiella, Citrobacter, Enterobacter, and Aeromonas, are not unique to human or animal feces but are commonly present in unpolluted surface waters (Boyd and Boyd 1962; Goodrich et al. 1970). Escherichia coli, enterococci, and, to a lesser degree, fecal coliforms are currently considered the best fecal indicators, because they most closely fit the above characteristics. Maximum acceptable concentrations of these indicator organisms are provided below.

The microbiological parameters selected for review are grouped as either fecal indicators (Sections 3.1 and 3.2), coliphages (Section 3.3) and pathogenic organisms (Section 3.4). The list is not intended to be comprehensive; some control agencies may wish to monitor additional parameters to consider regional interests.

3.1 Indicator Organisms for Fresh Waters

3.1.1 Escherichia coli and fecal coliforms

Maximum Limits

The geometric mean of at least 5 samples, taken during a period not to exceed 30 days, should not exceed 2000 E. coli/L. Resampling should be performed when any sample exceeds 4000 E. coli/L. The Working Group agreed that the tests to be used to fulfil these requirements are as follows:

1. When experience has shown that greater than 90 per cent of the fecal coliforms are E. coli, either fecal coliform or E. coli may be determined.
2. When less than 90 per cent of the fecal coliforms are E. coli, only E. coli may be determined.

Criteria

Description

For several years, recreational water quality experts in Canada have recognized E. coli as the indicator of choice for fecal contamination (Department of National Health and Welfare 1983). However, because its enumeration required complicated, time-consuming, and expensive techniques, total coliforms and, more recently, fecal coliforms were established as the indicators of fecal contamination. Fecal coliforms have all the properties of total coliforms, plus they are able to ferment lactose with the production of gas in 24 hours at an incubation temperature of 44.5°C. In practice, fecal coliforms are enumerated either by the most probable number (MPN) multiple tube fermentation test or by the membrane filtration (MF) method (American Public Health Association 1989) (see Chapter 5.0). The organisms most often recovered by these methods are E. coli and Klebsiella spp.

The development of the mTEC method specifically designed to enumerate E. coli (Shaw and Cabelli 1980; Dufour et al. 1981) has prompted a reevaluation of its use for monitoring recreational water quality. In recent tests, the mTEC method has proved to be more effective for E. coli isolation than the fecal coliform test (mFC method) for both freshwater samples (Pagel et al. 1982) and marine water samples, even when resuscitation steps were used with the mFC method (Rippey et al. 1987). The U.S. Environmental Protection Agency (1986) recently proposed the use of E. coli as an indicator for monitoring fecal contamination of fresh recreational waters. The incorporation of methylumbelliferone glucuronide (MUG) into various fecal coliform media has also been examined recently (Feng and Hartman 1982; Freier and Hartman 1987; Mates and Schaffer 1988). The enzyme b-glucuronidase, unique to E. coli and some strains of Salmonella and Shigella, will metabolize MUG to produce 4-methylumbelliferone, which fluoresces under longwave ultraviolet light.

Escherichia coli comprises about 97 per cent of the coliform organisms in human feces, with Klebsiella spp. comprising 1.5 per cent and Enterobacter and Citrobacter spp. together comprising another 1.7 per cent (Dufour 1977). Studies by Geldreich (1970) demonstrated that fecal coliforms represented 93 to 99 per cent of the coliforms in feces from humans, poultry, cats, dogs, and rodents. Dufour (1977) demonstrated that E. coli represented between 90 and 100 per cent of all coliforms in feces from 8 species of domestic animals, including chickens. Klebsiella spp. constituted a significant proportion of the coliforms only in feces from goats (8 per cent of the total coliforms) and pigs (6.8 per cent).

The fecal coliform and E. coli tests do not differentiate between animal and human fecal pollution (Wolf 1972; Dufour 1977). If this distinction is necessary, ancillary tests such as the speciation of fecal streptococcal isolates must be performed.

Soil has been shown to be variably contaminated with fecal coliforms due to animal fecal contamination (Geldreich et al. 1962; Van Donsel et al. 1967) and is known to contribute very significant pollution to storm water runoff (Environment Canada/Ontario Ministry of the Environment 1978; Qureshi and Dutka 1979) and recreational waters (Bastein et al. 1974; Cabelli 1977). Runoff from residential areas is usually as contaminated with fecal coliforms and pathogens as dilute sewage (Qureshi and Dutka 1979) and can thus be a significant health hazard when discharged in the vicinity of recreational waters. To counter this problem, some Canadian cities with urban beaches have adopted beach closure policies based on rainfall data. For example, beaches along the Rideau River in Ottawa are closed for 24 and 48 hours following a rainfall of greater than 10 and 20 mm, respectively, in the preceding 24 hours (Corber 1988). This approach has the advantage of posting beaches when the health risk to bathers is highest.

Although E. coli is undisputed as the fecal indicator of choice, some of the fecal coliform tests used will also enumerate Klebsiella spp., which are not restricted to fecal sources. Numerous studies have demonstrated the ability of Klebsiella spp. to survive and replicate in organic-rich environments, including waters receiving effluents from pulp and paper mills (Huntley et al. 1976; Rokosh et al. 1977; Bell et al. 1978) and textile industries (Dufour and Cabelli 1976; Vlassoff 1977). In the past, this disadvantage was somewhat tempered by the fact that K. pneumoniae was considered pathogenic for debilitated hosts and that environmental isolates of this species were virtually indistinguishable from clinical isolates (Matsen et al. 1974; Dufour and Cabelli 1976). However, a recent review of the Klebsiella literature by Duncan (1988) has convincingly demonstrated that there is no evidence to suggest that any community infections have resulted from exposure to Klebsiella spp. in the natural environment. The concerns that environmental strains of Klebsiella pose a health hazard seem to be based on hospital situations that cannot be applied to the community setting. In addition, Klebsiella may not represent a significant proportion of fecal coliforms in most Canadian recreational waters. Studies in Ontario by Vlassoff (1981) on more than 7700 water samples indicated that 91.4 per cent of fecal coliform isolates were identified as E. coli. However, this value could be too high in recreational waters receiving pulp and paper mill effluents. A study by Caplenas and Kanarek (1984) demonstrated that effluents from pulp and paper mills in northern Wisconsin contained up to 90 per cent non-fecal K. pneumoniae when fecal coliforms were enumerated by the mFC medium. In 1985, analyses of beach samples impacted by pulp and paper mill effluents in the vicinity of Thunder Bay revealed that most of the fecal coliforms, in some cases 100 per cent, were Klebsiella (Young 1989). In 1988, an examination of 162 samples from 19 beaches in the vicinity of St. Catharines that were impacted by pulp and paper mill effluents indicated that only 37 per cent of fecal coliforms were E. coli (Brodsky 1989).

Several jurisdictions and agencies have promulgated regulations or suggested limits for fecal coliforms. Many agencies, in addition to specifying a limit on the geometric mean, also specify that no more than a certain percentage (usually 10 per cent) should exceed a specific limit (usually twice the specified mean). Provincial authorities may wish to include this additional condition in their guidelines or regulations. It has been pointed out that fecal coliforms are usually sufficiently dispersed so that this percentage, rather than the established mean, is the factor most often exceeded (Fuhs 1975).

Association with Pathogens

Fecal coliforms are considered useful as indicators because they are present in virtually all warm-blooded animals, including humans, in numbers far exceeding the numbers of pathogens. Some attempts have been made to quantify the relationship of fecal coliforms to pathogens, with limited success. The pathogen most extensively studied in relation to indicator densities is Salmonella, mainly because the methodology has been available for a long time. Geldreich (1970) compiled the results of several studies in which the fecal coliform density per 100 mL and the frequency of Salmonella detection were compared. In fresh water, salmonellae were found in 27.6 per cent of samples where the fecal coliform density was below 200/100 mL, in 85.2 per cent of samples where the fecal coliform density was between 201 and 2000/100 mL, and in 98.1 per cent of samples where it was over 2000/100 mL. In estuarine waters, the results were not as definitive. When fecal coliform densities were less than 200/100 mL, the probability of finding Salmonella was 28.4 per cent; however, if fecal coliform densities were greater than 2000/100 mL, the probability of finding Salmonella was only 60 per cent.

A recent study has supported these results. Menon (1985), in an investigation of a tidal river in Nova Scotia receiving municipal and food processing effluents, reported that Salmonella spp. were always detected when fecal coliform levels were greater than 2000/100 mL and were occasionally detected when fecal coliform levels were greater than 200/100 mL. Conversely, Payment et al. (1982) found no relationship between the presence of Salmonella (most isolates were S. typhimurium) and other bacterial indicators, including fecal coliforms, at four freshwater bathing beaches in Quebec.

In general, samples containing high concentrations of fecal coliforms will likely also contain Salmonella, but the absence of fecal coliforms does not necessarily indicate that Salmonella or other pathogens will be absent. This relationship is also subject to considerable regional variation. Other comparative studies have been conducted with Pseudomonas aeruginosa (Cabelli et al. 1976; Sherry 1986), Vibrio parahaemolyticus (Robertson and Tobin 1983; Larsen and Willeberg 1984), Candida albicans (Sherry 1986), Aeromonas hydrophila (Seidler et al. 1980; Larsen and Willeberg 1984), and Campylobacter jejuni (Hill and Grimes 1984; Carter et al. 1987).

Enteric viruses have received considerable notoriety as a hazard associated with the use of recreational water; unfortunately, the relative incidence of virus and fecal coliforms has not been constant. In one sewage treatment plant studied extensively, ratios of fecal coliform to viruses ranged from 7500:1 to 2 900 000:1 over a 2-month period (Berg and Metcalf 1978). Sattar (1978a) found a large variation among the fecal coliform to enterovirus ratios from a variety of sources: raw sewage (between 1.1 × 10⁶ and 43 × 10⁶), chlorinated effluent (between 8.5 × 10² and 2.1 × 10⁴), and 2 beaches on the Ottawa River (between 1.2 × 10³ and 1.2 × 10⁶). Analysis of the data gathered from freshwater beaches in Quebec by Payment et al. (1982) did not reveal any correlations between the presence of enteric viruses and enteric bacteria, including fecal coliforms. It is generally agreed at this time that most bacterio-logical indicators do not correlate well with virus levels, even though the presence of large numbers of coliforms may indicate the probable presence of enteric viruses. However, the converse - that the absence of fecal coliforms indicates that enteric viruses are not present - cannot be ensured (Berg and Metcalf 1978).

Related Epidemiological Studies

In order to provide rational microbiological guidelines for recreational water, it is necessary to establish that there is some degree of health risk associated with a certain level of contamination. Again, because pathogens are sparse and difficult to quantify, fecal coliforms, including E. coli, have been used in all major epidemiological studies.

The first widely publicized epidemiological studies in North America were carried out by the U.S. Public Health Service (Stevenson 1953). These studies were conducted at 2 freshwater sites on Lake Michigan and the Ohio River and at 2 marine sites. Swimming-associated gastroenteritis was not observed at the marine sites or at the Lake Michigan site. However, the Ohio River study showed increased gastro-intestinal illness at median coliform densities of about 2300/100 mL. Data subsequently collected from the Ohio River site during the 1960s suggested that a level of 400 fecal coliforms was approximately equivalent to the threshold number of 2300 total coliforms (Geldreich 1966). Following application of a safety factor, the guideline of 200 fecal coliforms/100 mL was developed. Although the design of Stevenson's studies has been severely criticized, it is apparent that these studies have formed the basis of most guidelines in use today.

Since then, other microbiological-epidemiological studies have been conducted for recreational water based on fecal coliform concentrations. Seyfried et al. (1985a, 1985b) conducted an investigation on 10 beaches in Ontario. In total, 8402 swimmers and non-swimmers were interviewed. Crude morbidity rates for all illnesses, including respiratory, gastrointestinal, eye and ear, skin, and allergy illnesses, were roughly 2.4 times greater for swimmers than for non-swimmers, and 2.6 times greater for swimmers who immersed their heads than for those that did not. Morbidity among swimmers was shown to be related to fecal coliform counts (r = 0.284), even though the mean fecal coliform densities in the water (76/100 mL) were within the accepted guidelines. The total staphylococcal count was also related to morbidity among swimmers (r = 0.439). Levels of all groups of bacteria surveyed were at least 10 times greater in the sediment than in the surface waters, indicating that the resuspension of sediment may be a significant source of bacteria in recreational waters. The authors concluded that sediment samples should be analyzed during microbiological investigations.

During the summer of 1983, a microbiological-epidemiological study of three coastal beaches in Israel was conducted by Fattal et al. (1986). Although all beaches complied with Israel Ministry of Health bacteriological standards, analysis of the results indicated that the incidence of gastroenteritis among swimmers, particularly in the 0- to 4-year age group, was related to elevated densities of E. coli as well as enterococci and staphylococci. The incidence of ear infections in swimmers of all ages was also related to elevated densities of fecal coliforms, E. coli, and enterococci.

In the United States, a series of controlled epidemiological studies has been performed by the Environmental Protection Agency at both marine (Cabelli 1983) and freshwater (Dufour 1984) beaches. A summary of the studies and a presentation of proposed criteria for recreational waters were recently published (U.S. Environmental Protection Agency 1986). Two beaches were selected at each site, one relatively unpolluted and the other impacted with point or non-point sources of fecal contamination. In the freshwater studies, analyses were conducted for fecal coliforms, E. coli, and enterococci; in the saltwater studies, analyses were conducted for these indicators as well as several other waterborne pathogens. Regression coefficients were determined for levels of each of the parameters and the differential (swimmers minus non- swimmers) gastrointestinal symptom rates. On the basis of this pooled information, it was concluded that, for marine beaches, enterococci presented the best relationship with gastrointestinal symptoms (r = 0.75). At freshwater beaches, the best correlations were obtained with E. coli (r = 0.80) and enterococci (r = 0.74). Based on these investigations, the U.S. Environmental Protection Agency proposed that at freshwater beaches, the 30-day geometric mean should not exceed 126 E. coli/100 mL or 33 enterococci/100 mL.

From the pooled data, it can be calculated that in fresh waters, the seasonal risk of gastrointestinal illness per 1000 bathers (y) is related to the density of E. coli/100 mL (x) by the following relationship:

This investigation has been one of the most extensive microbiologicalepidemiological studies ever conducted, but it does have certain limitations. Although E. coli and enterococci are suitable for indicating the risk of gastroenteritis, this illness represents only about 30 per cent of overall swimmer-associated morbidity. Escherichia coli and enterococci are not suitable indicators for upper respiratory or dermal infections caused by Pseudomonas and Staphylococcus spp. Furthermore, the epidemiological information collected on the swimmers and non-swimmers and data on the water quality were averaged over an entire bathing season. Thus, the bacteriological data cannot be used to assess the risk of gastroenteritis on any specific day.

Other epidemiological studies have also noted elevated levels of fecal coliforms, including E. coli, associated with illness. An investigation by Philipp et al. (1985) demonstrated that snorkel swimming in fecally contaminated water presents a significant risk to health. In this British study, 27 per cent of the swimmers who participated in a competition experienced gastrointestinal symptoms within 48 hours of entering the water. The incidence of these symptoms was significantly greater than their incidence in the control populations. Escherichia coli concentrations in water samples collected during the event averaged 1800/100 mL. In light of the fact that this value still conformed with the European Economic Community guidelines, the authors suggested that an appraisal of these guidelines was warranted.

Dewailly et al. (1986) conducted an epidemiological study demonstrating certain health risks associated with windsurfing on water contaminated with sewage. Seventy-nine windsurfers and 41 controls were monitored over a 9-day competition for the occurrence of gastroenteritis, otitis, conjunctivitis, and skin infections. Relative risks were 5.5 for gastroenteritis and 2.9 for one or more of the above symptoms. This study also demonstrated that the relative risk increased with the reported number of falls in the water. Based on hydrodynamic simulation using tide levels and actual fecal coliform counts, mean densities of fecal coliforms at high tides during the competition were estimated to be 1000/100 mL.

A recent exercise in the United Kingdom (Brown et al. 1987) at two seaside resorts, although not a controlled epidemiological study, demonstrated that swimmers at resort A were more likely to develop stomach upsets, nausea, diarrhea, or headaches than either non-swimmers at the same resort or all vacationers at resort B. Furthermore, swimmers who had immersed their heads at resort A were most likely of all respondents to have reported gastrointestinal symptoms. The resorts investigated were chosen on the basis of existing microbio-logical data that indicated that fecal coliform concentrations at resort A were 440/100 mL and at resort B were only 10/100 mL.

Although all of the above epidemiological studies were comprehensive and detailed, they used different methods and produced data that are not reproducible or comparable.

Occurrence in the Aquatic Environment

Although most Canadian recreational waters are of high microbiological quality, certain waters are contaminated throughout part, or all, of the bathing season. Fecal coliform values range from nearly 0/100 mL in isolated areas to several thousand per 100 mL in areas directly impacted by sewage discharges (Payment et al. 1982; Ontario Ministry of the Environment 1984; Williamson 1988; Smith 1988). In temperate recreational waters, 63 to 100 per cent of the fecal coliforms appear to be E. coli, but this range can be affected by contamination from industrial effluents, particularly pulp and paper and textile mills.

Escherichia coli has been enumerated at a number of recreational beaches in Manitoba and Ontario (Sekla et al. 1987; Brodsky 1989; Palmateer 1989; Young 1989).

Summary

1. In fresh waters, E. coli is the best available indicator of fecal contamination from warm-blooded animals.
2. Klebsiella is not a good indicator of fecal contamination, but it may be present at high levels in certain industrial wastes (e.g., pulp and paper mills, food processing plants). It is not likely to cause infection or illness in healthy individuals.
3. Where there is evidence that greater than 90 per cent of the fecal coliforms are E. coli, the E. coli and fecal coliform tests will be considered equivalent.
4. The presence of E. coli is associated with bather-associated illness, but its absence cannot be equated with the lack of risk of illness.
5. Current microbiological-epidemiological studies are not sufficiently validated to allow calculation of risk levels. However, there is some evidence for increased risk of illness from bathing compared with non-bathing (i.e., wading or remaining on the beach).
6. The 1983 guidelines were, in principle, based on the definitive fecal coliform, E. coli. However, at that time, the more general fecal coliform test was considered the method of choice. The Working Group reaffirms that E. coli is the indicator of choice and recognizes that either the E. coli test or the fecal coliform test may be used to enumerate this organism, depending on the circumstances. The maximum acceptable concentration of 2000 E. coli (or fecal coliforms)/L can be calculated to correspond to a seasonal gastrointestinal illness of 1 to 2 per cent, based on the U.S. Environmental Protection Agency studies.

3.2 Indicator Organisms for Marine Waters

3.2.1 Enterococci

Maximum Limits

The geometric mean of at least five samples, taken during a period not to exceed 30 days, should not exceed 350 enterococci/L. Resampling should be performed when any sample exceeds 700 enterococci/L. However, if it can be demonstrated that E. coli or fecal coliforms can adequately demonstrate the presence of fecal contamination in marine waters, then the E. coli or fecal coliform maximum limit for fresh waters may be used. If there is any doubt, samples should be examined for both sets of indicators for extended periods to determine if a positive relationship exists.

Criteria

Description

Enterococci are large, ovoid, Gram-positive bacteria that are generally present in chains. The term enterococci refers to those species of the fecal streptococcal group that conform to the biochemical characteristics of the Sherman criteria (Clausen et al. 1977). They grow at temperatures between 10 and 45°C, survive exposure to 60°C for at least 30 minutes, and grow at pH 9.6 and in 6.5 per cent NaCl. This subgroup, which includes Streptococcus faecium and S. faecalis, occurs in significant quantities in both human and animal feces. Streptococcus avium and S. gallinarium, found principally in bird feces, are also classified as enterococci in Bergey's Manual of Systematic Bacteriology (1986). In the previous guidelines, the entire fecal streptococcal group, both enterococcal and non-enterococcal species, was addressed. Because the non-enterococcal subgroup includes species that normally occur only in animal feces (e.g., S. bovis, S. equinis), the more specific enterococcal subgroup will be considered in these guidelines. However, the presence of enterococci unique to animal feces may also indicate the presence of pathogenic microorganisms infectious to both humans and animals.

In the past, the main role of the fecal streptococci was in the use of the fecal coliform to fecal Streptococcus ratio as an indicator of the nature of the fecal source (Geldreich 1976; Clausen et al. 1977). However, many factors, for example, the differential die-off rates between these two groups in the natural environment, make the routine use of this ratio highly questionable, if not inaccurate. Recent experience indicates that the identification of enterococcal isolates is more useful in the determination of the type, source, and degree of fecal contamination (Rutkowski and Sjogren 1987).

Of all the microorganisms considered as suitable recreational water quality indicators, the enterococci most closely satisfy the desirable characteristics presented in the introduction to this chapter. Enterococci are exclusively associated with fecal wastes. They survive much longer than the other indicators in water and sediment (McFeters et al. 1974; Lessard and Sieburth 1983). Enterococci are also more resistant to sewage treatment, including chlorination, and thus may be more sensitive indicators of the survival of enteric pathogens and viruses (Cohen and Shuval 1973). As well, a strong correlation between the concentration of enterococci in marine waters and the risk of gastrointestinal infection has been demonstrated (Cabelli 1983). A membrane filtration method for the enumeration of enterococci in marine waters has recently been described in detail (U.S. Environmental Protection Agency 1985).

Epidemiological Studies

During the summer of 1983, a microbiological-epidemiological examination of 3 coastal beaches in Israel was conducted by Fattal et al. (1986)a. Although all beaches complied with Israel Ministry of Health bacteriological standards, analysis of the results indicated that the incidence of gastroenteritis among swimmers, particularly in the 0- to 4-year age group, was related to elevated densities of indicator organisms, most notably the enterococci (p <0.03).

A series of prospective epidemiological studies has also been performed in the United States at 3 marine and 2 freshwater sites (U.S. Environmental Protection Agency 1986). Two beaches were selected at each site, 1 relatively unpolluted and the other receiving point or non-point fecal contamination. Regression coefficients were determined for levels of each of the indicators measured and the differential (swimmers minus non-swimmers) gastrointestinal symptom rates. On the basis of results pooled for the entire season, enterococci displayed the best correlation with these symptoms at marine beaches (r = 0.75). At freshwater beaches, the best correlations were obtained with E. coli (r = 0.80) and enterococci (r = 0.74).

From the pooled data, it can be calculated that in marine waters, the seasonal risk of gastrointestinal illness per 1000 swimmers (y) is related to the densities of enterococci per 100 mL (x) by the following relationship:

Based on these investigations, the U.S. Environmental Protection Agency proposed that at marine beaches, the 30-day enterococci geometric mean should not exceed 35/100 mL.

This investigation has probably been one of the most extensive epidemiological studies conducted to date on recreational beaches, but it does have certain limitations. In developing the model, it was assumed that the non-swimming and swimming populations were equally susceptible to illness acquired from other sources. Although enterococci appear suitable for estimating the risk of gastrointestinal illness, they are not suitable indicators for the more prevalent upper respiratory or dermal infections. Furthermore, the data collected on both the participants and the water quality were averaged over the entire bathing season. Thus, the indicator data cannot be used to assess the risk of gastroenteritis on any specific day.

Occurrence in the Aquatic Environment

There have been a few published investigations in Canada on the distribution of fecal streptococci and enterococci in the marine environment. In a continuing study of the use of enterococci as a suitable indicator of water quality at marine recreational beaches, Gibson and Smith (1988) described the distribution of enterococci at 26 beaches in the Vancouver region. Only 1.6 per cent of the 30-day geometric means exceeded the 35 enterococci/100 mL limit proposed by the U.S. Environmental Protection Agency. In 1988, fecal streptococcal concentrations were also monitored at eight marine beaches along the Northumberland Strait, New Brunswick (Allen 1989). The overall geometric mean was only 3.5/100 mL, and fecal streptococci were absent in 60 per cent of the samples.

Summary

1. In marine waters, the enterococci group is the best available indicator of fecal contamination from warm-blooded animals.
2. Fecal coliforms do not survive well in marine waters and thus may not be reliable indicators of fecal contamination.
3. Enterococci survive longer than fecal coliforms in marine waters and thus are preferred when there is considerable time or distance between the source of fecal pollution and the bathing area.
4. There is a positive correlation between gastrointestinal illness and levels of enterococci in marine waters, but the absence of enterococci does not indicate a lack of risk.
5. Based on the U.S. Environmental Protection Agency epidemiological study, a seasonal geometric mean of 35 enterococci/100 mL corresponds to a seasonal gastrointestinal illness rate of 1 to 2 per cent. Because fecal coliforms do not survive well in marine waters, the use of the fresh water maximum limit may increase the risk of illness.

3.3 Coliphages

Maximum Limits

No limits are specified for coliphages in recreational waters.

Criteria

Description

Bacteriophages are virus-like entities that invade bacterial cells. Coliphage is the general name applied to bacteriophages that attack bacteria of the coliform group. Bradley (1967) described 6 morphological types of coliphages and suggested that the diversity among coliphages was greater than the known diversity among mammalian viruses. They differ in size, shape, site of attachment, and genetic material (e.g., single-stranded DNA, double-stranded DNA). The numbers of phages replicated may vary from 100 to several thousand.

Association with Pathogens

An excellent review of the literature on bacteriophages as indicators of bacterial and viral contamination has been published by Scarpino (1975). In the review, Scarpino reported that "correlations appear to exist in fresh and marine waters between bacterial pathogens such as Salmonella and Shigella species and fecal indication bacteria such as E. coli and their bacteriophages." Studies by Kott et al. (1974, 1978), Petrovicova et al. (1988), Dutka et al. (1987), and Borrego et al. (1987b) have also indicated a correlation between the presence of coliphage and other bacteriophages and the presence of viruses and pathogenic bacteria in rivers, lakes, and sewage effluents.

Occurrence in the Aquatic Environment

Recent research studies on four continents have found the presence of coliphages and bacteriophages in natural river waters, potable water, and coliform-free potable waters. Implications of these observations are that human enteric viruses could also survive in those waters. A Czechoslovakian study (Simkova and Cervenka 1981) has also shown that both coliphages and enteroviruses can survive for similar long periods of time in river water, and thus coliphages can be used as long-term indicators of enterovirus contamination, even in the presence of chemical pollution. The view of Russian researchers (Petrovicova et al. 1988) is that "an increase of coliphage numbers in sewage and surface and recreational (swimming pool) waters shall be considered as an indication for aimed virological studies."

It may be useful to include coliphage and bacteriophage enumeration as part of the continuing evaluation of the impact of fecal pollution on recreational waters. Dutka et al. (1987) suggested that coliphage levels in recreational fresh waters should not exceed 20 PFU (plaque-forming units)/100 mL.

Summary

1. No limit on coliphages can be established at this time. Monitoring and epidemiological studies are required to determine the levels of coliphages in water and the health effects associated with swimming in water containing coliphages.

3.4 Pathogenic Organisms

3.4.1 Pseudomonas aeruginosa

Maximum Limits

No numerical limit is proposed; however, it is recommended that Pseudomonas aeruginosa be used as a parameter to assist in interpreting the results of sanitary and microbiological surveys.

Criteria

Description

Pseudomonas aeruginosa is a motile, Gram-negative, rod-shaped bacterium that produces oxidase, pyocyanin, and fluorescein. This organism typically requires minimal growth factors and can multiply in mineral media containing very low levels of organic material. Pseudomonas aeruginosa exhibits extensive biochemical versatility and resistance to anti-microbial agents. As well as being a pathogen to man and animals and causing a variety of infections, including skin rashes and otitis externa, P. aeruginosa is active as a spoilage organism, attacking many common and exotic substrates (Hoadley 1977).

Pathogenicity

Pseudomonas aeruginosa has been identified as the causative agent in many infections. The organism seems capable of infecting plants, insects, birds, and mammals including humans. Pseudomonas aeruginosa infections are most frequent and dangerous in nurseries and among hospital patients with cancer, burns, and tracheotomies (Hoadley 1977); the organism is a common causative agent for nosocomial or hospital-acquired infections in patients that are in a debilitated state or have a compromised immune system, as well as where there is widespread use of antibiotics.

Pseudomonas aeruginosa is known to cause skin rashes (Kush and Hoadley 1980) and eye infections (Wilson and Ahearn 1977) and is the primary organism associated with external ear infections (otitis externa) (Cassisi et al. 1977). Ratnam et al. (1986) reported that P. aeruginosa in a hotel whirlpool had caused folliculitis in hotel guests that used the whirlpool. Cabelli et al. (1979) and McKee and Wolfe (1963) claimed that most reports of swimming-associated illness implicated non-enteric infections; P. aeruginosa has been implicated in non-enteric infections associated with bathing. Jones (1965) claimed that P. aeruginosa was the major etiological agent in otitis externa. Seyfried (1973) and Hoadley and Knight (1975) supported this finding. Seyfried isolated an identical serological strain from both a pool and an otitis externa-infected person who was using the pool. Hoadley and Knight, using telephone surveys, reported that the frequency of "earaches" among swimmers was 2.4 times higher than among non-swimmers; also, the chance of acquiring otitis externa, as reported by physicians, increased by five times for swimmers compared with the general population. Young and Armstrong (1972) reported the isolation of P. aeruginosa from skin and the outer ear. Craun (1976) reported the isolation of the same serotype of P. aeruginosa from a swimming pool and from the skin lesions of two affected bathers. McClausland and Cox (1975) reported that P. aeruginosa caused a rash outbreak among bathers at a pool; however, it was never proven whether the source of the organism was endogenous or exogenous (Cabelli et al. 1975).

Although Hoadley and Knight (1975) observed measurable health effects associated with swimming in sewage-polluted waters, Cabelli et al. (1979) were not able to show a good agreement between P. aeruginosa levels and the differential (swimmers minus non-swimmers) rate of gastrointestinal symptoms.

Pseudomonas aeruginosa in the water enters the ear canal and may lead to either colonization or infection. The actual process of infection appears to be, in part, related to the sensitivity of the individual to P. aeruginosa infections. Pseudomonas aeruginosa can be spread into the water if an infected ear is immersed while swimming (Seyfried et al. 1984). Data from a 4-year comprehensive Ontario study were used to develop a relationship between the concentration of P. aeruginosa in the bathing waters and the risk of ear infection (Ontario Ministry of the Environment 1984). In this study, 10 per cent of the sample bather population reported having ear infections in the first year of the study (1978), and 8 per cent reported ear infections in the second year. In both years, over 70 per cent of the cases were in children 14 years old or younger, and 90 per cent of these had reported previous ear infections.

Occurrence in the Aquatic Environment

Pseudomonas aeruginosa has been considered ubiquitous in U.S. waters. Cabelli et al. (1976) claimed that P. aeruginosa, "because of its poor reliability as an indicator of fecal pollution, cannot be used as a basis of water standards for the prevention of enteric diseases during the recreational use of surface waters." They based this statement on their observations of geographical and seasonal density fluctuations and of many non-fecal reservoirs of P. aeruginosa, as well as the ability of this organism to multiply in waters with low nutrient content.

Studies done on lakes in Ontario indicated that P. aeruginosa is most likely to be found in bathing areas impacted by high human activity. Both sewage and bathers are possible sources of this organism in recreational waters. In raw domestic sewage, concentrations of 105 to 106 P. aeruginosa/100 mL are common, as slightly in excess of 10 per cent of the healthy adults in the United States are intestinal carriers of P. aeruginosa. In addition, Pseudo-monas levels in excess of 100 organisms/100 mL can be measured in waters receiving surface drainage from urban areas (Ontario Ministry of the Environment 1984).

Pseudomonas aeruginosa survives longer in waters than do coliforms (Lanyi et al. 1966). Drake (1966) suggested that P. aeruginosa levels from 1 to 10/100 mL could be expected in rivers with low but definite sources of contamination. Levels of P. aeruginosa in Ontario recreational waters range from 0/100 mL to more than 100/100 mL. The median level is typically less than 1/100 mL (Ontario Ministry of the Environment 1984).

Summary

1. Pseudomonas aeruginosa is typically isolated from fresh recreational waters in low numbers. The levels of P. aeruginosa in a bathing area are influenced by density of bathers, especially individuals that are infected with P. aeruginosa or are carriers.
2. Levels of P. aeruginosa are influenced by sewage or urban drainage sources.
3. Pseudomonas aeruginosa has been associated with the occurrence of otitis externa in bathers.
4. One Ontario study has demonstrated that when levels of P. aeruginosa exceed 10/100 mL in at least 25 per cent of the seasonal samples, otitis externa may be expected to occur.

3.4.2 Staphylococcus aureus

Maximum Limits

No limits are specified for Staphylococcus aureus. Sampling for this pathogen should be carried out when there is epidemiological or other evidence of its presence in the water or in order to assess the hazards of excessive utilization of the water with possible person-to-person transfer of pathogens.

Criteria

Description

The Staphylococcus genus comprises Gram-positive, catalase-positive cocci that ferment glucose and grow aerobically and anaerobically. Staphylococcus aureus is identified by its ability to coagulate rabbit plasma in the presence of anti-coagulating agents. Other important diagnostic tests include the anaerobic fermentation of mannitol and the production of heat-stable nuclease (Evans 1977). Six biotypes exist that can be related to the animals that are their usual hosts (Dimitracopoulos et al. 1977).

Staphylococcus is not considered to be a natural inhabitant of environmental waters and is usually unable to grow there. It requires many organic nutrients to grow in water at about 20°C and grows little, if at all, below 10°C. It is, however, resistant to many environmental influences and can survive for relatively long periods of time. The presence of staphylococci in recreational waters is considered to be mainly due to discharges from the mouth, nose, and throat of swimmers, as well as from their skin surface. These organisms have been shown to be derived from bathers in studies on swimming pools (Mallman 1962; Favero et al. 1964; Paul 1972; Palmquist and Jankow 1973) and, to a lesser extent, natural waters (Oriz 1977). The 16th edition of Standard Methods for the Examination of Water and Wastewater (American Public Health Association 1989) identifies staphylococci as among the pathogens of human origin in natural bathing beaches. The extent and magnitude of the contribution of animals and land runoff to the total staphylococcal load have not been well documented.

Pathogenicity

Staphylococcus aureus is considered to be the major pathogen of this genus. It is responsible for most purulent infections, including boils and infected cuts and scratches (Evans 1977). Seyfried (1973) reported 1 case of ear infection associated with swimming in a contaminated pool. At present, there are no convincing data relating the frequency of illness to the degree of pollution of recreational waters.

In spite of the paucity of information that would allow an estimate of risk to be made for water of a particular quality, some maximum levels have been proposed. Limits suggested for safe bathing range from a "cocci" index of less than 15/100 mL (Seligmann 1951) to 100/100 mL (Favero et al. 1964). Seyfried (1987) found that total staphylococcus, fecal coliform, and fecal streptococci correlated best with swimming-associated morbidity. This was based on a 1983 study of freshwater beaches in southern Ontario and included an epidemiological component to the study. The study recommended that fecal coliforms, Escherichia coli, Pseudomonas aeruginosa, and total Staphylococcus be employed as recreational freshwater quality indicators.

Seyfried (1987) found geometric mean levels of total Staphylococcus of 142/100 mL for beaches in use and 96/100 mL for beaches that had been closed. The total staphylococci index has, in general, been favoured over a S. aureus index for purposes of health protection.

Occurrence in the Aquatic Environment

Limited data indicate that Staphylococcus is present at relatively low levels at Canadian beaches. One study revealed the presence of 10 to 40 staphylococci/100 mL in urban runoff (Environment Canada/Ontario Ministry of the Environment 1978). Studies performed by Seyfried (1973) demonstrated levels of between 30 and 90/100 mL on one of three Lake Erie beaches that were tested. Two of three Lake Ontario beaches were positive for S. aureus. Two of three Ontario conservation areas were positive for staphylococci, one species being coagulase-positive.

A more recent study (Seyfried 1980) of several Great Lakes beaches revealed the consistent presence of staphylococci. Of 12 Lake Ontario beaches studied, the modes ranged between 10 and 122/100 mL, with very few S. aureus being detected. In water from 10 Lake Erie beaches, modes ranged from 28 to 380/100 mL. In water from seven Lake Huron beaches, modes ranged from seven to 49/100 mL.

Summary

1. Staphylococcus aureus is known to be a major pathogen to man. It is responsible for boils, ear infections, and other purulent infections.
2. There appears to be a relationship between bather numbers and staphylococci levels in the water, but there does not appear to be a significant relationship between bather illness and concentration of S. aureus in the water.
3. For these reasons, no limit is being established for staphylococci at this time. Monitoring and epidemiological studies for this pathogen are recommended.

3.4.3 Salmonella

Maximum Limits

No limit is proposed for Salmonella concentrations in recreational waters. Instead, it is recommended that Salmonella be used as a parameter to assist in interpreting the results of sanitary and microbiological surveys. Because virtually all Salmonella species are pathogenic, a health hazard exists when Salmonella can be consistently isolated from a bathing area.

Criteria

Description

Salmonella species are members of the family Enterobacteriaceae and are Gram-negative, motile, straight, rod-shaped bacteria that ferment glucose but not lactose. Because of the antigenic complexity of the genus, organisms within this group are often identified by serotype as opposed to species. More than 1400 different Salmonella serotypes have been recognized (Bergey's Manual of Systematic Bacteriology 1986).

Pathogenicity

Salmonellosis is any disease in man or animal for which the causative agent is the Salmonella bacterium. The symptoms of this infection include acute gastroenteritis, enteric fever, and septicemia. Salmonellosis is a worldwide animal and human health problem that is compounded by the large number of serotypes and the widespread occurrence of the causative organism.

Dudley et al. (1976) developed a mathematical model to quantitatively estimate the recreational user health risk due to exposure to Salmonella in recreational waters. He concluded that more research was required to develop standardized quantitative Salmonella detection techniques for routine use with recreational waters before his model could be used effectively. In addition, Dutka and Bell (1973) suggested that all members of the genus are potentially pathogenic and should be considered a health hazard. There have been a number of reports of salmonellosis in communities where water supplies are considered to meet coliform standards (Dutka and Bell 1973). Van Donsel et al. (1967) indicated that S. typhimurium was still able to infect humans after surviving 280 days in contaminated soil.

Occurrence in the Aquatic Environment

Cherry et al. (1972) found that 65 per cent of the samples taken from moderately polluted waters contained Salmonella; 38 per cent of the samples taken from minimally polluted waters were positive for Salmonella; and 11 per cent of the samples from unpolluted streams were positive for Salmonella. A relationship between the presence of Salmonella and the levels of fecal coliforms in water has also been noted (Geldreich et al. 1968; Smith and Twedt 1971; Smith et al. 1973; Hoadley et al. 1974). It appears that Salmonella can be consistently isolated from surface waters where the fecal coliform levels are above 200/100 mL.

Menon (1985) isolated five Salmonella serotypes (S. infantis, S. typhimurium, S. saint-paul, S. tennessee, and S. heidelberg) from a tidal river in Nova Scotia that received both municipal and food processing plant effluents. The sources of Salmonella were traced to effluents from meat and poultry plants and several sewage treatment plants. Fecal coliform counts in the river also exceeded recommended recreational and shellfish harvesting guidelines. Palmateer (1980) also isolated Salmonella serotypes from surface water containing the effluent from poultry and meat packing plant operations. It was confirmed that four of the serotypes isolated (S. bareilly, S. infantis, S. schwarzengrund, and S. typhimurium) had been identified as the causative agents of several outbreaks of salmonellosis in Ontario. Salmonella thompson and S. typhimurium were also frequently recovered from a storm sewer outfall during periods of rainfall (Qureshi and Dutka 1979). Bell et al. (1978) isolated Salmonella species from the North Saskatchewan River that were resistant to five commonly used antibiotics. Qureshi (1977) isolated four Salmonella serotypes from urban storm water samples collected in Toronto and Guelph. These four serotypes are among those often isolated from humans during recent years.

Several studies have demonstrated that S. typhimurium can survive longer than Escherichia coli in fresh and estuarine waters (Gosselin 1979; McCam-bridge and McMeekin 1981). In Lake Ontario and Hamilton Harbour, S. thompson survived for at least 20 days (Dutka and Kwan 1980). Such survival data demonstrate that occasional discharges of storm water or sewage may have cumulative effects, so that natural temperate waters could contain pathogens and therefore constitute a potential health hazard.

Salmonella have been isolated in higher numbers from bottom sediments than from the overlying waters (Van Donsel and Geldreich 1971; Hendricks 1971). This was linked with prolonged survival of Salmonella in the sediments. Van Donsel and Geldreich (1971) suggested that Salmonella should be isolated from the sediments only when the fecal coliform levels reach 200/100 mL.

Summary

1. Salmonella organisms are pathogenic, and a health hazard exists if these organisms can be consistently isolated from a bathing area.
2. The methods for the isolation of Salmonella have not been standardized, and routine enumeration is not practical.
3. Salmonella can be considered as a support parameter to aid regulatory agencies in determining the health risk involved in using waters for recreation.

3.4.4 Shigella

Maximum Limits

No maximum limits are specified for Shigella in bathing areas. Sampling for these organisms in waters used for recreation should be carried out when there is epidemiological or other evidence of their presence in the water or in order to assess the hazards of excessive utilization of the water with possible person-to-person transfer of pathogens.

Criteria

Description

Members of the genus Shigella are typically Gram-negative, non-mo-tile, lactose-negative organisms that do not produce hydrogen sulphide. Shigellae are enteric bacilli and have been the main etiological agents of bacillary dysentery (Lund 1978).

Pathogenicity

Shigellosis, the infection caused by Shigella, can be transmitted through person-to-person contact, poor-quality drinking water, or contaminated food. The symptoms of shigellosis range from a mild transitory diarrhea to vomiting, abdominal pains, fever, and profuse bloody feces. On a few occasions, its transmission has been reported through bathing in contaminated waters. However, very few studies have been carried out to detect Shigella spp. in the aquatic environment, mainly because there are no standardized methods for detection or enumeration of Shigella in water. The methodologies that have been used to examine water for Shigella spp. are reported to have low sensitivity and can be considered qualitative only (Wang et al. 1966; Geldreich 1972).

McCabe and Craun (1975) reported that Shigella spp., especially S. sonnei, were the most commonly identified pathogens causing waterborne disease outbreaks during 1971 and 1972 in the United States and Canada. These outbreaks were associated with polluted drinking water. A severe outbreak of shigellosis associated with swimming in the section of the Mississippi River below Dubuque, Iowa, has been documented (Anonymous 1974; Rosenberg et al. 1976). As a result of the follow-up on this epidemic, Rosenberg et al. (1976) remarked that shigellosis - "a disease that can be caused by ingestion of only 10 to 100 organisms - can be contracted by swimming in polluted waters." This was the first epidemic of shigellosis that could be linked with bathing. Cabelli (1979), in reconstructing the epidemiological evidence from the shigellosis outbreak at Dubuque, concluded that, although the fecal coliform levels exceeded 200/100 mL for a period of several years, a critical level of Shigella carriers or ill individuals discharging into the environment had to occur before there would be an outbreak of shigellosis.

More recently, Makintubee et al. (1987) reported the occurrence of at least 62 cases (38 primary or co-primary and 24 secondary) of shigellosis (S. sonnei) associated with swimming in a natural reservoir. Although excessive concentrations of fecal indicator bacteria were present in the water, Shigella was not recovered.

Occurrence in the Aquatic Environment

Shigella can be isolated from the feces of warm-blooded animals and sewage. Lund (1978) suggested that the bacterium tends to survive for a relatively short period of time in the environment. Hendricks (1972), using autoclaved river water collected downstream from a sewage outfall, found that Shigella grows best at 30°C and reproduces at a rate 2 to 3 times slower than coliforms. He observed little or no growth at temperatures of 5 to 20°C. Wang et al. (1966) found that the organism survives longer in wastewater at 15°C than at 25°C. Andre et al. (1967) observed that Shigella persisted in untreated farm pond water for 12 days. Other survival characteristics, reviewed by Geldreich (1972), include an appreciable reduction in survival time by the use of aeration or by the reduction of the pH below 7.6.

Summary

1. Shigella organisms are pathogenic, and a health hazard exists if these organisms can be consistently isolated from a bathing area.
2. The methods for the isolation of Shigella have not been standardized, and routine enumeration is not practical.
3. Shigella can be considered as a support parameter to aid regulatory agencies in determining the health risk involved in using waters for recreation.

3.4.5 Aeromonas

Maximum Limits

No maximum limit has been proposed for Aeromonas in bathing areas. It is recommended that sampling for Aeromonas in waters used for recreation be considered only for epidemiological investigations.

Criteria

Description

Organisms belonging to the genus Aeromonas are facultatively anaerobic, Gram-negative rods possessing polar flagella. The temperature range for bacillus growth is between 0 and 41°C (Ewing et al. 1961). Aeromonas species are widely distributed in stagnant and flowing fresh waters, in sludge, and in sewage (Hazen et al. 1978). Bergey's Manual of Systematic Bacteriology (1986) and Popoff et al. (1981) recognize three species of Aeromonas found in clinical specimens: A. hydrophila, A. caviae, and A. sobria. However, Aeromonas can be divided into 9 to 12 DNA hybridization groups, and the taxonomy has yet to be clarified.

Pathogenicity

Human infections with Aeromonas occur predominantly during the period from May to November, probably because of the aquatic origin of the bacteria (Davis et al. 1978). Infections caused by the species have been divided into four categories (von Graevenitz 1985):

• cellulitis or wound infection related to exposure to water
• acute diarrheal disease of short duration
• septicemia, mostly in association with hepatic biliary or pancreatic disease
• other infections, such as soft-tissue infections, urinary tract infections, meningitis, peritonitis, otitis, and endocarditis, particularly in immuno-compromised people.

Aeromonas species cause an acute, self-limiting diarrheal illness in humans; this is supported by the finding of various exotoxins. One of them, enterotoxin, has been detected by infant mouse tests (Turnbull et al. 1984); it is a heat- and acid-labile molecule and causes cell lysis in tissue culture.

Carrier rates of 2 to 3 per cent for Aeromonas in feces have been observed in England, the United States, and Australia, with no associated gastrointestinal disease.

Occurrence in the Aquatic Environment

Aeromonas species can be isolated from the feces of warm-blooded animals, sewage, fresh waters, and salt waters that interface with fresh water. They have been found at pH values of 5.2 to 9.8 and at temperatures between 4 and 45°C. They are not considered halophilic, because their salt tolerance ranges from 0 to 4 per cent. They have also been isolated from soil and foodstuffs (Ewing et al. 1961).

Hanson et al. (1977) reported a case, in a previously healthy young man, of severe cellulitis that developed from a laceration that occurred during swimming. Aeromonas hydrophila was recovered in large numbers from the wound and the freshwater lake where the injury occurred.

Joseph et al. (1979) reported a case of primary soft-tissue infection caused by two species of Aeromonas (A. hydrophila and A. sobria) in a student diver conducting scuba operations in a freshwater lake.

Summary

1. As water appears to be the natural habitat of Aeromonas, this organism should not be used as an indicator of fecal pollution or as a sanitary indicator for bathing areas.
2. Aeromonas may be pathogenic, but sampling of recreational water should be considered for epidemiological investigations only.

3.4.6 Campylobacter jejuni

Maximum Limits

No limits are specified for Campylobacter jejuni in recreational waters. Sampling for this pathogen should be conducted when there is epidemiological or other evidence of its presence in the water. Sampling should also be considered to assess the hazards of excessive utilization of the water with possible person-to-person transfer of pathogens.

Criteria

Description

Campylobacter jejuni (C. fetus subsp. jejuni) and C. coli are now recognized as important enteric pathogens, often responsible for diarrhea in humans (Benenson 1985). Campylobacteriosis, also known as campylobacter enteritis or gastroenteritis, is the name of the illness caused by C. jejuni.

Campylobacter is thought to be responsible for a greater proportion of enteritis than either Salmonella or Shigella. Diagnosis is based on isolation of the organisms from feces using selective media, reduced oxygen tension, and an incubation temperature of 43°C. Visualization of motile curved, spiral, or S-shaped rods similar to those of Vibrio cholerae by phase- contrast or dark-field microscopy of feces can provide rapid presumptive evidence for campylobacter enteritis.

Source and Pathogenicity

Campylobacter jejuni has been isolated from water, mud, livestock, and dogs and cats. Birds, in particular, have been a well-documented reservoir, as small amounts of bird droppings in water can release many Campylobacter organisms (Benenson 1985; Sacks et al. 1986). Modes of transmission to humans include contact with animals, handling raw chicken, person-to-person contact, and consumption of contaminated food, raw milk, and water. The infective dose of Campylobacter in water is unknown. Comparatively, as few as 500 organisms in milk can be infectious (Robinson 1981).

Waterborne outbreaks of campylobacter enteritis have been associated with municipal water systems in North America (Sacks et al. 1986; Borczyk et al. 1987) and various European countries (Bolton et al. 1987).

Surface waters can contain Campylobacter spp., but their survival is temperature dependent. At 4°C, they can survive from 11 days to 4 weeks, and at 25°C, from 2 to 4 days (Mentzing 1981). In another study, C. jejuni inoculated into autoclaved mountain stream water remained viable for 33 days at 4°C. At 25°C, the organisms became non-viable within 4 days (Blaser et al. 1980). In an investigation of a waterborne outbreak, it was thought that, despite warm ambient temperatures, an accumulation of algae and scum in an open settling tank could have contributed to the organisms' extended viability (Sacks et al. 1986).

Occurrence in the Aquatic Environment

In a study by Taylor et al. (1982), C. jejuni was responsible for sporadic summertime diarrheal disease with the consumption of untreated surface waters in wilderness areas. Subsequently, C. jejuni was isolated from animals in the area. Another study (Taylor et al. 1983) involved a campylobacter enteritis outbreak that was epidemiologically linked to the community raw water system.

The survival of C. jejuni has been tested in drinking water, river water, and sewage; results indicate that its survival is restricted to a few days (Pickert and Botzenhart 1985). The concentration of oxygen and nutrients in the samples did not appear to affect survival, whereas temperature was demonstrated to be the most significant variable.

Campylobacter appears in the environment more commonly during summer. Studies of birds indicate an increase in Campylobacter carriage in the summer compared with the winter; this coincides with waterborne campylobacter outbreaks, which tend to occur only in summer and fall (Sacks et al. 1986).

In a study by Bolton et al. (1987) involving a river system that traversed rural and urban areas, the lowest frequency of isolation and the lowest counts obtained (10 Campylobacter/100 mL) were associated with samples collected from rural sites and areas with faster river flow. The greatest frequency of isolation and the highest counts (20 to 30 Campylobacter/100 mL) were associated with samples collected adjacent to, or downstream from, sewage works. A seasonal trend was demonstrated: the highest counts and most isolations were obtained in late autumn and winter, and the lowest counts and fewest isolations were obtained in the spring and summer. Heavy rainfall and the subsequent runoff from adjacent farmland were found to contribute to increased counts of Campylobacter in the river system.

Several studies have isolated Campylobacter spp. from surface water in association with Escherichia coli (Carter et al. 1987). These studies indicated that Campylobacter spp. were detected only in the presence of E. coli.

Summary

1. Recent improvements in environmental sampling techniques and means to distinguish strains have made environmental studies more feasible (Taylor et al. 1982).
2. Water is potentially an important reservoir of the thermophilic Campylobacter and is an established vehicle for the transmission of Campylobacter to man and domestic animals.
3. The use of a standard indicator of fecal pollution would be helpful in determining potential health hazards related to Campylobacter spp. as well as other bacterial intestinal pathogens.
4. All data to date link campylobacter enteritis to water consumption and not recreational contact. For this reason, no limit is being established at this time. Monitoring may be conducted where warranted on the basis of epidemiological or other data. A health hazard exists if Campylobacter can be consistently isolated from a bathing area.

3.4.7 Legionella

Maximum Limits

No limits are specified for Legionella spp. in recreational waters.

Criteria

Description

At least 22 species of Legionella have been identified. These are Gram-negative, non-spore-forming, aerobic bacilli, which grow at 35 to 37°C in special media (e.g., buffered charcoal yeast extract).

Pathogenicity

Since the outbreak of Legionnaires' Disease in Philadelphia in 1976, Legionellaceae have been recognized as an important cause of respiratory illnesses, causing legionellosis and Pontiac fever. Legionellosis is a multiple-system disease that can be fatal in immuno-compromised persons. Pontiac fever is a self-limited, flu-like illness, mainly affecting immuno-competent persons. The mode of infection is by inhalation of infected aerosols. Almost all outbreaks of endemic and epidemic, community-acquired, and nosocomial infections have been related to indoor plumbing and air conditioning systems (Benenson 1985).

Occurrence in the Aquatic Environment

All Legionella are aquatic bacteria, found mainly in water and mud (Edelstein 1985). Five reports can be found on legionellosis associated with recreational waters: two outbreaks of Pontiac fever associated with the use of whirlpool (Mangione et al. 1982; Goldberg et al. 1989); one case of a wound infection in a hydrotherapy tank (Brabender et al. 1983); one case associated with near drowning (Sekla et al. 1982); and 1 case following immersion in a river (Farrant et al. 1988).

Summary

1. Legionella are natural aquatic bacteria and cannot be used as an indicator of fecal pollution or as a sanitary indicator of bathing areas.
2. There are very few data available on legionellosis associated with recreational waters.
3. Routine testing of recreational waters for Legionella is not recommended.

3.4.8 Viruses

Maximum Limits

In Canada, no limits are specified for viruses in recreational waters. Sampling for viruses should be conducted when there is epidemiological or other evidence of their presence in the water or in order to assess the hazards of excessive utilization of the water with possible person-to-person transfer of pathogens.

Criteria

Description

Viruses are submicroscopic microorganisms that are unable to replicate outside their normal host. Among the more than 100 enteric viruses that are excreted in feces and could possibly be found in recreational waters, some can remain infective for several months in water and underlying sediments (Sattar 1981), including enteroviruses (polio, coxsackie, echo, and hepatitis A viruses), adenoviruses, rotaviruses, reoviruses, Norwalk viruses, caliciviruses, astroviruses, and coronaviruses. The infectious dose of some human enteric viruses can be as low as one tissue culture unit (Plotkin and Katz 1967; Ward and Akin 1984; Ward et al. 1986) and at least one order of magnitude lower than that of bacteria (Blaser and Newman 1982).

Pathogenicity

The diseases produced by the enteric viruses range from unapparent to severe. Enteric viruses can cause gastroenteritis, hepatitis A and hepatitis E (non-A, non-B hepatitis), fever, respiratory ailments, eye infections, central nervous system infections, poliomyelitis, etc. (Sattar 1978b; Gerba et al. 1985; Gust and Purcell 1987).

The lack of a central reporting of such infections coupled with rapid person-to-person transmissions have made it extremely difficult to determine the existence of waterborne viral diseases (Pipes 1978). The risk associated with bathing in virus-contaminated waters has been discussed by Payment (1984) and Craun (1986).

Epidemiology

Reports of recreational waterborne viral infections are rare (Paffenbarger et al. 1959; McLean 1965). Transmission of adenoviruses from swimming pools has been reported (Foy et al. 1968; Caldwell et al. 1974) and usually causes eye infections. Transmission of enteric viruses from lake waters has also been documented for coxsackievirus B3 (Hawley et al. 1973), coxsackie A16 (Denis et al. 1974), hepatitis A virus (Bryan et al. 1974), Norwalk virus (Koopman et al. 1982; Kappus et al. 1982), and, finally, echovirus (Walter-Offenhauser and Horn 1974).

Occurrence in the Aquatic Environment

A large number of enteric viruses may be found in the aquatic environment (Department of National Health and Welfare 1977; Bitton et al. 1985) as a result of pollution by animal wastes, municipal sewage, and other sources of human waste. In contrast to fecal coliform bacteria, which are present in all feces, viruses are excreted only from infected individuals (U.S. Environmental Protection Agency 1978a), who often are symptomless carriers in the under-15-years age group (Ramoz-Alverez and Sabin 1956). It is established that virus levels in water vary markedly on an hourly, daily, and seasonal basis (Berg and Metcalf 1978), whereas total and fecal coliform levels are generally more stable. For these reasons, the ratio of viruses to total coliforms of 1:65 000 does not always hold true (Scarpino 1975). Berg and Metcalf (1978) also demonstrated that fecal coliforms and enteric viruses do not occur in any constant ratio. Viruses have been found in the absence of detectable fecal coliforms (Berg 1978), and significant levels of viruses have been found in waters well within the bacteriological limits recommended for recreational waters (Goyal et al. 1978). Resistance to chlorine inactivation has been reported by Bates et al. (1977).

Several studies have investigated virus levels in sewage (Sattar and Westwood 1977, 1978; Subrahmanyan et al. 1979; Sekla et al. 1980), rivers (Subrahmanyan 1977; Sattar and Westwood 1977; Sekla et al. 1980; Payment et al. 1988), and lakes (McLean 1965; Subrahmanyan 1977). An extensive investigation of water in the Ottawa River (Sattar 1978a) revealed that, at one site, seven out of 16 samples were positive for virus; at a recreational beach, 11 of 20 samples were positive, with a range of 0.6 to 16.8 infective units/ 10 L. The sources of contamination were raw sewage inputs, as well as treated, chlorinated sewage effluents (Sattar and Westwood 1978).

Payment (1977) suggested a limit of one tissue culture infectious dose/40 L, whereas Sattar (1978a) proposed one infective unit/10 L when a 100-L sample is examined. In Israel, Shuval (1975) recommended that tentative limits be established. In the United States, Melnick (1976) suggested a limit of one detectable infectious virus unit/10 U.S. gallons (37.9 L) of recreational water. The State of Arizona adopted a surface water standard of not greater than one enteric virus/40-L sample (Gerba 1988).

The investigation of viral waterborne outbreaks requires special laboratory facilities. Occasional monitoring for viruses can be carried out to determine their distribution and relationship to disease incidence, but routine tests of recreational waters is not recommended.

Summary

1. Viruses are known to be pathogenic in low numbers. As few as one infective tissue culture unit can cause disease when ingested. Concentration and enumeration steps are too detailed to make routine monitoring practical.
2. Very few data are available on current virus levels in recreational waters.
3. There is no correlation between viral and bacterial counts in recreational waters.
4. No limit can be established for viruses at this time. Monitoring and epidemiological studies are needed to determine the levels of viruses in water and the health effects of swimming in virus-contaminated waters.

3.4.9 Protozoa

Maximum Limits

No limits are specified for pathogenic protozoa in recreational waters.

Criteria

Description, Pathogenicity, and Occurrence

A large number of pathogenic parasites can occur in the aquatic environment. Of potential importance in Canada are 4 protozoa (Giardia, Cryptosporidium, Naegleria, and Entamoeba histolytica) and 1 helminth (Schistosoma).

Giardia is currently the most common pathogenic intestinal protozoan in Canada and the United States. The ingestion of a few (10 to 100) viable cysts can cause a diarrheal illness (giardiasis). Transmission can be from person to person or via food or water. Waterborne giardiasis has received much attention lately, as outbreaks have been traced to pristine waters, as well as to sewage-contaminated potable waters (Lin 1985). Giardia are more resistant to chlorination than indicator organisms, pathogenic bacteria, and viruses (Sobsey 1989). Thus, fecal coliform counts cannot be used as indicators of protozoal contamination of recreational waters. Giardiasis associated with toddler swim classes has been discussed by Harter et al. (1984), and transmission in a swimming pool was reported by Porter et al. (1988). A water slide has also been incriminated in an outbreak of giardiasis (Greensmith et al. 1988).

Cryptosporidium, a newly recognized pathogenic protozoan, may be as important as Giardia (Rose 1988). The ingestion of low levels of viable oocysts can also result in a diarrheal illness known as cryptosporidiosis. Like Giardia, Cryptosporidium can also be transmitted from person to person or via food or water. The illness, which may be fatal in immuno-compromised patients, has occurred in Canada (Mann et al. 1986). No outbreaks have been linked to recreational waters, but major outbreaks in the United States and the United Kingdom have resulted from the ingestion of inadequately treated drinking water (D'Antonio et al. 1985; Rose 1988). Although information is only preliminary, it appears that Cryptosporidium is even more resistant to disinfection than Giardia (Sobsey 1989). Oocysts have been recovered in surface waters in British Columbia (Isaac-Renton et al. 1987) and the United States (J. Rose and C. Gerba, unpublished report).

Naegleria fowleri and other members of the freshwater amoeba group have caused more than 100 cases of an often-fatal primary amoebic meningoencephalitis (PAM). PAM has been reported from the United States, South America, Europe, Australia, and New Zealand in persons swimming in fresh waters, lakes, and ponds, and even in indoor pools filled with chlorinated heated river water, as these amoeba are very resistant to chlorine and thrive in warm waters (Griffin 1977; Hallenbeck and Brenniman 1989). Naegleria fowleri has been recovered from surface waters in Canada (Seyfried et al. 1984).

Entamoeba histolytica is found throughout the world, affecting 10 per cent of the world population. It can cause amoebic dysentery and liver abscesses. The cysts can survive well in water. The effect of temperature has been studied by Jones and Newton (1950). The best-studied outbreak of waterborne amoebiasis was related to the drinking of contaminated water in the United States, as reported by Le Maistre et al. (1956).

Schistosoma spp. are digenetic trematodes (also known as flukes or flatworms). The larvae (cercariae), released in water by infected aquatic snails, must enter the skin of a susceptible host to complete their life cycle. The species responsible for human schistosomiasis, a disease of considerable morbidity affecting 200 million people in Africa, South America, the Middle East, and parts of Asia, does not occur in North America.

Avian schistosomes can be found in Canada. The cercariae of bird and rodent schistosomes may penetrate the human skin, causing a dermatitis known as swimmers' itch. This itch occurs in bathers using recreational lakes in many parts of the world, including North America, as well as on certain coastal seawater beaches. These schistosomes do not mature in humans (Levy and Folstad 1969; Benenson 1985) but die just beneath the epidermis. Subsequent exposure to cercariae stimulates an allergic response. At the time of cercarial penetration, a prickling sensation is noted (Hoeffler 1977).

Prevention of swimmers' itch can be assured only by complete avoidance of aquatic sports in areas where the disease has been a problem. An acceptable alternative might be the elimination of the molluscan hosts, using approved molluscicides. Prevention of secondary infection by bacteria can be achieved by appropriate hygienic measures. Attempts to prevent the dermatitis during or after cercarial penetration are largely ineffective. Time-honoured measures such as rough towelling and alcohol rubdowns are without significant merit. Clothing and chemical repellents are of limited value. Treatment of the dermatitis is directed towards the relief of symptoms with the usual doses of antihistaminic and antipruritic medications.

Summary

1. Routine monitoring of waters for protozoa is not recommended. However, provincial laboratories should be able to participate in the investigation of documented waterborne outbreaks.

3.4.10 Toxic phytoplankton

Maximum Limits

No limits are specified for toxic phytoplankton in recreational waters. However, water containing a blue-green or turquoise scum is indicative of an algal bloom. Such waters should be avoided because of reduced clarity and the possible presence of algal toxins.

Criteria

Description

Phytoplankton, which are microscopic floating plants, can become a hazard and a nuisance in recreational waters, especially when they concentrate at the water surface in "blooms." This can be a natural phenomenon, but it is often caused by cultural eutrophication. The presence of certain species in a freshwater community is often a sensitive indicator of recreational water quality. Nutrient enrichment in bodies of water affects the density and diversity of the fauna and flora.

The algae of concern in lakes and ponds are usually blue-green algae. These algae are unusual because they share some typical features of both algae and bacteria. In some species, the algae are tiny single cells that cannot be seen with the naked eye; in most species found in Canadian lakes, the cells are grouped in colonies. The colonies may form strings, flakes, or globules and can reach a size of several millimetres. To the naked eye, they may look like fine grass clippings in the water or a homogeneous soupy mass.

In a typical summer, a lake water sample usually contains 20 or more blue-green algal species, along with dozens of other species of algae. The blue-green algae may form massive blooms under certain conditions during summer (Reynolds and Walsby 1975).

Blue-green algal cells contain small gas bubbles (vacuoles) that allow them to control their buoyancy. Usually the algae are well distributed in the zone where nutrients and light are in the optimum ranges. Blooms are not the result of a sudden surge in algal growth rate but occur when their buoyancy regulation is disrupted. Reynolds and Walsby (1975) stated that bloom formation requires that a substantial population of the algae already exists, that they have excess buoyancy, and that the water is calm enough to allow them to float up to the surface. The algae may develop excess buoyancy when turbulence sends them too deep, during the hours of darkness, when the concentration of carbon dioxide in the water becomes limiting, when the algal population is at the end of its growth cycle and is aging, or any combination of these factors. Blooms often occur in late August and in September when the algal population is aging and the hours of darkness are growing longer. Dillenberg and Dehnel (1960) and Senior (1960) reported a striking example of a sequence of meteorological events that illustrated these factors in a Saskatchewan lake.

Toxicity

Toxic algae are found in all aquatic environments and have been responsible for the death or illness of livestock, waterfowl, fish, and humans (Carmichael et al. 1985). The most important taxonomic phyla are Pyrrhophyta (dinoflagellates), Chrysophyta, and Cyanophyta (blue-green algae). Well-known instances of marine algal toxicity are associated with "red tides," which cause fish kills and poison edible shellfish and other bottom fauna (Tangen 1977; Cross and Southgate 1980). On the west coast of North America, clams and mussels become poisonous by ingesting the dinoflagellate Gonyaulax catenella and accumulating saxitoxin in the digestive tract. On the east coast, the associated poisoning is caused by Gonyaulax tamarensis, with a mixture of saxitoxin and three related toxins.

Blue-green algae have been known to cause animal toxicity in lakes, ponds, and dugouts for over 100 years. The three blue-green algae most often identified as the cause of poisoning are Anabaena flos-aquae, Aphanizomenon flos-aquae, and Microcystis aeruginosa (McLeod and Bondar 1952; Senior 1960; Aziz 1974; Moore 1977, Richard et al. 1983).

A number of toxins produced by freshwater blue-green algae have now been identified. Two neurotoxins and two hepatotoxins have been associated with Anabaena flos-aquae. Anatoxin-a is a potent postsynaptic, depolarizing, neuro-muscular blocking agent that causes death by respiratory arrest within minutes to a few hours depending on species, dosage and prior food consumption. The purified toxin intraperitoneal LD50 for mice is about 200 mg/kg body weight, with survival time of 4 to 7 minutes (Carmichael 1988). Another neurotoxin produced by Anabaena flos-aquae is anatoxin-a(s), an anticholinesterase, which also causes viscous salivation, lachrymation, urinary incontinence and defecation prior to death by respiratory arrest. Anatoxin-a(s) is about four times more toxic than anatoxin-a. The two hepatotoxins produced by Anabaena flos-aquae appear to be heptapeptides similar to two analogues produced by Microcystis aeruginosa (Carmichael 1988).

The toxins produced by Aphanizomenon flos-aquae consist mainly of two neurotoxic alkaloids that strongly resemble saxitoxin and neosaxitoxin (Sasner et al. 1984) These are more generally known as the paralytic shellfish poison associated with "red tides", and received their names from the Alaska butter clam Saxidomus giganteus (Schuett and Rapoport 1962). Three other neurotoxins have been detected that are labile and not similar to any known paralytic shellfish poisons (Carmichael 1988).

The hepatotoxins produced by Microcystis aeruginosa attack the liver, causing severe necrosis and hemorrhaging. Death from hypovolemic shock caused by interstitial hemorrhage into the liver may occur within 1 to 3 hours in acutely dosed and sensitive animals. In other animals, death may occur as long as 36 hours later, with the animals displaying symptoms of incoordination, muscular weakness and fibrillation, unsteady gait, recumbency, labouredrespi-ration, salivation, lacrimation, diarrhea, and in those suffering lingering deaths, icterus, photodermatitis, and loss of condition (Soll and Williams 1985). A diarrhea toxin has also been isolated from M. aeruginosa, and this may be a cause of gastroenteritis when no other known etiological agent can be identified (Aziz 1974).

Toxicity due to blue-green algae can occur only if there is a bloom dominated by the toxic strains of the bloom species. Toxic strains and non-toxic strains of a bloom species may occur at the same time in a lake; as a result, some parts of the lake could become toxic, while others could remain safe. Toxicity in a lake is normally transient, lasting only as long as the bloom or signs of the bloom persist.

The reason why toxic strains suddenly become more dominant than the non-toxic strains is not known (Carmichael et al. 1985). Consequently, toxicity due to blue-green algae is even less predictable than the blue-green blooms themselves. Lakes that have never had a problem can suddenly become toxic. Conversely, lakes that have shown toxicity in the past may never show it again. Routine monitoring of lakes where blooms have occurred is the best approach to identifying a problem.

When a lake becomes toxic as a result of a blue-green bloom, the only sign of a problem may be dead fish, waterfowl or other wildlife along the shoreline. Occasionally, domestic animals such as cattle or dogs may be poisoned if they have no other source of drinking water.

In Canada, animal poisonings have been reported for Alberta, Saskatchewan, Manitoba, and Ontario (Stewart et al. 1950; O'Donoghue and Wilton 1951; McLeod and Bondar 1952; Neil 1957; Senior 1960; Carmichael and Gorham 1978). Most algal poisonings in western Canada are associated with blooms of Anabaena flos-aquae.

There is some concern, particularly in the Prairie provinces, about possible poisoning of swimmers from blooms of blue-green algae. Humans are just as susceptible to blue-green toxins as animals, but it is unlikely that people would voluntarily drink much lake water during a bloom because of the objectionable appearance and odour of the water. However, people may suffer acute discomfort after accidentally ingesting or contacting blue-green algae or water containing toxins. Symptoms may include fever, headache, dizziness, stomach cramps, vomiting, diarrhea, skin and eye irritations, sore throat and swollen lips. Symptoms seldom persist for more than two to eight days. Children may be more intensely affected because they spend more time in the water than adults and are more likely to have accidentally swallowed contaminated water. In addition, they may have lower tolerances to the toxins than adults.

Dillenberg and Dehnel (1960) recorded the poisoning of a man who fell into a Saskatchewan lake containing a dense bloom of blue-green algae. Nausea, diarrhea, headache, cramps, and high temperature occurred. Microcystis and Anabaena, but no other pathogens, were found in his feces. Similar illness occurred among children who swam in another Saskatchewan lake also containing a bloom of Microcystis and Anabaena. Schwimmer and Schwimmer (1968) reported a near-fatal case of a boy who developed fever, laboured breathing, pneumonitis, generalized pains, and coma after swimming in water containing Microcystis aeruginosa. They also reported the poisoning of another boy who fell into a lake, accidentally swallowed water containing Aphanizomenon, and developed symptoms similar to those experienced by the first case described above. Schwimmer and Schwimmer (1968) cited reports of allergic responses to Microcystis, Anabaena, and Aphanizomenon, including papulo-vesicular eruption and acute allergic conjunctivitis, specifically with Anabaena. Billings (1981) also reported possible contact irritation associated with Anabaena. The recorded instances of illnesses in humans associated with blooms of toxic algae are few, but such illnesses may be more common than is thought because of lack of diagnosis.

Recreational lake users should take particular care where the water contains algae with the distinctive blue-green or turquoise colour and should treat any intense bloom with suspicion. Humans should not drink water from bloom-infested areas of lakes and reservoirs, nor should they swim or wade in water containing concentrated blue-green algal material. People should also take care to provide alternative water sources for domestic animals and pets. Bodies of water that have the green colouration of normal plants, like grass or the pond-dwelling plant duckweed, are most unlikely to be toxic regardless of the thickness of the scum.

Controlling Phytoplankton

Attempts at controlling blue-green algae by chemical treatment may be detrimental to lake water use and lake ecology. Hanson and Stefan (1984) reported short-term effects, such as the death of some algae, depletion of dissolved oxygen resulting in occasional fish kills, and the recycling of phosphorus from the lake bed, which fosters conditions that allow regrowth of the algae within 7 to 21 days. Further, because the toxins are released on the death of the algae, water that has been treated with copper or other chemicals to kill the algae may be especially dangerous for the first 24 hours after treatment.

The detrimental long-term effects of algal control are the accumulation of copper in the lake bed sediments, increasing resistance of the phytoplankton to copper, increasing domination of blue-green algae species, particularly Aphanizomenon, over green algae species and rough fish over game fish, the disappearance of macrophytes, and the loss of benthic macroinvertebrates. When the water overlying the bottom sediments becomes anoxic, the copper will re-enter the limnetic zone (Prepas and Murphy 1988). The use of copper suphate may be forbidden in some jurisdictions.

There are potential alternatives to the use of copper sulphate and aquatic herbicides such as Reglone A (Diquat). A finding that total phosphorus may be the best predictor of relative biomass of blue green algae (Trimbee and Prepas 1987) has lead to lake treatment based on decreasing the total phosphorus concentration in the water. Babin et al. (1989) report that the application of lime (calcium hydroxide) slurry has dramatically decreased both soluble reactive phosphorus and chlorophyll a (used to estimate algal biomass) in storm water retention ponds and in hypereutrophic natural lakes. Stable stratification of lake water also tends to favour the dominance of blue-green algae because of their buoyant properties. Experimental mixing has been successful in suppressing their growth in reservoirs previously subject to frequent blooms (Reynolds and Walsby 1975).

Summary

1. No limits are recommended for toxic phytoplankton, but swimming in waters containing blue-green algal blooms should be avoided.
2. Bather poisonings have occurred after immersion in lakes and ponds containing dense blooms of blue-green algae.
3. Sampling recreational waters for toxic phytoplankton should be considered only for epidemiological investigations.