PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [DEFF]On: 25 August 2010Access details: Access Details: [subscription number 910491762]Publisher Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Environmental Science and Health, Part BPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713597269

Fate and availability of glyphosate and AMPA in agricultural soilLouise Simonsenab; Inge S. Fomsgaarda; Bo Svensmarkb; Niels Henrik Spliida

a Faculty of Agricultural Sciences, Department of Integrated Pest Management, University of Aarhus,Slagelse, Denmark b Department of Chemistry, University of Copenhagen, Copenhagen, Denmark

To cite this Article Simonsen, Louise , Fomsgaard, Inge S. , Svensmark, Bo and Spliid, Niels Henrik(2008) 'Fate andavailability of glyphosate and AMPA in agricultural soil', Journal of Environmental Science and Health, Part B, 43: 5, 365— 375To link to this Article: DOI: 10.1080/03601230802062000URL: http://dx.doi.org/10.1080/03601230802062000

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

Journal of Environmental Science and Health Part B (2008) 43, 365–375Copyright C© Taylor & Francis Group, LLCISSN: 0360-1234 (Print); 1532-4109 (Online)DOI: 10.1080/03601230802062000

Fate and availability of glyphosate and AMPAin agricultural soil

LOUISE SIMONSEN1,2, INGE S. FOMSGAARD1, BO SVENSMARK2 and NIELS HENRIK SPLIID1

1Faculty of Agricultural Sciences, Department of Integrated Pest Management, University of Aarhus, Slagelse, Denmark2Department of Chemistry,University of Copenhagen, Copenhagen, Denmark

The fate of glyphosate and its degradation product aminomethylphosphonic acid (AMPA) was studied in soil. Labeled glyphosatewas used to be able to distinguish the measured quantities of glyphosate and AMPA from the background values since the soil wassampled in a field where glyphosate had been used formerly. After addition of labeled glyphosate, the disappearance of glyphosateand the formation and disappearance of AMPA were monitored. The resulting curves were fitted according to a new EU guideline.The best fit of the glyphosate degradation data was obtained using a first-order multi compartment (FOMC) model. DT50 values of 9days (glyphosate) and 32 days (AMPA) indicated relatively rapid degradation. After an aging period of 6 months, the leaching risk ofeach residue was determined by treating the soil with pure water or a phosphate solution (pH 6), to simulate rain over a non-fertilizedor fertilized field, respectively. Significantly larger (p < 0.05) amounts of aged glyphosate and AMPA were extracted from the soilwhen phosphate solution was used as an extraction agent, compared with pure water. This indicates that the risk of leaching of agedglyphosate and AMPA residues from soil is greater in fertilized soil. The blank soil, to which 252 g glyphosate/ha was applied 21months before this study, contained 0.81 ng glyphosate/g dry soil and 10.46 ng AMPA/g dry soil at the start of the study. Blank soilsamples were used as controls without glyphosate addition. After incubation of the blank soil samples for 6 months, a significantlylarger amount of AMPA was extracted from the soil treated with phosphate solution than from that treated with pure water. Todetermine the degree of uptake of aged glyphosate residues by crops growing in the soil, 14C-labeled glyphosate was applied to soil6.5 months prior to sowing rape and barley seeds. After 41 days, 0.006 ± 0.002% and 0.005 ± 0.001% of the applied radioactivity wasmeasured in rape and barley, respectively.

Keywords: Glyphosate; AMPA; aged residues; degradation; leaching; plant uptake; phosphate.

Introduction

As contact time between chemicals and soil increases,the non-extractable (bound) fraction of the chemicalsincreases.[1] The chemicals become progressively less avail-able for uptake by organisms,[2] and hence for biodegrada-tion. This process has been termed aging. It is believed that,during aging, chemicals bind strongly to the soil[3] or moveslowly into sites within the soil matrix that are not readilyaccessed by even the smallest of microorganisms.[4] Barra-clough et al.[5] discussed the issue of bound residues in soiland concluded that for single additions of individual com-pounds, the formation of bound residues probably does not

Address correspondence to Inge S. Fomsgaard, University ofAarhus, Faculty of Agricultural Sciences, Department of Inte-grated Pest Management, Forsøgsvej 1, 4200 Slagelse, Denmark;E-mail: Inge.Fomsgaard@agrsci.dkReceived January 21, 2008.

pose an environmental problem. However, the long-termsignificance of bound residues formed from multiple addi-tions of a number of compounds is less clear. Barracloughet al.[5] recommend that future work should emphasize thebiological significance of such residues and their release.

Glyphosate (N-(phosphonomethyl)glycine) is a commonherbicide that has been well-studied. However, the fatesof aged glyphosate residues and the glyphosate degrada-tion product aminomethylphosphonic acid (AMPA) havenot been studied in soil. Glyphosate is a foliar herbicideapplied to crops, but some leaches into the soil. The rateof glyphosate degradation depends on soil type, pH, andtemperature.[6−9] Glyphosate is generally considered to bedegraded at a high rate and it is strongly adsorbed tosoil,[10] so the risk of leaching is assumed to be low. Eventhough there is evidence to support this assumption,[11]

glyphosate still poses a risk to the groundwater.[12−14] Be-tween 1993 and 2003, glyphosate and AMPA were de-tected in 2.6% of samples from monitored groundwaterwells in Denmark.[15] In a lysimeter experiment, Foms-gaard et al.[16] measured glyphosate and AMPA in leachates

Downloaded By: [DEFF] At: 16:23 25 August 2010

366 Simonsen et al.

from treated soil (sampled from 1 m depth). Two yearsafter the last spraying, glyphosate and AMPA were de-tected in several samples of water at concentrations up toand above 0.1 µg/L. This indicates that not only newlyapplied glyphosate but also aged residues pose a risk togroundwater.

It is possible that phosphate has an influence on theamount of glyphosate leached from soil. Glyphosate bindsto soil through the phosphonic acid moiety[17] and there-fore competes with phosphate for binding sites.[18] Gimsinget al.[19] found that the addition of phosphate to soil de-creased the adsorption of glyphosate that had previouslybeen added to the soil. Dion et al.[20] added glyphosate tosoil solutions containing different concentrations of phos-phate and found that adsorption of glyphosate decreased asthe initial level of phosphate increased. However, the effectsof phosphate on leaching of aged glyphosate and AMPAresidues from soil have not been studied to date.

The uptake of aged glyphosate and AMPA residues bycrop roots is another area of concern. It is documented thatglyphosate can be taken up through the roots of a range ofplant species,[18,21−24] and more information about uptakeof aged residues of glyphosate and AMPA from field soilsis needed.

Traditionally, single first-order kinetics have been used todescribe the degradation of pesticides because the equationis simple. Furthermore, the endpoint values, including DTxvalues (the time taken for x percent of a substance to dis-appear from an environmental compartment by dissipationprocesses, after it has been applied to, formed in or trans-ferred to that compartment) are easy to calculate. Non-linear kinetics may, however, better fit the data.[25] Beulkeand Brown[26] found that the method used for modeling ex-perimental pesticide degradation might have a large impacton the modeling outcome, stressing the importance of ap-proaches to ensure that experimental data are evaluated ina consistent way. Due to the lack of a standardized pro-cedure for calculating kinetic endpoints, an EU guidelinecontaining recommendations for calculating the degrada-tion kinetics of parent and metabolites in soil, water andwater sediment has been published.[27]

It is common practice to study a single process in sev-eral soils.[19,28−30] In the present study, however, several pro-cesses were studied in a single soil. This allowed us to studythe kinetics taking place during the aging period, the exactage of the residues and the availability of each residue forextraction by different extraction methods and uptake byplants.

The purpose of this study was to investigate the fate ofaged residues of glyphosate and AMPA in soil. Degradationof glyphosate was followed for 6 months, and the obtainedresults were modeled according to the new EU guideline.[27]

After 6 months the uptake of the residual aged glyphosateand AMPA by crops was studied. Also the influence ofphosphate fertilization on the extractability of the residualaged glyphosate and degradation products was studied.

Materials and methods

Chemicals and solutions

The following products were used in this study: 13C-15N-labeled glyphosate (HOOC*CH2*NHCH2PO(OH)2)(purity > 98%) (Cambridge Isotope Laboratories, Inc,Andover, USA); 14C-labeled glyphosate (HOOCCH2NH*CH2PO(OH)2) (purity > 99%, specific activity of 5.155MBq/mg) (Izotop, Institute of Isotopes Co., Ltd.,Budapest, Hungary); unlabeled standards of glyphosate(purity 98.5%) and AMPA (purity 99.0%) (Dr. Ehren-storfer GmbH, Augsburg, Germany); 9-fluorenylmethylchloroformate (98%) (Acros Organics, Geel, Belgium); 2-propanol (≥99.7%) and glycerol (≥99.5%) (J.T. Baker,Deventer, Netherlands); dichloromethane (HPLC grade)and methanol (HPLC grade) (Rathburn Chemicals Ltd.,Walkerburn, Scotland); ammonia solution (purity 25%,GR for analysis), di-sodium tetraborate decahydrate (GRfor analysis), potassium hydroxide pellets (≥85%, GR foranalysis), anhydrous sodium sulphate (≥99%, GR for anal-ysis), acetone (≥99.8%, for LC), ammonium acetate (≥98%,GR for analysis), acetonitrile (≥99.9%, gradient grade forLC), sulphuric acid (95-97%, GR for analysis) (MerckKGaA, Darmstadt, Germany); Spec-Chec, Carbo-sorb E,14C-W and 14C-O standard capsules, Permafluor E+, andUltima Gold (PerkinElmer, Waltham, USA); potassiumphosphate (>99%) (Sigma-Aldrich Chemie GmbH, Stein-heim, Germany).

A borate buffer was prepared by dissolving 5 g disodiumtetraborate in 100 mL MilliQ water, and was used to preparea borate solution (pH 9) by diluting 85 mL buffer in 530 mLMilliQ water. A stock solution of derivatized glyphosateand AMPA was prepared in 2.5 mL glyphosate and AMPAsolution (containing 21.16 mg/L glyphosate and 19.17mg/L AMPA), 0.4 mL borate buffer and 2.5 mL FMOCreagent (16 mg 9-fluorenylmethyl chloroformate dissolvedin 100 mL acetone). The solution was left at room tempera-ture for 30 min before being diluted in 1:1 methanol:MilliQwater in volumetric flasks to give the standards for high-performance liquid chromatography (HPLC).

A 14C-glyphosate stock solution was prepared in a volu-metric flask to give a concentration of 28.32 mg/L MilliQwater and 8.76 × 109 dpm (disintegrations per minute)/L.For the degradation and leaching studies a 14C-glyphosatesolution was prepared by diluting the stock solution withMilliQ water to 14.44 mg/L and 4.47 × 109 dpm/L.A 13C-15N-glyphosate solution of 24 mg/L MilliQ waterwas prepared. A solution of unlabeled glyphosate for theplant uptake study of 280.20 mg/L MilliQ water was alsoprepared.

Soil

The soil was sampled from the 0–2.5 cm layer of a field,where reduced tillage had been practiced, in Sandved

Downloaded By: [DEFF] At: 16:23 25 August 2010

Glyphosate and AMPA in agricultural soil 367

Table 1. Soil characteristics.

% Clay (<2 µm)a 13.3% Silt (2–20µm)a 16.2% Coarse silt (20–63 µm)a 24.7% Fine sand (63–200 µm)a 26.9% Coarse sand (200–2000 µm)a 16.2% Humusa 2.7Density (g/cm3) 2.626% CaCO3 n.d.pH-H2O 6.5P (Al)b 7K (Al)b 20Mg (Al)b 10Ca (Al)b 140P (HCl)b 39.9

a% of oven-dry soil.bP, K, Mg, Ca (mg/100 g dry soil) were determined by extraction withammonium lactate (Al) or hydrochloric acid (HCL).

(eastern Denmark), late in March 2005, before fertiliza-tion. The last time glyphosate had been sprayed was inthe beginning of October 2003. At this time glyphosate wasapplied at a rate of 252 g/ha or, assuming that glyphosatespread into the upper 7 cm, 0.24 µg/g soil. Characteristicsof the soil are listed in Table 1. All results are calculated inoven-dry soil

Splitting of soil

The soil was split using the paper cone riffle splitting methoddescribed by Gerlach et al.[31] Several levels of splitting wereperformed to obtain a total of 72 samples of 5 g each forthe degradation and extractability study, and 6 samples of300 g each for the plant uptake study.

Preparation and incubation of samples with addition oflabeled glyphosate for the degradation and extractabilitystudies

Since the soil was sampled from a field where glyphosatehad previously been used, we added labeled glyphosate tothe soil, in order to distinguish the glyphosate and AMPAthat were added in this study from background levels. 14C-glyphosate solution (110 µL) and 13C-15N-glyphosate solu-tion (180 µL) were applied to each of 50 samples of 5 g drysoil, in Falcon tubes (38 samples for the degradation study,12 for the extractability study). This resulted in a total con-centration of 1.18 µg glyphosate/g dry soil and 98225.38dpm/g dry soil, corresponding to normal spraying practiceof 1 kg/ha or 1 µg/g soil. Both the 14C-glyphosate and the13C-15N-glyphosate solutions contained a certain percent-age of unlabeled glyphosate. This was taken into consider-ation in all the following measurements and calculations.

The samples were placed without cap in hermeticallysealed 1-L glass jars, together with a vial containing 1 mL

glycerol and another containing 1 mL KOH (1 M). Thesamples were incubated at 14.3◦C. On day 0, 7, 14, 21,35, 49, 77, and 105, four samples were discontinued andstored at −18◦C, while six samples were discontinued onday 179. Every time incubated samples were discontinued,potassium hydroxide (KOH) traps were sampled and forthe remaining soil samples they were replaced. Glyceroltraps were only sampled on the day the corresponding soilwas discontinued. Ultima Gold (5 mL) was added to theKOH or glycerol, and radioactivity was measured using aWallac 1409 liquid scintillation counter (LSC). The watercontent of the soil samples was maintained at 50% of thewater-holding capacity (WHC) throughout the experimentby periodically weighing the samples and by adding water asrequired.

Preparation and incubation of blank samples for thedegradation and extractability studies

MilliQ water (290 µL) was applied to 22 blank soil samples(18 for the degradation study, 4 for the extractability study).The incubation was performed in the same way as for thesamples with labeled glyphosate, except that on day 0, 7,14, 21, 35, 49, 77, 105, and 179, only two samples werediscontinued and stored at −18◦C.

Analysis of samples from the degradation study

Extractions were performed on all samples to determine theamount of glyphosate and AMPA. After addition of 40 mLborate solution, the soil sample was ultrasonicated for 20minutes, extracted using a Holm and Halby Master Mixerfor 40 minutes and then centrifuged at 4000 rpm for 10minutes in a Sorvall RT7 centrifuge. The supernatant wastransferred to a 1-L separating funnel. This was repeatedtwice, with an ultrasonic time of 10 minutes and a mastermixer time of 20 minutes.

To measure radioactivity in the extracts, 100 µL was sam-pled from each separating funnel. Then 5 mL Ultima Goldwas added and the samples were counted on a Wallac 1409LSC.

To derivatize glyphosate and AMPA, 100 mL FMOCreagent was added to each separating funnel and the so-lutions were left to react for 30 minutes at room temper-ature. To remove impurities the extracts were shaken with100 mL dichlormethane/2-propanol (3 + 1) and the or-ganic phase was removed. This was repeated with 50 mLdichloromethane/2-propanol (3 + 1).

The solutions were made acidic by adding 8 mL 2.5M sulphuric acid to each separating funnel; hence theglyphosate- and AMPA-FMOC derivatives became solu-ble in the organic phase. Liquid–liquid extraction was car-ried out twice with 100 mL dichloromethane/2-propanol(3 + 1). To remove water, the organic phase was fil-tered into a round-bottom flask through water-free sodiumsulphate placed in a funnel stopped with cotton cleaned in

Downloaded By: [DEFF] At: 16:23 25 August 2010

368 Simonsen et al.

dichloromethane. Afterwards, the funnels were rinsed with5 mL dichloromethane/2-propanol (3 + 1) to ensure thatno glyphosate or AMPA was retained in the sodium sul-phate or cotton.

Twenty drops of 25% ammonia solution were added toalkalize the filtrates and make the glyphosate- and AMPA-FMOC derivatives adsorb to the inside of the flask. Thesolvent was evaporated using a rotation evaporator at 60◦Cand a vacuum between 0.2 and 0.8 bar, adjusted during theevaporation. The flask was dried with nitrogen gas, washedwith dichloromethane and dried again. The inside of theflask was carefully washed with 1 mL 50% methanol/MilliQwater to dissolve the adsorbed glyphosate- and AMPA-FMOC derivatives. The solution was transferred with Pas-teur pipette into a vial.

The samples were analyzed by HPLC with a tan-dem mass spectrometry detector (MS/MS) on a HewlettPackard 1100 system coupled with a PE SCIEX, API 2000from Applied Biosystems. The method was run with adeclustering potential of −30.0 V. The collision energy was−30.0 V and the collision gas pressure was 1 arbitrary unit.The ion spray voltage was −4200 V and dry gas tempera-ture was 450◦C. ESI was used as the ionization method. AHypersil BDS C18 column, 250 × 2.1 mm with a particlesize of 5 µm was used. Mobile phase A (99% 10 mM ammo-nium acetate, 1% methanol) and mobile phase B (10% am-monium acetate, 90% acetonitrile) were used in a gradientsystem: 0–2 min. 10% B; 2–12 min. linear gradient to 100%B; 12–18 min. 100% B; 18–19 min. linear gradient to 10% B;19–30 min. 10% B. The flow was 0.2 mL/min and the injec-tion volume was 100 µL. Detection was performed in neg-ative mode with multiple reaction monitoring (MRM) ofm/z 390/168 (glyphosate-FMOC derivative), m/z 332/110(AMPA-FMOC derivative), m/z 392/170 (14C-labeledand 13C-15N-labeled glyphosate-FMOC derivative), m/z333/111 (15N-AMPA-derivative).

The limit of detection (LOD) was determined as threetimes the signal-to-noise ratio on soil samples spiked withglyphosate and AMPA. The LOD was 0.04 ng/g dry soilfor glyphosate and 0.02 ng/g dry soil for AMPA.

A recovery experiment for validation of the extractionmethod was performed by adding glyphosate and AMPAat three different levels (10, 100 and 1000 ng/g dry soil)to 6 replicate samples of 5 g dry soil at each level. Thesamples were placed without cap at 5◦C for a day for equi-libration before extraction. The recovery percentages ob-tained were taken into consideration in the calculations ofresults.

Total amounts of extracted glyphosate and AMPA (la-beled and unlabeled originating from the glyphosate addedin this study) were calculated based on measured amountsof 14C-glyphosate+13C-15N-glyphosate and of 15N-AMPA,using the initial ratio between the different glyphosate com-pounds added.

In extracts of the blank soil samples, the amounts of un-labeled glyphosate and AMPA were measured.

Extractability study

After 179 days of incubation, samples were extracted as de-scribed above, with the exception that the borate solutionwas replaced with other extraction agents. Half of the 12samples and 4 blank samples were extracted with MilliQwater, the other half with phosphate solution (6.46 mmolphosphate/L MilliQ water, pH 6). This concentration cor-responds to the concentration in rainwater, when 10 mmrain falls over a field fertilized with 20 kg phosphorus/ha.

Analysis of non-extractable residues

Soil samples, extracted in the degradation and extractabilitystudies with added labeled glyphosate, were dried at 60◦Cfor a day, ground and mixed in a Retsch Mortar GrinderRM 100 for 15 min and stored at −18◦C. Triplicate aliquotsof 0.1 g soil from each sample were combusted in a PackardOxidizer model 307 for 1 min. Evolved 14CO2 was capturedin 5 mL carbo-sorb E and mixed with 15 mL permaflourE in an Econo counting vial. Radioactivity was measuredon a Wallac 1409 LSC. A standard soil was prepared byadding 400 µL 930400 dpm/mL Spec-Chec to 20 g driedand ground blank soil. Six replicates of the standard soilwere combusted to determine recovery. Standard capsuleswere used to determine counting efficiency. Parameters wereas follows: recovery: 93.3 ± 0.96%; background radiation:47.3 ± 0.67 CPM; counting efficiency: 80.1 ± 0.08%; LOD:49.3 CPM; LOQ: 54.0.

Modeling

Degradation rates and DTx values were determined bymodeling the degradation data from samples with added la-beled glyphosate, according to the EU guideline on model-ing of degradation kinetics.[27] The modeling was performedusing ModelMaker version 4.0 (A.P. Benson). Three differ-ent models were used for fitting glyphosate: single first order(SFO), first order multi compartment (FOMC) and doublefirst order in parallel (DFOP). For metabolites, only SFOcan be used.[27]

SFO

Differential equation:dMdt

= −k × M

Integrated equation:M = M0 × e−k×t

where M0 is the amount of chemical present at time t =0,M is the amount of chemical present at time t , and k is thedegradation rate constant

FOMC

Differential equation:sdMdt

= −α

βM

(tβ

+ 1)

Downloaded By: [DEFF] At: 16:23 25 August 2010

Glyphosate and AMPA in agricultural soil 369

Integrated equation: M = M0( tβ

+ 1)α

where M0 is the amount of chemical present at time t =0,M is the amount of chemical present at time t , α is a shapeparameter, and β is a location parameter.

DFOP

Differential equation:dMdt

= −k1g exp(−k1t) + k2(1 − g) exp(−k2t)g exp(−k1t) + (1 − g) exp(−k2t)

M

Integrated equation:M = M0(ge−k1t + (1 − g)e−k2t)

where M0 is the total amount of chemical present at timet = s0, M is the total amount of chemical present at timet , g is the fraction of M applied to compartment 1 (the Mcompartment is divided into compartment 1 and compart-ment 2 in DFOP[27]) and k1and k2 are the rate constants incompartment 1 and compartment 2, respectively.

As a first step, results for glyphosate alone were fitted us-ing these models. After determination of the best-fit model,results for AMPA were added. To determine which modelbest fitted the data, the models were compared by visual as-sessment of curve fits and residual plots. As a supplementto the visual assessment, χ2 model errors were calculatedand compared.

χ2 =∑ (C − O)2

(err/100 × O)2

where C is the value calculated by ModelMaker, O is theobserved value, O is the mean of all observed values, anderr is the model error-%.

Because err is unknown, χ2 can not be calculated. Insteadthe minimum err is calculated. If χ2

calculated ≤ χ2tabulated at a

95% confidence level, the test is passed; hence the minimumerr at which the test is passed is the one corresponding toχ2

tabulated and is calculated by:

err = 100

√1

χ2tabulated

×∑ (C − O)2

O2

The model that produces the smaller level of error wastaken to better describe the data. Reliability of the estimatedparameters was determined based on a t-test.

Incubation of soil samples for the plant uptake study

Six samples of 260 g dry soil each were placed in hermet-ically sealed 1.5-L glass jars at 14.3◦C one day before ap-plication of glyphosate. One mL of 14C-glyphosate stocksolution and 1 mL of unlabeled glyphosate solution wasadded to the samples, resulting in a concentration of 1.19 µgglyphosate/g dry soil and 33872.67 dpm/g dry soil. The soilwas mixed with a spoon, which was washed over the soilwith 13 mL MilliQ water, added to give the soil a water con-tent of 50 % of the WHC. Two glass vials, one containing

3 mL KOH (1 M), the other containing 3 mL glycerol, wereplaced in each glass jar to absorb evolved CO2 and volatileorganic compounds, respectively.

To keep the water content at 50% of the WHC, the sam-ples were weighed periodically throughout the incubationperiod, and water was added when necessary.

The KOH traps were replaced after 2, 4, 6, 8, 10, 12and 24 hours. Thereafter they were replaced every dayin the first week, every second day of the second week,and then twice a week from week 3 to 10. From week 11onwards the traps were replaced about once a week. Thereplacement of KOH traps enabled replacement of air inthe glass jars. The glycerol traps were only sampled onday 197 when incubation was discontinued. A volume of1 mL was sampled from the 3 mL KOH or glycerol trapsand mixed with 5 mL Ultima Gold. The trapped 14C wasquantified using a Wallac 1409 LSC.

Plant uptake study

After 197 days of incubation, each sample of 260 g soil wastransferred to a plastic pot and placed in a greenhouse.Three pots were seeded with barley, and three with rapeseeds. A further two pots containing blank soil were alsoseeded with barley or rape. Shoots were later removed toallow the development of only a single plant in each pot.During the growing period, plants were irrigated with anappropriate amount of water containing 1‰ N-P-K 14-3-23fertilizer. Forty-one days after sowing, the plants were har-vested and stored at −18◦C. The plants without roots werefreeze-dried, in a Holm and Halby Heto Drywinner 6-85,for 24 hours and placed in liquid nitrogen for a few minutesbefore they were ground in plastic beakers with eight steelballs in a SPEX CertiPrep Model 2000 Geno/Grinder for40 seconds at 12 strokes per minute. Two 0.1 g replicates ofeach plant sample were combusted in a Packard Oxidizermodel 307 for 1 min. Evolved 14CO2 was captured in 5 mLcarbo-sorb E and mixed with 15 mL permaflour E. Ra-dioactivity was measured using a Wallac 1409 Liquid Scin-tillation Counter. LOD was 49.3 CPM and LOQ was 54.0CPM. Standard plant samples were prepared by adding25 µL Spec-Chec, diluted in methanol, corresponding to930.4 dpm to freeze-dried and ground plant samples of 0.1g. Six standard plant samples were combusted to determinethe recovery. Standard capsules were used to determine thecounting efficiency. Recovery was 90.2 ± 1.8%, backgroundradiation was 49.4 ± 0.85 CPM, counting efficiency was79.2 ± 0.2 CPM, LOD was 51.9 CPM, and LOQ was 57.9CPM.

Results

Recovery

We used three different concentrations in the recovery ex-periment to validate the extraction method used in thedegradation study. The recovery samples had been diluted

Downloaded By: [DEFF] At: 16:23 25 August 2010

370 Simonsen et al.

100×, 10×, and 0×, respectively, before measurement, andresulted in three levels of recoveries for glyphosate andAMPA, respectively: 40.3 ± 9.5% and 46.7 ± 11.3% at1000 ng/g dry soil; 81.8 ± 24.4% and 73.9 ± 5.1% at 100ng/g dry soil; 108.9±5.3 and 131.3±3.0 at 10 ng/g dry soil.The concentrations determined in the degradation studywere corrected for the recoveries corresponding to the de-gree of dilution of the samples before measurement (i.e.samples diluted 100x were corrected for the recovery foundfor the 1000 ng/g dry soil recovery samples, and so on).

Degradation in soil samples with added labeled glyphosate

The quantity of glyphosate in soil samples declinedthroughout the study period (Fig. 1). An initial high rate ofglyphosate degradation slowed down after 3 weeks. Untilday 77, the formation of AMPA was faster than its degra-dation, causing AMPA to accumulate in the soil. After day77, the formation became slower, and AMPA levels beganto decline.

SFO did not fit the glyphosate data satisfactorily, result-ing in a χ2 model error twice as large as that for FOMCand DFOP. The fit with DFOP resulted in an unreliableparameter. Based on visual assessment, χ2 statistics andt-test of parameters, the best-fit model for glyphosate wasfound to be FOMC. It was not possible to fit the forma-tion and dissipation of AMPA with SFO kinetics; henceglyphosate and AMPA were not fitted together. In caseslike this, the degradation rate for the metabolite can, ac-cording to Boesten et al.[27], be estimated separately fromthe decline curve. We used this approach for AMPA usingthe sampling time where the concentration was at maximum(day 77) as the starting point.

0

2

4

6

8

10

12

14

0 50 100 150 200

Time (days)

Con

cent

rati

on (

nmol

/g d

ry s

oil)

Fig. 1. The degradation of glyphosate (•) fitted with first ordermulti compartment (FOMC) kinetics, and the decline of AMPA(�) fitted with single first order (SFO) kinetics. Dots are datapoints, solid lines are kinetic models. The applied concentration ofglyphosate was 6.98 nmol/g dry soil, corresponding to 1.18 µg/gdry soil. The measured mean maximum amount of AMPA of3.96 nmol/g dry soil corresponds to 0.44 ng/g dry soil.

Table 2. Model parameters ± confidence interval and DTx

values for glyphosate fitted with first order multi component(FOMC) and the decline of AMPA fitted with single firstorder (SFO).

Compound Parameter Value

Glyphosate M0 7.8 ± 0.58 nmol/g dry soilα 0.85 ± 0.28β 7.1 ± 4.7DT50 9 daysDT90 101 daysM0 4.2 ± 0.28 nmol/g dry soil

AMPA k 0.022 ± 0.0038 days−1

DT50 32 daysDT90 106 days

Model parameters and DTx values are listed in Table 2.DT50 is about 3.5 times larger for AMPA than glyphosate,while the DT90 values for the two are almost equal; theDT90 is only 5% larger for AMPA than for glyphosate. Itshould, however, be pointed out that the DT90 for AMPAwas determined by extrapolation beyond the time period ofthe experiment.

Degradation of old residues of glyphosate and AMPA inblank soil samples

In the beginning of October 2003, 21 months before ourstudy, 252 g glyphosate/ha had been applied to the fieldfrom which we took soil samples. Decreasing quantitiesof residues from this application (and maybe earlier appli-cations) were detected in the blank soil samples through-out the experimental period (Table 3). On day 0 of thepresent study, 21 months after field application, 0.81 ng/gglyphosate (corresponding to 0.34% of the applied amount)and 10.46 ng/g AMPA (corresponding to 4.36% of appliedglyphosate, using molar conversion) still remained in thesoil. By the last day of the experiment 27 months after fieldapplication, amounts corresponding to 0.08% (glyphosate)

Table 3. Quantities of glyphosate and AMPA in blank soil samplesfound by extraction with borate solution.

Time Glyphosate AMPA(days) (ng/g dry soil) (ng/g dry soil)

0 0.81 10.467 0.61 7.33

14 0.50 5.8621 0.41 5.5935 0.46 5.0349 0.41 4.0477 0.32 3.87

105 0.31 4.45179 0.19 1.16

Downloaded By: [DEFF] At: 16:23 25 August 2010

Glyphosate and AMPA in agricultural soil 371

Table 4. Quantities of glyphosate and AMPA extracted from sam-ples with added labeled glyphosate (sample) and blank soil sam-ples (blank) using either MilliQ water or phosphate solution asextraction agent.

MilliQ water Phosphate solution

(ng/g (% of (ng/g (% ofdry soil) applied)a dry soil) applied)a

Glyphosate (sample)b 0.19 0.02 0.55 0.05AMPA (sample)b 0.13 0.02 0.38 0.05Glyphosate (blank) 0.01 <0.01 0.02 0.01AMPA (blank) 0.05 0.02 0.24 0.10

aBased on applied glyphosate (applied in this study and at the field in2003 for the samples and blank, respectively).bTotal quantities found in samples were calculated based on mea-sured amounts of labeled compounds, hence amounts originating fromglyphosate present in the soil prior to this study, were not included.

and 0.48% (AMPA) of applied glyphosate were found inthe soil.

Extractability of aged added residues

Significantly larger amounts (p < 0.05) of glyphosate andAMPA were extracted from the samples containing residuesaged for 6 months when phosphate solution was used as ex-traction agent, than when MilliQ water was used (Table 4).

Extractability of aged residues in blank soil

From the blank samples containing residues aged for 27months, significantly larger amounts (p < 0.05) of AMPAwere extracted with phosphate solution than with MilliQwater. The difference between amounts extracted usingMilliQ water and phosphate solution was not significant(p>0.05) for glyphosate in blank samples. However, themean extracted amount was twice as high when phosphate

solution was used as an extraction agent as when MilliQwater was used (Table 4). Using a higher number of repli-cates in future studies may reveal a significant difference.The largest amount of aged residues extracted, calculatedas percentage of added mol of glyphosate, was found forAMPA extracted from the blank soil samples with phos-phate solution.

Mass balance

Throughout the degradation and extractability studies,83.89–97.40% of applied radioactivity was recovered (Fig.2). The evolvement of CO2 was fastest during the firstthree weeks, after which it slowed down. After 179 days(6 months), 30.8% of applied radioactivity was evolvedas CO2. No radioactive volatile hydrophobic compoundswere detected in the glycerol vial. Even though the glyceroltraps were incubated with the soil samples for 179 days,very small amounts of radioactivity (<0.001% of applied)were detected in the glycerol. This was also the case for theglycerol incubated with soil for 197 days prior to the plantuptake study.

During the first two weeks, the amount of non-extractable labeled residues in soil (extracted with boratebuffer in the degradation study) increased to about 18–20%,whereupon it remained fairly constant. Extraction with bo-rate buffer left significantly smaller amounts (p < 0.05) ofnon-extractable residues in the soil, compared with extrac-tion with phosphate solution or MilliQ water. Significantlylarger amounts (p < 0.05) of non-extractable residues re-mained in the soil after extraction with MilliQ water thanwith phosphate solution.

Plant uptake

These data reflect the radioactivity that translocated to theparts of the plant that are above the soil. After a growth

Fig. 2. Radioactivity recovered in extracts ( ), evolved as 14CO2 ( ) and detected in soil as non-extractable residues ( ) throughoutthe degradation and leaching studies. The last two columns represent the samples extracted with MilliQ water (H2O) and phosphatesolution (P). All other samples were extracted with the borate solution.

Downloaded By: [DEFF] At: 16:23 25 August 2010

372 Simonsen et al.

Table 5. Uptake of aged residues by rape and barley.

Plantsample

Radioactivitymeasured

(% of applied/plant)

Equivalent toglyphosate(ng/plant)

Equivalent toglyphosate

(ng/g plant freshweight)

Rape 1 0.009 26.21 15.49Rape 2 0.004 13.15 11.10Rape 3 0.005 15.80 15.39Barley 1 0.005 16.34 12.63Barley 2 0.004 12.35 10.72Barley 3 0.006 19.79 11.24

period of 41 days, the rape and barley had taken up 0.006 ±0.002% and 0.005 ± 0.001% (mean values), respectively,of the radioactivity applied to the soil 6.5 months earlier(Table 5). This is equivalent to 14.0 ± 2.5 and 11.5 ± 1.0ng glyphosate/g plant fresh weight, for rape and barleyrespectively. The uptake by rape and barley did not differsignificantly (p > 0.05).

Discussion

Recovery

The lowest mean recoveries for glyphosate (40.3%) andAMPA (46.7%) were found at the 1000 ng/g-concentrationlevel. Higher mean recoveries were found at 100 ng/g(81.8% for glyphosate and 73.9% for AMPA) and at 10ng/g (108.9% for glyphosate and 131% for AMPA).

These recovery levels are acceptable compared with re-sults obtained in an intercalibration study[32] that was con-ducted specifically to assess the capabilities of analyticallaboratories in Europe to determine residues of glyphosateand AMPA in surface and groundwater. It is, however,worth reflecting on possible reasons for the low recov-ery from samples containing the largest glyphosate andAMPA concentrations, compared with higher recoveriesfrom samples containing lower concentrations. The mostlikely explanation is that one of the steps in the extractionand separation procedures has a limitation with respect tothe amount of glyphosate and AMPA it can extract. Fu-ture use of the method for soil samples containing largeamounts of glyphosate and AMPA should be initiated witha detailed examination of the method in order to detectand eliminate the limiting step or steps. It is however alsopossible that the lower recovery from samples containingthe largest glyphosate and AMPA concentration could bedue to the fact that the chemicals in higher concentrationson aging may have a better chance for getting irreversiblyabsorbed due to the intraparticle diffusion as well as gettingentrapped in the molecular voids of the organic matter.

Degradation in soil samples with added labeled glyphosate

Glyphosate degradation occurred initially in the readily ac-cessible regions of the soil, followed by degradation in the

least accessible portions of the soil. The degradation couldnot be described by SFO kinetics because the pattern was bi-phasic, with a fast initial decrease in pesticide concentrationfollowed by a slower decline. In the FOMC model, differentdegradation rates in individual regions of a soil sample areaccounted for by dividing the soil into a large number ofsub-compartments, each with a different first-order degra-dation rate constant.[27] The DT50 of 9 days determinedfor glyphosate is within the interval of 4-180 days reportedfor laboratory studies by the European Commission,[33] andcorresponds to the low end of the half-lives found under nat-ural conditions by Grunewald et al.[9] ranging from 11 to 17days. It is, however, significantly smaller than the half-livesof 19.2 days (sandy loam soil) and 309.7 days (muck soil)reported by Cheah et al.[8] The long half-life found in themuck soil by Cheah et al.[8] was probably due to strong sorp-tion. The DT90 value of 101 days determined for glyphosateis within the interval of 40-280 days reported for labora-tory studies by the European Commission.[33] The DT50values found for glyphosate and AMPA (Table 2) indicatethat AMPA is more persistent than glyphosate, support-ing the findings of Grunewald et al.[9] who also reportedthat the biological stability of AMPA is higher than thatof glyphosate. The DT90 values are almost similar for thetwo, indicating that the rate of degradation of glyphosateslows down so that the residence time of aged glyphosate be-comes similar to that of AMPA. Again it should be pointedout that our determination of DT50 and DT90 values forAMPA was based on the latter part of the curve, and thatDT90 was found by extrapolation beyond the time periodof the experiment. Extended study duration might have re-vealed bi-phasic degradation kinetics for AMPA and thusa prolonged DT90 value.

Degradation of old residues of glyphosate and AMPA inblank soil samples

Glyphosate had been sprayed onto blank soil 21 monthsbefore this study, at a rate of 252 g/ha (corresponding to0.24 µg/g soil under the assumption that glyphosate spreadin the upper 7 cm of the field soil). We measured 0.81 ngglyphosate/g dry soil and 10.46 ng AMPA/g dry soil at thestart of the study. Glyphosate and AMPA were still detectedin the blank soil samples at the end of this study, 27 monthsafter application to the field from which the soil was sam-pled. Even though the compounds had aged for so long inthe soil, they were still dissipating (Table 3) and were there-fore still bioavailable. The amounts of both glyphosate andAMPA found in the blank soil samples were above the cal-culated LOD (0.04 and 0.02 ng/g dry soil for glyphosate andAMPA respectively) throughout the experiment (Table 3).

These findings contrast with those of Grunewald et al.,[9]

who reported complete disappearance of glyphosate andAMPA 5 months after application (LOD 2 ng/g). The re-sults also contrast with results of a recent study by Landryet al.,[14] who found no glyphosate (LOD 0.13 ng/g) in

Downloaded By: [DEFF] At: 16:23 25 August 2010

Glyphosate and AMPA in agricultural soil 373

soil 1 year after the application of an amount (5.7 kgglyphosate/ha) several times higher than the amount ap-plied to the soil used in this study. The total amount ofAMPA found by Landry et al.[14] in the soil profiles, rel-ative to the amount of glyphosate originally applied, was0.007% to 0.011%. In this study, 0.48% AMPA was found inthe soil 27 months after application, relative to the amountof glyphosate applied.

Extractability

A borate solution was chosen as an extraction agent withthe purpose of maximizing the recovery of the analytes. Theamount of glyphosate and AMPA extracted with water andphosphate solution (which was meant to simulate leachingwith rain water, with and without presence of fertilizer)was significantly smaller (p < 0.05) than the amounts ex-tracted with borate solution. Thus the borate extraction wasnot suitable for determining the amount of glyphosate andAMPA available under natural conditions. Overestimationof the availability of pesticides in soil under natural condi-tions by vigorous extraction has also been reported by otherauthors.[4,34] Amounts of glyphosate and AMPA extractedwith phosphate solution were significantly (p < 0.05) largerthan the amounts extracted with pure MilliQ water – thusfertilization of soil with phosphate increases the risk ofleaching of the two compounds, probably due to compe-tition with phosphate for binding sites. It has been postu-lated that glyphosate binds to soil constituents through itsphosphonate group,[17] and therefore competes with phos-phate for binding sites.[19,30,35−36] In a recent study, Barjaand Afonso[37] found that when adsorbing to goethite, bothglyphosate and AMPA use the phosphonate group. This re-sult indicates that phosphate can also affect adsorption ofAMPA in soil, and that is confirmed by the present study.Our study also shows that this competition for binding sitesis of importance for the leaching of glyphosate and AMPA.This is of great relevance concerning normal agriculturalpractice, where soil is treated with both glyphosate and fer-tilizer. Ninety percent of the glyphosate sold in Denmark in2005 was used in agriculture;[38] hence the majority of soil inDenmark treated with glyphosate is also treated with phos-phate. Traditional leaching studies in which pure water ora CaCl2 solution were used as extraction agents have mostlikely underestimated the leaching potential of glyphosateand AMPA from fertilized field soil.

Both glyphosate and AMPA were found after extractionof the blank soil samples with MilliQ water and phosphatesolution (Table 4). The amount of AMPA extracted withphosphate solution was significantly (p < 0.05) larger thanthe amount extracted with pure MilliQ water from blanksamples. Even though the difference was not significant forglyphosate, the tendency was the same (Table 4). Hence,residues of glyphosate and AMPA that have aged in soil formore than two years are still affected by phosphate, and therisk of leaching increases when the soil is fertilized.

Non-extractable residues

Significantly (p < 0.05) larger amounts of non-extractableresidues remained in the soil following extraction with pureMilliQ water, compared with extraction with phosphatesolution, confirming that there is more leaching from soiltreated with water containing phosphate than with pure wa-ter. When a borate solution was used as extraction agent, theamount of non-extractable residues left in the soil increasedover time, until it reached 18–20% of applied radioactiv-ity. Thus it appears that the adsorption of glyphosate andits degradation products to soil, or its sequestration intominute voids, increase over time up to a certain level.

Plant uptake

Even though the applied glyphosate had aged for 6.5months in the soil before seeds were planted, radioactivitytranslocated into plants to the parts that were above the soil.

The possible effects of residual glyphosate on crops de-pend on the concentrations adsorbed. If we assume that theonly radioactive compound taken up by rape and barley inthis study was 14C-glyphosate (which is not likely, but usedas a worst-case scenario), the measured radioactivity wouldcorrespond to 14.0 ± 2.5 and 11.5 ± 1.0 ng glyphosate/gplant fresh weight for rape and barley, respectively. Theseconcentrations are small compared to concentrations foundto reduce plant fresh weight by 50% (ED50) in studies byAlister et al.[23] and Feng et al.[39]

In the apex of maize seedlings, ED50 was 260 ngglyphosate/g when the herbicide was applied to the root.For the whole plant, ED50 for glyphosate applied to theroots was 750 ng glyphosate/g.[23] Following foliar appli-cation of glyphosate, Feng et al.[39] determined the ED50 inroots and meristems of velvetleaf plants to be 230 and 210ng/g fresh weight, respectively, whereas in the basestem itwas 8400 ng/g fresh weight. They also found that the leastsensitive tissue determined the lethal dose for velvetleaf.

An average concentration of 110 ng glyphosate/g hasbeen calculated, in cereals, to result in pesticide intake byhumans corresponding to 0.04% of the Acceptable Daily In-take (ADI) established by the Food and Agriculture Orga-nization and World Health Organization (FAO/WHO).[40]

Therefore, the uptake of aged residues of glyphosate incrops measured in the present study must result in an intakeof <0.04% of ADI.

Our findings indicate that aged residues of glyphosateand AMPA originating from glyphosate applied duringnormal spraying practice do not seem to pose a risk for theyield of crops, and result in a consumer intake of glyphosateless than the ADI. The growth period (41 days) in this studywas short compared with the growth period used in agri-culture, and the longer the growth period, the more willprobably be taken up. But it is not likely that the uptakewill increase to an extent that will significantly influenceyield or intake by consumers.

Downloaded By: [DEFF] At: 16:23 25 August 2010

374 Simonsen et al.

Conclusion

The DT50 values of 9 and 32 days, determined forglyphosate and AMPA respectively, indicate fast degrada-tion in our laboratory study. Old residues of glyphosateand AMPA remained in blank soil more than two years af-ter application of 252 g glyphosate/ha. The degradation ofthese aged residues of glyphosate and AMPA during incu-bation in the laboratory was slower than the degradationof the newly added compounds. It is, however, remarkablethat 21-month-old glyphosate and AMPA residues are de-graded to the extent measured when the soil was broughtinto the laboratory.

The probability that aged residues of glyphosate andAMPA will leach from soil is enhanced when the soil is fer-tilized with phosphate. This is of great relevance because themajor use of glyphosate in Denmark, for example, is agricul-tural, and thus the majority of soil treated with glyphosateis also treated with phosphate. This also means that resultsobtained in traditional leaching studies, where pure wateror CaCl2 solution has been used, could underestimate theleaching potential of glyphosate and AMPA from fertilizedfield soil.

The non-extractable part of glyphosate and its degrada-tion products in soil increased over time up to a certainlevel, indicating that there are a limited number of sorp-tion sites and minute voids that promote the creation ofnon-extractable residues.

After 6 months, residues of glyphosate and AMPA werestill available for uptake by plants. However, the concentra-tion of residues in plant material found in this study doesnot seem to pose a risk for the yield of crops, and results ina consumer glyphosate intake lower than ADI.

Acknowledgments

The authors are grateful to senior scientist Solvejg K. Math-iassen for providing the soil and for her valuable advice re-garding the plant uptake study. We also thank Kirsten Hein-richson, Anne G. Mortensen, Bente Laursen and KirstenJensen for excellent technical support.

References

[1] Sharer, M.; Park, J.H.; Voice, T.C.; Boyd, S.A. Organic compoundsin the environment, aging effects on the sorption-desorption char-acteristics of anthropogenic organic compounds in soil. J. Environ.Qual. 2003, 32, 1385–1392.

[2] Kelsey, J.W.; Alexander, M. Declining bioavailability and inappro-priate estimation of risk of persistent compounds. Environ. Toxicol.Chem. 1996, 16, 582–585.

[3] Gevao, B.; Semple, K.T.; Jones, K.C. Bound pesticide residues insoils: a review. Environ. Pollut. 2000, 108, 3–14.

[4] Alexander, M. Aging, Bioavailability, and overestimation of riskfrom environmental pollutants. Environ. Sci. Technol. 2000, 34,4259–4265.

[5] Barraclough, D.; Kearney, T.; Croxford, A. Bound residues: envi-ronmental solution or future problem? Environ. Pollut.2005, 133,85–90.

[6] Heinonen-Tanski, H. The effect of temperature and liming onthe degradation of glyphosate in two arctic forest soils. Soil Biol.Biochem. 1989, 21, 313–317.

[7] Smith, A.E.; Aubin, A.J. Degradation of 14C-glyphosate insaskatchewan soils. Bull. Environ. Contam. Toxicol. 1993, 50, 499–505.

[8] Cheah, U.B.; Kirkwood, R.C.; Lum, K.Y. Degradation of four com-monly used pesticides in Malaysian agricultural soils. J. Agric. FoodChem. 1998, 46, 1217–1223.

[9] Grunewald, K.; Schmidt, W.; Unger, C.; Hanschmann, G. Behaviorof glyphosate and aminomethylphosphonic acid (AMPA) in soilsand water of reservoir Radeburg II catchment (Saxony/Germany).J. Plant Nutr. Soil Sci. 2001, 164, 65–70.

[10] Mamy, L.; Barriuso, E. Glyphosate adsorption in soils comparedto herbicides replaced with the introduction of glyphosate resistantcrops. Chemosphere 2005, 61, 844–855.

[11] Roy, D.N.; Konar, S.K.; Banerjee, S.; Charles, D.A.; Thomp-son, D.G.; Prasad, R. Persistence, movement, and degradation ofglyphosate in selected Canadian boreal forest soils. J. Agric. FoodChem. 1989, 37, 437–440.

[12] Veiga, F.; Zapata, J.M.; Marcos, M.L.F.; Alvarez, E. Dynamics ofglyphosate and aminomethylphosphonic acid in forest soil in Gali-cia, north-west Spain. Sci. Total Environ. 2001, 271, 135–144.

[13] Kjær, J.; Olsen, P.; Ullum, M.; Grant, R. Leaching of glyphosateand amino-methylphosphonic acid from Danish agricultural fieldsites. J. Environ. Qual. 2005, 34, 608–620.

[14] Landry, D.; Dousset, S.; Fournier, J.C.; Andreux, F. Leaching ofglyphosate and AMPA under two soil management practices inBurgundy vineyards (Vosne-Romanee, 21-France). Environ. Pollut.2005, 138, 191–200.

[15] Geological Survey of Denmark and Greenland (GEUS). Ground-water survey 1998–2003, 2004-report. GEUS: Copenhagen, Den-mark. http://www.geus.dk/publications/grundvandsovervaagn-ing/ (Accessed July 5, 2007).

[16] Fomsgaard, I.S.; Spliid, N.H.; Felding, G. Leaching of pesticidesthrough normal-tillage soil—a lysimeter study. II. Glyphosate. J.Environ. Sci. Health, Part B, 2003, 38, 19–35.

[17] Sprankle, P.; Meggit, W.F.; Penner, D. Adsorption, mobility, andmicrobial degradation of glyphosate in the soil. Weed Sci. 1975b,23, 229–234.

[18] Sprankle, P.; Meggit, W.F.; Penner, D. Rapid inactivation ofglyphosate in the soil. Weed Sci. 1975a, 23, 224–228.

[19] Gimsing, A.L.; Borggaard, O.K.; Bang, M. Influence of soil com-position on adsorption of glyphosate and phosphate by contrastingDanish surface soils. Eur. J. Soil Sci. 2004, 55, 183–191.

[20] Dion, H.M.; Harsh, J.B.; Hill, Jr, H.H. Competitive sorption be-tween glyphosate and inorganic phosphate on clay minerals and loworganic matter soils. J. Radioanal. Nucl. Chem. 2001, 249, 385–390.

[21] Cornish, P.S. Glyphosate residues in a sandy soil affect tomato trans-plants. Austral. J. Exp. Agric. 1992, 32, 395–399.

[22] Pline, W.A.; Wilcut, J.W.; Edmisten, K.L.; Wells, R. Physiologi-cal and morphological response of glyphosate-resistant and non-glyphosate resistant cotton seedlings to root adsorbed glyphosate.Pestic. Biochem. Physiol. 2002, 73, 48–58.

[23] Alister, C.; Kogan, M.; Pino, I. Differential phytotoxicity ofglyphosate in maize seedlings following applications to roots orshoot. Weed Res. 2005, 45, 27–32.

[24] Cornish, P.S.; Burgin, S. Residual effects of glyphosate herbicide inecological restoration. Restor. Ecol. 2005, 13, 695–702.

[25] Dyson, J.S.; Chapman, P.F.; Snell, R.J. Pesticide metabolites in soil:estimation of formation and degradation rates. In Human and Envi-ronmental Exposure to xenobiotics, Proceedings of the 11th Sympo-sium Pesticide Chemistry, Cremona, Italy, 11–15 September, 1999;del Re, A.A.M., Brown, C., Capri, E., Errera, G., Evans, S.P.,

Downloaded By: [DEFF] At: 16:23 25 August 2010

Glyphosate and AMPA in agricultural soil 375

Trevisan, M., Eds.; Facolta agraria, Universita Cattolica de SacreCuore: Cremona, Italy, 1999, 165–170.

[26] Beulke, S.; Brown, C.D. Evaluation of methods to derive degrada-tion parameters for regulatory modelling. Biol. Fert. Soils 2001, 33,558–564.

[27] Boesten, J.J.T.I.; Aden, K.; Beigel, C.; Beulke, S.; Dust, M.; Dyson,J.S.; Fomsgaard, I.S.; Jones, R.L.; Karlsson, S.; van der Linden,A.M.A.; Richter, O.; Magrans, O.; Soulas, G. Guidance Documenton Estimating Persistence and Degradation Kinetics from Environ-mental Fate Studies on Pesticides in EU Registration. Report ofthe FOCUS Work Group on Degradation Kinetics, EC DocumentReference Sanco/10058/2005 version 2.0. 2006. http://viso.jrc.it/focus/dk/docs/finalreportFOCDegKin04June06linked.pdf(Accessed May 31, 2008).

[28] Nomura, N.S.; Hilton, H.W. The adsorption and degradation ofglyphosate in five Hawaiian sugarcane soils. Weed Res. 1977, 17,113–121.

[29] Morillo, E.; Undabeytia, T.; Maqueda, C.; Ramos, A. Glyphosateadsorption on soils of different characteristics. Influence of copperaddition. Chemosphere 2000, 40, 103–107.

[30] de Jonge, H.; de Jonge, L.W.; Jacobsen, O.H.; Yamaguchi, T.;Moldrup, P. Glyphosate Sorption in Soils of Different pH and Phos-phorus Content. Soil Sci. 2001, 166, 230–238.

[31] Gerlach, R.W.; Dobb, D.E.; Raab, G.A.; Nocerino, J.M. Gy Sam-pling theory in environmental studies. 1. Assessing soil splitting pro-tocols. J. Chemom. 2002, 16, 321–328.

[32] Reichert, N. European round robin study for glyphosate residues(glyphosate and its metabolite AMPA) in surface water and ground

water. SGS INSTITUT FRESENIUS GmbH 2005, study no. IF-04/00162935. Taunusstein, Germany.

[33] European Commission. Review report for the active substanceglyphosate. Health and Consumer Protection Directorate-General2002;6511/VI/99-final. European Commission: Brussels, Belgium,2002.

[34] Kelsey, J.W.; Kottler, B.D.; Alexander, M. Selective Chemical ex-tractants to predict bioavailability of soil-aged organic chemicals.Environ. Sci. Technol. 1997, 31, 214–217.

[35] Hance, R.J. Adsorption of glyphosate by soils. Pest. Sci. 1976, 7,363–366.

[36] de Jonge, H.; de Jonge, L.W. Influence of pH and solution composi-tion on the sorption of glyphosate and prochloraz to a sandy loamsoil. Chemosphere 1999, 39, 753–763.

[37] Barja, B.C.; Afonso, M.D.S. Aminomethylphosphonic acid andglyphosate adsorption onto goethite: a comparative study. Environ.Sci. Technol. 2005, 39, 585–592.

[38] Danish Environmental Protection Agency. Bekæmpelsesmiddel-statistik 2005. Orientering fra Miljøstyrelsen nr. 5, Danish En-vironmental Protection Agency: Copenhagen, Denmark, 2006.www.mst.dk (Accessed July 5, 2007).

[39] Feng, P.C.; Chiu, T.; Sammons, R.D. Glyphosate efficacy is con-tributed by its tissue concentration and sensitivity in velvetleaf (Abu-tilon theophrasti). Pestic. Biochem. Physiol. 2003, 77, 83–91.

[40] Granby, K.; Vahl, M. Investigation of the herbicide glyphosate andthe plant growth regulators chlormequat and mepiquat in cere-als produced in Denmark. Food Addit. Contam. 2001, 18, 898–905.

Downloaded By: [DEFF] At: 16:23 25 August 2010