Contaminants can have a direct effect on amphibian populations by impairing physiological processes or resulting in mortality. In contrast to indirect effects, discussed later, direct effects are often easier to identify and to establish causal links.
40.3.1.1 Field Studies Showing Population Effects
Most of the studies that have examined the direct effects of contaminants on amphibian populations have dealt with acid deposition. Some of these studies also examined correlates of acidification including sulfates, dissolved aluminum and other metals, alkalinity, and cations. Many have been correlative in nature, i.e., drawing relationships between wetland chemistry and presence of species or relative abundance, but several have been experimental.
One correlative study examined 35 wetlands in Pennsylvania.42 The number of egg masses deposited by Jefferson salamanders (Ambystoma jeffersonianum) was positively correlated with pH and alkalinity and negatively correlated with aluminum concentration. In the same study, egg mass deposition by spotted salamanders (A. maculatum) correlated positively with pH and negatively with total cations (Na, K, Mg, Ca) and specific conductance. Another correlative study in Ontario, Canada found only weak relationships among species richness of amphibians and water character- istics among 180 wetlands.43 Although species richness correlated negatively with chloride, mag- nesium, and turbidity, multiple regression analysis only accounted for 19% of the variance in richness. However, those wetlands were well buffered and not prone to acidification.
In a field experiment in England, variation in egg fertility could be explained primarily by the concentration of monomeric aluminum with additional explanation provided by the concentration of zinc, silicon, and molybdenum.44 Monomeric aluminum also accounted for the greatest amount of variance in the frequency of abnormal yolk plugs in eggs. Aluminum is the most common metal in the earth’s crust, but acidic conditions cause it to dissolve and form toxic species of which monomeric (Al3+) is among the most toxic to amphibians.
Horne and Dunson45 used outdoor mesocosms to test the effects of low pH, dissolved organic carbons (DOC), and metals on survivorship in three species of amphibians. Exposure concentrations were set to reflect typical conditions in central Pennsylvania (pH 4.5 to 5.5, DOC 15–30 mg/L).
The time required for wood frog (R. sylvatica) larvae to metamorphose increased with high DOC and low pH. Delayed metamorphosis can be disadvantageous to species such as R. sylvatica that inhabit temporary wetlands by increasing the risk of dessication. Survival of A. jeffersonianum and A. maculatum was reduced by pH and metal concentrations in the same study.
The relative abundances of northern cricket frogs (Acris crepitans) and gray treefrogs (Hyla versicolor) were significantly lower in experimental wetlands that had been acidified to a pH of 5 to 5.5 than in wetlands with a pH range of 6 to 6.8.46 In this study, all the experimental wetlands were naturally colonized by four species of anurans from surrounding wetlands. For H. versicolor, there was a soil × pH interaction in that the greatest effects of pH were seen in wetlands with a clay base rather than a loam base. The soil effect could be explained by the fact that clay soils had approximately twice the concentration of metals as the loam soils and that acidification of sediment and water facilitates metal dissolution. No pH or soil effects were observed in green frogs (R.
clamitans) or southern leopard frogs (R. sphenocephala).
Some amphibians, such as the salamander genus Plethodon, spend their entire life cycle on land. Yet these species may also be limited by acid deposition. The distribution and population density of the red-backed salamander (P. cinereus) was restricted by soil pH ≤ 3.7.47 While this was a very acidic soil, it is not uncommon in coniferous and deciduous forests in the Northeast where humic acids combine with atmospheric deposition of acid-causing anions.
The effects of acidification on amphibian health and populations are well known, and acid deposition certainly may be a contributing factor to population declines in some areas. However,
as stated by Rowe and Freda,48 no studies have documented specifically that acid deposition has been involved with any long-lasting decline in amphibian populations. Hazards due to acid depo- sition would be greatest where subsoils are granitic, poorly buffered, with low acid-neutralizing capacity (ANC) and where anthropogenic sources of acid-causing anions such as sulfates and nitrates are downwind. Within North America, these areas tend to be most abundant in the northeastern United States and eastern Canada. The Rocky Mountains have poorly buffered soils but lack an abundant source of acid-causing anions. Florida has many ponds and wetlands with low pH, but the source of the acidity appears to be organic acids rather than anthropogenic sources,49 and the amphibians have probably adapted to low pH over long evolutionary periods. Bradford et al.50 did not find any relation between ANC and populations of anurans in the Sierra Mountains, and they did not find any regional differences in ANC or other water-chemistry parameters related to acid deposition. They concluded that acid deposition was not important in the decline of these anurans.
There are fewer experimental or correlative studies with other contaminants and amphibians under natural conditions. Another pervasive potential source of contaminants is nitrogen fertilizers.
Nitrogen is used or found in nature in various forms from ammonium ion to ammonia, nitrite, and nitrate. Of these, ammonia and ammonium are the most toxic forms to amphibians, followed by nitrites and nitrates. The use of nitrogen fertilizers in North America is tremendous, with application rates sometimes exceeding 8 metric tons/km2 on agricultural lands and more than 72 million tons worldwide;51 greatest use in the United States is in the East and in the Midwest. Aqueous concen- trations may exceed 100 mg/L in ponds and between 2 and 40 mg/L in streams.
In one study in Britain, Oldham et al.52 released adult common frogs (R. temporaria) into plastic arenas situated in the middle of a field that had been treated with ammonium nitrate at 10.8 g/m.2 Adults began displaying signs of toxicity — increased rate of breathing — and one died;
although these signs were considered predictive of death, frogs were removed before they suc- cumbed. Based on laboratory data, the authors estimated that the EC50 for R. temporaria was 6.9 g/m,2 well below the application rate. Fortunately, the toxicity of ammonium nitrate disappears rapidly as the solution is absorbed by moist soil. However, laboratory toxicity tests (see below) frequently conclude that lethal concentrations are below anticipated ambient levels.
A 20-year decline in amphibian numbers in Poland was ascribed to high levels of nitrogen in surface waters coming from the use of fertilizers on surrounding agricultural lands.53 In an agri- cultural study area in Ontario, Bishop et al.54 determined that habitat loss and nitrogen from fertilizers were more important than pesticides in restricting amphibian survival and species rich- ness. A recent review of the subject51 concluded that the problem of nitrogen-based fertilizers is extensive and toxic enough to represent one of the most pervasive contaminant threats to amphibian survival in North America and perhaps elsewhere. Of more than 8500 water samples collected from states and provinces around the Great Lakes, 19.8% had nitrate concentrations that exceeded those found to produce sublethal effects in amphibians. Sublethal and lethal effects can be caused by nitrate concentrations ranging from 2.5 to 100 mg/L, depending on species and study.
Chlorpyrifos is a highly toxic but widely used pesticide. In June 2000, the U.S. Environmental Protection Agency ordered a phase-out of chlorpyrifos use in several applications because of human health risks.* The pesticide is also very highly toxic to aquatic organisms, having an LC50 of around 9–15 àg/L. A field study by Moulton55 found adverse effects, including mortality in adult and larval Hyla femoralis, after exposure to environmentally realistic concentrations of chlorpyrifos.
Mazanti56 conducted a combined field and laboratory study on the effects of two commercial pesticide formulations: Lorsban 4E (44.9% chlorpyrifos) and Bicep II (27.4% atrazine, 35.6%
metolachlor). Both formulations are used extensively in agriculture, with Bicep II sprayed as a preemergent herbicide for broadleaved plants and Lorsban 4E applied about 3 weeks later to control insects. In the laboratory, a combination of both pesticides was far more toxic than either pesticide alone, and a high dose (2.0 mg/L atrazine, 2.5 mg/L metolachlor, 1.0 mg/L chlorpyrifos) killed
* http://www.epa.gov/pesticides/op/chlorpyrifos-methyl/summary.htm.
100% of H. versicolor tadpoles within 11 days. A low combined dose and high individual doses of insecticide or herbicide alone caused lethargy and reduced growth but did not increase mortality over controls. In naturally colonized, experimental wetlands, low-herbicide/low-insecticide (0.2 mg/L atrazine, 0.25 mg/L metoachlor, 0.1 mg/L chlorpyrifos) and high-herbicide/low-insecticide treatments resulted in almost complete, short-term extirpation of H. versicolor and R. clamitans tadpoles compared to controls (Figure 40.1). By the end of the summer, however, these prolonged breeding species had repopulated the treated wetlands. If species with contracted or explosive breeding seasons had been similarly affected, their reproduction would have been severely curtailed for the year.
A recent study implicates chlorpyrifos and other insecticides in the declines of ranid populations in the Sierra Mountains of California. The Central Valley of California is an intensely agricultural region that sits between the Pacific Coast and the Sierra Mountains. The most severely impacted sites for R. aurora, R. mucosa, and R. boylii are in the mountains east and above the Central Valley in what appear as pristine wetlands. However, 60% of California pesticides used in agriculture, thousands of kilograms of active ingredient pesticides, are aerially sprayed on the croplands of the Central Valley every year.* A large proportion of these pesticides are organophosphorus insecticides such as chlorpyrifos, malathion and diazinon which inhibit cholinesterase.
Figure 40.1 Growth (x ± SD) of gray treefrog (Hyla versicolor) tadpoles through time when exposed to low and high concentrations of the insecticide Lorsban (active ingredient is chlorpyrifos), low and high concentrations of the herbicide Bicep II (active ingredients atrazine and metalochlor) and low concentrations of both insecticide and herbicide. See text for concentrations. The primary size difference at 28 days is due to a lack of metamorphosis in high and low insecticide rather than to growth per se. All tadpoles exposed to combined high herbicide and high insecticide died by 11 days.
* http://www.cdpr.ca.gov/docs/pur/purmain.htm.
Other highly toxic pesticides such as endosulfan and trifluralin are also sprayed in large quan- tities. Pseudacris regilla occupies many of the same habitats as the declining ranid species but its numbers are not nearly as depressed as the ranids. Using P. regilla as a sentinel species, the authors found that cholinesterase activity was significantly depressed in tadpoles in the Sierra Nevada downwind of the Central Valley compared to coastal populations.199 The samples with the lowest cholinesterase values came from Yosemite and Sequoia National Parks where mean activity level was often less than 50% of reference means on the coast. In the mountains, more than 83% of the tadpoles and adults collected from Lake Tahoe region had measurable concentrations of endosulfan and 50% of those sampled in Yosemite and Sequoia National Parks contained diazinon or chlor- pyrifos residues (Figure 40.2). Both endosulfan and the organophosphorus pesticides have half-lives measured in days so detection of these pesticides in P. regilla signified very recent exposure.
Figure 40.2 Mean (± SD) concentrations (a) and frequencies of detection (b) for chlorpyrifos, diazinon and endosulfan in Pacific treefrog (Hyla [Pseudacris] regilla) adults and tadpoles collected from Cali- fornia, 1999.
Fenitrothion, an organophosphorus insecticide, used to be sprayed widely to control spruce budworm (Choristoneura fumiferana). The relative abundance of mink frogs (R. septentrionalis) among New Brunswick ponds was inversely related to the intensity of spraying in forest ponds.57 Ponds that had been sprayed in 3 of the preceding 4 years had significantly lower relative abundances of frogs than ponds that had been sprayed less often. In addition, the percent cover of submerged vegetation related positively to frog abundance, perhaps because it provided shelter from direct exposure to the insecticide.
Although DDT and most other chlorinated hydrocarbon pesticides are no longer used in developed countries, these compounds have been associated with die-offs of amphibians and could still be having some effect in Third World countries, where they are being used. For example, DDT was applied to forests to control tussock moths (Orgyia pseudotsugata). A die-off of western spotted frogs (R. pretiosia) occurred following a spraying of DDT with fuel oil as solvent at a rate of 0.72 kg DDT/ha. An unknown percentage of the adult frogs were dead, and those that were found dead had 5–10 times the concentration of DDT derivatives on a lipid basis than those that were still alive.58 There was 100% egg mortality in R. pipiens inhabiting a wetland adjacent to cropland that had been sprayed with aldrin.59 A mixture of endrin, aldrin, dieldrin, and toxaphene at 2.7 kg/ha used in mosquito control killed all the R. catesbeiana inhabiting a wetland.60
Whereas the number of documented cases of die-offs in amphibians due to contaminants is sparse, certainly the vast majority of cases either are unreported, or causal factors are not identified.
Both acid deposition, with the elevated concentrations of metals produced by increased acidity, and pesticides can cause local reductions or extirpations of amphibian communities, and in some cases, they may be related to regional declines.
40.3.1.2 Toxicity in the Laboratory at Ecologically Relevant Concentrations
As with field studies, more research on amphibian ecotoxicology has been conducted in con- junction with acid deposition than with any other contaminant. Studies have examined direct effects of elevated hydrogen ion concentration plus associated effects due to the dynamics of dissolved aluminum; mitigating effects of water hardness, dissolved organic content, and organic acids; effects of temperature; and other factors. Endpoints in these studies have included survivorship at various life stages (egg, embryo, tadpoles), ion regulation, metal uptake, respiration and energy conversion, and behavioral responses. More complete reviews of the effects on aquatic biota in general have been published.61,62 Here we focus on the direct effects of acid deposition and pesticides on factors that affect amphibian population dynamics — survivorship and reproduction.
Readers are referred to several reviews of the effects and tolerance limits of amphibians to acidity.48,61,63–65 Sensitivity to pH varies among species. Among the most sensitive species are the natterjack toad (Bufo calamatia), with a lower pH limit (at which substantial or total mortality occurs) of 4.5 and an upper effects level (where sublethal effects or mild mortality can be seen) of pH 4.7. In comparison, the more tolerant pine barrens treefrog (Hyla andersoni) can tolerate pH as low as 3.1.61,65 Most amphibian species are sensitive to pH = 4.5 or 4.6. In general, the toxicity of pH decreases as water hardness increases. Dissolved organic carbon can ameliorate the toxic effect of pH and Al but can itself be toxic to amphibians66,67 and can complex with Al to alter its availability. The speciation and toxicity of aluminum is complex as pH drops below 5.5, and three response categories to Al have been identified.68 These range from sensitivity only to pH below 4.5 to sensitivity to pH and aluminum between 4.5 and 5.5, to Al sensitivity only at pH > 4.5 and even an amelioration of pH toxicity by Al at pH < 4.5. Because of these dynamics, toxicity due to pH, Al, and DOC often cannot be totally distinguished under natural conditions.61
As a very brief synopsis of the relatively extensive literature on acid deposition and amphibians, embryos tend to be the most sensitive life stage, especially with elevated concentrations of Al69 that induce a characteristic curling of the tail due to alterations in the vitelline membrane of the
egg. Serious impairment of hatching due to ecologically realistic levels of pH and Al are well documented in several species.42,70–72 Similarly, tadpole mortality increases significantly at pH below 5.0.73–76 Acute or episodic exposure to reduced pH and elevated Al, such as what might occur with a rapid snowmelt, tends to be more toxic than chronic exposure.76 There is some evidence that amphibians may adapt physiologically or genetically to long-term exposure to reduced pH.77 Hardly anything is known about the effects of pH or Al on adult amphibians.48
Among the plethora of other contaminants, many occur in the environment at lethal concentra- tions as determined by laboratory LC50 tests (Table 40.1). In this table, ambient concentrations for most of the chemicals come from Eisler.78 However, organophosphorus, carbamate, and pyrethroid pesticides are not persistent, and environmental fate studies of these compounds often result in measurements below instrument detection levels. Thus, when measurable concentrations are detected, they may be a fraction of the maximum to which the amphibians were exposed. Because few reliable data exist, I used label application instructions79 to derive a rough approximation for some of these; amount per area was converted to concentrations in water based on wetlands with average depth of 0.5 m, under the presumption that most amphibians lay their eggs in shallow water. Realized concentrations can vary considerably from these estimates due to actual depth of eggs masses, applicator deviations from label application rates, and additional pesticides entering a wetland through runoff or wind. Lethal toxicity data come from many sources including some recent reviews.80–82
The table has several limitations and is intended only to illustrate that there are many compounds that potentially are found in the environment at higher than toxic concentrations. Limitations include the previously stated weaknesses in relying on laboratory toxicity tests to reflect natural exposures and differences in testing methodology among labs that may affect results. Life stages and duration of tests also are important because of differential sensitivity among stages and because length of exposure may greatly alter LC50 values. Whereas many laboratory toxicity tests end after 24 or 96 h, free-ranging animals may be exposed for much longer durations or be subject to repeated exposures through time. Third, the table does not report the many instances of nondetects for specific contaminants. The relative frequency of nondetects is biased against totally anthropogenic and short-lived chemicals compared to molecules such as metals that are persistent and also naturally occurring. Fourth, the presence of a substance in sediment may not reflect its bioavailability; it may be bound to sediment particles. Fifth, acute lethality, as demonstrated by the LC50 values, is the ultimate and strongest effect of contaminant exposure. Sublethal or chronic effects, which may reduce the survivorship or reproductive potential of an organism, can occur at concentrations much lower than those necessary for acute toxicity. In view of these caveats, perhaps the greatest value of the table is to clearly indicate the great need for ecologically relevant, cause-and-effect data on amphibian populations and contaminant exposure.
Despite these limitations, a few salient points may be developed:
• In contrast to the 75,000 chemicals manufactured in the United States each year, not including pesticides and the 3000 that are considered high-production-volume chemicals (> 1 million pounds produced),83 toxicity data for amphibians are known for only a very, very small proportion.
• Relationships between ambient concentrations of conventionally used pesticides and effects are poorly known.
• Ambient concentrations are often several times greater than the lethal concentrations, as determined by lab tests, and this difference extends across all categories of contaminants for which there are data.
• Formulations can make a difference. For example, at a given concentration of azinphos-methyl, the formulation Guthion 2S may result in amphibian die-offs, whereas use of the alternative formulation Guthion might not.
• Different life stages of amphibians vary in their sensitivity to the same contaminant. For example, tadpole Bufo arenarum are more sensitive than embryos to parathion, and adult R. pipiens are less sensitive to TFM than either tadpoles or embryos.