&EPA

United States Office ofEnvironmental ProtectionAgency

Research and DevelopmentWashington, DC 20460

EPA/600/6-9 l/003February 1991

Methods for AquaticToxicity IdentificationEvaluations

Phase I ToxicityCharacterizatidn Procedures

Second Editionc

EPA/600/6-9 l/O03February 1991

Methods for AquaticToxicity Identification Evaluations

Phase 1 Toxicity Characterization Procedures

(Second Edition)

Edited by

T.J. Norberg-KingD.I. Mount

E.J. DurhanG.T. An kley

L.P. BurkhardEnvironmental Research Laboratory

Duluth, MN 55804

JR. AmatoM.T. Lukasewycz

M.K. Schubauer-BeriganAScl Corporation

6201 Congdon BoulevardDuluth, MN 55804

L. Anderson-CarnahanRegion IV - Policy Planning & Evaluation Branch

Atlanta, GA 30365

National Effluent ToxicityAssessment Center

Technical Report 18-90

Environmental Research LaboratoryOffice of Research and DevelopmentU.S. Environmental Protection Agency

Duluth, MN 55804

@ Punted on Recycled Paper

Notice

This document has been reviewed in accordance with U.S. Environmental ProtectionAgency Policy and approved for publication. Mention of trade names or commercialproducts does not constitute endorsement or recommendation for use.

-

ii

Foreword-

This document is one in a series of guidance documents intended to aid dischargers andtheir consultants in conducting aquatic organism Toxicity Identification Evaluations (TIES)as part of Toxicity Reduction Evaluations (TREs). Such effluent evaluations may be requiredas the result of an enforcement action or as a condition of a National Pollutant DischargeElimination System (NPDES) permit. This document will also help to provide U.S. Environ-mental Protection Agency (EPA) and State Pollution Control Agency staff with the back-ground necessary to oversee and determine the adequacy of effluent TIES proposed andperformed by NPDES permittees. While this TIE approach was developed for effluents, themethods and techniques have direct applicability to other types of aqueous samples, suchas ambient waters, sediment pore waters, sediment elutriates, and hazardous wasteleachates.

The TIE approach is divided into three phases. Phase I (this document) containsmethods to characterize the physical/chemical nature of the constituents which cause ctoxicity. Such characteristics as solubility, volatility and filterability are determined withoutspecifically identifying the toxicants. Phase I results are intended as a first step in specificallyidentifying the toxicants but the data generated can also be used to develop treatmentmethods to removetoxicity without specific identification of the toxicants. Two EPA TREmanuals (EPA, 1989A; 1989B) use parts of Phase I in developing those approaches.

Phase II (EPA, 1989C) describes methods to specifically identify toxicants if they arenon-polar organics, ammonia, or metals. This Phase is incomplete because methods forother specific groups, such as polar organics, have not yet been developed. As additionalmethods are developed, they will be added.

Phase III (EPA, 1989D) describes methods to confirm the suspected toxicants. It isapplicable whether or not the identification of the toxicants was made using Phases I andII. Complete Phase III confirmations have been limited to date, but avoiding Phase III mayinvite disaster because the suspected toxicant was not the actual toxicant(

Phases I and II are intended for acutely toxic effluents. However, that limitation does notmean that effluents having chronic limits cannot be evaluated using these methods. TIEmethods to evaluate the cause of chronic toxicity in effluents are being developed (EPA,1991A).

These methods are not mandatory but are intended to aid those who need to character-ize, identify or confirm the cause of toxicity in effluents or other aqueous samples such asambient waters, sediments, and leachates. Where we lack experience, we have indicatedthis and have suggested avenues to follow. All tests need not be done on every sample; thetests are, in general, independent. However, experience has taught us that skipping testsmay result in wasted time, especially during the early stages of Phase I. An exception to thisis when one wants to know if only a specific substance, for example ammonia, is causingthe toxicity or if toxicants other than ammonia are involved. Otherwise, we urge the use of

, the whole battery of tests.

We welcome comments from users of these manuals so that future editions can beimproved. Comments can be sent to NETAC, ERL-Duluth, 6201 Congdon Boulevard,Duluth, MN 55804.

.*.III

Abstract

In 1988, the first edition “Methods for Aquatic Toxicity Identification Evaluations:Phase I Toxicity Characterization Procedures” was published (EPA, 1988A). This secondedition provides more details and more insight into the techniques described in the 1988document. The manual describes procedures for characterizing the physical/chemicalnature of toxicants in acutely toxic effluent samples, with applications to other types ofsamples such as receiving water samples, sediment pore water or elutriate samples, andhazardous wastes. The presence and the potency of the toxicants in the samples aredetected by performing various manipulations on the sample and by using aquatic organ-isms to track the changes in the toxicity. This toxicity tracking step is the basis of the toxicityidentification evaluation (TIE). The final step is to separate the toxicants from the otherconstituents in the sample in order to simplify the analytical process. Many toxicants mustbe concentrated for analysis.

The Phase I manipulations include pH changes along with aeration, filtration, sparging,solid phase extraction, and the addition of chelating (i.e., ethylenediaminetetraacetateligand (EDTA)) and reducing (i.e., sodium thiosulfate) agents. The physical/chemicalcharacteristics of the toxicants are indicated by the results of the toxicity tests conducted onthe manipulated samples.

Since the first document was developed, additional options or new procedures havebeen developed. For example, additional options are provided in the EDTA and sodiumthiosulfate addition tests, and in the graduated pH test. Also a discussion has been addedfor testing the effluent sample over time (weekly) to measure the rate of decay of toxicitywhich is used to detect the presence of degradable substances, particularly chlorine orsurfactants. Guidance for characterizing whether a toxicant removed by aeration issublatable is described, and techniques for characterizing filterable toxicity and adiscussionof C,, solid phase extraction elutable toxicity has been added. Use of multiple manipulationsis dtscussed and example interpretations of the results of the Phase I manipulations areprovided.

Additional manuals describe the methods used to specifically identify the toxicants(EPA, 19896) and to confirm whether or not the suspect toxicant is the actual toxicant(EPA, 19890).

Contents

Page

Foreword l . ..~..~.........~...~...............~....~.....~~............~............................................................... IllAbstract . . . . . . . . ..*........*..........................................*................................................*.*.......... ivContents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..~..................................................................... V

Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..~................................... viiTables . . .. . . . . . . . . . . . . . . . . . ..*..............................................*.............................*........................... VIII

Acknowledgments . . . . . . . . . . . . . ..*......................................................*..................................... ix

1.

2.

3.

4.

5.

6.

7.

8.

Introduction .............................................................................................................. l - l1.1 Background .................................................................................................... l - l1.2 Conventional Approach to TIES ..................................................................... l - l1.3 Toxicity Based Approach ............................................................................... l-3 l

Health and Safety .................................................................................................... 2-l

Quality Assurance .................................................................................................... 3-l3.1 TIE Quality Control Plans ............................................................................... 3-l3.2 Cost Considerations/Concessions ................................................................. 3-l3.3 Variability ........................................................................................................ 3-23.4 Intra-Laboratory Communication ................................................................... 3-23.5 Record Keeping ............................................................................................. 3-23.6 Phase I Considerations .................................................................................. 3-23.7 Phase II Considerations ................................................................................. 3-33.8 Phase III Considerations ................................................................................ 3-3_

Facilities and Equipment .......................................................................................... 4-l-

Dilution Water ........................................................................................................... 5-l

Effluent Sampling and Handling .............................................................................. 6-l6.1 Sample Shipment and Collection in Plastic versus Glass .............................. 6-3

Toxicity Tests ........................................................................................................... 7-l7.1 Principles ........................................................................................................ 7-l7.2 Test Species .................................................................................................. 7-l7.3 Toxicity Test Procedures ................................................................................ 7-27.4 Test Endpoints ............................................................................................... 7-37.5 Feeding .......................................................................................................... 7-57.6 Multiple Species ............................................................................................. 7-5

Phase I Toxicity Characterization Tests ................................................................... 8-l8.1 Initial Effluent Toxicity Test ........................................................................... 8-48.2 Baseline Effluent Toxicity Test ...................................................................... 8-58.3 pH Adjustment Test ....................................................................................... 8-88.4 pH Adjustment/Filtration Test ...................................................................... 8-158.5 pH Adjustment/Aeration Test ...................................................................... 8-21

v

Contents (continued)

Page

8.6 pH Adjustment/C,, Solid Phase Extraction Test ......................................... 8-278.7 Oxidant Reduction Test ............................................................................... 8-338.8 EDTA Chelation Test .................................................................................. 8-388.9 Graduated pH Test ..................................................................................... 8-44

9. Time Frame and Additional Tests a............................................................................9-l9.1 Time Frame for Phase I Studies ................................................................... 9-19.2 When Phase I Tests are Inadequate ............................................................. 9-19.3 Interpreting Phase I Results ......................................................................... 9-29.4 Interpretation Examples ............................................................................... 9-3 .

10. References . . . . . . . . . . . . . . . . . . . . . ..I..........................................................*.......................... 10-l10-l

vi

Figures

Number Page

l - l .1-2.6-1.7-l.

8-l.8-2.8-3.8-4.8-5.8-6.8-7.

8-8.

8-9.8-10.8-11.8-12.8-13.8-14.8-15.

8-l 6.

8-l 7.

8-l 8.

8-l 9.

Conventional approach to TIES ......................................................................... l-2Flow chart for toxicity reduction evaluations ...................... ..-............................. l-5Example data sheet for logging in samples ....................................................... 6-2Schematic for preparing effluent test concentrations using

simple dilution techniques ........................................................................... 7-4Overview of Phase I effluent characterization tests ............................................ 8-2Example data sheet for initial effluent toxicity test ............................................. 8-6Example data sheet for baseline effluent toxicity test ........................................ 8-7pE -pH diagrams for the CO,, H,O, and Mn-CO, systems (25°C) .................... .8-9Flow chart for pH adjustment tests .................................................................. 8-11Example data sheet for pH adjustment test ..................................................... 8-13Oven/iew of steps needed in preparing the filter and dilution water blanks

for the filtration and/or the C,, SPE column tests ...................................... 8-17Overview of steps needed in preparing the effluent for the filtration and/or

C,, SPE column tests ................................................................................ 8-18Example data sheet for filtration test ............................................................... 8-20Diagram for preparing pH adjustment/aeration test samples .......................... 8-23Example data sheet for aeration test ............................................................... 8-24Closed loop schematic for volatile chemicals .................................................. 8-26Step-wise diagram for preparing the C,, SPE column samples ..................... .8-29Example data sheet for effluent SPE test with and without pH adjustment.. . ..8-3 1Example data sheet for the oxidant reduction test when using a

gradient of sodium thiosulfate concentrations ........................................... 8-36Example data sheet for the oxidant reduction test when effluent

dilutions are used ...................................................................................... 8-37Example data sheet for EDTA chelation test when using a gradient of

EDTA concentrations ................................................................................. 8-41Example data sheet for the EDTA chelation test when

effluent dilutions are used .......................................................................... 8-42Example of data sheet for the graduated pH test when

effluent dilutions are used .......................................................................... 8-46

l

vii

Tables

Number Page

6-1. Volumes needed for Phase I tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38-1. Outline of Phase I effluent manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..‘......8-38-2. Acute toxicity of sodium chloride to selected aquatic organisms . . . . . . . . . . . . . . . . . . . . .8-148-3. Toxicity of methanol to several freshwater species . . . . . ..*.................................. 8-328-4. Toxicity of sodium thiosulfate to Ceriodaphnia dubia, Daphnia magna, and

fathead minnows . . . . . . . . ..~~.....~~~.~...~~..~....~~..............~....~.~~.........~.....~............ 8-348-5. Toxicity of EDTA to Ceriodaphnia dubia and fathead minnows in

water of various hardnesses and salinities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-398-6. The toxicity of the Mes, Mops, and Popso buffers to Ceriodaphnia dubia and

fathead minnows ~,,......,,.....,,,.,..~,,........,,..,,,..,.....,..,...,,..,....,.....,,..~.......... 8-48

_

-

Acknowledgments

We wish to acknowledge the assistance of many people from the EnvironmentalResearch Laboratory-Duluth (ERL-D) for their help and advice in preparing the manual. Theexperience referred to throughout the document is the collective experience of the individu-als in the National Effluent Toxicity Assessment Center (NETAC). During the developmentof this document and the first edition, that group has consisted of Don Mount, TeresaNorberg-King, Larry Burkhard, Liz Durhan, Gary Ankley, Shaneen Murphy (all ERL-D staff),Linda Anderson-Carnahan, (EPA, Region IV), Joe Amato, Marta Lukasewycz, GregPeterson, Jim Taraldsen, Jim Jenson, Mary Schubauer-Berigan, Art Fenstad, Doug Jensen,Steve Baker, Liz Makynen, Jo Thompson, Correne Jenson, Lara Andersen, LindaEisenschenk, Nola Englehorn, and Eric Robert, (all currently or formerly with AScl Corpo-ration, Duluth).

For this revision, several individuals were assigned sections to write or re-write. Thisgroup consisted of Don Mount, Teresa Norberg-King, Larry Burkhard, Liz Durhan, GaryAnkley, Marta Lukasewycz, Joe Amato, and Mary Schubauer-Berigan. Teresa synthesizedall the rewrites into similar styles, added additional sections, and updated the entiredocument. The assistance Debra Williams (AScl) provided to produce the graphics, toprepare the document, and assist in all aspects is greatly appreciated. Without her input theproduction of the document would have been slowed tremendously. Much of the data havebeen developed by a few people: Joe Amato generated much of the laboratory data that isdiscussed in this document, and others also generated the data for the tables (Jim Jenson,Doug Jensen, Shaneen Murphy, Greg Peterson, Gary Ankley, Mary Schubauer-Berigan,and Jo Thompson).

We also want to express our appreciation to Russ Hackett and Dave Mount (ENSR, FortCollins, CO) for their data on sodium thiosulfate and W. Tom Wailer (University of NorthTexas, Denton, TX) for his data and his technique for sulfur dioxide dechlorination.

As stated in the first edition of the TIE characterization document, the effluent groupwould not have been able to complete the work that is summarized in this report without thesupport and backing of Nelson Thomas, Senior Advisor for National Programs (ERL-D). Inaddition, Rick Brandes (EPA, Permits Division, Washington, D.C.) has been a strong voicein support of all the work upon which the manual is based. The support provided from theOff ice of Water through his impetus has enabled NETAC to become a well-established andwell-staffed center.

c

This manual is truly the result of the effort of many people. We welcome your suggestionsfor improvement so that any future revision can make the methods more useful.

Section IIntroduction

1. I BackgroundThe Clean Water Act

basis for control of toxic(CWA, 1972) provides thesubstances discharged to

waters of the United States. The Declaration of Goalsand Policy of the Federal Water Pollution Control Act of1972 states that ”. ..it is the national policy that thedischarge of toxic pollutants in toxic amounts be pro-hibited.” This policy statement has been maintained inall subsequent versions of the CWA.

It is the goal of the CWA that zero discharge ofpollutants to waters of the U.S. be achieved. Becausethis goal is not immediately attainable, the CWA allowsfor National Pollutant Discharge Elimination System(NPDES) permits for wastewater discharges. The fiveyear NPDES permits contain technology-based effluentlimits reflecting the best controls available. Where thesetechnology-based permit limits do not protect waterquality, additional water quality-based limits are includedin the NPDES permit in order to meet the CWA policyof “no toxic pollutants in toxic amounts.*’ State narrativeand numerical water quality standards are used in con-junction with EPA criteria and other toxicity databasesto determine the adequacy of technology-based permitlimits and the need for additional water quality-basedcontrols.

To insure that the CWA’s prohibitions on toxic dis-charges are met, EPA has issued a “Policy for theDevelopment of Water Quality-Based Permit Limita-tions for Toxic Pollutants” (Federal Register, 1984).This national policy recommends an integrated approachfor controlling toxic pollutants that uses whole effluenttoxicity testing to complement chemical-specific analy-ses. The use of whole effluent toxicity testing is neces-sitated by several factors including a) the limitationspresented by chemical analysis methods, b) inadequatechemical-specific aquatic toxicity data, and c) inabilityto predict the aggregate toxicity of chemicals in aneffluent.

To determine the toxicity of effluents to aquatic life,standardized methods for measuring acute and chronictoxicity have been developed by EPA (EPA, 1985A;EPA, 19886; EPA, 1989E). These cost-effective meth-ods facilitate the inclusion of whole effluent toxicitylimits and biomonitoring conditions in NPDES permitsfor facilities suspected of causing violations of statewater quality toxicity standards.

As a result of the increasing use of aquatic organ-ism toxicity limits and biomonitoring conditions in per-

mits, a substantial number of unacceptably toxic efflu-ents have been and continue to be identified. To rectifythese problems, permittees are being required, throughpermit conditions and administrative orders or otherenforcement actions, to perform effluent toxicity reduc-fion evaluations (TREs). The object of the TRE is todetermine which measures are necessary to maintainthe effluent’s toxicity at acceptable levels. Such evalua-tions, however, have often proven to be very compli-cated.

The goal of the TRE wilt be set by either the stateregulatory agency or EPA and will be dependent onstate standards that define acceptable levels of toxicityin the receiving water and effluent. Because of thts, andbecause specific TRE actions may also be required,communication between the regulators and TRE inves-tigators is crucial.

This document provides NPDES permittees withprocedures to assess the nature of effluent toxicity toaquatic organisms. It is intended for use by thosepermittees having difficulty meeting their permit for wholeeffluent aquatic organism toxicity limits or permitteesrequired, through special conditions, to reduce or elimi-nate effluent toxicity. This document does not addresshuman health toxicity concerns such as those frombioconcentration, water supplies and recreational uses.The methods are applicable to identifying the cause oftoxicity for samples other than effluents which displayacute toxicity, such as ambient water samples, elutriatesand pore waters from sediments, and possibly leachates.While we generally refer to effluents, the application ofthe techniques for any aqueous sample is implied.These methods may have applicability to effluents andother types of samples that exhibit chronic toxicity aswell.

1.2 Conventional Approach to TIESIn order to appreciate the complexities involved in a

typical effluent toxicity identification evaluation (TIE),one must first understand the drawbacks in what canbe considered the conventional approach to the prob-lem of controlling toxics. The following discussion ismeant, to exemplify the need for a logical approachwhich builds on the effluent data as they are beingcollected.

Traditionally, when an effluent has been identifiedas toxic or is suspected of being toxic to aquatic organ-isms, a sample of the wastewater is analyzed for the126 ‘*priority pollutants.” The concentration of each pri-

Toxic Eff bent

GenerateEff bent LCSO

t

Mass Balanceand

Comparison

Search Literaturefor Aquatic

ToxicityData on Effluent

Constituents

Figure l-l. Conventional approach to TIES.

ority pollutant present in the sample is subsequentlycompared to literature toxicity data for the pollutant, oris compared to EPA’s Ambient Water Quality Criteria orstate standards for aquatic life protection for that com-pound. The goal of this exercise is to determine whichpollutants in the wastewater sample are responsible foreffluent toxicity (Figure l-l). Unfortunately, determiningthe source of an effluent’s toxicity is rarely this straight-forward.

The first problem encountered in this course is oneof effluent variability. Because toxicity is a generic re-sponse, there is no way .to determine whether the

Conduct

Priority

PollutantAnalysis

Evaluate EffluentConstituents and

TheirConcentrations

toxicity observed over time is consistently caused by asingle constituent or a combination of constituents or anumber of different constituents, each acting periodi-cally to cause effluent toxicity. Experience has shownthat the latter may be a frequent occurrence especiallyin publicly owned treatment works (POTW) effluents.To further complicate the problem, the variability inconventional effluent monitoring parameters may notcoincide with variability in the effluent toxicant( Moni-toring methods for conventional parameters such asbiological oxygen demand (BOD) frequently are notresponsive to shifts in the toxicants because they are atrelatively low concentrations in the effluent or simply

l-2

because the toxicants are not amenable to analysis bythese procedures. For the conventional TIE approachto be successful, it is crucial that the same sample beanalyzed using both chemical and biological techniques,and that a number of samples over time be studied toassess the variability in the toxicant(

A second problem with the conventional approachinvolves the focus on the priority pollutants. Thesehave become known as the “toxic pollutants,” convey-ing an implication that they constitute the universe oftoxic chemicals but the priority pollutants are only a tinyfraction of all chemicals. Limiting the search to these126 compounds will result in failure to identify thecause of toxicity in most cases.

On the surface, solving this difficulty may seeminconsequential; the effluent analysis must include moni-toring techniques for “non-priority” as well as prioritypollutants. To analyze an effluent for every chemicalwould cost tens of thousands of dollars and therewould be no assurance that the detection levels wouldbe low enough. Determination of the composition of aneffluent is limited to the analyses used. For instancegas chromatography/mass spectrometry (GUMS) willnot identify cadmium and Inductively Coupled EmissionSpectroscopy (ICP) may not detect it when the concen-tration is low. The absence of a measurable quantity ofany substance at the method detection level is ofteninterpreted as meaning that it is not present in theeffluent at all or not at toxic levels.

The toxicants may be present at low concentrationsbecause only small concentrations of highly toxic chemi-cals are needed to produce toxicity. If this is true, thenlow concentrations must be measured. Such chemicalsare not easily found by examining system loadings. Forexample, if a chemical has an LC50 of 1 ug/L, 380 g(less than a pound per day) of the compound must bepresent to cause lethality in the effluent of a 100 milliongallons per day (mgd) treatment plant. With a removalefficiency of 99%, a loading of only 100 pounds per daywould be needed to produce a toxic effluent. Clearlythen, large loadings cannot be used to guide selectionof analytical techniques, and loads of a few pounds in acollection system producing 100 mgd may be next toimpossible to identify by the usual methods of estab-lishing loadings.

Many analytical methods are relatively limited intheir applicability. Even GUMS, an instrument heavilyrelied upon in typical wastewater analyses, is incapableof detecting about 80% of all synthetic organic com-pounds (G. Veith, personal communication, ERL-Duluth).This limitation is related to selection and efficiency ofsolvent extraction techniques, analyte volatility and ther-mal stability, detector specificity and sensitivity, andanalytical interferences and artifacts. The percentageof organics detected can be improved by derivatizationbut the results are much more difficult to interpret. Ingeneral, the broader spectrum methods are less sensi-tive and require higher concentrations of analytes fordetection and are costly. To detect lower concentra-

tions, more specific methods are usually more sensi-tive. To choose specific methods one must have knowl-edge of the toxicants-knowledge which does not exist,since that is the purpose of the analyses.

Surprisingly, even with these limitations, one usu-ally sees lengthy lists of effluent constituents whenanalyses are performed on wastewater. In the case ofGUMS chromatograms, large peaks of non-toxic efflu-ent constituents can overlap and hide smaller peaksthat may represent the toxicants of concern. Whenmany chemicals are present, the number of peaks thatcan be identified may be small. Failure to identify acomponent does not mean that the chemical is nottoxic. By using reference spectra, many peaks may betentatively identified as several different compoundswhich serves only to increase, not decrease, the num-ber of possibilities. No aquatic toxicity data will beavailable for most of these compounds, so toxicity datamust be generated during the study. Compounds mayneed to be synthesized in order to test them becausethey are not available commercially. For those com-pounds for which aquatic toxicity data are available, thedata may not include the species used for the TIE.Even if all this work is done, trying to pinpoint the causeof toxicity in such a complex mixture is likelyJo failbecause this approach does not include matrix effectsand toxicant bioavailability. For example, several met-als may be present in an effluent sample at concentra-tions well above the toxic threshold. These metals maynot be the source of the effluent’s toxicity, however,because they are not biologically available. Character-istics such as total organic carbon (TOC), total sus-pended solids (TSS), ionic strength, pH, hardness andalkalinity can change toxicity. The inability to quantitatethe effects these parameters have on toxicity furtherdecreases the chances for a successful TIE.

1.3 Toxicity Based ApproachThe approach described in this manual uses the

responses of organisms to detect the presence of thetoxicant during the first stages of the TIE. In this way,the number of constituents associated with the toxi-cants can be reduced before analyses begin and someknowledge of physical/chemical characteristics is gained.This approach simplifies the analytical problems andreduces cost. Some of the problems limiting the con-ventional approach can be used to enhance the suc-cess of this alternate approach.

There are two main objectives in the first step ofthis approach. First, characteristics of the toxicants(e.g., solubility, volatility) must be established. Thisallows them to be separated from other non-toxic con-stituents to simplify analyses and enhance interpreta-tion of analytical data. Secondly, throughout the TIE,one must establish whether or not the toxicity is consis-tently caused by the same substances. Failing to es-tablish the variability related to the toxicants could leadto control choices that do not correct the problem.

Knowledge of the physical/chemical characteristicsof the toxicants aids in choosing the appropriate ana-

l -3

lytical method. Such information also may be useful inselecting an effluent treatment method.

Figure l-2 is a flowchart representation of a TRE.This document details the toxicity characterization pro-cedures (Phase I). Phase II (toxicant identification) andPhase III (toxicant confirmation) usually follow Phase I.Two other EPA manuals (EPA, 1989A; 19898) can beconsulted for more information on bench scale and pilotplant effluent toxicity treatability studies and sourcecontrol options.

Phase I tests characterize the physical/chemicalproperties of the effluent toxicant using effluent ma-nipulations and accompanying toxicity tests. Each char-acterization test in the Phase I series is designed toalter or render biologically unavailable a group of toxi-cants such as oxidants, cationic metals, volatiles, non-polar organics or chelatable metals. Aquatic toxicitytests, performed on the effluent before and after theindividual characterization treatment, indicate the effec-tiveness of the treatment and provide information onthe nature of the toxicant( By repeating the toxicitycharacterization tests using samples of a particulareffluent collected over time, these screening tests willprovide information on whether the characteristics ofthe compounds causing toxicity remain consistent. Thesetests will not provide information on the variability oftoxicants within a characterization group. Knowing thatthe toxicants have similar physical/chemical propertiesmeans that they can be identified in Phase II usingsimilar techniques. With successful completion of PhaseI, the toxicants can be tentatively categorized as cat-ionic metals, non-polar organics, oxidants, substanceswhose toxicity is pH dependent, and others. Informa-tion on physical/chemical characteristics of the toxi-cants will indicate filterability, degradability, volatility,and solubility. Either of two choices is available in thesecond phase of testing, i.e., toxicant treatability ortoxicant identification studies.

Toxicant identification is described in Phase II (EPA,1989C). Phase II involves several steps, all of whichrely on tracking the toxicity of the effluent throughoutthe analytical procedure. Although effluent toxicantsare partially isolated in the first phase of the study,further separation from other compounds present in theeffluent is usually necessary. Techniques are availableto reduce the number of compounds associated withthe toxicants. Unlike Phase I procedures, Phase IImethods will be toxicant-specific. Currently availabletechniques in Phase II are for identifying non-polarorganics, EDTA chelatable metals, and ammonia.Enough information exists now to add a section for

surfactants. Additional procedures for other toxicantswill be added as they are developed. Once the toxi-cants have been adequately isolated from other com-pounds in the effluent and tentatively identified as thecausative agents, final confirmation (Phase III) can be-gin.

Like Phase I, Phase III (EPA, 19890) containsmethods generic to all toxicants. No single test pro-vides irrefutable proof that a certain chemical is caus-ing effluent toxicity. Rather, the combined results of theconfirmation tests are used to provide the “weight ofevidence” that the toxicant has been identified.

Once the toxicant has been identified, it can betracked through the process collection system usingchemical analyses. Toxicity cannot be used to find thesource for untreated wastes because toxicity from otherconstituents that are toxic in untreated waste but re-moved by treatment, will confuse the results. Of course,using bench- or pilot-scale systems and measuringtoxicity on treated waste, is feasible. c

TIES require that toxicity be present frequentlyenough and endure storage (that is, the toxicity is notrapidly degrading) so that repeated testing can char;ac-terize and subsequently identify and confirm the toxi-cants in Phases II and III. Therefore, enough testingstiould be done to assure consistent presence of toxic-ity before TIES are initiated. This is done not to validatea given test but to establish the sufficient and frequentpresence of toxicity.

The methods described herein are applicable pri-marily to acute toxicity. Chronic toxicity identificationmethods are being developed (EPA, 1991A). In somespecial cases in which toxicity can be concentrated (asin the non-polar organic section of Phase II) one maybe able to “convert” chronic toxicity to acute toxicity byconcentration and successfully identify what is causingthe chronic toxicity.

To be successful, TIES must be conducted bymultidisciplinary teams whose members must interactdaily so that toxicologists and chemists are aware ofthe many concerns that affect test results. Speed isusually important because effluents may decay duringstorage. Often subsequent tests cannot be designeduntil the results of the previous ones are known. Obvi-ously then, waiting a week for analytical or toxicologicalresults may preclude more work while the effluentsample undergoes changes during the waiting period. Ifthis happens, one must begin again on a new samplein which case resources are not being used effectively.

l-4

1 Effluent Sample 1

Phase I

Toxicant Characterization Tests

TREATABILITY

APPROACH or Identify Toxicant

IDENTIFY TOX

Phase IIToxicant Identification Analyses

Phase IllToxicant Confirmation Procedures

ICANT

Considerations

Control Method Selectionand Implementation

*,

Figure 1-2. Flow chart for toxicity reduction evaluations.

l-5l-5

Section 2Health and Safety

Working with effluents of unknown composition isthe nature of toxicity identification evaluations. There-fore safety measures must be adequate for a widespectrum of chemicals as well as biological agents.From the type of treatment used one may be able tojudge probable concerns. For example, extended aera-tion is likely to minimize the presence of volatile chemi-cals and chlorinated effluents are less likely to containviable pathogens.

Exposure to the wastewater during collection andits use in the laboratory should be kept at a minimum.Inhalation and dermal adsorption can be reduced bywearing rubber gloves, laboratory aprons or coats, safety.

glasses, and respirators, and by using laboratory hoods.Further guidance on health and safety for toxicity test-ing is described in Walters and Jameson (1984).

In addition to taking precautions with effluentsamples, a number of the reagents that might be usedduring Phase II toxicant identification and Phase IIItoxicant confirmation studies are known or suspectedto be very toxic to humans. Analysts should familiarizethemselves with safe handling procedures for thesechemicals (DHEW, 1977; OSHA, 1976). Use of thesecompounds may also necessitate specific waste dis-posal practices.

L

2-l

-

Section 3Quality Assurance

Quality assurance is composed of two aspects,quality verification and quality control. Quality verifica-tion entails a demonstration that the proposed studyplan was followed as detailed and that work carried outwas properly documented. Some of the aspects ofquality verification include chain of custody procedures,statements on the objective of the study and what isknown about the problem at its outset, instrumental logbooks, and work assignments. This aspect of qualityassurance ensures that a “paper trail” is created toprove that the work plan has been covered completely.The quality control aspect of quality assurance involvesthe procedures which take place such as the number ofsamples to be taken and the mode of collection, stan-dard operating procedures for analyses, and spikingprotocols.

No set quality assurance program can be dictatedfor a TIE; the formula to a successful study will beunique to each situation. However, adherence to somegeneral guidelines in formulating a Quality AssurancePlan (QAP) may increase the probability of success.

In preparing a QAP, enough detail should be in-cluded so that any investigator with an appropriatebackground could take over the study at any time.Cross checking of results and procedures should bebuilt into the program to the extent possible. Recordsshould be of a quality that can be offered as evidencein court. Generally, the QAP should be provided in anarrative form that encourages users to think aboutquality assurance. To be effective, the QAP must bemore than a paper exercise simply restating standardoperating procedures (SOPS). It must increase commu-nication between clients, program planners, field andlaboratory personnel and data analysts. The QAP mustmake clear the specific responsibilities of each indi-vidual. The larger the staff, the more important thisbecomes. While QAPs may seem to be ‘an inconve-nience, the amount of effort they require is commensu-rate with the benefits derived.

3. I TIE Quality Control Plans

A successful TIE is dependent upon a strong qual-ity control program. Obtaining quality TIE data is moredifficult because the constituents are unknown in con-trast to quality control procedures for a standard ana-lytical method for a specific chemical. In such an analy-sis, one knows the characteristics of the analyte andthe implications of the analytical procedure being uti-lized. Without knowledge of the physical/chemical char-

33

acteristics of the analyte, however, the impact of vari-ous analytical procedures on the compound in questionis not known. Further, quality control procedures arespecific to each compound; quality control proceduresappropriate to one analyte may be completely inappro-priate to another.

The problem of quality control is further exagger-ated because quality control procedures for aquatictoxicity tests may be radically different from those re-quired for individual chemical analyses. This additionaldimension to quality control requires a unique frame-work of checks and controls to be successful. Theimpacts of chemical analytical procedures on sampletoxicity must be included. Likewise, procedures used toinsure quality toxicity test results should not impactchemical analyses. For example, in performing stan-dard aquatic toxicity tests, samples with low dissolvedoxygen (DO) are usually aerated. This practice may,however, result in a loss of toxicity if the toxicant isvolatile or subject to oxidation.

3.2 Cost Considerations/ConcessionsThe quality control practices required in any given

experiment must be weighed against the importance ofthe data and decisions to be based upon it. The crucialnature of certain data will demand stringent controls,while quality control can be lessened in other experi-ments having less impact on the overall outcome.

Effluent toxicant identification evaluations require alarge number of aquatic toxicity tests. The decision touse the standard toxicity test methods described inEPA (1985A; 19918) (involving a relatively high degreeof quality control), must be weighed against the degreeof complexity involved, the time required and the num-ber of tests performed; all of these affect the cost oftesting. For this reason, toxicity tests used in the earlyphases of the evaluation generally do not follow thisprotocol, nor do they require exacting quality controlsbecause the data are only preliminary. Phase I, and toa lesser extent, Phase II results are more tentative innature as compared to the tests performed for theconfirmation of the effluent toxicant in Phase III.

The progression towards increasingly definitive re-sults is also reflected in the use of a single species inthe initial evaluation studies and multiple species in thelater stages. The use of several species of aquaticorganisms to assure that effluent toxicity has beenreduced to acceptable levels is necessary becausespecies have different sensitivities to the same pollut-11

ant. Quality control must relate to the ultimate goal ofattaining and maintaining the designated uses of thereceiving water. For this reason, final effluent test re-sults must be of sufficient quality to ensure ecosystemprotection. The use of dilution water for the toxicitytests which mimics receiving water characteristics (inhardness and pH) will help to ensure that the effluentwill remain non-toxic after being discharged into theenvironment. In the instances where the effluent domi-nates the receiving water, the dilution water shouldmimic the water chemistry characteristics of the efflu-ent. This is discussed in Section 5, Dilution Water. Inaddition, it is essential that the variability in the causeof effluent toxicity be defined during the course of theTIE so that appropriate control actions provide a finaleffluent safe for discharge.

3.3 VariabilityThe opportunities to retest any effluent to confirm

the quality of initial TIE results will be limited at best. Inaddition to the shifting chemical and toxicological na-ture of the diseharge over time, individual effluentsamples stored in the laboratory change. Effluent con-stituents degrade at unknown rates, as each compoundhas its own rate of change. The change in a sample’stoxicity over time represents the cumulative change inall of the constituents, plus that variation resulting fromexperimental error. Some guidelines for assessing andminimizing changes in sample chemistry and toxicityare discussed in Sections 6 and 8. Regardless of theprecautions taken to minimize sample changes, asample cannot be retested with certainty that it has notchanged.

3.4 lntra-Laboratory CommunicationQuality control procedures in chemistry and biology

can be quite different. For example, phthalates are afrequent analytical contaminant requiring special pre-cautions that are not of toxicological concern. The toxi-cological problem presented by the zinc levels typicallyassociated with new glassware are of no concern tothose performing organic analyses. The difference inglassware cleanup procedures is an example of manydifferences that must be resolved. Cleaning proceduresmust be established to cover the requirements of both.Time schedules for analyses must be detailed in ad-vance. One cannot assume compound stability; there-fore, time delays between the biological and chemicalanalysis of a sample cannot be tolerated.

3.5 Record KeepingThroughout the TIE, record keeping is an important

aspect of quality verification. All observations, includingorganism symptoms, should be documented. Detailsthat may seem unimportant during testing may be cru-cial in later stages of the evaluation. Investigators mustrecord test results in a manner such that preconceivednotions about the effluent toxicants are not unintention-ally reflected in the data. TIES required by state orfederal pollution control agencies may require that someor all records be reviewed.

3.6 Phase I ConsiderationsEffluent toxicity is “tracked” through Phases I, II

and III using aquatic organisms. Such tracking is theonly way to detect where the toxicants are until theiridentity is known. The organism’s response must beconsidered as the foundation and therefore, the toxicitytest results must be dependable. System blanks (blanksamples carried through procedures and analyses iden-tical to those performed on the effluent sample) areused extensively throughout the TIE in order to detecttoxic artifacts added during the effluent characterizationmanipulations. With the exception of tests intended tomake the effluent more toxic, or situations in which aknown amount of artifactual toxicity has been intention-ally added, sample manipulation should not cause theeff,iuent toxicity to change.

There are many sources of toxicity artifacts in PhaseI. These include: excessive ionic strength resulting fromthe addition of acid and base during pH adjustment,formation of toxic products by acids and bases, con-taminated air or nitrogen sources, inadequate mixing oftest solutions, contaminants leached from filters, pHprobes, solid phase extraction (SPE) columns, and thereagents added and their contaminants. The appropri-ate toxicity data for the reagent chemicals usti inPhase I and common aquatic test organisms are pro-vided as needed in subsequent sections of this docu-ment.

,

Frequently toxic artifacts are unknowingly intro-duced. For example, pH meters with refillable elec-trodes can act as a source of silver which can reachtoxic levels in the solutions being measured for pH.This is especially a problem where there is a need tocarefully maintain or track solution PH. Using pH elec-trodes without membranes avoids the silver problem(which can only be detected by profuse use of blanks).

Oil in air lines or from compressors is a source ofcontamination. Simple aeration devices, such as thosesold for use with aquaria are better as long as cautionis taken to prevent contamination of the laboratory airwhich is taken in by the pump.

Worst case blanks should be used to better ensurethat toxicity artifacts will be recognized. Test chambersshould be covered to prevent contamination by dustand to minimize evaporation. Since small volumes areoften used, evaporation must be controlled. Plastic dis-posable test chambers are recommended to avoid prob-lems related to the reuse of test chambers. Cups fromthe same lot should be spot-checked for toxicity.

Glassware used in various tests and analyses mustbe cleaned not only for the chemical analyses but sothat toxicity is not introduced either by other contami-nants or by residues of cleaning agents. Since theorganisms are sensitive to all chemicals at some con-centrations, all toxic concentrations must be removedand not just those for which analyses are being made.

Randomization techniques, careful observance oforganism exposure times and the use of organisms of

3-2

approximately the same age ensure quality data. Stan-dard reference toxicant tests should be performed withthe aquatic test species on a regular basis and controlcharts should be developed (EPA, 1985A; 1991B). Dur-ing Phase I it will not be known how much the toxicityof the reference toxicant varies over time compared tothe toxicant( When the toxicants are known, theyshould be used as the reference toxicant. Referencetoxicant tests should be performed to coincide with theTIE testing schedule.

3.7 Phase II ConsiderationsIn Phase II, a more detailed quality control program

is required. Interferences in toxicant analysis are for themost part unknown initially but as toxicant identifica-tions are made, interferences can be determined. Like-wise instrumental response, degree of toxicant separa-tion, and detector sensitivity can be determined asidentifications proceed.

3.8 Phase 111 ConsiderationsIn Phase III of a TIE, the detail paid to quality

control and verification is at the maximum. This phaseof the study responds to the compromises made todata quality in Phases I and II. For this reason, confi-dence intervals for toxicity and chemical measurements

must be calculated. These measurements allow thecorrelation between the concentration of the toxicantsand effluent toxicity to be checked for significance basedon test variability. Effluent manipulations prior to chemi-cal analysis and toxicity testing are minimized in thisphase in an effort to decrease the chance for produc-tion of artifacts. Field replicates to validate the precisionof the sampling techniques and laboratory replicates tovalidate the precision of analyses must be included inthe Phase III quality control program. System blanksmust be provided. Calibration standards and spikedsamples must also be included in the laboratory qualitycontrol program. Because an attempt will be made tocorrelate effluent toxicity to toxicant concentration, spik-ing experiments are important in determining recoveryfor the toxicant( These procedures are feasible inthis phase of the study because the identities of thesubstances being measured are known.

The toxicants being analyzed can be tested usingpure compounds, thereby alleviating the need for ageneral reference toxicant. Because the test organismalso acts as an analytical detector in the correlation ofeffluent toxicity with toxicant concentration, changesin the sensitivity of the test organisms must be known.This is best achieved by using the same cheFicalsidentified for the reference toxicants.

3-3

Section 4Facilities and Equipment

The facilities, equipment and reagents needed toperform an effluent TIE will depend on the phase of thestudy and the characteristics of the toxicant( Theequipment required for Phase I characterization tests isdescribed throughout Section 8. The facility and equip-ment needs in Phase II of the TIE will be site-specificand will depend both on the physical/chemical charac-teristics of the toxicants and on the choice of the PhaseII approach.

Phase I requires only basic analytical and toxicitytesting equipment which would be available in mostlaboratories where toxicity tests and the usual waterchemistry analyses are performed. Phase III require-ments are largely limited by equipment found in atypical toxicity testing lab and equipment necessary forthe analysis of the toxicant(

Because of the equipment needs and time requiredto conduct the evaluations, complete on-site effluent

TIES using a mobile laboratory are generally not fea-sible. Measurement of the loss of toxicity over time inseveral effluent samples will provide information uponwhich to base acceptable storage times. Usually, withmodern rapid sample shipment methods, off-site workis practical. The cost of shipment is usually far lessthan the cost of on-site work.

Large numbers of organisms and many tests areneeded for TIES. Ready availability of test organisms isimportant because often the test(s) needed are notpredictable. Only after the results of one experimentare known can the next test be planned. It is probablymore economical to culture many of the test speciesthat might be used in TIES than it is to purchas#them.A delay in testing caused by shipment time or lack ofavailability of test organisms could cost far more inwork loss than it would cost to maintain cultures formany weeks.

4-1

Section 5Dilution Water

The choice of dilution water will change with thepurpose of the tests and therefore the choice will oftenbe more varied in Phases I and II than in Phase III.Particularly for some toxicant groups in Phase II, somevery unusual dilution water is recommended in order toachieve the desired chemical conditions. Sometimesthe water may in itsetf be toxic! Such concepts areforeign to conventional toxicology and rightly so, butthis is not conventional toxicology.

Much of Phase I and parts of Phase II utilizeorganism tolerance and relative toxicity to accomplishthe objectives of the study. Methanol, hydrogen ionconcentration, and osmotic pressure may sometimesbe near lethal levels in order to test necessary condi-tions. In some cases, the dilution medium may causecomplete mortality in 48 h, but the point of interest iswhether treatment causes more rapid mortality. If so,one can say that one condition is more toxic thananother and obtain important information from the test.The key is to run sufficient numbers of system blanksso that the relative contribution to mortality is knownand toxicity is not attributed to an incorrect cause.These are examples of the previous statement thatthese methods “utilize tolerance and relative toxicity.”In reality, this approach is very much like the compari-son of the toxicity of two chemicals, A and B. If onedetermines LC5Os for A and B and concludes that A istwice as toxic as B, lethal conditions are being com-pared in order to say this. Controls are not involved inthe LC50 calculation and high control survival does notchange the data interpretation, The same concept ofrelative toxicity is used here. Chemical “A,, is the blankand chemical “B” is the treated sample and the ques-tion is, ‘,which is more toxic?“.

As these methods are built on tolerance (i.e., sur-vival), chronic toxicity endpoints cannot be used andthat is why these methods are primarily intended foracute toxicity. Obviously, if one wants to measure chroniceffects, the test organisms must be able to live longenough to display chronic effects. Many of the pHchanges and other manipulations used in these meth-ods do not allow sufficient survival time or health forreproduction or growth. For chronic TIES, more atten-tion has to be given to acclimation, feeding and generalliving conditions (EPA, 1991 A).

Many of the additives used in the Phase I manipu-lations change the mixture of the effluent much morethan the dilution water. In general, for Phase I, anywater which is of a consistent quality and which willsupport growth and reproduction of the test species issuitable. We have found the use of a dilution water thathas a hardness similar to that of the effluent or thereceiving water to be beneficial. A variety of dilutionwater choices are provided by EPA (1985A; 1989E)and any of these may be used for TIES.

In Phase III, where the objective is to confirm thetrue cause of toxicity, where artifacts are to be ex-cluded to the extent possible and where absolute toxic-ity is more important than relative toxicity, practicesincluding choice of dilution water, must follow conven-tional toxicological methodology. Tolerance to additivesmust not be necessary in order to provide the desiredresponse. Attention must be given to simulation of thedilution water into which the effluent is discharged.Some toxicant dose response relationships may betotally different as the water quality characteristicschange. These factors must be incorporated into PhaseIII where absolute toxicity is of the utmost concern. InPhases I and II, only relative differences are beingconsidered.

Perhaps a cautionary note is warranted regardingthe effects of dilution water on effluent toxicity. If highconcentrations of effluent are being tested (e.g., 80%)the physical/chemical characteristics will resemble thoseof the effluent. If low concentrations are tested (e.g.,5%) then characteristics will resemble those of thedilution water.

Little specific information can be given about theselection of dilution water in Phases I and II except thatthe desired tested conditions will often dictate its char-acteristics. For example, in Section 8.6, the same col-umn used for the blank may not be usable for theeffluent sample if receiving water is used as the dilutionwater. Secondly, sufficient numbers of blanks must beincluded to interpret the results. In Phase III, the choiceof the appropriate dilution water should be based onthe characteristics of the receiving water where thedischarge occurs.

5-15-1

.

Section 6Eff bent Sampling and Handling

A wastewater sample may be representative onlyof the discharge at the time of sampling. In effect, eachsample is a “snapshot” of the effluent’s toxicologicaland chemical quality over time. To determine whetherany effluent sample is typical of the wastewater mayrequire the collection of a large population of samples.Further, what constitutes a “representative” sample is afunction of the parameter of concern. Because effluentsvary in composition, sampling must be extensive enoughthat one is confident that the groups of samples repre-sent the discharge over time. Guidelines for determin-ing the number and frequency of samples required torepresent effluent quality are contained in the “Hand-book for Sampling and Sample Preservation of Waterand Wastewater” (Berg, 1982). However, since thisguidance is not based on toxicity, it should be usedwith caution.

Both quantitative (change in concentration) andqualitative (change in toxicants) variability commonlyoccur in effluents and both may affect toxicity. Changesin effluent toxicity are the result of varying concentra-tions of individual toxicants, different toxicants, chang-ing water quality characteristics (affecting compoundtoxicity) and analytical and toxicological error. Even ifthe toxicity of an effluent to an aquatic organism isrelatively stable, this does not mean that there is only asingle toxicant causing toxicity in any given sample oramong several samples.

Determining whether a sample is typically toxic isnot as simple as comparing the conventional pollutantsof the sample to long-term effluent averages. Effluenttoxicants often do not follow the same trends as BOD,TOC and TSS. The toxicant may be present at sucha low level that it does not significantly affect the quan-tity of the conventional pollutant, even though it ispresent in toxic concentrations.

Conventional parameters, BOD, TSS, and otherpollutants limited in the facility’s NPDES permit, willprovide an indication of the operational status of thetreatment system on the day of sampling. For industrialdischarges, information on production levels and typesof operating processes may be helpful. The condition ofthe facility’s treatment system at the time of samplingshould be determined by the individual collecting thesample. The type of sample, time of collection, andother general information on the facility should be re-corded. An example of a page of a log book is given inFigure 6-1.

Upon the arrival of the sample in the laboratory,temperature, pH, toxicity, hardness, conductivity, totalresidual chlorine (TRC), total ammonia, alkalinity, andDO should be measured. Toxicity should be measuredperiodically during storage to document any changes(cf., Section 8).

Investigators should not be surprised to find thatwell operated municipal and industrial treatment sys-tems discharge unacceptably toxic wastewaters. Eff lu-ent guideline-based limits which reflect best achievabletechnology, do not prescribe limits for more than a fewchemicals. Many compounds present in effluents arenot regulated because the discharger is not required toreport their presence in permit applications or theycannot be detected using typical methods for wastewa-ter analysis.

For chlorinated effluents, whether sampling shouldbe done before chlorination depends on the question tobe answered. Sometimes the question may be whetheror not there are toxicants other than chlorine present.Dechlorination prior to toxicity characterization may beneeded in order to distinguish toxicity from causesother than chlorine. Usual methods of dechlorinationmay remove more than toxicity from chlorine alone andcareful data interpretation is needed to understand theresults. Toxicity from more than one cause is often notadditive in effluents, so relative contributions from twoor more causes can be very hard to decipher.

The choice of grab or composite samples will de-pend on the specific discharge situation, (e.g., plantretention time) questions to be answered by the TIEand the stage of the TIE. In Phase I testing, samplesthat are very different from one another give resultsthat are difficutt to interpret; therefore composite samplesare more similar and are easier to. use. In Phase III,effluent variability is used to advantage; therefore, grabsamples are often best. If toxicity is low or intermittentlypresent, grab samples may be best during all phases.The additional difficulty of getting flow proportionalsamples should be balanced against their advantage ineach situation. While grab sampling may provide maxi-mum effluent toxicity, it is more difficult to catch peaksin toxicity and Phase I sampling may require more time.EPA (1985A; 1991 B) discusses the advantages anddisadvantages of grab and composite sampling andhave also detailed methods for sampling intermittentdischarges.

6-16-1

Figure 6-1. Example data sheet for logging in samples.

Sample Log No.: Sample Type: Cl Grab 0 Composite

Date of Arrival: 0 Glass 0 Plastic

Date and Timeof Sample Collection:

cl0cl

PrechlorinatedChlorinatedDechlorinated

Facility:

Location: Sample Conditions Upon Arrival:

NPDES No: TemperaturePH

Contact:

Phone Number:

Sampler:

Condition of treatment system at time of sampling:

Status of process operations/production (if applicable):

iota1 AlkalinityTotal HardnessConductivity/SalinityTotal Residual ChlorineTotal Ammonia

l

Comments:

6-2

If the TIE analyses are not conducted on-site,samples must be shipped on ice to the testing location.Effluent samples should not be filtered prior to testingunless it is necessary to remove other organisms.Sample filtration could affect the results of the charac-terization tests, one of which entails filtering the efflu-ent. Sample aeration should also be minimized duringcollection and transfer. Initial sample analysis shouldbegin as soon as practical after effluent sampling. PhaseII and especially Phase III may require specific types ofsample containers or the addition of preservative toaliquots of sample designated for chemical analyses.For a single Phase I test series, 3 L of effluent areneeded for analysis if test organisms such as daphnidsor newly hatched fathead minnows are used. The exactvolume required depends on the toxicity of the effluentand to a lesser extent, the test options chosen (cf.,Section 8). For other species different volumes may benecessary. Volumes frequently used in each character-ization test are supplied in Table 6-l.

The extent of the analyses carried out on anyindividual sample must be weighed against the cost of

Table 6-1. Volumes needed for Phase I tests

Volume for TotalCharacterization Step Each Step’ Volumes2 (mL)

Chemical analyses3 _- <500

pH 3 Adjustment 30 -300filtration 235solid phase extraction 200aeration 35

pH 11 Adjustment 30 -300filtration 235solid phase extraction’ 200aeration 35

Unadjusted pH effluent (pH ;)” -590initial test 40baseline test 80filtration 235solid phase extraction 200aeration 35EDTA additions 100sodium thiosulfate additions 100

Graduated pH . -120-1000PH 6 40-500PH 7 40-500PH 8 40-500

11 Amount is dependent on effluent characteristics.22 Total volume is -3 L; this is maximum needed, does not include

subsequent testing.33 These include temperature, pH, hardness, conductivity, TRC,

total ammonia, alkalinity, and DO.4 The pH is readjusted to pH 9 before it is put through the C,, SPE

column.5 The pH i of the effluent is the initial pH of the effluent sample. It

may be important to know the pH at the point of discharge aswell as the receiving water pH and to know the pH of theeffluent at air equilibrium.

additional sampling, the stability of the sample, samplerepresentativeness and the need to have samples ofdifferent toxicity. Clearly, the resources required forsuch TIES are too great to expend on a single sampleor on a few samples which do not represent the efflu-ent. Likewise, there is not a set number of sampleswhich should be analyzed in Phases I, II or III beforegoing on to subsequent phases of the study or takingfinal measures to control effluent toxicity. The numberof samples analyzed in each phase will be a function ofthe apparent variability in the effluent, the number oftoxicants, how persuasive the data are, the cost of theremedial action, regulatory deadlines and finally, thesuccess of each study phase.

6.7 Sample Shipment and Collection inPlastic versus Glass

Effluent samples often have been collected, shippedand stored in various types of plastic (e.g., polyethyl-ene) containers rather than glass. However, with a feweffluents, we have noted that samples shipped andstored in glass were more toxic and retained theirtoxicity longer than split samples shipped and stored inplastic. This effect appeared to be due to adsorption ofcertain types of toxicants (e.g., surfactants) to theslas-tic. For these instances the samples in glass weremore representative of the effluent, and thus for TIEpurposes were preferable to the samples in plastic.

An easy way to check whether or not there is adifference in the toxicity of samples shipped and storedin glass containers versus those shipped and stored inplastic containers, is to test two or three sets of effluentsamples. Effluent should be collected in glass or stain-less steel, then a portion shipped in glass and anotherportion shipped in plastic. Baseline toxicity tests (cf.,Section 8) are conducted on each, perhaps on days 4and 7 after receipt. If the initial toxicity of the sample issimilar for both the plastic and the glass containers,and the toxicity for samples from the two containers issimilar over time (i.e., over storage time), it is appropri-ate to have the effluent samples shipped and stored inplastic containers. However, if effluent shipped andstored in glass appears to be more toxic, and retainsthe toxicity longer than the effluent sample shipped andstored in plastic, glass containers should be used for allshipments and storage for that particular effluent. Thesesame considerations also apply to the sampling/collect-ing equipment. Collection, shipment, and storage ofeffluent samples in glass may involve more effort thanplastic containers. The use of glass containers forsamples that retain their toxicity longer might result inmore rapid and cost effective progress through the TIEbecause fewer samples might be required for identifica-tion of effluent(s). Since only certain classes of com-pounds are expected to adsorb to plastic containers(e.g., surfactants), if the effluent is more toxic in glass,this can be a useful piece of information for characteriz-ing the toxicants.

6-3

a

Section 7Toxicity Tests

7.1 Principles-

Acute lethality tests with aquatic organisms areutilized throughout the toxicity characterization proce-dures described in this manual as well as in Phases IIand III. Using toxicity for such evaluations is logicalsince toxicity triggers the TIE requirement. In thesetests the organism acts as the “detector” for chemicalscausing effluent toxicity. As such, they provide the trueresponse regardless of the outcome of other analyses.The toxicity test is the only analytical procedure thatcan be used to measure toxicity. Until the cause oftoxicity is known, chemical methods cannot be used toidentify and quantify the toxicants.

There are a number of consequences associatedwith this reliance on toxicity. The organism responds toevery constituent, provided that it is present above athreshold level either individually or collectively if theconstituents are additive. While this general responseto any compound presents an advantage as a broadspectrum test for toxicants, it requires considerableeffort to determine the primary cause of toxicity be-cause it is not specific. This non-specific responsenecessitates a generic chemical/physical characteriza-tion of toxicants during Phase I testing before Phase IIidentification is begun.

A further repercussion of this universal response isthe probability of artifactual toxicity. Because the ana-lyst is reliant upon the organism’s ability to track toxicitythroughout the effluent characterization steps, samplemanipulations are constrained. While characterizing theeffluent, no manipulation should change the toxicity ofthe sample in an unpredictable manner. “Toxicity-blanksand controls” are helpful but the difficulties associatedwith them are far greater than those connected withchemical analyses because of their non-specificity. Asa result many more blanks are employed in TIE testingthan in chemical analyses or standard toxicity testing.Negative blank toxicity cannot be assumed regardlessof past results. Quite unexpected sources of artifactualtoxicity will occur in the course of conducting an evalu-ation.

For some Phase I tests the corresponding blanks(treatments on the dilution water) do not provide com-pletely relevant information concerning the effect of themanipulation on the effluent. For example, blanks ofthe graduated pH test (Section 8.9) are not particularlyuseful whether the pH is adjusted with acids, bases, orCO,. The amount of the acid or CO, used to adjust and

maintain the same pH for an effluent sample and ablank are often radically different due to the differencesin the buffering capacity of each of the solutions. Sincethe matrix of the effluent and dilution water are differ-ent, the pH in each solution will change at differentrates during the toxicity tests. Therefore the blanks arenot representative of what is occurring in the effluenttest and the controls exposure does not provide infor-mation on the manipulation effect on the test organ-isms. The use of blanks for the other manipulationsteps is relevant, and they provide information on clean-liness of the acids and bases added, the air system,filter apparatus, and SPE columns.

7.2 Test Species l

Just as different analytical methods have differentdetection levels for the same chemical, different spe-cies have different sensitivities to the same toxicants.The major difference is that the toxicity measurement isnon-specific to chemicals and so for an unknown mix-ture (effluent, sediment pore water) one must deter-mine whether a different toxicity value for the sample iscaused by the organisms different sensitivity to thesame toxicant or to different toxicants.

The choice of species to use for the toxicity testcan change the conclusion reached. In addition to theobvious need to use species of an appropriate size,age, availability, and adaptability to test conditions,there are other important considerations. An effluenttoxic to two species, having equal or different LCSOsmay be toxic because of different toxicants. Differencesof 1,000x in sensitivity are common and differences of10,000x occur among species exposed to a singlechemical. Anyone involved in identifying the cause oftoxicity of an effluent will be concerned because some-one has found the effluent toxic to some organism. Ifthat is not the case, before a TIE is begun, one shoulddetermine to which organisms the toxicity concern isdirected.

Many effluents will be received for TIES becausethey have been found toxic to the cladocerans,Ceriodaphnia or Daphnia--species well suited to TIEmethods. TIE test species selection is obvious in theseinstances. Where toxicity concern is based on species(trout or mysid shrimp), that are not going to be the TIEtest species, one must demonstrate that the toxicity ofconcern has theby the speciesdepends on theI

7-17-1

same cause as the toxicity manifestedto be used in the TIE. The difficultyeffluent characteristics (especially tox-

icity variability), the number of TIE steps which affecttoxicity and the difference in sensitivity between thespecies being compared. Since this problem has notbeen one we have experienced frequently, our sugges-tions are certainly not all-inclusive. The final confirma-tion (Phase III) methods are designed to show whetherthe wrong toxicant was identified. However, many re-sources may be consumed before reaching that stageand earlier assurances should be obtained if reason-able, to save time and cost.

One approach is to compare the LC50 values ofwhole, unaltered effluent samples for the species origi-nally raising the toxicity concern and the selected spe-cies for the TIE. If the acute toxicity varies similarly foreach species among samples then there is evidencethat the two species are responding to the sametoxicant( If the LC50 values vary differently for thetwo species, there is evidence that the toxicants aredifferent. If the LC50 values among samples do notvary more than the precision of the test method thisapproach is useless for that effluent. Successful appli-cation of this approach does not require equal sensitiv-ity of the two species or the greatest toxicity of the TIEorganism, but rather sensitivity of the species to thesame toxicant.

If in Phase I, several steps (e.g., pH decrease,aeration, solid phase extraction) all changed toxicity,and if the direction and relative magnitude of changewas the same for both test species, then there isevidence that both are sensitive to the same toxicant. Ifone or more parameters are different, the evidence isstrong that the toxicants are different. This is not to saythat if a Phase I technique completely removes toxicityto one species, it will remove it to the same extent for

_ the other species. Because different species have dis-similar sensitivities to the same chemical, removal of90% of a compound in an effluent sample may lead toa non-toxic concentration to one species while onlyreducing the toxicity to another species. If the Phase Iprocedures that successfully remove or reduce effluenttoxicity differ by test species, it is unlikely that toxicity is

caused by the same chemical(s).

Symptom comparisons are useful, especially if oneis comparing similar organisms. Comparing fish symp-toms to Daphnia symptoms could be very misleadingbut comparing symptoms of Daphnia magna to those ofCeriodaphnia dubia should be relatively safe. If onefinds comparable symptoms, the evidence is not con-vincing because many toxicants cause specific symp-toms but if symptoms are distinctly different, the evi-dence is strong that the toxicants are different. This istrue only when symptoms are compared at effluentconcentrations that are the same multiple of the LC50for each species. For example, if two species haveLC50 values of 10% and 90%, comparing symptoms at100% concentrations could be misleading. At 100%effluent, the species with an LC50 of 10% might experi-ence the symptoms so fast that their sensitivity wouldappear completely different from those of the less sen-

sitive species. Experience will reveal additional tech-niques that can be used.

Freshwater discharges to saline receiving waterrequire separate considerations. Sea salts can be addedto raise the salinity of the effluent (EPA, 1991 B) enoughso that marine species can be used in the TIE. How-ever, the tolerances of marine organisms to the addi-tives and effluent manipulations have not been deter-mined. To do so is costly and time consuming and amore efficient method may be to use a freshwaterspecies in Phase I and II. If this is done, data must begathered to show that the freshwater species chosen issufficiently sensitive and is responding to the sametoxicant as the marine species. The principles ofdoing this are the same as described above for differ-ent freshwater species. When Phase III is reached,marine species should be used, but in that phase,manipulations and additives are minimal and little ancil-lary data are needed in order to use marine species.

For discharges with conductivities comparable tobrackish or marine water, caution is in order. Mostmethods for measuring “salinity” (conductivity or refrac-tion) are non-specific for NaCI, which is the principalcomponent of sea water. Marine organisms accomplishosmotic regulation by regulating sodium and chloride. Ifsalinity of an effluent is not caused by NaCI, marinespecies may be stressed as much as freshwater spe-cies by high concentrations of other dissolved salts.Unless the “salinity” of an effluent is known to becaused by NaCI, marine species cannot be used toavoid the salinity effects.

7.3 Toxicity Test ProceduresThe purpose of the toxicity test in Phase I is the

same as that of any analytical method--to measure(detect) the presence of the toxicants. This use is quitedifferent than conventional toxicity testing where theobjective is to accurately and quantitatively measurethe sensitivity of the organism to known concentrationsof a chemical or effluent. For this latter purpose, remov-ing stress (e.g., low DO) or other contaminants, andlack of space is important because such stresses maychange the sensitivity of the organism to the contami-nant of concern. In Phase I, relative sensitivity is used;that is, we compare whether one condition is more orless toxic than another but both may be toxic. There-fore, concern of documenting and/or removing otherstresses is not very important. It is important to be surethat these other stresses are similar for each conditionbeing compared, each time the manipulation and sub-sequent toxicity tests are performed.

The reason for this discussion under test methodsis that effort must be made to make the tests used inPhase I as inexpensive as possible, because for someeffluents, large numbers of tests may be needed. Forexample, we have used more than 100 tests on someeffluents in Phase I. If the effort usually expended inmeasuring all the required water chemistries for a wholeeffluent test (EPA, 1985A) had been done for these

7-2

tests, the cost would have been prohibitive. The readermay wonder whether data collected from such testscan be trusted. Confidence in the data hinges on care-ful assurance that the stresses are similar among com-parisons. For example, it does not matter if the testorganisms are acclimated to a pH change. It doesmatter that stress from lack of acclimation to pH changeoccurs in each treatment compared.

Sometimes, in order to achieve desired chemicalconditions, the stress from pH change cannot be madeuniform. In these situations, only gross differences inresponse may be dependable. In some cases, errone-ous conclusions will be reached. While these may causewasted effort, the error should be found in Phase III.That is why, in Phase III, careful quality control must beexercised and cost saving shortcuts are not acceptablebecause one of the purposes of Phase III is to catcherrors or artifacts that may occur in Phases I and II.

One need not use the standard acute methods(EPA, 1985A; 1991A) in Phase I for these reasons. Thefollowing mechanics of performing an acute test withcladocerans and newly hatched fathead minnows havebeen found by experience to be very cost effective andare offered as an aid to those doing Phase I testing.Specific volumes and sizes are used in this example forsimplicity, but of course, these are varied depending oneach test purpose.

For example, arrange a set of 12 plasticcups into six pairs. Fill 10 cups with 10 mLof dilution water using a disposable pipette.Add 10 mL of effluent to the two emptycups to make the high concentration, (e.g.,100%). Add 10 mL of effluent to the nextpair of test cups (duplicates labeled A andB) already containing 10 mL each of thedilution water (.Figure 7-1). The resultingconcentration is 50%. From each cup ofthe 50% solution, transfer 10 mL to thethird pair of test cups to produce the 25%concentration. Continue this process untilsufficient exposure concentrations havebeen prepared. One pair of cups in theseries contains only dilution water andserves as the control. Mixing the solutionsprior to the transfer of each aliquot is veryimportant. This can be accomplished bydrawing the solution into the pipette anddischarging it back into the cup severaltimes prior to transfer. Additional mixing oftest solutions should be done forexperiments in which reagents such assodium thiosulfate (Na,S,O,) and EDTA(Phase I), or effluent methanol eluateconcentrates (Phase I and II) are added toeffluent or dilution water.

‘The need for duplicates will depend on the accu-racy and precision required of the test results. Testsrequiring a measure of accuracy in the form of confi-dence intervals (Cls) should be run in duplicate. Testsdesigned to provide only’ an indication of positive or

negative toxicity need not be run in duplicate. Beyondthe initial and baseline effluent toxicity tests (Sections8.1 and 8.2) which are designed to define effluenttoxicity upon arrival in the laboratory and periodicallyduring the TIE with each effluent sample testing, re-spectively, Phase I toxicity tests usually do not requireduplicates.

The test organisms of uniform age should be placedat random in each test cup to insure valid results.Because the volume of test solution may be small, caremust be taken to minimize the volume added duringtest organism transfer. If the volume of water trans-ferred with the organism is reduced to a drop (50 pL),only five organisms are added to the test chamber anda 10 mL test volume is used, the resulting change intest solution volume will be 2.5%. Minimizing the changein volume is more critical as test solution volume isreduced. This is particularly important in the Phase IIexperiments, when limited volumes of effluent fractionconcentrates are available. Care should also be takento avoid chemical contamination between concentra-tions when test animals are being added.

We have stressed a relaxation of the usual waterchemistry requirements in these Phase I tests becausethey are not as necessary here as they are in Ph&e III.However, sometimes, in order to maintain the desiredconditions in the test (such as maintaining a specificpH) frequent specific repetitive measurements of thoseitems will be necessary. The distinction drawn here isto avoid measurements you don’t need (e.g., samplehardness) and concentrate on those that are important(e.g., pH). Effluents are often well buffered and pHsometimes will change quickly if equilibrium is not al-ready established. POTW eifluents are not in air equi-librium when discharged and as soon as they areexposed to air, the pH will rise. A typical POTW effluentpH is 7.2-7.4 when discharged but it will equilibrateafter contact with air and may stabilize at 8.2-8.5. If pHis important to test interpretation, pH must be moni-tored throughout the test. It will also be important todecide what the initial pH (pH i) of the effluent is sincethe pH at the discharge and/or the initial pH may bedifferent from the pH of the effluent at air equilibrium.

7.4 Test EndpointsLittle effort should be expended in calculating LC50

values for Phase I toxicity tests. There is no need toapply sophisticated and complex programs to the testresults. Several methods for estimating the LCSO fromthe acute toxicity data are described in EPA (1985A),however a method which is most easily and quicklyapplied to the data should be used. In many cases, thegraphical method entailing interpolation may prove tobe the most convenient. Differences resulting from thechoice of data analysis method should not impair theoutcome of Phase I studies. Phase III tests may requiremore sophisticated analyses.

Toxic units (TU) have a special utility in some partsof a TIE. The TU of whole effluent is 100% divided bythe LC50 of the effluent. For specific chemicals the TU

7-3

Figure 7-l. Schematic for preparing effluent test concentrations using simple dilution techniques. Two replicates are used for initial and baseline wholeeffluent toxicity test.

Effluent

aAdd 10mLtocupsAandBfor the high concentration.

0l Add 10 mL to next A and B cups

for the second high concentration.

Add 10 mL to eachreplicate except inthe high concentration

56

Mix

DilutionWater

6B

Control

HighCont.

* Serial Dilutions m

Waste

6A

Control

is equal to the concentration of the compound presentin the effluent divided by the LC50 of the compound(EPA, 1991 B). For example, if the LC50 of an effluentis 25%, the effluent contains 4 TU (100/25). If the 48-hLC50 of compound A is 3 mg/L, a solution of 1 mg/L ofthis compound contains 0.33 TU. By normalizing theconcentration term (such as the LC50) to a unit oftoxicity, the TU allows the toxicity of effluents and/orchemicals to be “summed,” provided that the test lengthand species used are the same in every test. Thiscannot be done using LC5Os because chemicals andeffluents each have different toxicity, and different con-centrations each equal one LC50. Phase III containsmore discussion about adding TUs; however one mustbe cautious in summing them. Unless it is known thatthe toxicants are strictly additive, simple summation ofTUs will be incorrect.

7.5 FeedingMost species used in acute tests are not fed during

the test. However, the acute effluent manual (EPA,1991 1991 B) has modified the effluent tests to allow clado-cerans to be fed before test initiation. We routinely addfood to all test waters (this includes the 100% effluent)for all Ceriodaphnia and Daphnia tests but only at theinitiation of each test. This practice is standard in PhasesI, II, and III. However, the decision to feed will be

species specific and dependent on the characteristicsof the effluent. Consistency throughout each phase ofthe TIE is most important. All tolerance data forCeriodaphnia given in Section 8 are based on tests inwhich animals were fed the yeast-cerophyll-trout food(YCT) mixture (EPA, 1989E: EPA, 1991C). The amountof YCT added was 66 uL of YCT per 10 mL and 5animals.

7.6 Multiple SpeciesA useful technique is to test two species together in

the same test chamber (e.g., 1 oz. plastic cup). This isvery beneficial in the initial toxicity test in order to selectthe most sensitive species for the Phase I tests or insituations where two species appear to be respondingto toxicity of the effluent differently. This type of testalso can be useful when conditions in tests with differ-ent organisms vary independently. For example, testingC. dubia and fathead minnows ~48 h old) togetherunder the same pH conditions is very, helpful in evaluat-ing the role of ammonia in an effluent’s toxicity. Bytesting the species together, the experimental condi-tions may change but both species experience identicalfluctuations. We have tested the following sets of spe-cies together: C. dubia and fathead minnows, C. dubiaand D. magna, and C. dubia and D. pulex. B

7-5

Section 8Phase I Toxicity Characterization Tests

The first phase of a TIE involves characterization ofthe toxic effluent. The characterization information gath-ered in Phase I forms the basis and direction for PhaseII identification of the specific toxicants or may beuseful for treatability evaluations. In Phase I, simplemanipulations for toxicity removal or alteration are per-formed on the whole effluent. Acute toxicity tests utiliz-ing aquatic organisms are used to determine whetherthe toxic chemicals have certain physical or chemicalcharacteristics. Two objectives are accomplished dur-ing the toxicity characterization phase: a) the physicaland chemical characteristics of the toxicant arebroadly defined and b) some information is gathered toindicate whether the toxicants are similar in effluentsamples taken over time. Several patterns of Phase Iresults are indicative of certain toxicants (See Section9.4) but otherwise Phase I only provides evidence ofcharacteristics of groups of chemicals that may be thetoxicants. This information can subsequently be used inthe second phase of the study, either in the develop-ment of bench-scale wastewater treatment processes(EPA, 1989A; 19898) or in choosing separation andanalytical procedures for toxicant identification as de-scribed in Phase II.

The tests described in this section are designedprimarily for acutely toxic effluents. Methods for chronictoxicity are being developed (EPA, 1991 A). The meth-ods in this section are based on the use of small testorganisms such as daphnids (Ceriodaphnia, Daphnia)and newly hatched fish (fathead minnows). If largerspecies are used, modifications to these methods willhave to be made.

Analysis of samples should begin as soon as prac-tical following collection. Until experience is gained withthe effluent, there is no way to predict how long samplescan be stored before substantial changes in toxicityoccur. In transit and in the laboratory, the bulk effluentshould be held below 4°C and kept headspace free.Minimizing the headspace for samples shipped in glassis not practical. Once in the laboratory, testing on indi-vidual samples of each effluent may continue indefi-nitely, provided that whole effluent toxicity stabilizes.The degree of toxicity can remain similar, while thecause of toxicity may change with age. Especially inthe early stages of the TIE, fresh samples should beused regardless of toxicant stability. The degree towhich any single sample is analyzed should be weighedagainst the cost of the analyses and the probability thatthe sample is an adequate representation of typical

effluent. Obviously, when several samples show that asingle class of compounds is responsible for effluenttoxicity, Phase II procedures should be initiated.

Each of the characterization tests described in Sec-tion 8 is designed to change the toxicity of groups ofconstituents (Figure 8-l). Toxicity before and after thecharacterization treatment will indicate for which groupsthe toxicity was changed. All but one (initial toxicitytest) of the characterization tests is performed at thesame time in order to minimize confounding effectsresulting from degradation of sample toxicity over time.While it is not critical that each characterization ma-nipulation be performed at exactly the same tipe, thetoxicity tests should be initiated at approximately thesame time. If more than one species is used, thePhase I results must be interpreted separately for eachbecause at this stage one cannot tell whether the sametoxicant is involved for all species.

Following receipt of the effluent sample, varioussteps to initiate Phase I are done (Table 8-l). Day 1 iswhen the sample arrives in the laboratory. On day 1,initial routine chemical measurements are taken for theeffluent sample and an initial toxicity test is started onan aliquot of the sample. This LC50 is used to set thedesired exposure concentrations for subsequent PhaseI toxicity tests and is referred to as the “initial” toxicitytest to distinguish it from the “baseline” toxicity testdescribed below. Other aliquots of the sample are ad-justed to pH 3 and 11, filtered, aerated and/orchromatographed using a C,, SPE column. Followingthese manipulations, each effluent aliquot is readjustedto the initial pH (pH i ) of the effluent. By pH i, wegenerally refer to the pH of the effluent at arrival in thelaboratory, which may or may not be the pH of theeffluent at air equilibrium. These aliquots and the re-mainder of the effluent are then covered to minimizeevaporation and held at 4OC overnight. However, uponwarming the solutions, supersaturation from dissolvedgases might occur. If the test organism to be used issensitive to supersaturation, then the supersaturationmust be removed. Generally, Ceriodaphnia are notvery “sensitive” to such situations, unlike newly hatchedfathead minnows.

Delaying the majority of the toxicity testing until thenext day (day 2) allows the test exposures to be set atconcentrations bracketing the 24-h LC50 of the day 1initial toxicity test. This procedure also allows pH ad-justed effluent aliquots more time to stabilize, and addi-tional pH adjustments can be made as necessary.

8-l8-l

Figure 8-1. Overview of Phase I effluent characterization tests. (Note: pH i stands for initial PH.)

t

-r

- Toxic Effluent SampleIt I

1

EDTAChelation

Test (Day 2) 1

Cl8 Solid PhaseExtraction Tests I

I OxidantReduction

Test (Day 2)

Graduated pHTests (Day 2)

Table 8-1. I effluent manipulations

Description Section

DAY 1 SAMPLE ARRIVAL:

Chemical analysesl PH*conductivity*total residual chlorine (TRC)*hardness*temperature*total ammonia*dissolved oxygen (DO)*alkalinity

6.0

Initial toxicity test

Sample Manipulations’:l pH adjustment (pH 3, pH i, pH 11)l pH adjustment/filtrationl pH adjustment/aerationl pH adjustment/C,, solid phase extraction

DAY 2 TOXICITY TESTING:

8.1

8.3-8.98.38.48.58.6

Warm effluent samples from day 1 and set-uptoxicity tests’

*baseline toxicityl pH adjustment samples*filtration samples*aeration samplesl C,,, solid phase extraction samples*sodium thiosulfate addition samplesl EDTA addition samples*graduated pH samples

Read 24-h mortality on initial toxicity test

DAYS 3 AND 4 MONlTORlNG TESTS:

8.28.38.4 .8.58.68.78.88.9

8.1

Read 48 h mortality initial toxicity test 8.1

Read 24 h and 48 h mortality on tests from day 2 8.2-8.9

11 These manipulations and toxicity tests can be performed on day2 after the presence of toxicity has been confirmed; see text fordetails.

However, the manipulations can be performed and thetest initiated the next day (day 2) rather than on samplearrival. This is useful when the toxicity of the effluent isunknown, and prevents conducting a Phase I on non-toxic samples. It is important that sufficient time isallowed so that the pH adjusted samples can stabilizeat the pH i. The following sections assume the manipu-lations were made on day 1.

On the second day the aliquots of whole andmanipulated effluent prepared on day 1 are diluted to4x-, 2x-, 1 x-, and 05x the 24-h LC50 of the effluent andsubsequently tested for toxicity. This dilution series isused so that for highly toxic effluents, smaller changesin toxicity can be detected than would be the case if100% effluent was used. (See Section 9 regardingmultiple toxicants and this dilution series.) A wholeeffluent toxicity test is begun using unaltered effluent,now 24 h old. The result of this test (and subsequentwhole effluent tests) is referred to as the “baseline”effluent LC50. Other toxicity tests involving the addition

of chelating or reducing agents and less severe pHadjustments are also conducted. For the EDTA additiontest, the time needed for the EDTA to complex anymetals present may be a function of the matrix of theeffluent. Therefore, the addition of EDTA should bemade first on day 2 and the sample held until all othermanipulations are complete before introducing the testorganisms (see Section 8.8 for details on EDTA test).

For one complete “Phase I” of a TIE as describedin this section, there are nine categories of toxicity teststhat are conducted. These are as follows: initial toxicitytest, baseline test, pH adjustment test, pH adjustment/filtration test, pH adjustment/aeration tests, pH adjust-ment/C,, SPE test, EDTA addition test, sodium thiosul-fate addition test, and graduated pH test. Toxicity testresults are read on subsequent testing days and de-pending on the outcome of the Phase I test series,additional toxicity tests designed to further define orconfirm the nature of the toxicants are conducted.

”For an experienced analyst the amount of time

required to conduct the sample manipulation tasksscheduled for day 1 is about half of one day. If at 24 h,less than 50% mortality of test organisms exposed tothe 100% day 1 effluent has occurred, the sample canbe discarded and a new sample collected with rela-tively little loss of resources or time. For this reason,waiting to perform the manipulations on day 2 is useful.Alternatively, the test can be continued to 48, 72 or96 h at which time the effluent may produce an LC50.In such cases, the baseline toxicity tests prepared onthe second day (day 2) following sample arrival are setup at exposure levels of 1 OO%, 50%, 25%, 12.5%,6.25% effluent.

For a highly toxic effluent sample with rapidly de-gradable toxicants, it may be prudent to override theuse of 4 x -24-h LC50 treatment level and opt for con-ducting the Phase I using 100% effluent. These rapidlydegradable compounds will be discovered only throughperiodic testing as the sample ages.

Several Phase I characterization tests are relativelybroad in scope, intended to include more than oneclass of toxicant. Therefore, if a significant change ineffluent toxicity is seen following these characterizationprocedures, additional tests are needed to further delin-eate the nature of the toxicity. The amount of testingbeyond the initial characterization of the sample willdepend on the stability of effluent toxicity, the nature ofthe toxicity, and previous Phase I results for the effluent(i.e., observed trends in the nature of the toxicity). A“significant reduction” in toxicity between aliquots of theday 2 whole effluent (baseline LC50) and treated efflu-ent must be decided based upon the laboratory’s testprecision. Usually a change in the LC50 equal to oneconcentration interval can be considered significant butwhen precision is good smaller differences can beused. This suggestion is arbitrary and should not re-place good judgement and experience. None of thesetests by themselves are conclusive, so the danger oftype I or type II errors is not great. Experience has

8-3

shown that for many effluents, at least one Phase Icharacterization test will be successful in substantiallyaltering effluent toxicity. If not all toxicity is removed,other groups of toxicants (not addressed by Phase Iprocedures) may be present in the effluent or a singletoxicant may be present in the effluent at such highconcentrations that only partial toxicity removal isachieved. Additional testing to resolve these findingsinvolves applying the successful Phase I test at ahigher level (i.e., increased degradation time, increasedaeration, larger C,, SPE column volume, increasedreagent concentratrons).

Another outcome of the Phase I characterizationtest series may be that several tests succeed in par-tially removing effluent toxicity. In this situation, onemay be dealing with several toxicants, each with differ-ent physical/chemical characteristics, or a single toxi-cant of such a nature as to be removed by more thanone Phase I test. These results may be resolved bytreating a single aliquot of the sample with all of thecharacterization tests that significantly reduced thebaseline toxicity of the effluent. If effluent toxicity re-moval is enhanced as compared to the reduction pro-vided by individual characterization tests, the samplemay contain more than one type of toxicant. If the finaltoxicity removal at the end of the series of characteriza-tion tests is approximately the same as that providedby the most efficient single Phase I test, then it is likelythat all of the test methods involved are successful inreducing the same toxicant to varying extents. Thisoutcome is also suggested when one or more Phase Itests completely remove toxicity while some number ofother tests partially reduce toxicity. Phase I tests over-lap somewhat in their abilities to remove groups oftoxicants. For example, increasing pH may cause ametal to precipitate and toxicity removed and EDTAmay also remove toxicity. In any case, results of thisnature are useful in selecting Phase II options. Use ofmultiple manipulations (Section 9.2) builds upon theseprinciples. When several treatments are applied to thesame sample, tests must be designed to ensure thattoxicity does not result from the additives used (e.g.,acid, base, EDTA) rather than from the effluent’st o x i c a n t (

The assumption must not be made that toxicantsare either additive or synergistic. Our experience showsthat independent action (one or more of multiple toxi-cants acting independently of the rest, as though theothers were not present) is not uncommon in effluents.Experience also shows that one should not use se-lected tests to confirm a suspicion that a certain toxi-cant is the cause of toxicity. Time and again, this leadsto wasted effort. There are so many possible causes oftoxicity that such guesses are rarely helpful and moreoften channel one’s thinking and delay the final solu-tion. On the other hand, if one wants only to knowwhether a certain chemical is the toxicant, these testscan be selected to accomplish that goal. Frequentlyone needs to know whether the toxicity is due to am-monia or whether there are toxicants present otherthan salt. These questions are quite different from the

former case where one is playing the “I’ll bet you thetoxicant is...” game.

No Phase I characterization test should be droppedfrom use on the basis that the toxicants it is designedto address are not likely to be present in the effluent. Inexcluding any Phase I test, the analyst may be limitingthe information that can be gained on effluent toxicants.The investigator should approach effluent characteriza-tion without a preconceived notion as to the cause oftoxicity.

There are two types of checks that can be used todetect artifact toxicity. A “toxicity blank” consists ofperforming the same (Phase I) test on dilution waterand measuring to determine whether any toxicity isadded by the test procedure. However, a toxicity blankdoes poorly in identifying artifact toxicity if toxicity isaffected by the effluents’ matrix (cf., Section 7.1). Forexample, the toxicity of the Phase I reagent, EDTA,may be completely different in dilution water and ineffluent. If so, a toxicity blank is inappropriate for the 4chelation test. A “toxicity control”, for many Phase Isteps involves a comparison of the toxicity of the ma-nipulated test solution and the baseline effluent twicity.In this case, the comparison must demonstrate that themanipulated effluent test solution has not become moretoxic than the unaltered effluent (baseline test). If it has,the test procedure has produced artifactual toxicity.The “toxicity control” for the pH adjustment/C,, test isthe filtered effluent sample at the respective pH. Forsome treatments, valid toxicity blanks or toxicity con-trols cannot be made. The use of toxicity blanks andtoxicity controls still requires the use of “regular” con-trols, which are always included to determine the per-formance of the test organism and dilution water. Dilu-tion water blanks for the EDTA addition test, sodiumthiosulfate addition test, and the graduated pH test arenot relevant (see Section 7.1 for more information).

No procedure should be assumed to be free ofartifactual toxicity. Many of the Phase I toxicity testsinvolve relatively severe or unorthodox effluent manipu-lations. Toxicity blanks and toxicity controls must beused consistently and conscientiously to detect theintroduction of toxic artifacts or other changes to theeffluent that increase sample toxicity.

For the following sections, the guidance for thevolumes required, apparatus, and test organisms isbased on test conditions using Ceriodaphnia, Daphniaand/or larval fathead minnows exposed in 10 mL testvolumes.

8. I lnitial Effluent Toxicity TestPrinciples/General Discussion:

The major purpose of the “initial” effluent test is toprovide an estimate of the 24-h LC50 for purposes ofsetting exposure concentrations in Phase I tests.

Volume Required:Initial toxicity test is performed in duplicate using 40

mL of effluent. I

0-4

Apparatus:Disposable 1 1 ozoz plastic cups or 30 mL glass bea-

k e r s , k e r s ,

species.

Procedure:7: mL i n d u p l i -

cate of 1 1 OO%, 50%, 25%, 6.25% effluent, and acontrol will suffice for most effluents. Obviously moretoxic effluents will require a lower range. If nothing isknown about the toxicity, more concentrations shouldbe included. A sample data sheet for the initial test isshown in Figure 8-2.

8.2 Baseline Effluent Toxicity Test

In order to determine the effects that the variousPhase I characterization tests have on effluent toxicity,the toxicity of the effluent sample, prior to any treat-ment in the laboratory, must be determined. The por-tion of the effluent sample, tested for toxicity the dayafter it arrives in the laboratory (day will be referred

LC50

the characterization tests. Such a comparison will dem-onstrate whether the removal or alteration of variousgroups of toxicants changes the effluent toxicity. Bycomparing these results, an indication of the physical/chemical nature of the toxicants can be obtained. If the

LC50 is substantially differentfrom the toxicity of the effluent when it arrived in thelaboratory (initial toxicity), one must decide whether theschedule suggested in these methods should be re-vised to reduce a delay in testing.

When Phase I testing is extended to additionaldays, baseline tests must be done each time on suc-ceeding days, and used for comparison to these addi-tional manipulation tests.

Volume Required:The baseline toxicity test is performed in duplicate.

The total volume necessary will depend, on the 24-hLC50mL s h o u l dbe adequate..

Apparatus:Disposable 1 1 ozoz plastic cups or 30 glass bea-

scope (optional).

Test Organisms:Test organisms, 60 or more, of the same age and

species.species.. .

Procedure:Day 2: Two concentration series will be used in

duplicate for the static acute toxicity test. In preparingthe test solutions for the day 2 baseline test, anyobvious physical changes (e.g., formation of precipi-tates, odors), which occurred during storage, should benoted.

The first test series will have exposure levels basedon the 24-h LC50 of the initial (day 1) toxicity and willinclude day 2 effluent concentrations at 4x-, 2x-, lx, lx,and 0.5x- the 24-h LC50. In this case, the method formaking dilutions described earlier may need to bechanged slightly. Most of the toxicity tests with thecharacterization solutions will also be performed usingthese same exposure concentrations. If the 24-h LC50of the initial effluent is greater than 25%, the seriesobviously begins at lOO%, and includes four exposureconcentrations. Of course, if the 24-h LC50 of the day 1initial effluent is greater than or equal to 25%, thesecond series will be unnecessary because this testfulfills the requirements for comparison to the initialeffluent test and characterization solution toxicity testresults.

The second test series will provide exposur s atseffluent dilutions of 1 OO%, 50%, 25%, 12.5% and 6. 5%

(and lower dilutions as appropriate if the effluent ismore toxic). This series will enable a comparison of theresults of the baseline (day 2) test to the initial effluentLC50 (cf., Section 8.1).

A sample data sheet is shown in Figure 8-3. Inorder to compare the baseline toxicity and the toxicityof the effluent aliquots subjected to characterizationtests, all of the day 2 toxicity tests should have the testorganisms added to test solutions at approximately thesame time.

The baseline toxicity test (toxicity control) must berepeated each time additional characterization tests areperformed on the sample after the initial Phase I bat-tery of tests. The baseline test will serve as the basisfor determining the effects produced by the additionalcharacterization tests, and will also provide informationon the degradation of sample toxicity. For effluentswhose initial toxicity is low (i.e., LC50 -6O-70%) andwhere the baseline toxicity is greatly changed com-pared to the initial toxicity of the sample, it may beadvisable to discard the remaining sample and collect afresh one.

Interferences/Controls and Blanks:The control treatment of animals in unaltered dilu-

tion water in this test is used for comparison to severalsubsequent tests and provides an important referencefor diluent water and organism acceptability. Mortalityin these controls will negate other work.

Results/Subsequent Tests:Baseline LCSO’s should be generated for as long

as the effluent sample is being used and a baselinetest (toxicity control) should be started every time the

8-58-5

Figure 8-2. Example data sheet for initial effluent toxicity test.

Test Type: Initial EmumatTest Initiation (Date & Time):

Investigator:Sample Log No., Name:Date of Collection:

Species/Age:No. Animals/No. Reps:Source of Animals:Dilution Water/Control:Test Volume:Other Info:

Cont. 0 h(Oh Effluent) 11 pH

12.5 II

6.25 II

-r24 h

A 8 pH 00

Survival Readings:

zqGz+K

Comments:

8-6-

effluent sample IS put through any characterization steps.(Note: similar procedures should be followed in PhasesII and III.)

8.3 pH Adjustment TestPrinciples/General Discussion:

The pH has a substantial effect on the toxicity ofmany compounds found in effluents. Therefore pH ad-justment is used throughout Phase I to provide moreinformation on the nature of the toxicants. Changes inpH can affect the solubility, polarity, volatility, stabilityand speciation of a compound, thereby affecting itsbioavailability as weil as its toxicity. Before describingthe pH adjustment test, some discussion on the effectof pH on various groups of compounds is warranted.

Two major groups of compounds significantly im-pacted by solution pH are acids and bases. To under-stand how organic and inorganic compounds of thistype are affected by pH changes, one must have abasic understanding of the thermodynamic equilibriumacidity constant, K,, for the proton transfer reaction:

HA + H,O = H,O+ + A*

Ka = [A-][H+]WI

H+: H,O+HA: protonated acidKp: thermodynamic equilibrium constant for

the acid

For example:

HCN + H,O = H,O+ + CN-

Ka= [CN-][H+] = 6.0 x lo-loWNI

The stronger the acid (i.e., the more it tends todissociate into its ionic state), the greater the value ofK,, and the smaller the -log,,K, or pK. In effect, theabo.ve reaction is shifted to the right for acidic com-pounds. For acids in water, when the pH of the solutionequals the pK, of the compound, equivalent amounts ofthe compound will exist in the ionized (A) and un-ionized (HA) forms. At a pH one unit lower than the pK,of the acid, approximately 90% of the compound will bein the un-ionized form with the remainder in the ionizedform. A solution pH two units below the acid’s pKd willresult in 99% in the un-ionized form and 1% in theionized form. Likewise, at one pH unit above the pK,,90% of the acid will be present in the dissociated(ionized) form and 10% present in the un-ionized form;at two pH units above the pKa, 99% of the acid is in thedissociated form while, 1% is present in the un-ionizedform. For example, at pH 4.2, the pK, of benzoic acid,50% of the compound is present as C,H,COOH and50% is present at C H COO-,H+. At pH 3.2, this ratioshifts to roughly 9O*L C H COOH:lO% C H COO-, H+while at 5.2 the ratio ne”ar’s 10% C,H,CdOk to 90%C,H,COO-,H’.

This relationship generally holds for diprotic andtriprotic acids (i.e., acids with two and three H atoms,respectively, that can dissociate from the molecule).This trend is not followed by multiprotic acids with pK,‘sless than three units apart (e.g., H,BO, with pIQ13.8and pK, *= 12.74). The amount of each dissociated spe-cies in such cases will not always follow the 90/i 0, 99/lrule stated above. For example, H,BO,, H,BO,-, andHBOa2- will be present at pH 13.5.

Basic compounds function in a similar fashion.

B + H,O = BH+ + OH-

K, = [BH;EH-1

B: unprotonated baseK,: thermodynamic equilibrium constant for the base

For example:

C,H,NH, + H,O = C,H,NH3* + OH-

K,= [C,H,NH,+][OH-] = 4.2 x lo-‘O

[C,yJq5

In the above reaction, BH+ can be considered the“conjugate acid of the base”, that is, the protonatedform of the base. Thus, the same reaction can beexpressed as follows:

BH++ H,O = H,O+ + B

K, = W’IPIW’l

Note: pK,+ pK,= 14

For example:

C,H,NH,+ + Hconjugate acid

0 = H,O+ + C,H,NH,

Ka= [H+][C6HSNH2] = 2.34 x 1O-5

[C,H,NH,‘l

This convention can be used to simplify dealingwith equilibrium constants for acids and bases.

As with acids, when the solution pH is equal to thepK, of the conjugate acid of a base, equal amounts ofthe base will exist in the ionized and un-ionized forms.For example, ammonia in an aqueous solution at pH9.25 (the pK, of ammonia) will be found as 50% NH,+and 50% NH,. At one pH unit above the pK, (i.e.,10.25) roughly 90% of the ammonia will be in the un-ionized form (NH,) and the remainder will be in theNH,+ form. At pH 8.25, one unit below the pK, ofammonia, approximately 90% of the ammonia will be inthe NH,+ form, and approximately 10% will be in theNH, form.

8-8

Figure 8-4.

4 6 8 IO I2 14PH

-

pE -pH diagram for the CO,, H,O, and Mn-CO, systems (25°C). Solid phases considered: Mn(OH) (8)(pyrochroite), MnCO,(s) (rhodochrosite), Mn,O,(s) (hausmannite), YMnOOH (manganite), YMna,(nsutite). (Reprinted with permission from Stumm 81 Morgan, 1981.)

The above can be summarized by the following:.PY

Orqanic InoroanicPH > PK

acid RCOO-, RCO- Abase RNH, B

pH < PKacid RCOOH, RCOH HAbase RNH,+ BH+

R = aliphatic or aromatic group

The effect of pH on the ratio of the ionized and un-ionized forms of acids and bases has a number ofimpacts on Phase I results. First, compounds may bemore toxic in the unionized form as compared to theionized form. For example, un-ionized ammonia (NH,)is generally recognized as the toxic form of ammoniawhile total ammonia (NH,+) is of far less concern (EPA,19858). A second implication of this effect relates totoxicant solubility. Unionized forms of acids and basescan be considered less polar than their ionized forms,which interact to a greater extent with water molecules.Consequently, un-ionized forms of acids and bases canbe more easily stripped from water using aeration (Sec-tion 8.5) or extraction with non-polar solvents or solid

phase column techniques (Section 8.6). Likewise,changes in compound solubility with pH change maymediate removal through filtration (Section 8.4).

Another implication of the pH effect involves metalion complexes. An example of how pH can alter theform of a metal in a natural water system is shown inFigure 8-4. Given a p& (the equilibrium electron activ-ity-in a simple sense, whether the system is aerobicor anaerobic), one can see how various forms of man-ganese are created and eliminated as’ pH shifts.

Each of the different forms of a metal will bemanifested differently in aquatic organism effects. Someforms of the metal will be relatively insoluble; theseforms may not affect toxicity. Likewise, as with acidsand bases, the toxicity of the soluble forms of the metalwill be a function of the actual species present (e.g.,the LC50 of Mn2* as compared to the LC50 of MnO:).The actual species formed will depend, in addition topH and p& on the other chemical constituents presentin the water. The hydrolysis rate of organics is greatlyaffected by pH, and pH changes may also alter organictoxicity.

Regardless of the speciation effect on toxicity,changes in solution pH may affect the toxicity of anygiven compound. The pH of the test solution may affect

8-9

membrane permeability at the cell membrane as wellas the chemistry of the toxicant. One might expect thatchanging the pH, only to return it to its original pH in ashort time, would not alter toxicity. Experience showsthat this is not the case and that this adjustment some-times results in reduction, loss or increase in toxicity. Ifthe kinetics of the pH driven reaction (on return to theoriginal effluent pH) are slow or irreversible, pH adjust-ment alone may be effective in evidencing toxicantsaffected by pH change. Some organics may also de-grade due to pH change.

Another purpose of the pH adjustment test is toprovide blanks (with both dilution water and effluent) forsubsequent Phase I pH adjustment tests performed incombination with other operations. This test will dem-onstrate whether toxic concentrations of ions have beenreached as a result of the addition of acid and base orwhether the reagent solutions are contaminated.

Comparable results for toxicity blanks are not ob-tained when the same volumes and same strengths ofacids and bases are added to the effluent. Effluentsalready contain substantial concentrations of major an-ions and cations that are not found in dilution water.Further, the volumes and strengths of the acid andbase necessary, for example, to lower an effluent witha pH i of 7.6 to pH 3 and raise it back to pH 7.6, arenot likely to result in the same final pH when added todilution water. However, it is necessary to conduct pHadjustment blanks to determine the cleanliness of theacid and base solutions used for the pH manipulations.

Volume Required:

To make pH adjustments, 680 mL of whole effluentis needed to have enough effluent for the four exposureconcentrations at each of the three pH’s. A 300 mLaliquot of the day 1 effluent sample is raised to pH 11,and the second 300 mL sample is lowered to pH 3. Analiquot of the pH i effluent (used for the baseline test:80 mL) is set aside for the duration of the manipulation(Figure 8-5). Approximately 30 mL will be needed forthe pH adjustment test but the actual amount dependson the 24-h LC50 of the initial effluent test. The remain-ing 270 mL of each of these solutions is reserved forthe “pH adjustment/filtration”, “pH adjustment/aeration”and “pH adjustment/C,B SPE” Phase I tests.

These pH adjustments must also be done usingdilution water for toxicity blanks for each test. To makethese adjustments, approximately 295 mL of dilutionwater will provide enough (and an excess) to test theblanks with one exposure level and one replicate. Onealiquot of 105 mL is adjusted to pH 3 and another 105mL is adjusted to pH 11. Only 10 mL is needed for thepH adjustment blanks but excess (~10 ml) is included.The, pH i dilution water blank is the control of thebaseline test. The remaining 85 mL of pH adjusted and85 mL of pH i dilution water are used for the “pHadjustment/filtration,” “pH adjustment/aeration” and “pHadjustment/C,, SPE” toxicjty blanks.

Apparatus:six glass stoppered bottles for acid and base solu-

tions, pH meter and probe, 2-500 mL beakers, 2-500 mLgraduated cylinders, 30 mL beakers or 1 oz plasticcups, stir plate, and stir bars (perfluorocarbon), auto-matic pipette, disposable pipette tips, eye dropper orwide bore pipette, light box and/or microscope (op-tional).

Reagents:1 .O, 0.1 and 0.01 N NaOH, 1 .O, 0.1 and 0.01 N

HCL (ACS grade in high purity water) and buffers forpH meter calibration.

Test Organisms:Test organisms, 40 or more, of the same age and

species.

Procedure:Day 7: The general procedure for the pH adjust-

ment test is shown in Figure 8-5.

Blank Preparation: The first step is to prepare thedilution blanks. These blanks are used as the controlsfor the other dilution water pH adjustment testsofaeration, filtration, and C,, SPE separation as well as todetermine whether the acid or base solutions are con-taminated. The pH i blank is the control while the pH 3or pH 11 adjustment blanks are treated in the mannerdescribed below under sample preparation.

Sample Preparation: Stirring constantly, 1 .O N NaOHis added dropwise to a 300 mL aliquot day 1 effluentuntil the solution pH nears 11. (Note: overshootingresults in the addition of more salts and a volumechange and should be avoided.) In order to minimizeany over-adjustment of the pH, 0.1 N NaOH is addeddropwise in the latter stages to bring the effluent aliquotto pH 11. The solution should be allowed to equilibrateafter each incremental addition of base. The amount oftime necessary for pH equilibration will depend on thebuffering capacity of the effluent. Caution should betaken to prevent any solution pH of greater than 11. IfpH 11 is exceeded, 0.1 N HCI must be used to lowerthe pH to 11. The goal of the pH adjustment step is toreach pH 11, while minimizing both the change inaliquot volume and the increase in ionic strength. Vol-umes and strengths of base (and any acid added)should be recorded. A 30 ml volume of effluent anddilution water is held for the same length of time ittakes to complete other Phase I manipulations with pH11 effluent.

Once other manipulation work has been completedwith the total volume of the pH 11 effluent, the 30 mLvolume at pH 11 is returned to the initial pH (pH i ) ofthe day 1 effluent. (The other aliquots of pH 11 effluentare also returned to pH i at this time.) This is accom-plished by the slow, dropwise addition of 0.1 N HCI firstand later 0.01 N HCL as the pH of the stirred solutionnears pH i. If pH i is exceeded, the pH must beappropriately increased with 0.01 N NaOH. Again, the

8-l 0

.

n-I0I

a940I

I

I

c0. -5

. ..-35a

kz0 0

volumes and strengths of acid and any base addedshould be recorded.

The pH of the solution should be checked periodi-cally throughout the remainder of the work day andreadjusted as necessary. Changes in the total volumeof acids and bases added should be recorded.

This procedure is repeated, except the pH is low-ered to pH 3 using the second 300 mL aliquot ofeffluent, and 1 .O N and 0.1 N t-U. As with the pH 11effluent, 270 mL of the pH 3 effluent is used for the pHadjustment/aeration, pH adjustment/filtration, and pHadjustment/C,, SPE tests. The remaining 30 mL of thepH 3 effluent is held until all of the work on all of the pH3 effluent has been completed. At this point, the pH ofthe 30 ml_ volume of pH 3 effluent is readjusted to pH iby the dropwise addition of 0.1 N and 0.01 N NaOH.Maintenance of pH i must be assured through check-ing and readjusting the sample periodically throughoutthe work day. All volumes and strengths of acid andbase added should be recorded.

Day 2: At the beginning of the work day (the dayafter the arrival of the effluent in the laboratory), the pHof both of the 30 mL volumes is again checked toensure that pH i has been maintained. Any additionalpH adjustments are made and the volumes of the acidand/or base added are recorded. The acute toxicity ofeach pH-adjusted solution is tested at 4x-, 2x-, lx-,0.5x-LCSO (the 24-h initial LC50) as described in Sec-tion 7. Test solution pH should be measured in allexposure concentrations and recorded at least every24 h. A sample data sheet is shown in Figure 8-6.

Interferences/Controls and Blanks:Controls prepared for the baseline toxicity test also

act as a check on the organisms, dilution water, andtest chambers for this test.

The baseline effluent test acts as a control for thepH adjustment test, indicating whether the addition ofNaCl in the form of the acid and base has increasedeffluent toxicity. This pH adjustment test acts as thecontrol for other Phase I tests entailing pH adjustment.In addition to serving as a control for other pH adjust-ment tests, increased toxicity following pH adjustment,not as a result of NaCl concentration, indicates a pHeffect on toxicity or contamination of acids or bases(see below).

Results/Subsequent Tests:If either the pH 3 or pH 11 adjustment effluent tests

have significantly greater toxicity than the baseline ef-fluent test, two possible sources of toxicity exist: 1) theions (Na*, Cl.) added by the acid and base have re-sulted in a solution with an ionic strength intolerable tothe test organism; or 2) chemical reactions driven bythe pW change have not reversed upon readjusting topH i. Neither of these phenomena would be detectedthrough the use of a blank (dilution water). To helpresolve this situation, the NaCl LC50 values for com-mon test organisms are provided in Table 8-2. Theminimum concentration of NaCl in the test solution (i.e.,

not including the concentration of NaCl originally presentin the effluent) can be calculated from the volumes andstrengths of the acid and base added and final solutionvolume. The data in the table can be used only as arough guide, however, because the toxicity of sodiumchloride depends on the other anions and cations aswell as the total osmotic pressure exerted by the dis-solved substances. The toxicity of the added NaCl isbest determined by adding that amount of NaCl directlyto the effluent and to see if the addition increasedeffluent toxicity.

If either the pH 3 and/or pH 11 adjustment testsindicate in a significant decrease in effluent toxicity, itcould result from volume changes by acid and baseadditions or from chemical reactions driven by the pHchange that may not have been re-established or areirreversible. To determine if the addition of acid or basediluted the sample due to their volume addition, add avolume of dilution water equivalent to the total volumeof acid and base originally added td the effluent vol-ume. If a similar loss of toxicity in the diluted wastewa-ter occurs, the pH adjustment test should be repeatedusing more concentrated acid and base solutions.

_

A reduction or loss of toxicity may also bmheresult of the degradation of toxicant at the altered pHvalues. In some cases, the toxicity could also be in-creased if the degradation product is more toxic thanthe original compound. Both organics and inorganicscan be so changed with a probable loss in toxicity.Inorganic and organic substances may precipitate dur-ing the process of pH adjustment. The precipitatedchemical may or may not be the toxicant. The precipi-tated chemical (which most often forms with the pH 11adjustment) removes from solution via the flocculationprocess, suspended solids, microbial growth, and col-loids, and via the adsorption process, metals and or-ganics. If the process of precipitation seems to removetoxicity, it is important to realize that the precipitatingchemical might not be the toxicant, but rather that thetoxicant may have been removed by the flocculationand/or adsorption processes. In some cases, the pre-cipitate may dissolve with the adjustment of the effluentback to pH i. The removal of toxicity when dissolutionof the precipitate occurs should be evaluated carefullysince the toxicant might be unavailable and/or notcompletely dissolved.

For most of the Phase I combination pH adjustmenttests (i.e., pH adjustment/filtration), the pH adjustmenttest will act as an equivalent or ‘Worst case” toxicitycontrol for changes in test solution ionic strength andvolume. In effect, most of the operations applied to thepH adjusted effluent in the following Sections (8.4-8.6)will either not affect pH or will drive it closer to the pH i.This may not be the case for the pH adjustment/aera-tion test, however. Because pH 3 and pH 11 must bemaintained throughout the aeration process and be-cause the loss oftowards pH i, morethese test solutionsonly solutions.

volatiles may result in pH shiftsacid and/or base may be added toas compared to the pH adjustment

8-l 2

Figure 8-6. Example data sheet for pH adjustment test.

Test Type: pH A@dment

Date of Collection:

Test Initiation (Date & Time):

Investigator:Sample Log No., Name:

Other Info:

Species/Age:No. Animals/No. Reps:Source of Animals:Dilution Water/Control:Test Volume:

pH/Concentration 0 h 24 h(Oh pH adjusted

effluent) I-pH A pH DO

/I 3/4x-LC50 II II

/I 3/2x-LC50

II 3/1x-LC50 II II

II 3 IO. 5x-LC50 II /I

II 3jblank II II

II 11/2x-LC50 II II

Survivai Readings:

48 h 72 hII

96 h

Note: See baseline data sheet for controi data.

Volumes and Strength of Solutions Added:NaOHHCI

-300 ml pH 3300 ml_ pH 11

30 mL pH 330 ml_ pH 11

Comments:

8-13

Table 8-2. Acute toxicity of sodium chloride to selected aquatic organisms

Water Ufe-Species Type 24 h

LCSO (g/L) (95% Cl)stage 48 h 72 h 96-K

Ceriodaphnla dubia SRW’2 124 h

SRW’ 2 124 h

SRW’ 2 124 h

SRW12 524 h

VHRW’,2 524 h

Daphnia magna

Pimephales promelas

Lepomis macrochirus SRW’

NR

NR

524 h

524 h

11 wk

s24 h

l-9 g

0iii

(2.012.6)2.8(-_)

3.3(NV6.4

(NR)

7.9(7.0-9.0)

(4.$8,

(;;I

(-1

NR

(2.2z.6)2.7

(27)

\FJ(2.0-2.6)

2.8(-_)

3.1(NR)

(:I$

7.9(7.0-9.0)

( 4 $ 6 )

(4ii.6)

NR

6.9

NR

4.6

‘2.8;47.3’(7.4-7.9)

(2.ii.7,

12.9(NW

’ Data generated at ERL-Duluth.water (DMW)).

2 Static, unmeasured test.3 Dowden and Bennett, 1965.

C. dubia were 424 h old at test initiation and fed. Water used was soft reconstituted water (diluted mineral

* Data generated at ERL-Duluth and values represent those from 7-d fathead minnow growth and survival tests and daily renewals.5 Adelman et al., 1976.* Pat&k et al., 1968.

Note: (-) = Confidence interval cannot be calculated as no partial mortality occurred. NR = Not reported; SRW = soft reconstituted water;VHRW = very hard reconstituted water; RW = reconstituted water; LSW = Lake Superior water.

There is another factor which must be consideredwhen carrying out pH adjustment tests. In those ma-nipulations where the pH is changed to pH 3 or pH 11and then readjusted to pH i, the pH may tend to driftover the course of the 48-h or 96-h toxicity tests. Thedrift can be very dissimilar among test manipulations.This is likely to occur even though the starting pH’s (ofsamples readjusted to pH i ) may be similar. This canlead to confusion in interpreting Phase I results if acompound whose toxicity is pH dependent is present inthe sample. An example of a manipulation in which thiseffect is encountered routinely is the pH 3 adjustment/aeration test (Section 8.5). For instance, an aliquot ofan effluent with a pH i of 7.5 is adjusted to pH 3 and iskept at that pH while the other manipulations are con-ducted. This pH 3 adjusted sample serves as a control

.for the pH 3 adjustments. Another portion of the pH 3adjusted effluent is aerated (Section 8.5). Both aliquotsare then readjusted to pH i (75) prior to toxicity test-ing. The pH of the pH 3 adjustment/aeration test solu-tion will probably not behave in a similar manner to thebaseline or pH 3 adjustment test. We have obsen/edthe pH in this test to go unchanged or drift downwardafter adjustment up to pH i of 7.5 over the course of the

toxicity test. However, the pH of the effluent in thebaseline test and the pH 11 adjustment test may driftupwards over the course of the toxicity test, from pH i(7.5) to as high as pH 8.5. By the end of the test, theanalyst may be confronted with interpreting the resultsof tests conducted at different pH values. If a com-pound whose toxicity is dependent upon pH (e.g., am-monia) is present in the sample (cf., Section 8.9 for adiscussion of the effects of pH on ammonia toxicity),the fact that pH either did not change, or even drifteddown in the pH 3 adjustment/aeration test sample (rela-tive to the baseline and/or pH 3 adjustment test), cancomplicate interpreting the test results. If ammonia werepresent (which is less toxic at a low pH), the samplewould appear to have lost toxicity in the pH 3 adjust-ment/aeration test, when the loss in toxicity may havebeen the result only of the differences in pH drift duringthe toxicity tests.

The pH 3 adjustment/aeration test is not the onlymanipulation that may cause differential pH drift overthe course of the toxicity test. Virtually all the manipula-tions in Phase I have the potential to cause this effect.For example, with some effluents any pH 3 manipula-tion (Le., pH adjustment, aeration, or filtration) followed

8-l 4

by readjustment to pH i, will cause pH to behavedifferently than the pH of the baseline test. Similarly,the pH 11 manipulations (i.e., pH adjustment, aeration,or filtration) can cause similar fluctuations, but they donot seem to occur as frequently as with the pH 3manipulations. Another manipulation that causes thisdifferential pH dnf-t is passing effluent over the C,, SPEcolumn at pH i, pH 3 or pH 9 (Section 8.6). Althoughnot as drastic as some of the effects observed with thepH 3 adjustment/aeration tests. the sample collectedafter passing the effluent over the C,, SPE column mayhave a slightly lower pH by the end of the toxicity testthan the pH of the baseline test (e.g., pH 8.2 as op-posed to pH 8.5). A final manipulation that has thepotential to cause acidic pH drift is the addition ofEDTA; details of this pH fluctuation are elaborated inSection 8.8.

Although ammonia is a commonly encounteredsample toxicant whose toxicity is pH dependent, it isnot the only compound whose toxicity can be affectedby different test pH’s (cf ., Phase II). We have observedpH dependence with several metals and the effects ofdifferential pH drift after various Phase I manipulationsshould be considered. pH should always be monitoredand recorded whenever mortality readings are made(e.g., 2 h, 24 h) as well as at the end of the test. It isparticularly important to record pH of the concentra-tions that determine the LC50, especially if greater than5 mg/L of total ammonia is present in the sample.Differential pH drift after manipulations can be over-come by closely monitoring the test pH, and adjustingthe pH in the manipulated samples to match the pH ofthe baseline toxicity test. These adjustments are donebefore animals are introduced.

8.4 pH Adjustment/Filtration lestPrinciples/General Discussion:

The filtration experiment provides information oneffluent toxicants associated with fitterable material. Toxicpollutants associated with particles may be less biologi-cally available. However, aquatic organisms can beexposed to these pollutants through ingestion of theparticles. This route of exposure may be significant forcladocerans and other filter feeders ingesting bacterialcells and other solids with sorbed toxicants. The de-gree to which any compound exists sorbed or in solu-tion depends on a number of factors including particlesurface charge (or lack thereof), surface area, com-pound polarity and charge, solubility and the effluentmatrix. By filtering particles from the effluent, an imme-diate cause or a sink of toxic chemicals may be re-moved.

In addition to determining the effect of filtration onthe toxicity of the whole effluent, the effects of pHadjustment in combination with filtration are also as-sessed with this manipulation. As discussed in Section8.3, changes in solution pH can result in the formationof insoluble complexes of metals (Figure 8-4). 8-4). Similarly,organic acids and bases existing in ionic form can betransformed into the non-ionic form by pH adjustment.

Shifts in effluent pH can also act to drive dissolvedtoxicants Onto particles in the effluent (e.g., shifting thedissolved/sorbed equilibrium away from the free form).Changes in toxicant polarity resulting from solution pHchange can make some particle/toxicant interactionsstronger. In other cases, the increase in effluent ionicstrength resulting from the shift in pH may force non-polar organic compounds onto uncharged surfaces to agreater extent.

By filtering pH adjusted aliquots of effluents, thosecompounds typically in solution at unadjusted pH butinsoluble or associated with particles to a greater ex-tent at more extreme pH’s, are removed. By removingthe toxicant-contaminated particles or precipitated com-pounds prior to readjustment of the sample to pH i,these toxicants are no longer available for dissolution inthe effluent. The pH change may also destroy or dis-solve the particles, thereby removing the sorption sur-faces or driving the dissolved/sorbed equilibrium in theopposite direction.

Positive pressure filtration is recommended. Use ofa vacuum to draw the effluent sample through the filtermay result in a loss of volatile compounds by de.gas-sing the solution during filtration. This problem is p@en-tially worsened in pH adjusted effluents if toxicantsbecome more volatile as a result of pH changes. Ifvacuum filtration is used and effluent toxicity is re-duced, subsequent tests must be performed to definethe nature of the toxicity loss. In this filtration step,whether pressure or vacuum filtering is done, it isimportant to avoid stainless steel housings for eitherthe pH 3 or pH 11 adjustment tests. Teflon, plastic orglass equipment does not have the associated toxicitythat the stainless steel has under acidic or basic condi-tions.

The pH adjustment/C,, SPE test (Section 8.6) re-quires the use of filtered effluent. Without knowledge ofthe effect of filtering on the effluent toxicity, it is notpossible to tell whether or not the SPE column or thefiltration removed the toxicity. Filtering may also beuseful in connection with other Phase I tests.

Volume Required:A 235 mL aliquot of pH i effluent is filtered. Also,

235 mL each of pH 3 and pH 11 effluent aliquots(Section 8.3) are filtered. The remaining 35 mL of eachsolution is reserved for the pH adjustment/aeration tests.A maximum volume of 30 mL of each of these threesolutions is needed to perform the filtration toxicitytests. The exact effluent volume required for the toxicitytest will be a function of the effluent toxicity (Section 7).Each test (pH 3, pH i, pH 11) 11) requires four exposureconcentrations (10 mL each). The remaining filteredeffluent volumes (+200 mL) of pH 3, pH 11, and thepH i solutions are each resewed for the C,, SPE tests(Section 8.6). Excess volume has been Included tocover losses occurring during the filtration operation.

For the blanks, 85 mL of dilution water is neededfor each pH test, of which 50 mi will be used in the pHadjustment/SPE test.

8-158-15

Apparatus:Six 250 ml graduate cylinders, six 250 mL bea-

kers, six 50 mL beakers, pump with sample reservoir,teflon tubing, in-line filter housing, ring stands, clamps.Alternatively, vacuum flask, filter stand, clamp, vacuumtubing, water aspirator or vacuum pump. Glass-fiberfilters (nominal size of 1 1 .O.O pm, without organic binder),stainless steel forceps, glass stoppered bottles for acidand base solutions, pH meter and probe, stir plate,perfluorocarbon stir bars, automatic pipette, disposablepipette tips, eye dropper or wide bore pipette, 30 mLbeakers or 1 oz plastic cups, light box and/or micro-scope (optional).

Reagents:Solvents and high purity water for cleaning pump

reservoir and filter, 1 .O, 0.1, and 0.01 N NaOH, 1 .O, 0.1,and 0.01 N HCI (ACS grade in high purity water),buffers for pH meter calibration.

Test Organisms:Test organisms, 75 or more, of the same age and

species.

Procedure:Day 7: First, the filters must be prepared. These

steps are outlined in Figure 8-7. After the filters areprepared, the dilution water at the appropriate pH isfiltered and collected for the toxicity blanks. Finally theeffluent samples at each of the three pH’s are filtered(Figure 8-8). Use of glass-fiber rather than cellulose-based filters should minimize the adsorption and loss ofdissolved non-polar organic compounds from the efflu-ent sample. Adsorption of toxic dissolved compoundsonto the filter or onto particles retained by the filter canlead to spurious results.

filter Preparation: To prepare the 1 .O pm glass-fiber filter for use, wash two 25 mL volumes of highpurity water through the filter. For the pH 3 effluentfiltration test, the filter must be washed with high puritywater adjusted to pH 3 using HCI. Likewise, the filterused with the pH 11 effluent sample must first bewashed with high purity water adjusted to pH 11 usinga concentrated NaOH solution. Washing the filters withwater adjusted to the same pH as the effluent shouldprevent sample contamination with water soluble toxi-cants contained on the filters.

Blank Preparation: The next step is to prepare filterblanks using dilution water (Figure 8-7). These blanksare used to detect the presence of any water solubletoxicants which may remain on the filter following thewashing process. The pH i filtration blank is simplyprepared by passing 50 mL of dilution water (where thepH is unadjusted) through a washed filter. The filtereddilution water is collected and 30 mL of this volume isreserved for the post-C,, SPE column toxicity blank(Section 8.6). The remaining 20 mL is used as a filtra-tion toxicity blank. Again, excess is included to coverany possible loss during rinses.

To prepare the pH 3 filtration blank, 105 mL ofdilution water is adjusted to pH 3 with HCI, cautionbeing. taken to minimize the increase in dilution waterionic strength. Of the pH 3 adjusted dilution water, 20mL is for the pt-l adjustment only test, 35 mL is re-served for use as a toxicity blank in the aeration test(Section 8.9, and 50 mC of pH 3 dilution water ispassed through a filter previously washed with pH 3rinse water. The filtered pH 3 dilution water is collectedand 30 mL of this volume is reserved for the pH 3filtration/C,, SPE toxicity test blank. The remaining 20ml is readjusted to the initial pH of the dilution waterusing NaOH, again taking care not to exceed the initialpH of the dilution water during the readjustment pro-cess. This solution is used in a single exposure toxicitytest as the filtered pH adjustment toxicity blank.

The pH 11 toxicity blank sample is prepared in asimilar fashion using 105 mL of dilution water adjustedto pH 11 with NaOH. Of the pH 11 dilution water, 20mL is reserved for use in the pH adjustment only testand 35 mL for the aeration test. The remaining volume(50 mL) is filtered using the filter previously washedwith pH 11 rinse water and 30 mL of the filtered pH 11d,ilution water is collected for use as the pHpH 11 filtration/C,, SPE blank. The remaining 20 mL is readjusted tothe initial pH of the dilution water with HCI and used asthe pH 11 filtered toxicity blank using a single expo-sure.

_

Sample Preparation: The same filter(s) used toprepare the pH i (or pH 3 or pH 11) dilution waterfiltration blank(s) is now used to filter the pH i (or pH 3or pH 11) effluent. First, a 235 mL aliquot of the pH ieffluent is passed through the pH i prepared filter andcollected; of which 200 mL is reserved for the C,, SPEtest. The remaining volume (approximately 30 mL) isheld for the pH adjustment/filtration toxicity tests.

Now, using the same filter used to prepare the pH3 filtration blank, 235 ml of pH 3 effluent (see Section8.3) is filtered and collected. The filtered pH 3 effluentis split into two aliquots (200 mL and approximately 30mL). The 200 mL aliquot is used in the pH 3 adjust-ment/C,, SPE test. The 35 mL filtered aliquot is read-justed to pH i using NaOH. Care must be taken tominimize both an increase in aliquot volume and ionicstrength. The pH readjusted 35 mL aliquot is held forthe toxicity testing.

Finally, the filtration step is repeated using 235 mLof the pH 11 effluent (Section 8.3) and the filter origi-nally used to filter pH 11 dilution water. Again, 200 mLof the pH 11 filtered effluent is used in the pH adjust-ment/C,,, SPE test and the filtered sample (approxi-mately 30-35 mL) is readjusted to the pH i of theeffluent with HCI and used to conduct a toxicity test onday 2.

In filtering effluent samples with high solids content,it may be necessary to use more than one filter for the235 mL of effluent. If so, the filter preparation step must

-8-168-16

Figure 8-7. Overview of steps needed in preparing the filter and the dilution water blanks for the filtration and/or the C,, solid phase extraction column tests.

Preparingthe Filter

Day 1

PreparingDilutionWaterBlanks

__

IDay 2

.-

NaOH

300 mL High Purity Water

HCI -J

150mLofpH3 1

ti

Filter

1 Discard water 1

50 mL of pH i(unadj.)

(Dilution1

‘P+cNaOH

50mLofpH 11.

Filter

Discard water

t - HCIu50mLo;pH3 t 50mLofpH i

t l

prepared filter

NaOH --+ t

v

Through C18SPE Column

IIC NaOH HCI -+ + - HCI

.

be repeated to provide additional filtration blanks orseveral filters can be prepared at one time by stackingthem together in the filter housing. Alternativeiy, it maybe possible to centrifuge samples high in ‘suspendedsolids and filter the supernatant through a single filter. Ifthis option is taken, the toxicity of the supernatant mustbe tested and compared to the toxicity in the baselineeffluent test.

In the above procedures, either separate effluentfiltration systems should be used or the filtration systemmust be cleaned between pH adjusted aliquots to pre-vent any carry-over of toxicity or particles. This meansall equipment should be thoroughly rinsed with 10%HNO,, acetone, and high purity water between aliquotsof effluent with the exception of stainless steel equip-ment where dilute solutions should be used.

The pH of the pH adjusted blanks and effluentaliquots, designated for day 2 toxicity tests, should bechecked periodically throughout the work day. Adjust-ments should be made as necessary in order to main-tain the pH i of these solutions.

Day 2: Prior to initiating the toxicity tests, the pH ofthe pH 3 and 11 blanks and filtered effluent aliquotsshould be measured and readjusted to pH i. Toxicitytests performed on all three (pH 3, pH i, and pH 11)filtration blanks involve testing without dilution. Basedon the 24-h initial LC50 of the day 1 effluent, toxicitytests performed on the effluent aliquots filtered at pH 3,pH 11, and pH i are set up at 4x-, 2x-, 1 x-, and 0.5x-LCSO as described in Section 8.2. Measurement ofexposure pH should be made daily on concentrationsaround the mortality and the highest tested concentra-tion, concurrently with survival readings. A sample datasheet for the filtration tests is shown in Figure 8-9.

Interferences/Controls and Blanks:Controls prepared for the baseline toxicity test serve

as a check on the quality of organisms, dilution waterand test conditions. Results of the pH adjustment test(Section 8.3) will indicate whether or not toxic levels ofNaCl have been produced through pH adjustment only.

Results of the effluent filtration tests at each pHshould be compared with the filtration dilution watertoxicity blank performed at the corresponding pH todetermine the validity of the toxicity test outcome. Nosignificant mortality should occur in any of the filtrationblanks. If unacceptable mortality of organisms occurs ineither the pH 3 or pH 11 adjusted filtration blanks,further investigation will be necessary to determinewhether lethality resulted from toxicants leached fromthe filter at pH 3 and/or pH 11, or whether the increasein dilution water ionic strength (via acid and base addi-tion) is responsible for the problem. Additionally, if thepH 3 and/or 11 filtration, aeration (Section 8.5) and CI,SPE (Section 8.6) dilution water blanks have approxt-

’ mately the same final concentration of acid and base,any ionic strength related toxicity should also be de-tected in them.

If a filtration toxicity blank shows unacceptable acutetoxicity but the corresponding filtered effluent is equallyor less toxic than the baseline toxicity test, it is possiblethat the dilution water toxicity blank removed the finaltraces of toxic filter artifacts. In some cases, the efflu-ent matrix may have also prevented the artifacts fromleaving the filter or masked their presence. Alternativelythe observed filtered effluent toxicity may represent thenet effect of toxicant removal via filtration plus contami-nation by filter artifacts.

Results/Subsequent Tests:The LCSOs for the aliquots of pH 3, pH i and pH 11

filtered effluent are compared to the baseline effluentLCSO to determine whether any of these processesresulted in a significant change in effluent toxicity.

If toxicity can be removed by filtration, either with orwithout pH change, one has a method for separatingthe toxicants from other material in the effluent. Thisknowledge itself provides an important advance be-cause further characterization and analyses will be lessconfused by non-toxic constituents. Usually further char-acterization will be the next step. Tests must be de-signed to determine whether the mechanisms causingremoval are precipitation, sorption, change in equilib-rium or volatilization. One necessary step is to trecoverthe toxicity from the filter. If this cannot be done and theloss is not by volatilization, then the whole experimentmay have little utility. Comparisons of pressure andvacuum filtration may reveal if volatilization is involved.If characterization of the toxicant is also achievedthrough other tests, filtration can be used to remove thetoxicants. Then if the suspected toxicant is the trueone, its concentration should be lower or zero aftertoxicity is removed by filtration.

If any or all of these pHIfiltration combinations re-sult in less effluent toxicity (not attributable to the ef-fects of pH adjustment alone), it may be possible toconfirm the findings of the test. A transfer of the solidscontained on the filter back into the filtrate at pH i canbe attempted by reversing the flow of the filtrate throughthe filter or by rinsing the solids off the filter with filtrate.The toxicity exhibited by this solution should be similarto that of the original effluent, provided that the finalconcentration of solids in the test solution approximatesthe solids level in the sample that was filtered. Forprecipitates formed as a result of pH changes or forcontamination of suspended solids facilitated by pHadjustment, time must be allowed for the precipitate toredissolve in the pH i filtrate or for a new equilibrium tobe set up between the contaminants on the solids andin solution. The results of this test are not likely to bequantitative due to the recovery problems inherent inthe process.

In order to determine whether the effluent matrixaffects the toxicity of filterable particles (e.g., its ionicstrength, dissolved organic carbon content), the filteredmaterial can also be added to a volume of pH i dilutionwater equal to the volume of effluent that was passed

8-I 9

Figure 8-9. Example data sheet for filtration test.

Test Type: Filtration Species/Age:Test Initiation (Date & Time): No. Animals/No. Reps:

Investigator:Sample Log No., Name:Date of Collection:

Source of Animals:Dilution Water/Control:Test Volume:Other Info:

Note: See baseline data sheet for control data.

nt-4 ?p’I ”

pH ipH 11

Volumes and Strength of Solutions Added:HCI NaOH

Comments:8-20

through the filter. The toxicity of this dilution water,spiked with effluent solids, can be compared to thetoxicity of the unfiltered (baseline) effluent and the fii-trate spiked with its own solids.

pH T h e a d d i t i o n a l t e s t s s u g g e s t e d h e r e i nmay or may not provide the relevant information.

t o r e c o v e r t h e toxicant fromi s t o sonicate w i t h a s o l v e n t ( e . g . ,

w a t e r ) . F o r s o m e e f f l u e n t s , 1 1 a d j u s t m e n t /

filtration test. These results may cause one to suspecteither cationic a s t h e toxicant

W e h a v e h a d s u c c e s s i n r e c o v e r i n gtoxicity from the glass fiber filter when the filter was

pH 3 a d j u s t e d d i l u t i o n w a t e r . T y p i -cally, 300 mL issonicated in 75 mL of pH 3 dilution water, this concen-trates the toxicity 4x (theoretically). The pH of theconcentrate is then readjusted to pH i before use in thetoxicity test. For some effluents the amount of a cat-ionic metal was reduced after filtration and the toxicityin the effluent removed after filtration. Toxicity wasrecovered in the pH 3 dilution water extract of the filter.If methanol is used, after sonication the methanol mustbe concentrated before use in toxicity tests (see Sec-tion 8.6 for methanol tolerances). Additional solventscan be used to recover toxicity off the filters (Schubauer-Berigan and Ankley, 1991).

8.5 pH Ad’usfmenf/Aerafion TestPrinciples/General Dlscussion:

The aeration test is designed to determine howmuch effluent toxicity can be attributed to volatile,sublatable, or oxidizable compounds. The test is per-formed with pH-adjusted and unadjusted (i.e., pH i)effluent. By comparing the toxicity test results for acidic,pH i and basic aerated samples, the toxicity may varyand this knowledge can be used for further character-ization. Some compounds can be removed or oxidizedmost easily at one pH, whereas others are most easilyremoved or oxidized at a different pH. Thus, the aera-tion is performed at several pH values.

Whether a constituent is completely removed, orsufficiently removed to reduce toxicity, depends on manychemical/physical conditions. At a minimum, one mustbe certain that the geometry of the sparging process isalways the same for a given effluent sample and thatthe duration is constant, otherwise the test is of littlevalue. The pH of many effluents will change, some-times rapidly (cf., Section 8.3) during sparging and sopH must be frequently checked and maintained duringthe entire aeration period.

Oxidation can change the constituents in manyways and one must determine if oxidation or spargingis the mechanism before additional tests can be de-signed. Water soluble constituents such as ammonia

8-21

and possibly cyanide are not readily stripped usingtechniques described in the Procedure below, and oneshould not assume that they will be removed. Underboth air and nitrogen sparging, a removal process, inaddition to volatilization, may occur. Sparging can re-move Surface active agents from solution by the pro-cess of sublation (lifting up, carrying away). Surfaceactive agents have a molecular structure that includesa polar end (either ionic or nonionic) and a relativelylarge non-polar, hydrocarbon end. Some examples ofsurface active agents are resin acids, soaps, deter-gents, charged stabilization polymers and coagulationpolymers used in chemical manufacturing processes.The process of sublation occurs because duringsparging, surface active agents congregate at the liq-uid/gas interface of the air or nitrogen bubbles and arecarried along with the gas bubbles to the surface of thesparged liquid. As the bubbles break up, they aredeposited and concentrated with continuous spargingat the surface of the sparged liquid, and the sides ofthe aeration vessel. After sparging, a faint deposit mayor may not be visible on the sides of the aerationvessel.

Air is used for sparging so that oxidation is in-cluded. Subsequent tests with nitrogen may betsed toseparate sparging from oxidation and tests describedu n d e i R e s u l t s / S u b s e q u e n t T e s t s can be used forsublatable compounds. We have grouped the tests toavoid many tests initially.

Volume Requlred:Thirty-five mL volumes of each pH 3, pH 11 (see

Section 8.3) and pH i effluent are needed for this test.A maximum volume of 30 mL of each of' of' these solutionsis required for the toxicity tests on aerated solutions. Anexcess volume has been provided to allow for lossesthrough aeration. Each toxicity test utilizes four expo-sure concentrations (10 mL each) without replication.The exact volume required for the toxicity test on eachpH adjusted or unadjusted aerated solution will dependon the toxicity of the effluent (the 24-h initial LC50).

The amount of dilution water that was used for thetoxicity blanks (35 mL) is kept the same as the effluentfor each pH. An excess volume has been provided toallow for any volume loss through aeration (cf., Section8.3).

Apparatus:Aeration device or compressed air system with a

0.22 urn filter, six air flow regulators, six glass diffusers,six 50 mL graduated cylinders with ground glass stop-pers, glass stoppered bottles for acid and base solu-tions, pH meter and probe, stir plate(s), penluorocarbonstir bars, automatic pipette, disposable pipette tips, eyedropper or wide bore pipette, 30 mL beakers or 1 ozplastic cups, light box and/or a microscope (optional).

Reagents:1.0, 0.1, and 0.01 N NaOH, 1.0, 0.1, and 0.01 HCI

(ACS grade in high purity water), buffers for pH metercalibration.

Test Organisms:Test organisms, 75 or more, of the same age and

species.

Procedure :

Day 7: Six different solutions are aerated in thistest; pH 3, pH i, and pH 11 effluent, and pH 3, pH i,and pH 11 dilution water, (cf., Sections 8.3 and 8.4,respectively for preparation of pH adjusted effluent anddilution water). A flow chart for the effluent samples ofthe pH adjustment/aeration test is shown in Figure 8-10. Each sample is transferred to a 50 mL cylindercontaining a small per-fluorocarbon stir bar. The diam-eter and length of the pH probe must be such that itcan be placed into the solution during aeration. Thetaller the water column and the smaller the bubbles, thebetter the stripping will be. Each solution should bemoderately aerated (approximately 500 mL air/min) fora standard time, such as 60 min. Formation of precipi-tates may or not be important and should be noted.

The pH of the acidic and basic effluent and dilutionwater aliquots should be checked every 5 min duringthe first 30 min of aeration and every 10 min thereafter.If the pH of any pH 3 or pH 11 solution drifts more than0.2 pH units, it must be readjusted back to the nominalvalue. The volume and concentration of additional acidand/or base added to the solutions should be recordedso that the final concentration of Na+ and Cl- in eachsolution can be calculated following final pH readjust-ment. Solutions should be stirred slowly during any pHreadjustment. Again, precautions must be taken in or-der to minimize the amount of acid and base added.Note that the aeration time does not include the timeintervals during which aeration is temporarily discontin-ued to readjust pH. A constant pH is not maintained inthe “pH i ” effluent because this solution represents thegeneralized effects of aeration on the effluent withoutregard to pH. Only slight changes in the pH of thedilution water at its initial pH are expected since suchwater is usually at air equilibrium before the start of the

. manipulation.

The sparged sample must be removed from thegraduated cylinder for toxicity testing so that any toxi-cant that may have been sublated is not redissolved inthe sparged sample. This may happen if the sample issimply poured from the cylinder, and sublation wouldnever be suspected. Therefore, one way to transfer theeffluent sample is by pipetting it out of the cylinder,exercising care to prevent any sample from contactingthe sides of the cylinder above the liquid level. Forexample, when using a 100 mL graduated cylinder forthe aeration vessel, a 50 mL pipette can be used toremove the 30 mL sample. At this point it may bepossible to recover a sublated toxicant from the sidesof the cylinder. This must be done at the end of theaeration step; see ResuWSubsequent Tests sectionbelow for details.

Sparging air contaminated with oil (droplets or va-por) or any other substance is unacceptable. Contami-

-

nated air is probable from air lines containing oil or incases where the source of the air is contaminated (e.g.,boiler room). Small air pumps, sold for home aquariaare adequate, but only if the room air is free of chemi-cals or contaminants. Chemistry laboratories where con-centrated chemicals and solvents are used often mightnot have suitable air quality.

Following aeration, the pH of each solution (includ-ing the 35 mL portions of pH unadjusted effluent (pH i )and dilution water) is returned to the pH of the initialeffluent or dilution water using the necessary volumesof NaOH and HCI. Returning all effluent solutions to theinitial pH of the wastewater will ensure that a validcomparison can be made with the baseline LC50. ThepH of each sample must be checked periodicallythroughout the remaining work day and readjusted asnecessary. If stable pH can be attained prior to toxicitytest initiation, the pH during the test is likely to changeless.

Day 2: Before initiating the toxicity tests, the pH ofall of the aerated effluent and blank solutions should bechecked and adjusted to pH i . Toxicity tests are per-formed on a single 100% concentration of all threedilution water blanks (pH 3, pH i and pH 11). Tbesedilution water blanks will provide information on toxicartifacts resulting from aeration.

Based on the 24-h initial LC50 of the day 1 effluent,toxicity tests are performed on each aerated effluentsolution at concentrations of 4x- (or 1 OO%), 2x-, lx-,and 0.5x-LC50 (cf., Section 8.2). The pH of each testconcentration should be measured and recorded daily.An example of the data sheet for the aeration test isgiven in Figure 8-l 1.

Interferences/Controls and Blanks:Dilution water controls prepared for the baseline

toxicity test also act as controls on organisms, dilutionwater and test conditions for this test. Results of the pHadjustment test (Section 8.3) will suggest whether ornot toxic levels of NaCl may have been reached as aresult of the addition of acids and bases to the effluent.

No significant mortality should occur in any of thethree aeration blanks. If there is significant mortality,the cause must be found and corrected before the testcan be meaningful. To determine which factor(s) causedblank toxicity, the toxicity of pH adjusted aerated dilu-tion water can be compared to that in the same pHadjusted, unaerated dilution water. Approximately thesame quantities and concentrations of acid and baseshould be added to both of these samples of dilutionwater to make them comparable. Blank toxicity in thepH adjusted and unadjusted aerated dilution water sug-gests contaminated air. Other possible causes includecontaminated equipment, such as electrodes or glass-ware (especially where low or high pH solutions were incontact), or the addition of too much acid or base.Another approach to this blank question involves evalu-ating the pH adjustment/filtration and pH adjustment/C,, SPE blanks (Sections 8.4 and 8.6). Assuming the

8-22

LY+ r-lc9

c

8-23

Figure 8-11. Example data sheet for aeration test.

Test Type: AerationTest Initiation (Date & Time):

Investigator:Sample Log No., Name:Date of Collection:

Species/Age: aNo. Animals/No. Reps:Source of Animals:Dilution Water/Control:Test Volume:Other Info:

Note: See baseline data sheet for control data.

nH -3pm. v

pH ipH 11 11

Volumes and Strength of Solutions Added:HCI NaOH

Comments:

8-24

concentration of acid and base in the final blank solu-tion is approximately the same in all dilution waters forthe three tests, toxicity in the aeration blanks but not inthe filtration or C,, SPE blanks suggests that aerationrather than pH adjustment has led to contamination.Compare the toxicity from the effluent baseline test tothe toxicity of all three aerated effluent samples. Whenthe baseline toxicity is significantly less than that of anyone of the aerated samples, toxicity was added orcreated during effluent manipulations. This check isespecially important because pH adjustment of aeratedeffluent may have required larger quantities of acid andbase as compared to the pH adjustment test (Section8.3).

If nitrogen was used for stripping, DO depletion islikely to have occurred. If a relatively large surface-to-volume ratio is used (such as the 10 mL volume in a 1oz plastic cup) during the overnight holding period, DOshould not be a problem.

Results/Subsequent Tests:The LCSOs for the aliquots of pH 3, pH i, and pH 11

aerated effluent are compared to the baseline effluentLC50 to determine whether any of these processesresulted in a significant change in effluent toxicity. If asubstantial reduction in toxicity is seen for any or all ofthe three aerated effluent solutions, one must nextdetermine whether the separation of effects was causedby sparging, oxidation, or sublation. This is done byrepeating those tests in which toxicity was reduced,substituting nitrogen for air in the stripping process.Use of nitrogen eliminates oxidation as a remo.vaI pro-cess. If side-by-side effluent stripping tests with air andnitrogen provide the same results, toxicant removal isprobably caused by the sparging process. If only thetest(s) conducted with air succeeds in reducing or re-moving effluent toxicity, oxidation is a probable cause.An effluent sample may contain toxicants removedthrough sparging and oxidation. An example of this iswhere aeration at pH 3 and pH i reduces toxicity, butnitrogen stripping removes the toxicity only in the pH 3effluent. Using the Procedure described above ammo-nia should not be air-stripped; however, if differentaeration vessels are used and greater surface area isused, ammonia can be reduced/stripped. If toxicity isreduced at pH 11 compared to pH i aerated, pH 3aerated samples, or the baseline test, measuring theammonia concentration after aeration can be informa-tive.

To determine if toxicity is due to sublatable com-pounds, toxicant recovery can be attempted by addingdilution water to the emptied cylinder used for sparging(preferably graduated cylinders with ground glass stop-pers), stoppering it .and shaking it vigorously, makingsure that the sides of the cylinder are thoroughly rinsedwith the dilution water. This dilution water is then testedfor toxicity. Recovery of the sublated material providesfurther evidence that a surface active compound waspresent. A more concentrated solution of the sublatedmaterial can be obtained by using a larger sample

:

volume for sparging, such as 90 mL, and less dilutionwater (i.e., 30 mL) to recover the toxicant( Thiswould result in a nominal concentration of 3x the wholeeffluent concentration of the toxicant( To avoid spuri-ous results in cases where sublated toxicity is recov-ered, a dilution water blank should be run. A dilutionwater sample should be subjected to the sparging stepand to the toxicant recovery step. In that way, if toxicityis inadvertently being added to samples during themanipulation (from contaminated glassware or contami-nated air or nitrogen supply) it should occur in the blanksample. As some sublated compounds are difficuft torecover. from the glass surface, a solvent rinse withmethanol may result in more efficient recovery. If sol-vents are used and because of the low concentrationfactor involved, most of the solvent will have to be aireddown in order to have an adequate water concentrationto perform the toxicity test. If solvents are needed, oneis wise to scale up the volumes to obtain higher con-centrations for testing. However, not all kinds of surfaceactive agents are removed to the same degree, andsome are not removed at all (Ankley et al., 1990). Thisis probably due to factors such as matrix effects andsolubility. Recovery of a sublated toxicant can be diffi-cult. Consequently, reduction of toxicity by spargingwith air and nitrogen can be an indication of a tUxicantwhich is a surface active agent, but a lack of toxicityremoval does not rule out the presence of these com-pounds.

In the pH adjustment/aeration test, removal of toxi-cants by precipitation resulting from pH change aloneshould also be detected by the pH adjustment/filtrationtest. Oxidation of compounds can cause precipitation. Ifoxidation is the cause, the pH adjustment/filtration testwill not change toxicity. If nitrogen sparging has re-moved the toxicant, the “volatile toxicant transfer” ex-periment described below may provide separation ofthe volatile toxicant from other constituents. Our experi-ence with this technique is limited to a few effluents. Toperform the “volatile toxicant transfer” experiment, aclosed loop stripping/trapping apparatus is used (Fig-ure 8-12). This apparatus consists of a pump which cancirculate air or nitrogen gas, two airtight fluid reservoirs,perfluorocarbon tubing, and diffusers. The arrangementshould be such that air or nitrogen can be passedthrough the effluent in one reservoir and then throughthe dilution water in the second reservoir before cyclingback to the first reservoir. The reservoir of the dilutionwater serves as a trap that will collect the volatilizedtoxicant( Of utmost importance to this experiment isan air tight system. The time to equilibrium of thevolatile toxicant will be dependent upon the efficiencyof the sparging process and the rate of volatilization ofthe toxicant which may be affected by pH. For ex-ample, using a glass or plastic pipette to aerate thesamples may not effectively sparge the entire volumeof sample. To optimize the toxicant recovery, use ofgas washing bottles (for example, 125 mL and 500 mLbottles from Kontes Glass Co., Vineland, NJ) fitted withglass frit diffusers is suggested because they spargethe sample volumes more effectively.

8-25

l b

Figure 8-12. Closed loop schematic for volatile chemicals.

Numerous operating conditions can be selected,each providing different information. This system should

so that if the trap is inefficient in removing the toxicantfrom the nitrogen, the toxicant will not be lost from

not be operated as a conventional purge and trap the sample. 8ecause conditions to optimize transfersystem. The reason is that since one does not knowthe identity of the toxicant( the conditions for trapping

cannot be selected until the chemical identity is known,

are not known. Initially, the objective should be to getlonger sparging times should be chosen,

measurable toxicity moved into the dilution water me- The first experiments should involve no pH changesdium in the trap. This will establish that there are at if any measurable change in toxicity occurred in theleast some volatile toxicants present. At this stage the earlier tests without the pH 3 or pH 11 adjustments.goal is not to move all the toxicant to the dilution The reason for this selection is that drastic changes inwater in the trap. If the same concentration of the pH can cause many unknown effluent changes, andtoxicant in the effluent can be transferred to the dilution artifacts are more likely to occur. Of course if pH changeswater as exists in the unaltered effluent, the data are are required to change toxicity, then pH will have to beeasiest to interpret. For this purpose the volume of altered. When pH is altered, then equilibrium objec-sparged effluent should be large and the dilution water tives, mentioned above, are not possible and the entirevolume in the trap small. The nitrogen gas is recirculated process takes on characteristics of more conventional

8-26

purge and trap experiments. The usual resin trapsdescribed in EPA methods are not suitable because

ously, relative degrees of water solubility exist, Many

the trap cannot be tested for toxicity. The trappinghighly toxic pollutants found in effluents at very lowconcentrations are not considered to be water soluble

medium must be, or be able to be, made into a toxicity despite the fact that they are present at toxic concen-testable water. trations.

In those instances when sparging affects toxicityonly when accompanied by a pH change (pH 3 orpH 1 1 1 ), the method to be used to operate it as aconventional purge and trap is as follows. The trap’sdilution water volume should be small relative to thesample volume and its pH should be opposite that ofthe sample pH (e.g., if the sample pH is 3, then the trappH should be 11). One can no longer conclude any-thing about the original effluent equilibrium, and theprocedure is one of separation. Toxicity in the trap mayor may not be caused by the same substance as thatwhich causes the original effluent toxicity. Obviously, allthe precautions mentioned above regarding NaCl addi-tion and other adjustments must be tracked with blanksjust as in any other experiment. We have not foundmany effluents where the transfer technique is useful,but for those effluents where it works, it is a powerfultool. We have found the volatile toxicant transfer ex-periment to be useful with some samples (i.e., sedimentpore water), where two pH dependent toxicants (e.g.,ammonia and hydrogen sulfide) are suspect. There issometimes an appreciable loss of toxicity after the pH 3aeration step in samples with ammonia toxicity, yet it isunknown whether the toxicity loss is due to volatiliza-tion of hydrogen sulfide (or some other pH dependenttoxicant) at low pH, or is an artifactual decrease ofammonia toxicity due to a downward pH drift in the test(cf., Section 6.3). In this case, a trap such as the onedescribed previously for transferring a volatile toxicantat altered pH is useful. Water in the trap that volumetri-cally concentrates the toxicant at two or more times itswhole sample concentration may be successfully testedfor toxicity. We suggest, in the case of suspected hy-drogen sulfide toxicity, testing the trap water at pH 6,as the toxicity of hydrogen sulfide is enhanced at thatpH. One caution in this setup is that the volatility ofsome pH dependent toxicants such as hydrogen sulfidemakes it imperative that the experiment be initiatedimmediately after adjusting the pH to minimize theirloss.

8.6 pH Adjustmen t/C,, Solid Phase

Obvi-

Compounds extracted by the C,, pH

or chloroform. The C,,

pH a n d a h i g h pH,

C,, S P E c o l u m n . B e c a u s e o f C,,column degradation, pH’sused. To ensure column integrity, the pHto be used on the SPE columns will be either pHor pH pH 11) in this manipulation.Manufacturer’s data should be consulted for tolerablecolumn pH ranges and for exact column conditioning

Of the 235 mL pH) filterEd8.4), 3 5 mLpH) i s h e l d f o r t h e pHtest. Now, the additional 200 mLthrough the C,, (pH 3 , pH i,and pHsure concentrations (10 mL

pH adjusted or pH i post-column effluent will dependon the toxicity of the effluent (the 24-h initial LC50).

For the blanks, 30 mL of pH adjusted (pH 3 or pH11) and/or filtered (pH i) dilution water is needed. Thelast 10 mL of the post-column water should be used forblank toxicity tests.

Apparatus:Six 250 mL graduated cylinders, eight 25 mL gradu-

ated cylinders, glass stoppered bottles for acid andbase solutions, pH meter and probe, stir plate,perfluorocarbon stir bars, pump with sample reservoir,perfluorocarbon tubing, ring stands, clamps, three 3 mLC, SPE columns (200 mg sorbent), automatic pipette(18 mL), disposable pipette tips (10 mL), eye dropperor wide bore pipette, 30 mL glass beakers of 1 ozplastic cups, light box and/or microscope (optional).

Reagents:HPLC grade methanol, high purity water, 1 .O, 0.1,

and 0.01 N NaOH, 1.0, 0.1, and 0.01 N HCI (ACSgrade in high purity water), buffers for pH meter calibra-tion, acetone and methanol for cleaning the pump andreservoir, and vials to collect methanol eluate.

Test Organisms:Test organisms, 135 or more, of the same age and

species.

a-27 .

Procedure:Day 7: This procedure is performed with effluent

samples adjusted to the various pH’s; however, themanipulations have three distinct steps (Figure 8-13)which are generally the same for each PH. Prior toattaching a new column to the apparatus, the reservoirand pump must be cleaned with acetone, methanol,and high purity water.

Step 1 involves conditioning the solid phase extrac-tion columns for each pH. Column conditioning proce-dures may vary with the manufacturer of the column.The procedures described below are modifications ofthe conditioning steps used with Bakep C,, SPE col-umns (J.T. Baker Chemical Company, Phillipsburg, NJ).

Using a flow-rate of 5 mUmin, 15 mL of HPLCgrade methanol is pumped through the column anddiscarded. Next 15 mL of high purity water, adjusted topH 3 with HCI, is placed in the sample reservoir. Caremust be taken in timing the addition of solutions afterthe methanol has passed through the column. Whilethe mixing of methanol with subsequent solutions mustbe minimized, the column must also be prevented fromgoing dry following the methanol wash and dilutionwater or sample application. The amount of time neededbetween introduction of solutions to prevent any col-umn drying will be unique to each investigator’s appa-ratus. This timing should be determined before per-forming this procedure with actual effluent samples. Ifthe column dries at any time after introduction of themethanol during conditioning, the column must be re-conditioned (with methanol).

As the last volume of pH 3 high purity water isentering the column (Step l), the pH 3 adjusted, filtereddilution water is placed into the reservoir (Step 2).Again, the column must not be allowed to dry beforethe pH 3 dilution water enters the column. The pH 3high purity water passing from the column should bemeasured to determine the point at which the dilutionwater begins to leave the column. This pH 3 high puritywater used to condition the column is discarded, Next,30 mL of the filtered pH 3 dilution water is collected,and the last 10 mL aliquot collected is used for thetoxicity blank to detect toxicity leached from the col-umn. This aliquot will have to be pH re-adjusted to theinitial pH of the dilution water using NaOH, and it isreserved for day 2 toxicity testing. Care should betaken to minimize changes in sample volume and ionicstrength during pH readjustment.

As the last several mL of filtered pH 3 dilution waterare entering the column, the 200 mL volume of filteredpH 3 effluent is placed in the sample reservoir (Step 3).Again, the column sorbent must not be allowed to drybetween the dilution water blank and the effluent. Col-lect a 30 mL aliquot of post-column effluent after 25 mLof the sample passes through the system. A secondpost-column 30 mL aliquot is collected after a total of150 mL of the sample passes through the column.Collection of the first post-column sample after 25 mLof sample has passed the column ensures that any

dilution water left in the system will not be present inthe post-column sample. The second subsample ofpost-column solution provides information on columnoverloading and toxicant t h e s e30 mL pH i u s i n g t h e

NaOH. T h e t o t a l v o l u m e o f NaOHnecessary for pH a d j u s t m e n t s h o u l d b e r e c o r d e d . Thesealiquots are reserved for day 2 toxicity testing. Columns

b u t s h o u l d b e s a v e d f o r s u b s e q u e n telution (see the Results/Subsequent Tests section).

Receiving water should not be used as the dilutionwater because trace organic and metal contaminantsor organics (such as humic acid) may be present. If forany reason, such a water is needed, the same columnshould not be used for concentrating the toxic sample.

mL d i l u t i o n w a t e r b l a n k , 3 0 mL

be checked for toxicity. The pH

c

For pH i, the above procedure is repeated using aclean reservoir and pump and a new conditioneM mLC SPE column for the filtered pH i effluent (Figure 8-l$ The pH of the post-column dilution water and post-column effluent should be measured.

In the final C,, SPE manipulation, pH 9 (readjustedfrom pH 11) dilution water and effluent are processedas described above. While use of pH 11 effluent offersthe likely advantage of shifting a larger number of basicorganic compounds farther towards the predominatelyunionized form, and therefore removal, the C SPEcolumn cannot withstand a pH above 10. For t& rea-son, the pH 11 filtered dilution water and sample aliquotsprepared in Sections 8.3 and 8.4 are readjusted to pH9 with HCI before they are put through the column. The15 mL of high purity water used to rinse the columnfollowing methanol conditioning must also be read-justed to pH 9 with NaOH. The 10 mL aliquot of post-column pH 9 dilution water and both 30 mL aiiquots ofpost-column pH 9 effluent are further adjusted to theirpH i’s respectively, prior to toxicity testing. The totalvolume of acid added for pH readjustment is recorded.The pH of all aliquots of the chromatographed dilutionwater and effluent should be checked and readjustedas appropriate throughout the remainder of the workday.

Day 2: The pH of all of the post-column dilutionwater and effluent aliquots should be checked andreadjusted if pH has drifted overnight. Toxicity tests areperformed on a single 100% concentration of all threeof the dilution water blanks. These blanks will provideinformation on the presence of toxicity leached from theC,, column at different pH’s.

The six 30 mL post-column effluent aliquots aretested for toxicity using an exposure series based onthe 24-h LC50 of the original effluent. Chromatographedeffluent aliquots are tested at concentrations of 4x-,

8-28

Figure 8-13. Step-wise diagram for preparing the C,, solid phase extraction column samples.

step 1

r

Prepare three Cl8 SPE’ Columns lll 15 15 mL methanoll 15 ml_ high purity water

00 NOT LET SOABENT GO DRY

Olscard Methanol 8Water After Rinses 1

step 2

step 3

L 4

Oilutlon Water

1 I\

HCI +, * NaOH

30mLatpH3Rltered Water

30mLatpH9Filtered Water

Collect 10 mL Sample

IDay I

Prepared Column2

1 If column will be eluted with 1 mL methanol (cf., Resu/&‘Subseque~ Tests), collect methanol column blankbefore dilution water is passed over column. Column should go to dryness and will have to bere-cxnditioned (Step 1) before proceeding to Step 2.

2 Use same column used with the dilution water unless receiving water is used (see text for details).3 Same test as depicted in Figure 8-8.

8-29

2x-, lx-, 0.5x-LC50 (cf., Section 8.2). The pH of eachsolutlon tested should be measured daily and recordedalong with organism survival. A sample data sheet forthe C,, SPE test is shown in Figure 8-14.

Interferences/Controls and Blanks:Controls on test organism performance, dilution

water quality, and test conditions are provided by thecontrol from the baseline toxicity test. The pH adjust-ment and filtration tests (Sections 8.3 and 8.4) provideinformation on the effects of pH adjustment and filtra-tion on eff iuent toxicity apart from any additional changescaused by C,, SPE. Effluent and blank test results fromthese two tests must be evaluated prior to interpretingthe results of the C,, SPE test, both in terms of identify-ing any toxic artifacts added during filtration and pHadjustment and in allocating toxicity reduction to thethree components potentially impacting effluent toxi-cants in the C,, SPE test.

Of those methods discussed so far, the C,, tech-nique requires the greatest manipulation. More prob-lems are likely to be encountered with toxic blanksbecause in addition to those factors associated with pHadjustment and filtering, the C,, method also involvesuse of resin and methanol. Blanks for toxicity must bechec:,-3d in the same manner as before for acid andbase addition, filter artifact toxicity, pH drift, as well astoxicity from the C,, column. In addition to these, someeffluents behave in a peculiar way after passing throughthe SPE column (cf., discussion below).

Results/Subsequent Tests:The above unique properties of some effluents and

the potential for column blank toxicity problems makeinterpretation of the test results more subjective.

If toxicity is not reduced in post-column effluent, nottoo much credence should be placed on the results.One needs to go back and sort through the possiblecauses.

If none of the Phase I treatments reduced toxicity(including the C,, SPE column) or if the toxicity wasreduced by the C,, column, it is useful to elute thecolumn with 100% methanol. Of course, a column blankmust also be evaluated; this is a methanol elutionfollowing the column conditioning with methanol andhigh purity water. The column must go to drynessbefore collecting the blank of methanol. After the metha-nol blank is collected, the column must be recondi-tioned as described in Step 1 in Figure 8-13 and thecolumn must not go to dryness before starting thedilution water over the column. If a 1 mL volume ofmethanol is used (for 200 ml of effluent on a 3 mLSPE column) and the sorption and elution efficiencyare loo%, any substances retained by the columns willbe concentrated 200x. To test the eluate, 150 PL of themethanol fraction is diluted to 10 mL with dilution water.The resultant methanol concentration is lS%, which isbelow the 48-h or 96-h LC50 for all species given inTable 8-3. This provides a concentration of effluentconstituents 3x whole effluent concentration. This small

a-30

amount of concentration over whole effluent allowsdetection even if some loss occurred either in sorptionor elution.

If the post-column effluent is not toxic or less toxic,and the methanol eluate is toxic, the next step is toproceed with the Phase II C,, SPE procedure to identifythe toxicant removed by the column. At this point itshould not be assumed that toxicity removed by the C,,SPE column is due to non-polar organic compounds.While metals are not non-polar organic chemicals, theycan be removed from some effluents using the C,, SPEcolumn. However, metals generally are not eluted withmethanol and therefore the fractions are not toxic. Met-als may be eluted with dilution water adjusted to pH 3or pH 11. Surfactants can be sorbed by the C,, SPEcolumn just as other non-polar organics, and someelute with methanol.

If neither the post-column effluent nor the methanolfraction is toxic, the toxicant is probably retained by theC,, SPE column but is not covered by the methanolelution. When interpreting these results it is important .to consider that when a sample is passed over the C,,column there can be other mechanisms besides re-verse phase SPE by which a toxicant can Qe re-moved. For example, the C!, column packing mayremove toxic compounds by filtering them out of solu-tion, e.g., the toxicant may be associated with solids inthe effluent, and the 40 pm C,, packing material mayremove the solids. The toxic compqunds could also bephysically adsorbed or ionically bound onto the surfaceof the column packing and the methanol elution cannotrecover the toxicant from the column. Perhaps thetoxicant has been removed by the reversed phaseSPE mechanism but the methanol does not recover thetoxicant because either methanol is too polar a sol-vent or the toxicants have too low a Volubility in metha-nol. In this case a different solvent system may beneeded to remove the toxicants, e.g., methylene chlo-ride, hexane or pH adjusted water. Both hexane andmethylene chloride are much more toxic than metha-nol, and if hexane or any other “toxic” solvent is used,solvent exchange or some other method must be usedto remove the solvent in order to effectively track toxic-ity throughout the procedure. Another possibility thatshould be considered is that the toxicant has decom-posed. Whatever the mechanism of toxicant removal,the interpretation of the loss of toxicity should be evalu-ated carefully.

After passage through the C,, column, some efflu-ents exhibit artifactual toxicity which is not observed inthe post-column dilution water blanks. Artifactual toxic-ity can arise from two sources: a) pH drift and/or b)biological growth in the post-column effluent. (Problemswith pH drift are discussed in Section 8.3; consult thatsection for further information.) Artifactual toxicity frombiological growth in the post-column effluent can be amajor problem for some effluents, particularly municipaleffluents. This growth has not been observed in alleffluents, but for the effluents where it did occur, it waspresent in nearly every post-column sample. The post-

.

Te ble 8-3. Toxicity of methanol to several freshwater rpecfer

Species Ufestage

Cenodaphnia dub/a G h’

24 h

>3.0

LCSO f%. v/v (95% Cl))48 h 72 h 96 h

A.0

124 h’ :-:(2.6 : 2.9)

z(2.6 : 2.9)

548 h’ 2.4 2.0(2.2-2.6) (1.9-2.2)

Daphnla magna 124 ht NR 3.2(2.5-3.7)

Daphnla pulex 114 h’ 2.55 NR(2.3-2.8)

Hyaleiia azteca juvenile4

(129-5258)NR NR NR

Salt-no gairdneri

Pimephaies promeias

juvenile’

524 h’

28-32 d6

(2.:::.7)2.5 NR

(2.5-2.7) (2.giz.7)

4.0( - )3.8

;+; (3.zli.2) , @i-:.2)NR

(3.7-3.9) (3.7-3.9) (3.63-37.9)

Lepomis macrochirus juvenile 6

(2.;:;.7)2.4 NR 1.9

(2.2-2.7) (1.8-2.3)

11 Data generated at ERL-Duluth.unmeasured.)

C. dubia were 124 h old at test initiation and fed. (Tested in soft reconstituted water (DMW); static Qod

z Randall and Knopp, 1980. (Tested in spring water: static and unmeasured.)3 48-h ECSO.4 Bowman et al., 1981. (Tested in well-water: static and unmeasured.)5 18-h LCSO.6 Poirier et al., 1986. (Tested in Lake Superior water; flow-through and concentrations measured.)

Note. (-) = Confidence interval cannot be calculated as no partial mortality occurred; NR = Not reported.

column samples exhibited a turbid, often filamentousgrowth and sometimes, lower than normal DO levels inthe toxicity tests. In one effluent, this growth was causedby methylotrophic bacteria. Methanol occurs in post-column effluent samples because the methanol used inconditioning (activating) the C,, SPE column is slowlyleached out of the column and into the effluent as itpasses through the column.

Methods for eliminating or controlling this type ofartifactual toxicity problem are currently limited. Forsome effluents, the most promising method appears tobe additional filtering of the post-column effluent througha 0.2 pm filter to remove bacteria prior to testing. Wholeeffluent filtered through a 0.2 pm filter serves as acontrol for the toxicity test with the post column/filtrationmanipulation. Filtration is easy to perform and allowsuseful post-column toxicity data to be obtained, pro-vided it does not alter or reduce toxicity in the post-column effluent. If toxicity is removed in the 0.2 urnfiltered post-column effluent, but not in the 0.2 pmfiltered whole effluent, repeat the experiment filteringthe whole effluent with 0.2 pm filter and testing thepost-C,, 0.2 pm filtered effluent. The post-column (0.2pm filtered effluent) may need to be filtered (0.2 pm)again. If toxicity is removed by filtration, see Section8.4 Results/Subsequent Tests. If this growth in thepost-column cannot be eliminated but toxicity occurs in

the methanol eluate, then proceed with Phase II identi-fication. If growth is not eliminated and no toxicityoccurs in methanol eluate, then use of different sol-vents to condition the C,, column may reduce growth(e.g., acetonitrile). Control of the turbid growth mayalso be possible by performing daily renewals withpost-column effluent. Initially, more post-column samplewould have to be collected (60 mL rather than 30 mL),and a portion should be refrigerated. If control of theartifactual toxicity caused by the turbid growth cannotbe achieved, other sorbents (e.g., XADs, activated car-bon) may have to be used. Another possible method ofcontrolling the growth may be by the use of antibioticsbut we have not investigated this approach.

Observation and judgement must be used to detectproblems occurring from artifactual toxicity and onlythrough experience can one recognize when they oc-cur. Failure to recognize them will result in the conclu-sion that the C,, SPE column did not remove toxicitywhen it in fact may have done so.have done so.

If toxicity occurred in the methanol eluate from aPOTW effluent, and the C. dubia were more sensitivethan fathead minnows, it might be cost-effective to tryadding a metabolic blocker, piperonyl butoxide (PSO),to the effluent and eluate. We have frequently foundnon-polar organics in POTW effluents and have identi-fied organophosphate pesticides (OP’s) as the toxicantfs)_Q?a-Jc

to C. dubia (Amato et al., 1992; Norberg-King et al.,1991). Most metabolic blockers used in aquatic toxicol-ogy have been used with fish; however, OP’s are gen-erally less acutely toxic to most fish than to cladocer-ans. PBO is a synthetic methylenedioxyphenyl that canblock the toxicity of various chemicals that need to bemetabolized in the cytochrome P450 cycle to be toxic.In tests with cladocerans, sublethal additions of PBO tothe whole effluent and/or the methanol eluate test havebeen useful for implicating some metabolically acti-vated OP’s as the toxicant (Ankley et al., 1991).Experiments showed that for C. dubia, D. magna, andD. pulex, PBO blocked the acute toxicity of parathion,methyl parathion, diazinon, and malathion, but not di-chlorvos, chlorfenvinphos, and mevinphos. For thoseOP’s where PBO reduced the toxicity, the reductionwas greater in the first 24 h; the toxicity of the OP’s inan effluent may be expressed after 24 h.

plexes with the thiosuffate anion (Giles and Danell,1983). Cationic metals that appear to have this poten-tial for complexation (based upon their equilibrium sta-bility constants) include <cadmium (*+), copper (Cu2+),silver (Agl*), and mercury (Hg2+) (Smith and Martell,1981). However, the rate of formation of the complex isspecific for various metals and some cationic metalsmay not be rendered non-toxic in the 48-h or 96-hperiod used for the toxicity test due to a slow complex-ation rate.

To perform the PBO addition test for cladocerans,a PBO water stock is prepared and microliter quantitiesare added at various sublethal concentrations (finalconcentrations of PBO are 500, 250, and 125 f_rg/L).(Note: the 48-h LCSO’s for C. dubia, 0. magna, and D.pulex are 1,000, 2,830, and 1,620 ug/L, respectively).The PBO additions can be set up in a similar manner tothe EDTA and oxidant reduction tests (Section 8.7 and8.8) using a 3 x 3 matrix of PBO and effluent concen-trations. Toxicity reduction with the addition of PBOwould suggest the presence of toxic levels of metaboli-cally-activated compounds such as OP’s. However iftoxicity was not changed, it does not mean those typesof compounds (i.e., OP’s) will not be present. Furthertests with PBO will be described in the second editionof Phase II.

Recent work using C. dubia has shown that sodiumthiosulfate (and EDTA) can remove the toxicity of sev-eral cationic metals (Hackett and Mount, in preparation)from dilution water and effluents. The toxicity of copper,cadmium, mercury,-silver and selenium (as selenate) at4x the 24-h LC50 of each in moderately hard reconsti-tuted water was removed by the levels of thiosulfatetypically added in this test. Mercury toxicity was re-moved with the addition of thiosuffate for 24 h but not48 h, indicating it may not have been completelycomplexed by the thiosulfate. In addition, tests withzinc, manganese, lead, and nickel and thiosulfate, indi-cated that the metal toxicity was not removed by thio-sulfate. However, with these metals and the addition ofEDTA, the toxicity to C. dubia was complexrd (cf.,Section 8.8, EDTA Test). Knowing which metals arebound by both thiosulfate and EDTA, and which metalsare complexed with only one or the other additives canbe very helpful in narrowing down the possible toxicant.

8.7 Oxidant Reduction TestPrinciples/General Discussion:

This test is designed to determine to what extentconstituents reduced by the addition of sodium thiosul-fate (Na,S,O,) are responsible for effluent toxicity. Chlo-rine, a commonly used biocide and oxidant, is fre-quently found at acutely toxic concentrations in munici-pal effluents. Chlorine is unstable in aqueous solutionsand decomposition is more rapid in solutions whenchlorine is present at low concentrations. Phase I initialaeration tests will provide information on chlorine toxic-ity as will the oxidant reduction test. However, thisoxidant reduction test does not simply affect chlorinetoxicity. Also neutralized in this test are other chemicalsused in disinfection (such as ozone, and chlorine diox-ide), chemicals formed during chlorination (such asmono and dichloramines), bromine, iodine, manganousions, and some electrophile organic chemicals. Fre-quently, the reduced form of the toxicant has a muchlower toxicity.

Data on the toxicity of sodium thiosulfate toCeriodaphnia dubia, Daphnia magna and fathead min-nows are given in Table 8-4. Data generated at ERL-Dshow that for Ceriodaphnia, both feeding and lowerhardness waters results in greater thiosulfate toxicity,and this trend appears to be the same for fatheadminnows (Table 8-4). In effluents, some of the addedthiosulfate will combine with certain oxidants present,thereby lowering the concentration of the reactive andtoxic thiosulfate. Therefore, the LCSO values indicatethat less toxicity due to thiosulfate (Table 8-4) might beexpected in effluents than in dilution water (i.e., recon-stituted water) where no oxidants are present to reactwith the thiosulfate. More importantly, when an effluentconcentration of 4x the LC50 is tested, toxic oxidantlevels should not be excessively high. As a result thereshould not be a need to add very large amounts ofthiosulfate to neutralize toxic oxidants in the test solu-tion.

Additions of sodium thiosulfate for this test can beapproached in either of two ways; a gradient of thiosul-fate concentration can be added to several test cham-bers containing the same effluent concentration or as adilution test where a 3 by 3 matrix of effluent concentra-tions and thiosulfate concentrations are used.

Although the thiosulfate addition test was initially For the gradient approach, concentrations of so-designed to determine if oxidants (such as chlorine) are dium thiosulfate equal to and lower than the thiosulfateresponsible for effluent toxicity, thiosulfate can also be LC50 for the test species being used are added toa chelating agent for some cationic metals. Conse- several containers with effluent at the 4x-LC50 (or 100%)quently, reductions in effluent toxicity observed with concentration (cf., Figure 8-l 5). If the test species isthis test may be due to the formation of metal com- not listed in Table 8-4, the thiosulfate LC50 will have to

8-33

be determined. Time to mortality may also be useful inaddition to observing mortality at a fixed time (i.e., 24,48 or 96 h). Time to mortality measurements are impor-tant when no dilutions of the effluent are used.

The dilution approach has the advantage in thatLCSOs can be calculated to see how much the toxicitywas reduced. For this test a matrix of three effluentconcentrations and three levels of thiosulfate concen-trations are used. The choice of the thiosulfate concen-trations to add to the effluent is based on the thiosulfateLC50 for the test species being used in an appropriatedilution water (Table 8-4). Three sets of effluent solu-tions (i.e., 4x-LC50, 2x-LC50, 1 x-LCSO or 1 OO%, 50%,25%) are prepared. To the first set, thiosulfate is addedto each test solution at one-half the thiosulfate LC50; tothe second set, thiosulfate is added at one-fourth (0.25x)

the LC50; and to the third set, thiosulfate is added at0.125x the LC50. In this approach the concentration ofthiosulfate remains constant over each effluent dilutionseries. The test results are compared to the baselinetest result to determine the amount of toxicity removal.

For cases where oxidants account for only part ofthe toxicity, sodium thiosulfate may only reduce, noteliminate the toxicity. The thiosulfate addition test isuseful even when chlorine appears to be absent in theeffluent. As discussed above, oxidants other than chlo-rine occur in effluents and this test should not beomitted just because the effluent is not chlorinated.Likewise, removal of toxicity by thiosulfate does notprove that chlorine was the cause of effluent sampletoxicity. Refer to the Results/Subsequent Tests sectionbelow for additional options.

Table 8-4. Toxicity of sodium thiosulfate to Ceriodaphnia dubia, Daphnla magna, and fathead minnows

Water Life LC50 (u/l I (95% Cl\.

Species Type Stage 24 h 48 h 72 h 96 h

Ceriodaphnia dub& SRW 524 h

SRW 124 h

SRW 124 h

SRW ~24 h

MHRW 124 h

HRW 124 h

VHRW ~24 h

VHRW 524 h

VHRW 524 h

D a p h n i a magn$

P i m e p h a l e s promelasl

SRW

SRW3

SRW’

MHRWJ

MHRW’

HRW3

HRW’

VHRWJ

VHRW’

NR

524 h

124 h

124 h

524 h

124 h

524 h

124 h

524 h

. 2.5

(12(1.0-l .7)

1.5(1.2-2.0)

2.0(152.7)

(1 Z2)6.6

(5.8-7.5)5.0

(3.8-6.5)6.6

(5.8-7.6)5.0

(3.9-6.4)

7 . 4( 6 . 3 - 8 . 9 )

;+j

I-_)11.6

(9.7-l 3.9)12.9

(11.8-14.1)13.3

(11.8-15.7)13.4

(11.8-15.3)

0 . 8 5( 0 . 7 2 - l .O)

0 . 8 8( 0 . 7 2 - l .l)

0 . 9 5( 0 . 8 3 - l .l)

(0.: -80499,0 . 9 8

( 0 . 6 2 - l .6)1 . 6

z(2.5i4.3)

1 . 8(1.4-2.3)

(0.;;: .6)

8 . 4( 7 . 6 - 9 . 3 )

(5.E.4)

(8.ti.7,

(“-“,

(7.ElLO)12.9

(11.8-14.1)

(6.:::.2)12.1

(10.8-l 3.4)

(7.Z.5)7.1

(5.9-8.4)

$2)9.8

0(6.9-9.1)

12.5(11.3-13.8)

6.7(5.5-8.2)

12.1(10.8-13.4)

(6.ki.3,

(4Z.9)

(3.ii.6)

Pi

(6.ip9.0)11.7

(10.4-13.0)

(4.Z.9)12.1

(10.8-13.4)

11 Data generated at ERL-Duluth; both species were 524 h old at test initiation and C. dubia were fed.2 Dowden and Bennett, 1965.3 Data generated at ERL-Duluth and values represent those from 7-d growth and survival tests and daily renewals.

Note: (-) = Confidence interval cannot be calculated as no partial mortality occurred; NR = Not reported; VSRW = very soft reconstitutedwater; SRW = soft reconstituted water; MHRW = moderately hard reconstituted water; HRW = hard reconstituted water; VHRW = veryhard reconstituted water.

8-34

Volume Required:A maximum volume of 100 mL effluent is required

for the oxidant reduction test. The exact volume re-quired will depend on the 24-h initial LC50. For thegradient addition, six effluent aliquots at 4x-LC50 or(100%) are required, each having different thiosulfateconcentrations. For the dilution test, three sets of threeeffluent solutions are prepared (Le., 4x-, 2x-, lx- 24-hLC50) and three different concentrations of sodiumthiosulfate (e.g., 0.25, 0.5, and 1 .O g/L) are each addedto a set of dilutions.

Apparatus:Glass stirring rods, glass volumetric flask for so-

dium thiosulfate, 1 mL glass pipettes, automatic pi-pette, disposable pipette tips, 10 - 1000 PL pipettes,eye dropper or wide bore pipette, 30 mL beakers or 1oz plastic cups, light box and/or microscope (optional),pH meter and probe.

Reagents:Regardless of whether the 1 x 6 gradient addition

test or the dilution test is to be done, the sodiumthiosulfate stock concentration should be 10x the so-dium thiosulfate LC50 concentration for the test spe-cies being used.

Test Organisms:Test organisms, 50 or more, of the same age and

species.

Procedure:Day 2: To perform the gradient thiosuifate addition

test, transfer six 10 mL aliquots of effluent diluted to 4x-LC50 (or 100%) into six test chambers. Add 1 .O, 0.8,0.6, 0.4, and 0.2 mL of the appropriate concentration ofthe thiosulfate stock to five aliquots and mix. Do notadd any to the sixth. The container receiving 1 mL ofthiosulfate should now contain the approximate con-centration of sodium thiosulfate equal to the LC50 ofthe test species. Figure 8-15 contains an example formfor recording the data. A suggested schedule for ob-serving time to mortality is shown on the data form.

To perform the thiosulfate dilution addition test,prepare three sets of effluent dilutions (i.e., 4x-LC50,2x-LC50, lx-LCSO) and add the appropriate amount ofthiosulfate (i.e., 0.5x, 0.25x, and 0.125x thjosulfate LC50)for the test species to each set of dilutions. Figure 8-16is an example data sheet for the thiosulfate additiontest using this effluent dilution approach. The baselinetest conducted at the same time will provide informa-tion on effluent toxicity without thiosulfate added.

interferences/Controls and Blanks:Controls prepared for the baseline toxicity test act

as a check on the general health of test organisms,dilution water quality and test conditions.

When the time to mortality in the various thiosulfateexposure concentrations in the gradient addition test iscompared to ,$he treatment without thiosulfate, one can

determine whether the addition of thiosutfate increasedthe time to mortality at some thiosulfate concentration.If, in all of the effluent exposures, the time to mortalitydecreases, then thiosulfate is affecting toxicity. If alltest solutions cause mortality in the thiosuifate effluentdilution test, but this trend does not occur in the baselinetest, the thiosulfate may be causing the toxicity. Ineither case, the test should be repeated with weakersodium thiosutfate additions. If the toxicity is unchanged,perhaps not enough sodium thiosutfate was added, andthe test can be repeated using a higher range of thethiosulfate additions.

If a significant loss in effluent toxicity is apparentover the first 24-h period after sample arrival in thelaboratory (i.e., initial LC50 c baseline LC50), it may benecessary to conduct future oxidant reduction tests forPhase I immediately upon arrival of the sample in thelaboratory.

Results/Subsequent Tests:If oxidants are causing toxicity, time to mortality

should increase somewhere in the range of testedthiosulfate additions or the toxicity should be reducedfrom the baseline LC50. No change in toxicity suggestseither no oxidant toxicity or not enough thiosulfate wasadded. The experiment should be repeated, increasingthe concentration of thiosulfate added.

When the LCSOs from the sodium thiosulfate addi-tion dilution test indicate toxicity was reduced whencompared to the baseline LC50, thiosulfate has eitherreduced or complexed the toxicant( If the highestaddition of thiosulfate increases the toxicity of thesample, the thiosulfate itself may be at a toxic concen-tration. However, if the LC50 for the next lower additioflof thiosulfate of the effluent dilutions reduced and/orremoved toxicity, then more tests for oxidants or metalsshould be explored.

If oxidant toxicity is evident, a measurement of freechlorine should be made and the concentration com-pared to the chlorine toxicity value for the test speciesused. For identification it may be necessary to measuremono and dichloramine since they have differenttoxicities than free chlorine (see Phase Iii for confirm-ing mixtures as toxicants). A comparison of the aera-tion and C,, SPE test results to the oxidant reductiontest results may provide even more information on thephysical/chemical nature of the oxidants.

For those effluents where chlorine is measurable,dechlorination may be achieved by the use of sulfurdioxide (SO,) gas. This technique used was developedby T. Wailer (personal communication, University ofNorth Texas, Denton, TX). (Note: Caution in handlingthe SO, should be exercised because it is an extremeirritant.) As with thiosuifate, SO may also reduce com-pounds other than chlorine. This information can beuseful when one needs to know if substances otherthan chlorine are causing the toxicity.

To dechlorinate using SO,, the following procedureis used. Place 10 mL of high quality distilled water into

8-35

Flgurm 6-15. Exrmplo data shoat for the oxldrnt toduction toet when urlng I gradlant of sodium thloaulfatoconcantrrtlona.

Test Type: Oxidant ReductionTest Initiation (Date & Time):

Investigator:Sample Log No., Name:Date of Collection:

Species/Age:No. Animals/No. Reps:Source of Animals:Dilution Water/Control:Test Volume:Other Info:4x-LC50: Of 100%TRC:

mL Stock II II SuMal Readlngs:

72 h 96 hApH DO A pH DO

l *

Note: See baseline data sheet for control data.

Stock Concentration - o/L Na2S203

Cgmments:

6-36

a graduated cylinder. Bubble the SO, gas directly intothe water for about 5 min to prepare SO,-saturatedwater. (Caution: This saturation procedure must bedone in a hood!) For a first attempt at the amount ofSO, to add without the TRC measured we use 2 PL ofthe SO,-saturated water per 100 mL of effluent withTRC values of O-5 mg/L). This amount of SO, is notacutely toxic and is effective at removing most com-monly encountered TRC concentrations. For measuredchlorine concentrations, proportional amounts of SO,-saturated water have been used as follows: for 0.02mglL TRC add 3.6 PL SO,-saturated water to 1 L ofsample; for 0.16 mg/L TRC, add 12 @_/L SO,-saturatedwater; for 1.3 mg/L TRC add 39 @L SO, saturatedwater: and for 2.1 mg/L TRC add 64 PUL SO -satu-rated water (T. Wailer, personal communication .f An-other technique to remove the chlorine is being ex-plored at present. The use of sodium bisulfite solutionsadded in the same way as sodium thiosulfate solutionsare added is being explored.

In cases where both the oxidant reduction test andEDTA chelation test reduce the toxicity in the effluentsample, there is a strong possibility that the toxicantmay be a cationic metal(s). For example, thiosutfateand EDTA both reduce the toxicity of copper, cadmium,and mercury, At this point, the Phase II methods foridentification for cationic metal(s) toxicants should beinvestigated.

8.8 EDTA Chelation TestPrinciples/General Dlscussion:

To determine the extent to which effluent toxicity iscaused by certain cationic metals, increasing amountsof a chelating agent (EDTA; ethylenediaminetetraacetateligand) are added to aliquots of the effluent sample.The form of the metal (e.g., the aquo ion, insolublecomplex) has a major effect on its toxicity to aquaticorganisms (Magnuson et al., 1979) and specific metalforms are more important in aquatic toxicity than thetotal quantity of the metal.

EDTA is a strong chelating agent, and its additionto water solutions produces relatively non-toxic com-plexes with many metals. The success of EDTA inremoving metal toxicity is a function of solution pH, thetype and speciation of the metal, other ligands in thesolution, and the binding affinity of EDTA for the metalversus the affinity of the metal for the tissues of theorganism (Stumm and Morgan, 1981). Because of itscomplexing strength, EDTA-metal complexes will oftendisplace other soluble forms such as chlorides andoxides of many metals. Among the cations typicallychelated by EDTA are aluminum, barium, cadmium,cobalt, copper, iron, lead, manganese (2+), nickel, stron-tium, and zinc (Stumm and Morgan, 1981). EDTA willnot complex anionic forms of metals such as selenides,chromates and hydrochromates, and forms relativelyweak chelates with arsenic and mercury. For thosemetals with which it forms relatively strong complexes,the toxicity of the metal to aquatic organisms is fre-quently reduced: EDTA has been shown to chelate the

toxicity to C. dubia due to copper, cadmium, zinc,manganese, lead; and nickel (Hackett and Mount, InPreparation) in both dilution water and effluents. HOW-ever, it was also found that EDTA did not complex thetoxicity of silver, selenium (either as sodium selenite orsodium selenate), aluminum (AI(OH chromium (ei-ther as chromium chloride or potassium dichromate), orarsenic (either sodium m-arsenite or sodium arsenate)when tested using moderately hard water and C. dubia.

Since EDTA chelates calcium and magnesium (al-beit weakly) the choice of the level of EDTA to add wasoriginally (EPA, 1988A) based on the premise thatcalcium and magnesium had to be chelated beforetoxic metals would be. However, recent work has shownthat the toxicity due to cationic metals was reducedregardless of water hardness. Therefore the mass ofchelating agent required should be approximated be-cause excess EDTA becomes toxic when present abovea certain concentration. The range of EDTA concentra-tions that will adequately bind the metals but is nottoxic appears to be smaller than that for sodium thiosul-fate and oxidants.

_

Table 8-5 contains LCSOs of EDTA for Ceriodaphniaand fathead minnows at various hardness and ainityvalues. Note that the concentration of EDTA toleratedby organisms increases directly with both water hard-ness and salinity. By measuring the hardness and sa-linity of the effluent, the range of EDTA concentrationsthat should not be toxic in an effluent sample can beestimated. “Salinity” not due strictly to NaCl will havedifferent effects on toxicity. This calculation, for predic-tion of the EDTA concentration, is more involved thanis at first apparent. The data in Table 8-5 indicate thatover the physiological range of hardness and salinity,hardness affects the toxicity of EDTA more than NaCI.The usual methods for measurement of salinity (con-ductivity meter, salinometer or refractometer) do notspecifically measure sodium chloride. The choice ofEDTA concentrations should always be based first onhardness and secondly on salinity when the salinity isknown. The particular combination of hardness andsalinity present in an effluent sample may have to betested to get an accurate EDTA LC50. If the salinity iscomposed of ions other than sodium and chloride, thehardness of the dilution water should be made equal tothe effluent hardness and the additional “salinity” addedin the form of other major cations and anions such aspotassium, sulfate and carbonate.

An EDTA LC50 value derived in a standard dilutiontest water (such as reconstituted water) is likely to bemuch lower than the LC50 of EDTA added to an efflu-ent. For example, the values contained in Table 8-5represent worst case conditions presented by EDTA indilution water. Likewise, the toxic concentration of EDTAin one effluent will probably not be the same as theconcentration causing toxicity in a different effluent oreven a different sample of the same effluent. For thisreason the concentrations of EDTA added to the efflu-ent should bracket the expected LC50 based on cleanwater with a similar hardness and salinity value as per

a-38

a graduated cyiinder. Bubble the SO, gas directly intothe water for about 5 min to prepare SO,-saturatedwater. (Caution: This saturation procedure must bedone in a hood!) For a first attempt at the amount ofSO, to add without the TRC measured we use 2 PL ofthe SO,-saturated water per 100 mL of effluent withTRC values of O-5 mg/L). This amount of SO, is notacutely toxic and is effective at removing most com-monly encountered TRC concentrations. For measuredchlorine concentrations, proportional amounts of SO,-saturated water have been used as follows: for 0.02mglL TRC add 3.6 PL SO,-saturated water to 1 L ofsample; for 0.16 mg/L TRC, add 12 plfL SO,-saturatedwater; for 1.3 mg/L TRC add 39 uL/L SO, saturatedwater; and for 2.1 mg/L TRC add 64 pUL SO -satu-rated water (T. Wailer, personal communication .f An-other technique to remove the chlorine is being ex-plored at present. The use of sodium bisulfite solutionsadded in the same way as sodium thiosulfate solutionsare added is being explored.

In cases where both the oxidant reduction test andEDTA chelation test reduce the toxicity in the effluentsample, there is a strong possibility that the toxicantmay be a cationic metal(s). For example, thiosulfateand EDTA both reduce the toxicity of copper, cadmium,and mercury. At this point, the Phase II methods foridentification for cationic metal(s) toxicants should beinvestigated.

8.8 EDTA Chelation TestPrlnclplesfGeneral Discussion:

To determine the extent to which effluent toxicity iscaused by certain cationic metals, increasing amountsof a chelating agent (EDTA; ethylenediaminetetraacetateligand) are added to aliquots of the effluent sample.The form of the metal (e.g., the aquo ion, insolublecomplex) has a major effect on its toxicity to aquaticorganisms (Magnuson et al., 1979) and specific metalforms are more important in aquatic toxicity than thetotal quantity of the metal.

EDTA is a strong chelating agent, and its additionto water solutions produces relatively non-toxic com-plexes with many metals. The success of EDTA inremoving metal toxicity is a function of solution pH, thetype and speciation of the metal, other ligands in thesolution, and the binding affinity of EDTA for the metalversus the affinity of the metal for the tissues of theorganism (Stumm and Morgan, 1981). Because of itscomplexing strength, EDTA-metal complexes will oftendisplace other soluble forms such as chlorides andoxides of many metals. Among the cations typicallychelated by EDTA are aluminum, barium, cadmium,cobalt, copper, iron, lead, manganese (**), nickel, stron-tium, and zinc (Stumm and Morgan, 1981). EDTA willnot complex anionic forms of metals such as selenides,chromates and hydrochromates, and forms relativelyweak chelates with arsenic and mercury. For thosemetals with which it forms relatively strong complexes,the toxicity of the metal to aquatic organisms is fre-quently reduced, EDTA has been shown to chelate the

toxicity to C. dub& due to copper, cadmium, zinc,manganese, lead; and nickel (Hackett and Mount, InPreparation) in both dilution water and effluents. HOW-ever, it was also found that EDTA did not complex thetoxicity of silver, selenium (either as sodium selenite orsodium selenate), aluminum (AI(OH) chromium (ei-ther as chromium chloride or potassium dichromate), orarsenic (either sodium m-arsenite or sodium arsenate)when tested using moderately hard water and C. dubia.

Since EDTA chelates calcium and magnesium (al-beit weakly) the choice of the level of EDTA to add wasoriginally (EPA, 1988A) based on the premise thatcalcium and magnesium had to be chelated beforetoxic metals would be. However, recent work has shownthat the toxicity due to cationic metals was reducedregardless of water hardness. Therefore the mass ofchelating agent required should be approximated be-cause excess EDTA becomes toxic when present abovea certain concentration. The range of EDTA concentra-tions that will adequately bind the metals but is nottoxic appears to be smaller than that for sodium thiosul-fate and oxidants.

.

Table 8-5 contains LC5Os of EDTA for Ceriodaphniaand fathead minnows at various hardness and s&n&yvalues. Note that the concentration of EDTA toleratedby organisms increases directly with both water hard-ness and salinity. By measuring the hardness and sa-linity of the effluent, the range of EDTA concentrationsthat should not be toxic in an effluent sample can beestimated. “Salinity” not due strictly to NaCl will havedifferent effects on toxicity. This calculation, for predic-tion of the EDTA concentration, is more involved thanis at first apparent. The data in Table 8-5 indicate thatover the physiological range of hardness and salinity,hardness affects the toxicity of EDTA more than NaCI.The usual methods for measurement of salinity (con-ductivity meter, salinometer or refractometer) do notspecifically measure sodium chloride. The choice ofEDTA concentrations should always be based first onhardness and secondly on salinity when the salinity isknown. The particular combination of hardness andsalinity present in an effluent sample may have to betested to get an accurate EDTA LC50. If the salinity iscomposed of ions other than sodium and chloride, thehardness of the dilution water should be made equal tothe effluent hardness and the additional “salinity” addedin the form of other major cations and anions such aspotassium, sulfate and carbonate.

An EDTA LC50 value derived in a standard dilutiontest water (such as reconstituted water) is likely to bemuch lower than the LC50 of EDTA added to an efflu-ent. For example, the values contained in Table 8-5represent worst case conditions presented by EDTA indilution water. Likewise, the toxic concentration of EDTAin one effluent will probably not be the same as theconcentration causing toxicity in a different effluent oreven a different sample of the same effluent. For thisreason the concentrations of EDTA added to the efflu-ent should bracket the expected LC50 based on cleanwater with a similar hardness and salinity value as per

a-38

Table 8-5. EDTA t o Ceriodaphnlo and fathead minnow@ a n d sailnitlee

Hardness

SpeciesWaterType

a sCaCO,)

I Salinity’(PPt) 24 h

LC50 tall) (95%

dubraVSRW lo-13

SRW 40-48

80-100

HRW 160-180

VHRW

0

00

00

0

0

0.5

1

2

3

0

0

0

0

0

0.5

1

2

3

(O.i&)(O.i&)0.12

(0.104.13)0.23

(0.21-0.27)

(0.402SO)0.71

(0.58-0.87)0.05(-)0.12

(0.10-0.13)0.33

(0.27-0.41)0.44( - )

(O.C!&4)0.14

(0.12-0.18)

(O.E35)

(0.4Z6)0.81

(068-0.97)

0.03(-_)0.11( - )0.22

AZ(-)0.41

(0.36-0.47)0.05(-)0.11(-)0.23

(0.21-0.27)0.32

(0.23-0.45)

0.03(0.03-0.04) (O&&4)

0.14 0.11(0.12-0.18)(0.08-0.14)

0.27 0.27(0.22-0.33) (0.22-0.33)

(O.E62) (O.&O)0.81 0.81

(068-0.97) (0.68-0.97)

(O.cz?%4)(0.&%9,

0.25(0.20-0.31)

(O.Zk)0.81

(0.52-0.83)0.11(-)0.17

(0.13-0.21)0.23

(0.17-0.32)0.37

(0.28-0.48)

l Brine from evaporated seawater used as source of salinity. All data generated at EPA ERL-Duluth. All C. dubia were 524 h old and thefathead minnows were all 136 h old at test initiation. Ceriockphnia were fed; see section on toxicity tests for details.

Note: (-) = Confidence interval cannot be calculated as no partial mortality occurred: NR = Not reported; VSRW = very soft reconstitutedwater; SRW = soft reconstituted water; MHRW = moderately hard reconstituted water; HRW = hard reconstituted water; VHRW = veryhard reconstituted water.

the above discussion. The complexation of metals withEDTA may not be immediate after the addition of EDTA.Therefore, it is recommended that the EDTA test solu-tions be set up first and these solutions allowed to sitfor the duration of pH adjustments and other manipula-tions before the introduction of test organisms. This isat least 2 h.

As with the oxidant reduction test, the EDTA canbe added in two ways; a gradient of EDTA can beadded to replicate of one effluent concentration or threeconcentrations of EDTA can be added to three sets ofeffluent dilutions. The effluent itself is used as a controlrather than a blank based on dilution water as in the pHadjustment test. The gradient addition test is done byadding increasing concentrations of EDTA to severalaliquots of the effluent (4x-LC50 or 100%). The goal ofthis test is to add enough EDTA to reduce metal toxic-ity. At some EDTA addition the metals will be chelated

and the EDTA will not be present at toxic concentra-tions. At lower EDTA additions the metal toxicity is notremoved; in the midrange of the EDTA additions themetals will be rendered non-toxic by the EDTA, and atthe high end of the range of EDTA additions, theunreacted EDTA is itself toxic. By using an effluentconcentration of 4x-LC50 (or 100% if the LC50 is greaterthan 25%) the potential for exceeding the binding capa-bility of the added EDTA is lessened, especially forvery toxic effluents (LC50 ~10%).

To conduct the EDTA test using effluent dilutions,three addition levels of EDTA (using one stock) areselected (based on the LC50 of EDTA for the speciesof choice). Each of these three EDTA levels are thenadded to effluent dilution tests in a 3 x 3 matrix. TheEDTA is added to the 4x-, 2x-, and lx-LCSO test cupsafter the effluent solutions are prepared so that thethree EDTA concentrations are constant across each

a-39

set of effluent dilutions. For example, 0.2 mL of anEOTA s o l u t i o n i s a d d e d t o t e s t c u p s c o n t a i n i n g

4xLC50, 2x-LC50, lx-l_CSO of 1 OO%, 50%,and 25%. To the next set of test cups of same effluentdilution sequence, 0.05 mL of the EDTA stock solutionis added and likewise 0.0125 mL is added to the thirdset of test cups.

To determine the amount of EDTA to add, one canuse the hardness titration, the measurement of calciumand magnesium concentrations, or the concentration ofEDTA at the EDTA LC50 for the species of interest.These are described in the Procedure Procedure below.

Volume Required:A volume of 100 mL effluent usually is required for

the EDTA chelation test. The exact volume needed willdepend on the 24-h initial LC50 and the particularoption chosen to determine the EDTA addition.

Apparatus:Glass stirring rods, glass volumetric for EDTA stock

solution, automatic pipette, disposable pipette tips, 10,100, and 1000 uL pipettes, eye dropper or wide borepipette, 30 ml beakers or 1 of plastic cups, light boxand/or microscope (optional).

Reagents:EDTA (disodium salt, Na, EDTA) stock solution

(see discussion under Procedure), Procedure), reagents for deter-mination of effluent hardness and salinity (APHA, 1980;Methods 314 314 and 210).

Test Organisms:

Test organisms, 50 or more, of the same age andspecies.

Procedure:

Day 2: There are three ways to determine theconcentration of EDTA stock to prepare.

The first and the most accurate approach (when itcan be used) is to measure the hardness of the 4x-24-hLC50 effluent concentration (or 100% when the LCSO is>25%) using the standard method for measuring hard-ness (APHA, 1980). The concentration of EDTA thatproduced the endpoint in the hardness titration of theeffluent sample is the concentration of EDTA needed atthe 0.2 mL addition for either the EDTA gradient test orthe effluent dilution test. An example illustrates thiscalculation. In a 36% effluent sample (4x-LC50), 5 mLof 0.01 M EDTA was needed to titrate the hardness(100 ml sample size). For the gradient test, 7 EDTAconcentrations will be added to several test chambers,all containing one concentration of effluent (4x-LC50).The concentration of EDTA required for the hardnesstitration is the highest additive concentration. For a 10mL test volume of 36% effluent, when 0.5 mL of 0.01 MEDTA stock was added the resultant EDTA concentra-tion is that which is desired at the 0.2 mL addition. Toprovide this EDTA concentration at the 0.2 mL addition(minimizing the volume addition), increase the 0.01 M

EDTA concentration (concentration of EDTA used intitration) by 0.510.2 or 2.5x = 0.025 M EDTA stock.(Note: Molecular weight (MW) of Na,EDTA is 372.3 g.)

The second approach is used when the hardnessmeasurement endpoint cannot be discerned becauseof interferences. If the hardness cannot be titrated,measure the calcium (Cal*) and magnesium (Mg2+) ofthe sample using atomic absorption procedures, andcalculate the amount of EDTA needed to chelate thecalcium and magnesium. EDTA binds with both Cal+and Mg2+ on a 1:l molar basis. The combined numberof moles of Ca2* (MW=40.1 g) and Mg2* (MW=24.3 g) in10 mL of effluent at 4x-LC50 equals the number ofmoles of EDTA needed for the 0.2 mL addition for a10 mL sample. This calculated concentration should beadded at the 0.2 mL addition for either the gradient ordilution test. The calcium and magnesium should bemeasured at 100% effluent if the LC50 is greater than25%.

The third approach, and the one we use mostfrequently, is to use the EDTA LC50 concentration toselect addition levels. This approach allows for either agradient of EDTA additions to 100% (4x-LC50) solu-tions or EDTA additions to effluent dilution tests. Choiceof the EDTA LC50 must be based on effluent hardiness(and salinity). It may be necessary to determine theEDTA LCSO for the particular combination of effluenthardness and salinity and test organism used.

After the concentration for the stock solution ofEDTA has been determined, the EDTA gradient testcan be set up. The EDTA LC50 is generally set at the0.2 mL addition. To perform the gradient EDTA addi-tion test, 7 aliquots of the effluent are prepared at aconcentration equal to 4x-LCSO, or 100% effluent wherethe initial 24-h LC50 is greater than 25%. Next, 0.4 mLof the appropriate EDTA stock is added to the first 10mL aliquot of the effluent, 0.2 mL is added to thesecond 10 mL sample of effluent, 0.1 mL to the third,and so on until the sixth 10 mL effluent sample hasreceived 0.0125 mL. The seventh is an effluent blankused to compare treatment effects on time to mortality(see Figure S-17). A microliter syringe will be neededfor the smaller additions. If the effluent has a lowtoxicity (LCXO = 50-100%) a series of dilution blanksmay be necessary to check for the dilution effect of theEDTA stock addition. No more than 10% dilution of theeffluent aliquots should be allowed unless a dilutionblank series is included.

To perform the EDTA additions using the effluentdilution test, three sets of three effluent concentrationsare prepared (1 OO%, 50%, 25%, or 4x-, 2x-, 1 x-LCSO),while the baseline test serves as the toxicity blank.After the concentration of stock solution of EDTA hasbeen established, add the EDTA using a 3 x 3 matrix.This means that the 0.2 mL addition is added to theloo%, 50%, 25%, 0.05 ml is added to another set, and0.0125 mL is added to the third effluent set (see Figure8-18). The EDTA is added after all the dilutions areprepared. To allow the EDTA time to complex the

8-40

-

Figure 8-17. Example data sheet for EDTA chelation test when uslng a gradlent of EDTA concentrations.

Test Type: EDTA ChelationTest Initiation (Date & Time):

Investigator:Sample Log No., Name:Date of Collection:

Species/Age:No. Animals/No. Reps:Source of Animals:Dilution Water/Control:Test Volume:Other info:4x-LC50: or 100%

.

0.40.2

0.1

0.05

0.0250.0125

0.0

b

Note: See baseline data sheet for

Stock Concentration - !J/L

Comments:

control data.

EDTA

8-418-41

8-43

metals, these samples should be prepared first. Thetest solutions should not have test organisms addeduntil all other manipulations are performed.

The complexation of metals by EDTA proceeds ata rate which may vary according to the sample matrix.In studies of some aqueous samples containing metaltoxicity, better success in demonstrating chelation ofmetals by EDTA may occur if samples spiked withEDTA remain refrigerated overnight. The next day theyare warmed to the test temperature and the pH isadjusted to pH i before placing test organisms in thechambers. This allows time for the EDTA to chelateany metals which may be in the sample. The solutionsshould be mixed thoroughly after spiking with EDTAand before adding the test organisms.

For both the gradient EDTA addition and EDTAdilution tests, the pH of the effluent after addition ofEDTA should be checked. Since EDTA is an acid,additions of this reagent will lower the pH of the efflu-ent. The amount of change in solution pH will dependupon the buffering capacity of the effluent and theamount of reagent added. If the pH of the effluent haschanged, readjustment of the test solution pH to pH ishould be performed. When stable pHs are obtainedand all other manipulations have been completed, testorganisms are added to the test chambers.

Interferences/Controls and Blanks:Controls prepared for the baseline toxicity test pro-

vide quality control for test organisms, dilution waterand test conditions. Either the zero mL EDTA additionin the gradient test or the baseline effluent test servesas a blank for use in determining the presence of EDTAtoxicity.

For the EDTA gradient test, time to mortality maybe recorded at each EDTA addition and then comparedto the untreated effluent. If time to mortality is shorter inall treatments than in the untreated effluent, repeat thetest using lower EDTA concentrations. If the baselinetest has less toxicity than the EDTA additions in thedilution test, then the EDTA may be causing toxicity. Iftime to mortality or toxicity is not reduced in any treat-ment, it may be wise to repeat the additions using ahigher range of EDTA concentrations. Erratic patternsin mortality cannot be used, and when this occurs itsuggests that this test is not appropriate for the particu-lar effluent being studied.

The addition of EDTA to the sample often lowersthe pH of the test solution to as low as pH 4.0. If thepresence of a pH-dependent toxicant such as ammoniais suspected, then the results of this test must beinterpreted cautiously before attributing losses of toxic-ity to chelation of metals. For instance, a sample whichcontains ammonia toxicity and an undetermined amountof metal toxicity (perhaps none) may show a loss oftoxicity at some EDTA concentrations. A closer lookmay reveal that pH drifted in these samples to 7.5 orlower, rendering ammonia non-toxic in the sample, and

that metal chelation may have had no role in reducingtoxicity. In the same way, the presence of compoundswhose toxicity is exacerbated at low pH (e.g., hydrogensulfide) may confound interpretations of this test. Insuch a sample, metal toxicity may indeed be reducedby chelation at some EDTA additions: however, theincreased toxicity of hydrogen sulfide at the lower pHcould mask the metal toxicity loss. In cases such asthese, we have returned pH to the initial value, andsuccessfully employed a simple method of pH control(e.g., closing the test vessel) to avoid misleading pH-dependent toxicant interferences. This strategy mayneed to be attempted to ferret out interferences andobtain useful information from the EDTA addition test.

In certain efflueNs, EDTA may reduce the toxicityof cationic surfactants. This reduction may appear as adelay in time to mortality. If the EDTA test result is notlikely caused by cationic metals, other Phase I proce-dures, such as sublation during the aeration step, mayalso indicate surfactant toxicity (cf., Section 8.3 aera-tion test).

Results/Subsequent Tests:For the EDTA gradient test, if the appropriate EDTA

concentration range is utilized, the time to mortaTtty willnot change from that seen in the exposure 4x-24-hLCSO of unaltered effluent at low additions of EDTA. Inthe 0.2 mL addition, toxicity should be reduced and athigher additions of EDTA, toxicity will be as high orhigher than the whole effluent itself due to unboundEDTA toxicity and effluent toxicants other than chelat-able metals if present. Time to mortality must be usedto detect partial toxicity removal. Toxicity may be re-moved at all exposures if the lowest addition of EDTAremoves metal toxicity and the highest addition doesnot cause E.DTA toxicity. If toxicity is not reduced in anytreatment, either the effluent has no chelatable metaltoxicity or not enough EDTA was added. Increasedtoxicity over the toxicity of untreated effluent suggestsEDTA toxicity and a lower EDTA range should betested.

For the EDTA dilution test, if the effluent is lesstoxic (i.e., LC50 is greater than baseline LC50) in anyof the three EDTA addition dilution tests, then theindication is that EDTA removed or reduced the toxicityand therefore metal toxicity is present. If in all threetests the effluent is more toxic (i.e., treatment LCSOsare lower than baseline LC50), then the possibilityexists that EDTA itself is causing toxicity and the testshould be repeated using lower EDTA addition concen-trations. If no LC50 of any of the three additions indi-cates less toxicity than in the baseline test, the possibil-ity of the presence of cationic metals causing toxicity inthe effluent is low, but additions of EDTA at higherlevels may need to be explored.

If toxicity is reduced in a systematic manner, pro-ceed to Phase II methods for specific identification ofthe metal(s).

.

8.9 Graduated pH TestPrinciples/General Dlscussion:

This test is designed to determine whether effluenttoxicity can be attributed to compounds whose toxicityis pH dependent. The pH dependent compounds ofconcern are those with a pK, that allow sufficient differ-ences in dissociation to occur in a physiologically toler-able pH range (pH 6-9). Also, the two forms of thecompound (ionized versus un-ionized) must have de-tectable toxicity differences to the TIE organism. Theionizable compounds commonly found in municipal andindustrial discharges include ammonia, hydrogen sul-fide, cyanide, and some organic compounds (e.g., pen-tachlorophenol). In addition, pH differences can affectmetal toxicity through changes in solubility and specia-tion. The effect of pH on ammonia toxicity might bemore readily observed than the effect of pH on thelevels of toxicity of metals, hydrogen sulfide, cyanide,and ionizable ofganics.

concentration of ammonia occurring in the toxic formwith increasing pH is greater than the decrease in itstoxicity. The net result is an increased toxicity of agiven total ammonia concentration with increased pH.Temperature also affects the dissociation of ammonia,but since the temperature is held constant in thesetoxicity tests for Phase I, it can be ignored.

Ammonia is frequently present in effluents at con-centrations of 5 mg/L to 40 mg/L (and higher). (Theammonia is measured upon arrival of the sample (Sec-tion 6) and this information will be helpful for the gradu-ated pH test.) Levels of 5 mg/L to 40 mg/L are likely tocause toxicity when several other effluent conditionsoccur. Effluent parameters to consider are pH, tem-perature, DO, CO, content, and TDS. Of these param-eters, pH has the largest effect on ammonia toxicity,and for many effluents (especially with POTW efflu-ents) the pH of a sample rises upon contact with air.Typically, the pH at air equilibrium ranges from 8.0 to8.5. Literature data on ammonia toxicity (EPA, 19858)can be used only as a general guide because of thelarge effect of very slight pH changes. The pH valuesfor most ammonia toxicity tests are usually not mea-sured or reported fully enough to be useful.

One might expect ammonia to be removed duringthe pH 11 adjustment/aeration test. Based on our expe-rience, however, ammonia is not substantially removedby the method described in Section 8.5. Other tech-niques which can be used to remove ammonia relatedtoxicity may also displace metals or other toxicants withcompletely different physical and chemical characteris-tics. For example, ion exchange resins (e.g., zeolite)removes ammonia, cationic metals, and possibly or-ganic compounds through adsorption. For these rea-sons, the graduated pH test is most effective in differ-entiating toxicity related to ammonia from other causesof toxicity, if it is the dominant toxicant.

Effluent toxicity related to metals may also be de-tected by the graduated pH test, although these effectsare less well documented in effluents than those asso-ciated with ammonia toxicity. Acidification of a samplemay increase the bioavailable portion of a metal, and insome cases (i.e., cadmium, copper, and zinc) this iscountered by a decreasing toxicity of the metal as thetest pH decreases. It-is known, however, that aluminumtoxicity increases as pH diverges from neutral. In ex-periments in the pH range of 5 to 7 (Campbell andStokes, 1985) the toxicities of cadmium, copper andzinc were shown to increase with increasing pH whilethe toxicity of lead decreased with increasing pH. Wehave found lead and copper to be more toxic to C. _dubia at pH 6.5 than at pH 8.0 or 8.5, (in very hardreconstituted water) and nickel, zinc, and cadmiumwere more toxic at pH 8.5 than at 6.5. Since thesecompounds are also chelatable by EDTA, the resuqs ofboth tests (the graduated pH test and the EDTA addi-tion test) can give information about whether it is anionizable compound or a pH sensitive cationic metal.Other metals have exhibited some degree of pH de-pendence, but these are not as well defined. Results ofthe graduated pH test should be considered in conjunc-tion with the EDTA addition test (Section 8.8). Whetherthe metal toxicity can be discerned will depend in largepart on the concentration of other pH dependent toxitcants in the sample. In order to detect metal toxicity,one must be cautious when selecting a dilution waterwhen the test solutions are at low effluent dilutionsbecause artifactually enhanced toxicity due to metalsmay be created if the hardness of the dilution water ismuch different than that of the effluent. This effect maybe magnified for metals when coupled with the pHchange. A dilution water similar in hardness to theeffluent must be used for this test to reveal metal-caused toxicity. If more than one pH dependent toxi-cant is present, the pH effects may either cancel orenhance one another.

Ammonia acts as a basic compound in water. Theun-ionized, more toxic form (NH,) predominates abovepH 9.3 and the ionized, essentially non-toxic form (NH,+)is most abundant below pH 9.3 at 25°C. Through thepH range of 6-8.5, the percent of ammonia in the toxicform increases 250x over this range. Importantly, as pHincreases, the percentage of the toxic form becomesgreater but the toxicity of the toxic form is less, andconversely, as pH decreases, the percentage of ammo-nia (NH,) decreases, but the toxicity of the NH, in-creases (EPA, 19858). However, the increase in the

Hydrogen sulfide (H,S) occurs in wastewaters, andits toxicity can be detected by the graduated pH test.Dissolved sulfide exists in two forms, H,S and HS. Thepredominant form depends on both pH and tempefa-ture, but since temperature is held constant in thesePhase I toxicity tests it essentially can be ignored. Theun-ionized form (H,S) is more toxic to aquatic ofgan-isms, and at pH 6 it comprises over 90% of the dis-solved sulfide, while at pH 7, 50% is unionized. At apH of 8.5, less than 5% of the dissolved sulfide ispresent in the unionized form. Since H,S is the moretoxic form, one would expect to observe an increase intoxicity relative to a decrease in solution pH. Whenconsidering results of this type it is wise to checktoxicity alteration by the pH adjustment/aeration tests

8-44

(Section 8.5). H,S is readily oxidized and also removedthrough volatilization; therefore if H,S is the predomi-nating toxicant, a significant reduction of toxicity shouldbe observed in the pH adjustment/aeration tests.

The effects of pH on toxicity can be used to detectthe presence of these pH dependent toxicants. Byconducting three effluent tests, each at a different pH,the effluent toxicity can be enhanced, reduced or elimi-nated. For a typical example (at 25°C) where ammoniais the primary toxicant, when the pH is 6.5, 0.180% ofthe total ammonia in solution is present in the toxicform (NH,). At pH 7.5, 1.77% of the total ammonia ispresent as NH, and at pH 8.5, 15.2% is present asNH,. Similar changes in the percent ammonia as NHfor pH’s 6.5, 7.5 and 8.5 occur at other temperatures:for example, the percentages of unionized ammonia at20°C for pH’s 6.5, 7.5, and 8.5 are 0.130%, 1.24% and11.2%, respectively (EPA, 1979). This difference in thepercentages of unionized ammonia is enough to makethe same amount of total ammonia about three timesmore toxic at pH 8.5 as at pH 6.5. Whether or nottoxicity will be eliminated at pH 6.5 and the extent towhich toxicity will increase at pH 8.5 will depend on thetotal ammonia concentration. If the graduated pH testsare done at dilutions symmetrical about the LC50, oneshould see toxicity differences between pH 6.5 and 8.5(cf., Phase II discussion on equitoxic test). The effluentLC50 (expressed as percent effluent) should be lowerat pH 8.5 than pH 6.5 if ammonia is the dominanttoxicant.

The most desirable pH values to choose will de-pend upon the characteristics of the particular effluentbeing tested. For example, if the air equilibrium pH ofthe effluent at 4x the 24-h LC50 is 8.0 it may be moreappropriate to use pH’s 6.0, 7.0, and 8.0. The gradua-tion scheme that includes the air equilibrium (the pHthe effluent naturally drifts to) will allow a comparison oftreatments to unaltered effluent (i.e., baseline test). ThepH’s of many POTW effluents rise to 8.5 or higher, so agradient of pH’s such as 6.5, 7.5 and 8.5 is moreappropriate. In any case, it will be necessary to conductthe test at more than one effluent concentration (4x-,2x-, lx-, 24-h LC50) or with a different graduated pHscheme to determine what role, if any, the pH depen-dent compounds play in toxicity.

Perhaps the greatest challenge faced in this gradu-ated pH test is that of maintaining a constant pH in thetest solution. This is a necessity if the ratio of ionized tothe un-ionized form is to remain constant and the testresults are to be valid. In conducting toxicity tests oneffluents, it is not unusual to see the pH of the testsolutions with effluent concentrations of 212% drift 1 to2 units over a 48 to 96-h period (see Procedure forsuggestions on pH control).

Volume Required:The volume needed

sign chosen to conductsize, number of dilutions,will dictate this; however,suffice for all three pH’s.

is dependent on the test de-this test. The test chamberand the toxicity of the effluent200 mL of test volume’should

Apparatus:Test chambers such as 78 mm L x 50 mm W x 50

mm H or 1 oz plastic cups; Hamilton 1 L gas syringe(Model S-1000, Reno, NV); 35 mm x 14 mm H Corningplastic petri dish bottoms, rubber stoppers, eye drop-pers or wide bore pipette, 30 mL beakers or 1 ozplastic cups, light box, and/or microscope (optional).

Reagents:Cylinder tank of CO , 1 .O, 0.1, and 0.01 N HCI

1 .O, 0.1, and 0.01 N NaaH (ACS grade in high ‘puritywater), buffers for pH meter calibration.

Test Organisms:Use 5 for each of three dilutions of the whole

effluent (4x-, 2x-, lx-, 24-h LC50 or 1 OO%, 50%, and25%) and for each test pH (e.g., pH 6, 7, 8) (Figure 8-19) as well as a control.

Procedure:Day 2: Either CO or HCI (or the combination of

both) can be used to r’ower the pH of the sample. ThepH of most natural waters and some effluents is con-trolled by the bicarbonate buffering system. Surfacewaters normally contain ~10 mg/L of free CO,. ,;

For the CO, pH controlled tests, the pH is adjustedwith CO, by varying CO, content of the gas phase overthe water or effluent sample. It is necessary to maintainconstant pH’s in the static acute test throughout the 48or 96-h tests. By usiirg closed headspace test cham-bers, the CO, content of the gas phase can be con-trolled. The amount of CO, needed to adjust the pH ofthe solution is dependent upon sample volume, the testcontainer volume, the desired pH, the temperature, andthe effluent constituents (e.g., dissolved solids). Whendilutions of an effluent have the same hardness andinitial pH as the effluent, the same amount of CO willusually be needed for each dilution, but somet!mesmore is needed in the higher effluent concentrations.Use of a dilution water of similar hardness as theeffluent makes the CO, volume adjustments easier.

In our laboratory, a rectangular chamber (measur-ing 78 mm L x 50 mm W x 50 mm H) with a smalldiameter hole (approximately 20 mm) on one end hasworked well for the CO, graduated pH test. The testsolution volume should be about 10% of the headspacevolume to maintain a large surface to volume ratioshould be maintained. For a 20 mL test volume, withthe CO, gas flushed into air space of the test chamber,pH’s have reached equilibrium in about 1 h. In mostinstances, the amount of CO, produced by the inverte-brates has not caused further pH shifts, but with larvalfathead minnows, the pH can drop from the amount ofCO, they respire as well as decomposition of food.Therefore, in fish tests, the headspace must be reflusheddaily.

The exact amount of CO td inject for pH’s 6.0, 7.0and 8.0 must be determined through experimentationwith each effluent before the graduated pH test begins.

8-45

The amount of CO, added to the chamber assumesthat the liquid volume to gas volume ratio remains thesame. Generally, as the alkalinity increases, the con-centration of CO, that is needed to maintain the pHalso increases. inject the CO using a gas tight syringeand quickly close the test chamber tightly. Place thetest chamber in a position that maximizes the surfaceto volume ratio. To prepare the test solutions, use adilution water of a similar hardness to the effluent andtransfer the effluent solutions to the test container andrandomly add the test organisms. Then add the pre-determined amount of CO, to obtain the desired pH’sand close the container. For pH values from pH 8.5 to6, O-10% CO, has been needed. If more than 10% CO,is needed, adjust the solutions with acids and bases(described below) and flush the headspace with CO,.Again, the necessary concentration of CO, to use must ’first be determined experimentally with effluent testsolutions already adjusted to the appropriate pH. Thismay require the test to be set up one day later than theother Phase I tests.

For some effluents adequate pH control can beobtained by adjusting the pH with acid or base andtightly covering the test container (no headspace pHtest). A technique that we use has the 1 oz plastic cupscovered with plastic tissue culture dishes (see Appara-tus for details). This technique works well with effluentsthat have adequate DO content, and’where the BOD isnot high. The procedure for using plastic cups withtissue culture dish covers is as follows. Adjust threealiquots of the effluent and the dilution water to theappropriate pHs. Next, prepare the appropriate dilu-tions for testing (i.e., 4x-, 2x-, 1x-24-h LC50, or 1 O%,50%, 25%) and check the pH in one-half hour. If thepH’s have drifted, readjust them with the appropriateacid or base. Transfer about 35 mL of each into the 1oz plastic cups, and randomly add the test organisms.Carefully place the cover onto the cup; care must beexercised because some test water will be displaced bythe lid, and organisms can be lost. Ensure that no air istrapped under the lid during the sealing process. If airis trapped, remove the cover, count the number oforganisms, and add an additional small amount of theappropriate pH adjusted test solution. The test organ-isms can readily be observed through the clear coveror the sides of the plastic cup. The cover should beremoved only when all the animals have died as thetight seal cannot be obtained after initially setting upthe test without adding more test water. Once animalshave died or the test is over, remove the cover andmeasure the pH and DO. It is important to measure theDO because toxicants such as ammonia have differenttoxicities when DO is low (EPA, 1985B). Keep in mindthat if all of the test animals have been dead for awhile, the pH and/or DO of the test water could havechanged.

Methods that use continuous flow of a COdairmixture, such as tissue cell incubators, may be prefer-able and give better pH control. At this time we havenot attempted to use a continuous flow of CO, andcannot recommend a system to use.

Maintaining pH above the air equilibrium pH (gen-erally above 8.3) is difficult to achieve. The pH controlin this high range is much more difficuft because theconcentration of CO2 must be very low and the micro-bial respiration can Increase the CO, levels in the testchamber. Use of CO,-free air in the headspace maywork or bubbling a mtx of CO,-free air and normal airthrough the headspace or test solution may be needed.Because such small CGZ concentrations are neededand because CO, evolutron by microorganisms or testorganisms can significantly alter the CO, concentration,more frequent flushing of the headspace in static testswill be needed.

Since many plastics are permeable to CO,, glasscontainers may need to be used. Measurements of pHmust be made rapidly to minimize the CO exchangebetween the sample and the atmosphere. dvoid vigor-ous stirring of unsealed samples because at lower pHvalues, the CO loss during the measurement can causea substantial pk rise.

For the CO, pH controlled tests, the pH should bemeasured at 24, 48, 72, and 96 h and at each reading,one may need to re-flush the headspace with CO,. Asmall amount of experimentation will determiOe theamount of CO, needed for this step. For tf?e noheadspace pH tests conducted in cups with covers, airbubbles may start to appear after 12 h, and this cancause the pH to change. An excess of each test solu-tion may need to be prepared to be added to the testcup at each 24 h interval to prevent the formation of airpockets which contribute to pH drift.

We also have been exploring the use of hydrogenion buffers to maintain the pH of effluent test solutions.Efforts to use phosphate buffers were unsuccessfuldue to the toxicity of the phosphates themselves. Threehydrogen ion buffers were used by Neilson et al. (1990)to control pH in toxicity tests in concentrations rangingfrom 2.5 to 4.0 mM. These buffers were chosen basedon the work done by Ferguson et al. (1980). Thebuffers are: 2-(N-morpholino) ethane-sulfonic acid (Mes)(PK. = 6.15)) 3-( N-morpholino) propane-sulfonic acid(Mops) (PK, = 7.15), and piperazine-N,N’-bis (2-hydroxypropane) sulfonic acid (Popso) (pK, = 7.8).

The acute toxicity of these buffers is low to bothC. dubia and fathead minnows (Table 8-6) and suble-thal levels can be added to hold the pH of test solu-tions. For example, 6.25 mM (1.2 g/L) of the Mes bufferhas been adequate to maintain the pH of one effluentto within +O.l pH units. However when used in a sedi-ment pore water, more buffer was needed (i.e., 25 mMor 4.9 g/L) but these levels are still below the acutetoxicity for the buffer. Likewise for the Mops buffer, 6.25mM (1.3 g/L) held the pH of the effluent at kO.1 pHunits, but 50 mM (10.5 g/L) was needed for the porewater. The Popso buffer held the pH at 8.2 or 8.5 using6.25 or 12.5 mM (2.3 or 4.5 g/L, respectively) of bufferfor both the effluent or pore water. The addition ofthese buffers did not change the toxicity of a non-toxiceffluent or change the toxicity of a toxic effluent andsediment pore water.

Table 8-6. The toxicity of the Mea, Mopr, and Popao buffers to Cerfodaphnlo dubia and fathead minnowa

Buffsr Specie8WaterType 24 h

50 @,/!_I48 h 72 h 96 h

Mes C. dubia

Mes C. dubia

LSW

VHRW

15.0

17.4

Mops

Mops

C dubia

C. d&a

LSW

VHAW

16.1

B20.9

13.0

11.9

Pops0

Pops0

C. dubia

C. dubia

LSW

VHRW

>2.3

12.7

~2.3

8.3

Mes

Mes

P. promelas

P. promelas

LSW

VHRW

13.9

B19.5

13.9 13.9 13.9

B19.5 B19.5 z-19.5

Mops

Mops

P. promelas

P. promelas

LSW

VHRW

B20.9 17.2 16.1 16.1

>20.9 B20.9 >20.9 B20.9

Pops0

Pops0

P. promelas

P. promelas

LSW

VHRW

32.3

~36.2

32.3 27.9 27.9

~36.2 ~36.2 fl6.2

Note: The pH was held to at least f 0.1 pH unit of desired pH for all tests. Mes buffer tests were at pH 6.2, Mops buffer tests were at pH7.2, and Popso buffer tests were at pH 8.2. LSW = Lake Superior water; VHRW = very hard reconstituted water.

While these buffers serve to prevent the pH fromdrifting, their addition alone does not actually adjust thepH value to the desired pH. The buffers are weighedout and added to the aliquots of whole effluent anddilution water and both are then pH adjusted with baseto the appropriate values. Serial dilutions are made,and test organisms are added. While our experiencewith the buffers is limited, we have found the amount ofany buffer needed to hold any pH is effluent specific.Experiments will need to be done to determine thelowest concentration of buffer needed to maintain thedesired pH. The test solutions need not be coveredtightly to maintain pH: however, pH should be mea-sured at each survival reading at all dilutions.

In all graduated pH tests, the pH should be mea-sured at least in the chambers that bracket the LC50concentration as soon all the animals die. If the pHdrifts more than 0.2 pH units, the results may not beusable and better pH control must be achieved.

Interferences/Controls and Blanks:The controls in the CO, chamber or closed cup,

and the baseline test act as checks on the generalhealth of the test organisms, the dilution water andmost test conditions. If the effluent pH in the baselinetest (at the LCSQ) is close to the pH of the pH adjustedtest solutions (at their respective LCZO’s), the toxicity

expressed in the two tests should be similar. Signifi-cantly greater toxicity may suggest interference fromother factors such as the ionic strength related toxicityif the pH was adjusted with either HCL or NaOH (cf.,Section 8.3), or CO, toxicity. Dilution water blanks atthe various pH’s are not used because such blanks arenot appropriate since the effluent matrix may differ fromthat of the dilution water. The cleanliness of the acidsand bases is checked in the blanks of the pH adjust-ment test. Other compounds with toxicities that in-crease directly with pH may lead to confounding resuttsor may give results similar to ammonia. Phase II con-tains a suggested test (called the equitoxic test) toidentify ammonia as the cause of toxicity. Monitoringthe acid and base additions may be useful to determineif amfactual toxicity resulted from the addition of thesalts. Monitoring conductivity of the effluent solutionsafter the addition of the acids and bases may also behelpful in determining artifactual toxicity.

Results/Subsequent Tests:For the graduated pH test, the pHs selected must

be within the physiological tolerance range for the testspecies used (which generally is a pH range of 6 to 9).In this pH range, the amount of acid or base added isnegligible, and therefore the likelihood of toxicity due toincreased salinity levels is low.

8-48

.

When ammonia is the dominant toxicant, the efflu-ent LC50 of the pH 6.5 test solution should be higherthan in the pH 7.5 test, which in turn, should be higherthan the pH 8.5 test. However, ammonia is not the onlypossible cause of toxicity. Using the pH at the baselineeffluent LC50, the relative toxicity of each pH adjustedsolution can be predicted if ammonia is the sole causeof toxicity. For example, if in the baseline effluent toxic-ity test, the average pH was 8.0 in the 100% concentra-tion in which no organisms survived and the averagepH for the 50% concentration was 7.5 and all organ-isms survived, the estimated pH at the LC50 (71%)could be approximated at 7.7. One would expect greater

than 50% mortality in the pH 8 test solution and signifi-cantly less in the pH 7 solution. Therefore, if this occursone should proceed to Phase II to identify the pHsensitive toxicant.

If ammonia is one of several toxicants in an efflu-ent, this procedure may pose problems. For this rea-son, if effluent total ammonia levels are greater than 20mg/L, it may be appropriate to include a pH 6 effluenttreatment interfaced with other Phase I tests (cf., Sec-tion 9). Methods for further identifying and confirmingammonia as the toxicant can be found in Phases II andIII.

8-49

Section 9Time Frame and Additional Tests

9.1 Time Frame fur Phase I StudiesThe amount of time necessary to adequately chat-

acterize the physical/chemical nature of, and variabilityin, an effluent’s toxicant will be discharge specific.Among the factors affecting the length of Phase I stud-ies for a given discharge is the appropriateness ofPhase I tests to the toxicants, the existence of long- orshort-term periodicity in individual toxicants and to alesser extent, the variability in the magnitude of toxicity.An effluent which consistently contains toxic levels of asingle compound that can be neutralized by more thanone characterization test, should be moved into PhaseII more quickly than an ephemerally toxic effluent withhighly variable constituents, none of which are im-pacted by any of the Phase I tests. The decision as towhen to go beyond Phase I should be based in part onthe regulatory implications and resources involved insubsequent actions. Where a great amount of resourcesis involved, it is crucial that Phase I results be ad-equate.

There are no clearly defined boundaries betweenPhase I and Phase II. The section Results/SubsequentTests of the characterization tests in Section 8 providefurther tests to conduct and may be thought of asintermediate studies between Phases I and II. In termsof guidance for the time frame of the TIE, severalsamples should be subjected to the Phase I character-ization test battery but not all manipulations have to bedone on all subsequent samples. The decision to dosubsequent tests on these samples to confirm or fur-ther delineate initial results is a judgement call and willdepend on whether or not the results of Phase I areclear-cut. The time required to perform a completePhase I battery on a sample will depend on manycircumstances, not the least of which is how well orga-nized and experienced the performing lab is at doingTIES.

If the Phase I characterization tests needed toremove or neutralize effluent toxicity vary by the sample,the number of tested samples must be increased andthe frequency of testing should be sufficient to includeall major variability. We cannot provide a time frame orthe number of samples to evaluate. Again, judgementwill have to be used but the differences seen amongsamples can be used to decide when further differ-ences are not being found. Phase I characterization

months this may take--each problem for every dis-charger is unique. The LC50 of samples can be verydifferent but the same screening tests must be SUC-cessful in removing and/or neutralizing effluent toxicity.

The individual Phase I tests which were previouslysuccessful in changing toxicity should be used as astarting point for Phase II identification. The first step inPhase II will often be to reduce the number of constitu-ents accompanying the toxicants. These efforts mayreveal more toxicants than suggested by Phase I test-ing. In Phase II one may discover that toxicants of quitea different nature are also present but were not inevidence in Phase I. More Phase I charactarizationmay then be needed. b

Phase I results will not usually provide informationon the specific toxicants. Therefore, if effluent toxicity isconsistently reduced, for example through the use ofCre SPE, this does not prove the existence of a singletoxicant. In fact, several non-polar organic compoundsmay be causing the toxicity in the effluent over time,but use of the C,, SPE technique in Phase I detects thepresence of these compounds as a group. Recognizingthis lack of specificity is very important for subsequentPhase II toxicant identification.

9.2 When Phase I Tests are InadequateFor some effluents, the Phase I tests described

above will provide few or no clues as to the characteris-tics of the toxicants. For such effluents, other approachesmust be tried. Some additional approaches are givenbelow in much less detail than tests in Section 8 be-cause our experience with them is limited. In addition tothese, one should not hesitate to use originality andinnovation to develop other approaches. As long astoxicity is used to track the changes, any approach maybe helpful.

Use of Multlple Phase I ManipulationsOur experience suggests that independent action

and less than additive action are much more commonthan we realize, at least in effluents. When these inter-actions occur, interpreting Phase I data may be difficultand in some instances (especially with independentaction) no apparent effect on toxicity will be seen un-less Phase I tests are clustered or used in a series.These steps do not begin until all Phase I manipula-

r tions have been completed and the results evaluated.testing should continue until there is reasonable cer-tainty that new types of toxicants are not appearing. No Tests are continued to further separate and concen-guidance can be given as to how many weeks or trate the toxicant( . ’

9-19-1

The pH of effluents plays an amazingly powerfulrole in how it affects both the form of toxicants and theirtoxicity. Including pH adjustments to different valuesthan is suggested in Phase I may be helpful. Forexample, if the C, SPE column has partially removedthe toxicity, then flhase I manipulations with the post-column sample may be possible (cf., Section 8.6 pHadjustment/solid phase extraction test). For this mul-tiple manipulation, the post-C,, SPE column effluentcan be treated as whole effluent, and several of thePhase I steps conducted on the post-column effluenthave been found to be useful in further characterizingadditional effluent toxicants. These include the EOTAaddition test, the thiosulfate addition test, and the gradu-ated pH test. Another combined test is to test the post-column effluent that has been spiked with the 100%methanol eluate to see if the toxicity is equal to that ofthe whole effluent. However, this can be a tricky ma-nipulation as the post-column effluent is not the sameas the original effluent, and spiking methanol into thesample may lower the DO as well as cause quickbacterial growths, which may result in erratic mortali-ties.

We have used aeration/fiItration/pH adjustment/C,,SPE in various combinations to decipher the changesthat occur. The presence of more than one toxicantmay often require such combinations. For instanceswhere there are multiple toxicants and aeration andEDTA have both removed some toxicity, the addition ofEDTA in the post-aeration sample may help character-ize whether a metal(s) is causing the toxicity that is notremoved through aeration.

If the C,, SPE column has partially removed toxic-ity, it may be possible to pass the post-column effluentover an ion exchange column to determine the charac-teristics of the remaining toxicity. If a non-polar toxicantand ammonia are suspected, then passing the sampleover the C,, SPE column and then over zeolite (cf.,Phase II), may assist in accounting for all of the toxicity.Likewise, passing the effluent over zeolite and thenover the C,, SPE column may provide additional in-sight. To gain this knowledge it is essential that toxicitytests be performed after each manipulation and not juston the multiple manipulated sample.

A special effect occurs when an effluent, whichcontains two toxicants at very different concentrations,is diluted. Suppose that toxicant A would produce anLC50 at 50% effluent and toxicant B causes an effluentLC50 at 5%. In most cases, only toxicant t3 will materi-ally affect toxicity because the effect of A will be “di-luted out” long before the LC50 of B is reached. Unlessthe toxicity degrades rapidly, the Phase I tests for suchan effluent would be performed near the LC50 of B(20% is 4x-LC50) in which case the toxicity of A will notbe yoticed. If one finds toxicity at effluent concentra-tions in the very low range (such as loo/,) additionalPhase I testing at higher effluent concentrations shouldsubsequently be done. Cases such as these should becaught in Phase III, but earlier detection will be morecost-effective.

The two objectives which usually move the TIEalong more rapidly are to separate and concentrate thetoxicant( Anything that can be done in Phase I toachieve these goals will speed the process.

Activated CarbonThe use of carbon has been limited because it is

much less selective than ion exchange or C,, SPEcolumns, and extraction is less precise and more diffi-cult. However, carbon’s non-selectivity can be an ad-vantage in some situations. When a rather wide arrayof more specific methods have failed and the Phase Itests above have not changed toxicity, a “chemicalsponge” may be useful. In order to start, one must beable to alter toxicity somehow in order to tell whatchanges are occurring. A second objective in earlywork is a way to remove the toxicants from the sample(i.e., to concentrate them). Carbon has a high capabilityto do both. Furthermore, the knowledge about carbonsorption and extraction is extensive and help can befound in the literature. While it is true that carbon mayalter some chemicals, many are not affected by it. Wemust recognize that other conventional methods suchas ion exchange are also not specific. Ion exchangecolumns can sorb non-polar organics and C,, SPEcolumns can sorb metals. *

Otfter Specific Ion Columns

_

Many other types of resin columns are availablethrough commercial sources. Many of these have “in-surmountable” blank toxicity problems but some showpromise. Mixed bed ion exchange columns appear prom-ising because pH is not drastically altered as the samplepasses through the resin bed and the blanks appear tobe acceptable. Of course with any of these lesser usedmethods, the organism’s tolerance to dilution waterpassed over the resins and the eluate(s) must be deter-mined.

Other LIgandsEDTA reduces toxicity for only some of the cationic

metals, and other ligands may help.

9.3 interpreting Phase I ResultsAfter the suite of Phase I tests has been com-

pleted, the results will usually show that some manipu-lations increased toxicity, some decreased it, and oth-ers effected no change. Rarely is there no effect fromany manipulation. Frequently more than one manipula-tion affects toxicity. Even if toxicity is affected by onlyone manipulation, one still does not know whether ornot there are multiple toxicants. When several manipu-lations affect toxicity, it still does not ensure that thereare multiple toxicants. There is also no way to tell atthis stage if there are multiple toxicants, whether or notthey are additive, partially additive or independent. Inour experience with about 80 different effluents, wehave not found synergism but independent action hascommonly been found. Some toxicants identified ineffluents have been additive, but more often thesehave been only partially additive. In regard to multipletoxicants, refer to the above section Use of Multiple

9-2

Phase I Manipulations, regarding complications of de-termining toxicant interactions in effluents.

After Phase I is completed on a sample, the investi-gator must carefully evaluate the data, draw conclu-sions, and make decisions about the next steps thatare needed. Sometimes the next step is obvious, atother times the outcome will be confusing and the nextstep will not be obvious. Several general suggestions,based on our experience to date, may provide somehelp.

As a matter of principle, where multiple toxicantsare involved, experience shows that once one toxicantis identified, identification of subsequent toxicants be-comes easier because:

1. The toxicity contribution of the identifiedtoxicant can be established for each sample.

2. The number of Phase I manipulations thatwill affect the toxicity of the known toxicantcan be determined.

3. One can determine whether the identifiedand the unidentified toxicant are additive.

4. If some manipulations affect the toxicitydue only to the unidentified toxicants, some Iof their characteristics can be inferred.

5. One can determine if the relative toxicitycontributions of identified and unidentifiedtoxicants varies by sample. Such informationcan be used to design tests to elucidateadditional physical/chemical characteristics.

Another suggestion, is that when some Phase Ioutcomes are understandable and others are not, con-centrate on the one or the few that seem to be the mostclear-cut and which have a major effect on toxicity. Forexample, if an effluent has 10 TU and 2 TU are re-moved by the addition of EDTA, 1 TU is removed bythe C,, column and 5 TU are removed by the aerationmanipulation, begin identification on the toxicity removedby aeration. In another example, suppose the filtrationmanipulation reduced the toxicity by 1 TU, both pH 3and pH 11 adjustment tests showed that the toxicityincreased by 2 TU, the graduated pH test at pH 7decreased toxicity by 2 TU and the post-c,, SPE col-umn effluent (at pH i ) had 2.5 TU less toxicity than thewhole effluent. of th8 2.5 TU removed by the column,1.7 TU could be eluted with the 100% methanol. Thenext step then is to begin the Phase II identification onthe SPE extractable toxicity because:

1. Widely accepted methods are available foranalyses of many non-polar organiccompounds.

2. The method exists for both separating andconcentrating such toxicants (cf., Phase II).

3. This C, extractable toxicity manipulationbehaved as expected.

4. Many effluents have non-polar toxicity, andbased on those probabilities, that non-polartoxicity is likely to be real.

In the latter example, the unexplainable pH andfiltration effects might be a result of the behavior of thenon-polar toxicant or could be caused by some as-sociated artifact. If the non-polar toxicity is identified,then the results of the pH adjustment and filtrationsteps may be explainable.

The third suggestion is to concentrate on thosemanipulations affecting toxicity in which the toxicant isremoved from other effluent constituents. In the aboveexample, the SPE column separated the toxicantfrom other non-sorbable constituents. Other examplesof where the toxicant is removed from the other con-stituents are the filtration and the aeration manipula-tions.

Separating the‘toxicant(s) from non-toxicant(s), andconcentrating the toxicant are usually the most produc-tive efforts to pursue before identification (analyses)begins. Attempts to begin analysis for suspect toxicantwithout thisstep is frequently a mistake, and can becostly.

9.4 Interpretation ExamplesIn this section, various examples of Phase I results

are given with interpretation suggestions. TheseShouldbe used only as guides to thinking and not as definitivediagnostic characteristics. Since almost any toxicantcan be present in effluents, clear-cut logic is not totallydependable in interpreting results. Rather, one mustuse the weight of evidence to proceed, and be awarethat artifacts cannot at this point always be identified.

One should avoid making categorical assumptionsto every extent possible. For example, to assume thatthe toxicity is due to a non-polar toxicant becausethe toxicity in the post-c,, SPE column effluent wasremoved often is an error. Metals may also be thetoxicant adsorbed by the SPE column. However, as inthe example in Section 9.3, if the toxicity can be recov-ered in the methanol fraction (see Section 8.6, Results/Subsequent Tests for elution and Phase II for moredetails), then the theory that a non-polar toxicant iscausing the toxicity is better substantiated. Metals donot elute with methanol and therefore do not producetoxicity in the methanol fraction toxicity test (cf.,Phase II).

Example 1. Nun-polar toxicant( The Phase I re-sults implicating non-polar toxicants are:

1. All toxicity in the post-c,, SPE columneffluent was removed.

2. The toxicity removed was recovered in themethanol elution of the SPE column.

The above discussion (cf., Section 9.3) has pro-vided most of the interpretative rationale for thesePhase I results which are typical of non-polar organics.As stated above, toxicants other than non-polar com-pounds may be retained by the SPE column but theyare less likely to be eluted sharply. Also, as discussed

9-3

Example V. Ammonia. Ammonia concentrations canbe measured easily, and because it is such a commoneffluent constituent, determining the total ammonia con-centration in the whole effluent is a good first step (seeSection 6). If more than 5 mg/L of total ammonia ispresent, additional evaluations should be done. Soledependence on analyses is not advisable because thereis little or no additivity between ammonia and someother toxicants (e.g., such as surfactants). Even thoughthe ammonia concentration is sufficient to cause toxic-ity, other chemicals may be present to cause toxicity ifthe ammonia is removed.

Three indicators of ammonia toxicity are:

1. The concentration of total ammonia is5 mg/L or greater.

2. Toxicity increases as the pH increases.3. The effluent is more toxic to fathead

minnows than to Ceriodaphnia or Daphnia.

Example VI. Oxidants. In effluents, oxidants otherthan chlorine may be present. Measurement of a chlo-

-

rine residual (TRC) is not enough to conclude that thetoxicity is due to an oxidant.

In general, oxidants are indicated by the following:

1.

2.

3.

4.

The addition of sodium thiosulfate to theeffluent reduced or removed the toxicity.Aeration without any pH adjustmentremoved or reduced toxicity.The sample is less toxic over time whenheld at 4°C (type of container is not anissue here).Ceriodaphnia are more sensitive thanfathead minnows.

Of course, TRC greater than 0.1 mg/L at the efflu-ent LC50 concentration (and depending on test spe-cies) would indicate chlorine as the oxidant causing thetoxicity. In addition, the dechlorination with SO, pro-vides evidence of chlorine toxicity in the same manneras the sodium thiosulfate addition test.

9-5

Section IOIO

References

Adelman, I .R., I-L. Smith Jr., and G.D. Siesennop. Third Edition. EPA-600/4-85/013. Environmental1976. Acute Toxicity of Sodium Chloride, Monitoring and Support Laboratory, Cincinnati, OH.Pentachlorophenol, Guthion@, and Hexavalent EPA. 19858. Ambient Water Quality Criteria forChromium to Fathead Minnows (Pimephales Ammonia. EPA-440/5-85-001. Environmentalpromelas) and Goldfish (Carassius auratus). J. Fish. . Research Laboratory-Duluth, MN, and the CriteriaRes. Board Can. 33:203-208. and Standards Division, Washington, D.C.

Amato, JR., D.I. Mount, E.J. Durhan, M.T. Lukasewycz,G.T. Ankley and E.D. Robert. 1992. An Example ofthe identification of Diazinon as a Primary Toxicantin an Effluent. Environ. Tox. and Chem: In Press.

Ankley, G.T., J.R. Dierkes, D.A. Jensen and G.S.Peterson. 1991. Piperonyl Butoxide as a Tool inAquatic Toxicological Research WithOrganophosphate Insecticides. Ecotoxicol. Environ.Safety. 21:266-274.

EPA. 1988A. Methods for Aquatic Toxicity IdentificationEvaluations: Phase I Toxicity CharacterizationProcedures. EPA-600/3-88-034. EnvironmentalResearch Laboratory-Duluth, MN.

EPA. 19888. Short-Term Methods for Estimating theChronic Toxicity of Effluents and Receiving Watersto Marine and Estuarine Organisms. EPA-600/4-87-028. Environmental Monitoring and &pportLaboratory, Cincinnati, OH.

Ankley, G.T., M.T. Lukasewycz, G.S. Peterson andD.A. Jenson. 1990. Behavior of Surfactants inToxicity Identification Evaluations. Chemosphere.21:3-l 2.

EPA. 1989A. Toxicity Reduction Evaluation Protocol forMunicipal Wastewater Treatment Plants. EPA/GOO/2-881062. Water Engineering Research Laboratory,Cincinnati, OH.

APHA. 1980. Standard Methods for the Examination ofWater and Wastewater. 15th Edition. AmericanPublic Health Association, American Water WorksAssociation Water Pollution Control Federation,Washington, D.C.

EPA. 19898. Generalized Methodology for ConductingIndustrial Toxicity Reduction Evaluations (TREs).EPA/600/2-88/070. Water Engineering ResearchLaboratory, Cincinnati, OH.

Berg, E. 1982. Handbook for Sampling and SamplePresentation of Water and Wastewater. EPA-600/4-82-019. Cincinnati, Ohio.

EPA. 1989C. Methods for Aquatic Toxicity IdentificationEvaluations: Phase II Toxicity IdentificationProcedures. EPA/600/3-88/035. EnvironmentalResearch Laboratory, Duluth, MN.

Bowman, M.C., W.L. Oiler, T. Cairns, A.B. Gosnell andK.H. Oliver. 1981. Stressed Bioassay Systems forRapid Screening of Pesticide Residues. Part I:Evaluation of Bioassay Systems. Arch. Environ.Contam. Toxicol. 10:9-24.

EPA. 19890. Methods for Aquatic Toxicity IdentificationEvaluations: Phase III Toxicity ConfirmationProcedures. EPA/600/3-88/036. EnvironmentalResearch Laboratory, Duluth, MN.

Campbell, P.G.C. and P.M. Stokes. 1985. Acidificationand Toxicity of Metals to Aquatic Biota. Can. J.Fish Aquat. Sci. 42:2034-2039.

DHEW. 1977. Carcinogens - Working With Carcinogens.Public Health Sewice, Center for Disease Control,National Institute of Occupational Safety and Health.Department of Health, Education and Welfare.Publication No. 77-206.

Dowden, B.F. and H.J. Bennett. 1965. Toxicity ofSelected Chemicals to Certain Animals. J. WaterPollut. Control Fed. 37(9):1308-l 316.

EPA. 1979. Aqueous Ammonia Equilibrium - Tabulationof Percent Un-ionized Ammonia. EPA-600/3-79/091. Environmental Research Laboratory, Duluth,MN.

EPA. 1989E. Short-Term Methods for Estimating theChronic Toxicity of Effluents and Receiving Watersto Freshwater Organisms. Second Edition, EPA/600/4-89/001 and Supplement EPA/600/4-89/001A.Environmental Monitoring and Support Laboratory,Cincinnati, OH.

EPA. 1991A. Toxicity Identification Evaluation:Characterization of Chronically Toxic Effluents,Phase I. EPA/600/6-911005. EnvironmentalResearch Laboratory, Duluth, MN.

EPA. 1985A. Methods for Measuring the Acute Toxicityof Effluents to Freshwater and Marine Organisms.

EPA. 1991 B. Methods for Measuring the Acute Toxicityof Effluents to Freshwater and Marine Organisms.Fourth Edition. EPA-600/4-90/027, EnvironmentalMonitoring and Support Laboratory, Cincinnati, OH.

EPA. 1991C. Short-term Methods for Estimating theChronic Toxicity of Effluents and Receiving Watersto Freshwater Organisms. Third Edition. EPA/600/4-91/002. Environmental Monitoring and SupportLaboratory, Cincinnati, OH.

Federal Register. 1984. U.S. EPA: Development ofWater Quality Based Permit Limitations for ToxicPollutants; National Policy. EPA, Volume 49, No.48, Friday, March 9, 1984.

Ferguson, W.J., K.I. Braunschweiger, W.R.Braunschweiger, JR. Smith, J.J. McCormick, C.C.Wasmann. N.P. Jarvis, D.H. Bell, and N.E. Good.1980. Hydrogen ion buffers for biological research.Anal. Biochem. 104: 300-310.

Giles, M.A. and R. Danell, 1983. Water Dechlorinationby Activated Carbon, Ultraviolet Radiation andSodium Sulphite. Water Res. 17(6): 667-676.

Hackett, J.R. and D.R. Mount. In Preparation. Use ofMetal Chelating Agent to Differentiate AmongSources of Toxicity. Manuscript.

Magnuson, V.R., D.K. Harriss, M.S. Sun, D.K. Taylor,G.E. Glass. 1979. Relationships of Activities of ’Metal-Ligand Species to Aquatic Toxicity. ACS,Symposium Series, No. 93. Chemical Modeling inAqueous Systems, E.A. Jenne, Editor. pp. 635656.

Neilson, A.J.-, A.S. Allard, S. Fischer, M. Malmberg, andT. Viktor. 1990. Incorporation of a Subacute Testwith Zebra Fish into a Hierarchical System forEvaluating the effect of Toxicants in the AquaticEnvironment. Ecotox. and Environ. Safety 20:82-97.

Norberg-King, T.J., E.J. Durhan, G.T. Ankley and E.Robert. 1991. Application of Toxicity IdentificationEvaluation Procedures to the Ambient Waters ofthe Colusa Basin Drain. Environ. Tox. and Chem.10(7):891-900.

OSHA. 1976. Occupational Safety and HealthAdministration. OSHA Safety and Health Standards,General Industry. 29 CFR 1910. OSHA 2206(Revised).

Patrick, R., J. Cairns Jr., and A. Scheier. 1968. TheRelative Sensitivity of Diatoms, Snails, and Fish toTwenty Common Constituents of Industrial Wastes.Prog. Fish-Cult.:1 37-140.

Poirier, S.H., M.L Knuth, C.D. Anderson-Buchou, L.T.Brooke, A.R. Lima, and P.J. Shubat. 1986.Comparative Toxicity of Methanol and N,N-Dimethylformamide to Freshwater Fish andInvertebrates. Bull. Environ. Contam. Toxicol.37(4):615-621.

Randall, T.L. and P.V. Knopp. 1980. Detoxification ofSpecific Organic Substances by Wet Oxidation. J.Water Pollut. Control Fed. 52(8):2117-2130.

Schubauer-Berigan, M.K. and G.T. Ankley. 1991. TheContribution of Ammonia, Metals and Non-polarOrganic Compounds to the Toxicity of SedimentInterstitial Water from an Illinois River Tributary.Environ. Toxicol. and Chem. 10(7):925-939.

Smith, R.M. and A.E. Martell. 1981. Critical StabilityConstants. Volume 4: Inorganic Complexes. PlenumPress, NY.

Stumm, W. and J.J. Morgan. 1981. Aquatic Cher%try -An Introduction Emphasizing Chemical Equilibria inNatural Waters. John Wiley & Sons, Inc., NewYork, NY.

Walters, C.I. and C.W. Jameson. 1984. Health andSafety for Toxicity Testing. Butterworth Publ.,Woburn, MA.

1 o-2

* L

,

TECHNICAL REPORT OATA(Pltosr rrod lnrmrnont on fhr rtvtne brforw comulrnn~j I

1 ntpont NO. a. ( 3. ALC4?lE~T’S ACCtSflON NO* IEPA/600/6-91,'003 PEW-100PEW-100 072 072 II

,,,, TITLETITLE ANOSUB~~~L~ANOSUB~~~L~ 9.RCrOnTOATl9.RCrOnTOATl ii

~ktiods~ktiods forfor -ati& mxicity Identificdion Evaluation: February 1991Phase 1 Toxicity Characterization Procedures (Second ~.~E~~OR~INGORC~~~Z~TI~~CC=)~ I

Edition)7 *UtHORISJ 1 3 1

Nor&p-King, T.J.I,lbbunt, D/, Durban, &,' Ank$ey,G.T. , Burkhard, L. I Amato, J.*, Lukasewycz, M. I

0, ~R~OMHG ORGANIZATION NAME AN0 AODRESSI

10. PROGRAM ELSMENT two,

knited States Env~onmental Protection Agency,Environmntal Reseych Laboratory, 6201 Congdon glvd. 11. CONTRACTIGRANT NO.

Duluth, MI3 55804; AScI Carp, Duluth, MN 55804; U.S.

12. SPONSORING AGENCY NAME ANO ADOmESS

Environmental Research Environmental Research LabdratoryLabdratoryIfficeIffice of Research and Development of Research and Development

13. T Y P E OC REPORT AN0 PER,00 COVEAEO

14. SPONSOR1NG A G E N C Y COOE

j.S.j.S. Environmental Protection Agency Environmental Protection AgencyIuluth,Iuluth, MN 55804 MN 558045.5. SUPPLEMENTARY NOTES

EPASeries Research Reprt .

.

EPA-600/03EPA-600/03.

I

ABSTRACTf

In 1988, the first edition “Methods for Aquatic Toxicity Identification Evaluations:

Phase I Toxicity Characterization Procedyres” was publiShed (EPA, 1988A). This

second edition provides more details and ‘more insight into the techniques described inin

the 1988 document. The manual describes procedures for characterizing the

physical/chemical nature of toxicants in acutely toxic effluent samples, with

applications to other types of samples such as receiving water samples, sediment

pore water or elutriate samples, and hazardous wastes. .d

5

. &LY wOROS AN0 OOCUMINT ANALYSIS

OasCnICtOns b.tocfuT~FlliRS/OPIN ENOEO TERMS

Author(s) Contmued: c

Anderson<amahan, L.3.

Performing Org. Name and Address Continuak

Environmental Protection Agency, -onIv-Policy Planning and Evaluation Branch,345 courtland, Atlanta, GA 30365.

1. ~~ST~~~UTIO~ S~ATL.MENT rs. SECURITY CLASS lthtr Report}

eleaseelease to public to public UnclassifiedUnclassified,120. SECURITY CLASS /Thu pot*)

. COSAT4 FeldiCroup

21. NO. OF PAG

22. PRICE

)) Unclassified Unclassified I

trr Pwc* ZflO-I (Rev. 4-11) rrCvlO&Jl COlT~O*) 1% omrobc~s* U . S . c o - m m a : 1991.5ra-le)rabo2