Council Regulation (EC) No 440/2008 of 30 May 2008 laying down test methods pursuant to Regulation (EC) No 1907/2006 of the European Parliament and of the Council on the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) (Text with EEA relevance)

Any water in which the test species shows suitable long-term survival and growth may be used as a test water. It should be of constant quality during the period of the test. The pH of the water should be within the range 6,5 to 8,5 , but during a given test it should be within a range of ± 0,5 pH units. Hardness above 140 mg/l (as CaCO3) is recommended. In order to ensure that the dilution water will not unduly influence the test result (for example by complexion of test substance), samples should be taken at intervals for analysis. Measurements of heavy metals (e.g. Cu, Pb, Zn, Hg, Cd and Ni), major anions and cations (e.g. Ca, Mg, Na, K, Cl and SO4), pesticides (e.g. total organophosphorus and total organochlorine pesticides), total organic carbon and suspended solids should be made, for example, every three months where a dilution water is known to be relatively constant in quality. If water quality has been demonstrated to be constant over at least one year, determinations can be less frequent and intervals extended (e.g. every 6 months). Some chemical characteristics of an acceptable dilution water are listed in Appendix 2.

1.6.3.   Test solutions

Test solutions of the chosen concentrations are prepared by dilution of a stock solution.

The stock solution should preferably be prepared by simply mixing or agitating the test substance in the diluition water by using mechanical means (e.g. stirring or ultrasonication). Saturation columns (solubility columns) can be used for achieving a suitable concentrated stock solution.

The use of solvents or dispersants (solubilising agents) may be required in some cases in order to produce a suitably concentrated stock solution. Examples of suitable solvents are acetone, ethanol, methanol, dimethylsulfoxide, dimethylformamide and triethyleneglycol. Examples of suitable dispersants are Cremophor RH40, Tween 80, Methylcellulose 0,01 % and HCO-40. Care should be taken when using readily biodegradable agents (e.g. acetone) and/or highly volatile compounds as these can cause problems with bacterial built-up in flow-through tests. When a solubilising agent is used it must have no significant effects on the fish growth nor visible adverse effects on the juvenile as revealed by a solvent-only control.

For flow-through tests, a system which continually dispenses and dilutes a stock solution of the test substance (e.g. metering pump, proportional diluter, saturator system) is required to deliver a series of concentrations to the test chambers. The flow rates of stock solutions and dilution water should be checked at intervals, preferably daily, during the test and should not vary by more than 10 % throughout the test. A ring-test (2) has shown that, for rainbow trout, a frequency of water removal during the test of six litres/g of fish/day is acceptable (see Section 1.8.2.2).

For semi-static (renewal) tests, the frequency of medium renewal will depend on the stability of the test substance, but a daily water renewal is recommended. If, from preliminary stability tests (see Section 1.4), the test substance concentration is not stable (i.e. outside the range 80-120 % of nominal or falling below 80 % of the measured initial concentration) over the renewal period, consideration should be given to the use of a flow-through test.

1.6.4.   Selection of species

Rainbow trout (Oncorhynchus mykiss) is the recommended species for this test since most experience has been gained from ring-test with this species (1)(2). However, other well documented species can be used but the test procedure may have to be adapted to provide suitable test conditions. For example, experience is also available with zebrafish (Danio rerio) (3)(4) and ricefish (medaka, Oryzias latipes) (5)(6)(7). The rationale for the selection of the species and the experimental method should be reported in this case.

1.6.5.   Holding of fish

The test fish shall be selected from a population of a single stock, preferably from the same spawning, which has been held for at least two weeks prior to the test under conditions of water quality and illumination similar to those used in the test. They should be fed a minimum ration of 2 % body weight per day and preferably 4 % body weight per day throughout the holding period and during the test.

Following a 48 h setting-in period, mortalities are recorded and the following criteria applied:

• — mortalities of greater than 10 % of population in seven days: reject the entire batch;

• — mortalities of between 5 % and 10 % of population: acclimation for seven additional days; if more than 5 % mortality during second seven days, reject the entire batch;

• — mortalities of less than 5 % of population in seven days: accept the batch.

Fish should not receive treatment for disease in the two weeks preceding the test, or during the test.

1.7.   TEST DESIGN

The ‘test design’ relates to the selection of the number and spacing of the test concentrations, the number of tanks at each concentration level and the number of fish per tank. Ideally, the test design should be chosen with regard to:

• — the objective of the study;

• — the method of statistical analysis that will be used;

• — the availability and cost of experimental resources.

The statement of the objective should, if possible, specify the statistical power at which a given size of difference (e.g. in growth rate) is required to be detected or, alternatively, the precision with which the ECx (e.g. with x = 10, 20, or 30, and preferably not less than 10) is required to be estimated. Without this, a firm prescription of the size of the study cannot be given.

It is important to recognise that a design which is optimal (makes best use of resources) for use with one method of statistical analysis is not necessarily optimal for another. The recommended design for the estimation of a LOEC/NOEC would not therefore be the same as that recommended for analysis by regression.

In most of cases, regression analysis is preferable to the analysis of variance, for reasons discussed by Stephan and Rogers (8). However, when no suitable regression model is found (r² < 0,9 ) NOEC/LOEC should be used.

1.7.1.   Design for analysis by regression

The important considerations in the design of a test to be analysed by regression are:

• — The effect concentration (e.g. EC10,20,30) and the concentration range over which the effect of the test substance is of interest, should necessarily be spanned by the concentrations included in the test. The precision with which estimates of effect concentrations can be made, will be best when the effect concentration is in the middle of the range of concentrations tested. A preliminary range-finding test may be helpful in selecting appropriate test concentrations.

• — To enable satisfactory statistical modelling, the test should include at least one control tank and five additional tanks at different concentrations. Where appropriate, when a solubilising agent is used, one control containing the solubilising agent at the highest tested concentration should be run in addition to the test series (see Sections 1.8.3 and 1.8.4).

• — An appropriate geometric series or logarithmic series (9) (see Appendix 3) may be used. Logarithmic spacing of test concentration is to be preferred.

• — If more than six tanks are available, the additional tanks should either be used to provide replication or distributed across the range of concentrations in order to enable closer spacing of the levels. Either of these measures are equally desirable.

1.7.2.   Design for estimation of a NOEC/LOEC using Analysis of Variance (ANOVA)

There should preferably be replicate tanks at each concentration, and statistical analysis should be at the tank level (10). Without replicate tanks, no allowance can be made for variability between tanks beyond that due to individual fish. However, experience has shown (11) that between-tank variability was very small compared with within-tank (i.e. between-fish) variability in the case examined. Therefore a relatively acceptable alternative is to perform statistical analysis at the level of individual fish.

Conventionally, at least five test concentrations in a geometric series with a factor preferably not exceeding 3,2 are used.

Generally, when tests are performed with replicate tanks, the number of replicate control tanks and therefore the number of fish should be the double of the number in each of the test concentrations, which should be of equal size (12)(13)(14). On the opposite, in absence of replicate tanks, the number of fish in the control group should be the same as the number in each test concentration.

If the ANOVA is to be based on tanks rather than individual fish (which would entail either individual marking of the fish or the use of ‘pseudo’ specific growth rates (see Section 2.1.2)), there is a need for enough replication of tanks to enable the standard deviation of ‘tanks-within-concentrations’ to be determined. This means that the degrees of freedom for error in the analysis of variance should be at least 5 (10). If only the controls are replicated, there is a danger that the error variability will be biased because it may increase with the mean value of the growth rate in question. Since growth rate is likely to decrease with increasing concentration, this will tend to lead to an overestimate of the variability.

1.8.   PROCEDURE

1.8.1.   Selection and weighing of test fish

It is important to minimise variation in weight of the fish at the beginning of the test. Suitable size ranges for the different species recommended for use in this test are given in Appendix 1. For the whole batch of fish used in the test, the range in individual weights at the start of the test should ideally be kept to within ± 10 % of the arithmetic mean weight and, in any case, should not exceed 25 %. It is recommended to weight a subsample of fish before the test in order to estimate the mean weigh.

Food should be withheld from the stock population for 24 h prior to the start of the test. Fish should then be chosen at random. Using a general anaesthetic (e.g. an aqueous solution of 100 mg/l tricaine methane sulphonate (MS 222) neutralised by the addition of two parts of sodium bicarbonate per part of MS 222), fish should be weighted individually as wet weights (blotted dry) to the precision given in Appendix 1. Those fish with weights within the intended range should be retained and then should be randomly distributed between the test vessels. The total wet weight of fish in each test vessel should be recorded. The use of anaesthetics likewise handling of fish (including blotting and weighing) may cause stress and injuries to the juvenile fish, in particular for those species of small size. Therefore handling of juvenile fish must be done with the utmost care to avoid stressing and injuring test animals.

The fish are weighed again on day 28 of the test (see Section 1.8.6). However, if it is deemed necessary to recalculate the food ration, fish can be weighed again on day 14 of the test (see Section 1.8.2.3). Other method as photographic method could be used to determine changes in fish size from which food rations could be adjusted.

1.8.2.   Conditions of exposure

1.8.2.1.   Duration

The test duration is ≥ 28 days.

1.8.2.2.   Loading rates and stocking densities

It is important that the loading rate and stocking density is appropriate for the test species used (see Appendix 1). If the stocking density is too high, then overcrowding stress will occur leading to reduced growth rates and possibly to disease. If it is too low, territorial behaviour may be induced which could also affect growth. In any case, the loading rate should be low enough in order that a dissolved oxygen concentration of at least 60 % ASV can be maintained without aeration. A ring-test (2) has shown that, for rainbow trout, a loading rate of 16 trouts of 3-5 g in a 40-litre volume is acceptable. Recommended frequency of water removal during the test is 6 litres/g of fish/day.

1.8.2.3.   Feeding

The fish should be fed with an appropriate food (Appendix 1) at a sufficient rate to induce acceptable growth rate. Care should be taken to avoid microbial growth and water turbidity. For rainbow trout, a rate of 4 % of their body weight per day is likely to satisfy these conditions (2)(15)(16)(17). The daily ration may be divided into two equal portions and given to the fish in two feeds per day, separated by at least 5 h. The ration is based on the initial total fish weight for each test vessel. If the fish are weighted again on day 14, the ration is then recalculated. Food should be withheld from the fish 24 h prior to weighing.

Uneaten food and fecal material should be removed from the test vessels each day by carefully cleaning the bottom of each tank using a suction.

1.8.2.4.   Light and temperature

The photoperiod and water temperature should be appropriate for the test species (Appendix 1).

1.8.3.   Test concentrations

Normally five concentrations of the test substance are required, regardless of the test design (see Section 1.7.2). Prior knowledge of the toxicity of the test substance (e.g. from an acute test and/or from range-finding studies) should help in selecting appropriate test concentrations. Justification should be given if fewer than five concentrations are used. The highest tested concentration should not exceed the substance solubility limit in water.

Where a solubilising agent is used to assist in stock solution preparation, its final concentration should not be greater than 0,1 ml/l and should preferably be the same in all test vessels (see Section 1.6.3). However, every effort should be made to avoid use of such materials.

1.8.4.   Controls

The number of dilution-water controls depends on the test design (see Sections 1.7-1.7.2). If a solubilising agent is used, then the same number of solubilising-agent controls as dilution-water controls should also be included.

1.8.5.   Frequency of analytical determinations and measurements

During the test, the concentrations of test substance are determined at regular intervals (see below).

In flow-through tests, the flow rates of diluent and toxicant stock solution should be checked at intervals, preferably daily, and should not vary by more than 10 % throughout the test. Where the test substance concentrations are expected to be within ± 20 % of the nominal values (i.e. within the range 80-120 %; see Sections 1.6.2 and 1.6.3), it is recommended that, as a minimum, the highest and lowest test concentrations be analysed at the start of the test and at weekly intervals thereafter. For the test where the concentration of the test substance is not expected to remain within ± 20 % of nominal (on the basis of stability data of the test substance), it is necessary to analyse all test concentrations, but following the same regime.

In semi-static (renewal) tests where the concentration of the test substance is expected to remain within ± 20 % of the nominal values, it is recommended that, as a minimum, the highest and lowest test concentrations be analysed when freshly prepared and immediately prior to renewal at the start of the study and weekly thereafter. For tests where the concentration of the test substance is not expected to remain within ± 20 % of nominal, all test concentrations must be analysed following the same regime as for more stable substances.

It is recommended that results be based on measured concentrations. However, if evidence is available to demonstrate that the concentration of the test substance in solution has been satisfactorily maintained within ± 20 % of the nominal or measured initial concentration throughout the test, then the results can be based on nominal or measured values.

Samples may need to be filtered (e.g. using a 0,45 μm pore size) or centrifuged. Centrifugation is the recommended procedure. However, if the test material does not adsorb to filters, filtration may also be acceptable.

During the test, dissolved oxygen, pH and temperature should be measured in all test vessels. Total hardness, alkalinity and salinity (if relevant) should be measured in the controls and one vessel at the highest concentration. As a minimum, dissolved oxygen and salinity (if relevant) should be measured three times (at the beginning, middle and end of the test). In semi-static tests, it is recommended that dissolved oxygen be measured more frequently, preferably before and after each water renewal or at least once a week. pH should be measured at the beginning and end of each water renewal in static renewal test and at least weekly in flow-through tests. Hardness and alkalinity should be measured once each test. Temperature should preferably be monitored continuously in at least one test vessel.

1.8.6.   Observations

Weight: at the end of the test all surviving fish must be weighed as wet weights (blotted dry) either in groups by test vessel or individually. Weighing of animals by test vessel is preferred to individual weights which require that fish be individually marked. In the case of the measurement of individual weights for determination of individual fish specific growth rate, the marking technique selected should avoid stressing the animals (alternatives to freeze marking may be appropriate, e.g. the use of coloured fine fishing line).

The fish should be examined daily during the test period and any external abnormalities (such as hemorrhage, discoloration) and abnormal behaviour noted. Any mortalities should be recorded and the dead fish removed as soon as possible. Dead fish are not replaced, the loading rate and stocking density being sufficient to avoid effects on growth through changes in number of fish per tank. However, the feeding rate will need to be adjusted.

2. DATA AND REPORTING

2.1.   TREATMENT OF RESULTS

It is recommended that a statistician be involved in both the design and analysis of the test since this test method allows for considerable variation in experimental design as for example, in the number of test chambers, number of test concentrations, number of fish, etc. In view of the options available in test design, specific guidance on statistical procedure is not given here.

Growth rates should not be calculated for test vessels where the mortality exceeds 10 %. However, mortality rate should be indicated for all test concentrations.

Whichever method is used to analyse the data, the central concept is the specific growth rate r between time t1 and time t2. This can be defined in several ways depending on whether fish are individually marked or not or whether a tank average is required.

Image omitted; see official source

Image omitted; see official source

Image omitted; see official source

where:

r1	=	individual fish specific growth rate
---	---	---
r2	=	tank-average specific growth rate
---	---	---
r3	=	‘pseudo’ specific growth rate
---	---	---
w1, w2	=	weights of a particular fish at times t1 and t2, respectively
---	---	---
loge w1	=	logarithm of the weight of an individual fish at the start of the study period
---	---	---
loge w2	=	logarithm of the weight of an individual fish at the end of the study period
---	---	---
loge w1	=	average of the logarithms of the values w1 for the fish in the tank at the start of the study period
---	---	---
loge w2	=	average of the logarithms of the values w2 for the fish in the tank at the end of the study period
---	---	---
t1, t2	=	time (days) at start and end of study period

r1, r2, r3 can be calculated for the 0-28 days period and, where appropriate (i.e. when measurement at day 14 has been done) for the 0-14 and 14-28 days periods.

2.1.1.   Analysis of results by regression (concentration-response modelling)

This method of analysis fits a suitable mathematical relationship between the specific growth rate and concentration, and hence enables the estimation of the ‘ECx ’ i.e. any required EC value. Using this method the calculation of r for individual fish (r1) is not necessary and instead, the analysis can be based on the tank-average value of r (r2). This last method is preferred. It is also more appropriate in case of the use of smallest species.

The tank-average specific growth rates (r2) should be plotted graphically against concentration, in order to inspect the concentration response relationship.

For expressing the relationship between r2 and concentration, an appropriate model should be chosen and its choice must be supported by appropriate reasoning.

If the numbers of fish surviving in each tank are unequal, then the process of model fitting, whether simple or non-linear, should be weighted to allow for unequal sizes of groups.

The method of fitting the model must enable an estimate of, for example, the EC20 and of its dispersion (either standard error or confidence interval) to be derived. The graph of the fitted model should be shown in relation to the data so that the adequacy of the fit of the model can be seen (8)(18)(19)(20).

2.1.2.   Analysis of results for the estimation of the LOEC

If the test has included replication of tanks at all concentration levels, the estimation of the LOEC could be based on an analysis of variance (ANOVA) of the tank-average specific growth rate (see Section 2.1), followed by a suitable method (e.g. Dunnett's or Williams' test (12)(13)(14)(21)) of comparing the average r for each concentration with the average r for the controls to identify the lowest concentration for which this difference is significant at a 0,05 probability level. If the required assumptions for parametric methods are not met — non-normal distribution (e.g. Shapiro-Wilk's test) or heterogeneous variance (Bartlett's test), consideration should be given to transforming the data to homogenise variances prior to performing the ANOVA, or to carrying out a weighted ANOVA.

If the test has not included replication of tanks at each concentration, an ANOVA based on tanks will be insensitive or impossible. In this situation, an acceptable compromise is to base the ANOVA on the ‘pseudo’ specific growth rate r3 for individual fish.

The average r3 for each test concentration may then be compared with the average r3 for the controls. The LOEC can then be identified as before. It must be recognised that this method provides no allowance for, nor protection against, variability between tanks, beyond that which is accounted for by the variability between individual fish. However, experience has shown (8) that between-tank variability was very small compared with within-tank (i.e. between fish) variability. If individual fish are not included in the analysis, the method of outlier identification and justification for its use must be provided.

2.2.   INTERPRETATION OF RESULTS

The results should be interpreted with caution where measured toxicant concentrations in test solutions occur at levels near the detection limit of the analytical method or, in semi static tests, when the concentration of the test substance decreases between freshly prepared solution and before renewal.

2.3.   TEST REPORT

The test report must include the following information:

2.3.1.   Test substance:

• — physical nature and relevant physical-chemical properties;

• — chemical identification data including purity and analytical method for quantification of the test substance where appropriate.

2.3.2.   Test species:

• — scientific name, possibly

• — strain, size, supplier, any pre-treatment, etc.

2.3.3.   Test conditions:

• — test procedure used (e.g. semi-static/renewal, flow-through, loading, stocking density, etc.),

• — test design (e.g. number of test vessels, test concentrations and replicates, number of fish per vessel),

• — method of preparation of stock solutions and frequency of renewal (the solubilising agent and its concentration must be given, when used),

• — the nominal test concentrations, the means of the measured values and their standard deviations in the test vessels and the method by which these were attained and evidence that the measurements refer to the concentrations of the test substance in true solution,

• — dilution water characteristics: pH, hardness, alkalinity, temperature, dissolved oxygen concentration, residual chlorine levels (if measured), total organic carbon, suspended solids, salinity of the test medium (if measured) and any other measurements made,

• — water quality within test vessels: pH, hardness, temperature and dissolved oxygen concentration,

• — detailed information on feeding, (e.g. type of food(s), source, amount given and frequency).

2.3.4.   Results:

• — evidence that controls met the validity criterion for survival, and data on mortalities occurring in any of the test concentrations,

• — statistical analytical techniques used, statistics based on replicates or fish, treatment of data and justification of techniques used,

• — tabulated data on individual and mean fish weights on days 0, 14 (if measured) and 28 values of tank-average or pseudo specific growth rates (as appropriate) for the periods 0-28 days or possibly 0-14 and 14-28,

• — results of the statistical analysis (i.e. regression analysis or ANOVA) preferably in tabular and graphical form and the LOEC (p = 0,05 ) and the NOEC or ECx with, when possible, standard errors, as appropriate,

• — incidence of any unusual reactions by the fish and any visible effects produced by the test substance.

3. REFERENCES

• (1)Solbe J.F. de LG (1987) Environmental Effects of Chemicals (CFM 9350 SLD). Report on a UK Ring Test of a Method for Studying the Effects of Chemicals on the Growth rate of Fish. WRc Report No PRD 1388-M/2.

• (2)Ashley S., Mallett M.J. and Grandy N.J., (1990) EEC Ring Test of a Method for Determining the Effects of Chemicals on the Growth Rate of Fish. Final Report to the Commission of the European Communities. WRc Report No EEC 2600-M.

• (3)Crossland N.O. (1985) A method to evaluate effects of toxic chemicals on fish growth. Chemosphere, 14, p. 1855-1870.

• (4)Nagel R., Bresh H., Caspers N., Hansen P.D., Market M., Munk R., Scholz N. and Höfte B.B., (1991) Effect of 3,4 -dichloroaniline on the early life stages of the Zebrafish (Brachydanio rerio): results of a comparative laboratory study. Ecotox. Environ. Safety, 21, p. 157-164.

• (5)Yamamoto, Tokio., (1975) Series of stock cultures in biological field. Medaka (killifish) biology and strains. Keigaku Publish. Tokio, Japan.

• (6)Holcombe, G.W., Benoit D.A., Hammermeister, D.E., Leonard, E.N. and Johnson, R.D., (1995) Acute and long-term effects of nine chemicals on the Japanese medaka (Oryzias latipes).Arch. Environ. Conta. Toxicol. 28, p. 287-297.

• (7)Benoit, D.A., Holcombe, G.W. and Spehar, R.L., (1991) Guidelines for conducting early life toxicity tests with Japanese medaka (Oryzias latipes). Ecological Research Series EPA-600/3-91-063. U.S. Environmental Protection Agency, Duluth, Minesota.

• (8)Stephan C.E. and Rogers J.W., (1985) Advantages of using regression analysis to calculate results of chronic toxicity tests. Aquatic Toxicology and Hazard Assessment: Eighth Symposium, ASTM STP 891, R C Bahner and D J Hansen, Eds., American Society for Testing and Materials, Philadelphia, p. 328-338.

• (9)Environment Canada, (1992) Biological test method: toxicity tests using early life stages of salmonid fish (rainbow trout, coho salmon, or Atlantic salmon). Conservation and Protection, Ontario, Report EPS 1/RM/28, p. 81.

• (10)Cox D.R., (1958) Planning of experiments. Wiley Edt.

• (11)Pack S., (1991) Statistical issues concerning the design of tests for determining the effects of chemicals on the growth rate of fish. Room Document 4, OECD Ad Hoc Meeting of Experts on Aquatic Toxicology, WRc Medmenham, UK, 10-12 December 1991.

• (12)Dunnett C.W., (1955) A Multiple Comparisons Procedure for Comparing Several Treatments with a Control, J. Amer. Statist. Assoc., 50, p. 1096-1121.

• (13)Dunnett C.W., (1964) New tables for multiple comparisons with a control. Biometrics, 20, p. 482-491.

• (14)Williams D.A., (1971) A test for differences between treatment means when several dose levels are compared with a zero dose control. Biometrics 27, p. 103-117.

• (15)Johnston, W.L., Atkinson, J.L., Glanville N.T., (1994) A technique using sequential feedings of different coloured food to determine food intake by individual rainbow trout, Oncorhynchus mykiss: effect of feeding level. Aquaculture 120, p. 123-133.

• (16)Quinton, J. C. and Blake, R.W., (1990) The effect of feed cycling and ration level on the compensatory growth response in rainbow trout, Oncorhynchus mykiss. Journal of Fish Biology, 37, p. 33-41.

• (17)Post, G., (1987) Nutrition and Nutritional Diseases of Fish. Chapter IX in Testbook of Fish Health. T.F.H. Publications, Inc. Neptune City, New Jersey, USA. p. 288.

• (18)Bruce, R.D. and Versteeg D.J., (1992) A statistical procedure for modelling continuous toxicity data. Environ. Toxicol. Chem. 11, p. 1485-1494.

• (19)DeGraeve, G.M., Cooney, J.M., Pollock, T.L., Reichenbach, J.H., Dean, Marcus, M.D. and McIntyre, D.O., (1989) Precision of EPA seven-day fathead minnow larval survival and growth test; intra and interlaboratory study. report EA-6189 (American Petroleum Institute Publication, No 4468). Electric Power Research Institute, Palo alto, CA.

• (20)Norbert-King T.J., (1988) An interpolation estimate for chronic toxicity: the ICp approach. US Environmental Protection Agency. Environmental Research Lab., Duluth, Minesota. Tech. Rep. No 05-88 of National Effluent Toxicity Assesment Center. Sept. 1988. p. 12.

• (21)Williams D.A., (1972) The comparison of several dose levels with a zero dose control. Biometrics 28, p. 510-531.

Appendix 1

FISH SPECIES RECOMMENDED FOR TESTING AND SUITABLE TEST CONDITIONS

Species	Recommended test temperature range (oC)	Photoperiod (hours)	Recommended range for initial fish weight (g)	Required measurement precision	Loading rate (g/l)	Stocking density (per litre)	Food	Test duration (days)
Recommended species:
Oncorhynchus mykiss rainbow trout	12,5 -16,0	12-16	1-5	to nearest 100 mg	1,2 -2,0	4	Dry propietary salmonid fry food	≥ 28
Other well documented species:
Danio rerio zebrafish	21-25	12-16	0,050 -0,100	to nearest 1 mg	0,2 -1,0	5-10	Live food (Brachionus Artemia)	≥ 28
Oryzias latipes ricefish (Medaka)	21-25	12-16	0,050 -0,100	to nearest 1 mg	0,2 -1,0	5-20	Live food (Brachionus Artemia)	≥ 28

Appendix 2

SOME CHEMICAL CHARACTERISTICS OF AN ACCEPTABLE DILUTION WATER

Substance	Concentrations
Particulate matter	< 20 mg/l
Total organic carbon	< 2 mg/l
Unionised ammonia	< 1 μg/l
Residual chlorine	< 10 μg/l
Total organophosphorus pesticides	< 50 ng/l
Total organochlorine pesticides plus polychlorinated biphenyls	< 50 ng/l
Total organic chlorine	< 25 ng/l

Appendix 3

Logarithmic series of concentrations suitable for toxicity test (9)

Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]	Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]	Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]	Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]	Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]	Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]	Column (Number of concentrations between 100 and 10, or between 10 and 1) [^E0139]
1	2	3	4	5	6	7
100	100	100	100	100	100	100
32	46	56	63	68	72	75
10	22	32	40	46	52	56
3,2	10	18	25	32	37	42
1,0	4,6	10	16	22	27	32
	2,2	5,6	10	15	19	24
	1,0	3,2	6,3	10	14	18
		1,8	4,0	6,8	10	13
		1,0	2,5	4,6	7,2	10
			1,6	3,2	5,2	7,5
			1,0	2,2	3,7	5,6
				1,5	2,7	4,2
				1,0	1,9	3,2
					1,4	2,4
					1,0	1,8
						1,3
						1,0

A series of five (or more) successive concentrations may be chosen from a column. Mid-points between concentrations in column (x) are found in column (2x + 1). The values listed can represent concentrations expressed as percentage per volume or weight (mg/l or μg/l). Values can be multiplied or divided by any power of 10 as appropriate. Column 1 might be used if there was considerable uncertainty on the toxicity level.

C.15.   FISH, SHORT-TERM TOXICITY TEST ON EMBRYO AND SAC-FRY STAGES

This test method has been deleted, as it is no longer recognised as being appropriate for generating information on the ecotoxicological properties of chemicals for the purpose of Regulation (EC) No 1907/2006. The applicable test methods for the endpoint in question appear in Table 3 of Part 0.

C.16.   HONEYBEES — ACUTE ORAL TOXICITY TEST

1. METHOD

This acute toxicity test method is a replicate of the OECD TG 213 (1998).

1.1.   INTRODUCTION

This toxicity test is a laboratory method, designed to assess the oral acute toxicity of plant protection products and other chemicals, to adult worker honeybees.

In the assessment and evaluation of toxic characteristics of substances, determination of acute oral toxicity in honeybees may be required, e.g. when exposure of bees to a given chemical is likely. The acute oral toxicity test is carried out to determine the inherent toxicity of pesticides and other chemicals to bees. The results of this test should be used to define the need for further evaluation. In particular, this method can be used in step-wise programmes for evaluating the hazards of pesticides to bees, based on sequential progression from laboratory toxicity tests to semi-field and field experiments (1). Pesticides can be tested as active substances (a.s.) or as formulated products.

A toxic standard should be used to verify the sensitivity of the bees and the precision of the test procedure.

1.2.   DEFINITIONS

Acute oral toxicity: is the adverse effects occurring within a maximum period of 96h of an oral administration of a single dose of test substance.

Dose: is the amount of test substance consumed. Dose is expressed as mass (μg) of test substance per test animal (μg/bee). The real dose for each bee can not be calculated as the bees are fed collectively, but an average dose can be estimated (totally consumed test substance/number of test bees in one cage).

LD 50 (Median Lethal Dose) oral: is a statistically derived single dose of a substance that can cause death in 50 % of animals when administered by the oral route. The LD50 value is expressed in μg of test substance per bee. For pesticides, the test substance may be either an active substance (a.s.) or a formulated product containing one or more than one active substance.

Mortality: an animal is recorded as dead when it is completely immobile.

1.3.   PRINCIPLE OF THE TEST METHOD

Adult worker honeybees (Apis mellifera) are exposed to a range of doses of the test substance dispersed in sucrose solution. The bees are then fed the same diet, free of the test substance. Mortality is recorded daily during at least 48 h and compared with control values. If the mortality rate is increasing between 24 h and 48 h whilst control mortality remains at an accepted level, i.e. ≤ 10 %, it is appropriate to extend the duration of the test to a maximum of 96 h. The results are analysed in order to calculate the LD50 at 24 h and 48 h and, in case the study is prolonged, at 72 h and 96 h.

1.4.   VALIDITY OF THE TEST

For a test to be valid, the following conditions apply:

• — the average mortality for the total number of controls must not exceed 10 % at the end of the test,

• — the LD50 of the toxic standard meets the specified range.

1.5.   DESCRIPTION OF THE TEST METHOD

1.5.1.   Collection of bees

Young adult worker bees of the same race should be used, i.e. bees of the same age, feeding status, etc. Bees should be obtained from adequately fed, healthy, as far as possible disease-free and queen-right colonies with known history and physiological status. They could be collected in the morning of use or in the evening before test and kept under test conditions to the next day. Bees collected from frames without brood are suitable. Collection in early spring or late autumn should be avoided as the bees have a changed physiology during this time. If tests must be conduced in early spring or late autumn, bees can be emerged in an incubator and reared for one week with ‘bee bread’ (pollen collected from the comb) and sucrose solution. Bees treated with chemical substances, such as antibiotics, anti-varroa products, etc., should not be used for toxicity test for four weeks from the time of the end of the last treatment.

1.5.2.   Housing and feeding conditions

Easy to clean and well-ventilated cages are used. Any appropriate material can be used, e.g. stainless steel, wire mesh, plastic or disposable wooden cages, etc. Groups of 10 bees per cage are preferred. The size of test cages should be appropriate to the number of bees, i.e. providing adequate space.

The bees should be held in the dark in an experimental room at a temperature of 25 ± 2 ^oC. The relative humidity, normally around 50-70 %, should be recorded throughout the test. Handling procedures, including treatment and observations may be conducted under (day) light. Sucrose solution in water with a final concentration of 500 g/l (50 % w/v) is used as food. After given test doses, food should be provided ad libitum. The feeding system should allow recording food intake for each cage (see Section 1.6.3.1). A glass tube (approximately 50 mm long and 10 mm wide with the open end narrowed to about 2 mm diameter) can be used.

1.5.3.   Preparation of bees

The collected bees are randomly allocated to test cages, which are randomly placed in the experimental room.

The bees may be starved for up to 2 h before the initiation of the test. It is recommended that the bees are deprived of food prior to treatment so that all bees are equal in terms of their gut contents at the start of the test. Moribund bees should be rejected and replaced by healthy bees before starting the test.

1.5.4.   Preparation of doses

Where the test substance is a water miscible compound this may be dispersed directly in 50 % sucrose solution. For technical products and substances of low water solubility, vehicles such as organic solvent, emulsifiers or dispersants of low toxicity to bees may be used (e.g. acetone, dimethylformamide, dimethylsulfoxide). The concentration of the vehicle depends on the solubility of the test substance and it should be the same for all concentrations tested. However, a concentration of the vehicle of 1 % is generally appropriate and should not be exceeded.

Appropriate control solutions should be prepared, i.e. where a solvent or a dispersant is used to solubilise the test substance, two separate control groups should be used: a solution in water, and a sucrose solution with the solvent/carrier at the concentration used in dosing solutions.

1.6.   PROCEDURE

1.6.1.   Test and control groups

The number of doses and replicates tested should meet the statistical requirements for determination of LD50 with 95 % confidence limits. Normally, five doses in a geometric series, with a factor not exceeding 2,2 , and covering the range for LD50, are required for the test. However, the dilution factor and the number of concentrations for dosage have to be determined in relation to the slope of the toxicity curve (dose versus mortality) and with consideration taken to the statistical method which is chosen for analysis of the results. A range-finding test enables the choice of the appropriate concentrations for dosage.

A minimum of three replicate test groups, each of 10 bees, should be dosed with each test concentration. A minimum of three control batches, each of 10 bees, should be run in addition to the test series. Control batches should also be included for the solvents/carriers used (see Section 1.5.4).

1.6.2.   Toxic standard

A toxic standard should be included in the test series. At least three doses should be selected to cover the expected LD50 value. A minimum of three replicate cages, each containing 10 bees, should be used with each test dose. The preferred toxic standard is dimethoate, for which the reported oral LD50-24 h is in the range 0,10 -0,35 μg a.s./bee (2). However, other toxic standards would be acceptable where sufficient data can be provided to verify the expected dose response (e.g. parathion).

1.6.3.   Exposure

1.6.3.1.   Administration of doses

Each test group of bees must be provided with 100-200 μl of 50 % sucrose solution in water, containing the test substance at the appropriate concentration. A larger volume is required for products of low solubility, low toxicity or low concentration in the formulation, as higher proportions in the sucrose solution have to be used. The amount of treated diet consumed per group should be monitored. Once consumed (usually within 3-4 h), the feeder should be removed from the cage and replaced with one containing sucrose solution alone. The sucrose solutions are then provided ad libitum. For some compounds, at higher concentrations rejection of test dose may result in little or no food being consumed. After a maximum of 6 h, unconsumed treated diet should be replaced with the sucrose solution alone. The amount of treated diet consumed should be assessed (e.g. measurement of volume/weight of treated diet remaining).

1.6.3.2.   Duration

The duration of the test is preferably 48 h after the test solution has been replaced with sucrose solution alone. If mortality continues to rise by more than 10 % after the first 24 h, the test duration should be extended to a maximum of 96 h provided that control mortality does not exceed 10 %.

1.6.4.   Observations

Mortality is recorded at 4 h after starting the test and thereafter at 24 h and 48 h (i.e. after giving dose). If a prolonged observation period is required, further assessments should be made at 24 h intervals, up to a maximum of 96 h, provided that the control mortality does not exceed 10 %.

The amount of diet consumed per group should be estimated. Comparison of the rates of consumption of treated and untreated diet within the given 6 h can provide information about palatability of the treated diet.

All abnormal behavioural effects observed during the testing period should be recorded.

1.6.5.   Limit test

In some cases (e.g. when a test substance is expected to be of low toxicity) a limit test may be performed, using 100 μg a.s./bee in order to demonstrate that the LD50 is greater than this value. The same procedure should be used, including three replicate test groups for the test dose, the relevant controls, the assessment of the amount of treated diet consumed, and the use of the toxic standard. If mortalities occur, a full study should be conducted. If sublethal effects are observed (see Section 1.6.4), these should be recorded.

2. DATA AND REPORTING

2.1.   DATA

Data should be summarised in tabular form, showing for each treatment group, as well as control and toxic standard groups, the number of bees used, mortality at each observation time and number of bees with adverse behaviour. Analyse the mortality data by appropriate statistical methods (e.g. probit analysis, moving-average, binomial probability) (3)(4). Plot dose-response curves at each recommended observation time and calculate the slopes of the curves and the median lethal doses (LD50) with 95 % confidence limits. Corrections for control mortality could be made using Abbott's correction (4)(5). Where treated diet is not completely consumed, the dose of test substance consumed per group should be determined. LD50 should be expressed in μg of test substance per bee.

2.2.   TEST REPORT

The test report must include the following information:

2.2.1.   Test substance:

• — physical nature and relevant physical-chemical properties (e.g. stability in water, vapour pressure),

• — chemical identification data, including structural formula, purity (i.e. for pesticides, the identity and concentration of active substance(s)).

2.2.2.   Test species:

• — scientific name, race, approximate age (in weeks), collection method, date of collection,

• — information on colonies used for collection of test bees including health, any adult disease, any pre-treatment, etc.

2.2.3.   Test conditions:

• — temperature and relative humidity of experimental room,

• — housing conditions including type, size and material of cages,

• — methods of preparation of stock and test solutions (the solvent and its concentration must be given, when used),

• — test design, e.g. number and test concentrations used, number of controls; for each test concentration and control, number of replicate cages and number of bees per cage,

• — date of test.

2.2.4.   Results:

• — results of preliminary range-finding study if performed,

• — raw data: mortality at each dose tested at each observation time,

• — graph of the dose-response curves at the end of the test,

• — LD50 values with 95 % confidence limits, at each of the recommended observation times, for test substance and toxic standard;

• — statistical procedures used for determining the LD50,

• — mortality in controls,

• — other biological effects observed or measured e.g. abnormal behaviour of the bees (including rejection of the test dose), rate of consumption of diet in treated and untreated groups,

• — any deviation from the test procedures described here and any other relevant information.

3. REFERENCES

• (1)EPPO/Council of Europe (1993) Decision-Making Scheme for the Environmental Risk Assessment of Plant Protection Products — Honeybees. EPPO bulletin, vol. 23, N.1, 151-165. March 1993.

• (2)Gough, H. J., McIndoe, E.C., Lewis, G.B., (1994) The use of dimethoate as a reference compound in laboratory acute toxicity tests on honeybees (Apis mellifera L.) 1981-1992. Journal of Apicultural Research, 22, p. 119-125.

• (3)Litchfield, J.T. and Wilcoxon, F., (1949) A simplified method of evaluating dose-effect experiments. Jour. Pharmacol. and Exper. Ther., 96, p. 99-113.

• (4)Finney, D.J., (1971) Probit Analysis. 3rd ed., Cambridge, London and New-York.

• (5)Abbott, W.S., (1925) A method for computing the effectiveness of an insecticide. Jour. Econ. Entomol., 18, p. 265-267.

C.17.   HONEYBEES — ACUTE CONTACT TOXICITY TEST

1. METHOD

This acute toxicity test method is a replicate of the OECD TG 214 (1998).

1.1.   INTRODUCTION

This toxicity test is a laboratory method, designed to assess the acute contact toxicity of plant protection products and other chemicals to adult worker honeybees.

In the assessment and evaluation of toxic characteristics of substances, determination of acute contact toxicity in honeybees may be required, e.g. when exposure of bees to a given chemical is likely. The acute contact toxicity test is carried out to determine the inherent toxicity of pesticides and other chemicals to bees. The results of this test should be used to define the need for further evaluation. In particular, this method can be used in step-wise programmes for evaluating the hazards of pesticides to bees, based on sequential progression from laboratory toxicity tests to semi-field and field experiments (1). Pesticides can be tested as active substances (a.s.) or as formulated products.

A toxic standard should be used to verify the sensitivity of the bees and the precision of the test procedure.

1.2.   DEFINITIONS

Acute contact toxicity: is the adverse effects occurring within a maximum period of 96 h of a topical application of a single dose of a substance.

Dose: is the amount of test substance applied. Dose is expressed as mass (μg) of test substance per test animal (μg/bee).

LD 50 (Median Lethal Dose) contact: is a statistically derived single dose of a substance that can cause death in 50 % of animals when administered by the contact. The LD50 value is given in μg of test substance per bee. For pesticides, the test substance may be either an active substance (a.s.) or a formulated product containing one or more than one active substance.

Mortality: an animal is recorded as dead when it is completely immobile.

1.3.   PRINCIPLE OF THE TEST METHOD

Adult worker honeybees (Apis mellifera) are exposed to a range of doses of the test substance dissolved in appropriate carrier, by direct application to the thorax (droplets). The test duration is 48 h. If the mortality rate is increasing between 24 h and 48 h whilst control mortality remains at an accepted level, i.e. < 10 %, it is appropriate to extend the duration of the test to a maximum of 96 h. Mortality is recorded daily and compared with control values. The results are analysed in order to calculate the LD50 at 24 h and 48 h, and in case the study is prolonged at 72 h and 96 h.

1.4.   VALIDITY OF THE TEST

For a test to be valid, the following conditions apply:

• — the average mortality for the total numbers of controls must not exceed 10 % at the end of the test,

• — the LD50 of the toxic standard meets the specified range.

1.5.   DESCRIPTION OF THE TEST METHOD

1.5.1.   Collection of bees

Young adult worker bees should be used, i.e. bees of the same age, feeding status, race etc. Bees should be obtained from adequately fed, healthy, as far as possible disease-free and queen-right colonies with known history and physiological status. They could be collected in the morning of use or in the evening before test and kept under test conditions to the next day. Bees collected from frames without brood are suitable. Collection in early spring or late autumn should be avoided, as the bees have a changed physiology during the time. If tests have to be conduced in early spring or late autumn, bees can be emerged in an incubator and reared for one week with ‘bee bread’ (pollen collected from the comb) and sucrose solution. Bees treated with chemical substances, such as antibiotics, anti-varroa products, etc., should not be used for toxicity test for four weeks from the time of the end of the last treatment.

1.5.2.   Housing and feeding conditions

Easy to clean and well-ventilated cages are used. Any appropriate material can be used, e.g. stainless steel, wire mesh, plastic, disposable wooden cages, etc. The size of test cages should be appropriate to the number of bees, i.e. providing adequate space. Groups of 10 bees per cage are preferred.

The bees should be held in the dark in an experimental room at a temperature of 25 ± 2 ^οC. The relative humidity, normally around 50-70 %, should be recorded throughout the test. Handling procedures, including treatment and observations may be conducted under (day) light. Sucrose solution in water with a final concentration of 500 g/l (50 % w/v) should be used as food and provided ad libitum during the test time, using a bee feeder. This can be a glass tube (approximately 50 mm long and 10 mm wide with the open end narrowed to about 2 mm diameter).

1.5.3.   Preparation of bees

The collected bees may be anaesthetised with carbon dioxide or nitrogen for application of the test substance. The amount of anaesthetic used and time of exposure should be minimised. Moribund bees should be rejected and replaced by healthy bees before starting the test.

1.5.4.   Preparation of doses

The test substance is to be applied as solution in a carrier, i.e. an organic solvent or a water solution with a wetting agent. As organic solvent, acetone is preferred but other organic solvents of low toxicity to bees may be used (e.g. dimethylformamide, dimethylsulfoxide). For water dispersed formulated products and highly polar organic substances not soluble in organic carrier solvents, solutions may be easier to apply if prepared in a weak solution of a commercial wetting agent (e.g. Agral, Cittowett, Lubrol, Triton, Tween).

Appropriate control solutions should be prepared, i.e. where a solvent or a dispersant is used to solubilise the test substance, two separate control groups should be used, one treated with water, and one treated with the solvent/dispersant.

1.6.   PROCEDURE

1.6.1.   Test and control groups

The number of doses and replicates tested should meet the statistical requirements for determination LD50 with 95 % confidence limits. Normally five doses in a geometric series, with a factor not exceeding 2,2 , and covering the range for LD50, are required for the test. However, the number of doses has to be determined in relation to the slope of the toxicity curve (dose versus mortality) and with consideration taken to the statistical method which is chosen for analysis of the results. A range-finding test enables the choice of the appropriate doses.

A minimum of three replicate test groups, each of 10 bees, should be dosed with each test concentration.

A minimum of three control batches, each of 10 bees, should be run in addition to the test series. If an organic solvent or a wetting agent is used three additional control batches of each 10 bees for the solvent or the wetting agent have to be included.

1.6.2.   Toxic standard

A toxic standard must be included in the test series. At least three doses should be selected to cover the expected LD50 value. A minimum of three replicate cages, each containing 10 bees, should be used with each test dose. The preferred toxic standard is dimethoate, for which the reported contact LD50-24 h is in the range 0,10 -0,30 μg a.s./bee (2). However, other toxic standards would be acceptable where sufficient data can be provided to verify the expected dose response (e.g. parathion).

1.6.3.   Exposure

1.6.3.1.   Administration of doses

Anaesthetised bees are individually treated by topical application. The bees are randomly assigned to the different test doses and controls. A volume of 1 μl of solution containing the test substance at the suitable concentration should be applied with a microapplicator to the dorsal side of the thorax of each bee. Other volumes may be used, if justified. After application, the bees are allocated to test cages and supplied with sucrose solutions.

1.6.3.2.   Duration

The duration of the test is preferably 48 hours. If mortality increases by more than 10 % between 24 h and 48 h, the test duration should be extended up to a maximum of 96 h provided that control mortality does not exceed 10 %.

1.6.4.   Observations

Mortality is recorded at 4 h after dosing and thereafter at 24 h and 48 h. If a prolonged observation period is required, further assessments should be made, at 24 h intervals, to a maximum of 96 h, provided that the control mortality does not exceeding 10 %.

All abnormal behavioural effects observed during the testing period should be recorded.

1.6.5.   Limit test

In some cases (e.g. when a test substance is expected to be of low toxicity) limit test may be performed, using 100 μg a.s./bee in order to demonstrate that the LD50 is greater than this value. The same procedure should be used, including three replicate test groups for the test dose, the relevant controls, and the use of the toxic standard. If mortalities occur, a full study should be conducted. If sublethal effects are observed (see Section 1.6.4) these should be recorded.

2. DATA AND REPORTING

2.1.   DATA

Data should be summarised in tabular form, showing for each treatment group, as well as, control and toxic standard groups, the number of bees used, mortality at each observation time and number of bees with adverse behaviour. Analyse the mortality data by appropriate statistical methods (e.g. probit analysis, moving-average, binomial probability) (3)(4). Plot dose-response curves at each recommended observation time (i.e. 24 h, 48 h and, if relevant, 72 h, 96 h) and calculate the slopes of the curves and the median lethal doses (LD50) with 95 % confidence limits. Corrections for control mortality could be made using Abbott's correction (4)(5). LD50 should be expressed in μg of test substance per bee.

2.2.   TEST REPORT

The test report must include the following information:

2.2.1.   Test substance:

• — physical nature and physical-chemical properties (e.g. stability in water, vapour pressure),

• — chemical identification data, including structural formula, purity (i.e. for pesticides, the identity and concentration of active substance(s)).

2.2.2.   Test species:

• — scientific name, race, approximate age (in weeks), collection method, date of collection,

• — information on colonies used for collection of test bees including health, any adult disease, any pre-treatment, etc.

2.2.3.   Test conditions:

• — temperature and relative humidity of experimental room,

• — housing conditions including type, size and material of cages,

• — methods of administration of test substance, e.g. carrier solvent used, volume of test solution applied anaesthetics used,

• — test design, e.g. number and test doses used, number of controls; for each test dose and control, number of replicate cages and number of bees per cage,

• — date of test.

2.2.4.   Results:

• — results of preliminary range-finding study if performed,

• — raw data: mortality at each concentration tested at each observation time,

• — graph of the dose-response curves at the end of the test,

• — LD50 values, with 95 % confidence limits, at each of the recommended observation times, for test substance and toxic standard,

• — statistical procedures used for determining the LD50,

• — mortality in controls,

• — other biological effects observed or measured and any abnormal responses of the bees,

• — any deviation from the test method procedures described here and any other relevant information.

3. REFERENCES

• (1)EPPO/Council of Europe (1993) Decision-Making Scheme for the Environmental Risk Assessment of Plant Protection Products — Honeybees. EPPO bulletin, vol. 23, N.1, p. 151-165. March, 1993.

• (2)Gough, H. J., McIndoe, E.C., Lewis, G.B., (1994) The use of dimethoate as a reference compound in laboratory acute toxicity tests on honeybees (Apis mellifera L.), 1981-1992. Journal of Apicultural Research 22, p. 119-125.

• (3)Litchfield, J.T. and Wilcoxon, F., (1949) A simplified method of evaluating dose-effect experiments. Jour. Pharmacol. and Exper. Ther., 96, p. 99-113.

• (4)Finney, D.J., (1971) Probit Analysis. 3rd ed., Cambridge, London and New-York.

• (5)Abbott, W.S., (1925) A method for computing the effectiveness of an insecticide. Jour. Econ. Entomol. 18, p. 265-267.

C.18.   ADSORPTION/DESORPTION USING A BATCH EQUILIBRIUM METHOD

1. METHOD

This method is a replicate of the OECD TG 106, for the Determination of Soil Adsorption/Desorption, using a Batch Equilibrium Method (2000).

1.1.   INTRODUCTION

The method takes into account a ring test and a workshop for soil selection for the development of an adsorption test (1)(2)(3)(4) and also existing guidelines at national level (5)(6)(7)(8)(9)(10)(11).

Adsorption/desorption studies are useful for generating essential information on the mobility of chemicals and their distribution in the soil, water and air compartments of the biosphere (12)(13)(14)(15)(16)(17)(18)(19)(20)(21). The information can be used in the prediction or estimation, for example, of the availability of a chemical for degradation (22)(23), transformation and uptake by organisms (24); leaching through the soil profile (16)(18)(19)(21)(25)(26)(27)(28); volatility from soil (21)(29)(30); run-off from land surfaces into natural waters (18)(31)(32). Adsorption data can be used for comparative and modelling purposes (19)(33)(34)(35).

The distribution of a chemical between soil and aqueous phases is a complex process depending on a number of different factors: the chemical nature of the substance (12)(36)(37)(38)(39)(40), the characteristics of the soil (4)(12)(13)(14)(41)(42)(43)(44)(45)(46)(47)(48)(49), and climatic factors such as rainfall, temperature, sunlight and wind. Thus, the numerous phenomena and mechanisms involved in the process of adsorption of a chemical by soil cannot be completely defined by a simplified laboratory model such as the present method. However, even if this attempt cannot cover all the environmentally possible cases, it provides valuable information on the environmental relevance of the adsorption of a chemical.

See also General Introduction.

1.2.   SCOPE

The method is aimed at estimating the adsorption/desorption behaviour of a substance on soils. The goal is to obtain a sorption value which can be used to predict partitioning under a variety of environmental conditions; to this end, equilibrium adsorption coefficients for a chemical on various soils are determined as a function of soil characteristics (e.g. organic carbon content, clay content and soil texture and pH). Different soil types have to be used in order to cover as widely as possible the interactions of a given substance with naturally occurring soils.

In this method, adsorption represents the process of the binding of a chemical to surfaces of soils; it does not distinguish between different adsorption processes (physical and chemical adsorption) and such processes as surface catalysed degradation, bulk adsorption or chemical reaction. Adsorption that will occur on colloids particles (diameter < 0,2 μm) generated by the soils is not taken into account.

The soil parameters that are believed most important for adsorption are: organic carbon content (3)(4)(12)(13)(14)(41)(43)(44)(45)(46)(47)(48); clay content and soil texture (3)(4)(41)(42)(43)(44)(45)(46) (47)(48) and pH for ionisable compounds (3)(4)(42). Other soil parameters which may have an impact on the adsorption/desorption of a particular substance are the effective cation exchange capacity (ECEC), the content of amorphous iron and aluminium oxides, particularly for volcanic and tropical soils (4), as well as the specific surface (49).

The test is designed to evaluate the adsorption of a chemical on different soil types with a varying range of organic carbon content, clay content and soil texture, and pH. It comprises three tiers:

• Tier 1: preliminary study in order to determine:
• — the soil/solution ratio,
• — the equilibrium time for adsorption and the amount of test substance adsorbed at equilibrium,
• — the adsorption of the test substance on the surfaces of the test vessels and the stability of the test substance during the test period.

• Tier 2: screening test: the adsorption is studied in five different soil types by means of adsorption kinetics at a single concentration and determination of distribution coefficient Kd and Koc.

• Tier 3: determination of Freundlich adsorption isotherms to determine the influence of concentration on the extent of adsorption on soils. Study of desorption by means of desorption kinetics/Freundlich desorption isotherms (Appendix 1).

1.3.   DEFINITIONS AND UNITS