United States
Environmental Protection
Agency
Environmental
Research Laboratoiy
Athens, GA 30613-7799
Research and Development
EPA/600/S3-90/038 Aug. 1990
SEPA Project Summary
Freshwater Assay Using Soil
Eluates as Sample Material
(Single Laboratory Evaluation)
W.R. Lower, M.W. Thomas, B.M. Judy, and W.W. Sutton
The Chlorophyta assay, which uses soil
as sample material, has been a useful
bioassessment technique for screening
hazardous waste site problems. An
eluate is prepared from a 125-gram soil
sample and then diluted into three
separate concentrations prior to being
tested using Selenastrum capricornutum.
The work reported here has attempted to
determine the procedure's capability for
data quality, to provide a basis for decid-
ing whether the assay merits collabora-
tive testing, and to define more clearly the
method's potential for inclusion as part of
an operational monitoring network.
The soil used for most of this evaluation
was a homogenized clay loam (charac-
terized as being 22% sand, 51% silt, and
27% clay). Samples were chemically
analyzed using ICAP, GC/MS, and Kjel-
dahl nitrogen techniques to confirm that
a reasonably homogenized soil had been
selected for the evaluation. Soil contain-
ing either sodium fluoride or 2,4-
dichlorophenol was tested using the algal
assay, and several tests were conducted
to confirm the dose/response curve using
the positive control compound, zinc
chloride. A known test response also
was established using a white silica sand
spiked with sodium fluoride.
Although considerable progress was
made toward a standardization of this
procedure, some difficulties remain for
collaborative testing. The known test
response provides a value with a fairly
high standard error, which, in some ways,
probably detracts from the future applica-
tion of this technique when accuracy
and/or systematic error estimates are
needed. The current evaluation also
revealed an apparently poor capability for
sensitivity and a somewhat limited range
of reliable measurement. However, the
procedure will definitely detect the
presence of certain chemicals and, thus,
can be effectively used as a screening
technique. Consequently, the observed
method sensitivity should not necessarily
be considered detrimental to further
method standardization or to future ap-
plication as part of a monitoring network.
This Project Summary was developed
by EPA's Environmental Research
Laboratoiy, Athens, GA, to announce key
findings of the research project that is
fully documented in a separate report of
the same title (see Project Report order-
ing information at back).
Introduction
Assessing the significance of current and
potential problems at uncontrolled hazard-
ous waste sites has been very difficult be-
cause (1) complex chemical mixtures are
found at most sites and (2) toxicity and en-
vironmental fate data are limited for many of
the compounds present at the sites, espe-
cially for those that are by-products of either
organic synthesis or of organic degradation.
Biological tests can be important com-
ponents of waste site monitoring programs.
Benefits include their ability to predict the
potential hazard posed by unknown mix-
tures of chemicals, to identify the bioactive
fractions of unknown chemical mixtures,
and similarly, to detect toxic substances not
normally reported by standard chemical
analysis.
If biological methods are to be used as
part of the waste site monitoring effort, the
selected procedure(s) must be capable of
-------
providing reliable and repeatable results.
Consequently, it Is necessary to determine
the procedure's capability for data quality in
a single laboratory and, by way of a col-
laborative test, in multiple laboratories.
This research concerns an evaluation of a
modified freshwater assay. The assay is
based on a procedure developed by EPA's
Environmental Research Laboratory, Cor-
vallls, OR, as a bioassessment method for
hazardous waste sites. The assay method
has been tested at actual waste sites and
found to be one of the better screening tech-
niques for detecting the presence of poten-
tially hazardous chemicals. An aqueous
eluate is acquired initially from the original
soil sample, and this eluate is subsequently
filtered and diluted prior to toxicity testing.
This algal assay procedure addresses
neither non-water soluble chemicals nor
volatile compounds. In addition, proce-
dures associated with the collection of soil
samples (e.g., representative sampling of a
waste site area) are not addressed. Based
on tha ecological importance of unicellular
algae, however, this relatively simple techni-
que can provide an environmental hazard
assessment for those chemicals that are
most likely to be transported to surface and
ground waters.
This work seeks to establish the data
quality that can be achieved within a single
laboratory. The objective of this evaluation
also provides a basis for deciding whether
this modified algal procedure merits col-
laborative testing and to define more clearly
tha assay's potential for inclusion as part of
an operational monitoring network.
Procedure
Single Laboratory Evaluation
Phases of the single laboratory test in-
clude identification of procedural variables
that must be carefully controlled (rugged-
ness testing), evaluation of method sen-
sitivity, Identification of the limits of reliable
measurement, evaluation of systematic
error (bias), and identification of method
precision and accuracy.
Ten successive analyses (i.e., aseriesthat
yields ten valid responses by following the
method protocol) are typically conducted for
several phases of the single laboratory test.
Tha determination of method precision, for
example, requires that ten successive inde-
pendent analyses be conducted on the
same sample material. Multistage calcula-
tions to determine the required number of
analyses might be conducted during the
single laboratory test as more information
becomes available on the expected
variance. However, 10 analyses will allow
the test laboratory to estimate the standard
deviation to within 45 percent of its true value
(at a 95 percent confidence interval). The
single laboratory test cost will increase
rapidly as more of these successive
analyses are required because each addi-
tional value must represent a valid test
response and, therefore, will include
whatever quality control analyses (blanks,
replicates, etc.) are required in the original
method protocol to ensure a valid.test
response.
Consensus Review
The method protocol was reviewed by
persons familiar with both this specific assay
and with many of the problems encountered
at uncontrolled hazardous waste sites. A
consensus review is always necessary as an
attempt to achieve consensus agreement
prior to beginning the method stand-
ardization process. The protocol was slight-
ly revised based on the review comments
and reviewed a second time by the same
scientists. However, as is often the case,
complete consensus agreement for each
step of the assay was not totally achieved.
Outline of Assay Procedure
Soil Preparation
A blender is initially used to homogenize
the soil which is then smoothed to an even
layer in a ceramic pan. The soil is then air
dried before a '125-gram portion is used for
the assay. Rocks or gravel are manually
removed prior to drying. The 125-gram soil
sample is transferred to a 1-liter teflon-lined
container and 500 ml of deionized water are
added. The container then is capped and
mechanically shaken. The liquid is
decanted from the container and transferred
to 50-ml centrifuge tubes and then
centrifuged. The total amount of soil eluate
recovered from the container will vary some-
what between samples but at least 400 ml
should be available for analysis.
Eluate Dilutions
This leachate (soil eluate) is filtered, and
the pH is adjusted prior to testing. Several
preliminary filtering steps (i.e., 2.7 ^m; 1.6
fim; 1.0 fim; etc.) may be required depend-
ing on the characteristics of any precipitate
noted in the resulting soil eluate. After
centrifugation, the eluate typically is trans-
ferred to a Buchner funnel that contains a
0.7-^m, nonsterile glass fiber filter. Sub-
sequently, the eluate is filtered through a
0.45-^m nylon filter to obtain a sterile eluate.
Although the growth response of
Selenastrum capricornutum is not affected
over a wide range of pH values, the potential
array of chemicals and soil types en-
countered when using the procedure oc-
casionally will require pH alterations for the
leachate. Therefore, several pH determina-
tions are routinely conducted and, when re-
quired, pH adjustments are made using 1N
sodium hydroxide and 1N hydrochloric acid.
In the next step, the soil eluate is diluted to
a series of three separate concentrations
using the growth medium preparation. Algal
cellsare transferred from the stock culture to
each dilution flask. A negative control also
is prepared using the growth media prepara-
tion and the algal cell addition. Zinc chloride
is added in addition to the growth medium
and algal cells when preparing the positive
control. Ninety-six hours after the flasks are
inoculated, the cell concentration in each
flask is determined using a Coulter counter.
Duplicate flasks are prepared routinely for
the negative control, positive control, 80%
soil eluate, 10% soil eluate, and 1% soil
eluate, respectively. In the case of the posi-
tive control, separate response values are
tabulated. The first step of the final calcula-
tion is to determine the eluate effect for each
of the three dilutions and the second step is
to extrapolate an ECso value which then ser-
ves as the final test result. The eluate effect
is tabulated using the initial inoculum cell
count and the 96-hour cell count determined
for each eluate dilution and for the negative
control. The proportion of growth medium
used in a given dilution flask also is included
in the tabulation. A positive control
response also is routinely calculated for
each separate assay.
Cell Culture
A pure monoculture of algal cells is re-
quired for the assay. The algal cells are first
transferred from the stock culture to an in-
oculum flask, which should contain 50 ml of
previously sterilized deionized water. When
the assay flasks are prepared, the cell trans-
fer should contain a sufficient number of
cells so that the resulting concentration (in
the 50 ml of soil eluate plus growth medium)
will be 1 x 104 ± 1 x 103 cells/ml.
Nutrient Composition
Total consensus was never reached con-
cerning the composition of the nutrient
medium. The nutrient media used obviously
will affect the algal response, but persons
experienced with the assay simply do not
agree on the concentration of some medium
components. The growth medium used
throughout this evaluation, and the medium
composition specified in the "consensus"
method protocol, is given in Table 1. The
medium is prepared by adding 1.0 ml from
each of the Table 1 solutions (macronutrient
and micronutrient) to 900 ml of sterilized
deionized water. The final volume is taken
to 1 liter.
-------
Table 1. Composition of Algal Growth Media
Macronutrients
Stock Solution3
Nutrient Composition
Prepared Medium
Compound
NaNOs
NaHCOa
K2HP04
Concentration
(9/1) .
25.500
15.000
1.044
Element
Stock Solution 1
N
Nab
C
Stock Solution II
K
P
Concentration
(rng/l)
4.200
11.001
2.143
0.469
0.186
Stock Solution III
MgSC-4 7HzO
14.700
1.911
Stock Solution IV
MgCLz CHaO
CaC122HzQ
12.164
4.410
Mg°
Ca
2.904
1.202
Micronutrients
Stock Solution
Nutrient Composition
Prepared Medium
Compound
H3B03
MnC1z 4HzO
ZnC12d
CoC126H2Od
CuC12 2HzO d
Na2MoO.t2HsOd
Fed 3 SHaO
NazEDTA 2HsO
Concentration
(mg/l)
185.520
415.610
3.271
1.428
0.012
7.260
160.000
300.000
Element
Stock Solution V
B
Mn
Zn
Co
Cu
Mo
Fe
—
Concentration
tern
32.460
115.374
1.570
0.354
0.004
2.878
33.051
—
a Other forms of the salt may be used if the resulting element concentrations are the same as those listed.
b Includes A/a from NaNOg.
0 Includes Mg from MgSC-4 7HaO.
d Prepare at 100 times concentration and add in designated amount to stock solution V.
Positive Control
Zinc chloride was used as the positive
control compound but many other chemi-
cals undoubtedly could be selected. The
purpose of the positive control is to provide
information on whether the algal assay is
responding to a known toxic material and,
after several assays have been conducted,
the various positive control responses can
provide data that are used to assess the
variability occurring between assays. The
positive control response is not used when
tabulating the ECgo value for the final test
result, but the positive control values ob-
viously are used to help determine whether
a test response is accepted as a valid test
result.
The positive control contains the growth
medium solution, the algal cell inoculum,
and the zinc chloride. It should be em-
phasized that the positive control com-
pound is added directly to the assay flask
and not to a soil sample. Two positive con-
trol flasks are prepared for each assay. The
average response of the two positive con-
trols must show at least a 15 percent inhibi-
tion of cell concentration as compared to the
negative control. In addition, the cell counts
from the two positive control flasks must be
within 50 percent of each other in order to
achieve a valid test result. If these condi-
tions do not exist for the positive control
result, the assay results will not constitute a
valid test response. The suggested zinc
chloride concentration of the positive control
is 128^g/l. The suitability of this concentra-
tion, however, should be verified by the test-
ing laboratory prior to beginning an actual
series of assays.
Sample Material
Water eluates were prepared from a series
of soil types and tested using Selenastrum
capricornutum prior to the selection of a clay
loam soil (characterized as being 22% sand,
51% silt, and 27% clay) for use during the
evaluation. After the soil was selected, a
-------
suitable quantity (approximately 275 kg) was
collected so that enough soil would be avail-
able for the entire single laboratory effort.
The soil was homogenized thoroughly, first
by grinding and then by mixing in a mechani-
cal soil mixer for several hours. The soil and
the subsequently prepared soil eluate were
chemically analyzed primarily to determine
whether a sufficiently homogenized
preparation had been achieved for use in the
method evaluation. Analyses also were
conducted to confirm the transport of
fluoride from the spiked soil (sodium
fluoride) to the eluate and to compare this
transport with the eventual ECso test result.
Results from this analytical effort are shown
in Tables 2 and 3.
Results and Discussion
Ruggedness Test
The first step in the evaluation was to iden-
tify those procedural steps that must be
carefully controlled. If the Chlorophyta pro-
cedure is "rugged" it will not be susceptible
to the inevitable, modest departures in
routine that occur, and the final test result will
Tab/o 2. Element Analysis of Soil Subsamples Taken from the Homogenized Soil Collection Used for the Chlorophyta
Evaluation
Element Composition (ftglg)
Subsample
Number
1
2
3
4
5
6
7
8
9
10
X
±S.D.
C.V."
Mg
27.3
29.5
27.5
27.3
24.7
26.7
27.6
26.3
27.7
29.5
27.4
1.4
5.2
K
21.5
22.3
20.8
20.1
19.0
20.3
20.2
20.0
21.8
22.8
20.9
1.2
5.7
Zn
7.5
7.5
7.5
6.8
6.4
6.8
7.1
7.7
6.6
7.4
7.1
0.5
6.3
Fe
76.2
73.6
71.0
68.9
65.4
66.0
68.7
73.6
64.9
74.6
70.5
4.4
6.2
Mn
71.8
71.6
67.4
64.1
59.8
59.4
67.0
71.6
63.5
60.2
65.6
5.0
7.6
Cu
0.76
0.48
1.24
1.53
1.52
0.49
1.05
0.48
0.55
0.63
0.9
0.4
49.4b
Ca
195.5
202.5
180.5
184.5
170.0
176.0
199.5
182.5
189.5
185.2
185.3
11.3
6.1
SC-4-S
7.5
7.0
7.0
6.8
7.0
7.0
9.3
7.5
7.3
5.5
7.2
0.9
12.9
pH
5.5
5.5
5.4
5.4
5.4
5.4
5.6
5.5
5.5
5.5
5.5
0.1
1.2
In addition to the results shown above, a few nitrogen analyses were conducted (Kjeldahl nitrogen) using the soil eluates. These
analyses revealed an essentially consistent concentration (ftg/ml) occurring between eluates from different soil subsamples. Some
attempts at organic compound identification also were conducted, i.e., eluate samples were analyzed using GC/MS and LC/MS
procedures. Few specific organics were detected and those that were identified were detected at low concentrations (ng/ml for
eluates and ng/g for the actual soil). Tentatively identified organic compounds included benzole acid, hydroxy methoxy benzal-
dehyde, and methyl hexanone. The various chemical analyses have helped to confirm that a reasonably homogenized soil was
selected for the current Chlorophyta evaluation.
With the exception of potential analytical error at these relatively low concentrations, no explanation is presented for this unusually
high coefficient of variation.
Table 3. Approximate Transfer of Sodium Fluoride from Soil to Eluate Using the Chlorophyta Assay Procedure
Amounts of sodium fluoride added to 125-g soil sample (mg)
Sodium fluoride concentration in soil (mg/g)
Sodium fluoride concentration in eluate (ft/ml)
Total recovery In eluate (mg)
Approximate recovery of spike (%)
X ECso for assay
Standard deviation ±
Coefficient of variation (%)
556
4.5
378
151.2
27
-0.7
—
—
900
7.2
442
176.8
20
56.6
6.5
11.4
1090
8.7
770
308.0
28
50.9
7.5
14.8
1112
'9.0
693
277.2
25
44.6
2.6
5.8
2224
17.8
1714
685.6
31
32.7
8.2
25.1
-------
not be altered by these slight variations. If the
results are altered by small procedural varia-
tions, it is important to emphasize in the
protocol that a specific step must be strictly
followed or, in some cases, to indicate the
limits of allowable variability. The ex-
perimental design used during the rugged-
ness test (based on the design by W. J.
Youden In The Collaborative Test) did not
seek to study each separate variable in an
individual sequential fashion, but rather to
provide for the simultaneous introduction of
multiple protocol variations (Tables 4 and 5).
As noted in Table 4, eight complete as-
says were conducted as part of the rugged-
ness evaluation. Each test sample
contained 100 mg of the organic compound
sodium dodecyl sulfatethat had been added
to the 125-gram aliquot soil samples imme-
diately prior to the test. The "protoco1
directed" conditions were designated as A
through G, and the varied conditions were
designated as a through g. The evaluation
was concerned with identifying respective
variations in the final test result due to the
specific procedural differences, i.e., A-a, B-
b, C-c, D-d, E-e, F-f, and G-g. Each of the
eight trials consisted of a single analysis
conducted using eight aliquots of the test
material (100 mg of sodium dodecyl sulfate).
The final test results were indicated as s, t, u,
v, w, x, y, and z.
Table 4. Experimental Design for a Seven Variable Ruggedness Test *
Test
Number
Combination of
Variables
Test
Result
1
2
3
4
5
6
7
8
ABCDEFG
ABcDefg
AbCdEfg
AbcdeFG
aBCdeFg
aBcdEfG
abCDefG
abcDEFg
s
t
u
v
w
X
y
z
1 Based on W.J. Youden, 1969, The Collaborative Test, p. 151-158. In Precision Measure-
ment and Calibration. H.H. Ku, Editor. U.S. Department of Commerce, National Bureau
of Standards. 436 pp. For the Chlorophyta evaluation, each test sample contained 100
mg of the organic compound sodium dodecyl sulfate that had been added to the 125-gram
aliquot soil samples immediately prior to the test.
Table 5. Test Variables Used to Evaluate Ruggedness of Chlorophyta Procedure
Directed Instruction Altered Instruction
A. 125-g soil aliquot used as sample material for assay.
B. at start of assay, each flask contains 1 x104 algal
cells per ml.
C. original soil collection is first air dried for 24 hours
(room temperature).
D. 500 ml of water added to the aliquot sample of air
dried soil.
£ soil eluate is pH adjusted to between 6.0 - 8.5 prior
to the assay.
F. eluate dilution series consists of 1%, 10%, and 80%
soil eluate.
a. 110-g soil aliquot used as sample material for the assay.
b. at start of assay, each flask contains 0.75x104 algal cells per ml.
c. original soil collection is first air dried for 20 hours (room
temperature).
d. 550 ml of water added to the aliquot sample of air dried soil.
e. soil eluate is pH adjusted topHW prior to the assay.
f. eluate dilution series consists of 1%, 30%, and 80% soil eluate.
G. algal assay terminated at 96 hours.
g. algal assay terminated at 92 hours.
-------
The average ofA = (s + t + u + v)/4,
compared with the average of a = (w + x +
y + z)/4, serves as a rapid means of assess-
ing the effect of changing variable A to a.
Because each of the two groups of four
determinations contains the other six vari-
ables (twice at the upper case level and twice
at the lower case level), the effect of these
variables (if present) tends to cancel out,
leaving only the effect of changing variable
A to a The relative effect of the other vari-
ables was estimated by examining the fol-
lowing averages:
(s +1 + w + x) (u + v + y + z)
B = b =
(s + v + x + y) (t + u + w + z)
G= g=
(s + u + w +• y) (t + v + x + z)
C- c-
(s +1 + y + z) (u + v + w + x)
D- d=
(s + u + x + z) (t + v + w + y)
: 9 =
(s + v + w + z) (t + u + x + y)
., f =
After tabulating the above averages, the
differences between each respective vari-
able was computed, e.g.,
A-a = -
(s +1 + u + v) (w + x + y + z)
Most of the modest procedural alterations
should typically have little or no effect on the
test result. However, a comparison of the
respective differences (A-a, B-b, etc.)
provides considerable information on which
variables, if any, are having the greater ef-
fects.
The analysis showed that moderate pro-
cedural variations definitely altered the final
Chlorophyta test result when they occurred
at certain critical steps. Steps shown to be
most sensitive to variation involved variables
D, E, and F, i.e., the amount of water added
to the dry soil sample, the pH of the resulting
soil eluate that is actually used during the
assay, and the specific eluate concentra-
tions prepared for the assay dilution series.
The Chlorophyta protocol consequently
was revised to emphasize strict adherence
to these critical instructions.
Precision
The precision evaluation was conducted
using the ruggedness tested, and rugged-
ness revised, Chlorophyta method protocol.
Each sample contained 8.9 mg of sodium
fluoride per gram of homogenized clay loam
soil. This sodium fluoride concentration was
selected after several concentrations were
tested and after it was shown that the assay
response was close to the calculated ECso
concentration. The final result seemed to
represent the best attainable precision when
using the Chlorophyta procedure.
Ten separate tests using aliquots of the
same sample were conducted. Each
separate test shown in Table 6 represents a
valid test as required by the Chlorophyta
method protocol. Based on the 10 separate
determinations using 8.9 mg sodium
fluoride per gram of soil, the assay's single
laboratory capability for precision (ex-
pressed as a coefficient of variation) is
presented as 5.8 percent. An attempt was
made to conduct the individual determina-
tions on alternate days, i.e., an interval of at
least one day between the completion of one
assay and the start of the next. While an
exactly identical interval was not maintained
between individual assays due to the eluate
preparation time and problems of routine
laboratory scheduling, the individual
analyses shown in Table 6 were conducted
sequentially over a total time interval that
covered several days.
Table 6. Chlorophyta Assay's Single Laboratory Capability for Precision
(Sample Material Consisted of 8.9 mg Sodium Fluoride per Gram of Soil. Valid Test Result Was Achieved for Each Assay.)
Positive Control
Assay
Number
1
2
3
4
5
6
7
8
9
10
X
±S.D.
C.V.(X)
Calculated
ECso
41.0
47.4
41.1
42.1
46.6
44.8
46.0
43.5
45.5
48.2
44.6
2.6
5.8
#1
-53.0
-51.0
-51.0
-58.8
-62.9
-57.7
-72.0
-89.5
-93.6
-72.0
#2
-59.3
-69.0
-59.5
-50.2
-72.9
-65.7
-80.4
-91.1
-94.8
-76.1
%
Difference
8.9
26.1
14.3
14.7
13.9
12.2
10.0
2.0
1.3
5.4
Negative
Control
2.9x106/ml
2.8x10e/ml
2.7x10e/ml
2.6x106/ml
2.6x106/ml
2.6x106/ml
2.7x106/ml
2.6x10e/ml
2.5x106/ml
2.5x106/ml
Note: Each Individual soil sample weighed 125 grams and 1112mgof sodium fluoride was added to each sample.
As directed by the method protocol, the assay result was not considered to be a valid test result unless (1) the
negative control had achieved at least 1.0 x106 cells/ml at the 96-hour time interval, (2) an average of the two
positive control results demonstrated at least a 15 percent inhibition In cell growth, and (3) tf?e two positive
control results did not differ from each other by more than 50 percent.
-------
Response to Organic Compound
Soil containing 2,4-dichlorophenol also
was tested using the Chlorophyta proce-
dure. Chlorinated hydrocarbons frequently
are present at hazardous, waste sites and
they definitely represent one of the waste site
chemical groups that cause environmental
concern. This particular chemical was
tested as a representative organic pollutant
that is somewhat resistant to microbial
degradation.
The dichlorophenol was added to the clay
loam test soil to achieve a concentration of
QOfig/g. Separate determinations were con-
ducted using ten identical soil samples that
contained this concentration of
dichlorophenol. Each of the ten determina-
tions represented a valid test response as
directed by the Chlorophyta method
protocol. It has already been established
that the Chlorophyta procedure was respon-
sive to organic pollutants as well as to actual
waste site samples that contain low con-
centrations of several chemicals. This
method responsiveness was certainly
present when the soil samples contained
dichlorophenol. As is noted in Table 7, how-
ever, the observed precision is not very im-
pressive when compared to the test results
achieved using sodium fluoride. Since
preliminary testing preceded selection of the
SQfiQ/g dichlorophenol concentration, these
data may suggest a responsive procedure
but a somewhat less precise assay than
expected when the sample contains an or-
ganic pollutant (i.e., 46.5 percent coefficient
of variation when testing dichlorophenol).
Method Sensitivity
For purposes of a single laboratory
evaluation, a method's sensitivity is defined
as the method's capability to detect (or dis-
tinguish between) small changes in sample
concentration, i.e., concentrations of
analyte. The specific concentrations for the
sensitivity phase usually are selected se-
quentially based on results from previously
selected concentrations. The same sample
material (sodium fluoride in clay loam soil)
used during the precision test phase was
used for the method sensitivity evaluation.
Table 8 presents a series of assay results
that were ultimately used to assess the
method's capability not only for sensitivity
but also for the limits of reliable measure-
ment.
The single laboratory evaluation typically
uses ten independent assays for each new
concentration, i.e., ten separate 'valid
responses acquired by following the written
protocol. Under routine operating condi-
tions, however, the assay would only be
conducted one time per sample. Conse-
quently, when tabulating data for sensitivity,
non-overlapping standard deviations (rather
than non-overlapping standard errors) are
used to indicate whether the method can
distinguish between the different samples.
As noted in Table 8, the Chlorophyta pro-
cedure does not have a particularly impres-
sive capability for sensitivity. Concentra-
tions of sodium fluoride between 17.8 mg/g
and 48.0 mg/g of soil produced the same
test response. In addition, concentrations of
sodium fluoride between 8.9 mg/g and 7.2
mg/g of soil produced the same test
response. Unfortunately, this total con-
centration range also covers most of the
response range. The Chlorophyta proce-
dure, therefore, clearly will detect the
presence of the pollutant compound but, for
the most part, is only sensitive to order of
magnitude concentration differences.
Limits of Reliable Measurement
The same sample material should be used
for reliability measurement as was originally
used during the method precision and
method sensitivity evaluations. The evaluat-
ing laboratory typically should select two
concentrations of sample material. One of
these concentrations should be near the
upper extreme of the method's detection
range and the other should be near the lower
extreme of the detection range. Ten
analyses would be conducted on each con-
centration to provide precision data (ex-
pressed as a coefficient of variation).
However, the previously acquired data con-
firming the poor capability for sensitivity and
the somewhat limited total assay response
range made this stage of the evaluation less
difficult than is usually the case.
As was the case for the above-mentioned
method sensitivity, the limits of reliable
Table 7. Assay Response Using Soil Samples Containing Either Sodium Fluoride or 2,4-Dichlorophenol
(Precision Capability, Expressed as a Coefficient of Variation, in Response to Soil Containing
Either an Inorganic or an Organic Pollutant Compound)
Calculated ECso
Assay
Number
1
2
3
4
5
6
7
8
9
10
XECso
±SD
CV(%)
1112mgNaFper
125 g soil (8.9 mgjg)
41.0
47.4
41.1
42.1
46.6
44.8
46.0
43.5
45.5
48.2
44.6
2.6
5.8
10 mg dichlorophenol
per 125 g soil (80ftg/g)
55.9
57.3
49.8
25.7
93.4
40.2
94.7
46.1
44.5
116.2
62.4
29.0
46.5
-------
Table 8. Response of Chlorophyta Assay Procedure to Various Soil Concentrations of Sodium Fluoride
mg sodium
fluoride
per 125 g
of soil
556
700
800
900
1090
1112
2224
3336
4448
6000
7500
mg sodium
fluoride per
gram of clay
loam soil
4.5
5.6
6.4
7.2
8.7
8.9
17.8
26.7
35.6
48.0
60.0
Mean
ECso
-0.7
-181.7
54.3
56.6
50.9
44.6
32.7
31.9
22.6
16.6
-0.9
±SD
_
455.3
6.9
6.5
7.5
2.6
8.2
9.1
8.1
10.9
29.5
CV(%)
~
250.6
12.7
11.4
14.8
5.8
25.1
28.5
35.8
65.5
-
Range of
ECso±SD
—
—
47.4 - 61.2
50.1 - 63.1
43.4 - 58.4
42.0-47.2
24.5-40.9
22.8-41.0
14.5-30.7
5.7-27.5
; •
Note: Each mean ECso value is the result of ten separate assays using the indicated concentration of sodium fluoride
per gram of soil. Each separate assay was conducted as directed by the method protocol and included the
required control samples. Concentrations of 4.5, 5.6, and 60.0 mg/g apparently are outside the limits of
reliable measurement for the Chlorophyta assay.
measurement are not particularly impressive
oiherthan that the assay does show a defini-
tive response to the chemical over a con-
centration range of 6.4 mg/g to 48.0 mg/g
(i.e., before complete inhibition/lethality).
For this phase, the basic objective was to
verily that the method capabilities for sen-
sitivity, precision, and accuracy (when ap-
plicable) do not deteriorate at the upper and
lower extremes of the detection range.
Although the procedure's sensitivity is not
very good at the extremes of the detection
range, it is not any less impressive than the
method's best sensitivity. The method's
deterioration in precision provides the most
information concerning limits of reliable
measurement. As mentioned above, ten
analyses were conducted at each of several
concentrations (Table 8) so that, among
other reasons, precision comparisons could
be made. The coefficient of variation fre-
quently wilt show a dramatic increase at the
extreme limits of detection and, therefore,
precision data provide a distinct indication
of the limits of reliable measurement. The
results of this effort suggest a very restricted
concentration range of 6.4 mg/g to 35.6
mg/g of soil. The coefficient of variation at
the greater concentrations increases rapid-
ly, especially when compared with the 5.8
percent coefficient of variation achieved at
the 8.9 mg/g concentration. The 65.5 per-
cent coefficient of variation noted at 48.0
mg/g might well be considered as being
beyond a range of reliable measurement.
However, the assay clearly shows a
response to sodium fluoride atthe increased
concentrations and, considering that the
overall procedure is only a screening tech-
nique, the increased coefficient of variation
should not really affect method usability.
Accuracy
To determine a method's single laboratory
capability for accuracy (and for systematic
error), the testing laboratory must have both
a standard reference material and a known
method response (true response) to this
reference material. When a method calls for
analysis of biological tissue or biological
fluid, there will usually be a standard refer-
ence material available to the testing
laboratory, e.g., samples with certified com-
pound concentrations or perhaps with cer-
tified enzymatic activity levels. In these
instances, the method's single laboratory
capability for accuracy can be assessed by
determining the differences between the ob-
served single laboratory result, using the
reference sample, and the known true value.
In the case of a method such as this algal
assay that uses the soil eluates, no "true
response" is available for a reference
material; hence the method's single
laboratory capability for accuracy (orfor sys-
tematic error) cannot be determined. Under
these conditions, the testing laboratory
should first select a reference material and
then determine an average test response for
a single concentration of the reference
sample. Comparison of test data with the
results of a reference method can be used
in certain situations. For the purposes of this
evaluation, however, the use of a reference
method is not really possible and is not
typically done for a single laboratory evalua-
tion.
Although not technically a standard, this
evaluation used a white silica sand that is
recommended by the American Society for
Testing and Materials (ASTM) for use in cer-
tain cement preparations and, more sig-
nificantly, can be acquired easily by other
laboratories that are using the assay proce-
dures. The sand has a particle size of 850 to
600 /4m and receives a series of agitation
washes prior to use. It is subsequently
spiked with sodium fluoride to achieve a 4.8
mg/g concentration.
A greater recovery of toxic chemical will
typically occur in the eluate when sand is
used and, consequently, the known test
response is being established at a different
sodium fluoride concentration than was
used during most of the current evaluation.
The known test response to 4.8 mg of
sodium fluoride per gram of washed ASTM
silica sand was 43.4 ± 9.7 when expressed
as the calculated ECso ± the standard error.
Results from the individual determinations
that have provided this known value are
shown in Table 9. However, these data
again emphasize an apparent poor
capability for precision which, in some ways,
probably detracts from the future application
of this response for use in accuracy and/or
systematic error estimates.
Conclusion
Although considerable progress has been
made with the standardization of this proce-
dure, some difficulties remain if one con-
siders a continuation to collaborative
testing. The poor capability for sensitivity
and the somewhat limited range of reliable
measurement might be considered as being
discouraging. However, the greatest prob-
lem confronting a continuation to collabora-
tive testing may be the recurring lack of
-------
Table 9. Determinations to Establish a Known Test Response to a
Reference Sample
(Each Final Result Expressed as Calculated ECso)
Assay
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
XECso
±SE
±SD
CV%
600 mg NaF per 125gASTM
silica sand (4.8 mg/g)
48.2
48.5
52.5
62.0
43.5
29.8
15.1
25.4
62.2
21.9
54.2
23.2
39.8
45.5
62.1
50.3
45.3
43.5
42.6
51.6
43.4
9.7
13.8
31.8
consensus noted among knowledgeable
biologists for several phases of the method
protocol. However, many frequently used
procedures have never enjoyed total con-
sensus agreement and various organiza-
tions such as the Association of Official
Analytical Chemists (AOAC) and the
American Society for Testing and Materials
(ASTM) frequently publish method revisions
as well as interim protocols to allow for on-
going improvements. The authors of this
report recommend that the current algal
assay be collaboratively tested as a next
step toward more frequent application of the
procedure for hazardous waste monitoring.
William R. Lower, Mark W. Thomas, and Barbara M. Judy are with the University of
Missouri, Columbia, MO 65203. William W. Sutton is with the Environmental
Research Laboratory, Athens, GA and is the EPA Project Officer (see below).
The complete report, entitled "Freshwater Assay Using Soil Eluates as Sample Material
(Single Laboratory Evaluation)", (Order No. PB 90-203 4561 AS; Cost: $17.00,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: (703) 487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
College Station Road
Athens, GA 30613-7799
-------
-------
-------
m
TJ
•jo
§ 3
m Q.
CO
CD
•e»
CO
s
1:
0)
CD
a
i
o-
I H
-D|
af
col
------- |