nvimnme~| •A,-c•iv•s cont~mi~ion of

Arch. Environ. Contam. Toxicol. 17,449-457 (1988)

9 1988Springer-VerlagNew York[nc.

Effects of Atrazine on Freshwater Microbial Communities J. R. Pratt l*, N. J. Bowers**, B. R. Niederlehner**, and J. Cairns, Jr.** *School of Forest Resources, The Pennsylvania State University, University Park, Pennsylvania 16802 and **University Center for Environmental and Hazardous Materials Studies, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061

Abstract. A multispecies toxicity test system using naturally derived microbial communities on polyurethane foam substrates was used to evaluate the toxic effects of the herbicide atrazine. Both structural (e.g., protozoan species number, biomass) and functional (e.g., colonization rate, oxygen production) community responses were measured. Oxygen production and the ability of communities to sequester magnesium and calcium were the most sensitive measures of toxic stress due to atrazine ( m a x i m u m allowable toxicant c o n c e n t r a t i o n s [MATCs] = 17.9 ~xg/L). Dissolved oxygen was 33% lower, and there was 15% less calcium and magnesium in communities at and above 32.0 pog/L atrazine compared to controls. Species richness and estimates of biomass (total protein and chlorophyll a) were less sensitive (MATCs = 193) to atrazine. At the highest atrazine concentration (337 Ixg/L), species numbers were 30% lower than controls, and protein and chlorophyll a content of communities were reduced by 38 and 91%, respectively. Low levels of atrazine (3.2-32.0 g~g/L) resulted in a 46% increase in species numbers and a greater concentration of total protein and chlorophyll a (41 and 57%, respectively). Results compared well with other estimates of ,chronic toxicity for effects of atrazine on aquatic communities. Reported MATCs ranged from 70.7 to 3,400 ~g/L. The results from this test emphasize the importance of monitoring both structural and functional measures of community integrity in toxicity testing with multispecies.

I Address correspondence to Dr. J. R. Pratt, 213 Ferguson Building, The Pennsylvania State University, University Park, Pennsylvania 16802

Recent attention has focused on the limitations of single species toxicity tests in predicting adverse environmental effects of chemicals (Cairns 1983). Single species tests lack a significant degree of environmental realism and may fail to predict indirect or system level responses to toxicants, such as changes in predation, competition, succession, and nutrient cycling (National Research Council 1981). Multispecies or microcosm tests can include ecologically important interactions, evaluate effects on complex processes, and may provide a better prediction of environmentally ~'safe" levels of a potentially toxic compound. The herbicide atrazine (2-chloro-4-[ethylamino]6-[~sopropylamino]-s-triazine) is one of the most heavily used herbicides in the United States, representing 17.3% of the total poundage of herbicide used in 1982 (alachtor was used the most at t9.1% and trifluralin was third at 8.1%; Council on Environmental Quality 1984). Considerable evidence from single species tests and limited evidence from microcosm tests confirm its effects on aquatic ecosystems. Atrazine inhibits photosynthesis through blockage of electron lransport in the Hill reaction of photosystem II (Moreland 1980) and is used to control pest plant species in corn and soybean fields. Effects of atrazine on nonphotosynthetic components of a community may result, due to ab terations in food chains and other important community level interactions. This research examined the effect of atrazine on naturally derived microbial communities comprised of bacteria, fungi, protozoa, algae, and micrometozoans collected on polyurethane foam (PF) substrates. Researchers have successfully used PF artificial substrates to study freshwater microbial communities (Pratt and Cairns 1985; Cairns et at. 1979)u Communities colonized on these substrates

450

are easily sampled, can be adequately maintained in the laboratory, and have previously been used as test communities for toxicological studies (Cairns et al. 1985, 1986; Niederlehner et al. 1985, 1986; Pratt et al. in press a, in press b). Results of previous studies have shown that sensitivity of microbial communities to toxicants is comparable to effect levels in chronic, single species toxicity tests. Use of complex communities allows simultaneous testing of interacting species and, therefore, provides a higher level of biological organization not available in single species toxicity tests.

Methods

J . R . Pratt et al. SPE octadecyl (C18) solid phase disposable columns (J. T. Baker; Steinheimer and Brooks 1984). Atrazine was eluted from the column with 2.5 ml of HPLC grade acetonitrile (ACN), resulting in a 100-fold increased concentration. Standards were extracted in the same manner each time an analysis was done. Sample collection and concentration were always identical in method, and analysis was carried out the following day. Concentrated samples were stored in glass vials with teflon caps at 4~ Samples were analyzed with a high performance liquid chromatograph with a Beckman Ultrasphere ODS column (5 txm mesh, 4.6 mm ID • 250 ram). A mixture of water and ACN (67:33) at 1.5 ml/min was used to elute the atrazine. Absorbance at 220 nm was used to monitor peak height, and concentration was calculated from standards. Precision of these determinations based on replicate measures was 2.04% as a coefficient of variation. Average recovery calculated from spiked samples was 104.1%. The minimum detectable level of atrazine, using this technique, was 0.5 txg/L.

Test System The test units (microcosms) were high density polyethylene tubs (35 x 28 x 15 cm) with an inlet tube to deliver a mixture of diluent water and toxicant at one end and three drain holes at the other end. Flow rate was set to give approximately seven turnovers of the 7-L test volume per day. Triplicates of five concentrations of atrazine, a diluent control, and a solvent (methanol) control were tested. True replication was obtained by using 21 toxicant reservoirs with appropriate concentrations of atrazine, diluent, or solvent pumped to 21 mixing chambers. Diluent water from a head box was fed into the mixing chambers via a flow splitter. From the mixing chambers, test solutions flowed directly into the microcosms. Diluent water was tap water (source New River, Virginia; hardness - 7 0 ppm CaCO3, alkalinity ~45 ppm CaCO3, pH - 8 . 4 ) dechlorinated by passage through activated charcoal. Lighting was provided by Vita Lite bulbs (CRI >90, Duro Test Corp.) located - 3 0 cm above the test systems (intensity -5000 lux). Lights were set on a 16L:8D photoperiod. Water temperature was uncontrolled and ranged from 13.5~ to 15.0~ during the test period.

Toxicant Technical grade atrazine (98.6%, FL851205, Ciba-Geigy) was used for toxicant stocks, and analytical grade atrazine (98.8%, Code #$85-0653-3, Ciba-Geigy) was used as standard for analytical purposes. The primary stock was made by dissolving 200.84 mg of atrazine in 100 ml of methanol and diluting in distilled water to 1 L. Secondary stocks were made by diluting the appropriate amount of primary stock with diluent water. Both primary and secondary stocks were made once at the start of the test and again on day 10. Nominal atrazine concentrations tested were 3, 10, 30, 100, and 300 txg/L. The solvent control consisted of diluent water amended with methanol (0.015% final concentration). Diluent and solvent controls were analyzed separately for all variables measured, but there were never any significant differences between these two groups. Toxicant concentrations in all test systems were measured at the start of the test and after 10 and 21 days. Approximately 500 ml of water were collected from each system and filtered through a 0.45 ~xm GA-6 Metficel membrane filter (Gelman). Atrazine was extracted from 250 mg of filtered water, using 10-

Test Organisms Test communities collected on polyurethane foam (PF) substrates (6.0 x 5.0 x 4.0 cm) in Pandapas Pond (Montgomery County, Virginia) were used as test units. These communities included at least 150-200 species of bacteria, protozoa, algae, fungi, and small metazoans. The PF substrates placed in Pandapas Pond were collected after 14 days of exposure and returned to the laboratory. Two PF units were placed in each of the 21 test chambers as species sources. Five barren PF substrates (islands) were also placed in each test chamber to provide colonizable habitat. One of these island substrates was removed from each system after 3, 7, 14, and 21 days and examined for protozoan species richness. Three reference substrates were examined prior to the start of the test to determine initial species richness. At the end of the test, the species sources were removed and examined in the same manner to evaluate species survival. The PF substrates were sampled by removing one substrate from each test chamber and placing it into randomly numbered sterile Whirl-Pak | bags. Substrates were then squeezed to remove as much of the contents as possible, and the substrate was discarded. Samples were allowed to settle and then repeatedly subsampled and examined at 200-400 x for identification of protozoa (see Cairns et al. 1979). Protozoa were identified to lowest practical taxonomic unit (usually species). When identification to species proved impossible, taxa were identified to genus. The total number of taxa was recorded, and sample identifications were matched with original random sample numbers.

Physical and Chemical Measures Temperature, pH, dissolved oxygen, and conductivity were measured weekly (USEPA 1985) in all test systems - 1 hr after lights on and prior to substrate collection. Hardness and alkalinity were measured weekly in the diluent control, solvent control, and highest test concentration (USEPA 1985). In addition to determination of species richness, several nontaxonomic variables were evaluated in day 21 samples. Protein concentration was used as an estimate of total biomass. Samples were concentrated by centrifugation (1500 rpm for 10 rain), and

Atrazine Effects on Microbial Communities protein was extracted with 0.5 N NaOH (Rausch 1981). Protein concentration in the extracts was measured using the method of Bradford (1976). An aliquot of each sample was fixed to pH <2 with trace pure H N O 3 for analysis of calcium, potassium, and magnesium (USEPA 1985). Metals were analyzed by flame atomic absorption spectrophotometry on a Perkin-Elmer model 603 atomic absorption spectrophotometer. Chlorophyll a was extracted in 90% acetone overnight and concentrations determined by the trichromatic method (American Public Health Association 1985).

Statistical Analysis Accrual of protozoan species on initially barren PF substrates over time was fitted to the MacArthur-Wilson (MacArthur and Wilson 1967) equilibrium model S = Seq(1 - e -Gt) where S = species number at time t, Seq = species equilibrium number, G = colonization rate, and t = time, using nonlinear least squares regression to estimate Seq and G parameters (Draper and Smith 1981; Helwig and Council 1979). Colonization curves for different treatments were compared, using 'dummy' variable analysis (Kleinbaum and K u p p e r 1978). Species numbers and nontaxonomic measures were compared on each sample day using analysis of variance (ANOVA). When ANOVA was significant (p < 0.05), multiple comparisons were made using Fisher's LSD (Sokal and Rohlf 1983). A no-observable-effect concentration (NOEC), the highest tested concentration where the response is not significantly different than control, and the lowest-observable-effect concentration (LOEC), the lowest concentration tested where the response is significantly different than control, were defined for all responses. The maximum allowable toxicant concentration (MATC), defined as the geometric mean of the NOEC and LOEC, was also calculated.

451

Colonization curves were significantly different among t r e a t m e n t s (F12,70 = 11.02, p < 0.0001). Equilibrium species numbers were significantly greater at 3.2-32.0 Ixg/L and significantly lower at 337 ixg/L. Colonization rate (G) was also significantly affected by atrazine at 3.2-110 ~g/L. Examination of epicenters at the end of the test showed no difference in survival among treatments. The results of nontaxonomic measures are shown in Table 3. Protein biomass was significantly elevated at low atrazine concentrations (3.2-10.0 jxg/L) and significantly reduced at the highest atrazine concentration (337 txg/L). Chlorophyll a concentrations were - 5 0 % greater in the lower atrazine concentrations (3.2-10.0 pxg/L) than in the control, but were reduced to 9% of control values at 337 ~xg/L atrazine. Midmorning dissolved oxygen, used as an index of primary productivity, was depressed at t>10.0 Ixg/L atrazine throughout the test and, on day 21, was significantly lower than controls at >32.0 ixg/L. Concentrations of potassium, magnesium, and calcium in PF substrate samples on day 21 are shown in Table 3. Atrazine significantly reduced the content of magnesium and calcium in communities at 32.0 ~xg/L, and the concentration of potassium was reduced at 337 ~g/L. Unlike species number and protein and chlorophyll a content, no significant increase in concentration of these metals was observed at the low atrazine concentrations.

Discussion Results

Atrazine concentrations were stable in all test systems (Table 1). Water chemistries showed no deviation from normal ranges for dechlorinated tap water with respect to pH, temperature, hardness, alkalinity, and conductivity. Protozoan colonization of PF substrates is shown in Figure 1. There was a significant treatment effect on every day except day 7. Low atrazine concentrations (3.2-32.0 jxg/L) were associated with increased species number on days 14 and 21. Initially, 110 and 337 Ixg/L atrazine decreased species numbers, but, by day 21, only the highest concentration resulted in a significant decrease in species numbers. At all concentrations, protozoan colonization of barren substrates was adequately described by the MacArthur-Wilson equilibrium model. Results of n o n l i n e a r r e g r e s s i o n using 'dummy' variable analysis are shown in Table 2.

Responses reflecting the metabolic status of test communities (i.e., dissolved oxygen, calcium, magnesium) were the most sensitive indicators of adverse effects of atrazine (MATC = 17.9 ~xg/L, Table 4). Taxonomic richness and biomass measures (i.e., equilibrium species number, protein, chlorophyll) were less sensitive, but responded similarly (MATC = 193 Ixg/L). These results are consistent with other atrazine studies that found impaired metabolic function at atrazine levels lower than those affecting community structural changes (Kemp et al. 1985; Malanchuk and Kollig 1985; Stay et at. 1985; Brockway et al. 1984; de Noyelles et al. 1982). Atrazine significantly d e c r e a s e d equilibrium species numbers in protozoan communities at 337 p~g/L and reduced colonization rate (G) at 3.2 txg/L. Decreased protozoan equilibrium species numbers and altered colonization rates due to toxic stress have been reported :for other toxicants such as

452

J . R . Pratt et al.

Table 1. Measured atrazine concentrations in microcosms. Values are means (SD) Nominal atrazine concentration (~g/L)

Measured atrazine concentration (ixg/L) Day 0

Day 10

Day 21

Mean

Control Solvent Control 3

<0.5 <0.5 3.0 (0.il) 9.2 (0.23) 29.7 (2.31) 102 (8.02) 324 (8.02)

<0.5 <0.5 3.6 (0.33) 11.3 (0.58) 33.3 (1.34) 102 (10.4) 336.0 (17:0)

<0.5 <0.5 3.1 (0.00) 9.6 (1.10) 33.1 (4.82) 126 (14.4) 350 (14.4)

<0.5 <0.5 3.2 (0.34) 10.0 (1.15) 32.0 (3.26) 110 (15.6) 337 (16.4)

10 30 100 300

r ~0

Table 2. Effects of atrazine on colonization estimates based on MacArthur-Wilson model. Estimates of equilibrium species number and colonization rate based on fit to model equation using nonlinear regression. Numbers in parentheses are asymptotic standard errors. Values significantly different from control are marked with an asterisk

Control o-e SolventC~O

9r, 30

10I

3

I

7

I

Days

14

]3.2ug/L~

I

21

I

I

I

] II0

Treatment

Seq a

Gb

Control

33.2 (2.34) 34.4 (1.92) 42.4* (2.65) 52.6* (5.06) 61.7" (12.26) 34.8 (3.05) 24.1" (1.18)

0.349 (0.104) 0.309 (0.068) 0.167' (0.030) 0.122' (0.028) 0.081" (0.031) 0.138" (0.031) 0.266 (0.048)

I

9

Solvent control 3.2 10.0

I

I

I

I

32.0 110

~r 1 337

~

337 a Equilibrium species number b Colonization rate

Fig. 1. Colonization of polyurethane foam substrates by protozoa over time. Curves were fitted, using the MacArthurWilson equilibrium model. Points marked with asterisks are significantly different from controls

heavy metals (Niederlehner et al. 1985, 1986; Pratt et al. in press b), mixed effluents (Shen et al. 1986), and a lampricide (McCormick et al. 1986). At the highest atrazine concentration (337 Ixg/L), we observed a 91% reduction in algal biomass (expressed as chlorophyll a concentration). Decreased algal chlorophyll a in response to atrazine treatment has also been reported by Larsen et al. (1986)

at concentrations >200 Ixg/L. Johnson (1986) found decreased algal biomass (dry weight) in microcosms treated with 1,000 Ixg/L atrazine. Measures of photosynthetic function, such as dissolved oxygen and in vivo fluorescence, have also been used to evaluate the effect of atrazine on photosynthetic organisms. A significant reduction in dissolved oxygen at 32.0 ixg/L was observed, which was lower than that at which chlorophyll a levels were reduced (337 ixg/L), suggesting that photosynthetic function may be more sensitive than chlorophyll a biomass as an indicator of stress. Rocchio and Malanchuk (1986) applied single, daily

Atrazine Effects on Microbial Communities

453

Table 3. Selected nontaxonomic measures on day 21. Shown are mean (SD) and results of one-way analysis of variance. Values signlJicantly different from control are marked with an asterisk

Treatment Control Solvent 3.2 10.0 32.0 110 337 p

Protein (rag/L)

Chlorophyll (p.g/L)

DO ~ (mg/L)

K (mg/L)

Mg (mg/L)

Ca (mg/L)

35.7 (7.4) 31.9 (8.0) 49.7* (11.0) 57.2* (20.6) 47.0 (5.30) 42.7 (2.28) 21.2* (5.10) O.0072

1,326.0 (491.8) 732.0 (241.5) 2,247.3* (711.3) 2,356.0* (1330.8) 1,037,0 (216,1) 538.7 (50.4) 88.3* (124.1) O.0005

9.23 (0.60) 8.67 (0.81) 8.90 (0.10) 7.47 (1.63) 6.00* (1.08) 6.30* (1.55) 6.73* (0.15) O.0060

2.04 (0.44) 2.04 (0.54) 1,73 (0.10) 2.56 (0.90) 2.17 (0.52) 1.66 (0.11) 1.14" (0.19) O.0345

3.82 (0.36) 3.69 (0.49) 4.17 (0.45) 4.17 (0.45) 3,15" (0.06) 3,45" (0.20) 3.00* (0.11) O.0043

24.1 (2.19) 23.0 (0.29) 23.0 (0.29) 21.8 (0.82) 20.2* (0.29) 22.7 (2.50) 20.0* (0.00) O.0271

Dissolved oxygen

Table 4. Calculated NOEC, LOEC, and MATCs based on structural and functional responses. All values are txg/L Response

NOEC a

LOEC b

MATC c

Seq a Protein Chlorophyll a Potassium Magnesium Calcium Dissolved oxygen

110 110 110 110 10.0 10.0 10.0

337 337 337 337 32.0 32.0 32.0

193 193 193 193 17.9 17.9 17.9

a No-observable effect concentration b Lowest-observable effect concentration r Maximum allowable toxicant concentration d Species equilibrium number

doses of atrazine at 50-150 ~g/L in aquatic systems. Dissolved oxygen dropped to zero in all treatments, but, after termination of dosing, all treatments recovered within 48 hr and the investigators could not detect any permanent ecosystem damage. Stay et aI. (1985) observed decreased dissolved oxygen in Taub microcosms treated with 60 Ixg/L atrazine and increased chlorophyll a biomass. De Noyelles et al. (1982) found that photosynthetic function (measured as in vivo fluorescence) was inhibited at concentrations as low as 1 Ixg/L in communities tested in the laboratory and decreased dissolved oxygen at 100 Ixg/L in ponds treated with atrazine (de Noyelles and Kettle 1985). Brockway et al. (1984) tested microcosms composed of plankton and Aufwuchs and found decreased dissolved oxygen at 50 ixg/L atrazine. Potassium, magnesium, and calcium are seques-

tered by organisms, and these metals were used as another indicator of biomass. Magnesium and calcium content were the most sensitive, and both were significantly reduced at 32.0 Ixg/L. Potassium was less sensitive to atrazine and was reduced only at 337 Ixg/L. Decreased potassium levels in PF communities in response to copper treatment has also been reported (Pratt et al. in press b). Low levels of atrazine produced a stimulatory effect on species richness and biomass. Species numbers were elevated 22-46% at 3.2-32.0 txg/L. At 3.2-10.0 Ixg/L atrazine, total protein and chlorophyll a were elevated as much as 41 and 57%, respectively. Similar increases in biomass at sublethal atrazine concentrations have been reported in the literature. Kuemmerlin (1986) found increased protein, chlorophyll a, and RNA content in planktonic algae treated with 5-200 Ixg/L atrazine. Larsen et al. (1986) conducted single species algal assays with eight different species of algae. They measured 14C-uptake and found a 10-fold range in sensitivity with EC50s ranging from 37-308 ixg/L. This suggests that loss of sensitive species and resultant increase in growth of insensitive species might be responsible for the observation of increased chlorophyll a at low atrazine concentrations. Stay et al. (1985) found increased chlorophyll a at low atrazine concentrations, but primary production expressed as 14C-uptake/chlorophyll a was significantly reduced compared to controls, suggesting an impairment of photosynthetic function despite the apparent increase in algal biomass. Stress responses in individual species are often realized as a stimulation of activity or production at

454

J . R . Pratt et al.

MATCs, this study

Reported MATCs Altered life-cycle of C ~ r i n o d e n (Ward and Ballantine 1985)

1000

Decreased species equilibrium number, d 21 species number, chlorophyll a, and total protein

30 mJ~m.

I( Decreased potassium content

Decreased macrophyte biomass (Kemp et al, 1985)

Decreased 14C-uptake and DO (Stay et al. 1985); Decreased algal growth and macrophyte biomass (Johnson 1986) Decreased in vivo fluorescence (Shaw et al. 1985) Altered life-cycle of M ~ d o p s i s (Ward and Ballantine 1985) Altered life-cycle of rainbow trout (Macek 1976) Decreased macrophyte biomass (Kemp et al. 1985)

30 Decreased magnesium and calcium content; decreased dissolved oxygen

!

10

Atrazine (ug/L) Fig. 2. Maximum allowable toxicant concentrations predicted from this study and taken from the literature.

sublethal concentrations of toxic compounds (hormesis; Fedtke 1982). Interpreting such responses in complex systems is more problematic. It seems unlikely that a photosynthetic inhibitor such as atrazine could result in greater production of chloro-

phyll b i o m a s s c o n c o m i t a n t with increases in numbers of microbial species such as protozoa that include a great number of photosynthetic forms. In our experiments, low levels of atrazine (3.2-10.0 ~xg/L) resulted in the near doubling of chlorophyll

Atrazine Effects on Microbial Communities

content, but this biomass elaboration was not translated into greater rates of primary production as measured by midmorning dissolved oxygen concentrations. These unusual responses of biomass e l a b o r a t i o n and i n c r e a s e d n u m b e r s of e x t a n t species may be related to shifting nutrient dynamics in systems that are photosynthetically stressed. Alternatively, elaboration of chlorophyll biomass may be an irritation response caused by the interruption of electron flow due to atrazine. The source of energy for elaboration of chlorophyll at low atrazine doses is uncertain but may come from stored reserves or facultative heterotrophic activity. In many respects, the observed nonlinear responses bear striking resemblance to compensating mechanisms in information overloads (Meier 1972). While this nonlinear response to stress has been termed a subsidy-stress gradient by Odum and colleagues (1979), it seems unlikely that photosynthetic inhibition could result in a subsidy to the system. Rather, increased biomass and numbers of extant species represent initial stress responses and probably reflect the breakdown of normal feedback mechanisms controlling nutrient dynamics and species interactions in communities. As such, these responses are not reflective of a subsidy but are the initial signs of stress that will be followed by cascading events leading to significant simplification of the community if the stress is increased. The persistence of atrazine in aquatic environments is influenced by a number of environmental factors, such as s e d i m e n t type, m a c r o p h y t e growth, and bacterial activity. Half-lives for atrazine in laboratory microcosms have ranged from 3-12 days (Jones et al. 1982) to 3-4 months (Kemp et al. 1985). Reported atrazine concentrations in natural systems range from 0 to 74 txg/L (Richard et al. 1987; Kemp et af. 1985). Kemp et al. (1985) measured atrazine concentrations at several sites in the Chesapeake Bay and found high concentrations were usually transient, due to agricultural runoff in the spring and levels were reduced to <5 Ixg/L within 6-24 hr. Atrazine is not normally supplied continuously to ecosystems and, accordingly, appears as a pulse following application of atrazine containing herbicides in agricultural systems. Agricultural runoff into impounded waters with low flow could result in atrazine concentrations approaching levels that have been predicted to produce adverse ecological effects (Albanis et al. 1986). Where atrazine concentrations persist, it is possible that increased algal biomass may result. It is unclear whether low levels of atrazine may induce greater algal biomasses in natural ecosystems, but the potential for algal biomass elaboration is present at environmentally realistic concentrations

455

of atrazine. Although not examined in our study, other investigators have supported recovery after short-term exposure to low levels of atrazine (Johnson 1986; Jones et al. 1986; Larsen et al. 1986; de Noyelles and Kettle 1985; Stay et al. 1985; Brockway et al. 1984; de Noyeltes et aI. 1982), suggesting that aquatic communities may not be at risk due to exposure of natural levels of atrazine as long as the exposure time is limited. The MATCs predicted from a variety of community response measured in this study ranged from 17.9-193 ~g/L atrazine (Table 4). Chronic MATCs found by other investigators ranged from 71-3,400 ixg/L (Figure 2), indicating a wide range of sensitivities by both species and systems (Johnson 1986; Kemp et al. 1985; Shaw et aI. 1985; Stay et al. 1985; Ward and Ballantine 1985; Macek 1976). Excluding fish and invertebrate data, the range of MATCs is 70.7-707.7. In the majority of cases, the most sensitive measure was one of function and not structure. This was also the case with our results; oxygen production and calcium and magnesium content were the most sensitive variables measured. The greater sensitivity of functional responses compared with structural responses, however, is not universally true for all toxic compounds. We have found with other toxicants that structural rather than functional responses are more sensitive predictors of stress, such as with copper (Pratt et al. in press a) and chlorine (Pratt et al. in press b). These results emphasize the need to measure both structural a n d functional responses to toxic stress. The results from this study support the use of multispecies toxicity tests. Because of the broad range of species sensitivities to toxic chemicals, single species toxicity tests may not be adequate in predicting a safe level for the protection of complex aquatic communities.

Acknowledgments. This work was supported by grants from the United States Environmental Protection Agency, Office of Research and Development (R-812813-01-0) and the Air Force Ofrice of Scientific Research (AFOSR-85-0324). This paper has not been subject to agency review, and no official endorsement should be inferred. The authors would like to thank Dr. Homer M. LeBaron, Senior Research Fellow at Ciba-Geigy for supplying the atrazine and Darla Donald for editorial assistance and Betty Higginbotham for typing the final manuscript.

References Albanis TA, Pomonis PJ, Sdoukos AT (1986) Seasonal fluctuations of organochlorine and triazine pesticides in the aquatic system of I o a n n i n a basin (Greece). Sci Total E n v i r o n 58:243-253

456 American Public Health Association, American Water Works Association, Water Pollution Control Federation (1985) Standard Methods for the Examination of Water and Wastewater, 16th edn. Washington DC Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein, utilizing the principle of protein-dye binding. Anal Biochem 72:248-254 Brockway DL, Smith PD, Stancil FE (1984) Fate and effects of atrazine in small aquatic microcosms. Bull Environ Contam Toxicol 32:345-353 Cairns J Jr (1983) Are single species tests alone adequate for estimating environmental hazard? Hydrobiologia 100:47-57 Cairns J Jr, Kuhn DL, Plafkin JL (1979) Protozoan colonization of artificial substrates. In: Weitzel RL (ed) Methods and Measurements of Periphyton Communities: A Review. American Society for Testing and Materials, Philadelphia PA, pp 34-57 Cairns J Jr, Pratt JR, Niederlehner BR (1985) A provisional multispecies toxicity test, using indigenous organisms. J Test Eval 13:316-319 Caims J Jr, Pratt JR, Niederlehner BR, McCormick PV (1986) A simple cost-effective multispecies toxicity test, using organisms with a cosmopolitan distribution. Environ Monit Assess 6:207-220 Council on Environmental Quality (CEQ) (1984) Environmental Quality, 15th Report. Washington DC de Noyetles F Jr, Kettle, WD, Sinn DE (1982) The responses of plankton communities in experimental ponds to atrazine, the most heavily used pesticide in the United States. Ecology 63:1285-1293 de Noyelles F Jr, Kettle WD (1985) Experimental ponds for evaluating the bioassay predictions. In: Boyle TP (ed) Validation and Predictability of Laboratory Methods for the Assessment of Fate and Effects of Environmental Contaminants in Aquatic Ecosystems, STP 865. American Society for Testing and Materials, Philadelphia PA, pp 91-103 Draper NR, Smith H (1981) Applied Regression Analysis. Wiley, New York Fedtke C (1982) Biochemistry and Physiology of Herbicide Action. Springer-Verlag, New York Helwig J, Council K (1979) SAS User's Guide. SAS Institute Inc, Cary NC Johnson BT (1986) Potential impact of selected agricultural chemical contaminants on a northern prairie wetland: A microcosm evaluation. Environ Toxicol Chem 5:473-485 Jones TW, Kemp WM, Stevenson JC, Means JC (1982) Degradation of atrazine in estuarine water/sediment systems and soils. J Environ Qual 11:632-638 Jones TW, Kemp WM, Estes PS, Stevenson JC (1986) Atrazine uptake, photosynthetic inhibition, and short-term recovery for the submersed vascular plant, Potamogeton perfoliatus L. Arch Environ Contam Toxicol 15:277-283 Kemp WM, Boynton WR, CunninghamJJ, Stevenson JC, Jones TW, Means JC (1985) Effects of atrazine and linuron on photosynthesis and growth of the macrophytes Potamogeton perfoliatus L. and Myriophyllum spicatum L. in an estuarine environment. Mar Environ Res 16:255-280 Kleinbaum D, Kupper L (1978) Applied Regression Analysis and Other Multivariable Methods. Duxbury Press, North Scituate MA Kuemmerlin R (1986) Interactions between the sorption of the herbicide atrazine and the physiological condition of planktonic algae. Arch Hydrobiol Suppl 66:495-571

J.R. Pratt et al. Larsen DR de NoyeUes F, Stay F, Shiroyama T (1986) Comparisons of single-species, microcosm and experimental pond response to atrazine exposure. Environ Toxicol Chem 5:179-190 MacArthur RH, Wilson EO (1967) The Theory of Island Biogeography. Princeton University Press, Princeton NJ, 203 PP Macek KJ (1976) Chronic Toxicity of Atrazine to Selected Aquatic Invertebrates and Fishes. US Environmental Protection Agency Ecological Research Series, EPA-600/3-76047 Malanchuk JL, Kollig HP (1985) Effects of atrazine on aquatic ecosystems: A physical and mathematical modeling assessment. In: Boyle TP (ed) Validation and Predictability of Laboratory Methods for Assessing the Fate and Effects of Contaminants in Aquatic Ecosystems, STP 865. American Society for Testing and Materials, Philadelphia PA, pp 204-228 McCormick PV, Pratt JR, Cairns J Jr (1986) Effect of 3-trifluoromethyl-4-nitrophenol on the structure and function of protozoan communities established on artificial substrates. In: Cairns J Jr (ed) Community Toxicity Testing, STP 920. American Society for Testing and Materials, Philadelphia PA, pp 224-240 Meier RL (1972) Communication stress. Ann Rev Ecol Syst 3:289-314 Moreland DE (1980) Mechanisms of action of herbicides. Annu Rev Plant Physiol 31:597-638 National Research Council (1981) Testing for Effects of Chemicals on Ecosystems. National Academy Press, Washington DC Niederlehner BR, Pratt JR, Buikema AL Jr, Cairns J Jr (1985) Laboratory tests evaluating the effects of cadmium on freshwater protozoan communities. Environ Toxicol Chem 4:155-165 (1986) Comparison of estimates of hazard derived at three levels of complexity. In: Cairns J Jr (ed) Community Toxicity Testing, STP 920. American Society for Testing and Materials, Philadelphia Pa, pp 30-48 Odum EP, Finn JT, Franz EH (1979) Perturbation theory and the subsidy-stress gradient. BioScience 29:349-352 Pratt JR, Cairns J Jr (1985) Long-term patterns of protozoan colonization in Douglas Lake, Michigan. J Protozool 32:95-99 Pratt JR, Bowers NJ, Niederlehner BR, Cairns J Jr (In press a) Effect of chlorine on microbial communities in naturally derived microcosms. Environ Toxicol Chem Pratt JR, Niederlehner BR, Bowers NB, Cairns J Jr (in press b) Prediction of permissible concentrations of copper from microcosm toxicity tests. Tox Assess Rausch P (1981) The estimation of microalgae protein content and its meaning to the evaluation of algal biomass, V: Comparison of methods for extracting protein. Hydrobiologia 78:237-251 Richard F, Ripley BD, Braun HE, Clegg BS, Johnston R, O'Neill TJ (1987) Survey of farm wells for pesticide residues: Southern Ontario, Canada. Arch Environ Contam Toxicol 16:1-8 -

-

Rocchio PM, Malanchuk JL (1986) The effects of atrazine on dissolved oxygen and nitrate concentrations in aquatic systems. Environ Int 12:597-602 Shaw DR, Peiper TF, Nofziger DL (1985) Comparison of chloro-

Atrazine Effects on Microbial Communities phyll fluorescence and fresh weight as herbicide bioassay techniques. Weed Sci 33:29-33 Shen Y-F, Buikema AL Jr, Yongue WH Jr, Pratt JR, Cairns J Jr (1986) Use of protozoan communities to predict environmental effects of pollutants. J Protozool 33:146-151 Sokal RR, Rohlf FJ (1983) Biometry. WH Freeman, NY, 859 pp Stay FS, Larsen DP, Kato A, Rohm CM (1985) Effects of atrazine on community level responses in Taub microcosms. In: Boyle TP (ed) Validation and Prediction of Laboratory Methods for Assessing the Fate and Effect of Contaminants in Aquatic Ecosystems, STP 865. American Society for Testing and Materials, Philadelphia PA, pp 75-90 Steinheimer TR, Brooks MG (1984) Development and eva!uation

457 of a gas chromatgraphic method for the determination of triazine herbicides in natural water samples. Int J Environ Anal Chem 17:97-111

United States Environmental Protection Agency (1985) Methods for the Chemical Analysis of Water and Wastewater, EPA600/4-79-020. National Technical Information Service, Springfield VA Ward GS, Ballantine L (1985) Acute and chronic toxicity of atrazine to estuarine fauna. Estuaries 8:22-27

Manuscript received August 13, 1987 and in revised form November 23, !987.