Chemosphere 63 (2006) 853–859

www.elsevier.com/locate/chemosphere

Effects on hawthorn the year after simulated spray drift

Christian Kjær a, Morten Strandberg a,*, Mogens Erlandsen b

a National Environmental Research Institute, Department of Terrestrial Ecology, P.O. Box 314,

Vejlsøvej 25, DK-8600 Silkeborg, Denmarkb University of Aarhus, Department of Biostatistics, Vennelyst Boulevard 6, DK-8000 Arhus C, Denmark

Received 14 March 2005; received in revised form 7 July 2005; accepted 20 July 2005Available online 16 September 2005

Abstract

This paper describes the effect of a herbicide applied at levels consistent with off target movement on hawthorn theyear following exposure. In the experiment, metsulfuron-methyl was applied in five dosages to individual trees in sevendifferent hawthorn hedgerows. Spraying was conducted both at the bud stage and at the early flowering. Five endpoints(i.e. leaves, buds, flowers, green berries and mature berries) were sampled and counted. All were significantly reducedwith increasing exposure. Present day risk assessment of effects on non-target plants is therefore likely to overlooksignificant effects on perennial non-target plants in the spray drift zone due to the focus on results from short-termlaboratory test studies. The significance of the present study is underlined by the fact that the effects observed weresignificant, even though other influential factors such as herbivory and differential pollination were not eliminatedand that experiments were conducted in multiple locations.� 2005 Elsevier Ltd. All rights reserved.

Keywords: Hawthorn; Crataegus monogyna; Delayed effects; Test methods; Non-target plants; Metsulfuron-methyl; Herbicide

1. Introduction

In most West European countries, agriculture is thepredominant form of land use. Agricultural fields areusually treated with pesticides. Due to wind drift, evap-oration and deposition, neighbouring habitats also re-ceive small amounts of pesticides. Pesticide exposure ofhedgerows is therefore a widespread and common phe-nomenon. Hedgerows are important habitat elementsin terms of biodiversity as they consist of trees and herbsof which many are very different from agriculturalplants. The plant species and the more stable perennial

0045-6535/$ - see front matter � 2005 Elsevier Ltd. All rights reservdoi:10.1016/j.chemosphere.2005.07.058

* Corresponding author. Tel.: +45 8920 1759/1461; fax: +458920 1413.

E-mail address: mts@dmu.dk (M. Strandberg).

environment in hedgerows form a habitat for many ani-mal species including arthropods, mammals and birds.Hedgerows may serve both as a refuge for species sensi-tive to soil treatment and as overwintering habitats forspecies reproducing in the field.

Hawthorn hedgerows have a rich arthropod fauna(Kennedy and Southwood, 1984). Woody species inhedgerows are important and often obligatory for theoccurrence of a number of bird species in arable land.Hawthorn provides not only breeding sites but berriesare an important food source to birds during autumnand winter (Snow and Snow, 1988).

At present, the assessment of effects on non-targetplants is based on studies of short-term effects on annualplant species. These standard tests include seedlingemergence, root elongation vegetative vigour and earlyseedling growth tests (OECD 208 (1984); ISO 11269-1

ed.

854 C. Kjær et al. / Chemosphere 63 (2006) 853–859

(1993); ASTM (1994); ISO 11269-2 (1995) and US EPA(1996)). The test species are usually crop species. Nostandardised tests use plant reproduction as a measureof effects, although it is recognised that the fitness of an-nual plant species largely depends on the reproductiveoutput (Harper, 1977). The choice of annual plant spe-cies as test species for risk assessment is primarily basedon the ease of testing and the economical importance ofcrop plants. In contrast, most natural and semi-naturalhabitats are dominated by perennial plant species.Perennial species differ from annual species because theycan directly carry effects from one year to the next,whereas effects on annual plant species are only mani-fested the year after exposure if seedling recruitment isseed limited in the habitat of interest.

The impact of a common sulfonylurea herbicide met-sulfuron-methyl (Ally� (Dupont de Nemours)) on thecommon hedgerow tree species hawthorn was investi-gated. The sulfonylurea herbicides are widely used, how-ever, they have been shown to affect both growth andreproduction of trees as well as annual plant species inthe year of exposure (Al-Khatib et al., 1992; Fletcheret al., 1993; Bhatti et al., 1995; Fletcher et al., 1995;Fletcher et al., 1996). The acute effect of the herbicidemetsulfuron-methyl on hawthorn within the same grow-ing season has been reported in (Kjær et al., 2005). Ahighly significant reduction in number of berries wasobserved. The reduction was close to 100% at exposurelevels observed as spray drift under normal sprayconditions.

In this study the trees were revisited the year afterherbicide exposure and the effect of the herbicide onfruit production was measured.

2. Materials and methods

This study consists of two parts. The first part is anin situ assessment of herbicidal effects of metsulfuron-methyl the year after exposure. We measured bothvegetative and reproductive endpoints because fieldassessments with sulfonylurea herbicides often show apoor relationship between foliar injury and yield (Obrig-awitch et al., 1998). The second part assesses the impor-tance of herbivorous insects on the endpoints measured.This is performed because hawthorn is attacked by alarge number of herbivorous insects, some of whichattack flowers and berries causing loss of these repro-ductive structures.

2.1. Herbicide trials

In a series of spray experiments conducted in haw-thorn hedgerows on seven different locations, the effectof the sulfonylurea herbicide metsulfuron-methyl wasmeasured the year after exposure.

The study was designed as a block-randomised sprayexperiment with four blocks in each of seven hedgerows.In each block metsulfuron was applied in five dosages torandomly chosen trees, and each dosage was applied tofour trees chosen at random in one block and three treeschosen at random in three blocks. The design was appliedtwice, i.e. May 13–22 and June 19–21 2002 in the samehedgerows but in separate blocks. These periods coincidewith flowers bud stage and late flowering stage. In eachhedgerow, metsulfuron was applied in five dosages toindividual trees. Five endpoints (number of leaves, buds,flowers, green berries and mature berries per side shoot)were sampled the year after herbicide exposure. Thenumber to be sampled of each type of endpoint wasbased on a statistical power analysis from a pilot study(Kjær et al., 2002). For each dosage and spraying timethe following samples were taken:

• 4 samples of leaves from four different trees in each ofseven hedgerows,

• 10 samples of flowers from 10 different trees in eachof seven hedgerows,

• 13 samples of green berries from 13 different trees ineach of seven hedgerows,

• 13 samples of red berries from 13 different trees ineach of seven hedgerows.

This adds up to a total number of samples of (4 leaves ·7 hedgerows · 2 Spraying times · 5 doses) 280 for leavesand similarly 700 for flowers, 910 for green berriesand also 910 for mature berries. For green and matureberries this means that one sample of each type wassampled from each of 910 trees. In each hedgerow,the spraying and sampling were separated into four sec-tions (blocks) with 4, 3, 3, and 3 replicates per section,respectively. At the time for sampling mature berries,many side shoots had lost all berries. It was thereforedecided to increase sample size by counting the num-ber of mature berries within an area of 0.1225 cm�2

(35 cm · 35 cm) with the randomly chosen side shootas centre.

2.2. Description of hedgerows use in the experiments

All hedgerows consisted entirely of hawthorn trees.Hedgerows numbered from 1 to 4 were all part of anintensively cultured farmland area whereas the hedge-row 5–7 were located adjacent to extensively used fieldswithout herbicide use the years before the experiment.

2.3. Spray procedure

The hedgerows were sprayed with an experimentalazo-sprayer equipped with Hardy flat fan nozzles no4110-16. The sprayer was used with a working pressure

Table 1Regression analyses of the spray deposition (lg cm�2) inrelation to the targeted dosage (% of maximum label rate)

Spray time Hedgerownumber

b ± SE (· 10�4) N r2

Spring 1 3.162 ± 0.332 65 0.5862 0.976 ± 0.096 65 0.6133 0.322 ± 0.042 65 0.4774 1.245 ± 0.211 65 0.3515 1.620 ± 0.120 65 0.7146 1.113 ± 0.242 65 0.2487 0.567 ± 0.103 65 0.320

Summer 1 1.619 ± 0.158 65 0.5712 2.033 ± 0.188 65 0.6033 1.999 ± 0.217 65 0.5224 0.468 ± 0.038 65 0.6445 1.116 ± 0.120 65 0.5296 1.323 ± 0.142 65 0.5297 0.895 ± 0.107 65 0.474

Data were fitted to the equation Y = b · dose, where Y is thespray deposition. b is the slope of the regression line. Nointercept was included in the analyses. The dosage is % ofmaximum label rate. N is the number of observations.

C. Kjær et al. / Chemosphere 63 (2006) 853–859 855

of 2 bar and administered so that an application rate of200 l ha�1 was obtained by adjusting the walking speed.The sprayer was held vertical in order to spray a beltcovering the hawthorn with spray solution from theground and up to approximately 2.5 m above ground.The hedgerows were treated with five dosages includinga no-spray control. The dosages were 0.05, 0.1, 0.2 and0.4 times the maximum label rate for spring barley (4 ga.i. metsulfuron ha�1). These dosages were selectedbased on a pilot trial (Kjær et al., 2002) in order to ob-tain effect levels between 0% and 100%. Actual deposi-tion was measured for each tree by means of a glycinetracer. Glycine was added to the spray solution in a con-centration of 10 g l�1. Immediately after spraying, threeleaves, placed next to the shoots marked for sampling,were collected and subsequently placed in demineralisedwater. Back in the laboratory the leaf area and the con-tent of glycine in the water phase was measured accord-ing to the method of Babcock et al. (1990). This way,deposition could be related to leaf area for each sprayedtree.

2.4. The importance of insects on reproduction

The importance of herbivory for the fruit set wasstudied in a single hedgerow. In a part of hedgerownumber five not treated with metsulfuron, 15 trees weretreated with the insecticide alpha-cypermethrin at inter-vals of approximately 14 days. Another 15 trees werenot treated with anything. Through the season the num-ber of buds, flowers and berries were counted 5 times onrandomly chosen shoots on treated and unsprayed trees.The responses of treated and unsprayed trees were com-pared. The experiment was conducted with 5 replicatesper treatment and each replicate consisted of one shootfrom each of three trees. Treated and control blockswere alternating along the hedgerow separated byapproximately 5 m wide buffer zones. All trees werecompletely separated from the herbicide treatment.The trees were sprayed with the recommended field rateof alpha-cypermethrin (0.2 l ha�1). The first sprayingwas conducted in early May and the last when green ber-ries appeared. Before the first spraying inflorescenceswithout visible insect damage were randomly selectedfor later sampling. The number of intact reproductiveunits per inflorescence was subsequently registered regu-larly throughout the season.

2.5. Statistical analyses

For the herbicidal trial, the number of sampled buds,flowers and leaves, respectively, could be described bynormal distributions and were analysed by means ofANOVA with random effects. �Spray dosage� (lg metsul-furon cm�2) was considered a fixed effect (continuousvariable), whereas �hedgerow� and �block� both was

analysed as random effects (variance components). The�block� factor was nested to �hedgerow�. The two variancecomponents are presented by the proportion that can beattributed to that component out of the total variance.

The number of sampled green and mature berries perside shoot could not be described by standard paramet-ric statistical distributions, because numbers were toolow to make a reasonable statistical analysis. Instead,the number of berries counted within the 35 · 35 cm2

frame was tested by means of Jonckheere–Terpstra testfor ordered alternatives (Siegel and Castellan, 1988).

In the insect exclusion experiment the differences be-tween number of reproductive units for untreated (con-trol) and insecticidally treated trees were analysed bymeans of a MANOVA (Multivariate analysis of vari-ance) with 15 independent samples (one for each markedshoot) each sampled 5 times over the season. In order toassess the effect of treatment for each sampling time, aBonferoni correction was applied.

3. Results

3.1. Herbicidal effect assessment

In all the following analyses, the effects were relatedto the measured spray deposition in each hedgerow. Ac-tual exposure was established from a regression analysisbetween the nominal dosage and the spray depositionmeasured in each hedgerow. The actual exposure rangedbetween 0.00016 and 0.01265 lg metsulfuron-methylcm�2. The analyses are presented in Table 1.

Table 2Relationship between number of buds, flowers and leaves sampled and the dosage of metsulfuron (lg metsulfuron cm�2) for a springand summer exposure, respectively

Endpoint Spray time ab b p DF Hedgerow, % Block (hedgerow), % Single trees, %

Buds Spring 6.62 �480.67 <0.0001 451 19.2 0 80.7Summer 6.78 �507.05 <0.0001 452 22.0 0 78.0

Flowers Spring 5.62 �440.08 <0.0001 432 17.8 1.5 80.7Summer 5.83 �297.51 0.0011 433 18.4 1.9 79.7

Leaves Spring 6.51 �300.94 <0.0001 411 8.9 3.3 87.8Summer 6.57 �374.24 <0.0001 431 7.3 3.4 89.2

The last three columns give the percentage of the total variation accounted for by each of the random effect factors. The linearrelationship between exposure and effect is described by Y = b · depos + ab, where Y = endpoint, and depos = the measured depo-sition (lg metsulfuron cm�2). DF is the total degrees of freedom.

Table 3Analysis of the effect of metsulfuron exposure on the number ofmature berries sampled per 0.1225 m2

Spraytime

Hedgerownumber

N Z p (one-sided)

Spring 1 65 0.1833 0.4272 65 �1.0168 0.1553 65 �1.6356 0.0514 65 �2.7190 0.0035 65 �1.7449 0.0416 65 �1.4430 0.0757 65 �1.8471 0.032

Summer 1 65 �0.1326 0.4472 65 �1.3880 0.0833 65 �2.0147 0.0224 65 �1.7788 0.0385 65 �2.5669 0.0056 65 �0.3151 0.3767 65 �2.7159 0.003

The analysis were done by means of a Jonckheere–Terpstra testfor ordered alternatives. N is the number of observations and Z

is the test value of the Jonckheere–Terpstra test.

856 C. Kjær et al. / Chemosphere 63 (2006) 853–859

It was found that the number of leaves, buds, andflowers per side shoot were negatively correlated toherbicide exposure (Table 2). The relationship was notsignificantly (p = 0.80, a = 0.05, ANOVA) different be-tween spring and summer treatment. The separation ofthe hedgerow into blocks along the hedgerow did notcontribute much to the overall variation (0–4.7%). Thevariation attributed to general differences betweenhedgerows varied between 2.6% and 26.8% for differentmeasurement endpoints. The remaining variation, i.e.variation between trees, accounted for the major partof the overall variation namely 73.2–97.2%. The sampleddata for mature berries were analysed by means of thenon-parametric test for ordered alternatives (Jonckhe-ere–Terpstra test). In order to accept the alternativehypothesis (h1 P h2 P h3 P h4 P h5, where hi is thespray dosage) the value of the response variable for atleast one spray dosage should be significantly larger thanthe next dosage. A significant reduction in number ofmature berries with increasing metsulfuron exposurewas observed (p < 0.0001), when tested across all sevenhedgerows. As this test does not account for variabilitybetween hedgerows, the analyses were made also for eachhedgerow separately. Seven of the fourteen relationshipsstudied showed significant negative effects at a 6 0.05(Table 3). For three hedgerows, namely 1, 2 and 6, nei-ther spring nor summer treatment evoked a significantresponse with increasing herbicide exposure. Fig. 1shows the data behind these analyses. It is seen that forhedgerow 1 and 2 a low number of berries in the controland low-dose treatments gave rise to the non-significantresult. Hedgerow 6 shows a trend toward reduced num-ber of berries with increasing herbicide dosage for bothspray timings, however, due to a large relative variationcompared to the mean values no dosage had significantlylarger effect than the next dosage. The low reproductiveoutput in the control treatment was not observed earlierin development, i.e. for buds and flowers (Fig. 2). For allthe other hedgerows the productivity in the early devel-opment corresponded to the production in the laterdevelopmental stage (number of berries).

3.2. Insects

There was a significant loss of reproductive entities(buds, flowers and berries) during the growing season.The loss was significantly larger for unprotected (un-sprayed) trees than for insecticide sprayed trees. The dif-ference could be seen from the onset of flowering andthroughout the season (Fig. 3). During the period fromJune 6th to July 16th the number of reproductive unitsper side shoot were from 2 to 145 times higher in thesprayed trees.

Two weevil species (Anthonomus pedicularis L. andAnthonomus sorbi Germ.), attacking the buds were com-monly observed in hedgerow no. 5 in the late May earlyJune (insects was not sampled in the other hedgerows).A second generation of these two species appeared inmid July. These oviposit in the flower buds, in which

0

4

8

12

16Hedgerow 1

0

4

8

12

Hedgerow 3

Spring Summer

0

2

4

6

8

10

12 Hedgerow 4

0

2

4

6

8

10 Hedgerow 5

Num

ber

of m

atur

e be

rrie

s pe

r 0.

1225

m2

0

1

2

3Hedgerow 6

0

5

10

15

20

25

30

Hedgerow 2

0 10-4 10-3 10-2

0

1

2

3

4 Hedgerow 7

Deposition, g metsulfuron cm-2

0 10-4 10-3 10-2

Fig. 1. Effect of metsulfuron spray deposition on the number of mature berries per 0.1225 m2 the year after exposure. Vertical barsrepresent ± one standard error of mean. The shaded area visualise the deposition equal to between 1% and 10% of the maximumrecommended field rate.

C. Kjær et al. / Chemosphere 63 (2006) 853–859 857

the entire larval development takes place. Ultimately,the buds are lost and pupation takes place within thebuds on the soil. Furthermore, the hawthorn leaf beetleLochmea cratagei (Forster) were observed in high num-bers feeding on the flowers in the late May coincidentwith the onset of flower loss.

4. Discussion

Present day risk assessment is likely to overlook sig-nificant effects on perennial non-target plants in thespray drift zone due to the focus on short-term labora-tory testing. We observed and quantified pronouncedeffects on both growth indices and reproductive mea-

sures the year after exposure. Our results are supportedby general experience that sulfonylurea herbicides usedfor brush control do not cause ultimate effects untilthe year after exposure (Hager and Nordby, 2003).

In the present study, the observed effects were signif-icant even though variation due to other factors such asherbivory and differences in pollination were not elimi-nated. The present study is further strengthen by beingconducted at multiple locations.

In two of seven experimental hedgerows, namelyhedgerow 1 and 2, no significant effects of the herbicidetreatment could be demonstrated on the number of sam-pled mature berries. In these trials, very few mature ber-ries were observed in the control and at the low dosagetreatments (5% of recommended maximum field rate). A

02468

101214

Hedgerow 1

SummerSpringHedgerow 1

Buds Flowers

0 10-4 10-3 10-202468

1012 Hedgerow 2

Exposure, g metsulfuron cm-2

Num

ber o

f bud

s an

d flo

wer

s

0 10-4 10-3 10-2

Hedgerow 2

Fig. 2. Effect of metsulfuron spray deposition on the number of buds and flowers in hedgerow number 1 and 2 the year after exposure.The left column represents spraying in the spring and the right column the summer spraying. Vertical bars represent ± one standarderror of mean.

5 May 26 May 16 Jun 7 Jul

0

2

4

6

8

10

12

**

*

flower harvestbud harvest

Ave

rage

num

ber o

f rep

rodu

ctiv

e un

its

per s

ide

shoo

t

Date

Control Insecticide

Fig. 3. Number of reproductive units (buds, flowers andberries) per side shoot over time for insecticidally treated andunsprayed hawthorn. A significant difference between treat-ments on the same date is marked with *. Vertical barsrepresent ± one standard error of mean.

858 C. Kjær et al. / Chemosphere 63 (2006) 853–859

similar low number of reproductive entities was not ob-served when (buds and flowers were counted earlier onthe same branches (Fig. 2). Apparently, the young fruitshave been lost, probably due to insect infestations. Anumber of herbivorous insects attack the reproductivestructures and consequently the plant may loose them.Insects are not evenly distributed within the hawthornhedgerow (unpublished data) and therefore by chanceberries in the control treatments have been heavilyattacked by such organisms. Taking into account thatsulfonylurea herbicides can result in delayed flowering(for example Gealy et al. (1995)), flowers from herbicide

treated trees may have escaped herbivory because theywere not present at the time of insects attack. The insectexclusion experiment (Fig. 3) supports that insects canhave dramatic consequences on the reproductive outputand that these effects are manifested late in development.

For tree species, the production of flowers and leavesin the following year is based on structures formed in theearly autumn the year before. Therefore, the reducedleaf numbers observed in this study, suggest that, inthe case of hawthorn, reproduction will be affected alsotwo years after spraying because the leaves accumulateenergy for reproduction the next season. A precedingstudy on the effects of metsulfuron-methyl on hawthornwithin the spraying season showed that the vegetativestructures were not affected, but reproduction was se-verely affected (Kjær et al., 2005). Even though the num-ber of leaves were unaffected in the year sprayedreproduction was reduced the following year. This maybe due to secondary effects of sulfonylurea herbicides.These herbicides have been found to reduce the transferof assimilation products out of the treated leaves of sus-ceptible plants (Bestman et al., 1990). It means that theresources available for fruit production is reduced, butthe treated leaves may not be affected (Kjær, 1994). Ifthis is the reason, then resources for the next years pro-duction of leaves and reproductive entities are capturedin the treated leaves and lost. Therefore, all endpointswill be affected the next year as seen in this study.

Our test was conducted with an experimental sprayerpointed directly into the hedgerow, which is an artificialsituation. Therefore, the deposition of spray solutionwas measured, so that the effects can be translated tolevel of deposition observed in studies performed under

C. Kjær et al. / Chemosphere 63 (2006) 853–859 859

conditions more similar to normal practice. In thisstudy, severe reduction in the number of mature berriessampled was observed at dosages greater than 1% of themaximum recommended field rate (Fig. 1). Existing datafor spray drift from field spraying into hedgerows indi-cate that drift values for flat fan nozzles range between0.8% and 22% of the applied spray rate, dependent onboom height, working pressure and wind speed (Nordbyand Skuterud, 1975; Longley et al., 1997; Weisser et al.,2002). These data suggest that the effect levels observedin this study are relevant for spray drift. The use of flatfan nozzles may be considered a worst case for theassessment of spray drift in general. If drift reducingspray equipment is used spray deposition will be consid-erable smaller (Weisser et al., 2002; Wolf et al., 2004).

Acknowledgements

This study was supported by the Danish Environ-mental Protection Agency grant MS 7041-0471. Weare grateful to Niels Elmegaard and Hans Løkke forconstructive comments on an earlier version of thepaper. Thanks to Ingelise Lauridsen for co-ordinationof and technical assistance with the field work.

References

Al-Khatib, K., Parker, R., Fuerst, E.P., 1992. Sweet cherry(Prunus avium) response to simulated drift from selectedherbicides. Weed Technology 6, 975–979.

ASTM, Method E1598-94, 1994. Practice for Conducting EarlySeedling Growth Tests.

Babcock, J.M., Brown, J.J., Tanigoshi, L.K., 1990. Volume andcoverage estimation of spray deposition using an aminonitrogen colorimetric reaction. Journal of Economic Ento-mology 83 (4), 1633–1635.

Bestman, H.D., Devine, M.D., Vanden Born, W.H., 1990.Herbicide chlorsulfuron decreases assimilate transport outof treated leaves of field pennycress (Thlaspi arvense L.)seedlings. Plant Physiology 93, 1441–1448.

Bhatti, M.A., Al-Khatib, K., Felsot, A.S., Parker, R., Kadir, S.,1995. Effects of simulated chlorsulfuron drift on fruit yieldand quality of sweet cherries (Prunus avium L.). Environ-mental Toxicology and Chemistry 14 (3), 537–544.

Fletcher, J.S., Pleeger, T.G., Ratsch, H.C., 1993. Potentialenvironmental risks associated with the new sulfonylureaherbicides. Environmental Science and Technology 27 (10),2250–2252.

Fletcher, J.S., Pleeger, T.G., Ratsch, H.C., 1995. Chlorsulfuroninfluence on garden pea reproduction. Physiologia Planta-rum 94, 261–267.

Fletcher, J.S., Pleeger, T.G., Ratsch, C., Hayes, R., 1996.Potential impact of low levels of chlorsulfuron and other

herbicides on growth and yield of nontarget plants. Envi-ronmental Toxicology and Chemistry 15, 1189–1196.

Gealy, D.R., Boerboom, C.M., Ogg, A.G.J., 1995. Growth andyield of pea (Pisum sativum L.) and lentil (Lens culinaris L.)sprayed with low rates of sulfonylurea and phenoxyherbicides. Weed Science 43, 640–647.

Hager, A.G., Nordby, D., 2003. Brush control in Illinois. In:Illinois Agricultural Pest Management Handbook—2004.University of Illinois Extension, College of Agricultural,Consumer and Environmental Sciences, pp. 297–305.

Harper, J.L., 1977. Population Biology of Plants. AcademicPress, London.

ISO 11269-1, 1993. Determination of the Effects of Pollutantson Soil Flora—Method for the Measurement Growth.

ISO 11269-2, 1995. Determination of the Effects of Pollutantson Soil Flora—Effects of Chemicals on the Emergence andGrowth of Higher Plants.

Kennedy, C.E.J., Southwood, T.R.E., 1984. The number ofspecies of insects associated with British trees: A reanalysis.Journal of Animal Ecology 53, 455–478.

Kjær, C., 1994. Sublethal effects of chlorsulfuron on blackbindweed (Polygonum convolvulus L.). Weed Research 34,453–459.

Kjær, C., Strandberg, M., Elmegaard, N., 2002. Effekter afsprøjtemiddelafdrift pa bærbærende buske og træer somindikator for biodiversitetsforandringer. Bekæmpelsesmidd-elforskning 60.

Kjær, C., Strandberg, M.T., Erlandsen, M., 2005. Metsulfuronspray drift reduces fruit yield of hawthorn (Crataegus

monogyna L.). Science of the Total Environment, in press.Longley, M., Cilgi, T., Jepson, P.C., Sotherton, N.W., 1997.

Measurement of pesticide spray drift deposition into fieldboundaries and hedgerows. 1. Summer applications.Environmental Toxicology and Chemistry 16, 165–172.

Nordby, A., Skuterud, R., 1975. The effects of boom height,working pressure and wind speed on spray drift. WeedResearch 14, 385–395.

Obrigawitch, T.T., Cook, G., Wetherington, J., 1998. Assess-ment of effects on non-target plants form sulfonylureaherbicides using field approaches. Pesticide Science 52 (3),199–217.

OECD Guideline for Testing of Chemicals 208, 1984. Terres-trial Plants, Growth Test.

Siegel, S., Castellan, N.J.J., 1988. Nonparametric Statisticsfor the Bahavioural Sciences. McGraw-Hill Internationaleditions.

Snow, B., Snow, D., 1988. Birds and Berries—A Study of anEcological Interaction. T&A D Poyser, Calton.

US EPA, OPPTS Ecological Effect Guideline, 850 Series, 1996;850.4100: Terrestrial Plant Toxicity, Tier I (Seedling Emer-gence); 850.4230: Early Seedling Growth Toxicity Test.

Weisser, P., Landfried, M., Koch, H., 2002. Off-crop driftsediments on plant surfaces—exposure of non-target organ-isms. Aspects of Applied Biology 66, 225–230.

Wolf, T.M., Caldwell, B.C., Pederson, J.L., Caessna, A.J.,2004. Interaction of riparian vegetation and nozzle type fordrift deposit reduction. Aspects of Applied Biology 71, 183–190.