COMPARATIVE STUDIES OF INSTRUMENTAL AND

BIOASSAY METHODS FOR THE ANALYSIS OF

HERBICIDE RESIDUES
D. O. EBERLE and H. R. GERBER
CIBA-GEIGY Ltd. Agrochemicals Division,
Basle, Switzerland

A study of the quantitative analysis of herbicide residues by both chemical and

bioassay methods in soils is presented.

Field and laboratory residue trials were carried out with a representative member of
the following groups of herbicides: ureas, triazines, diphenylethers, phenoxyacetic
acids, and dithiophosphates. Representative samples were taken at different time
intervals, and degradation curves were established both by chemical methods and by
two types of bioassay. Chemical analysis either separated active ingredient and
metabolites by chromatographic techniques or comprised total residues. Bioassays
were performed using either monocotyledons and dicotyledons or algae.

The results obtained by chemical and bioassay analysis for the degradation rates of
chlorotoluron, ametryn, 2,4-D and C 19490 showed a correlation coefficient of 0.914,
indicating that the two methods gave almost identical results. Especially with the
highly adsorbed urea and triazine herbicides, the uptake of biologically active material
by test plants was slightly less than the solvent-extractable parent compound plus its
metabolites, and so the absolute level of residues obtained by bioassay was lower. In
the case of fluorodifen, the correlation between the methods was not established. The
bioassay showed higher residues and slower degradation than chemical analysis. Vari-
ous factors which could explain this anomalous result are discussed.

H e r b i c i d e s r a n k first in the list o f pesticides a p p l i e d on the basis o f worldwide

c o n s u m p t i o n ( M e t z g e r 1972). The effectiveness o f a s o i l - a p p l i e d herbicide depends
p r i m a r i l y on its intrinsic selective p h y t o t o x i c i t y and secondarily on its availability
and persistence in soil.

Residues o f h e r b i c i d e s and their m e t a b o l i t e s are of concern because of the

f o l l o w i n g aspects:

-- c o n t a m i n a t i o n o f a n i m a l feed and human f o o d ,

-- e n v i r o n m e n t a l p o l l u t i o n , e. g . , residues in soil and water, and effects on
wildlife
-- p h y t o t o x i c i t y to the treated c o m m o d i t y or to rotational crops.

R e s i d u e s o f h e r b i c i d e s and their degradation products can be d e t e r m i n e d either by

c h e m i c a l a n a l y s i s or by b i o a s s a y techniques. Both methods have their advantages
and s h o r t c o m i n g s a c c o r d i n g to the o b j e c t i v e o f the experiment.

Archives of Environmental Contamination 101

and Toxicology Vol. 4, 101-118 (1976)
1976 by Springer-Verlag New York Inc.
 102 D. O. Eberle and H. R. Gerber

Principles of instrumental and bioassay m e t h o d s

A detailed discussion of existing methodology for herbicide residue analysis by

direct instrumental measurement and by bioassay techniques is beyond the scope of
this paper. Main emphasis will be laid on comparison and significance of results
obtained with several classes of herbicides by both techniques.

Bioassays. Many variations of the experimental design are described depending

on the mode of action of the herbicide the particular aspect under study and the test
organism used (Behrens 1970, Horowitz 1974). In the present bioassay studies with
monocotyledons and dicotyledons, the same experimental parameters were chosen as
described by Ebner (1966) for field degradation studies and Bieringer and Gerber
(1975) for greenhouse degradation studies. This type of bioassay was carded out in
the following way:

Various herbicide dosages are uniformly incorporated into airdried soil passed
through a two-mm sieve to give concentrations increasing in geometrical order, e.g.,
concentrations of 0.62, 1.25, 2.5, 5.0, 10.0, 20.0 ppm based on weight of air-dried
soil. The treated soil is brought to uniform moisture level and then incubated in the
dark at constant temperature. Soil samples are ventilated twice per week by shaking
and at the same time moisture is controlled by weighing. After various time
intervals, e.g., O, 15, 30, 60, 90 and 120 days after treatment two 200-g subsamples
are taken and stored at -25~ in order to stop any further degradation. As a rule six
concentrations and six time intervals are used yielding 36 subsamples. After the
final subsampling the soil is thawed and bioassayed using sensitive test plants. Plant
growth is assessed and/or fresh weights are recorded at a certain time interval after
planting, e.g., 12 to 15 days afterwards for growth inhibitor herbicides and 18 to 21
days afterwards for photosynthesis inhibitors, respectively.

From the results of the bioassay six individual dosage-response lines are ob-
tained, one for each individual time interval (Figure 1). These dosage-response lines
are plotted as shown in Figure 2, i.e., plant fresh weight (or assessment scores)
against log of concentration. From dosage-response lines established equieffective
concentrations of herbicides are determined for any individual timing. The loss of
original activity occurring between succeeding dosage-response lines is due to the
degradation of the herbicide. Therefore, the increase of concentration required for
equal effectiveness (e.g., 50%) at consecutive time intervals is calculated as percen-
tage of original activity at the time zero to yield dissipation of herbicidal activity.
Thus the dissipation curve of the herbicide examined is installed.

Bioassays with Chlorella pyrenoidosa, were performed in soil and water by

measuring the reduction in oxygen production caused by photosynthesis-inhibiting
herbicides in an illuminated Warburg apparatus, as described by Guth et al. (1970).

Basically, the validity of biological assay procedures rests on two assumptions.

They are: 1) that a given test plant-response to a herbicide increases proportionally
 Comparison of Methods for Herbicide Residue Analysis 103

ppm

~.~.OF--II--]I-II-I O
"---] Normal growth

r-] ] Partly damaged

O i Kill

20.
i KI
4~ II II II II II i
0 15 30 60 90 120

Days after treatment

Fig. i. Scheme of bioassay indicating dissipation of herbicides (6 timings x 6 concentra-

tions)

Plant
fresh
weight(g)
\ \
\
\ \
\ \
\\
I
I
I
I
I
I
\
i
I
I B C D E F
I
I
I }1
Log concentration (ppm)

Fig. 2 Dosage response lines of 6 individual timings. A-F are used to determine equieffective
concentrations of herbicide (Scheme)
 104 D. O. Eberle and H. R. Gerber

to the dose and 2) that, within the limits of sample variation and test conditions, this
response is reproducible.

Biotests have a number of limitations that should be recognized if a suitable assay

is to be developed. (1) Biological assays are not well suited to the determination of a
herbicide present in extracts of biological materials. Almost invariably there are
interfering components in the extract which influence the response of the test
organism. Therefore, the quantitative analysis of herbicide residues by plant biolog-
ical assay methods is restricted to soil samples, where the test organism can be
grown directly in the commodity to be analyzed. (2) Relatively narrow ranges of
environmental conditions and concentrations of the test compound can be included if
the biological assay is to be successful. (3) Some biological assays require a
considerable length of time to grow the test organism to a stage where responses will
be significant. For example, in herbicide bioassays with plants, a test period of 30
days is common. During this period, the plant may be exposed to a continually
changing concentration of the test compound which is being deactivated or lost from
the soil throughout the test. (4) The response of organisms in a biological assay is an
indirect measurement of the herbicide concentration. Significant amounts of the test
compound may be adsorbed by the soil so that it is not available to the test plant.
This may be advantageous or detrimental depending on the objective of the experi-
ment. If the objective is to determine the phytotoxic residue left in soil, i.e., to
measure the amount of test compound available to affect plant growth, it is advan-
tageous, while if the objective is to determine the total residue of applied herbicide
and metabolites, it is detrimental. Adsorption of the test compound by soil directly
influences the limit of detection that can be obtained with bioassay techniques.
Strongly adsorbed herbicides can be detected in soil only if residues are o f the order
of one to five ppm whereas weakly adsorbed compounds become detectable at the
0. I ppm level or below.

By the bioassay technique using Chlorella pyrenoidosa as test organism, some of

the above limitations can be overcome. The Chlorella test can be performed on a
routine basis with soil or water samples within a few hours, with a minimum of
laboratory or greenhouse space. Frequently, crude sample extracts of soil and water
are directly amenable to Chlorella-bioassay without any cleanup, a fact which
contributes considerably to the significance and importance of this method. Unfor-
tunately it is restricted to photosynthesis-inhibiting herbicides, and the problem of
low extraction efficiency of the extraction solvent represents a main limiting factor.

Instrumental techniques. Currently, excellent surveys on existing instrumental

methodology for residue analysis are available in the literature (Gunther and Blinn
1955, Zweig 1963, 1964, 1967, and 1973, Gould 1971, Jarczyk 1974). Spec-
trophotometric and chromatographic methods both alone and in combination, e.g.,
thin layer chromatography, gas chromatography with element specific detectors,
high pressure liquid chromatography (Brown 1973) and mass fragmentography
(Biros 1971) are residue detection methods of prime importance today, combining
high sensitivity with specificity.
 Comparison of Methods for Herbicide Residue Analysis 105

A problem of special significance in modern residue analysis is the determination

of metabolites. These products, formed by chemical, photochemical, or biochemical
alterations of the applied chemical, can be of reduced or increased mammalian
toxicity and phytotoxicity, or may persist in soil and in wildlife. Metabolites are
usually chemicals of greater polarity than the original pesticide, and tend to form
complexes with natural plant or soil constituents. As a practical consequence for
residue analysis, more drastic conditions of extraction in regard to polarity of
solvents and duration and temperature of equilibration have to be selected. In
addition, more extended and complicated cleanup steps become necessary and
frequently the final determination by chromatographic techniques is only possible
after derivative formation of the metabolite molecule.

Modern instrumental techniques are much more specific and generally more
sensitive than bioassays. Residue analysis by chemical assay is also much more
versatile in regard to the commodities which are amenable to analytical investiga-
tions. Furthermore, automation of instrumental techniques for monitoring residues
o f herbicides and metabolites has been successfully achieved by Guth and Voss
(1971) and HSrmann and Eberle (1971), so that large sample series can be analyzed
within short time periods. Residue data so obtained from soil, plants, water and
wildlife are well suited to give information on the ultimate fate of herbicides in the
environment by quantifying the sum of active ingredient and metabolites, i.e., the
total residue at measured time intervals after application.

The two main limitations of instrumental techniques are: i) analytical data in soil,
expressed in terms of ppm, give no information on the phytotoxic significance of the
residue for target weeds or rotational crops, and 2) the interpretation of residue data
obtained by instrumental methods relies mainly on the assumption that the extraction
procedure removes the total residue, both free and bound, from the commodity. This
assumption is ussually thought to be justified if control samples, spiked with
herbicide prior to the extraction, can successfully be subjected to the analytical
procedure. If recoveries are above 70%, the method is usually considered as valid.
However, it must be borne in mind that this approach is only suitable to estimate the
validity of cleanup and instrumental assay procedures, but not of the efficiency of
the extraction step. Complete extraction can only be proven by analyzing samples
treated with radiolabelled material and measuring the percentage of -radioactivity
recovered in the stripping solution. This fact has frequently been overlooked in the
past.

Depending on the type of herbicide under investigation, several of the instrumen-

tal techniques mentioned were applied in the following residue studies; therefore no
generally applicable instrumental technique can be recommended as it can in the
bioassay studies.

Results and discussion

In the present study field and laboratory experiments were carried out with the
following types of herbicides: ureas, triazines, phenoxyacetic acids, dithiophos-
Table I. Chemical structure, names, mode of action and analytical techniques for determination of residues of herbicides
included in the present study

Bioassay test organism and

Chemical name Mode of action instrumental techniques applied

3-(3-chloro-4-methylphenyl)- 1,1 - Photosynthesis- Arena s., Cl, lorella p.,

dimethyl-urea inhibitor colorimetry, TLC
C 2242, chlorotoluron
9
2-ethylamino-4-isopropylamino-6- Photosynthesis- Arena s., Chlorella p., t-rl
~r"
methylthio-s-triazine inhibitor GLC, TLC, liquid chromatography
G 34162, ametryn
I:L
2-chloro-4-eth~lamino-6-t- Photosynthesis- Chlorella p.,
butylamino-s-triazi ne inhibitor t4C-scintillation counting
GS 13529, terbuthylazine ~o

2,4-dichlorophenoxy-acetic acid Hormone type Sinapis a.,

2,4-D GLC after methylation

S-(2-methyl- 1-piperid ylcarbonyl methyl)- Germination Lolium p.,

O,O-di-n-propyl dithiophosphate inhibitor GLC
C 19490

2-nitro-ot,a,a-trifluoro-p-tolyl- Inhibition of Sinapis a.,

4-nitrophenyl ether ATP-production GLC
C 6989, fluorodifen
 Comparison of Methods for Herbicide Residue Analysis 107

phates and dinitrophenyl ethers. Chemical names, mode of action and analytical
techniques used in the present study are summarized in Table I.

In the field experiments, representative soil samples were obtained by a special

sampling technique described by Hrrmann et al. (1973) at pre-arranged time inter-
vals after application. Degradation curves were established, both by chemical
methods and bioassay techniques using as test organism either dicotyledons (Sinapis

ppm O.- ~ O Bioassay, avena

50 X - - --x Bioassay, chlorella
+--+ Chemical analysis, unchanged chlorotoluron
C)~C) Chemical analysis, total residues

0.5

0.1 .
I 0 ! ! , !
I0 90 120 150 180
Days after application

Fig. 3. Rate of disappearance of chlorotoluron in the 0 to 5 cm layer of plant free Swiss

sandy loam under fiel.d conditions. (Guth et al. 1970). Soil composition: 20.7% clay, 13.7%
silt, 65.6% sand, 3.2% humus, pH 5.6. Chlorotoluron applied at a rate of 6 kg a.i./ha. Total
3-chloro-4-methylaniline-containing residues determined chemically by subjecting soil sam-
ples to alkaline hydrolysis, steam distillatiorf, and colorimetric measurement after diazotiza-
tion and coupling of anilines to form a red azo-dye. (Friestad 1967). Limit of detection: 0.05
ppm. Structurally unchanged herbicide measured after exhaustive soil extraction with
acetone/hydrochloric acid and separation of the herbicide from its potential metabolites by
thin layer chromatography. (Guth et al. 1970). Limit of detection: 0.05 ppm. Bioassay
carried out with Avena sativa seedlings (detection limit 0.2 ppm) and Chlorella pyrenoidosa.
(Extraction of soil by acetone/hydrochloric acid, detection limit 0.1 ppm).
t08 D. O. Eberle and H. R. Gerber

alba ) and monocotyledons (Avena sativa, Lolium perenne ) or Chorella pyrenoidosa.

Greenhouse studies were performed by treating soil with a range of concentrations
of the herbicides, formulated as wettable powders. The soil was then kept in the
greenhouse at 70% of water capacity. Representative subsamples wereLtaken after
different time intervals and subdivided for chemical analysis and bioassay.

Figure 3 represents the result of a field degradation study performed with the urea
herbicide'chlorotoluron. Total residues were determined by an automated instrumen-
tal method (Friestad 1967). The extraction-efficiency of the hydrolysis/extraction
step was proven by using soil samples from a field study, where chlorotoluron,
labelled with C 14 in the methyl group attached to the phenyl ring was applied at a
rate of two kg a.i./ha. Sixty and 90 days after treatment, 78% and 81% respectively
of the total radioactivity was extractable and determined colorimetrically by the total
residue method. This method determined both the unaltered ureaherbicide and all
degradation products still containing the aniline-moiety. These are formed according
to the following general degradation scheme (Figure 4); by stepwise demethylation,
hydrolysis, and/or conjugation (Geissbfihler 1969).

The herbicidal activity of products at different levels in this degradation scheme

decreases from A to C, i.e., from dialkyl-substituted through monoalkyl-substituted
to unsubstituted ureas. The latter demonstrate a weaker photosynthesis-inhibiting
effect; the final aniline (D) possesses no herbicidal activity (Blass and Voss 1972).

0 RZ

| \R2

o
II 7 H ~ II ~R].
| - N"--C-N \

9 H
.@ IJ J
R3 - NH--C--N

~
(H20)
~'H

9
R3~L'~ -NH 2 + C 0 2 + N H 3

Fig. 4. General pathway of degradation of urea herbicides.

R,, = CH3, OCH3
R2 = Alkyl
R3 = H, Alkyl or Halogen
 Comparison of Methods for Herbicide Residue Analysis 109

Therefore soil degradation curves of urea herbicides as obtained by chemical

analysis of the total residue must be interpreted with care; Figure 3 demonstrates
that unchanged chlorotoluron disappears considerably faster than the total of all
3-chloro-4-methylaniline-containing residues. Both biotest methods yield approxi-
mately the same degradation curves as the chemical assay of the unaltered molecule.
This means that the acetone/hydrochloric acid extraction step used in the latter
method removes practically all of the herbicidally-active soil residues.

From the close agreement between bioassay results and those of analysis of parent
chlorotoluron, it can further be concluded that monoalkyl-substituted and unsubsti-
tuted ureas are not major soil metabolites. The percentage of total residue which is
not parent compound and not herbicidally active must consist of metabolites which
are not extractable by acetone/hydrochloric acid, but which are convertable to
aniline-type compounds by alkaline hydrolysis.

ppm ~ ~0 Bioassay, chlorella

X m=x Bioassay, avena
10
+ , ~" Chemical analysis, unchanged ametryn
5 ~ ~ ~ Chemica' anaJ~sisf t~ residues

0.5 x--x..~ \
- "~ ~%7~'~- " " " " " ~ "--X

+"~,~.~
0.05

I I I I
100 200 300 400
Days after application

Fig. 5. Rate of disappearance of ametryn in the 0 to 5 cm layer of a plant free Swiss clay
loam under field conditions. Soil composition: 12% clay, 49% silt, 38% sand, 3.6% humus,
pH 6.1. Ametryn applied at a rate of 5 kg a.i./ha. Residues of ametryn and dealkylated
metabolites still containing the methylthio-group determined chemically by extraction with
methanol/water or acetonitrile/water, alumina column chromatography and gas chromatog-
raphy (Ramsteiner et al. 1971 and Ramsteiner et al. 1974) both having a limit of detection of
0.04 ppm. Hydroxy triazines determined after methanol extraction and ion exchange column
cleanup by thin layer chromatography or high pressure liquid chromatography using spec-
trophotometric detection at 240 nm. Total residues obtained by addition of individual con-
centrations at each time interval. Bioassay carried out with A v e n a (limit of detection 0.2
ppm) and C h l o r e l l a (limit of detection 0.02 ppm).
 110 D. O. Eberle and H. R. Gerber

Figure 5 contains results of a similar dissipation study performed with the triazine
herbicide ametryn..In soils, the molecule of the active ingredient is subjected to
photolytic, hydrolytic and microbial degradation, whereby two basic pathways are
followed (Knuesli et al. 1969): 1) desalkylation of the amino sidechains in the 3-
and 6-position and 2) removal of the methylthio-group at the 2-position followed by
hydroxylation. (Figure 6). As long as the -methylthio-substituent in the 2-position is
unaltered, triazine metabolites demonstrate herbicidal activity, independently of the
extent of side-chain-dealkylation. Yet this activity disappears completely upon for-
mation of the 2-hydroxy-triazines.

With this general picture of the triazine metabolism in mind, the data in Figure 5
favor the following interpretation. Both bioassay techniques which detect the sum
of unaltered herbicide and side-chain dealkylated methylthio triazine metabolites
yield almost identical soil dissipation curves. Both curves closely follow the degra-
dation curve of unchanged herbicide as established by chemical analysis. This
correspondence leads to the conclusion that dealkylated methylthio-triazines repre-
sent only minor metabolites of ametryn, and that the acetonitrile/water extraction-
procedure used for the chemical assay extracts quantitatively the sum of phytotoxic
residues available to the test organisms in the soil. The remaining percentage of the
total residue which is not detectable by bioassays consists of hydroxytriazines,
mainly of 2-ethylamino- 4-isopropylamino-6- hydroxytriazine, as shown by separate
chemical analysis of this class of metabolites.

The validity of the chemical assay method was demonstrated by extracting soil
samples from a radioactive field dissipation study with 14C-ring-labelled ametryn.
Twenty weeks after application of two kg a.i./ha, 70% of the total radioactivity was
extractable by the above residue method (Ramsteiner et al. 1974) and 67% was
identified as either parent compound or dealkylated and hydroxylated triazine-
metabolites.

Bioassays are not restricted to herbicide dissipation studies in soil. A different

type of study is presented in Table II. This model-experiment was designed to study
the leaching behavior of the triazine herbicide terbuthylazine in three types of soil
by using soil columns (Gerber and Guth 1973). After application of 14C-labelled
terbuthylazine on top of the columns and percolation of an amount of water corres-
ponding to 200 mm of rain, two-cm segments were cut and analyzed separately by
scintillation counting and Chlorella-bioassay. The percentages of applied herbicide
found in the individual two cm segments by either method demonstrate again the
good correspondence of data obtained by instrumental and bioassay techniques in
such types of study whereby the bioassay method is inferior in regard to sensitivity.

Whereas the Chlorella-bioassay, as mentioned, is restricted to photosynthesis-

inhibiting herbicides, other biotests using higher plants as test organisms can suc-
cessfully be extended to degradation studies with compounds of different modes of
action. Figure 7 presents results of a greenhouse and a field study performed in the
same type of soil with 2,4-D. This chemical is known to act as a hormone-type
 X

N~---.N X
O
R1-HNL N//!~NH--R2
NH--R 2

oJ

\ x/ O

.J%. N-'~N
OH H2N N~/" "-NH_R2
NH--R 2 t~

E
t~

OH
R,_.NI~N~I'-.N.2 \ x J
~N E
t.-

H2"~N~ NH2 >

2N NH2
Fig. 6. General pathways of degradation of triazine herbicides, whereby X = CI, CH3S, CHzO; R~ = Alkyl, Methoxyalkyl. right:
Side-chain dealkylation, left: Hydroxytriazine-formation.
 112 D. O. Eberle and H. R. Gerber

Table l I . Leaching of terbuthylazine in three soil types, determined by

Chlorella-bioassay and 14c-scintillation counting (Blass and Voss 1972). Numbers
represent percent of applied herbicide found in consecutive vertical two-cm layers

cii1 Chlorella C ~4 C hlorella C 14 C hlorella C 14

0
64 78 46 61 27 22
2
<5 I 27 24 8 8
<5 <0.2 <5 0.6 18 14
6
<5 <0.2 <5 0.4 25 24
8
<5 <0.2 <5 <0.2 21 17
10
<5 <0.2 <5 <0.2 <5 4
12
<5 <0.2 <5 <0.2 <5 0.4
14
<5 <0.2 <5 <0.2 <5 <0.2
16
<5 <0.2 <5 <0.2 <5 <0.2
18
<5 <0.2 <5 <0.2 <5 <0.2
20
<5 <0.2 <5 <0.2 <5 <0.2
22
<5 <0.2 <5 <0.2 <5 <0.2
24
<5 <0.2 <5 <0.2 <5 <0.2
26
<5 <0.2 <5 <0.2 <5 <0.2
28
<5 <0.2 <5 <0.2 <5 <0.2
30
Sandy Loam Clay Loam Sandy Soil

herbicide. The compound was applied in the field experiments at a rate of five kg
a.i./ha. Samples from both studies were analyzed simultaneously by chemical
analysis and by bioassay with Sinapis alba.

All dissipation curves demonstrate a rapid dissipation of 2,4-D, whereby the

degradation rate is identical in the greenhouse experiment and in the 0-10 cm soil
layer in the field study, independently of the analytical method applied.

The main mechanisms of metabolic alteration of 2,4-D in soil are known to be

side-chain oxidation and ring hydroxylation (Maier-Bode 1971). Both pathways
represent potent detoxication mechanisms, because none of the metabolites formed
possesses herbicidal activity (Loos 1969). This was confirmed again in the present
study by the correspondence of bioassay and chemical analysis. The fact that the
rate of degradation is identical under greenhouse and field conditions further de-
monstrates that the main degradation factors must be hydrolytic and biochemical
breakdown. Photolysis and leaching, usually more pronounced in the field and
therefore causing a faster dissipation than in the greenhouse, seemed to play only a
minor role in the above study.
 Comparison of Methods for Herbicide Residue Analysis 113

Residuesin ppm
l~ ~=~O Greenhousestudy, bioassay,sinapis
~k~ tP----Q Greenhousestudy, chemical analysis
5
~'~ ~ Field study, bioassay,sinapis
~ X X F=eldstudy chem=calanalys=s

1 -

0.5

0.2,

0.1
I I I
10 20
Days after application

Fig. 7. Rate of dissipation of 2,4-D in the 0-10 cm layer of Swiss sandy loam under field and
greenhouse conditions. Soil composition as in Fig. 3. Herbicide applied at a rate or 5 kg/ha.
Residues determined chemically according to Maier-Bode (1971) by extracting soil with ! N
NaOH, reextraction of the acidified solution with methylene chloride and determination of
2,4-D by GLC after methylation with diazomethane. Limit of detection: 0.02 ppm. Bioassay
carried out with Sinapis alba, detection limit 0.5 ppm

A further comparative study of instrumental and bioassay methods was performed

with the new rice herbicide C 19490, which chemically belongs to the group of
dithiophosphates. Its mode of action is probably connected with inhibition of
germination (Figure 8).

With this compound bioassays in soil can be performed using Lolium perenne as
test plants, whereas chemical analysis is based on gas chromatography of unchanged
herbicide. Analytical results with both methods demonstrate that somewhat lower
residues were detected by the chemical assay. Although no information is available
so far on the identity and biological effects of soil metabolites, the above degrada-
tion curves provide some evidence either that the chemical method gives no quan-
titative extraction of parent compound, or that certain metabolites still have some
herbicidal activity.

The same differences between bioassay and chemical results but in more pro-
nounced form, arise from a similar greenhouse study with fluorodifen. (Figure 9).
114 D. O. Eberle and H. R. Gerber

ppm O===mO 8ioassay,Ioleum

10 - ~ X~X Chemicalanalysis

1
I I I I 7
20 40 60 90 140
Days after application
Fig. 8. Rate of dissappearance of C 19490 in a Swiss clay loam in a greenhouse study. Soil
composition: 27% clay, 32% silt, 31% sand, 3.2 % humus, pH 6.4. Residues of C 19490
determined chemically by GLC after extraction of soil with methanol/water. Limit of detec-
tion 0.002 ppm. Bioassay with Lolium perenne: Limit of detection 0.5 ppm

ppm o==,===o Bioassay,sinapis

10 -~ ~ , x~x Chemicalanalysis

/
5 x

1
I I l I | I

20 40 60 80 100 120
Days after application
Figure 9. Rate of dissipation of fluorodifen in a Swiss clay loam in a greenhouse study. Soil
composition as in Fig. 6. Residues of fluorodifen determined chemically by GLC after
extracting soil with methanol/water. Limit of detection: 0.02 ppm. Bioassay tried with
Sinapis alba. Limit of detection 2 to 5 ppm
 Comparison of Methods for Herbicide Residue Analysis 115

Soil residue data as established by the two methods - - bioassay with Sinapis and
c h e m i c a l analysis by gas chromatography - - were no longer identical, but de-
m o n s t r a t e d significant differences in regard to the initial residue and the speed of
d e g r a d a t i o n . In addition to the explanation postulated earlier - - herbicidal activity
o f m e t a b o l i t e s not detected by the chemical assay - - we must bear in mind that the
c h e m i c a l method was not valid in regard to the efficiency of extraction. In a field

ppm
10 + 0

9 84

8- X

.o 6

.D

~ 5 +

n~
ox~
3

8
2'

! ! I ! ! I I ! I
1 2 3 4 5 6 7 8 9 ppm
x = Residues by chemical analysis

Fig. 10. Comparison of soil residue data.obtained both by bioassays and instrumental
methods. Bioassay data mentioned in this study plotted against the corresponding residues of
parent herbicide, determined by chemical methods.
o = chlorotoluron + -- 2,4-D,
x = ametryn O --- C 19490
Regression line y = 1.003 x + 0.086 obtained by using Wang 700 desk top computer;
fluorodifen data not included in calculation. Correlation coeficient of regression line 0.914
116 D. O. Eberle and H. R. Gerber

study with 14C-radiolabelled fluorodifen, only 36% of the total radioactivity present
in soil were extractable by methanol/water 120 days after application of five kg/ha
fluorodifen. In addition considerable experimental difficulties were encountered
with the flurodifen - - bioassay method and contributed to the unsatisfactory agree-
ment of results. So far, no test organism has been found which fulfills one of the
basic requirements for successful performance of bioassays, namely an increase of
the test plant response proportional to the concentration of herbicide. Therefore both
methods cannot be considered as valid in the fluorodifen study and any interpreta-
tion of data without knowledge of those experimental limitations results in unrealis-
tic conclusions.

Conclusions

The above comparison of several instrumental and bioassay methods performed

with herbicides of different chemical classes and modes of action demonstrate the
significance and validity of both techniques in herbicide/soil studies. As long as the
basic requirements for their successful application are fulfilled and the limitations
recognized, bioassays and instrumental techniques are considered as important com-
plementary analytical tools, allowing significant conclusions in regard to the persis-
tence and mode of action of metabolites to be drawn.

In many cases, the determination of the parent compound by instrumental

techniques and of phytotoxic residues by biotest yield an identical soil dissipation
picture. This is demonstrated in Figure 10 where chemical and bioassay results from
the above studies are plotted on a linear scale. A regression line was calculated and
the coefficient of correlation found to be 0.914, showing excellent correlation
between the two techniques, with the exception of the data on fluorodifen. This
means that dissipation of herbicides in soil usually follows a detoxification route,
leading from herbicidal compounds to non-phytotoxic metabolites, either by micro-
bial cleavage of the original molecule, or by adsorption or complex-formation
mainly with organic soil constituents.

References

Behrens, R.: Quantitative determination of triazine herbicides in soils by bioassay.

Residue Reviews 32, 355 (1970).
Bieringer, H. and H. R. Gerber: Weed Res. 14, in preparation (1974).
Biros, F. J.: Recent applications of mass spectrometry and combined gas chromatog-
raphy - - mass spectrometry to pesticide residue analysis. Residue Reviews 40,
1 (1971).
Blass, W. and G. Voss: Ein Chlorella-Biotest zur Bestimmung photosynthese-
hemmender Herbizide in B5den und Gew~issern. Schr. Reihe Ver. Wass.-
Boden-Lufthyg. Berlin, Dahlem, A. 37, Stuttgart, 21 (1972).
 Comparison of Methods for Herbicide Residue Analysis 117

Brown, P. R.: High pressure liquid chromatography. P1, Academic press, N.Y.,
London (1973).
Ebner, L.: Studies on the inactivation of Patoran (N-p-bromphenyl)-N'-methoxyurea
in two different soil types using bioassay. Z. Pfl. Krank. Pfl. Path. Pfl Schutz
73, 458 (1966).
Friestad, H. O.: Determination of linuron in soil by application of an automated
diazotisation and coupling procedure. Bull. Environ. Contam. Toxicol. 2,236
(1967).
Geissbiihler, H.: The substituted ureas. In Degradation of herbicides, 79, P. C.
Kearney and D. D., Kaufman Eds. Marcel Dekker, Inc. N.Y. (1969).
Gerber, H. R. and J. A. Guth: Short theory, techniques and practical importance of
leaching and adsorption studies. Proc. Eur. Weed. Res. Coun. Symp.
Herbicides-Soil. 51 (1973).
Gould, R. F.: Pesticides identification at the residue level. Adv. in Chemistry Series
104, 1 (1971).
Gunther, F. A. and R. C. Blinn: Analysis of insecticides and acaricides. Intersci-
ence publishers, Inc. N.Y. (1955)
Guth, J. A., G. Voss, and L. Ebner: Der Abbau des Getreide-Herbizides Dicuran in
B6den, bestimmt dutch zwei verschiedene Biotests und eine automatische
colorimetrische Methode. Z. Pflkr. Pflsch. Sonderheft v, (1970).
Guth, J. A. and G. Voss: Automated colorimetric procedure for the determination of
total and unchanged urea herbicide residues in soil. Weed Res. 11, 111
(1971).
H/Srmann, W. D. and D. O. Eberle: Mechanization of the gas chromatographic
determination of pesticide residues in soil, food and feeds. Proc. 2nd Int.
IUPAC Congress of Pesticide Chem., Tel Aviv, 4, 525 (197l).
H6rmann, W. D., B. Karlhuber, K. A. Ramsteiner and D. O. Eberle: Soil sampling
for residue analysis. Proc. Eur. Weed Res. Coun. Symp. Herbicides Soil. 129
(1973).
Horowitz, M.: Bioassay techniques for the analysis of herbicides. Proc. of the 3rd
Int. IUPAC Congress of Pesticide Chem., Helsinki, in press. Georg Thieme
Verlag, Stuttgart (1974).
Jarczyk, H. J.: Instrumental techniques for the analysis of herbicides. Proc. of the
3rd Int., IUPAC Congress of Pesticide Chem. Helsinki, in press. (1974).
Kn/.isli, E., D. Berrer, G. Dupuis, and H. Esser: s-Triazines. In Degradation of
herbicides, 51, P. C. Kearney and D. D. Kaufman Eds., Marcel Dekker, Inc.
N. Y. (1969).
Loos, M. A., Phenoxyalkanoic acids: In Degradation of herbicides, l. P. C.
Kearney and D. D. Kaufman Eds., Marcel Dekker, Inc. N. Y. (1969).
Maier-Bode, H.: Herbizide und ihre Riicks6inde, 62-84. Eugen Ulmer, Stuttgart
(197l).
118 D. O. Eberle and H. R. Gerber

Metzger, H.: Pflanzenschutz 1980 aus der Sicht der Industrie. "VCI Schriftenreihe
Chemie und Fortschritt" 2, 3 (1972).
Ramsteiner, K. A., W. D. H/Srmann, and D. O. Eberle: Bestimmung von
Triazinmetaboliten-Rfickstiinden in Boden. Meded. Rijksfac. Landbouw-
wetensch., 36, 1119 (1971).
Ramsteiner, K. A., W. D. H6rmann, and D. O. Eberle: Multiresidue method for the
determination of triazine herbicides in field-grown agricultural crops, water,
and soil. J. Ass. Offic. Anal. Chem. 57, 192 (1974).
Zweig, G., Analytical methods for pesticides, plant growth regulators, and food
additives. Vol. I., Academic press, N. Y., London (1963).
Ibid, Vols. II, III, IV (1964).
Ibid, Vol. V (1967).
Ibid, Vol. VI (1973).

Manuscript received September 18, 1974; accepted December I0, 1974.