Priority Substances List Assessment Report for Ammonia in the Aquatic Environment

2.0 Summary of Information Critical to Assessment of "Toxic" Under CEPA 1999 (Continued)

2.4 Effects characterization (continued)

2.4.3 Sublethal effects on freshwater organisms

Sublethal effects occur at concentrations and over extended periods that do not result in acute lethality to the organism, but can affect the population of species and community characteristics. The most evident responses are integrative and are exhibited by reductions in growth (length or weight) or are related to reproductive success (egg production, hatching, larval survival). Other effects, such as behavioural responses, tissue damage (e.g., pathological changes in the tissue of the gills, liver and kidney of fish) or biochemical or physiological changes, can affect the individual but in most cases are reversible and will not necessarily change the character of the community. The concentrations at which these sublethal responses occur are presented in Table 6.

2.4.4 Acute and sublethal effects on saltwater organisms

Available acute and chronic ammonia toxicity data for saltwater organisms are more limited than those for freshwater organisms. The U.S. EPA (1989) published a review on the saltwater toxicity of ammonia, much of which is reported in Table 7.

Table 6 Summary of mean sublethal endpoints in freshwater species

Enlarge image

2.4.4.1 Acute toxicity

Cheung and Wong (1993) found that relatively unpolluted and heavily polluted sediments dredged from around Hong Kong were both highly toxic to the marine clam, Tapes philippinarium. A correlation coefficient of 0.99 at
p < 0.001 was determined between mortality and ammonia concentrations in the seawater in tests with the relatively unpolluted sediments, and from 0.92 to 0.96 at p < 0.001 for the heavily polluted sediments. The total ammonia concentration in seawater at the ET⁵⁰ (time to 50% effect; 14-15 days) was 10-11 mg/L in both sediment systems.

Table 7 Ranked mean acute toxicity values for un-ionized ammonia

Species	Mean LC50 or EC50 (mg NH3/L)	Temperature (°C)	Salinity (‰)	pH
Winter flounder, Pseudopleuronectes americanus	0.49	7.5	31	8.0 (7.9-8.1)1
Red drum, Sciaenops ocellatus	0.55	25 (25-26) 1	29 (28-30) 1	8.1 (8.0-8.2)1
Sargassum shrimp, Latreutes fucorum	0.77	23.4	28	8.07
Prawn, Macrobachium rosenbergii	0.78	28	12	7.6 (6.8-8.3)1
Planehead filefish, Monocanthus hispidus	0.83	23.4	28	8.07
Copepods:
Eucalanus elongatus	0.87	20.3	34	8.0
Eucalanus pileatus	0.79	20.5	34	8.2
Morone spp.:
Striped bass, M. saxatilis	0.48	19.3 (15-23)1	12.9 (5-34)1	(7.2-8.2)1
White perch, M. americana	2.13	16	14	8.0
Mysid, Mysidopsis bahia	1.02	23.2 (19.3-26.5)1	21.4 (10-31)1	(6.8-9.2)1
Spot, Leiostomus xanthurus	1.04	20.4	9.3	79.2
Silversides:
Inland, Menidia beryllina	1.32	25.3 (18-32.5)1	26.1 (11-31.5)1	(6.9-9.1)1
Atlantic, Menidia menidia	1.05	20.3 (10.8-24.8)1	11.6 (8.5-29.8)1	8.0 (7.0-9.0)1
Striped mullet, Mugil cephalus	1.54	21.8 (21.0-23.3)1	10	8.1 (8.0-8.1)1
Grass shrimp, Palaemonetes pugio	1.65	19.9 (19.3-20.4)1	19.2 (10-28.4)1	8.1 (7.9-8.1)1
Sea bream, Sparus auratae	1.88	22.5 (17.9-27)	37.5 (34.5-40.5)	8.1
At dissolved oxygen levels:
93% saturation	1.93	27	40.5	8.1
61% saturation	1.28
33% saturation	0.97
26% saturation	0.41
Lobster, Homarus americanus	2.21	21.9	33.4	8.1
Sheepshead minnow, Cyprinodon variegatus	2.74	21.0 (10.3-32.5)1	19.4 (9.8-32.5)1	(7.6-8.1)1
Three-spined stickleback, Gasterosteus aculeatus	2.93	19 (15-23)1	26.3 (11-34)1	(7.6-8.2)1
Turbot, Scophthalmus maximus	2.96	17.9 (17-18.8)	34.3 (34.0-34.5)	8.15
Brackish water clam, Rangia cuneata	3.08	20.2	9.2	7.95
Mudskippers:
Periophthalmodon schlosseri	9.13	2	15	N/A
Boleophthalmus boddaertid	1.02	25	15	N/A
Quahog clam, Mercenaria mercenaria	5.36	20	27	(7.7-8.2)1
Lake Magadi tilapia, Oreochromis alcalicusgrahamic	11.472	34	N/A	9.9
Eastern oyster, Crassostrea virginica	19.10	20	27	(7.7-8.0)1

¹Mean (range) values for temperature, salinity or pH calculated when values from original text were given individually. When only ranges were given in original text, mean was not calculated.
²Average of LC₅₀ values at 24 and 48 hours.

In published literature, mean LC₅₀ values for marine invertebrate species found in North American waters range from 0.77 to 19.1 mg NH₃/L; for marine fish species, they range from 0.49 to 2.9 mg NH₃/L (see Table 7). The winter flounder (Pseudopleuronectes americanus) had the most sensitive acute toxicity value of 0.49 mg/L. The remaining genera tested have mean acute values within an order of magnitude of that for the winter flounder. The three most tolerant North American species reported by the U.S. EPA (1989) are molluscs. Species mean acute values of 3.08, 5.36 and 19.1 mg/L were reported for the brackish water clam (Rangia cuneata), the quahog clam (Mercenaria mercenaria) and the eastern oyster (Crassostrea virginica), respectively. Except for these molluscs, there is no phyletic pattern in acute sensitivity to ammonia. Fishes and crustaceans are well represented among both the more sensitive and the more tolerant species tested.

Few consistent trends or patterns were evident in the acute toxicity values with respect to biological or environmental variables. Contributing to this, in part, is test variability. Variability in acute toxicity values may reflect differences in condition of the test organisms, changes in the exposure conditions, particularly pH, during testing, and variance incurred through calculation of un-ionized ammonia concentrations. Few differences are evident in acute toxicity at different salinities in tests with similar life stages and similar pH and temperature conditions. Temperature also has little influence on acute ammonia toxicity to most saltwater animals. There are few differences in acute toxicity with respect to differences in life stage or size of the test organism (U.S. EPA, 1989). Several data sets on the effect of pH on the toxicity of un-ionized ammonia suggest that, unlike the data on freshwater species, the pH-toxicity relationship is not consistent between species.

The U.S. EPA (1989) concluded that there was insufficient information to conclude that any of these factors, when acting alone, has a consistent major influence on the acute toxicity of NH₃ to saltwater organisms.

2.4.4.2 Sublethal toxicity

The U.S. EPA (1989) also reported on unpublished chronic toxicity tests, but with only two saltwater species, neither of which is native to Canada. A life cycle toxicity test has been conducted with the mysid, Mysidopsis bahia, and an early life stage test has been completed with the inland silverside (Menidia beryllina).

The M. bahia test lasted 32 days and was reported in Cardin (1986). Survival was reduced to 35% of controls, and length of test organisms was significantly reduced at 0.33 mg NH₃/L.

The effect of ammonia on survival and growth of the inland silverside M. beryllina was assessed in an earlylife stage test lasting 28 days (Poucher, 1986). Fry survival was reduced to 40% at 0.38 mg/L. Average weights of surviving fish at concentrations above 0.074 mg/L were significantly less than that of controls.

2.4.5 Ecosystem effects

This section focuses on effects of ammonia on whole ecosystems, where the impact is not direct toxicity of ammonia. The two major processes found are acidification of soft waters and eutrophication of aquatic and terrestrial ecosystems.

A well-documented effect of human impact upon aquatic ecosystems is eutrophication, a multifaceted term generally associated with increased productivity, structural simplification of biotic components, and a reduction in the ability of the metabolism of the organisms to adapt to imposed changes (reduced stability). In this condition of eutrophication, excessive inputs commonly seem to exceed the capacity of the ecosystem to be balanced. In reality, however, the systems are out of equilibrium only with respect to the freshwater chemical and biotic characteristics desired by humans for specific purposes (Wetzel, 1983).

Acidification of water by ammonium sulphate deposition is a strong reaction, stronger than the addition of sulphuric acid. This is due to the nitrification of the ammonium molecule, releasing hydrogen ions, in addition to the release of the acidic sulphate molecule.

2.4.5.1 Freshwater eutrophication

The material presented on aquatic and marine coastal eutrophication is from a review of the literature on the causes and conditions of aquatic eutrophication in Canada, prepared for Environment Canada (Chambers et al., 2000).

Nutrients are essential to lakes because they provide the raw material for the growth of algae, which are the food sources of zooplankton, which, in turn, are eaten by fish. The concentration of nutrients in a lake is determined by the interplay of the magnitude, timing and bioavailability of the nutrient load, the rate of water supply compared with the volume of the lake (flushing time) and the depth of the lake.

In lakes and rivers not affected by nutrient inputs, the nutrient cycling processes are typically in balance. However, with the excessive input of nutrients, like ammonia and dissolved phosphorus, these processes become unbalanced, usually resulting in large standing crops of algae and plants. The phytoplankton have a high total respiration demand, reducing oxygen concentrations and generating toxins (depending on the algal species present) that can suppress herbivorous plankton. The inevitable die-off of algae in late summer increases the bacterial populations tremendously, which can also increase toxins in the water and will severely decrease the dissolved oxygen content to the point where fish can be killed. Over the long term, elevated eutrophic rates can alter the biological community towards organisms more tolerant of shaded, oxygen-deficient waters.

In most Canadian lakes, phosphorus is the nutrient that is most in demand, and algal growth in the majority of lakes is therefore said to be phosphorus limited. Discovering this relationship led to the significant reduction in releases of phosphorus from municipal water treatment plants, largely through the regulation of phosphorus in detergents and the chemical precipitation of phosphorus in the plants prior to discharge. Major improvements in water quality resulted from these actions.

The role of ammonia in aquatic eutrophication is as a source of nitrogen for the generation of nitrates that are directly usable by algae and aquatic plants. In lakes that receive continuous inputs of ammonia and phosphorus (secondary sewage treatment facilities typically release both), the nitrification process can be operating at a maximum in summer, so that the lakes are enriched in nitrates and phosphorus, leading to lush growths of algae and rooted plants. In the fall, the die-off of algae and plants depletes oxygen and creates a harsh environment for fish. In water systems that receive ammonia but not phosphorus, the nitrification process will still be at a maximum, but, due to the deficiency in phosphorus, they have limited algal and macrophytic growth. In these cases, eutrophication does not proceed, but the concentrations of ammonia and nitrates can rise to toxic levels and can still lead to severely depleted oxygen conditions from the nitrification process. Still other lakes are truly limited by nitrogen; one such system in Canada is the Qu'Appelle Lakes in southern Saskatchewan (see Section 3.1.2.2.3).

2.4.5.2 Coastal marine eutrophication

In the last 20 years, the causes and extent of coastal eutrophication have been increasingly recognized as a global problem (Howarth, 1988; Vollenweider, 1992; NRC, 1993; UNEP, 1995; Paerl, 1997). Coastal areas, including fjords, estuaries, lagoons, continental shelves and inland seas, comprise 1-2% of the total area of the ocean, yet are responsible for 20% of global primary production (Duarte, 1995). These regions receive the bulk of their nutrient inputs from freshwater sources (i.e., terrestrial runoff, rivers and groundwater). The natural background levels of nutrient concentrations of these inputs are normally much higher than those of even the most eutrophic seawater (Dederen, 1992).

Nitrogen is generally the nutrient limiting primary production in the open ocean, in contrast with fresh waters, where phosphorus is typically the limiting nutrient (Howarth, 1988; Vollenweider, 1992). It is in the coastal zone where nutrient-rich freshwater inputs are diluted into the nutrient-poor saline environment of the open ocean. In these highly dynamic transitional waters, either phosphorus or nitrogen limitation can occur, depending on a set of complex interactions.

In recent decades, nitrogen and phosphorus transport to coastal waters has increased (Howarth et al., 1996) and is correlated with various indices of human activity in the watershed (Cole et al., 1993; Caraco, 1995; Howell et al., 1996; Vitousek et al., 1997). If nitrogen is measured, then most of the inputs to coastal waters are derived from non-point sources, typically as nitrate (NRC, 1993).

Evidence from the northern hemisphere indicates that over-enrichment of coastal waters has created a niche occupied by a diverse group of dinoflagellates and diatoms that, like their counterparts in eutrophic lakes (the blue-green algae), produce toxic chemicals (Burkholder et al., 1992). Marine algae have been found responsible for at least four different illnesses in human consumers of molluscs as well as massive mortality of fish, birds and marine mammals (Paerl, 1997). The occurrence of these "harmful algal blooms" has resulted in the closure of shellfisheries, resulting in large economic impacts on coastal communities. The exact cause of these blooms is not clear, although they tend to follow periods of intense rainfall, runoff and intense irradiation from sunlight (Smayda, 1997).

Eutrophication of Canada's east and west coasts is not occurring at present. There are some indications that coastal areas around Vancouver and Halifax are impacted as a result of sewage effluents, but these are not eutrophication issues. This situation will likely remain as long as anthropogenic nutrient loading does not increase substantially (Chambers et al., 2000).

2.4.5.3 Terrestrial eutrophication

Among the mineral elements, nitrogen is required in the largest amount by plants; very often growth is limited by its supply. When more nitrogen is added, plants grow more rapidly, and the nitrogen in the increased plant biomass is effectively retained by the ecosystem. In addition, plants can accumulate nitrogen, as nitrate, in tissues in excess of the specific nutritional requirements. Ecologically, this may be an adaptation to deal with a chronically low nitrogen supply. In effect, plant growth responds to increased nitrogen supply until nitrogen is no longer the limiting factor for growth. Nitrogen-deficient ecosystems can tolerate, even benefit from, periodic excessive doses of nitrogen; however, metabolic imbalances can occur if the excessive nitrogen levels occur for too long.

Nitrogen addition has the potential to affect many attributes of the terrestrial environment, not all of which are well understood. Among the indirect effects, increased leaching of nitrate from soils is one of the more obvious. The concept of "nitrogen saturation" has been used to describe the level of nitrogen in an ecosystem that maximizes the retention within the ecosystem (Aber et al., 1989). Additions above this limit result in nitrate leaving the ecosystem in amounts that could be detrimental downstream. This concept is based on the observation that ecosystems cycle nitrogen very efficiently.

Critical loads of ammonia were established in Europe to avoid two general types of effects. One was the leaching of nitrogen, typically as nitrate, from ecosystems that normally are very conservative in nitrogen cycling. The other general effect to be avoided is the shift in dominance among species, especially in nitrogen-poor environments (Schulze et al., 1989; Bobbink et al., 1992; De Vries, 1992). Some other indirect effects are subtle, such as the loss of mycorrhizal fungi associated with conifer tree roots (Pérez-Soba et al., 1995). The critical loads for nitrogen promulgated by the Dutch Priority Programme on Acidification ranged from 9.8 to 42 kg/ha per year, with the lowest values of this range associated with avoiding changes in species composition in coniferous forests (Lekkerkerk et al., 1995).

2.4.5.4 Acidification

Deposition of ammonium sulphate, the most common form of ammonia particulate, will generate considerable quantities of acid, as eight hydrogen ions may be released during nitrification. The Dutch, Belgians, Norwegians and Germans have found that excessive quantities of ammonium sulphate are having adverse impacts on poorly buffered soils and waters in close proximity to large sources (Schuurkes, 1986; Schuurkes et al., 1986; Gjessing, 1994).

In long-term, indoor, soft-water ecosystem studies, Brouwer et al. (1997) showed that acidification of an ecosystem was greater when ammonium sulphate was deposited in rainfall than when sulphuric acid was deposited. This is due to the nitrification of the ammonium, releasing extra hydrogen ions into the ecosystem. Increased levels of dissolved metals were detected, as well as shifts in the plant community. Plants typical of soft waters declined and were overgrown by Sphagnum species and Juncus bulbosus. The recovery of the impacted ecosystems was also different; the sulphuric acid system recovered quickly, but the ammonium sulphate ecosystems did not fully recover after 10 years of clean water. The ecosystems most sensitive to such acidic inputs are found on the Canadian Shield throughout much of eastern Canada. There has been little in the way of ammonium particulate monitoring within the Canadian acid monitoring program, so the contributions of ammonia to acidification in Canada are not known.

2.4.6 Abiotic effects mediated through the atmosphere

Ammonia is the most prevalent alkaline gas in the atmosphere, as well as the third most common form of nitrogen in the troposphere. Because of its high reactivity, ammonia readily combines with acidic chemical species, such as hydrochloric acid, nitric acid or sulphuric acid, forming ammonia aerosols. Klemm and Gray (1982) determined that the acidity of rainfall in Alberta was determined as much by the presence of alkaline species (calcium and ammonium ions) as by the absence of acidic species (sulphur and nitrogen oxides). Even so, un-ionized ammonia as an atmospheric gas itself is rather passive: it either deposits quickly near sources or is converted to particulate form. As a result, particulates can be transported long distances, affecting tropospheric aerosol loading and thus issues of visibility, smog and climate.

An important connection for air quality issues, therefore, is the conversion of ammonia gas into the aerosol form, increasing tropospheric loading of respirable particulate matter, PM₁₀ and PM_2.5 Ammonia in the atmosphere can determine the type and quantity of fine particulate matter. The chemically preferred form for sulphate is ammonium sulphate, solid or aqueous. However, competition between sulphate and nitrate for the available ammonia produces complicated aerosol behaviour. In areas with low concentrations of atmospheric ammonia, most particulate matter will be acidic, as there is insufficient ammonia to neutralize the available sulphate. In areas with high ammonia concentrations, however, any ammonia that does not react with sulphate will be able to react with available nitrate, forming ammonium nitrate aerosols. PM₁₀ and PM_2.5 have been determined to be "toxic" to humans under CEPA 1999, so effects of ammonia-containing particulate matter are not considered in this report. Table 8 lists other air issues that are connected to atmospheric ammonia.

2.5 Toxicokinetics, mode of action and metabolism

2.5.1 Freshwater fish

This information is derived from an unpublished review (Randall, unpublished), used with the author's permission.

Most biological membranes are permeable to NH₃ , but not NH₄⁺. Ammonia is excreted by diffusion across the body surface of most aquatic animals, usually the gills, although there may be some carrier-mediated excretion of NH₄⁺ in some species. The rate of NH₃ excretion is determined by the magnitude of the NH₃ gradient between blood and water (Wilson et al., 1994). Ammonia excretion is augmented by acidic conditions in the water, because any NH₃ excreted into the water is rapidly converted to and trapped as NH₄⁺, maintaining the NH₃ gradient across the gills and augmenting ammonia excretion. Many freshwater fish actively excrete protons, forming an acid boundary layer next to the gill surface (Lin and Randall, 1991), and this augments ammonia excretion (Wright et al., 1989). Above water pH 9.0, ammonia excretion is reduced because of the absence of ammonium ion trapping (Wright et al., 1989), resulting in elevated plasma ammonia levels (Yesaki and Iwama, 1992). Thus, many animals have difficulty excreting ammonia when exposed to alkaline conditions.

2.5.2 Marine fish

Ammonium ion diffusion across the gills may be significant in seawater teleost fish, where ionic permeability is high (Evans, 1984).

The body surface of marine animals is generally more permeable to ions than that of freshwater animals (Evans, 1984). Thus, the passive flux of ammonium ions is likely to be greater in marine animals. There is also evidence for the active excretion of ammonium ions in the mudskipper, Periophthalmodon schlosseri (Randall et al., 2000).

Table 8 Ammonia connections to other air issues

Air issue	Major atmospheric gases	Ammonia role	Sensitivity
Stratospheric ozone depletion	Chlorofluorocarbons, hydrochlorofluorocarbons	Too reactive aspecies forstratospheric chemistry	Minimal
Climate change	Carbon dioxide, nitrous oxide, methane, water	May influence globalaerosol concentrations	Minimal
Ground-level ozone (smog)	Oxides of nitrogen, volatile organic compounds	Participates innitrogen chemistry	Minimal
Acid deposition	Sulphates, nitrates, ammonia	Soil/water acidifierthrough nitrogen cycle;participates in sulphurdioxide oxidation byozone	Moderate
Hazardous air pollutants (toxics)	Organic compounds and heavy metals	May enhance particulatematter pathway fordispersion of hazardousair pollutants	Moderate
Particulate matter (PM2.5 and PM10)	Sulphates, nitrates, ammonia, volatile organic compounds	Reacts with acid gasesto form hygroscopic salts	High

There is no clear evidence that water pH is modulating toxicity in marine species. It is possible that, because of the increased ammonium ion permeability, the relationship between water pH and ammonia toxicity is minimal. That is, there is no a priori reason to assume that pH will modulate ammonia toxicity in the marine environment. There is a paucity of data on the effects of water pH on ammonia toxicity in the marine environment.

2.5.3 Accumulation in aquatic organisms

Accumulation of ammonia in the body can be due to either the inability to excrete or convert nitrogenous wastes or a net influx of NH₃ from the environment. Externally, the concentration of NH₃ , rather than NH₄⁺ , is of concern, as biological membranes are permeable to NH₃ but much less so to NH₄⁺. Consequently, NH₃ , but not NH₄⁺ , diffuses readily across the external surface into the body. As a result, if NH₃ levels are high in the environment, ammonia levels in exposed animals increase as well. In acid water, nearly all ammonia is as NH₄⁺ , and the rate of ammonia entry into the fish is low. As pH increases to more alkaline conditions and water pH approaches the pK (9.2-9.5) of the ammonia/ammonium ion reaction, toxicity increases significantly for many species due to the shift in equilibrium to the more diffusable NH₃ form. Water of pH above 9.5 can be toxic, even though it contains little or no ammonia, because ammonia levels rise to toxic levels in the fish as a result of impaired excretion.

2.5.4 Factors affecting the aquatic toxicity of ammonia

Several factors have been shown to modify the acute toxicity of ammonia to freshwater organisms. Some factors alter the concentration of NH₃ in the water by affecting the aqueous ammonia equilibrium, while other factors affect the toxicity of NH₃ itself, either ameliorating or exacerbating its effects. Factors that have been shown to affect ammonia toxicity include temperature, pH, dissolved oxygen concentration, ionic concentration, previous acclimatization to ammonia, fluctuating or intermittent exposure, carbon dioxide concentration, salinity and the presence of other toxic substances. The best studied of these is pH; the acute toxicity of NH₃ has been shown to decrease as pH decreases (becomes more acidic). Data on temperature effects on acute toxicity are limited and variable; the U.S. EPA (1998) recently released revised water quality guidelines for ammonia for which they reviewed the data on temperature. The effects of dissolved oxygen are probably more important than the effects of temperature, with increased toxicity at lower dissolved oxygen concentrations (Thurston et al., 1981a). All of these factors may come into play in any water body. The pH of most rivers fluctuates with season, as does temperature. Dissolved oxygen will inversely follow the temperature variations, with less oxygen dissolved at high temperatures, exacerbating the toxicity effect from temperature. In Canadian waters, pH values usually rise in summer as the temperature increases and the dissolved oxygen content decreases. Downstream of municipal outfalls, there is often an oxygen sag as nitrification of ammonia and other biological processes use up the available oxygen, making the in-plume region more hazardous for organisms.

2.5.4.1 pH

The toxicity of aqueous solutions of ammonia and ammonium compounds to fish has been attributed to NH₃ present in the solution. The pH correlation with toxicity of ammonia was assumed to be based on the aqueous ammonia equilibrium. Thurston et al. (1981b) tested the toxicity of ammonia to rainbow trout (O. mykiss) and fathead minnows (Pimephales promelas) in 96-hour flow-through bioassays at different pH levels within the range 6.5-9. Results showed that the toxicity of ammonia, in terms of NH₃ , increased at lower pH values and could also increase at higher pH values. It was concluded either that NH₄⁺ exerts some measure of toxicity or that increased hydrogen ion concentration increases the toxicity of NH₃ . The U.S. EPA (1998) reviewed the extant toxicity data and came to the conclusion that "all of the datasets show a strong trend of total ammonia LC₅₀ s decreasing with increasing pH." This confirms the concept that ammonia is more toxic at basic pHs.

2.5.4.2 Temperature

Information on the correlation between temperature and toxicity of ammonia is varied, but the two appear to have an inverse relationship. The toxicity of ammonia is greater at colder temperatures, the reverse of what would be expected based solely on the aqueous ammonia equilibrium. After the U.S. EPA (1998) reviewed the data for their recent water quality criterion document on ammonia, they concluded that temperature had a minor effect on toxicity and decided that they would not use it in their calculation of a water quality criterion. Thurston and Russo (1983) reported an inverse relationship between temperature and toxicity for rainbow trout (O. mykiss) over the temperature range 12-19ºC. Thurston et al. (1983) reported a similar decrease in toxicity with increasing temperature in fathead minnow (P. promelas) over the temperature range 12-22ºC. A similar relationship was found by Reinbold and Pescitelli (1982) in rainbow trout, bluegill (Lepomis macrochirus) and fathead minnow, while Colt and Tchobanoglous (1978) found a similar relationship in channel catfish (Ictalurus punctatus).

At a temperature of 19°C and a pH of 8.5, it takes less than 0.4 mg total ammonia/L to generate a potentially toxic condition, while at 19°C and pH 7, it takes over 11 mg/L. At a temperature of 4°C and pH 8.5, it takes just over
1 mg total ammonia/L to generate this condition, while at 4°C and pH 7, it takes over 35 mg/L (Emerson et al., 1975).

2.5.4.3 Dissolved oxygen concentration

The dissolved oxygen concentration of water has long been known to affect the toxicity of ammonia to fish (Merkens and Downing, 1957; Vamos and Tasnadi, 1967; Alabaster et al., 1979). Thurston et al. (1981a) conducted a detailed study of this phenomenon and showed the potential impacts of reduced dissolved oxygen levels on the acute toxicity of ammonia. The 96-hour LC₅₀ of un-ionized ammonia to rainbow trout (O. mykiss) was tested in various concentrations of dissolved oxygen, from 2.6 to 8.6 mg/L. The former concentration was the lowest at which 90% or more of the control fish survived. There was a positive linear correlation between LC₅₀ and dissolved oxygen over the entire dissolved oxygen range tested: ammonia toxicity increased as dissolved oxygen decreased. Un-ionized ammonia LC₅₀ values were also computed for 12, 24, 48 and 72 hours: the correlation with dissolved oxygen (DO) was greater the shorter the time period. The 96-hour LC₅₀ values varied from 0.7 mg/L at 8.6 mg DO/L to 0.3 mg/L at 2.6 mg DO/L. The estimated correlation coefficient was 0.93, with an estimated regression line of LC₅₀ = 0.1903 − 0.06712(DO) (Thurston et al., 1981a).

The analysis of dissolved oxygen versus LC₅₀ over the entire 96-hour test period showed a clear trend: the shorter the time period, the more pronounced the positive relationship between acute toxicity and dissolved oxgyen. This suggests either that individual fish that were sensitive to ammonia succumbed early or that those fish that do survive become increasingly acclimated to ammonia and oxygen conditions. These tests show that any reduction in dissolved oxygen reduces the tolerance of rainbow trout fingerlings to acutely toxic concentrations of ammonia: the estimated tolerance at 5.0 mg DO/L is 30% less than at 8.5 mg DO/L.