Factors Contributing to First-Year Recruitment Failure of Fishes in Acidified Waters with Some Implications for Environmental Research

Although fish kills caused by acidic precipitation have seldom been reported (Harvey and Lee 1982; Reader and Dempsey 1989), many populations of fishes in freshwater lakes have been severely affected by environmental acidification (Beamish 1974a; Haines 1981; Rosseland et al. 1986). The paucity of such reports probably does not mean that fish kills rarely occur but rather that they go unnoticed because the fish involved are small (embryos and larvae; Craig and Baksi 1977; McCormick et al. 1989a) or because observation is obscured by ice and snow cover (Eaton et al. 1992; McCormick and Jensen 1992). Typically, acidification progresses gradually over a period of years, and fish losses are usually restricted to the most sensitive early life stages or, in more northerly latitudes, to overwintering first-year juveniles (McCormick and Jensen 1992). Usually only the most severe episodic exposures result in mortality of larger individuals that may be noticed and reported as fish kills (Muniz and Leivestad 1980).

Therefore, a population can persist within an environment undergoing anthropomorphic acidification without a demonstrated fish kill, but the population may one day no longer be present. The loss of a population, in such cases, becomes complete with the disappearance of the last of the more acid-resistant larger individuals, but it is the ultimate result of successive years of age-1 recruitment failure (Beamish et al. 1975).

Consequently, though “acid rain” fish kills are seldom reported, populations have been affected and sometimes eliminated by environmental acidification. Most reports of adverse effects of environmental acidification on fishes are, therefore, observations of the absence of populations from habitats where they had once been present or of populations made up entirely of older age-groups (Beamish and Harvey 1972; Beamish 1974a; Howells 1985; Schindler et al. 1985; Schindler 1988). Such population declines are most often attributed to reproductive failure (Beamish 1976). If reproduction is broadly defined as the process by which one generation produces another successfully reproducing generation, reproductive failure seems, indeed, to be the case. But if reproduction merely implies yearly production of a new set of individuals, reproduction may be successful but mortality may occur at some time during the first year of life and the year-class is lost (Mohr et al. 1990; Eaton et al. 1992; McCormick and Jensen 1992).

What is responsible for the peculiar vulnerability of first-year recruitment? The purpose of this paper is to integrate the findings of several studies, emphasizing those of the present authors, that describe the particular vulnerability to acid precipitation of biological processes necessary for completion of a fish's first year of life. (For pertinent reviews, see Peterson et al. 1982 and Morris et al. 1989.) For this discussion, biological processes that occur during life stages from gametogenesis through the following spring as age-1 individuals will be stressed. For convenience, the presentation will be ontogenetic.

Because poorly buffered waters dominate those that become acidified by atmospheric fallout, environmental acidification, as herein presented, will assume poorly buffered waters and thus, low Ca⁺² availability. In addition, because metal concentrations, particularly aluminum, tend to be elevated in acidified natural water (Haines 1981; LaZerte 1984; Hutchinson and Sprague 1986), this presentation will assume aluminum mobilization into the water column (Baker 1982; Neville 1985; Booth et al. 1988; Eaton et al. 1992). Also, because episodic pH fluxes tend to be more extreme in stream environments than in lakes and because our research emphasized a lake environment, some of the following discussion may be more applicable to lakes than to streams.

Gametogenesis

Working with the fathead minnow Pimephales promelas in the laboratory (McCormick et al. 1989a; Leino et al. 1990) and in artificial streams (McCormick et al. 1989c) and rock bass Ambloplites rupestris in Little Rock Lake (Leino and McCormick 1993b), we have shown that both oogenesis and spermatogenesis may be retarded or even halted by environmental acidification. These results are similar to those reported by others including Mount (1973) and Ruby et al. (1977, 1978). Impaired gametotogenesis occurred at pH values that were not lethal to the adults (Ruby et al. 1977; Ruby et al. 1978; Freeman and Sangalang 1985; McCormick et al. 1989a). When oogenesis is interrupted, the process often proceeds through vitellogenesis (deposition of yolk in the oocyte) but ceases before ovulation (Stacey 1984; McCormick et al. 1989c). However, inhibited vitellogenesis has also been reported (Ruby et al. 1977; Leino et al. 1990; Leino and McCormick 1993b). The result of interrupted oogenesis is an accumulation of unspawned or incompletely matured oocytes that eventually become atretic and are reabsorbed (McCormick et al. 1989c; Leino et al. 1990; Leino and McCormick 1993b). Vitellogenesis suppression may be the result of altered cortisol secretion, reduced gonadotrophin secretion, or Ca⁺² insufficiency (Freeman and Sangalang 1985; Tam et al. 1987; Mount et al. 1988). Low serum [Ca⁺²] may be a consequence of accelerated ion losses or of inhibition of ion uptake in acidified waters with low [Ca] and, in some situations, with elevated [Al], of decreased mobilization of calcium from somatic storage sites, such as scales (Leino and McCormick 1993a). Calcium is required for hepatic production of vitellogen, which is transported to the ovary to form yolk (Urist and Schjeide 1961; Ruby et al. 1977). Lockhart and Lutz (1977) were among the first to report that the usual blood Ca⁺² elevation during the vitellogenic phase of oogenesis failed to occur in fish from an acidified environment where reproductive failure was evident, but the pervasiveness and effects of this phenomenon are still unclear (see Mount et al. 1988; Valtonen and Laitinen 1988; Munkittrick, 1991; Vuorinen and Vuorinen 1991; Vuorinen et al. 1992).

With calcium insufficiency, other calcium-related functions, such as gill membrane permeability, ionoregulation, nerve conduction, and muscle control, may also be adversely affected, (Carafoli and Penniston 1985; Dunstan et al. 1985; Playle et al. 1989; McCormick and Jensen 1992). All of these factors could contribute to increased vulnerability of the female during the spawning cycle, when stored calcium is in particular demand. An extreme example would occur in anadromous species experiencing enhanced ionoregulatory challenges because spawning requires migration from ion-rich salt water to ion-poor acidified freshwater (Muniz and Leivestad 1980). The ionoregulatory difficulties encountered by such fish are further exacerbated by diminished ion intake due to the fasting (Phillips et al. 1956) that usually accompanies spawning migration. Fasting has been linked to lowered activity of Na⁺-K⁺-ATPase, which is necessary for ionoregulation (Virtanen and Soivio 1985; Potts and McWilliams 1989).

Decreased spermatogenesis has also been observed at pH levels approximately the same as those inhibiting oogenesis (Ruby et al. 1978; Leino et al. 1990). However, the effect of a partial reduction in sperm production on reproductive success is not quantitatively known. In fathead minnows chronically exposed to low pH in the laboratory under breeding conditions, the numbers of males with mature testes were reduced by 25% at pH 5.5 and by 100% at pH 5.2 (Leino et al. 1990). This suggests that loss of sperm production could reduce zygote production (even should egg production be unaffected) and that the combined male and female effect could be substantial, which agrees with the conclusion of Ruby et al. (1978) that both female and male gamete production should be examined when assessing the impacts of depressed pH on reproduction.

Spawning and Fertilization

Spawning can take place at pH levels where recruitment failure occurs (McCormick et al. 1989a; Eaton et al. 1992). In serial spawners (spawning several times throughout the year), intervals of restored higher pH might be adequate for population maintenence. For example, a few days of restored near-neutral pH enabled renewed spawning in fathead minnows after this activity had been halted by exposure to low pH (J. H. McCormick, unpublished data). However, anything that lengthens the interval between spawnings will ultimately reduce seasonal fecundity. This is particularly important to piscivorous species whose well-being is dependent on the fecundity of prey species, which are often serial spawners (Mills et al. 1987).

Daye and Glebe (1984) and Duplinsky (1982) have reported adverse effects of low pH on sperm motility. In vitro studies of the spermatozoa of chinook salmon Oncorhychus tschawytscha have shown that acid pH levels tend to cause accelerated swimming rates and shortened motility times (J. H. McCormick, unpublished data). These effects may contribute to reduced fertilization (Ginzburg 1972; Daye and Glebe 1984).

Courtship and spawning appear to be less vulnerable to acidification than gametogenesis. We have observed male fathead minnow courtship behavior and spawning coloration at pH levels well below those that caused death at the larval stage. Spawning of largemouth bass Micropterus salmoides and rock bass was also observed in the acidified water (pH 4.7) of Little Rock Lake, Wisconsin, where hatching success was severely inhibited or nonexistent (Eaton et al. 1992).

Largemouth bass and rock bass were found to produce viable zygotes at pH levels that were lethal to the newly hatched larvae (Eaton et al. 1992). Viability was confirmed by transferring zygotes from the acidified spawning environment of Little Rock Lake's acidified basin to its reference basin (pH 6.1), where they hatched and produced larvae capable of surviving for 14 d or more. This observation is consistent with Kennedy's (1980) report of apparently normal fertilization but poor hatching success of the eggs of lake trout Salvalinus namaycush from an acidified environment. Thus, though spawning and fertilization may be affected by low pH in some instances, these processes generally may not be the most critical in establishing a new age-1 year-class (Eaton et al. 1992; McCormick and Jensen 1992).

Embryogenesis and Hatching

At pH levels where spawning and fertilization are successful, embryogenesis is usually also successful (McCormick et al. 1989a; Mohr et al. 1990; Eaton et al. 1992). There have, however, been reports of acid-engendered delays in hatching in some species (Peterson et al. 1980; Waiwood and Haya 1983; Ingersoll et al. 1985; Chulakasem et al. 1989; Li and Zhang 1992). Such delays have been attributed to inhibited activity of the hatching enzyme, chorionase, or to a relatively inactive embryo as hatching nears (Yamagami 1973; Runn et al. 1977; Peterson et al. 1980; Nelson 1982). For fathead minnows (McCormick et al. 1989a), as well as largemouth bass and rock bass (Eaton et al. 1992), hatching was not significantly delayed at pH levels where hatching success was unaffected. However, if hatching is delayed, it might increase the hazard to the embryo because of prolonged exposure to (1) chorionase, which is thought to reduce embryo viability (Hayes 1930) and (2) predation by egg-eating predators. In Little Rock Lake we observed groups of yellow perch Perca flavescens distracting the nest-guarding male rock bass while other yellow perch consumed the eggs. Losses of embryos to yellow perch predation appears to be one of the most important agents determining reproductive success among the nest-building centrarchids in Little Rock Lake. Consequently, acid-engendered hatching delays in similar situations could conceivably have a major effect on recruitment because of increased egg predation.

The generally low vulnerability of the embryonic stage to the effects of environmental acidification is probably not due to resistance of the embryo itself but to the protection afforded by the ion-selective perivitelline membrane (Gray 1932; Krogh and Ussing 1937; Hayes 1949; Wedemeyer 1968; Ruby and Potts 1969). However, H⁺ is probably not restricted by this membrane (Peterson et al. 1980). It is more likely that the perivitelline membrane minimizes loss of essential ions from the embryo (Shephard 1987), at the same time restricting infiltration of toxic metals (Gray 1920) that are often elevated during environmental acidification.

Hatching through First Feeding

Hatching through first feeding encompasses the critical period (Hjort 1914; Braum 1967) for fish in natural environments. It is at this time that the fish must make the critical transition from endogenous nutrition to exogenous feeding. During this period, fish exhibit particular vulnerability to environmental acidification (Mount 1973; Johansson and Milbrink 1976; Craig and Baksi 1977; Daye and Garside 1977; 1979; McCormick et al. 1989a; Farag et al. 1993). Histomorphological examinations of yolk sacs, livers, and swim bladders of larval fish exposed to various acidic pH levels and aluminum concentrations provide evidence of altered physiological processes that could account for the particular sensitivity of this life stage (Leino et al. 1990). First, yolk energy reserves are not effectively utilized. Yolk consumption is slowed in newly hatched fish in acidic water (Menendez 1976; Korwin-Kossakowski 1988; Leino et al. 1988; 1990); possibly, this may produce starvation before the yolk is exhausted. Retarded yolk absorption would also delay onset of the free-swimming and first-feeding stage, resulting in increased time of relative immobility and vulnerability to predation. Second, liver glycogen of newly hatched fish from acidified water is often reduced. At the time when feeding usually begins in fish at near-neutral pH, the liver glycogen reserves remain abundant; in acid-exposed larvae, they may be exhausted (Leino et al. 1990). These observations suggest that bioenergetics is important to the survival of larval fish during this critical transitional stage. Third, swim bladder inflation is inhibited under acidic conditions (Korwin-Kossakowski 1988; Leino et al. 1988; 1990). First feeding in most fishes is synchronized with the swim-up stage and the swim-up stage with inflation of the swim bladder (Braum 1967). This association is probably not coincidental because the primary function of the swim bladder is to aid in achieving neutral buoyancy and more energetically efficient food searching behavior (Tait 1960). If low pH interferes with swim bladder inflation and if available energy is severely limited in acid-exposed larvae (Leino et al. 1990), then it is not surprising to find that the newly hatched larvae are particularly vulnerable to the effects of environmental acidification. In fact, this stage has most frequently been reported to have the greatest acid sensitivity (Craig and Baksi 1977; Nelson et al. 1988; McCormick et al. 1989a).

If, however, transition to exogenous feeding is successful, some of the adverse metabolic effects of acidification seem to be compensated for, and liver glycogen is soon restored (Leino et al. 1990). These individuals may then survive the first growing season without further peril from acid exposure. Nevertheless, the energy cost of physiological compensation for life in an acidic environment could lower growth rates when food is limited (Beamish 1974b; Shuter et al. 1989). Retarded growth may occur through lowered food conversion efficiency (Prosser and Brown 1961; Swenson et al. 1989), appetite suppression (McCormick et al. 1989c), or both. Reduced growth at low pH has, in fact, been reported, even in fish with ample food available (Menendez 1976; Muniz and Leivestad 1979; Mount et al. 1988). On the other hand, in some instances, growth is unaffected (Wiener and Hanneman 1982) or even greater (Rask and Raitaniemi 1988) in acidified environments, possibly because of density-dependent factors (Wedemeyer et al. 1984). However, when growth is retarded, increased size-dependent predation may reduce year-class strength (Larkin 1979).

Overwintering

Once the transition to exogenous feeding is completed, there may be little observable difference between fish from environments at low pH and those at near-neutral pH (Swenson et al. 1989, Leino and McCormick 1993b). However, a short growing season and an acidified environment may limit the accumulation of essential ions or energy reserves necessary to sustain the fish through its first winter (Shuter et al. 1980; Johnson and Evans 1990; Shuter and Post 1990; McCormick and Jensen 1992; Leino et al. 1993a). In Little Rock Lake, this effect was particularly noticeable in largemouth bass that failed to produce a year-class at a pH level of 5.1 (1987) but did so the following year (1988) when the pH was the same but the growing season was unusually long and warm (Eaton et al. 1992). Fall fingerling largemouth bass were substantially larger in 1988 than in 1987, which may account for the overwintering survival of some 1988 fish (see also LeCren et al. 1977).

Overwintering is the critical period for many northern fishes because energy and ion intake (feeding and active transport processes) at cold temperatures may fail to keep pace with metabolic demand and ion loss (Virtanen and Soivio 1985; McCormick and Jensen 1992), resulting in death. Death occurs because the individual does not possess sufficient energy reserves or (we emphasize) sufficient ion reserves (Leino and McCormick 1993a) to sustain it until the environment warms and feeding is resumed the following spring. Therefore, even though larger individuals have consistently been found to be less vulnerable to the direct effects of low pH than smaller ones (Robinson et al. 1976; McCormick and Naiman 1984; McCormick et al. 1989a), first-year juveniles at cold overwintering temperatures may be more at risk to environmental acidification, especially at northern latitudes, than newly hatched larvae present only at warmer spring temperatures.

Although size alone is not enough to successfully predict pH sensitivity (McCormick and Jensen 1992), the smaller members of the first-year juvenile cohort would be more vulnerable to overwintering mortality than the larger individuals (Shuter et al. 1980; 1989; McCormick et al. 1989c). Consequently, effects of acid water on growth during the first growing season can have a profound influence on overwintering mortality and thus recruitment to the age-1 year-class. After the first winter, survival is almost assured, and the threat of subsequent overwintering mortality diminishes with each succeeding growing season.

Individual size is important to survival success during chronic low pH exposure because of its relation to ionoregulatory homeostasis (Bernstein 1971; Mount et al. 1988; Wood 1989; McCormick and Jensen 1992). Size is significant because body mass relates to ion pools, to relative areas of ion exchange surfaces, to energy reserves, and to metabolism. Smaller individuals have greater relative gill (the major site of ion loss) surface areas (Muir 1969), higher metabolic rates (Prosser and Brown 1961), and smaller reserves of energy and ions to draw upon during overwinter fasting (McCormick and Jensen 1992).

The relative abilities of various species to maintain ionoregulatory homeostasis in laboratory acidification experiments was related to their relative recruitment success in Little Rock Lake. The poorest osmoregulator of the major Little Rock Lake fishes, the rock bass, was the first to suffer gill damage during laboratory exposures (Leino and McCormick 1993a) and the first to experience recruitment failure (and the only one to experience ionoregulatory failure in adult fish) as the pH of Little Rock Lake was lowered over several years (Eaton et al. 1992). The agreement between laboratory and field observations of osmoregulatory competency and age-1 recruitment also occurred with largemouth bass. Largemouth bass were slightly more competent osmoregulators than rock bass (McCormick et al. 1989b), and they were the next species to experience overwintering year-class failures in Little Rock Lake as the pH was reduced. This hierarchy continued with yellow perch, the most effective ionoregulator studied in the laboratory and the last of the Little Rock Lake study fish to suffer overwintering problems with pH reduction (McCormick et al. 1989b).

Other factors may contribute to low overwintering survival in particular acidified environments. Elevated aluminum concentrations tend to exacerbate gill damage and electrolyte insufficiency (Leino and McCormick 1993a), perhaps in part by interfering with Ca⁺²-dependent cell adhesion and metabolic functions (Wood and McDonald 1987; McCormick and Jensen 1992) and with adequate mobilization of Ca⁺² reserves (Leino and McCormick 1993a). In addition, species that endure long periods of drastically reduced food intake at low temperatures (e.g., rock bass and largemouth bass) tend to be less pH tolerant than species that can feed regularly under such conditions (Kwain et al. 1984; Virtanen and Soivio 1985; McCormick et al. 1989a). As a result of these factors, fish may become particularly vulnerable to acid pulses that occur at the end of winter when ion pools or energy reserves are depleted (Jefferies et al. 1979).

The intensity of the problems that fish encounter while enduring the harsh conditions of overwintering is, of course, related to latitude. These problems particularly involve species that spawn late in the spring and thus avoid exposure of their most sensitive early life stages to acid pulses during snowmelt. The benefit gained is at the expense of a shortened period of first-year growth. The resulting small juveniles have more difficulty surviving the winter at low pH as this season becomes longer in more northerly habitats (and at higher altitudes). Conversely, the threat of overwintering mortality diminishes in more southerly habitats. The prime concern in this situation is for pH effects on newly hatched prefeeding larvae.

Seasonal Synchrony of Life Stage and Physiological Adaptations to Naturally Occurring or Anthropogenic pH Fluctuations

A number of survival mechanisms have evolved among fishes (many undoubtedly due to environmental challenges other than acidification) that allow them to persist in acidified environments that would otherwise be untenable for one or more life stages. In some species, this includes spawning after the snowmelt has produced its usual acid pulse (Johnson et al. 1987). For example, spring acid pulses occur primarily in the nearshore epilimnion (Gunn and Keller 1986; Gunn et al. 1990; Porcella et al. 1995) as the cold (<4°C), acidic meltwater floats on the warmer (∼4°C), more dense main lake water until mixing and pH buffering occurs. At this time, 1-year-old and older largemouth bass (Coutant 1975) and most other species tend to be offshore and at greater depths, and shoreline spawning generally will not yet have occurred. Thus acid-sensitive newly hatched larvae are not exposed to the lowest pH in the annual cycle. Another life history pattern allows a longer growing season before the juvenile fish must endure their first winter at low pH, as in fall spawners whose eggs hatch in early spring. In this case, the eggs must survive a long exposure to cold temperatures and low pH but are protected by the ion-selective perivitelline membrane (Gray 1932; Krogh and Ussing 1937; Wedemeyer 1968).

Not all species have life histories that allow them to avoid the exposure of sensitive stages to seasonal acid pulses and the osmoregulatory challenges they present. Instead, some have developed pH tolerance through physiological means. For example, yellow perch spawn soon after ice-out, exposing the newly hatched larvae to the spring acid pulse before natural buffering can occur. Yet yellow perch are among the last species to be lost from environments undergoing acidification (Harvey 1980; Rahel and Magnuson 1983; Gunn and Keller 1984; Ek et al. 1995). Their tolerance is apparently due to an exceptionally effective ionoregulatory capacity. Histological studies of yellow perch have found distinct differences in the anatomy of the gills as compared with those of other less acid-tolerant species from the same environments, such as centrarchids (Leino et al. 1987a; McCormick et al. 1989b), and to other freshwater teleosts (e.g., Kikuchi 1977; McDonald et al. 1991; Perry 1997). Yellow perch gills have higher numbers of chloride cells, the principal ionoregulatory cells, than do gills of most fish. Besides being scattered individually over the gill surface as in other teleosts, in yellow perch, many of these cells are found in gland-like clusters (Leino et al. 1987a). These chloride cell clusters form a lumen that fills with a protective mucosubstance. This anionic mucosubstance appears to act as a buffer that titrates H⁺ and prevents high concentrations of H⁺ from damaging the chloride cells (see Philpott 1968). In addition, the polyanionic mucus may attract cations such as Na⁺ and Ca²⁺, increasing their concentration in the immediate vicinity of the membrane containing uptake channels for these ions (Menendez 1976). The importance of chloride cell-related mucosubstances is also demonstrated by their association with apical pits. These invaginations of the exposed surface of chloride cells develop in some freshwater species in response to acid stress. The pits also fill with a polyanionic mucosubstance, suggesting an adaptation to acid challenge that is similar to, but less extensive, than that seen in yellow perch (Leino and McCormick 1984; Leino et al. 1987b). The ability of chloride cells to remain viable increases the capacity to limit ion loss in the face of low pH, which is a hallmark of acid-tolerant species: acid-sensitive species exhibit chloride cell damage and death at low pH (Evans et al. 1988; Leino et al. 1990; Leino and McCormick 1993a). In addition, acid-tolerant species may have rapid chloride cell replacement rates that help to assure sufficient active cells to carry on the enhanced ion uptake required in an acid environment (Wendelaar Bonga et al. 1990).

Ordering of Life Stage and Species Sensitivity

Ordering the sensitivity of life stages and of species is helpful for estimating the impacts of various degrees of environmental acidification on fish communities. For example, we found the fathead minnow most vulnerable to acidification, followed by rock bass, largemouth bass, and yellow perch. Study of the direct effects of environmental pH on fish is relevant because observations involving Little Rock Lake (Swenson et al. 1989; Brezonik et al. 1993; Leino and McCormick 1993b) and other ecosystems (Beamish 1974b; Beamish 1976; Leivestad 1982; Leino et al. 1987b; Mills et al. 1987) have indicated that fish are at least as acid sensitive as other biota. That is, direct effects of acidification on fishes are probably as important, or even more so, than indirect effects due to food chain disruptions (Webster et al. 1992).

Our studies affirm the belief that, with equivalent exposure conditions, fish are most sensitive to acidification during the period between hatching and first feeding. However, several factors may combine to make other life stages the first to disappear. These situations can occur when episodes of extreme environmental stress affect even the more resistant life stages (Palmer et al. 1989; Leino et al. 1992; McCormick and Jensen 1992). As a consequence, any ordering of species by sensitivity is ecologically irrelevant unless it is linked to the exposures each species is likely to experience at each life stage in its specific environment.

We propose four categories of life history and environmental acid stress for the purpose of ordering life stage sensitivities (Figure 1): (1) fall spawners with late winter or early spring hatchings (coregonids and some salmonids), (2) early spring spawners with hatching occurring shortly after ice-out (Esocidae and percids), (3) late spring to early summer spawners with hatching after spring warming and pH buffering (Centrarchidae), and (4) summer spawners with repeated spawning throughout the warmer months (some Cyprinidae). In each category, acid-sensitive life stages are present at different times of the year and have different likelihoods that they will be exposed to the most acidic conditions. In addition, seasonal growth dynamics differ for each category, resulting in dissimilar effects of acidity on growth and resultant vulnerability to overwintering mortality.

Once the life history and environmental acid stress category has been identified, the relative ionoregulatory capacity of first-year juveniles of a species becomes more relevant (Trout and Dunson 1985; McCormick et al. 1989c; Eaton et al. 1992). As an example, species from northern habitats with a relatively low ionoregulatory capacity may survive environmental acidification by spawning after the acid pulse associated with snowmelt. But in doing so, they have a shortened growing season and, accordingly, will be small in size as fall fingerlings. Thus, these species are progressively more vulnerable to overwinter mortality as winter becomes longer. Alternatively, the same species at a more southern latitude have a longer growing season, which results in cold-induced fasting and ionoregulatory stresses being less hazardous. Moreover, there is no acidic snowmelt pulse at latitudes with no snow accumulation. However, acid pulses related to heavy seasonal rains are also a problem, especially for streams, and must be considered when evaluating the probable impact of acid precipitation (Harvey and Lee 1982; Baker et al. 1996).

Conclusions

The particular vulnerability of first-year fishes to environmental acidification has long been known and is probably the result of a number of factors (e.g., low pH and Ca levels, high levels of Al and other toxic cations) that interfere with their ability to maintain a positive bioenergetic or ionoregulatory balance under a specific set of environmental conditions. These factors exert their influence with greatest effect on newly hatched larvae when ionoregulatory capacities are least well developed and the change from endogenous to exogenous sources of nutrition is occurring or during winter when juveniles are most encumbered by seasonal stresses. Therefore, when attempting to estimate the effect of environmental acidification on fishes in a particular environment, it is necessary to consider the sensitivity of the life stage present at each point in the seasonal pH cycle, as well as other seasonal conditions of the specific environment. We suggest that essential data for this purpose would include (1) the timing of the embryo–larval period of the species of concern (examined through first feeding and initial growth to ensure successful transit through the critical first feeding period) and (2) overwintering sensitivity when the water body is at a high latitude or altitude. Sensitivity evaluations should be obtained under exposure conditions similar to those of the water body of concern at the time when that life stage will be present (i.e., temperature, calcium levels, and aluminum speciation). Ordering of sensitivity of overwintering juveniles can be determined by comparing the pH at which they lose ionoregulatory control when chronically exposed to winter conditions.

Success or failure of a species in a particular ecosystem depends on how well the physiological capabilities of the organism match up with seasonal changes in environmental conditions. The life stage present is linked to the season, and the season determines the severity of environmental stress.

The present discussion is based on studies performed specifically to understand the effects of environmental acidification on fish. We believe, however, that the challenges presented by other potential toxicants relative to the season, water chemistry, and life history stage of organisms merits increased attention in both laboratory and field studies.

Acknowledgments

The authors thank the reviewers and especially Charles Coutant for many helpful comments and suggestions regarding this manuscript.