higher tier laboratory aquatic toxicity testing · cranfield centre for ecochemistry 2 foreword...

Cranfield Centre for EcoChemistryCranfield University, Silsoe, Beds MK45 4DT, UK

Tel: 01525 863000 Fax 01525 863253E-mail: [email protected] Web: http://www.cranfield.ac.uk/ecochemistry

Higher Tier Laboratory Aquatic Toxicity Testing

A Boxall, C Brown & K Barrett

Final report

March 2001

DETR Contract Reference No. EPG 1/5/131

Cranfield Centre for EcoChemistry Contract No. JA4317E

Cranfield Centre for EcoChemistry

2

FOREWORD

This report reviews the role of higher tier laboratory aquatic toxicity studies in the risk

assessment process. Results from a survey of regulators, industry and academics are presented

along with a review of available literature describing the use of higher tier studies. Guidance

is given on the role of higher tier studies in the risk assessment process and recommendations

are made for future work.

Reference to this report should be made as follows:

BOXALL, A. BROWN, C. & BARRETT, K., (2001). Higher tier laboratory aquatic toxicity

testing. Cranfield Centre for EcoChemistry research report No. JF 4317E for DETR, 70p.

Opinions expressed within the report are those of the authors and do not necessarily reflect the

opinions of the sponsoring organisation.


3

TABLE OF CONTENTS

FOREWORD ........................................................................................................................................................2

TABLE OF CONTENTS .....................................................................................................................................3

1. INTRODUCTION.......................................................................................................................................4

2. LABORATORY-BASED HIGHER TIER TOXICITY STUDIES.........................................................6

2.1 REALISTIC EXPOSURE SCENARIOS .............................................................................................................62.1.1 Time-to-event analyses....................................................................................................................72.1.2 Variable duration exposure studies ................................................................................................92.1.3 Pulsed-exposure studies..................................................................................................................92.1.4 Inclusion of dissipation processes.................................................................................................152.1.5 Conclusions and recommendations ..............................................................................................20

2.2 STUDIES USING ADDITIONAL SPECIES ......................................................................................................232.2.1 Selection of additional test species ...............................................................................................232.2.2 Analysis of data.............................................................................................................................272.2.3 Current limitations to the approach..............................................................................................292.2.4 Conclusions and recommendations ..............................................................................................29

2.3 ORGANISM RECOVERY ............................................................................................................................322.4 SYSTEM RECOVERY .................................................................................................................................332.5 SENSITIVE LIFE STAGE STUDIES ...............................................................................................................362.6 POPULATION LEVEL STUDIES...................................................................................................................382.7 INDOOR MULTI-SPECIES STUDIES .............................................................................................................41

3. APPLICATION OF HIGHER TIER LABORATORY STUDIES IN ENVIRONMENTAL RISKASSESSMENT....................................................................................................................................................43

3.1 USE OF FORMULATED PRODUCT OR ACTIVE INGREDIENT IN HIGHER TIER STUDIES ..................................483.2 IN-LIFE MONITORING OF CHEMICAL FATE ................................................................................................493.3 REPLICATION...........................................................................................................................................503.4 USE OF STATISTICS ..................................................................................................................................503.5 QUALITY CONTROL .................................................................................................................................52

4. RECOMMENDATIONS FOR FURTHER WORK ..............................................................................52

ACKNOWLEDGEMENTS ...............................................................................................................................55

REFERENCES ...................................................................................................................................................56

APPENDIX A – ORGANISATIONS CONTACTED DURING THE STUDY*...........................................66

APPENDIX B – AVAILABLE SINGLE SPECIES ECOTOXICITY METHODS......................................67


4

1. INTRODUCTION

Registration schemes for plant protection products require registrants to assess the potential

ecological risk of their products using a tiered approach. Standard tests are used at the lower

tiers and clearly defined methodologies are available for assessing the potential environmental

risks (e.g. toxicity exposure ratios or risk quotients). The lower tier single species studies rely

upon a number of assumptions (Crane, 1997):

§ that the response of selected organisms in single-species laboratory tests will correspond

to those of a larger array of organisms in natural systems;

§ that the chosen endpoints in single-species tests are more sensitive than any other at any

level of organisation;

§ that a species shown to be sensitive to a few toxicants will be equally sensitive to a much

wider range of substances.

As none of these assumptions will always hold true, uncertainty factors are incorporated into

the risk assessment process. In order to refine the assessments, it will be appropriate to

critically evaluate the base dataset. However, if these assessments indicate that a compound is

likely to pose a risk to the environment, then impacts can be determined using more

environmentally realistic conditions.

A number of approaches have been used in the past to address concerns arising from the

preliminary assessments, including the use of pond mesocosms (e.g. Neugebaur et al., 1990;

Fairchild et al., 1992; Webber et al., 1992; Farmer, et al., 1995; Giddings et al., 1996 and

1997; Peither et al., 1996; Hanazato, 1998), artificial streams (e.g. Pusey et al., 1994; Ward et

al., 1995; Debus et al., 1996; Dyer and Belanger, 1999; Kostel et al., 1999; Wang et al.,1999;

Belanger et al., 2000), field monitoring studies (Hamilton et al., 1988; Kreiger et al., 1988;

Dyer and Belanger, 1999) and experimental ditches (e.g. Kersting and Van Wijngaarden,

1999; VanGeest et al., 1999). These approaches have a number of advantages over single-

species investigations, including the ability to assess: a) endpoints at higher levels of

biological organisation; b) species interactions; and c) indirect ecological effects. The

approaches also allow the assessment of population and community recovery. Furthermore,

model ecosystems can be used to simultaneously integrate fate processes with exposure and

consequently include the effects of transformation products. Model ecosystems have been

sucesssfully applied to the assessment of pesticides and surfactants (e.g. Belanger, 1992;

Fairchild et al., 1993; Belanger et al., 2000).


5

There are however a number of limitations associated with model ecosystem approaches,

most notably that the results are difficult to interpret and the extrapolation to other situations

can be problematic. Reasons for this include the fact that no-effect concentrations are

dependent on the choice of test concentrations; studies are performed at different times and

necessarily different starting conditions; studies vary in duration; sensitivities may vary across

experimental systems; and measured endpoints are not always consistent or comparable. The

high background variation associated with the studies means that the discriminatory power of

the tests can be low. Moreover, in order to simulate natural conditions and to ensure the

maintenance of basic ecological characteristics and functions (i.e. different trophic levels,

possibilities for organisms to find new habitats, energy input and nutrient cycles), the test

systems need to be of a sufficient size and complexity (Debus et al., 1996). This means that

the studies can be costly.

It has been recognised that other approaches that are intermediate between standard aquatic

toxicity tests and field and microcosm studies might provide valuable data for use in the risk

assessment of plant protection products. The majority of approaches that have been

investigated have involved further analysis of core data or have been performed in the

laboratory using single species. A number of simple multi-species studies have also been

used. The objectives of these laboratory-based higher tier studies include to:

• account for the effects of non-continuous exposure;

• better assess the sensitivity of species to contaminants;

• assess the effects of contaminants on sensitive life stages and populations;

• determine the potential for organisms, populations and systems to recover after exposure

to a pesticide;

• determine indirect effects of compounds on biological communities.

Pesticide regulators and registrants have identified the need to evaluate these different

approaches in terms of what methods are available and how each approach could be used in

the risk assessment process. This study was therefore performed to review current research

and practice with respect to higher tier laboratory studies, to recommend methodological

approaches within a tiered assessment and to identify any further work required to produce

guideline protocols. The report draws upon the results of studies documented in the literature

along with a survey of regulators and laboratories (Appendix A) involved in the testing and

environmental risk assessment of chemicals.


6

2. LABORATORY-BASED HIGHER TIER TOXICITY STUDIES

A range of laboratory approaches has been reported that are intermediate in complexity

between standard ecotoxicity tests and mesocosm/field studies. These include: the

incorporation of realistic exposure scenarios into laboratory studies; studies into effects on

sensitive life stages and populations; studies into recovery of organisms, populations and

environmental systems; species sensitivity studies; and simple multispecies studies. This

section reviews the different approaches used. It is assumed that a full and critical evaluation

of the baseline fate and effects data package precedes any move to higher-tier studies.

2.1 Realistic exposure scenarios

Preliminary acute risk characterisation methods typically use the initial predicted

environmental concentration (PEC) along with the results from standard ecotoxicity studies.

However, the method of entry and rate at which a compound dissipates in the environment

may have a significant impact on the relevance of standard ecotoxicity studies used. For

example, for compounds that dissipate rapidly, a comparison of the initial exposure

concentration with data from standard ecotoxicity tests may result in an over-estimate of the

potential effects. In such cases it may be appropriate to calculate a time-weighted average

(TWA) PEC which takes into account the dissipation of the chemical (Campbell et al., 1999).

The use of TWA PECs is most applicable to chronic toxicity studies on compounds whose

mode of action is time dependent or reversible. However, where dissipation is very rapid it

may also be appropriate to apply the approach to acute data.

In some instances, the use of TWA PECs may be inappropriate and the approach generally

results in an over-estimation of toxicity. For example, if organisms are exposed to a

contaminant for a short period, there may not be a potential for an effect, even when

compounds are shown to be toxic in standard studies. Pulsed exposures may result in

cumulative individual effects; for example, the first pulse may weaken an organism by

causing cumulative damage such as an irreversible interaction with a receptor site.

Standard studies also tend to be performed using well characterised media (e.g. reconstituted

water). The differences between standard test media and natural media may mean that the

bioavailability of a compound is either reduced or increased in natural systems compared to

standard test systems.


7

A number of approaches have therefore been proposed that account for differences in

exposure scenarios, dissipation and bioavailability between natural systems and standard

toxicity studies (Table 7). These include:

• time-to-event analyses;

• pulsed exposure studies;

• short-term exposure studies;

• simulation of spray drift;

• suspended sediment/water studies;

• studies using natural water.

The rationale behind these different approaches and the methodologies that have been used

previously are outlined below.

2.1.1 Time-to-event analyses

Both exposure concentration and duration determine the effect of a toxicant. The probability

of an effect increases if either of these variables is increased up to a point where 100%

mortality is reached. However, the main approach to quantifying the effects of a toxicant

currently focuses on the intensity of exposure (e.g. determination of LC50) and as the tests are

performed over a set time period, they neglect the effect of exposure duration.

Time-to-event (TTE) approaches characterise the toxicity of a compound by modelling the

exposure concentration and duration of exposure for a specified event (e.g. Miller, 1981; Cox

and Oakes, 1984; Marubini and Valsecchi, 1995; Allison, 1995; Newman and McCloskey,

1996). Time to event analyses have been used since the earliest days of toxicity testing. The

approaches draw upon experimental designs where groups of exposed individuals are

monitored through time and the intervals to some event, usually death, are recorded for each

individual. The resulting data can then be analysed using a range of approaches, including:

Kaplan-Meier methods; life table methods; the semi-parametric Cox proportional model; or

the exponential Weibull Gempertz model (e.g. Hendley and Giddings, 1999; Table 1). Each of

the approaches can be used to predict time to events and to test the significance of a co-variate

(e.g. exposure concentration) on a time to event.


8

Table 1 Time-to-event methods applicable to toxicity test results

(adapted from Hendley and Giddings, 1999)

Method Basis

Kaplan-Meier method Non-parametric method; statistical differences between

classes can be tested using several non-parametric tests (e.g.

log rank; Wilcoxon).

Life tables Non-parametric method.

Cox proportional model Semi-parametric method; assumes no underlying distribution

for the mortality curve but that hazard remains proportional

among groups such as males and females; uses maximum

partial likelihood estimation to fit data.

Exponential Weibull Gompertz Fully parametric method with specified underlying

distributions for mortality and specific functions describing the

influence of some covariate; uses maximum likelihood

estimation to fit data to specified model.

The use of TTE methods provides a number of advantages over traditional methods of

analysis for ecotoxicity data. These include:

• exposure duration is explicitly included in the assessment;

• TTE information enhances the power of statistical tests as more data are extracted per test

treatment;

• conventional endpoints (e.g. LC50) can still be estimated;

• the effects of covariates can be measured and included in predictive models;

• results can be incorporated directly into ecological, epidemiological and toxicological

models.

Whilst summary data from standard toxicity tests are usually reported and used in the risk

assessment process, the use of standard methods (e.g. those recommended by OECD) means

that additional data will have been recorded throughout the duration of the test. It may often

be possible to use this supplementary data for TTE assessments without the need for further

experimental study (Campbell et al., 1999).


9

2.1.2 Variable duration exposure studies

For compounds that dissipate rapidly, the duration of exposure to a pesticide may be very

short. Studies can be performed with varying exposure intervals to determine the impact, if

any, of such short exposures (e.g. Maund et al., 1998). For example Gammarus pulex were

exposed to lamda-cyhalothrin for either 1, 3, 6, 12 or 96 h (Maund et al., 1998). After

exposure, organisms were transferred to clean water and effects were recorded at 96 h. There

was a highly significant relationship between effects and duration of exposure. The effect

concentration with exposure for 1h was 18 times larger (less toxic) than that for 96-h exposure

(Maund et al, 1998). Similar results have been obtained for Hyallela azteca using the

pyrethroid cypermethrin (Maund et al., 1998).

2.1.3 Pulsed-exposure studies

Pesticide contamination typically occurs in pulses as a result of spray drift, overland flow and

drainage inputs (e.g. Solomon et al., 1996). Continuous exposure, as used in standard tests,

may therefore not provide a true estimate of the ecotoxicity of compounds and could result in

both over- and under-estimations of ecotoxicity. For example, pulsed exposure will result in a

lower nett dose. Conversely, organisms exposed to multiple pulses may be more likely to

survive than organisms exposed continuously. Possible reasons for this include:

• organisms are able to detoxify or depurate any accumulated test compound during the

exposure interval (e.g. Wright, 1976; Breck, 1988);

• induced individual tolerance – the first pulse may strengthen survivors through

acclimation or induction of detoxification enzymes (e.g. ; Parrott et al., 1995; Burnison et

al., 1996);

• individual selection – weaker individuals may be removed by the first pulse resulting in a

selection of more robust individuals and an apparent reduction in responses to future

exposure;

• ecological recovery – after exposure, an impacted population may or may not recover,

depending on the characteristics of the species that are affected (methods to assess system

recovery are described in Section 2.4).

Data from standard toxicity studies can be used to provide an initial assessment of the effects

of intermittent exposure to a contaminant (Handy, 1994). A number of approaches are

available: 1) estimation of the average concentration of toxicant on the basis of an exposure

profile and reading off toxicity data from existing continuous exposure data; 2) the cumulative


10

mortality from a single pulse can be assessed and the remaining pulses assumed to be

additive; and 3) the construction of lethality threshold curves. These approaches assume that

the toxicity of an intermittent event is the same as a continuous exposure test of equivalent

dose and that survival time will be longer than the periodicity between exposure peaks.

The effects of intermittent exposure can also be assessed using ‘custom-designed’ tests. A

number of studies have investigated the effects of pulsed exposure on a range of organisms

(including daphnids, amphibian tadpoles, caddisfly, chironomids and fish) for a range of

pesticides (including fenoxycarb, fenitrothion, tebufenozide, chlorpyrifos and fenvalerate)

(e.g. Hansen and Kawatski, 1976; Wright, 1976; Abel, 1980; Pascoe and Shazili, 1986;

McCahon and Pascoe, 1988; Handy, 1994; Holdway et al., 1994; Berrill et al., 1995;

Kallander et al., 1997; Hosmer, et al., 1998; Pauli et al., 1999; Naddy et al., 2000).

Two approaches have generally been used, namely static exposure studies or flow-through

studies. Virtually any temporal exposure pattern can be created using a static renewal toxicity

test. Static renewal studies are particularly useful for generating ‘square pulses’ (i.e. constant

exposure for a set duration). This situation is most relevant when real-world exposure is

expected to be dominated by water flow – for example, where a pulse of chemical travels

quickly down a small stream.

The use of flow-through exposure systems allows pulsed exposure to be more realistically

simulated with gradual changes in concentration. This approach has been used to simulate the

effects of degradation of a herbicide on Selenastrum capricornutum (Grade et al., 2000).

Three compounds were used in the simulation, namely the parent compound and two major

metabolites and the pulses of each mimicked the fate of the compound in the natural

environment.

The effects of pesticides studied depend on a number of variables including pesticide

concentration, pulse duration, interval between pulses and pulse frequency (e.g.Table 2).

Pulsed exposure experiments can provide information on the potential for organisms to

recover, the development of resistance to pesticide exposure and any latent effects.

Comparison of the results from standard and pulsed exposure tests indicate that the standard

tests can be either over-protective (e.g. Hosmer et al., 1998) or under-protective (e.g. Hansen

and Kawatski, 1976).

As a wide range of possibilities exist for the intensity and duration of exposure to pulsed

release of contaminants, no testing regime can cover all pulse discharge scenarios.


11

Mechanistic models have therefore been proposed and tested for predicting the toxicity of

time-varying exposures (e.g. Mancini, 1983; Wang and Hanson, 1985; Breck, 1988; Meyer et

al., 1995; Table 3). The approaches used include:

• concentration x time models;

• Mancini uptake-depuration models which are based on simple first-order kinetics of

uptake and depuration;

• damage-repair models.

Investigations into the effects of monochloroamine on rainbow trout and common shiners

demonstrated that concentration x time and Mancini type models can predict an LC50 to

within ± 50% (Meyer et al., 1995). The Mancini model performed better for longer term

pulse exposures.


12

Table 2 The effect of exposure frequency on the lethal toxicity of intermittent pollution events (adapted from Handy, 1994)

Species Pollutant Duration of exposure:recovery Number ofexposures

Percent oftime

exposed

LT50 LC50 (µg l-1) Reference

Mosquitolarvae

permethrin 2h:0h1h:6h

continuous2

10025

- 180250

Parsons andSturgeoner, 1991

Rainbow trout sulphuric acid 96h:0h12h:12h24h:24h

continuous42

1005050

- 65.1-84.952.7-57.846.8-55.8

Curtis et al., 1989

Brook trout acid/Al pH4.4 24 d:0d2d;4d

continuous4

10033.3

- 500700

Siddens et al.,1986

Brook trout acid/Al pH 4.9 24 d: 0d2d: 4d

continuous4

10033.3

- 460400

Siddens et al.,1986

Rainbow trout copper 78d:0d4.5 h: 19.5 h

continuous78

10018.75

- 10272

Seim et al., 1984

Rainbow trout ammonia 96h:0h12h:12h6h:6h6h:18h

continuous484

1005050

33.3

- 29617622199

Thurston et al.,1981

Rainbow trout sulphuric acid 90d:0d4d;4d

continuous11

10050

38d84d

- Curtis et al., 1989

Harlequin industrial effluent -300 min:300 min

continuousmultiple

10050

225 min262 min

- Alabaster andAbraham, 1965

Rainbow trout ammonia continuouscontinuous with 1 h fluctuationscontinuous with 2 h flucutations

100100100

>700 min>700 min

370

- Brown et al., 1969


13

Table 2 Continued

Species Pollutant Duration of exposure:recovery Number ofexposures

Percent oftime

exposed

LT50 LC50 Reference

Rainbow trout zinc contiunuouscontinuous with 2 h fluctuationscontinuous with 4 h fluctuations

100100100

2400 min2340 min2960 min

- Brown et al., 1969

Rainbow trout thiocyanate continuous12 h: 24 h18 h: 24 h24 h: 24 h48 h: 24 h72 h: 24 h

continuoussinglesinglesinglesinglesingle

100 ------

1151180018721222611243

Kevan and Dixon(1996)


14

Table 3 Mechanistic models for assessing the effects of intermittent exposure to contaminants

Model Generalized model Inputs

Concentration x timed

ywm tCA .=

Am = a species specific constant for m% mortality (µg/l/h)

Cw = toxicant concentration in exposure water (µg/l)

td = time to death (exposure time required for m% mortality; h)

y = a power term

Mancini uptake-depuration model

soadepusoa

CkCKdt

dC y

w.. −=

Cw=toxicant concentration in exposure water (µg/l)

Csoa= concentration of toxicant at site of action (µg/g)

Ku=uptake rate constant (h-1)

Kdep=depuration rate constant (h-1)

t=exposure time (h)

y=a power term

Breck damage-repairmodel

+

−+=

−

−

50,

.

ln1

ln)ln(..1

lnL

damtk

repw

DK

ke

CydM

M repM=the observed mortality proportion

d=slope of logit response as a function of ln (Cw)

Cw=toxicant concentration in exposure water (µg/l)

Kdam=damage rate constant (h-1)

Krep=repair rate constant (h-1)

DL,50=lethal damage at 50% mortality

t=exposure time (h)

y= a power term


15

2.1.4 Inclusion of dissipation processes

Once a compound has entered the aquatic environment, its distribution and hence potential to

cause adverse effects will depend on a number of factors including the potential for abiotic

(e.g. hydrolysis, oxidation, photolysis) and biotic (aerobic and anaerobic) degradation and the

partioning behaviour of the compound to sediment and air. A number of studies have

therefore investigated the incorporation of one or more of these processes into laboratory

ecotoxicity studies. The processes simulated have included: spray drift, effects of suspended

sediment, effects of pesticide entering the environment in runoff and the effects of bed

sediment (e.g. Maund et al., 1998; Kusk, 1996; Table 4)

By adding sediment to the standard test system, the degradation and adsorption processes

occurring in the environment can be simulated (e.g. Hamer et al., 1992). Studies using the

hydrophobic synthetic pyrethroids, lamda cyhalothrin (log Kow 6.8), esfenvalerate (log Kow

6.22) and fenvalerate (log Kow 5.0) have demonstrated that the presence of soil or sediment

in the test system results in a significant reduction in the observed effect of a pesticide (Table

4). This observation can be explained by the fact that very little pesticide (i.e. <1% for lamda-

cyhalothrin) is available in the water column due to partitioning of the compound from the

aqueous phase to sediment (Hamer et al., 1999). The addition of sediment had less effect on

the hydrophilic compounds isoproturon (log Kow 2.5) and pirimicarb (log Kow 1.7).


16

Table 4 Changes in observed toxicity caused by a range of modifications to water-only exposure studies (Clark et al., 1989; Maund et al., 1998; Shillabeer

et al., 2000; Schulz and Liess, in press)

Reduction factor in toxicity or BCF relative to water-only study

Compound/test species Simulation ofspray drift

Simulation of spraydrift and suspended

sediment

Application to soilprior to addition of

water

Simulation ofsuspendedsediment

Simulation of bedsediment (not

stirred)

Simulation ofsediment (stirred)

Simulation ofeffects of soil

lambda cyhalothrin

Daphnia magna 3 40 175 - 120-140 81-280 -

glyphosate

Daphnia magna - - - 3.3 - - -

esfenvalerate

Daphnia magna

Lepomis macrochirus

- - - - 3.0

5.0

- -

fenvalerate

Limephilus lunatus

grass shrimp

- - - 10-100

-

- -

9,700

-

pirimicarb

Daphnia magna (BCF) - - - - 0.7 - -

isoproturon

Scenedesmus subspicatus - - - - - 2 -

herbicide 1

Selenastrum capricornutum

herbicide 2

Navicula pelliculosa

-

-

-

-

-

-

-

-

>17

>900

-

-

-

-


17

Water/sediment systems have also been used to assess the impact of herbicides on

macrophytes. For example, Forney and Davis (1981) investigated the effects of herbicides on

aquatic plants. Studies with atrazine in soil/water systems demonstrated that atrazine

associated with the soil was not toxic because submerged plants lack a transpiration stream

whereas atrazine in the water column exerted a toxic effect.

The studies using the hydrophilic carbamate pesticide, pirimicarb (log Kow 1.7) have

demonstrated that whilst the presence of sediment had no effect on concentrations of the study

compound in the water column, the bioconcentration factor for Daphnia magna was reduced

(Kusk, 1996). This observation was possibly due to the effects of dissolved organic carbon

released from the sediment.

In the natural environment, water contains dissolved or colloidal organic matter (DOM).

These substances have been shown to interact with organic compounds resulting in effects on

bioavailability because only freely dissolved compounds are generally assumed to be

accumulated by organisms. Previous studies using pesticides, surfactants and polycyclic

aromatic hydrocarbons have demonstrated that the presence of dissolved organic carbon

reduces accumulation and toxicity (e.g. Landrum, et al., 1985; McCarthy et al., 1985; Black

and McCarthy 1988; Kukkonen and Oikari, 1991; Kukkonen and Pellinen, 1994; Oris et al.,

1990; Weinstein and Oris, 1999). However, at low concentrations of DOM (i.e. less than 10

mg/l) bioaccumulation may be enhanced (e.g. Haitzer et al., 1998).

In order to simulate the effects of natural concentrations of dissolved organic matter on

ecotoxicity, studies have also been performed using water samples collected in the field (e.g.

Isnard et al., 2000; S. Marshall, personal communication). Studies on a range of compounds

have demonstrated that generally the apparent toxicity of chemicals is only slightly reduced in

tests using natural waters (Isnard et al., 2000; Table 5). This suggests that the use of

laboratory water does not necessarily introduce a significant source of error. The use of this

type of study as a higher tier test may not therefore be particularly useful.


18

Table 5 Mean ratios of EC50s obtained using tests on 15 natural waters to EC50s obtained using

ISO standard media (taken from Isnard et al., 2000)

Log Kow Raphidocelis

subcapitata

microplates

Raphidocelis

subcapitata

erlens

Brachionus

calyciflorus

Daphnia

magna

zinc sulphate na 5.7 3.0 2.1 1.2

4-nonylphenol 5.92 0.6 0.7 1.2 0.9

phosalone 4.30 0.7 1.2 1.1 1.7

pentachlorophenol 5.12* 2.7 2.1 1.2 3.1

2,4,5-trichloroaniline 3.45 1.9 2.2 2.1 1.0

Mean 2.3 1.8 1.5 1.6

Standard ecotoxicity studies usually require that the test concentration at the end of the study

is 80% of the starting concentration. For compounds that are volatile or which are degraded

abiotically (i.e. via hydrolysis) this can mean that the test system is covered and that a flow-

through test system is required or that the test solution is replaced regularly. Thus the impact

of abiotic degradation processes on toxicity could readily be assessed using a static exposure

system with no replacement of test solution.

The test matrix used in dissipation studies (e.g. sediment or natural water) should be selected

based on available data from fate studies and modelling work and should represent a realistic

‘worst case’. Two types of test matrix could be used, namely an artificial matrix (e.g. artificial

sediment) or a natural test matrix. Ideally, test sediment should be uncontaminated by

chemicals in toxic amounts, free of biota, and possess characteristics similar to the sediment

type of interest (Suedel et al., 1996). Formulated sediments have therefore been developed to

simulate field-collected sediments (e.g. Suedel and Rogers, 1994). Studies with copper

(Suedel et al., 1996) indicate that formulated sediments can be used as reference sediments.

The selection of the test matrix type will be dependent on the objectives of the higher tier test.

Natural materials will mimic actual conditions in the field more closely, whereas artificial

matrices are standardised and thus facilitate comparison of results with those from similar

studies. At present, higher tier regulatory studies seek to increase the realism of the test

system and the use of natural water and sediment materials might be preferred on this basis.

Single test systems should be based on sediment with realistic worst-case organic carbon

contents (1-2%), whereas studies with two sediment types (as for fate water-sediment

experiments) allow assessment of effects under a broader range of conditions. Artificial

sediments are continually improving and studies with such matrices are routinely accepted.


19

If a natural test matrix is to be used, then it should be obtained from sites that are

uncontaminated and that have not previously been exposed to the test substance. It must also

be able to support the survival of the test organisms. The sampling methodology for both

water and sediment should be considered and involve minimum disturbance. If possible, the

test matrix should be initially characterised on-site in terms of dissolved oxygen and pH. The

matrix should be used as soon as practicable after collection. On return to the laboratory,

samples should be fully characterised and the characterisation should also reflect the objective

of the study. For example, if the effects of humic acids in the water column are being

investigated, then the dissolved organic carbon concentration of the test solution should be

determined.

2.1.4.1 Time-varying toxicokinetic models

The results of modified exposure studies can be used in conjunction with toxicokinetic models

to assess toxicant effects under non-steady state conditions. A range of approaches is

available (Table 6), including: compartment based (these describe the movement of toxicants

between compartments); physiologically based (describe accumulation, elimination and

distribution); and bioenergetic based (describe accumulation and loss in terms of the

organism’s energy requirements). To apply the models, information is required on body

residue levels and toxic effects. This is a neglected area in ecotoxicology and further work is

required before it can be incorporated into the risk assessment process.

Table 6 Toxicokinetic models that can be used in the assessment of results from modified

exposure studies

Model Model type Major inputs Reference

FGETS Bioenergetic-

based

MW, MP, Log Kow, fish

species, fish weight, water

temperature

Barber et al., 1988

PULSETOX Compartment

based

Kow, MW, HLC, organism

volume, lipid fraction, BCF,

uptake and depuration rates,

test concentration, water and

toxicant flow rate

Landrum et al., 1992;

Hickie et al., 1995

DEBtox Bioenergetics

based

Concentration of compound

in solution; uptake and

depuration rates

Kooijman and

Bedaux, 1996


20

2.1.5 Conclusions and recommendations

A wide range of approaches is available for assessing the impact of fate processes on the

toxicity of a substance. These include:

1. analysis of core data to assess the time to effect – this approach is most appropriate for

chemicals that enter the environment in pulses or which dissipate rapidly;

2. standard tests without maintenance of test substance concentration – this approach is

appropriate for volatile compounds or compounds that readily hydrolyse;

3. pulsed exposure studies – these are appropriate for chemicals that enter the environment

in pulses;

4. simulation of spray drift;

5. simulation of the effects of sediment or suspended sediment – for compounds that readily

partition to solids and/or which are readily biodegradable; and

6. tests using natural water – appropriate for highly hydrophobic substances that may

associate with dissolved organic carbon.

Differences between the results of standard and modified exposure studies can be used to

demonstrate the influence of a particular fate process or dissipation of a compound. Generally,

the results should be expressed as an initial toxicant concentration and then compared to an

initial PEC. Care should be taken to ensure that the process studied has not been incorporated

into the PEC calculation. The uncertainty factors associated with the preliminary risk

characterisation should be used.


21

Table 7 Modified exposure studies used for the assessment of chemicals that were identified in the survey

Test title Species Culture source Organism age Brief methodology Rationale for test QA Procedures +statistics

References

Modified exposure Daphnia magna Laboratory < 24 h old Acute 48 h in awater/sedimentsystem

Incorporatesadsorptionprocesses

GLP with EC50

determinationConfidential report

Modified exposure Daphnia magna Laboratory <24 h old 21 d reproductionin a water/sediment system


GLP with EC50 andNOECdetermination

Confidential report

Modified exposure Daphnia magna Laboratory Day 6 gravid 21 d reproductionin a water/sediment system

Incorporatesadsorption process


Confidential report

Modified exposure Oncorhynchusmykiss

Laboratory Juvenile Acute 96 h in awater/sedimentsystem


GLP with LC50


Modified exposure Lemna minor Laboratory Mature plant 14 d exposure andrecovery in awater/sedimentsystem

Incorporatesadsorptionprocesses andrecovery potential


Confidential report

Modified exposure Glyceria fluitans Laboratory Emergent seedling 14 d exposure andrecovery in awater/sedimentsystems



Confidential report

Modified exposure Glyceria maxima Laboratory Mature plant 14 d exposure andrecovery in awater/sedimentsystems



Confidential report


22


References

Modified exposure Gammaru pulex Laboratory 2 months old Mortality over 48 ina sediment watersystems

Incorporatesadsorptionprocessess

GLP test, duplicatevessels at testconcentration

Confidential report

Modified exposure Daphnia magna Laboratory neonates Mortality over 48 hin a river watersystem

Simulates toxicityunder realconditions i.e.presence of humicacids


Confidential report

Modified exposure Daphnia magna Laboratory neonates Mortality over 48 hin a river watersystem

Simulates toxicityunder realconditions i.e.presence of humicacids


Confidential report


23

2.2 Studies using additional species

There is no species that is ‘most sensitive’ to all classes of contaminants (Cairns, 1986; Sloof

et al., 1986; Sloof and Canton, 1993). Species can vary in their sensitivity to chemicals by

orders of magnitude. For example, variation in algal EC100s for 18 compounds and 13 algal

species ranged from less than an order of magnitude to three orders of magnitude (Blanck et

al., 1994). Narrow variation spans were observed for Scenedesmus obtusiusculus, Chlorella

emersonii and Selenastrum capricornutum, whereas wider spans were found for Monodus

subterraneus, Raphidonema logiseta and Synechococcus leopoliensis which were sensitive to

some compounds and insensitive to others. Similar results have been recorded elsewhere. The

preliminary assessments of pesticide risk to the aquatic environment therefore incorporate

factors of either 10 or 100 to cover uncertainties regarding the relative sensitivities of the

standard test organisms compared to the range of species that are likely to be present in the

field (as well as uncertainties in other aspects of the test system).

One approach to reducing these uncertainties is to test additional organisms in order that the

uncertainty factors can be reduced and/or a distribution of sensitivity can be defined

(Campbell et al., 1999). The testing of additional organisms also generates more ecological

information for use in site-specific analyses. A wide range of test methods is available for a

large number of species, e.g. algae, protozoans, vermes, molluscs, crustaceans, insect larvae

and fish (Appendix B) and a number of additional species have been used in the assessment of

chemicals for regulatory purposes (Table 11). Generally, these approaches assess the acute

effects of a compound.

2.2.1 Selection of additional test species

A number of approaches have been proposed for selecting additional test species. Some of

these are outlined below. The approach used will depend on a number of factors including the

nature of the compound being tested and the availability of data for compounds with a similar

mode of action.

2.2.1.1 Approach proposed by the Aquatic Dialogue Group, 1994

For compounds that are not selective to organisms, the Group proposed that the test species

should include at least two species of fish, one invertebrate and one aquatic plant (macrophyte

or algae) plus four other species. The other four test species should be selected from the

group(s) that is/are shown to be the most sensitive after the initial studies have been


24

completed (Aquatic Dialogue Group, 1994). If chronic studies are to be performed, it is

recommended that one fish, one invertebrate and one plant be used. A further organism

should be selected from the most sensitive group identified in acute tests (Aquatic Dialogue

Group, 1994). Opportunities for extrapolation should not be overlooked for compounds with a

strong similarity to an existing substance. For example, the number of additional organisms

required to show the range of species sensitivity might justifiably be reduced on the basis of

data for a similar compound.

2.2.1.2 Approach proposed by the USEPA, 2000

Recently, the USEPA (2000) have recommended that additional data should include:

invertebrate acute and chronic tests, sediment toxicity tests, rooted plant testing and

amphibian testing. The increase in the number and type of invertebrate toxicity tests reflects

an attempt to provide some consideration of the range of potential responses. In assessing

impact on aquatic invertebrates, risk assessments generally focus on Daphnia magna, a

parthenogenic, short-lived crustacean. The following should be considered when selecting

additional species: 1) different and often variable life spans; 2) unpredictable lengths of life

stages and different metamorphic stages; 3) indeterminate growth in some species; and

4) significant differences in adipose stores (USEPA, 2000). Examples of additional species

for invertebrate acute toxicity tests could include stoneflies and amphipods, whereas chronic

tests should address effects on sexually reproducing invertebrate species including copepods

and chironomids. The addition of an acute amphibian test reflects the need to consider effects

on amphibian specie.

2.2.1.3 Approach proposed by Vaal et al., 2000

The sensitivity of individual species is primarily dependent on the compound’s mode of

action and species from the same taxonomic group (either phylum or class) generally respond

to contaminants in a similar manner (Vaal et al., 2000). For chemicals with a non-polar mode

of action, variation in species sensitivity is small. However, for compounds that are either

reactive or have a specific mode of action, the variation can be as large as 105 – 106. The

following testing strategy is therefore recommended:

1. test one species in screening programs in order to prioritize substances;

2. classify compounds for the effects assessment procedure as early as possible using

toxicologically-based schemes (e.g. Verhaar et al., 1992; Russom et al., 1997);

3. if a chemical belongs to class I (non-specific narcotics), use QSARs for baseline toxicity

to estimate PNEC;


25

4. if a chemical belongs to class II (polar narcotics), III (reactive compounds) or IV

(specific-acting compounds), use at least three acute toxicity data for preliminary

assessment and take care when applying assessment factors for classes III and IV;

5. if a chemical belongs to class II, III or IV use at least five chronic toxicity data in a

refined effects assessment.

2.2.1.4 Mode of action approach proposed by ECOFRAM (Hendley and Giddings, 1999)

Based on an understanding of the mode of toxic action of a compound, it may be possible to

identify and group sensitive and less sensitive organisms. This allows the testing strategy to

be focused on groups at high risk. For example, Cuppen et al. (2000) recommended the use of

non-arthropod macroinvertebrate species at an advanced stage of the risk assessment of

fungicides. This is based on the assumption that the spectrum of effects of fungicides in

aquatic systems is similar to results obtained for carbendazim and pentachlorophenol.

Similarly, tests on insects would be recommended for substances with an insecticidal activity.

It should be noted that mode of toxic action is often not well understood, particularly for new

compounds.

Whilst the mode of action of a pesticide is an important criterion for grouping of organisms,

habitat, reproductive strategy and life cycle may also affect the sensitivity of species.

ECOFRAM therefore proposed a possible system that could be used for grouping test species

which is shown in Figure 1.

Figure 1 Possible groupings for toxicity sensitivity datasets (Hendley and Giddings, 1999)

DATA

Freshwater Saltwater

Fish Athropods OtherInverts

Phytoplankton Multicellularplants

Crustacea Insects Greenalgae

Bluegreenalgae


26

2.2.1.5 The use of a test battery of additonal species

Recent studies (Girling et al., 2000; Pascoe et al., 2000) have assessed the suitability of a

battery of ecotoxicity tests that evaluate the sublethal and lethal effects of chemicals on algae,

protozoans, rotifers, crustaceans and dipterans (Table 8). The aim of the studies was to

identify tests that could be adopted in the risk assessment process. Comparison of the results

from the tests with those from stream and pond mesocosm studies for lindane, 3,4-

dichloroaniline, atrazine and copper demonstrated that lowest observed effect concentrations

obtained in the laboratory were generally similar to results obtained using pond mesocosms

and artificial streams (i.e. within a factor of 6). The exception to this was 3,4-dichloroaniline

(DCA) where the pond mesocosm was 200 times more sensitive. Similar studies on

isoproturon (Traunspurger et al., 1996) demonstrated that single species tests using a range of

organisms were more sensitive than a microcosm study.

Table 8. Test organisms and endpoints studied by Girling et al. (2000)

Taxon Species Endpoint

alga Chlamydomonas reinhardi

Scenedesmus subspicatus

Euglena gracilis

growth inhibition-flow throughgrowth inhibition-staticeffective photosynthesis rate

growth inhibition-staticeffective photosynthesis rate

growth inhibition in lightgrowth inhibition in dark

protozoa Tetrahymena pyriformis growth inhibition

rotifera Brachionus calyciflorus mortalityswimming activityfeeding activitypopulation growthfull life cyclepartial life cyclemultigeneration life cycle

crustacea Gammarus pulex juvenile mortalityjuvenile immobilisationneonate growthprecopula separationpopulation growth

diptera Chironomus riparius second instar mortalitysecond instar growthpercentage egg hatchabilitylife cycle: increasing hatching timelife cycle: slower larval developmentlife cycle: reduced emergencelife cycle: delayed emergence


27

2.2.1.6 Guidance Document on Higher Tier Risk Assessment (HARAP)

The HARAP Guidance Document (Campbell et al., 1999) suggests that for compounds that

are not selective to aquatic organisms, eight species should be tested as a minimum to

describe the distribution of sensitivities of aquatic organisms. In cases where it is known that

a specific group of organisms is particularly sensitive then the additional test species should

be selected from the most sensitive group. However, if fish are the most sensitive group,

fewer test species are required and five species are probably sufficient. This is for animal

welfare reasons and because distributions of sensitivity are usually narrow for fish. By using

at least eight additonal test species (five for fish), there is a high probability that a good

estimate (i.e. within an order of magnitude) of the most sensitive endpoint will be obtained

(Campbell et al., 1999). The use of eight additional species also allows a probabilistic

approach to be used for effects characterisation (Aquatic Dialogue Group, 1994).

2.2.2 Analysis of data

One of the objectives of laboratory ecotoxicity testing is to assess the impact of a chemical on

natural ecosystems. Rather than using uncertainty factors, the data from single species studies

can be extrapolated to field conditions using either regression techniques or distribution

models (Roman et al., 1999). A wide range of approaches are available that are based on the

same general principle (i.e. they assume that experimental ecotoxicity data fit a given

distribution), but differ in the statistical distributions that they use (e.g. Blanck, 1984; Slooff

et al., 1986; Kooijman, 1987; Van Straalen and Denneman, 1989; Suter et al., 1995; Solomon,

1996).

The outputs from the methods vary and include:

• hazardous concentrations – i.e. a hazardous concentration for a specified percentage of

species (HCx where x is the specified percentage; Van Straalen and Denneman, 1989;

Wagner and Lokke, 1991);

• concern levels – levels that cause an environmental effect at least 95% of the time

(USEPA, 1984);

• final chronic values – a threshold concentration for unacceptable effects that is calculated

from the lowest of three chronic values (the final chronic value, the final plant value and

the final residue value). The final chronic value is calculated from the NOEC values for at

least eight families (Stephan et al., 1985). There may be problems with availability of

internationally accepted methods for this number of chronic tests (e.g. for invertebrates).


28

There are significant challenges in extrapolating laboratory results to the field, particularly as

many test organisms are selected for practical reasons (e.g. ability to grow well in the

laboratory) rather than for direct relevance to field conditions. Nevertheless, the results of

single species distribution methods have been compared with those from multispecies studies

(e.g. Okkerman et al., 1993; Van den Brink et al., 2000; Table 9). In most cases, the

extrapolation procedures lead to lower ‘safe’ values than NOECs from multispecies studies

and demonstrate that field effects can be predicted if uncertainties related to mode of action

(e.g. species groups tested and exposure regimes) are accounted for. The studies have

demonstrated that the larger the number of species, the smaller the difference between the

results of multispecies studies and the statistical techniques.

Table 9. Comparison of multispecies NOECs with results obtained using extrapolation

techniques on single-species toxicity data (Okkerman et al., 1993; Van den Brink et al., 2000;

Girling et al., 2000). All values are µµg l-1

Compound MS NOEC Aldenberg

and Slob

(1993)

50%

Aldenberg

and Slob

(1993)

95%

Wagner

and Lokke

50%

Wagner

and Lokke

95%

SSD (method

not reported)*

95%

USEPA

lindane 0.22 0.12 0.13

dichloroaniline 1.0-12 120 120 0.067

atrazine <3.0-5 0.82-2.2 0.12 0.87-2.1 0.17 0.15

copper 1.1 0.28 0.30

isoproturon 2 0.63

atrazine 20 13

azinphos-methyl 0.2 0.037

diflubenzuron 0.3 0.18

chlorpyrifos 0.1 0.044

diflubenzuron 0.1 - - - - 0.001

parathion 0.1 0.013 0.00013 0.011 0.00023 0.002

azinphosmethyl 0.25 0.15 0.085 0.077 0.018 0.01

methylparathion 0.1 - - - - 0.024

pentachlorophenol 20 3.2 0.53 2.6 0.66 0.32

1,2,4-trichlorobenzene 57 44 1.1 39 1.7 19

trichloroethylene 2.8 3100 65 3,100 97 130

trifluralin 0.5 - - - - 0.2

dichlobenil 5.0 - - - - 7.8

permethrin 0.023 - - - - 0.066

toxaphene >1.5 0.003 <0.0001 0.003 <0.0001 0.0025

* Species sensitivity distribution (Van den Brink et al., 2000)


29

There is conflict over which value from a range of species sensitivities is most appropriate to

protect the aquatic environment. Van den Brink et al. (2000) suggest that the 5th percentile

value is a generically applicable criterion that is protective for exposed communities (the

higher the number of species tested, the lower the difference between the NOECecosystem and

the 5th percentile). Whilst the 5th percentile is therefore the accepted norm, previous studies

with pyrethroids (Giddings, 1997) and atrazine (Solomon et al., 1996) have proposed that the

10th percentile effect concentration is adequate. It is recommended that the 5th percentile value

continues to be used.

More recently, the use of bootstrapping to derive hazardous concentrations has been

investigated (Newman et al., 2000). The approach involves random sampling (with

replacement) of an ecotoxicity dataset for a compound to create a resampling dataset (e.g of

100 observations). The resulting dataset is then ranked and the 5th percentile value is selected

as the HC5. The resampling is performed a number of times (e.g. 10,000) to produce a number

of HC5 estimates. These are then ranked and the value corresponding to 50% is taken as the

best estimate of HC5. Estimates at 2.5% and 95% are used as the 95% bootstrap confidence

limits. Large datasets are required to derive reliable HC5 estimates because knowledge of the

mode of action of the substance and sensitive species is not accounted for.

2.2.3 Current limitations to the approach

Whilst the use of additonal test species provides useful data, there are a number of limitations

with the approaches compared to standard ecotoxicity studies. For many species, laboratory

culturing procedures are not available and the test species will need to be collected from the

field. This can mean that the age and quality of the organisms cannot be guaranteed (Persoone

& Jansen, 1993). Where possible, organisms collected from the field should be taken from an

uncontaminated site that has not been exposed to the test substance or chemicals with a

similar toxic mode of action. In some instances, the test organism may not be available

throughout the year. Therefore, when designing such studies, either species should be

selected that are available all of the year or a wider range of possible test animals should be

available so that a selection can be made at any time. The use of Geographical Information

Systems (GIS) to identify the proportion of a particular species and habitat exposed to a given

pesticide could also assist in the identification of additonal species for testing (USEPA, 2000).

2.2.4 Conclusions and recommendations

Studies to date indicate that if additional test data are to be used in the risk assessment

process, then a minimum of eight additional species should be tested. If available data


30

indicate that a particular group of organisms is sensitive, then the testing should focus on

representative species from this group. If the most sensitive group is fish, then five additional

tests are probably adequate.

Data on toxicity of a substance to additonal species can be incoporated into the risk

assessment process in two ways:

1. the uncertainty factor used in the risk assessment process can be reduced by up to an order

of magnitude, depending on the data produced (Campbell et al., 1999); or

2. statistical approaches can be used to derive hazardous concentrations that can be

incorporated into a probabilistic assessment.


31

Table 11. Additional species used for the regulatory assessment of chemicals identified from the survey


References

Modified exposureand speciessensitivity

Elodea canadensis Field Mature plant Variable exposurelength

Incorporates fatedata andestablishment ofspecies sensitivity


Confidential report

Modified exposureand speciessensitivity

Lemna gibba Laboratory Mature plant Variable exposurelength

Incorporates fatedata andestablishment ofspecies sensitivity


Confidential report

Species sensitivity Colomba sp Field Mature plant 7 d exposure Establishesspecies sensitivity

GLP with EC50


Species sensitivity Hygrophillapolysperma

Field Mature plant 7 d exposure Establishesspecies sensitivity

GLP with EC50


Species sensitivity Chironomusriparius

Laboratory 1st instar Acute 48 h Establishesspecies sensitivity

GLP with EC50


Species sensitivityAsellus aquaticusDaphnia pulexNotonecta spGammarus pulexBaetis rhodaniOstracod sp.Chaoborus sp.Chydorodoe sp.Cyclopoid sp.Calanoid sp.Limnea stagnalis

Field Juvenile Acute 48 h Establishesspecies sensitivity

GLP with EC50



32

2.3 Organism recovery

A number of approaches have been used to assess the potential for an organism to recover

after exposure to a toxicant (e.g. Table 12). Assessment of recovery is a standard component

of the lower tier toxicity tests for algae and the higher plant, Lemna. An aliquot of previously

exposed cells or fronds is placed into clean media to establish whether cell division is

reversibly or irreversibly retarded. However, the principle can be extended to aquatic plants,

invertebrates and to sub-lethal effects in fish (e.g. swimming abnormalities, colouring,

lethargy or impaired feeding and growth). A few studies have been reported for pesticides,

including organophosphorous and carbamate pesticides and for organisms including daphnids,

chironomids and blackfly (e.g. Kallander et al., 1997; Sanchez et al., 1999).

The potential for recovery will be determined by the duration before further exposure and the

toxic mode of action of the compound being considered. For example, studies into the effects

of carbamate pesticides on chironomids indicated that toxic effects were reduced if animals

were allowed to recover for more than 6 hours (Kallander et al., 1997). This observation was

probably due to the reactivation of acetylcholinesterase during the recovery period –

carbamate pesticides have been shown to be reversible inhibitors of AChE. No such recovery

was observed for organisms exposed to organophosphorus insecticides which are considered

irreversible inhibitors (Kallander et al., 1997). RADAR is a useful tool to evaluate duration

of pesticide exposure and time for recovery which has been proposed by ECOFRAM

(Hendley & Giddings, 1999).

The results of organism recovery studies could be used to refine the toxicity value used in the

risk assessment process. For example, if a compound is shown in standard toxicity studies to

affect a test organism at a particular concentration but is able to recover when subjected to a

more realistic exposure scenario, the effect value used in the assessment could be raised. One

problem with this approach is in demonstrating the relevance of recovery in the laboratory for

a whole community and this could be particularly so where species replacement occurs.


33

2.4 System recovery

Once a system has been impacted by a contaminant, its rate of recovery will be determined by

the persistence of the contaminant and the ecology, physiology and biochemistry of organisms

in the system and in proximity to the system.

One approach to predict the likely recovery of a system is based on the dissipation half-life of

the compound, the initial exposure concentration and the hazardous concentration for 5% of

species (Van Straalen et al., 1992). If recovery is required within a year after a single

application, then the chemical’s half-life should meet the condition:

<

5

1/2

ln

2ln

HC

CT

o

where T1/2 is measured in years, C0 is the initial concentration and HC5 is the hazardous

concentration for 5% of species. Although degradation will be necessary for complete

recovery, under some conditions it may not be sufficient and ecological recovery may lag

behind the disappearance of the chemical. This lag will depend on a range of factors

including: accessibility of shelters; survival of resistant life stages; and presence of untreated

areas from which recolonisation can occur. In addition, ability to recolonise will depend on

habitat selection and life history

A number of experimental studies have been proposed for determining whether an impacted

system is likely to be re-colonised. This is particularly relevant for mobile organisms such as

fish which may be resident in a section of a water body for a short time and may be able to

move to leave a toxic system. The study design consists of adding the test organism at the

start of the study and at regular intervals thereafter. A similar approach can also be used with

those invertebrate species with a high potential for re-colonisation, with the re-introduction of

individuals into a previously treated system, and monitoring of their subsequent survival and

performance. For example, Crane et al. (1999) investigated the toxicity of pirimiphos-methyl

to Gammarus pulex. Beakers were spiked with the study compound at day 0 and toxicity to

Gammarus pulex was assessed over 24 h at 1, 4, 8 and 12 d after application. Generally

mortality decreased with increasing time from spiking, probably as a result of pesticide

degradation.


34

The factors that influence the recovery of a population after a significant perturbation are

complex. One important factor affecting recovery is the life history of the individual species

(Sherratt et al., 1999). Aquatic organisms can vary widely in their reproductive and dispersal

characteristics. For example some species may reproduce continuously whereas others may

reproduce at discrete times of the year. Generation times can vary from days (e.g. daphnids)

to years (e.g. odonates). The ability for recolonisation can also vary - many aquatic insects

have adult life stages with wings whereas other taxa (e.g. crustaceans and molluscs) will rely

on recolonisation by more passive forms of dispersal (wind, flooding, transportation by birds).

The method for inclusion of system recovery in the risk assessment process is likely to be

dependent on the nature of the organisms most at risk. For example, if a particularly sensitive

group has a high potential for recolonisation, then the refinement of the effect concentration

used in the risk assessment process may be justified. If there is low potential for

recolonisation, then the original effect concentration should be used. A wide range of factors

will affect recolonisation potential and many of these are not fully understood. Significant

work is therefore required in this area before useful guidelines on application to the risk

assessment process can be developed.


35

Table 12. Higher tier laboratory studies identified in the survey that are performed to assess the potential for organisms to recover following exposure to toxicants


References

Recovery Selenastrumcapricornutum

Laboratory Exponential growth Cell densitydetermined during4 d exposure toseveral testconcentrations.Aliquot of culture isthen taken fromselected testconcentrations anddiluted at leat 100xin fresh media andalgal density isthen meaured overseveral days

Determines if algalgrowth resumesonce testconcentration fallsbelow a NOEC

3 replicates,calculate areaunder growth curveand growth rate c.f.untreated control

Internal industryreport

Recovery Lemna minor Laboratory Mature plant 7 and 14 dexposure andrecovery

Incorporatesrecovery potential


Confidential report

Recovery Lemna gibba orLemna minor

Laboratory Rapidly dividing Frond numberdetermined over 4-14 d exposure toselected testconcentrations.Equal number offronds/plantsremoved to freshmedia and frondnumberdetermined forseveral days

Determine whetherincrease in frondnumber resumesafter testsubstanceremoved

3 replicates,calculate growthrate, c.f untreatedcontrol

Internal industryreport


36

2.5 Sensitive life stage studies

Sensitivity to toxicants is well known to vary amongst species but sensitivity can also vary

between the different development stages of an organism (e.g. Beattie and Pascoe, 1978;

Green et al., 1986; Shazili and Pascoe, 1986; Williams et al., 1986; McCahon and Pascoe,

1988). Standard toxicity studies are typically performed using neonate or juvenile test

organisms as it is generally recognised that these are the most sensitive life-stages. By testing

sensitive life stages, it is expected to obtain a protective assessment of the effect of a chemical

on aquatic organisms and thus derive safe concentrations for a substance over the full life

cycle of a test population. For fish, it is generally recognised that early life stages are more

sensitive to contaminants than adults, athough the egg itself may be more resistant (e.g.

McKim, 1977; Macek and Sleight, 1977; McKim, 1985; Kevan and Dixon, 1996). For

example, studies with fish have demonstrated that early life stage tests could predict effects on

complete lifecycles to within a factor of two (e.g. McKim, 1985). Analyses of the ECETOC

Aquatic Toxicity (EAT) database (ECETOC, 1993) of toxicity values for a range of chemical

classes and organisms has supported these findings and indicated that, with the exception of

fish embryos, younger fish tend to be more sensitive to toxicants than older fish

(larvae>juveniles>adults) (Hutchinson et al., 1998; Table 13).

Similar results have been obtained for invertebrates (Hutchinson et al., 1998; Table 13). For

example, studies with D. magna (Kusk, 1996) demonstrated that newborn animals were as

sensitive to pirimicarb as older (1 week old) animals. Studies by Stuijfzand et al. (2000)

investigated the effects of diazinon on the different developmental stages of Hydropsyche

angustipennis and Chironomus riparius. First instars for both species were more sensitive

than last instars by up to two orders of magnitude.


37

Table 13. Comparison of sensitivities of organism life stages to chemical substances (adapted

from Hutchinson et al., 1998)

Life stages Hazen percentile ( percentage of

substances for which toxicity value for

a is greater than that for b)

Mean Sensitivity ratio

a vs b EC50 NOEC EC50 NOEC

Fish – embryos vs

larvae

99 31.8 0.12 – 2.04 0.72 – 11.1

Fish – larvae vs

juveniles

71.4 83.3 0.05 – 9.09 0.05-4.35

Fish – juveniles vs

adults

92.2 - 0.02-12.5 -

Fish - early life

stage vs life cycle

- 42.3 - 0.30-21.3

Invertebrates –

larvae vs juveniles

66.1 - 0.80-333 -

Invertebrates –

juveniles vs adults

54.3 90.7 0.07-2000 0.03-3

There are instances where older organisms have been shown to be more sensitive to a

particular toxicant. Studies with chlorpyrifos (Naddy et al., 2000) have indicated that older

organisms are more sensitive than neonates. One possible reason for this is that the older

animals have an increased filtration rate and received a higher dose as chlorpyrifos is likely to

be associated with food. Berrill et al. (1998) showed that 2-week old tadpoles were more

sensitive to endosulfan than newly hatched tadpoles because endosulfan affected the post-

hatch development of the neuromuscular system. By understanding both the significance of

different routes of uptake for different classes of chemicals and the physiology of an

organism, it may be possible to predict these types of differences in species sensitivity and

hence identify the most appropriate life stage to test. The testing of sensitive life stages for

some species may be problematic due to the lack of test methodologies and culturing methods

for certain species.

Standard ecotoxicity studies generally focus on neonate or juvenile animals which are likely

to be the most sensitive life stage. The use of sensitive life stage studies is therefore unlikely


38

to have a major impact on the risk assessment of a particular substance. However, in instances

where it is known that a particular substance is likely to be more toxic to a life stage not

studied in the standard tests (e.g. based on the results of studies with related compounds or

based on the mode of action of the substance), it may be appropriate to test this additional life

stage and as a result incorporate a smaller uncertainty factor into the risk assessment process.

2.6 Population level studies

As standard toxicity studies are generally performed on the most sensitive life-stages, results

may over-estimate effects on populations. Life-cycle and population level studies can be

useful tools for assessing the likely impact of a pesticide on populations of aquatic organisms

(e.g. Table 14). The structure of a population at the time of toxicant exposure may be an

important factor in determining toxicant impact (Stark and Banken, 1999). Populations in

nature consist of a mixture of life stages, yet toxicological studies usually only evaluate one

starting life stage. Studies into the effects of the pesticides azadirachtim and dicofol on three

differently structured populations of two terrestrial arthropods (Tetranychus urticae and

Acrythosiphon pisum) have demonstrated that initial population structure does influence the

effects of the pesticides on populations (Stark and Banken, 1999). Consequently, age/stage

structure should be considered when evaluating toxicant effects.

Both modelling and experimental approaches can be used to determine population level

effects. Laboratory experimental studies have been performed using plant and invertebrate

species (Campbell et al., 1999) and have ranged from simple studies evaluating effects on a

limited number of life stages to complex multi-life stage studies that simulate natural

population dynamics (e.g. Maund et al., 1992; Taylor et al., 1992; Chandler and Green, 1995;

Blockwell et al., 1999). This type of test is normally restricted to invertebrates and plants and

may not be possible with groups prone to cannibalism unless the individuals are held

separately. However, experiments with fish may be possible under field conditions (e.g.

Johnson et al., 1994; Shaw et al., 1995).

The data generated from population-level studies and from simple toxicity studies can be used

to model the effects of compounds on populations in the field (e.g. Hallam and Lassiter, 1994;

Sibly, 1996; Calow et al., 1997; Maund et al., 1997). Population models describe the

dynamics of a finite group of individuals through time and have been used extensively in

ecology and fisheries management. These approaches are well developed for use on fish

(Barnthouse et al., 1987; De Angelis et al., 1991) and invertebrates (e.g. Kooijman and Metz,


39

1984; Gurney et al., 1990; Acevedo et al., 1995; Roex et al., 2000). By using models, the

results of laboratory studies can be used to evaluate a range of scenarios (Campbell et al.,

1999) and can assist in identifying short and long-term changes in population structure.

Proper use of population models requires an understanding of the natural history of the

species under consideration, as well as a knowledge of how the toxicant influences its

biology. Model inputs can include somatic growth rates, physiological rates, fecundity,

survival rates of various classes within populations, and how these change in response to the

study compound (USEPA, 1998). In addition, population density effects may be important.

These can be addressed in the risk assessment process using uncertainty analysis.

The results of single species population dynamics can be put into context by linking them to

less detailed food-web and carbon/nutrient cycling models to get a better indication of the

ecological consequences of pesticide usage (Baveco and De Roos, 1996).


40

Table 14. Higher tier population level studies


References

Population effects Daphnia magna Laboratory 6 and 14 dexposure andrecovery

Acute 48 h Incorporatespopulationconsiderations


Confidential report

Population effects Daphnia magna Laboratory Mixed age Life cycle test Incorporatespopulation leveleffects


Confidential report


41

2.7 Indoor multi-species studies

Single-species laboratory studies may be inadequate for predicting the effects of chemicals on

ecological communities. There are two approaches to overcome this problem: development of

test systems with two or more interacting species; and development of population dynamic

simulations incorporating selected species of an ecosystem and the impact of varying abiotic

factors (Axelsen et al., 1997).

By using multispecies studies it is possible to demonstrate: 1) indirect trophic level effects; 2)

compensatory shifts within a trophic level; and 3) seasonal responses to chemicals. Outdoor

multispecies studies are generally well established and are used in the risk assessment of

pesticides (e.g. Giddings et al., 1996 and 1997; see Introduction). Indoor multispecies systems

have a number of advantages over outdoor mesocom studies in that they can be undertaken

throughout the year and experimental conditions can be closely controlled. The scale of many

of these studies means that less chemical is used and the set up cost can be significantly less.

A number of systems have been used including: 1) simple multispecies tests that are designed

to illustrate the influence of a particular biotic factor on a test endpoint (e.g. Day and Kaushik,

1987; Ham et al., 1995; Taylor et al., 1995; Kluttgen et al., 1996; Gomez et al., 1997; Hamers

and Krogh, 1997); 2) defined microcosm tests involving the inoculation of small systems with

a well defined assemblage of small organisms (e.g. Taub, 1969; Leffler, 1981; Kersting,

1991); and 3) semi-realistic microcosm tests that represent natural assemblages of organisms

and that are generally constructed from samples of natural ecosystems (e.g. Breneman and

Pontasch, 1994; Fliedner et al., 1997; Van den Brik et al., 1997; Barry and Logan, 1998). The

simple and defined systems have the advantage that they are simple to perform and that they

have a high degree of reproducibility. However, they do represent a relatively abstract system

with tenuous links to the real world. In contrast, systems based on natural assemblages and

media provide a closer approximation to natural ecosystems, but are harder to replicate.

Simple multispecies laboratory studies could be beneficial in the risk assessment process. As

the studies become more complex, it is probably more appropriate to perform pond mesocosm

and artificial stream studies. The simple studies are most appropriate where a substance

impacts a known key species within a food chain. A simple study into how effects are

transferred from this species to organisms higher up the food chain may then be appropriate.

A detailed analysis of results from mescososm studies and the use of food web modelling

could assist in the design of these approaches. Ultimately the results of these studies could be

used in conjuction with a reduced trigger value.


42

Table 15. Indoor multispecies test systems used in the assessment of contaminants

System Microcosm

characteristics

Compounds studied Reference.

Vascular plants 0.0375 m3; 6 kgsediment; Zannichelliapalustris; Potamogetonpectinatus; Vallisneriaamericana; Zosteramarina

atrazine Correll and Wu, 1982

Macrophytedominated system

0.85 m3; 10 cm layer ofnatural sediment; 600 lwell water; planktoncommunities andbenthicmacroinvertebratesintroduces + Elodeanuttallii,macroinvertebrates andzooplankton from Dutchditches

carbendazim Van den Brink et al.,1997; Cuppen et al.,2000; Van den Brink etal., 2000

Benthic invertebratesand algae

0.00008 m3; flowthrough system usingsite water; benthicinvertebrates +phytoplankton slides

atrazine Gruessner and Watzin,1996

Genericmicroecosystem

7.5L; 18°C; 3 trophiclevels (algae,herbivores (daphniamagna); anddecomposers(bacteria)) kept inseparate subsystemsconnected by arecirculating flowsystem.

chlorpyrifos Leeuwangh et al., 1994

Benthicmacroinvertebrates(estuarine)

144 or 400 cm2 with 5cm deep sediment layerand sewater with 1.6L/min/tube flow.

chlorpyrifos Flemer et al., 1997

Communityrepresentative ofepilimnon of amesotrophictemperate lake

3 l; 3 species of algaemix of lake bacterialspecies

atrazine Genoni, 1992

Alagae + benthicinvertebrates

120 l re-circulatingsystem; field collectedinvertebrates andperiphyton

atrazine Gruessner and Watzin,1996

Algae + daphniamagna

100 ml of watercontaining 1.2 x 105

algal cells/ml + 1daphnid

chromium Gorbi and Corradi,1993


43

3. APPLICATION OF HIGHER TIER LABORATORY STUDIES IN

ENVIRONMENTAL RISK ASSESSMENT

Generally, higher tier studies should be carried out in response to indications of adverse risk

at lower tiers. It is assumed that a full and critical evaluation of the baseline fate and effects

data package precedes any move to higher tier studies. Higher tier laboratory methods can be

used to:

• reduce the magnitude of uncertainty factors in hazard/risk assessment;

• provide a more realistic assessment of effects on a particular species;

• set dose rates, duration and identity of key species of concern for multi-species testing;

• provide input data to ecological models;

• assess compounds with a high Koc and/or affinity for binding to sediment/humic acids;

• provide information on the potential duration of effects;

• assess whether effects are reversible.

The exact nature of the test to be performed will be dependent on a number of factors

including: the results of lower tier studies; the physicochemical properties of the compound of

interest; and the degradability of the study compound. There are currently no guidelines

describing the incorporation of higher tier laboratory studies into the risk assessment process.

However, it would seem appropriate to use a stepped system, involving the initial

interrogation of core data, followed by the identification of sensitive species and finally a

combination of modified exposure studies and population studies on the identified species. A

summary of the major higher tier studies identified is provided in Table 16. The combination

of a number of the approaches (e.g. use of population studies coupled to pulsed exposure and

dissipation studies) may also be appropriate.

Higher tier laboratory approaches have a number of advantages over the use of standard

ecotoxicity tests and/or mesocosm/field studies. These include:

• the endpoints are more pertinent to actual environmental exposure;

• they provide scope for the integration of ‘realistic’ measures of fate and exposure into the

risk assessment process;

• they should provide more confidence when predicting actual effects in the environment;


44

• they can be designed to determine effects on specific endpoints, e.g. mortality, growth,

behaviour, population level;

• they can be used to identify more clearly those species that may be at risk and assist in the

targeting of multi-species tests;

• the results may help explain observed effects (e.g. mode/specificity of action);

• recovery can be assessed for both individuals and populations;

• there are none of the complications that are associated with the interpretation of data from

multi-species studies;

• relative to mesocosm studies, they can be performed with less regard to the season, the

test organisms are usually readily available and the system is more controlled;

• sublethal effects can be determined.

However the approaches also have limitations:

• they cannot provide information on species interactions;

• results can be more difficult to interpret and compare than those from standard single

species tests;

• the tests may be more costly than standard single species studies;

• the test methods are undefined;

• the quality of available test species cannot always be guaranteed.


45

Table 16. Higher tier toxicity methods

Test type Rationale for test Design of test Data analysis Output Advantages/Disadvantages

Testing ofadditonalspecies

To reduce uncertaintyfactors and/or developspecies sensitivitydistributions

Test a minimum of 8 additionalspecies: –

From a range of groups

or

For compounds whose mode ofaction indicates that a particulargroup will be sensitive, testsshould be performed on speciesfrom this group

or

If fish are likely to be sensitive,then only 5 additional speciesshould be tested

Lower uncertainty factorsby up to an order ofmagnitude or use statisticalapproaches to generatespecies sensitivitydistributions

Revised TER or HCx • identification of most sensitivespecies

• more relevance than standardspecies

• refinement of TERs• allows probabilistic risk

assessment

• no guidance available onuncertainty factors to use

• no guidance available on whatlevel of population protection isacceptable

• lack of standard test methods foradditional species

• may need to obtain organismsfrom the field

Time-to-eventanalysis

To assess the toxiceffects of exposure topesticides overdurations shorter thanthe standard testduration

Use data recorded over time forstandard ecotoxicity tests

Compare time-to-eventdata with expectedexposure duration

If duration < time toeffect then reduceeffect estimate

• more realistic assessment ofecotoxicity

• resolution of observations instandard tests may not beappropriate

• no guidance on how effectsestimate can be reduced


46


Short-termexposurestudies

To assess effects ofshort exposuredurations on pesticidetoxicity

Perform test over exposureduration predicted for the field

Use standard toxicity teststatistical procedures

Revised effectmeasurement for usein risk characerisation

• more realistic assessment ofexposure

• does not consider delayed effects

Pulsed-exposurestudies

To simulate pulsedexposure conditionsthat are likely in the field

Static renewal or flow-throughsystem, depending on detailrequired

Use standard toxicity teststatistical proceduresand/or toxicokinetic models



• includes potential for organismsto recover

• results affected by a number ofvariables (including pulseduration, time between pulsed,organism recovery) so difficult tointerpret

Dissipationstudies

To assess the impact ofdissipation processeson toxicity

Generally performed usingsediment/water systems.Studies could also beperformed to assess effects ofphotodegradation (byperforming studies in the light)or volatilisation.




• can include effects of metabolites• guidelines not available on

selection of test matrices

• not suitable for all species (e.g.algae)

• methods not available forassessing photodegradation andvolatilisation


47


Studies usingnatural matrices

Account for differencesbetween bioavailabilityin natural and laboratorysystems

Perform toxicity studies usingwater samples collected fromthe field



• assessment of bioavailabilityunder natural conditions

• no guidance available onselection of test matrix

Effects onsensitive lifestages

To determine whether aparticular life stage issensitive or not

A range of approaches areavailable for a number ofspecies.


Toxicity to sensitivelife stages – possiblereduction inuncertainty factor

• reduction of uncertainty factors

Populationstudies

Perform studies on apopulation of aparticular species overa prolonged time period

Experimental and modellingapproaches are available

Use population modellingapproaches to interpretresults

More ecologicallyrelevant assessmentof effects – possiblereduction inuncertainty factor

• includes effects of recovery• includes selection of tolerant

organisms

• may be more ecologicallyrepresentative

• difficult to perform on fish

Recoverystudies

To determine whetherpopulations can recoverafter exposure to apesticide

Expose organisms and transferto clean media to determinetime to recovery

Assessment of timerequired fororganisms to recover– within 2 monthsmay be acceptable(CLASSIC, in prep.)


48

When designing higher tier tests for use in the risk assessment of chemicals, a number of

factors should be considered. These include: whether formulated or technical product should

be use; in-life monitoring of chemical concentrations in the test; replication and statistics; and

quality control. The approaches used to address each of these issues will vary according to the

nature of the test.

3.1 Use of formulated product or active ingredient in higher tier studies

A number of studies have investigated the effects of co-formulants on pesticide toxicity (e.g.

Zitko et al., 1979; Bradbury et al., 1985; Holdway et al., 1994). The results indicate that

formulations can be either more or less toxic than the active substance. For example, technical

grade esfenvalerate and fenvalerate were 1.7 and 2.3 times more toxic than emulsified

formulations containing these active substances. The results were probably due to a slower

rate of uptake of the pesticides from the formulated product (Holdway et al., 1994). In

contrast, studies on one emulsified pesticide indicated that this was more toxic than the

technical grade pesticide, possibly due to a longer half-life as a result of the emulsifier. If the

formulated product is more toxic, it should be used in the higher tier studies.

Generally, it is expected that formulated material will be used for higher tier studies as it will

better mimic drift inputs to surface waters. Dosing to nominal concentrations above the limit

of solubility makes interpretation of test results difficult and may result in physical effects (e.g

formulated material in excess of solubility may form a film on the water surface); such doses

are only appropriate in support of a particular application of concern. Formulated product is

also preferred where the formulation has been shown to have a significant impact on exposure

and/or effects. For example, it may affect the stability and residence time of the compound in

the water column and thus increase bioavailability. If formulated product is used, testing of a

formulation blank may be appropriate where levels of toxicity are observed that were not

anticipated. However, use of such a blank will normally be triggered at lower tiers of the

testing procedure.

The technical product may be easier to work with in some instances. A co-solvent can be

used to aid dispersion. In addition, the results of studies on technical product can be

extrapolated to different formulations of the same product. Technical product is likely to be

preferred for species sensitivity tests as this will allow for a direct comparison with the

available Tier I data, and again for extrapolation to a range of formulations. Where toxicity

tests are mimicking an input of chemical to surface water via overland flow or drainflow, it is

recommended that technical product should be used. Although it is not clear from fate studies


49

how long a chemical and its formulation are likely to remain associated after introduction into

the environment, the current assumption is that the two are instantaneously decoupled upon

reaching the soil and that the formulation plays no further part in the behaviour of the active

substance.

3.2 In-life monitoring of chemical fate

Monitoring of chemical fate over the course of a particular higher tier laboratory toxicity test

is generally required as at lower tiers. This is true for tests both with maintained

concentrations (e.g. life cycle tests, species sensitivity tests) and those with a degradation or

removal process. Monitoring will show the effect of any modified design on the behaviour of

a compound and also allow comparisons between results of laboratory toxicity tests at lower

and higher tiers.

Frequency of in-life monitoring of chemical fate needs to be appropriate to the test system.

For example, more frequent monitoring will be required if chemical is added in a series of

pulses (e.g. to simulate drift from multiple applications) than if a single dose of chemical is

added. In certain cases (e.g. the introduction of a sediment layer), the higher tier toxicity test

will closely match a fate and behaviour study (e.g. a water-sediment study) and in-life

monitoring of fate may be reduced to the minimum necessary to show similar behaviour of

the chemical. This is only the case if the system designs are closely related.

Where tests examine effects of a chemical at a particular concentration (e.g. a PEC), it would

be appropriate to define a target threshold whereby aqueous concentrations should be within a

certain percentage range of the target value. Use of radiolabelled compound may prove an

effective way of monitoring fate and distribution at low levels where sufficiently sensitive and

specific methods of analysis may not be available. It is important to remember that for many

systems the size will not allow for removal of adequate sample for analysis. This may mean

that analysis has to be limited to the start and end of the experiment, or alternatively that

additional units should be set up and taken for destructive analysis at timed intervals.


50

3.3 Replication

As a general principle, the replication specified in existing standardised guidelines for lower

tiers of testing should be adhered to, wherever the study designs are comparable. However,

there may be practical issues which prevent this – for example, the availability of some field

collected animals (e.g. dragon fly larvae) may prohibit testing on this species with the same

replication as prescribed for a standard invertebrate test such as Daphnia. In addition, their

carnivorous nature may dictate a different holding regime of individually held animals.

Generally, as the test system becomes more natural and complex, the response is likely to

become more variable and the level of replication required is likely to increase. A decision on

replication required should be made in consultation with a statistician.

An important, related point is that acceptance criteria in the form of mortality thresholds

should not be prescriptive for higher tier tests because of additional difficulties in handling.

Decisions on the acceptability and validity of a study should be based on comparison with an

appropriate control system run in parallel. As indicated earlier, the levels of background

mortality experienced with these non-standard higher tier tests is likely to be greater and this

should be taken into consideration. Statistical aspects of replication are briefly considered in

Section 3.4 below.

3.4 Use of statistics

Two statistical activities are involved in ecotoxicity testing, namely: experimental design and

data analysis. Of these, experimental design is the most important and the following should

be considered when higher tier studies are being designed (Chapman et al., 1996):

Randomisation – treatments should be allocated to experimental units in a random fashion as

the experimental variation in higher tier studies may be large (due for example to the use of

natural populations or test matrices). Moreover, all handling of experimental units after the

start of a study should also be performed in a random fashion.

Replication – replication allows the power or precision of a test to be controlled. Larger

numbers of replicates will lead to more precise estimates or tests with greater power. If

ANOVA is the likely form of data analysis, then the higher tier test procedure should specify

the required precision and power so that the most appropriate number of replicates can be

chosen. If a dose response model is to be fitted, then the precision of the numbers of


51

concentrations tested will also affect an estimate of effective concentrations (e.g. EC50,

LC50).

Number of organisms per unit - Having more than one organism per experimental unit will

usually improve the power and precision of the test. The extent of the improvement will be

dependent on the size of the within-unit variation. An assessment of the within-unit variation

associated with a range of higher tier approaches would therefore be useful in order to identify

the most approproate number of test organisms to be used in a particular test.

Number and spacing of concentrations – the choice of concentrations, both their number and

their value, will affect the precision of the LC and EC estimates. Standard test guidelines

currently require four or five geometrically spaced concentrations. Five test concentrations are

also likely to be appropriate for higher tier studies.

Optimum times for taking measurements – standard guidelines often require measurements to

be made at more than one time point. The timing and frequency of measurements in a higher

tier study may well be more critical. For example a pulsed exposure may require many

observations.

Blind assessing – the endpoints used in certain higher tier studies may be quite subjective.

For example, evaluation of effects such as lethargy or swimming abnormalities in sublethal

tests. Under these circumstances, blind assessment of test vessels is recommended.

Use of linear regression to generate a dose response is generally preferred to an ANOVA

approach to experimental design, although both options are available in the OECD guideline

for mesocosm studies. ANOVA requires a larger number of replicates and the value of

NOECs has recently been questioned due to the different design requirements of dose-

response modelling and hypothesis testing (e.g. Skalski, 1981; Kooijman, 1981; Pack, 1993).

Estimation of an effective concentration (EC) using regression analysis has a number of

advantages over the derivation of NOECs: the EC is not restricted to being one of the test

concentrations; its value does not depend on the precision of the experiment; its precision can

be estimated; its interpretation is straightforward; and regression modelling uses data from all

concentrations (Chapman et al., 1996).


52

3.5 Quality control

Higher tier studies should always be conducted to GLP. The standard 10% and 20% mortality

thresholds are likely to be exceeded in some non-standard studies. However, acceptable

thresholds cannot be established without ring tests to investigate inter- and intra-laboratory

variation. This is considered impractical because of the inherent flexibility and wide range of

available study designs at higher tiers. In the absence of appropriate mortality thresholds for

higher tier tests, expert judgement should be used in deciding the acceptability of a given

study relative to an appropriate control system. Results should be adequate to provide a valid

statistical evaluation.

A potential tool in quality control which is little used at present is the parallel testing of a

reference toxicant. In attempting to make higher-tier tests more like the real world, there is a

clear need to evaluate the success of a particular system. Reference toxicants offer the

opportunity to compare behaviour in the test system with known behaviour in the wider

environment and thus to establish the validity of the higher tier study. A suite of reference

toxicants would be required covering a range of physico-chemical properties and degradation

rates so that the chemical with properties most similar to the test compound could be selected.

4. RECOMMENDATIONS FOR FURTHER WORK

Higher tier laboratory tests provide a useful intermediate between standard toxicity studies

and semi-field / field tests. The review presented here has shown that there are a wide range

of higher tier studies available and that a significant body of work has been generated on

most, but not all, of these methods. Three main priorities are identified to take this area of

work forward and improve the regulatory acceptability of higher tier laboratory tests:

1. There is currently no guidance available to support the use of higher tier laboratory tests

and a priority is to put in place agreed documentation to do this. Aspects which would

need to be covered within the guidance include: a) a scheme of how and when to use each

particular test; b) major constraints and methodological pointers for each test type; c) how

to interpret the results; d) acceptability of the data; and e) how the results should be

incorporated into the risk assessment process, including indications of appropriate safety

factors to apply. Flexibility is an intrinsic component of higher-tier testing and it is


53

envisaged that the documentation would set broad constraints around the use and

interpretation of tests rather than a rigid set of guidelines. Given the complexity of the

issues involved and the range of tests available, it is likely that most benefit will be gained

through a joint initiative between scientists from regulatory agencies, industry and

research organisations. SETAC or a similar body may offer a useful, common forum for

discussion as with the HARAP workshop.

2. A number of higher tier tests are available and there is a need to validate these methods in

a consistent and systematic way. To do this, comparisons should be undertaken between

data and risk assessments generated from standard laboratory testing, higher-tier

laboratory testing and mesocosms. This would allow evaluation of the advantages and

limitations of laboratory higher tier testing compared to standard studies and mesocosm

studies. Costs involved in carrying out validation could be minimised if case studies were

performed on well studied compounds where some laboratory higher tier data are already

available (e.g. atrazine, lamda-cyhalothrin, chlorpyrifos). Supplementary data could then

be generated to fill any data gaps before the risk of the compound was characterised.

3. It would be greatly beneficial if a means could be found to pool expertise and knowledge

on working with higher tier tests and with particular aspects such as studying non-

standard species. It is not realistic to expect ring-tested methods to be available for higher

tier tests even in the medium term. However, informal meetings could be held in a

manner similar to the Joint Initiative on Beneficial Insects which led to the production of

a series of guidance notes. The aim would be to produce guidance documents to improve

the quality of data produced and harmonise aspects of procedure where appropriate. One

possibility for the UK would be to re-start the UK Aquatic ad-hoc group to address this

issue. Pesticides Safety Directorate (PSD) and DETR could play an important role in

driving the process, but the main sources of practical experience will lie within industry

and contract laboratories.

A number of secondary requirements for further work are set out below:

1. For assessments of species sensitivity, the number of additional species for testing has

been identified. However, the majority of available test methods (Appendix B) have

generally been developed in North America. The suitability of these tests relative to UK

aquatic systems needs to be assessed and recommendations should be provided on the

most appropriate test species to use for additional studies. Test methods for these

additional species may need developing. The development of geographical information


54

systems on the ecology of water bodies in agricultural areas should also be considered.

These systems would enable sensitive species to be readily identified for testing and assist

in setting priorities for method development.

2. Statistical analyses of species sensitivity datasets result in a hazard concentration for a

fixed percentage of species. There is some conflict in the literature with published studies

indicating that either a 5% or a 10% level is appropriate to protect aquatic systems. A

small study should be undertaken to support the selection of a common value and to

demonstrate its potential to protect the environment. A large volume of data (e.g. the

AQUIRE database, the EAT database, industry held data) is available that could enable

this work to be performed.

3. Datasets on the ecotoxicity of chemicals to species should be further analysed based on

information on the mode of action of a compound. The development of lists of sensitive

species associated with a particular mode of action would be highly beneficial when

selecting organisms for testing.

4. The results of modified exposure studies are likely to be highly dependent on the test

matrix. For sediment/water studies, research should focus on the effects of variability in

sediment characteristics on test results so that guidelines to standardise properties can be

put in place.

5. When a large dataset is available on the ecotoxicity of a compound, statistical approaches

should be used to assess effects levels for use in the risk assessment process. If few data

are available, then the assessment factor approach should be used. This latter approach

has a number of limitations including a lack of stability and no improved precision when

more species are tested. Work is required to develop a better approach for analysing

small datasets. One possibility would be a method based on assessment factors derived

from results of statistical approaches (Roman et al., 1999).

6. Current methods for assessing effects on communities from distributions of species

sensitivity assume that all species are equally important. This is probably not the case and

consideration should be given to identifying key species in aquatic ecosystems. In part,

this may be an output from a current MAFF/PSD project on “Aquatic ecosystems in the

UK agricultural landscape”. Approaches should be developed for ensuring that key

species are protected by risk assessments.


55

7. A range of modelling approaches has been identified to support/complement higher tier

toxicity studies. There is a need to validate models in order to allow better predictions of

effects and extrapolations to other scenarios in the future.

8. A number of multispecies studies have been identified . It may be possible to refine these

methods based on a knowledge of major interactions observed in the field or large

mesocosm studies. The use of food web models to assess the likely impact of a compound

should also be considered.

9. In situ toxicity tests are outside the scope of this review. A number of recent publications

suggest that such studies may provide a useful supplement for higher tier laboratory tests.

ACKNOWLEDGEMENTS

The authors would like to thank Dr Mike Collins and Dr Steve Norman for useful dialogue

during the course of the study and Professor Allen Burton, Dr Andy Girling, Dr David Pascoe

and Dr Martin Streloke for their constructive comments on the draft final manuscript for this

project. This study was funded by the Department of Environment, Transport and the

Regions.


56

REFERENCES

1. ABEL, P.D. (1980) A new method for assessing the lethal impact of short-term high-level discharges ofpollutants on aquatic animals. Prog Water Technol. 13: 347-352.

2. ACEVEDO, M.F., WALLER, W.T., SMITH, D.P., POAGE, D.W., MCINTYRE, P.B. (1995) Modellingcladoceran population to stress with particular reference to sexual reproduction. Nonlinear World 2: 97-129.

3. ALABASTER, J.S., ABRAM, F.S.H. (1965) Development and use of a direct method of evaluating toxicityto fish. In The Proceedings of the second international water pollution research conference, Tokyo, 1964.Pergamon Press, NY.

4. ALDENBERG, T., SLOB, W. (1993) Confidence limits for hazardous concentrations based on logisticallydistributed NOEC toxicity data. Ecotoxicology and Environmental Safety 25(1): 48-63.

5. ALLISON, P.D. (1995) Survival analysis using the SAS system. A practical guide. SAS Institute, Cary,NC.

6. AMERICAN SOCIETY FOR TESTING MATERIALS (1992) Standard guide for conducting acute toxicitytests with fishes, macroinvertebrates and amphibians. Designation E 729-88a Annual book of ASTMmethods, Vol. 11.04:403-422.

7. AQUATIC DIALOGUE GROUP (1994) Pesticide risk assessment and mitigation. SETAC Press,Pensacola, FL.

8. AXELSEN, J.A., HOLST, N., HAMERS, T., KROGH, P.H. (1997) Simulations of the presator-preyinteractions in a two species ecotoxicological test system. Ecological Modelling 101: 15-25.

9. BARBER, M.C., SUAREZ, L.A., LASSITER, R.R. (1988) Bioconcentration of nonpolar organics by fish.Environ. Toxicol. Chem. 7: 545-558.

10. BARNTHOUSE, L.W., SUTER, G.W., ROSEN, A.E., BEAUCHAMP, J.J. (1987) Estimating responses offish populations to toxic contaminants. Environ Toxicol Chem 6: 811-824.

11. BARRY, M.J., LOGAN, D.C. (1998) The use of temporary pond microcosms for aquatic toxicity testing:direct and indirect effects of endosulfan on community structure. Aquatic Toxicology 41: 101-124.

12. BAVECO, J.M., DEROOS, A.M. (1996) Assessig the impact of pesticides on lumbricid populations: anindividual-based modelling approach. Journal of Applied Ecology 33: 1451-1468.

13. BEATTIE, J.H., PASCOE, D. (1978) Cadmium uptake by rainbow trout, Salmo gairdneri eggs and alevins.J. Fish Biol 13: 631-637.

14. BELANGER, S.E. (1992) Use of mesocosms in predicting ecosystem risk from cationic surfactantexposure. In: Cairns, J., Neiderlehner, B.R., Orvos, D.R. (Eds) Predicting ecosystem risk. PricetonScientific, Princeton, NJ.

15. BELANGER, S.E., GUCKERT, J.B., BOWLING, J.W., BEGLEY, W.M., DAVIDSON, D.H., LEBLANC,E.M., LEE, D.M. (2000) Responses of aquatic communities to 25-6 alcohol ethoxylate in model streamecosystems. Aquatic Toxicology 48: 135-150.

16. BERRILL, M., BERTRAM, S., PAULI, B., COULSON, D., KOLOHON, M., OSTRANDER, D. (1995)Comparative sensitivity of amphibian tadpoles to single and pulsed exposures of the forest-useinsecticide fenitrothion. Environmental Toxicology and Chemistry. 14(6): 1011-1018.

17. BERRILL, M., COULSON, D., MCGILLIVRAY, L., PAULI, B. (1998) Toxicity of endosulfan to aquaticstages of anuran amphibians. Environ. Toxicol. Chem. 17(9): 1738-1744.

18. BLACK, M.C., MCCARTHY, J.F. (1988) Dissolved organic macromolecules reduce the uptake ofhydrophobic organic contaminants by the gills of rainbow trout (Salmo gardneri). EnvironmentalToxicology and Chemistry 7: 593-600.

19. BLANCK, H. (1984) Species dependent variation among aquatic organisms in their sensitivity tochemicals. Ecol. Bull. 36: 107-119.


57

20. BLANCK, H., WALLIN, G., WANGBERG, S.A. (1984) Species-dependent variation in algal sensitivity tochemical compounds. Ecotoxicology and Environmental Safety 8: 339-341.

21. BLOCKWELL, S.J., MAUND, S.J., PASCOE, D. (1999) Effects of the organochlorine insecticide lindane(γ-C6H6Cl6) on the population responses of the freshwater amphipod Hyalella azteca. EnvironmentalToxicology and Chemistry 18(6): 1264-1269.

22. BRADBURY, S.P., COATS, J.R., MCKIM, J.M. (1985) Differential toxicity and uptake of two fenvalerateformulations in fathead minnows. Environmental Toxicology and Chemistry 4: 533-541.

23. BRECK, J.E. (1988) Relationships among models for acute toxicity effects: applications to fluctuatingconcentrations. Environmental Toxicology and Chemistry 7: 775-778.

24. BRENEMAN, D.H., PONTASCH, K.W. (1994) Stream microcosm toxicity tests: predicting the effects offenvalerate on riffle insect communities. Environ. Toxicol. Chem. 13: 381-387.

25. BROWN, V.M., JORDAN, D.H.M., TILLER, B.A. (1969) The acute toxicity to rainbow trout offluctuating concentrations and mixtures of ammonia, phenol and zinc. J. Fish. Biol. 1: 1-9.

26. BURNISON, B.K., HODSON, P.V., NUTTLEY, D.J., EFLER, S. (1996) A bleached-kraft mill effluentfraction causing induction of a fish mixed function oxygenase enzyme. Environ. Toxicol. Chem. 15:1524-1531.

27. CAIRNS, J. (1986) The myth of the most sensitive species. BioScience 36(10): 670-672.

28. CALOW, P., SIBLY, R.M., FORBES, V.E. (1997) Risk assessment on the basis of simplified life historyscenarios. Environ. Toxicol. Chem. 16: 1983-1989.

29. CAMPBELL, P.J., ARNOLD, D.J.S, BROCK, T.C.M., GRANDY, N.J., HEGER, W., HEIMBACH, F.,MAUND, S.J. & STRELOKE, M. (1999). Guidance document on higher-tier aquatic risk assessment forpesticides (HARAP). SETAC-Europe, Brussels, 179p.

30. CHANDLER, G.T., GREEN, A.S. (1995) A 14-day harpacticoid copepod reproduction bioassay forlaboratory and field contaminated muddy sediments. In: Ostrander, G. (ed) New techniques in aquatictoxicology. Lewis Publishers, Boca Raton.

31. CHAPMAN, P.F., CRANE, M., WILES, J., NOPPERT, F., MCINDOE, E. (1996) Improving the quality ofstatistics in regulatory ecotoxicity tests. Ecotoxicology 5(3): 169-186.

32. CHAPMAN, P.F., CRANE, M., WILES, J.A., NOPPERT, F., MCINDOE, E.C. (1996) Asking the rightquestions: ecotoxicology and statistics. SETAC-Europe, Brussels.

33. CLARK, J.R., GOODMAN, L.R., BOTHWICK, P.W., PATRICK, J.R., CRIPE, G.M., MOODY, P.M.,MOORE, J.C., LORES, E.M. (1989) Toxicity of pyrethroids to marine invertebrates and fish: a literaturereview and test results with sediment-sorbed chemicals. Environ Toxicol Chem 8: 393-401.

34. CORRELL, D. L., WU, T. L. (1982) Atrazine toxicity to submerged vascular plants in simulated esturinemicrocosms. Aquat. Bot. 12: 151-158.

35. COX, D.R., OAKES, D. (1994) Analysis of survival data. Wiley and Sons, New York.

36. CRANE, M. (1997) Research needs for predictive multispecies tests in aquatic toxicology. Hydrobiologia346: 149-155.

37. CRANE, M., ATTWOOD, C., SHEAHAN, D., MORRIS, S. (1999) Toxicity and bioavailability of theorganophosphorous insecticide pirimiphos methyl to the freshwater amphipod Gammarus pulex L. inlaboratory and mesocosm systems. Environmental Toxicology and Chemistry 18(7): 1456-1461.

38. CUPPEN, J.G.M., VAN DEN BRINK, P.J., CAMPS, E., UIL, K.F., BROCK, T.C.M. (2000) Impact of thefungicide carbendazim in freshwater microcosms.1. water quality, breakdown of particulate organicmatter and responses of macroinvertebrates. Aquatic Toxicology 48: 233-250.

39. CURTIS, L.R., SEIM, W.K., SIDDENS, L.K., MEAGER, D.A., CARCHMAN, R.A., CARTER, W.H.,CHAPMAN, G.A. (1989) Role of exposure duration in hydrogen ion toxicity to brook (Salvelinusfontinalis) and rainbow trout (Salmo gardneiri). Can J. Fish Aquat. Sci. 46: 33-40.

40. DAY, K., KAUSHIK, N.K. (1987) The adsorption of fenvaerate to laboratory glassware and the algaChlamydomonas reinhardii, and its effects on uptake of the pesticide by Daphnia galeata mendotae.Aquatic Toxicology 10: 131-142.


58

41. DE ANGELIS, D., GODBOUT, L.L., SHUTER, B.J. (1991) An individual-based approach to predictingdensity-dependent compensation in smallmouth bass populations. Ecological Modelling 57: 91-115.

42. DEBUS, R., FLIEDNER, A., SCHAFERS, C. (1996) An artificial stream mesocosm to simulate fate andeffects of chemicals: technical data and initial experience with biocenosis. Chemosphere 32(9): 1813-1822.

43. DYER, S.D., BELANGER, S.E. (1999) Determination of the sensitivity of macroinvertebrates in streammesocosms through field derived-assessments. Environmental Toxicology and Chemistry, 18(12), 2903-2907.

44. ECETOC (1993). Aquatic toxicity data evaluation. Technical Report No. 56. ECETOC, Brussels.

45. FAIRCHILD, J.F., LAPOINT, T.W., ZAJICEK, J.L., NELSON, M.K., DWYER, F.J., LOVELY, P. (1992)Population, communiy and ecosystem level responses of aquatic mesocosms to pulsed doses of apyrethroid insecticide. Environmental Toxicology and Chemistry 11: 115-129.

46. FAIRCHILD, J.F., DWYER, F.J., LAPOINT, T.W., BURCH, S.A., INGEROLL, C.G. (1993) Evaluationof laboratory-generated NOEC for linear alkylbenzene sulfonate in outdoor experimental streams.Environmental Toxcology and Chemistry, 12: 1763-1776.

47. FAIRCHILD, J.F., RUESSLER, D.S., CARLSON, A.R. (1999) Comparative sensitivity of five species ofmacrophytes and six species of algae to atrazine, metribuzin, alachlor and metolachlor. EnvironmentalToxicology and Chemistry 17(9) 1830-1834.

48. FARMER, D., HILL, I.R., MAUND, S.J. (1995) A compaarison of the fate and effects of two pyrethroidinsecticides (lamda-cyhalothrin and cypermethrin) in pond mesocosms. Ecotoxicology 4: 219-244.

49. FLIEDNER, A., REMDE, A., NIEMANN, R., SCHAFERS, C. (1997) Effects of the organotin pesticideazocyclotin in aquatic microcosms. Chemosphere 35: 209-222.

50. FLEMER, D.A., RUTH, B.A., BUNDRICK, C.M.,MOORE, J.C. (1997) Laboratory effects of microcosmsize and the pesticide chlorpyrifos on benthic macroinvertebrate colonization of soft estuarine sediments.Marine Environmental Research 43(4): 243-263.

51. FORNEY, D. R., DAVIES, D. E. (1981) Effects of low concentrations of herbicides on submersed plants.Weed Sci. 29: 677-685.

52. GENONI, G.P. (1992) Short-term effect of a toxicant on scope for change in ascendency in a microcosmcommunity. Ecotoxicology and Environmental Safety 24: 179-191.

53. GIDDINGS, J.M., BIEVER, R.C., ANNUNZIATO, M.F., HOSMER, A.J. (1996) Effects of diazinon on largeoutdoor pond microcosms. Environmental Toxicology and Chemistry 15(5): 618-629.

54. GIDDINGS, J.M. (1997) Aquatic mesocosm studies and field studies with pyrethroids: observed effects andtheir ecological significance. Springborn Laboratories Report 97-6-7014.

55. GIDDINGS, J.M., BIEVER, R.C., RACKE, K.D. (1997) Fate of chlorpyrifos in outdoor pond microcosmsand effects on growth and survival on bluegill sunfish. Environmental Toxicology and Chemistry 16(11):2353-2362.

56. GIRLING, A.E., PASCOE, D., JANSSEN, C.R., PEITHER, A., WENZEL, A., SCHAFER, H.,NEUMEIER, B., MITCHELL, G.C., TAYLOR, E.J., MAUND, S.J., LAY, J.P., JUTTNER, I.,CROSSLAND, N.O., STEPHENSON, R.R., PERSONNE, G. (2000) Development of methods forevaluating toxicity to freshwater ecosystems. Ecotoxicology and Environmental Safety, 45(2): 148-176.

57. GIRLING, A.E., TATTERSFIELD, L., MITCHELL, G.C., CROSSLAND, N.O., PASCOE, D.,BLOCKWELL, S.J., MAUND, S.J., TAYLOR, E.J., WENZEL, A., JANSSEN, C.R., JUTTNER, I.(2000) Derivation of predicted no-effect concentrations for lindane, 3,4-dichloroaniline, atrazine andcopper. Ecotoxicology and Environmental Safety 46: 148-162.

58. GOMEZ, A., CECCINE, G., SNELL, T.W. (1997) Effects of pentachlorophenol on predator-preyinteractions of two rotifers. Aquatic Toxicology 37: 271-282.

59. GORBI, G., CORRADI, M.G. (1993) Chromium toxicity on two linked trophic levels. Ecotoxicology andEnvironmental Safety 25: 64-71.

60. GRADE, R., GONZALEZ-VALERO, J., HOCHT, P., PFEIFLE, V. (2000) A higher tier flow-throughtoxicity test with the green alga Selenastrum capricornutum. Sci. Total Environ. 247: 355-361.


59

61. GREEN, D.W.J., WILLIAMS, K.A., PASCOE, D. (1986) The acute and chronic toxicity of cadmium todifferent life history stages of the freshwater crustacean Asellus aquaticus (L). Archives EnvironmentalConatmination and Toxicology 15: 465-471.

62. GRUESSNER, B., WATZIN, M.C. (1996) Response of aquatic communities from a Vermont stream toenvironmentally realistic atrazine exposure in laboratory microcosms. Environmental Toxicology andChemistry 15(4), 410-419.

63. GURNEY, W.S.C., MCCAULEYE., NISBET, R.M., MURDOCH, W.W. (1990) The physiological ecologyof Daphnia: a dynamic model of growth and reproduction. Ecology 71: 716-732.

64. HAITZER, M., HOSS, S., TRAUNSPURGER, W., STEINBERG, C. (1998) Effects of dissolved organicmatter (DOM) on the bioconcentration of organic chemicals in aquatic organisms: a review. Chemosphere37(7): 1335-1362.

65. HALLAM, T.G., LASSITER, R.R. (1994) Individual-based mathematical modelling approaches inecotoxicology: a promising direction for aquatic population and community ecological risk assessment.In Kendall, R.J., Lacher, T.E. (eds) Wildlife toxicology and population modelling. Lewis Publishers,Boca Raton.

66. HAM, L., QUINN, R., PASCOE, D. (1995) Effects of cadmium on the predator-prey interaction betweenthe turbellarian Dendrocoelum lacteum (Muller, 1774) and the isopod crustacean Asellus aquaticus (L).Archives of Environmental Contamination and Toxicology 29: 358-365.

67. HAMER, M.J., MAUND, S.J., HILL, I.R. (1992) Laboratory methods for evaluating the impact ofpesticides on water/sediment organisms. Proceedings of the British Crop Protection Council Conference(Pests and diseases), Brighton, UK.

68. HAMER, M.J., GOGGIN, U.M., MULLER, K., MAUND, S.J. (1999) Bioavailability of lamda-cyhalothrinto Chironomus riparius in sediment-water and water-only systems. Aquatic Ecosystem Health andManagement 2: 403-412.

69. HAMERS, T., KROGH, P.H. (1997) Predator-prey relationships in a two-species toxicity test system.Ecotox. Environ. Safety 37: 202-212.

70. HAMILTON, P.B., JACKSON, G.S., KAUSHIK, N.K., SOLOMON, K.R., STEPHENSON, G.L. (1988)The impact of two applications of atrazine on the plankton communities of in situ enclosures. AquaticToxicology 13: 123-140.

71. HANAZATO, T. (1998) Response of a zooplankton community to insecticide application in experimentalponds: a review and the implications of the effects of chemicals on the structure and functioning offreshwater communities.

72. HANDY, R.D. (1994) Intermittent exposure to aquatic pollutants: assessment, toxicity and sublethalresponses in fish and invertebrates. Mini review: Comp Biochem Physiol C 107: 171-184.

73. HANSEN, C.R., KAWATSKI, J.A. (1976) Application of the 24 h post-exposure observation to acutetoxicity studies with invertebrates. J Fish Res Board Can 33: 1198-1201.

74. HENDLEY, P., GIDDINGS, J. (1999) Draft report of the Aquatic Workgroups of ECOFRAM (EcologicalCommittee on FIFRA Risk Assessment) - Aqex_ecofram_Peer01_may499.doc.

75. HICKIE, B.E., MCCARTY, L.S., DIXON, D.G. (1995) A residue-based toxicokinetic model for pulseexposure toxicity in aquatic systems. Environ. Toxicol. Chem. 14: 2187-2197.

76. HINMAN, M. L., KLAINE, S. J. (1992) Uptake and translocation of selected organic pesticides by therooted aquatic plant Hydrilla verticillata Royle. Environ. Toxicol.Chem. 26: 609-613.

77. HOLDWAY, D.A., BARRY, M.J., LOGAN, D.C., ROBERTSON, D., YOUNG, V., AHOKAS, J.T. (1994)Toxicity of pulse-exposed fenvalerate and esfenvalerate to larval australian crimson-spotted rainbow fish(Melanotaenia fluviatilis). Aquatic Toxicology 28(3-4): 169-187.

78. HOSMER, A.J., WARREN, L.W., WARD, T.J. (1998) Chronic toxicity of pulsed-dosed fenoxycarb toDaphnia magna exposed to environmentally realistic concentrations. Environ. Toxicol. Chem. 17: 1860-1866.

79. HUTCHINSON, T.H., SOLBE, J., KLOEPPER-SAMS, P.J. (1998) Analysis of the ECETOC aquatic toxicity(EAT) database.III- Comparative toxicity of chemical substances to different life stages of aquaticorganisms. Chemosphere 36:129-142.


60

80. ISNARD, P., VASSEUR, P., GRAFF, L., NARBONNE, J.F., CLERANDEAU, C., BUDZINSKI, H.,AUGAGNEUR, S., BASTIDE, J., CAMBON, J.P., CELLIER, P., ROMAN, G. (2000) Comparing theecotoxicity of chemicals in standard media and natural waters. Poster presented at the 3rd SETAC WorldCongress, 21-25 May, 2000, Brighton, UK.

81. JOHNSON, P.C., KENNEDY, J.H., MORRIS, R.G., HAMBLETON, F.E. (1994) Fate and effects ofcyfluthrin (pyrethroid insecticide) in pond mesocosms and concrete microcosms. In Graney, R.L., Kennedy,J.H., Rodgers, J.H. (eds) Aquatic mesocosm studies in ecological risk assessment. Lewis Publishers,Michigan.

82. KALLANDER, D.B., FISHER, S.W., LYDY, M.J. (1997) Recovery following pulsed exposure toorganophosphorous and carbamate insecticides in the midge, Chironomus riparius. Archives ofEnvironmental Contamination and Toxicology 33(1): 29-33.

83. KAY, S. H. HALLER, W. T., GARRARD, L. A. (1984) Effects of heavy metals on water hyacinths(Eichhornia crassipes (Mart.)Solms). Aquat. Toxicol. 5: 117-128.

84. KERSTING, K. (1991) Microecosystems state and its response to the introduction of a pesticide. Verh. Int.Verein. Limnol. 23: 1641-1646.

85. KERSTING, K., VAN WIJNGAARDEN, P.A. (1999) Effects of a pulsed treatment with the herbicide afalon(active ingredient linuron) on macrophyte-dominated mesocosms. I. responses of ecosystem metabolism.Environmental Toxicology and Chemistry. 18(12): 2859-2865.

86. KEVAN, S.H., DIXON, D.G. (1996) Effects of age and coion (K+ and Na+) on the toxicity of thiocyanate torainbow trout (Oncorhynchus mykiss) during pulse or continuous exposure. Ecotoxicology andEnvironmental Safety 35: 288-193.

87. KLUTTGEN, B., KUNTZ, N., RATTE, H.T. (1996) Combined effects of 3,4-dichloroaniline and foodconcentration on life table data of two related clacocerans, Daphnia magna and Ceriodaphnia quadrangula.Chemosphere 32: 2015-2028.

88. KOOIJMAN, S.A.L.M. (1981) Parametric analyses of mortality rates in bioassays. Water Research 17:747-759.

89. KOOIJMAN, S.A.L.M. (1987) A safety factor for LC50 values allowing for differences in sensitivityamongst species. Water Research 21: 269-276.

90. KOOIJMAN, S.A.L.M., METZ, J.A.J. (1984) On the dynamics of chemically stressed populations: thededuction of population consequences from effects on individuals. Ecotoxicol. Environ. Safety 8: 254-274.

91. KOOIJMAN, S.A.L.M., BEDAUX, J.J.M. (1996) The analysis of aquatic toxicity data (includes DEBtox,vers. 1.0) VU University Press, Amsterdam.

92. KOSTEL, J.A., WANG, H., AMAND, A.L.S., GRAY, K.A. (1999) Use of a novel laboratory streamsystem to study the ecological impact of PCB exposure in a periphytic biolayer. Water Research 33(18):3735-3748.

93. KREIGER, K.A., BAKER, D.B., KRAMER, J.W. (1988) Effects of herbicides on stream aufwuchsproductivity and nutrient uptake. Archives of Environmental Contamination and Toxicology 17: 299-306.

94. KUKKONEN, J., OIKARI, A. (1991) Bioavailability of organic pollutants in boreal waters with varyinglevels of dissolved organic material. Water Research 25: 455-463.

95. KUKKONEN, J., PELLINEN, J. (1994) Binding of organic xenobiotics to dissolved organicmacromolecules:comparison of analystical methods. Science of the Total Environment 152: 19-29.

96. KUSK, K.O. (1996) Bioavailability and effect of pirimicarb on Daphnia magna in a laboratoryfreshwater/sediment system. Archives of Environmental Contamination and Toxicology 31(2):252-255.

97. LANDRUM, P.F., REINHOLD, M.D., NIHART, S.R., EADIE, B.J. (1985) Predicting the bioavailability oforganic xenobiotics to Pontoporeia hoyi in the presence of humic and fulvic materials and naturaldissolved organic matter. Environ. Toxicol. Chem. 4, 459-467.

98. LANDRUM, P.F., LEE, H., LYDY, M.J. (1992) Toxicokinetics in aquatic systems: model comparisons anduse in hazard assessment. Environ. Toxicol. Chem. 11: 1709-1725.


61

99. LEEUWANGH, P., BROCK, T.C.M., KERSTING, K. (1994) An evaluation of four types of freshwatermodel ecosystem for assessing the hazard of pesticides. Human and Experimental Toxicology, 13: 888-899.

100. LEFFLER, J.W. (1981) Aquatic microcosms and stress criteria for assessing environmental impact oforganic chemicals. Subcontract No. T6411 (7197) 025. Report. USEPA, OPPTS, Washington, DC.

101. MACEK, K.J., SLEIGHT, B.H. (1977) Utility of toxicity tests with embryos and fry of fish in evaluatinghazards associated with the chronic toxicity of substances to fishes. In Mayer, F.L. and Hamelink, J.L.(eds) Aquatic Toxicology and Hazard Evaluation. ASTM STP, Philadelphia.

102. MANCINI, J.L. (1983) A method for calculating effects on aquatic organisms of time-varying exposureconcentrations. Water Researh 17: 1355-1362.

103. MARUBINI, E., VALSECCHI, M.G. (1995) Analysing survival data from clinical trials and observationalstudies. John Wiley and Sons, NY.

104. MAUND, S.J., TAYLOR, E.J., PASCOE, D. (1992) Population responses of the freshwater amphipodcrustacean Gammarus pulex (L) to copper. Freshwater Biol. 28:29-36.

105. MAUND, S.J., SHERRATT, T.N., STICKLAND, T., BIGGS, J., WILLIAMS, N., SHILLABEER, N.,JEPSON, P. (1997) Ecological considerations in risk assessment for pesticides in aquatic ecosystems.Pesticide Science 49: 185-190.

106. MAUND, S.J., HAMER, M.J., WARINTON, J.S. AND KEDWARDS, T.J. (1998) Aquatic ecotoxicologyof the pyrethroid insecticide lamda-cyhalothrin: considerations for higher-tier aquatic risk assessment.Pesticide Science 54(4): 408-417.

107. MCCAHON, C.P., PASCOE, D. (1988) Cadmium toxicity to the freshwater amphipod Gammarus pulex(L.) during the molt cycle. Freshwater Biol 19: 197-203.

108. MCCARTHY, J.F., JIMINEZ, B.D., BARBEE, T. (1985) Effect of dissolved humic material onaccumulation of polycyclic aromatic hydrocarbons: structure-activity relationships. Aquatic Toxicology,7:15-24.

109. MCKIM, J.M. (1977) Evaluation of tests with early life stages of fish for predicting long-term toxicity.Journal of the Fish Research Board of Canada 34: 1148-1154.

110. MCKIM, J.M. (1985) Early life stage toxicity tests. In Rand, G.M. and Petrocelli, S.R. (Eds.) Fundamentalsof aquatic toxicology, First edition. Hemisphere Publishing Corporation, New York.

111. MEYER, J.S., GULLEY, D.D., GOODRICH, M.S., SZMANIA, D.C., BROOKS, A.S. (1995) Modelingtoxicity due to intermittent exposure of rainbow trout and common shiners to monochloroamine.Environmental Toxicology and Chemistry 14(1): 165-175.

112. MILLER, R.G. (1981) Survival Analysis. Wiley and Sons, NY.

113. NADDY, R.B., JOHNSON, K.A., KLAINE, S.J. (2000) Response of daphnia magna to pulsed exposures ofchlorpyrifos. Environ. Toxicol. Chem. 19(2): 423-431.

114. NEUGEBAUR, K., ZIERIS, F.J., HUBER, W. (1990) Ecological effects of atrazine on two outdoorartificial freshwater ecosystems. Wasser Abwasser Forsch 23: 11-17.

115. NEWMAN, M.C., McCLOSKEY, J. (1996) Time-to-event analysis of ecotoxicity data. Ecotoxicology 5:187-196.

116. NEWMAN, M., OWNBY, D.R., MEZIN, L.C.A., POWELL, D.C., CHRISTENSSEN, T.R.L., LERBERG,S.B., ANDERSON, B.A. (2000) Applying species-sensitivity distributions in ecological risk assessment:assumptions of distribution type and sufficient numbers of species. Environ. Toxicol. Chem. 19(2): 508-515.

117. OKKERMAN, P.C., PLASSCHE, E.J.V.D., EMANS, H.J.B., CANTON, J.H. (1993) Validation of someextrapolation methods with toxicity data derived from multiple species experiments. Ecotoxicology andEnvironmental Safety 25: 341-359.

118. ORIS, J. T., HALL, A. T. & TYLKA, J. D. (1990) Humic acids reduce the photo induced toxicity ofanthracene to fish and Daphnia. Environmental Toxicology and Chemistry 9: 575-584.

119. PACK, S. (1993) A review of statistical data analysis and experimental design in OECD aquatic toxicologytest guidelines. Report to the OECD, Paris.


62

120. PARROT, J.L., HODSON, P.V., SERVOS, M.R., HUESTIS, S.L., DIXON, D.G. (1995) Relative potencyof polychlorinated dibenzo-p-dioxins and dibenzofurans for inducing mixed-function oxygenase activityin rainbow trout. Environ. Toxicol. Chem. 14: 1041-1050.

121. PARSONS, J.T., STURGEONER, G.A. (1991) Acute toxicity of permethrin, fenitrothion, carbaryl andcarbofuran to mosquito larvae during single or mutliple-pulse exposures. Environmental Toxicology andChemistry 10: 1229-1233.

122. PASCOE, D., SHAZILI, N.A.M. (1986) Episodic pollution – a comparison of brief and continuousexposure of rainbow trout to cadmium. Ecotoxicol Environ Saf 12: 189-198.

123. PASCOE, D., WENZEL, A., JANSSEN, C., GIRLING, A.E., JUTTNER, I., FLIEDNER, A.,BLOCKWELL, S.J., MAUND, S.J., TAYLOR, E.J., DIEDRICH, M., PERSOONE, G., VERHELST, P.,STEPHENSON, R.R., CROSSLAND, N.O., MITCHELL, G.C., PEARSON, N., TATTERSFIELD, L.,LAY, J-P., PEITHER, A., NEUMEIR, B., VELLETTI, A-R. (2000) The development of toxicity testsfor freshwater pollutants and their validation in stream and pond mesocosms. Water Research 34(8):2323-2329.

124. PAULI, B.D., COULSON, D.R., BERRILL, M. (1999) Sensitivity of amphibian embryos and tadpoles tomimic 240 lv insecticide following single or double exposures. Environ. Toxicol. Chem. 18(11) 2538-2544.

125. PEITHER, A., JUTTNER, I., KETTRUP, A., LAY, J-P. (1996) A pond mesocosm study to determine thedirect and indirect effects of lindane on a natural zooplankton community. Environmental Pollution93(1): 49-56.

126. PERSOONE G. & JANSSEN C.R. (1993) Freshwater Invertebrate Tests. In Handbook ofEcotoxicologyVolume 1. Ed. P. Calow. Pages51-65, Blackwell Science Ltd, London.

127. PUSEY, B.J., ARTHRINGTON, A.H., MCCLEAN, J. (1994) The effect of a pulsed application ofchlorpyrifos on macroinvertebrate communities in outdoor artitificial stream system. Ecotoxicology andEnvironmental Safety 27: 221-250.

128. ROEX, E.W.M., VANGESTEL, C.A.M., VAN WEZEL, A.P., VANSTRAALEN, N.M. (2000) Ratiosbeween acute aquatic toxicity and effects on population growth rates in relation to toxicant mode ofaction. Environmental Toxicology and Chemistry 19(3): 685-693.

129. ROMAN, G., ISNARD, P., JOUANY, J.M. (1999) Critical analysis of methods for assessment of predictedno-effect concentration. Ecotoxicology and Environmental Safety, 43(2): 117-125.

130. RUSSOM, C.L., BRADBURY, S.P., BRODERIUS, S.J., HAMMERMEISTER, D.E., DRUMMOND, R.A.(1997) Predicting the modes of toxic action from chemical structure: acute toxicity in the fatheadminnow (Pimephales promelas). Environmental Toxicology and Chemistry 16(5): 948-967.

131. SANCHEZ, M., FERRANDO, M.D., SANCHO, E., ANDREU, E. (1999) Assessment of the toxicity of apesticide with a two generation reproduction test using Daphnia magna. Comparative Biochemistry andPhysiology Part C 124: 247-252.

132. SCHULZ, R., LIESS, M. (in press) Toxicity of aqueous-phase and suspended-particle-associatedfenvalerate: chronic effects following pulse-dosed exposure of Limnephilus lunatus (trichoptera).Environ. Toxicol Chem.

133. SEIM, W.K., CURTIS, L.R., GLEN, S.W., CHAPMAN, G.A. (1984) Growth and survival of developingsteelhead trout (Salmo gairdneri) continuously or intermittently exposed to copper. Can J. Fish. Aquat.Sci. 41: 433-438.

134. SHAW, J.L., MAUND, S.J., HILL, I.R. (1995) Fathead minnow (Pimephales promelas Rafinesque)reproduction in outdoor microcosms: an assessment of the ecological effects of fish density. EnvironToxicol Chem 14: 1763-1772.

135. SHAZILI, N.A.M., PASCOE, D. (1986) Variable sensitivity of rainbow trout (Salmo gairdneri) eggs andalevins to heavy metals. Bulletin Environmental Contamination and Toxicology 36: 468-474.

136. SHERRATT, T.N., ROBERTS, G., WILLIAMS, P., WHITLFIELD, M., BIGGS, J., SHILLABEER, N.,MAUND, S.J. (1999) A life-history approach to predicting the recovery of aquatic invertebratepopulations after exposure to xenobiotic chemicals. Environmental Toxicology and Chemistry 18(11):2512-2518.


63

137. SHILLABEER, N., SMYTH, D.V., TATTERSFIELD, L. (2000) Higher tier risk assessment ofagrochemicals, incorporating sediment into algal test systems. Proceedings of th BCPC Conference –Pests and Diseases, Brighton, 2000. pp 359-364

138. SIBLY, R.M. (1996) Effects of pollutants on individual life histories and population growth rates. InNewman, M.C., Jagoe, C.H. (eds) Ecotoxicology: a hierarchical approach. Lewis Publishers, BocaRaton, Fl.

139. SIDDENS, L.K., SEIM, W.K., CURTIS, L.R., CHAPMAN, G.A. (1986) Comparison of continuous andepisodic exposure to acidic aluminium-contaminated waters of brook trout (Salvelinus fontinalis). Can. J.Fish. Aquat. Sci. 43: 2036-2040.

140. SKALSKI, J.R. (1981) Statistical incosistencies in the use of no-observed-effect levels in toxicity testing.In: Branson, D.R. and Dickson, K.L. (eds) Aquatic Toxicology and Hazard Assessment: FourthConference ASTM STP737. ASTM, Philadelphia.

141. SLOOFF, W., CANTON, J.H. (1983) Comparison of the susceptibility of 11 freshwater species to 8chemical compounds. 2. (semi)chronic toxicity tests. Aquatic Toxicology 4(3): 271-282.

142. SLOOFF, W., VAN OERS, J.A.M., DEZWART, D. (1986) Margins of uncertainty in ecotoxicologicalhazard assessment. Environ Toxicol Chem. 5: 841-852.

143. SOLOMON, K.R. (1996) Overview of recent developments in ecotoxicological risk assessment. RiskAnalysis 16(5): 627-633.

144. SOLOMON, K.R., BAKER, D., RICHARD, P.R., DIXON, K.D., KLAINE, S.J., LA POINT, T.W.,KENDALL, R.J., WEISSKOPF, C.P., GIDDINGS, J.M., GIESY, J.P., HALL, L.W., WILLIAMS, W.M.(1996) Ecological risk assessment of atrazine in north american surface waters. Environ. Toxicol. Chem.,15(1) 31-76.

145. STARK, J.D., BANKEN, J.A.O. (1999) Importance of population structure at the time of toxicantexposure. Ecotoxicology and Environmental Safety 42: 282-287.

146. STEPHAN, C.E., MOUNT, D.I., HANSEN, D.J., GENTILE, J.H., CHAPMAN, G.A., O’NEILL, R. (1985)Guidelines for deriving numerical national water quality criteria for the protection of aquatic organismsand their uses. USEPA, PB85-227049.

147. STUIJFZAND, S.C., POORT, L., GREVE, G.D., VANDERGEEST, H.G., KRAAK, M.H.S. (2000)Variables determining the impact of diazinon on aquatic insects: taxon, developmental stage, andexposure time. Environmental Toxicology and Chemistry 19(3): 582-587.

148. SUEDEL, B.C., ROGERS, J.H. (1994) Development of a formulated reference sediments for freshwaterand estuarine sediment toxicity testing. Environmental Toxicology and Chemistry 13: 1163-1175.

149. SUEDEL, B.C., DEAVER, E., RODGERS, J.H. (1996) Formulated sediment as a reference and dilutionsediment in definitive toxicity tests. Archives of Environmental Toxicology and Chemistry 30(1): 47-52.

150. SUTER, G.W., BARNTHOUSE, L.W., BRECK, J.E., GARDNER, R.H. AND O’NEILL, R.V. (1985)Extrapolation from the laboratory to the field: how uncertain are you? In: Cardwell, R.D., Bahner, R.C.(Eds) Aquatic Toxicology and Hazard Assessment. American Society for Testing of Materials,Philadelphia, Pa.

151. TAUB, F.B. (1969) A biological model of a freshwater community: a gnotobiotic ecosystem. Limnol.Oceanogr. 14: 136-142.

152. TAYLOR, E.J., BLOCKWELL, S.J., MAUND, S.J., PASCOE, D. (1992) Effects of lindane on the lifecycle of a freshwater invertebrate Chironomus riparius Meigen (Insecta: Diptera). Arch. Environ.Contam. Toxicol. 24:145-150

153. TAYLOR, E.J., MORRISON, J.E., BLOCKWELL, S.J., TARR, A., PASCOE, D. (1995) Effects of lindaneon the predator-prey interaction between Hydra oligactis Pallas and Daphnia magna Strauss. ArchivesEnvironmental Contamination and Toxicology 29: 291-296.

154. THURSTON, R.V., CHAKOUMATOS, C., RUSSO, R.C. (1981) Effect of fluctuating exposures on theacute toxicity of ammonia to rainbow trout (Salmo gardneiri) and cut throat trout (S. clarki). WaterResearch 15: 911-917.


64

155. TRAUNSPURGER, W., SCHAFER, H., REMDE, A. (1996) Comparative investigation on the effect of aherbicide on aquatic organisms in single species tests and aquatic microcosms. Chemosphere 33(6): 1129-1141.

156. USEPA (1984). Estimating concern levels for concentrations of chemical substances in the environment.Environmental Effects Branch, Health and Environmental Review Division, USEPA.

157. USEPA (1998) Guidelines for ecological risk assessment. USEPA Risk Assessment Forum, Washington,DC.

158. USEPA (2000) Technical report of the implementation plan for probabilitic ecological assessments: aquaticsystems. USEPA, Duluth, MN.

159. VAAL, M.A. VANLEEUWEN, C.J., HOEKSTRA, J.A., HERMENS, J.L.M. (2000) Variation onsensitivity of aquatic species to toxicants: Practical consequences for effect assessment of chemicalsubstances. Environmental Management 25(4): 415-423

160. VAN DEN BRINK, P.J., HARTGERS, E.M., FELTWEIS, U., CRUM, S.J.H, VAN DONK, E., BROCK,T.C.M. (1997) Sensitivity of macrophyte-domiated freshwater microcosms to chronic levels of theherbicide linuron. Ecotoxicology and Environmental Safety 38(1): 13-24.

161. VAN DEN BRINK, P.J.,HATTINK, J., BRANSEN, F., VAN DONK, E., BROCK, T.C.M. (2000) Impactof the fungicide carbendazim in freshwater microcosms. II. zooplankton, primary producers and finalconclusions. Aquatic Toxicology 48, 251-264.

162. VAN DEN BRINK, P.J., PSTHUMA, L., BROCK, T.C.M. (2000) Verification of the SSD-concept: fieldrelevance and implications for ecological risk assessment. Poster presented at the SETAC WorldCongress, Brighton, May 2000.

163. VAN GEEST, G.J., ZWAARDEMAKER, N.G., VAN WIJNGAARDEN, R.P.A., CUPPEN, J.G.M. (1999)Effects of a pulsed treatment with the herbicide afalon (active ingredient linuron) on macrophyte-dominated mesocosms. II. structural responses. Environmental Toxicology and Chemistry 18(12): 2866-2874.

164. VAN STRAALEN, N.M., DENNEMAN, G..A.J. (1989) Ecotoxicological evaluation of soil quality criteria.Ecotoxicol Environ Safety 18: 241-251.

165. VAN STRAALEN, N.M., SCHOBBEN, J.H.M., TRAAS, T.P. (1992) The use of ecotoxicological riskassessment in deriving maximum acceptable half-lifes of pesticides. Pesticide Science 34: 227-231.

166. VERHAAR, H.J.M., VAN LEEUWEN, C.J., HERMENS, J.L.M. (1992) Classifying environmentalpollutants 1: structure-activity relationships for prediction of aquatic toxicity. Chemosphere 25: 471-491.

167. WAGNER, C., LOKKE, H. (1991) Estimation of ecotoxicological protection levels from NOEC toxicitydata. Water Research 25(10): 1237-1242.

168. WANG, M.P., HANSON, S.A. (1985) The acute toxicity of chlorine on freshwater organisms: time-concentration relationships of constant and intermittent exposures. In: R.C. Bahner and D.J. Hansen(Eds) Aquatic toxicology and hazard assessment: eigth symposium. STP891. ASTM, Philadelphia, PA.

169. WANG, W., WILLIAMS, J. M. (1990) The use of phytotoxicity tests (common duckweed, cabbage andmillet) for determining effluent toxicity. Envionment Monitor Assess 14: 45-58.

170. WANG, H., KOSTEL, J., AMAND, A.L.S., GRAY, K.A. (1999) The response of a laboratory streamsystem to PCB exposure: study of periphytic and sediment accumulation patterns. Water Research33(18): 3749-3761.

171. WARD, S., ARTHINGTON, A.H., PUSEY, B.J. (1995) The effects of chronic application of chlorpyrifoson the macroinvertebrate fauna in an outdoor artificial stream system. species responses. Ecotoxicologyand Environmental Safety 30(1): 2-23.

172. WEBBER, E.C., DEUTSCH, W.G., BAYNE, D.R., SEESOCK, W.C. (1992) Ecosystem-level testing of asynthetic pyrethroid insecticide in aquatic mesocosms. Environmental Toxicology and Chemistry 11: 87-105.

173. WEINSTEIN, J.E., ORIS, J.T. (1999) Humic acids reduce the bioaccumulation and photoinduced toxicityof fluoranthene to fish. Environmental Toxicology and Chemistry 18(9): 2087-2094.


65

174. WILLIAMS, K.A., GREEN, D.W.J., PASCOE, D., GOWER, D.E. (1986) The acute toxicity of cadmiumto different larval stages of Chironomus riparius (diptera: chironomidae) and its ecological significancefor pollution regulation. Oecologia 70: 362-366

175. WRIGHT, A. (1976) The use of recovery as a criterion for toxicity. Bull Environ Contam Toxicol 15: 747-749.

176. ZITKO,V., MCCLEESE, D.W., METCALFE, C.D., CARLSON, W.G. (1979) Toxicity of permethrin,decamethrin and related pyrethroids to salmon (Salmo salar) and lobster (Homerus americanus). Bull.Environ. Contam. Toxicol. 21: 338-343.


66

APPENDIX A – ORGANISATIONS CONTACTED DURING THE

STUDY*

Alterra Green World Research American Cyanamid

AstraZeneca Ltd Aventis Crop Science

BASF Aktiengesselschaft Bayer-AG

BBA CEFAS

Covance Inc Crop Protection Association

DETR Dow AgroSciences

DuPont Crop Protection Environment Agency

Fraunhofer Institut Health & Safety Executive

Huntingdon Life Sciences INIA

Inveresk Research International Monsanto Company

Novartis Crop Protection AG Pesticides Safety Directorate

RIVM SafePharm Laboratories Ltd

Shell Chemicals Stirling University

Unilever University of London

University of Sheffield USEPA

Veterinary Medicines Directorate Water Research Centre

Wright State University Zeneca Agrochemicals

* Company names at time questionnaire issued


67

APPENDIX B – AVAILABLE SINGLE SPECIES ECOTOXICITY

METHODS

Table A1: Invertebrate species commonly used for fresh water toxicity tests

(Persoone and Janssen, 1993, ASTM 1992)

Group Species Authority recomendation

Protozoans CilatesTetrahymena pyriformis APHA

Vermes PlatyhelminthsDugesia tigrinaAnnelidsLimnodrilus hoffmeisteriTubifex tubifexBranchiura sowerbyiStylodrilus heringianus

ASTM

APHA, FAOAPHA, FAOAPHA, FAOAPHA

Molluscs GastropodsPhysa integraPhysa heterostrophaAmnicola limosa

ASTMASTMASTM

Crustaceans BranchiopodsDaphnia magnaD. PulexD. pulicariaDaphnia spp.Ceriodaphnia spp.RotifersBrachionus calyciflorusB. rubensB. plicatilisAmphipodsGammarus lacustrisG. pseudolimnaeusG. fasciatusHyalella aztecaPontoporeia affinisHyalella spp.MysidsMysis relictaDecapods (crayfish)Plaaemonetes cummingiPalaemonetes kadakiensisGambarus spp.Orconectes rusticusOrconectes spp.Procambarus spp.Pacifastacus lenisculus

APHA, ASTM, FAO, US EPA, OECDASTM, US EPA, OECDASTMOECDUS EPA

ASTMASTMASTM

APHA, ASTM, FAO, US EPAAPHA, ASTM, US EPAAPHA, ASTM, FAO, US EPAAPHA, FAOAPHAUS EPA

APHA, FAO

APHAAPHAAPHA, US EPA, FAO, ASTMAPHAUS EPA, ASTMASTMASTM, US EPA


68

Group Species Authority recomendation

Insect Larvae PlecopteransPteronarcys dorsata (stonefly)P. californica (stonefly)Pteronarcys spp. (stonefly)Hesperoperla lycoriasH. pacificaIsogenus fontalisIsogenus spp.Perlesta placidaParagnetina mediaParagnetina spp.Phasganophora capitataPhadganophora spp.Acroneuria californicaAcroneuria spp.Ephemeropterans (mayflies)Hexagenia bilineataH. limbataH. rigidaHexagenia spp.Ephemerella subvariaE. cornutaE. grandisE. doddsiE. NeedhaniiE. TuberculataEphemerella spp.Stenonema ithacaStenonema sppBaetis spp.TrichopteransBrachycentrus americanusB. occidentalisBrachycentrus spp.Clistoronia magnificaHydropsyche bettiniH. bifidaHydropsyche spp.Macronemum zebratumMacronemum spp.DipteransChironomus plmosusC. attenuatusC. tentansC. californicusChironomus spp.Glyptochironomus labiferusGoeldichironomus labiferusG. holoprasinusTanypus grodhausiTanypus Spp.Tanytarsus dissimilisTanytarsus spp.

APHAAPHAASTM, FAOAPHAAPHAAPHAFAOAPHAAPHAFAOAPHAFAOAPHAFAO

APHA, ASTM, US EPAAPHA, ASTM, US EPAAPHAFAOAPHAAPHAAPHAAPHAAPHAAPHAASTM, FAO, US EPAAPHAFAOASTM, US EPA

APHAAPHAFAOAPHAAPHAAPHAFAOAPHAFAO

APHAAPHAAPHAAPHAASTM, FAO, US EPA, OECD (draft)APHAAPHAAPHAFAOAPHAAPHAFAO


69

Table A2: Fish species commonly used for fresh water toxicity tests (ASTM 1992)

Common name Latin name

Brook trout Salvelinus fontinalis

Coho salmon Oncorhynchus kisutch

Chinook salmon Oncorhynchus tshawytscha

Rainbow trout Oncorhynchus mykiss

Gold fish Carassius auratus

Common Carp Cyprinus carpio

Fathead minow Pimephales promelas

White Sucker Catostomus commersoni

Channel catfish Ictalurus punctatus

Bluegill sunfish Lepomis macrochirus

Green sunfish Lepomis cyanellus

Northern Pike Esox lucius

Threespine stickleback Gasterosteus aculeatus

Zebra fish Brachydanio rerio

Guppy Poecilia reticulata

Table A3. Ampibian species used in ecotoxicity studies

Common name List name

frog Rana sp.

Toad Bufo sp.

Table A4. Plant species used in ecotoxicity studies

Species Reference

Elodea canadenis (water weed) Forney and Davis, 1981; Fairchild et al., 1999

Myriophyllum spicatum (Eurasin water

milfoil)

Forney and Davis, 1981

Vallisneria americana (wild celery) Forney and Davis, 1981

Hydrilla vertcillata Hinman and Klaine, 1992

Ceratophyllum demersum Fairchild et al., 1999

Eichhornia crassipes (water hyacinths) Kay et al., 1984

Lemna minor Fairchild et al., 1999

Ceratophyllum demersum Fairchild et al., 1999

Myriophyllum heterophyllum Fairchild et al., 1999

Najas sp. Fairchild et al., 1999

Brassica oleracea Wang and Williams, 1990

Panicum maliceum Wang and Williams, 1990


70

Table A5. Available ecotoxicity tests for algal species

Organism Reference

Chlamydomonas dysosmos Blanck et al; 1984

Chlorella emmersonii Blanck et al; 1984

Kirchnereilla contorta Blanck et al; 1984

Monoraphidium pusillum Blanck et al; 1984

Scenedesmus obtusiusculus Blanck et al; 1984

Selenastrum capricornutum Blanck et al; 1984

Klebsormidium marinum Blanck et al; 1984

Raphidonema longiseta Blanck et al; 1984

Brumilleriopsis filiformis Blanck et al; 1984

Monodus sunterraneus Blanck et al; 1984

Tribonema aequale Blanck et al; 1984

Synechococcus leopoliensis Blanck et al; 1984

higher tier laboratory aquatic toxicity testing · cranfield centre for ecochemistry 2 foreword...

Documents