ARTICLE IN PRESS

0261-2194/$ - se

doi:10.1016/j.cr

�CorrespondE-mail addr

Crop Protection 26 (2007) 1303–1314

www.elsevier.com/locate/cropro

Eradication of an invasive alien pest, Thrips palmi

R.J.C. Cannona,�, L. Matthewsa, D.W. Collinsa, E. Agalloua, P.W. Bartletta,K.F.A. Waltersa, A. Macleoda, D.D. Slawsonb, A. Gauntc

aCentral Science Laboratory, Sand Hutton, York, YO41 1LZ, UKbDefra, Plant Health and Seeds Inspectorate, Foss House, Kings Pool, York YO1 7PX, UK

cDefra, PHSI, 10-11 City Business Centre, Basin Road, Chichester, West Sussex, PO19 8DU, UK

Received 24 August 2006; received in revised form 9 November 2006; accepted 9 November 2006

Abstract

Thrips palmi Karny, the melon thrips, is a polyphagous pest that has spread widely in tropical and subtropical regions. It is absent

from Europe, although outbreaks have occurred in the Netherlands (1988–98), the UK (2000–01) and most recently Portugal (2004). An

outbreak of T. palmi occurred in Sussex in 2000. It was already well established when the UK Plant Health Service first notified. High

populations were discovered in two glasshouse sites on the same nursery, growing all-year-round chrysanthemums. An intensive, largely

chemically based eradication programme was carried out, with applications of aerosol ‘space’ treatments, systemic and foliar insecticides.

Other measures included the use of sticky yellow sheets, methyl bromide fumigation of flowerbeds and plastic mulches. Together, these

controls resulted in the collapse of the outbreak, within 7 months. Eradication was subsequently declared after freedom from the pest had

occurred over two complete cropping cycles.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Thrips palmi; Eradication; Insecticides; Control measures

1. Introduction

1.1. Pest status

Thrips palmi Karny (Thysanoptera, Thripidae), themelon thrips, is listed in Annex IAI of the EC PlantHealth Directive (2000/29/EC) and as such requireseradication wherever it is found in the EU. Similarly,introduction of T. palmi into Great Britain is banned underthe Plant Health (England) Order 2005 (Anon, 2005).T. palmi is a vector of alien tospoviruses (Moritz et al.,2004), and has been implicated in the transmission of atleast six plant viruses (Jones, 2005). If these viruses wereintroduced to the UK, other tospovirus vectors that arealready present—such as the western flower thrips,Frankliniella occidentalis (Pergande)—could potentiallyredistribute them. Furthermore, there is the potential for

e front matter r 2006 Elsevier Ltd. All rights reserved.

opro.2006.11.005

ing author. Tel.: +441904 462218; fax: +44 1904 462250.

ess: r.cannon@csl.gov.uk (R.J.C. Cannon).

T. palmi to become an additional vector of other plantviruses already present in the UK.

1.2. Life-cycle and damage symptoms

T. palmi adults are small (42–3mm), yellow-wingedinsects and both sexes are capable of flight. The females arecapable of sexual or asexual reproduction, producing manyeggs that are deposited into plant tissues. Eggs give rise totranslucent first instar ‘larvae’ that are barely visible withthe naked eye; this stage is largely responsible for theuptake of tospoviruses from infected host plants. Togetherwith adults and first instars, the larger, yellow, secondinstar nymphs are mobile and responsible for direct feedingdamage. In chrysanthemums, second instar nymphs arefound in the small curled leaf buds where the fullyexpanded leaves join the main stem. When the secondinstar has completed development, it becomes a propupaand subsequently, a pupa. These relatively sedentary, non-feeding stages are found in loose soil, growing substrates,

ARTICLE IN PRESS

Plate 1. Photograph of T. palmi damage to chrysanthemum leaf.

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–13141304

plant debris, on host plants, or in moist, enclosedenvironments. The winged adults subsequently emergeand begin to feed, and reproduce soon after.

Larvae and adults feed by piercing individual plant cellsand sucking the cell contents. This gives damaged tissues aspeckled appearance (Plate 1: chrysanthemum leaf da-mage). Where the cells coalesce, a shiny scar may develop,which can be sufficient to reduce the marketability andvalue of ornamental crops or fruits. If sufficient damageoccurs at early stages of plant growth, then distortion,scarring and malformation of tissues can result andproductivity can be greatly reduced. Leaves and terminalshoots become stunted and fruit can become scarred anddeformed (Smith et al., 1997).

When T. palmi populations are high, feeding typicallycauses a silvery or bronzed appearance on the surface ofthe leaves, especially along the midrib and veins. In Japan,T. palmi was a problem in autumn chrysanthemums, andalthough only occurring at low densities, damage wasreportedly severe (Kawai, 1990; Miyashita and Soichi,1993). Adults aggregate to the growing points of thechrysanthemum plant before flower differentiation andthen disperse to other parts later (Miyashita and Soichi,1993).

High developmental and reproductive rates of T. palmi

at glasshouse temperatures allow rapid build-up ofpopulations, even from small numbers of females. Todevelop from egg to adulthood, T. palmi requires 194 day-degrees above a thermal threshold of 10.1 1C, anddevelopment from egg to adult takes between ca. 10 (at30 1C) and 40 days (at 15 1C), (McDonald et al., 1999).

2. Materials and methods

2.1. Identification

In the UK outbreak, adult thrips were sent to the CentralScience Laboratory (CSL), for identification, on the stickytraps on which they were trapped. The traps were scanned

under a binocular dissecting microscope (Wild M10, LeicaMicrosystems, Wetzlar, Germany) at 80 magnifications.Putative T. palmi were recognised using characters such as:the number, size and location of setae on the head,pronotum, forewing, abdominal sternites and abdominaltergite II; as well as colouration characteristics.However, in order to confirm the identity of the putative

T. palmi, individual thrips were removed from each trap(the smaller the number of thrips caught on any one trap,the greater the proportion that were removed and slidemounted) and microscope slide-mounted in Heinz Mount-ing Medium (Heinze, 1952). Mounting the thrips in HeinzMounting Medium rather than Canada balsam allowedmore thrips to be critically examined at high magnificationmore quickly. This was essential with sticky traps fromboth sites being sent to the laboratory for examination on aregular basis throughout the eradication campaign, fort-nightly at the height of the campaign. The thrips werefloated free from the trap in a drop of white spirit,transferred to fresh white spirit in a watch glass, agitatedgently to get rid of the remaining glue, transferred to adilute detergent solution, gently heated for 15min at 70 1C,slide mounted, and heated at 70 1C for 30min. The thripswere then examined at magnifications of up to 500 using aLeitz Diaplan compound microscope (Leica Microsys-tems). Individual specimens were confirmed as T. palmi

using diagnostic characters as outlined in the literature(e.g., Sakimura et al., 1986; zur Strassen, 1989; Anon,2001). All other thrips species were also identified fromslide-mounted specimens using the keys provided byMound et al. (1976). In all cases, only adult material wasexamined and confirmed to species.The numbers of T. palmi present were recorded by

counting the specimens caught on each trap. When only afew thrips had been caught, all were removed for slidemounting. When very large numbers were being trapped,an estimate was made, based on counting the numbersof putative T. palmi present on just one third of thesticky trap and multiplying that figure by three. Only a verysmall proportion of thrips were in too poor a condition onthe traps for a putative identification to be assigned tothem. In comparison to the very large numbers of thripsindividually confirmed as T. palmi between April andDecember 2000, only very small numbers of the verysimilar species Thrips flavus Schrank were ever recorded ateither site. This suggests that the estimates were reasonablyaccurate and not inflated by erroneous inclusion of thesimilar species.

2.2. Crops and site

Chrysanthemum production at the nursery in Sussexwhere T. palmi was reported in 2000, accounted for ca.30% of the total UK production at the time: ca. 100,000stems per week. The company operated two glasshousesites—referred to hereafter as Site 1 and Site 2—which wereseparated by ca. 100m at their nearest points. There were

ARTICLE IN PRESS

Table 1

The standard cropping pattern at Site 2 (all-year-round chrysanthemums, var. Lineker)

Sequence of events Timing Comments

Cuttings stuck in peat blocks containing

Intercept granules

Week 1 Propagation in imidacloprid (Intercept 5GR granules containing 5%

imidacloprid) incorporated compost

Rooting cuttings receive first insecticide spray

and then covered by plastic sheeting to

improve humidity

Week 1 Teflubenzuron (‘Nemolt’, 150 g l�1 SC; 13.6%w/w)a and L. muscarium

(‘Mycotal’, 16.1% w/w WP) Treatments)b

Small plants in peat blocks transferred to

growing bays

Week 2 ‘Planting out’ after 7–10 days

‘Spot’ insecticide treatments along crop edges

(4.5� 8m)

Week 3–4 Aldicarb (‘Temik10 G’ granules, 10%) applied 2–3 weeks after planting

out

Limit of persistence of imidacloprid granules Week 4 Based on first appearance of the pest

Applications of foliar insecticdes commence Week 6 Foliar sprays of abamectin (‘Dynamec’, 18 g l�1 EC; 1.88% w/w)a

teflubenzuron (‘Nemolt’, 150 g l�1 SC), imidacloprid (Admire, 70%

w/w WG)c and others

Last foliar spray Week 9–13 Short pre-harvest intervals, e.g. 7 d for malathion (600 g l�1 EC;

56.6% w/w)d

Cutting of stems Week 10–14 Depending on whether ‘winter’ or ‘summer’ crop

aApplied at a high volume rate: 0.5 l product in 1000 l water (0.5%); e.g. 50ml of Dynamec per 100 l water.bApplied at an average rate of 1.5 kg product per 1500 l water (0.1%) per ha; between 1000 l of spray liquid per hectare when the crop is low, to

2,000 l ha�1 when crop is high.cApplied at a rate of 10 g and then 20 g a. i. 100 l�1 of water (see Table 2).dApplied at a rate 190ml product per 100 l water in chrysanthemums.

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–1314 1305

about 2.1 million chrysanthemum stems growing in bothsites at the time of the outbreak, �1.1m in Site 1 and 1m inSite 2. In 2000, the company harvested �200 stems/m2.Cultivars and husbandry practices varied considerablybetween the two sites.

The plants were grown to a winter and summer spacingregime: with the closest spacing—60 stems/m2—from Week22 (29 May 2000) to Week 37 (11 September 2000), and theleast dense spacing—40 stems/m2—from Week 46 (13November 2000) to Week 10 (5 March 2001). For furtherdetails on the economics of the nursery and all-year-roundchrysanthemum production and sales, see MacLeod et al.(2004).

With a total growing area of 24,600m2 (2.5 ha), Site 1was a series of 16, linked, clear-span glasshouses,connected by an enclosed, wide central pathway. Each‘house’ was effectively a single, large (ca. 0.18 ha) glass-house bay. Sodium lighting was installed in 2000.Differently coloured cultivars of var. Reagan were grownin 14 beds each week, together with 7 beds of a whitedouble flowered var. Euro. The crops in the entire series ofhouses were grown in rotation, with about 1.5 houses cuteach week.

Site 2 was a 20,500m2 (2.1 ha), single continuous multi-span glasshouse growing Goldstock Breedings’s var. Line-ker for supermarket sale only. Under artificial lighting inwinter, var. Lineker took a maximum of 11.5 weeks tomature, whereas var. Reagan took at least 14 weeks.Rooted cuttings—destined for both sites—were housed intwo propagation bays on either side of the centralwalkway; all cuttings were of Kenyan origin. The cuttingsreceived a spray of Mycotal, (Lecanicillium muscarium

(Petch) [formerly Verticillium lecanii (Zimm.) Viegas]) and

Nemolt (teflubenzuron), and were then covered by plasticsheeting to improve humidity (Tables 1 and 2).The remainder of the very large area at Site 2 was

operated as a single rotating crop, using various colours ofvar. Lineker. Bays 19–36 on the north side, and bays 1–18to the south side, of the central walkway. At any giventime, only a small proportion of the beds (e.g. three out ofthe 36) were fallow, either being harvested or undergoingbed preparation for the next crop. The cropping cycle(ca.10 weeks in the summer; 12–14 weeks in winter) movedin the same direction (i.e. away from the propagation houseand entrance) on both sides of the house (18 bays), butthese were not in synchrony. There were heating pipes(35 1C) at the base of the plants, which acted as secondarysupports, being raised as the crop grew. Well above thecrop, another loop of heating pipes were run at 80 1C. Also,attached to the rising heating frame were thin polythenetubes carrying CO2 enrichment into the crop. In additionto trickle irrigation into the soil, there was heavy mistinginto the var. Lineker crop. The irrigation system delivered105 l of water per nozzle per hour, the equivalent of 5 l/m2

every 17min.At Site 2, spraying and watering was automatic, and

workers only re-entered the crop on two occasions afterplanting: the first time, two weeks prior to harvesting, tocarry out ‘de-budding’; and the second time, to cut thecrop. Prior to the involvement of the Plant Health Service,detailed weekly pest monitoring records had been main-tained for the period August to December 1999, but afterthis time, recording ceased as a result of managerialchanges. Numbers of ‘thrips’, i.e. not identified to speciesor type, had been recorded by the growers for each bay, ona weekly basis.

ARTICLE IN PRESS

Table 2

Timetable of T. palmi control measures and other significant events in the eradication campaign at a nursery site growing all-year-round chrysanthemums,

Sussex, UK

Control measure/event Date/timing Comments

Notice served on Site 2 26 April 2000 Total ‘hold’ notice

Intensive eradication programme required

under notice

Late April 2000 Initial recommendations provided by CSL

Notice served on Site 1 3 May 2000 Prohibit movement of plants apart from

retail sale of cut flowers after treatment

Incorporation of Intercept 5GR granules to

peat-based growing media commenced

Mid-May 2000 Whole of the Site 2 treated by 25 July 2000

Revised notice served 25 July 2000 Stricter regime introduced for ‘rouging out’

plants

Statutory notice revised regarding frequency

of spray applications

18 Aug 2000 Revised recommendations following

discussions on 17 Aug 2000

Methyl bromide (MeBr) fumigation

commenced

14 Aug 2000 Bays 11, 12 and 13 treated (Site 2)

MeBr treatments 2 Oct 2000 House 6 treated (Site 1)

MeBr treatments 4 Sept 2000 Houses 1 and 16 treated (Site 1)

MeBr treatments 4 Sept 2000 Bays 11–22 treated (Site 2)

MeBr treatments 9 Oct 2000 Bays 19–36 completed Whole of the Site 2

done

New spray adjuvant introduced 31 July 2000 Codicide oils first tested and introduced in

Aug 2000

New Conejet nozzles fitted to automated

sprayer

August 2000 To improve crop penetration

‘Winter spacing’ of plants introduced 3 September 2000 (Site 2) A 25% reduction in plant numbers

Peak numbers of adults on sticky traps 29 Aug–11 Sept (Site 2) and 4–18 Sept 2000

(Site 1)

Average numbers of 1.7 (Site 1) and 48 (Site

2) T. palmi per trap per day

Population ‘crash’ 18 Sept–2 Oct (Site1) and 11–25 Sept 2000

(Site 2)

Marked decline between 2-week trapping

periods

Plastic sheeting (mulches) introduced First used in bay 32 on 9 Oct 2000 (Site 2) To prevent/inhibit T. palmi mergency from

soil

Mastertrap introduced into propagation

house

10 Oct 2000 Measure to prevent re-infestation of cuttings

Final ‘space’ treatments carried out. 19 Oct (Site 2) and 3 Nov (Site 1) 2000 Dichlorvos no longer used: inefficacous

against T. palmi

Propoxur smokes used in Site 1 as ‘space’

treatment

28 Nov 2000 Only used on newly planted crops due to

scorch.

Last T. palmi adult(s) trapped 3–15 Jan 2001 (Site 2) and 27 Nov–11 Dec

2000 (Site 1)

Numbers were at zero after these 2-week

periods

Switch to higher, 0.02% rate (20 g a.i. 100 l�1

of water) for Admire applicationsa8 Jan 2001 Based on CSL recommendations

‘Bait’ plants introduced 12 June 2001 Removed after 2 weeks

Eradication campaign review date set in

advance

31 March 2001 Crop destruction an option if outbreak still

on-going

Pest freedom for 140 d (Site 2) 4 June 2001 No pest found for c. 7–8 T. palmi generations

Eradication achieved 26 June 2001 Pest freedom for two full crop cycles

(2� 81 d)

Eradication formally declared 17 July 2001 Statutory notices removed

aAdmire was initially applied at a rate of 10 g a.i. 100 l�1 of water (i.e. 0.01%).

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–13141306

A battery-powered, high-volume robotic sprayer (Brink-man Alumaster, Brinkman Techniek BV1)—originallyfitted with flat fan nozzles, but changed to hollow conesin August 2000, to improve penetration of the crop—waspresent at Site 2. The sprayer ran along the heating pipesand took �3 h to spray the 20,500m2 house (Gress, 2003),spraying on both the outward and return run. A change of

1/http://www.brinkman.nl/pdf/Gewasbeschermingstechniek/Alumas-

ter.pdfS

spray adjuvant (to rapeseed oil) also improved spraypenetration. Thorough coverage was thereby achieved,avoiding run-off.These chrysanthemum varieties reportedly performed

best at average temperatures between 20 and 22 1C. Atboth sites, the thermostats were set for 19.5 1C at night and18.5 1C during the day, however solar radiation increasedthe daytime temperatures upwards, such that temperaturesaveraged between 21 and 22 1C, over a 24-h period.However, average temperatures may have increased

ARTICLE IN PRESSR.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–1314 1307

beyond this level on particularly hot, sunny days, when theability of the vents and screens to regulate glasshousetemperatures was challenged.

2.3. Monitoring

Weekly pest monitoring record sheets were obtainedfrom the growers for Site 2, for the period 9 August to 10December 1999. The nursery only kept records for thislimited period and they were estimates of numbers of‘thrips’, i.e. not identified to species or type, norindividually counted.

From April 2000, the Plant Health and Seeds Inspecto-rate placed blue sticky traps (Oecos Blue 100mm�200mm) in each of the 16 houses on Site 1, and in,approximately, every other one of the 36 bays in Site 2. Thetraps were located �10–15m from the central walkways.Two traps were also present in each of the two propagationhouses. As both sites are �2–2.5 ha in area, the trappingdensity of the traps was, in effect, 18/20,500m2 (or ca. 1trap per 1139m2 in Site 2); and 16/24,600m2 (or ca. 1 trapper 1538m2 in Site 1). There were also two sets of stickytraps placed outside the house at Site 2: 12 were locatedaround the base of the greenhouse and another 12 weresited ca. 10m away from the house.

3. Results

3.1. The situation prior to involvement of the Plant Health

Service

Analysis of Plant Health Service sticky traps receivedfrom both sites in April 2000 (see below), revealed highlevels of T. palmi, indicating a substantial outbreak. Inaddition, previous pest monitoring records kept at thenursery, suggested that the pest had probably been presentfor some considerable time. These indicated that very highnumbers of a leaf-feeding thrips had been present in at leastone site (i.e. Site 2) throughout the period August toDecember 1999. According to the records examined,numbers peaked on 15 November 1999 (Week 46), with amaximum average weekly sticky trap catch of 4600 thripsper trap in one bay.

3.2. Plant health services sticky trap catches

Trapping periods at both sites were somewhat irregularto begin with, but the overall catches were averaged out interms of numbers per trap per day. For example, T. palmi

numbers for the four consecutive periods, from 26 April to11 July, were: 0.5, 1.4, 0.4 and 0.1 per trap per day. Thepopulation peaked between 4 and 18 September 2000 inSite 1, but overall, numbers were considerably less than atSite 2. Trap catches from 18 September to 2 October 2000,decreased markedly, with just three individual T. palmi

caught in two houses. It will be argued below (Section 3.8)that this population crash was as a result of the

intensification of insecticide applications. By 6 December2000, only a single thrips was trapped.Numbers of T. palmi in Site 2 reached a peak, 29 August

to 11 September, when a total of 9995 adults were captured(Fig. 1), or ca. 770 per day on average, on 16 sticky traps.This was a 110% increase on the previous period, 15–28August, with numbers increasing in 10 out of the 16monitored bays. The highest numbers were recorded inbays where flower harvesting was occurring. Very largenumbers of T. palmi were trapped in certain bays, whilst inothers, none, or very low numbers occurred; a highlyuneven distribution. The thrips were most prevalent at thetwo ends of the glasshouse, but high numbers also occurredin bays that were ‘in bud’.Trap catches then declined markedly, to ca. 4T. palmi

per trap per day, 12–24 September (Fig. 1). Numberscontinued to decline, to 0.83 (25 September–9 October),and then to 0.25 per trap per day (9–23 October 2000).There followed a small, probably insignificant, increase to0.43 (23 October–6 November) and then from 6 to 21November, remained at about the same level, with 0.45T. palmi per trap per day.From 21 November to 4 December 2000, only 11

T. palmi were captured, i.e. 0.05 per trap per day, onaverage, and for the first time more F. occidentalis (n ¼ 17)were present on the traps. During the next period, 18December 2000 to 3 January 2001, numbers remained verylow—0.01 T. palmi per trap per day on average—a total ofonly four individuals (Fig. 1). Seven T. palmi were trapped,3–15 Jan 2001, the last individuals to be captured at Site 2.

F. occidentalis numbers were generally very low at Site 2.Across both sites, small numbers of indigenous thrips wererecorded, presumably caught on the sticky traps afterentering the glasshouse units from outside. Though theabsolute numbers of these additional species were low,many different species were recorded from both sites,2000–01. In total, 30 native species (not including WFT)were recorded; these species included Thrips tabaci Linde-man, but not Thrips nigropilosus Uzel, another speciesknown as a pest in UK glasshouses.

3.3. Outdoor traps and nearby sites

In August 2000, T. palmi adults were captured on two of24 traps placed out-of-doors in the vicinity of Site 2: onefrom an inner circle (1m out) and one from an outer circle,10m out. During the eradication campaign, the PlantHealth and Seeds Inspectorate (PHSI) inspected other localnursery premises surrounding the outbreak. Initially, thisinvolved inspecting all neighbouring nurseries within 1 km,every 4 weeks; those within 1–3 km, every 12 weeks; andthose between 3 and 5 km, every 26 weeks. However, thiswas subsequently modified, such that nurseries within the1 km radius were monitored every 6 weeks. On 4 May 2000,the first sweep of eight nurseries within a 1 km radius wascompleted and in the following months, inspections wereextended outwards to include another 42 premises within a

ARTICLE IN PRESS

0 0 0 0 0 0 0 0 0

121

282

242

3030

234

106

888

3

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

10-26Apr

26 Apr -9 May

12 May-23 Jun

23 Jun -11 Jul

11-24 Jul 24 Jul -21 Aug

21 Aug -4 Sep

4-18 Sep 18 Sep -2 Oct

16 Oct -6 Nov

6-13 Nov 13-27Nov

27 Nov -11 Dec

11-25Dec

25Dec -8 Jan

8-22 Jan 22 Jan -5 Feb

5-19Feb 19 Feb -5 Mar

5-19March

Sampling period

Av

era

ge

no

. o

fT. p

alm

i p

er

tra

p p

er

da

y

Fig. 1. T. palmi sticky trap catches at Site 1: average numbers per day (16 traps) for different sampling periods: 10 April 2000 to 19 March 2001. Data

labels on columns are the total number caught in each period.

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–13141308

3 km radius; and then another 68 premises within 5 km.Repeat inspection of premises within a 1 km radius werecarried out between 3 and 24 September 2000 and in March2001. No T. palmi were found on sticky traps from any ofthe premises included in these inspections.

3.4. Control measures

The Plant Health Service established with the UKPesticide Safety Directorate (PSD) that the use of Interceptgranules in peat-based media used for propagatingchrysanthemum cuttings, whilst not a label recommenda-tion, could be utilised against this quarantine pest in thisemergency. Thus, from mid-May 2000 onwards, rootedcuttings for both sites were grown in peat containingimidacloprid granules (Table 2). The persistence of thischemical within the plugs—i.e. after the young plants hadbeen planted out in the crop bays—was not determined,but was probably limited to ca. 4 weeks, and was therefore,an expensive option (MacLeod et al., 2004).

The Plant Health Service also submitted an applicationto the Pesticides Safety Directorate, on 19 December 2000,for the emergency off-label use of ‘Admire’ (i.e. a waterdispersible granule (WG) formulation containing 70% w/wimidacloprid) on protected, all-year-round chrysanthe-mums, for the control of T. palmi. The application madethe case that some alternative chemicals were ineffective,due to resistance, and that others were in ‘use-up’ priorto revocation; also there were no plans to grow cropsother than ornamentals on the sites of application (seebelow).

Off-label approval for the use of ‘Admire’ againstT. palmi was obtained on 12 March 2001 (PSD SOLAno. 0584/2001). The authorisation stipulated that ‘Admire’was subject to use only on protected (i.e. a crop grownunder a permanent or temporary cover) chrysanthemums,at a rate of 178.5 g product per hectare, and could only beused if growers had been issued with a Plant Health Orderto control T. palmi on chrysanthemum. ‘Admire’ ismarketed in the UK for use as a drench to controldamson/hop aphids, with a recommended rate of 0.049 gproduct (70% w/w WDG) per plant. Imidacloprid wasinitially applied as a foliar spray at a rate of 10 g a.i. 100 l�1

of water (i.e. 0.01%), but extrapolating from recommendedrates for thrips on labels for other imidacloprid products,such as ‘Confidor’ (not registered in the UK), it wasrecommended that the rate be increased to 0.02% (or20 g a.i. 100 l�1) (Table 2).Biological control was considered, but was not thought

to a viable option for eradication purposes, mainlybecause T. palmi-specific biological control agents wereunavailable in the UK. The risks to wildlife and theenvironment from the use of insecticides in these sites werelow, because the nurseries were contained glasshousesgrowing non-food crops. Cox et al. (2006) reviewed thepotential of natural enemies for use in the eradication ofT. palmi, and as part of potential IPM strategies againstthis pest in the UK.Steaming with portable steamers was not practical

because the area to be covered (4.6 ha) was much toolarge. Methyl bromide fumigation of Site 1 had beencarried out in 1999, over a ca. 11-week period, as the beds

ARTICLE IN PRESS

Table 3

Insecticide categories applied Site 2, between 21 June and 18 July 2000

Date Insecticide

21 June Carbamate

21 June Entomopathogen 1a

21 June Entomopathogen 2b

21 June Carbamate/translaminar avermectin

26 June Carbamate

26 June Entomopathogen 2b

28 June Carbamate

28 June Entomopathogen 1a

28 June Entomopathogen 2b

28 June Carbamate/translaminar avermectin

30 June Organophosphate

2 July Carbamate/translaminar avermectin

2 July Translaminar avermectin

4 July Carbamate

4 July Entomopathogen 2b

5 July Organophosphate

6 July Entomopathogen 1a

6 July Carbamate

8 July Translaminar avermectin

12 July Entomopathogen 2b

12 July Organophosphate/Entomopathogen 2b

12 July Entomopathogen 1a

18 July Carbamate

18 July Entomopathogen 2b

19 July Carbamate/Entomopathogen 2b

19 July Entomopathogen 1a

*No UK approval for the use of this formulation in this situation.

/ ¼ separates ingredients applied as mixtures.aBauveria bassiana (experimental formulation).bLecanicillium muscarium (formerly Verticillium lecanii).

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–1314 1309

became progressively bare after cropping. Fumigation isexpensive (MacLeod et al., 2004) and effectively delays re-planting by about 1 week. Nevertheless, fumigation of bothsites was repeated in 2000, from late August-early October,taking ca. 12 weeks to treat each site (Table 2).

The use of roller traps (sheets of yellow plastic withglue on both sides), a form of mass trapping, was restrictedat Site 2 as a result of heating pipes and an overheadspraying system. Other alternatives, such as water traps(Layland et al., 1994), would also have been subject tothe same logistical constraints, and were therefore notemployed.

Leaf damage caused by thrips feeding was most apparentat about the 7–8 week stage—i.e. prior to flowering—andwas visible along the margins of the crop at Site 2. Damagewas largely localised to individual plants (‘stems’), and thepattern of damage was consistent with the thrips beingdisplaced off mature chrysanthemum stems during cutting,and moving back along the house to infest the less matureplants. Roguing of ‘hot spots’ was carried out andprobably contributed to the overall eradication: up to90% of stems were removed from the first few metres ofsome heavily infested bays.

Commencing on 9 October 2000, the whole soil surfaceof Site 2 was covered with white polythene (and notsprayed with rapeseed oil). Significant improvements incrop production occurred in the following months, whichwere attributed by the growers to the use of the mulch.However, this measure produced a 20% increase in labourcosts because additional alterations to the size anddistribution of the holes (for the plant stems) were requiredas the crop matured. Therefore, the use of the mulch wassuspended in early 2001.

3.5. Control measures prior to the involvement of the Plant

Health Service

In the months prior to the serving of Statutory Noticesby the Plant Health Service, dichlorvos ‘space’ treatments(i.e. automated fumigation with dichlorvos delivered as afog) had been applied, but usually to only two or four ofthe 12 glasshouses on a given day. From 9 February to 28April 2000, insecticides were applied 17 times at Site 1, butagain usually to only three or four (sometimes six) of the 12glasshouses, approximately once a week on average. Thesame insecticides were used at both sites (see: Table 3 andMacLeod et al., 2004). In other words, prior to theinvolvement of the Plant Health Service, insecticidetreatments were infrequent and intermittent, and thepartial coverage of the site was insufficient to eradicatethe pest. Nevertheless, given the relative degree ofseparation of the individual glasshouses at Site 1, partialcoverage was less of a problem than at Site 2.

Applications of L. muscarium and teflubenzuron weremade to the chrysanthemum cuttings in the propagationhouse at Site 2 (Table 1), and during the first 6 weeks of anew crop. All applications were made according to UK

label recommendations for each product (Table 1). ‘Spot’applications of aldicarb were also made, 2–3 weeks afterplanting out, predominantly along the ‘edges’ (a strip,8� 4.5m) of each bay (Table 1). Applications of otherinsecticides—including abamectin, deltamethrin and ma-lathion—followed, predominantly after 6 weeks (Table 1).Dichlorvos treatments (a 600 g/l EC formulation applied asa thermal fog at a rate of 100–150ml per 2500m3) were alsocarried out, but were infrequent—e.g. between 10 Febru-ary–16 April 2000, only one such ‘space’ treatment hadbeen applied—and to only half of the house (i.e. bays19–36). Foliar applications were similarly infrequent andincomplete. A variety of mixtures as well as single chemicalapplications were being used, but only six or eight out ofthe 36 bays were being treated on any single day.Heptenophos was used, up until 31 April 2001, the post-revocation, deadline for of a ‘use-up’ period for existingstocks; nicotine or teflubenzuron were alternatives whichcould be used instead of heptenophos.

3.6. Plant Health Service treatment recommendations

The control measures being carried out on the nursery,prior to the involvement of the Plant Health Service, wereunlikely to eradicate the pest and a concerted eradication

ARTICLE IN PRESSR.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–13141310

programme was needed. Due to the rapid reproductive rateand short development time of T. palmi, the optimumstrategy for eradication was a highly intensive, insecticidetreatment programme, including:

�

Ta

Nu

ins

Da

19/

2–1

16–

30/

13–

27/

11–

25/

8–2

22/

6–1

20/

03–

17–

31/

14–

28/

a

per

‘space’ treatments: alternate applications of dichlorvosor nicotine, every two or three days for at least 10 days;

� systemic treatments: imidacloprid or aldicarb, to target

cryptic life stages within the plant; and

� foliar treatments: such as abamectin, on a weekly basis,

plus an additional foliar treatment if the plants were nottreated with a systemic.

In the ca. 2 months immediately after the StatutoryNotices were served, there was a marked increase in therate of application of dichlorvos ‘space’ treatments, andwhole-house applications were made. Nevertheless, thegrowers did not manage to achieve the required, 2–3 dayfrequency, with an average of about once treatment every 5days. However, this soon became academic, since efficacytests carried out at Central Science Laboratory demon-strated that although abamectin and imidacloprid wereeffective against adult T. palmi collected from this site,dichlorvos was largely ineffective. Since the dichlorvostreatments were having minimal effect, applications ceasedafter 19 October 2000 at Site 2, and after 3 November2000 at Site 1. Applications of foliar insecticides alsoremained rather irregular, with different mixtures beingapplied to varying proportions of the house, such that one,three, five, six, eight, 15 or 18 beds were being treated withdifferent combinations of insecticides at different times(Table 4).

ble 4

mbers of bays (out of a total of 36) at Site 2 receiving one or more (1–5)

ecticide applications per 2-week sampling intervals, 2000–01

te Nos. of treatments

0 1 2 3 4 5 Totala

07–01/08/00 11 1 2 15 7 — 78

5/08/00 9 4 1 18 4 — 76

29/08/00 8 15 9 4 — — 45

08–12/09/00 — 6 2 4 24 — 118

26/09/00 7 5 2 1 3 18 114

09–10/10/00 12 4 8 3 9 — 29

24/10/00 13 3 17 3 — — 46

10–07/11/00 19 6 8 3 — — 31

1/11/00 3 10 17 6 — — 62

11–05/12/00 11 5 — — 20 — 85

9/12/00 12 1 — 5 18 — 88

12/00–02/01/2001 24 3 9 — — — 21

16/01/01 9 5 3 4 15 — 83

30/01/01 10 3 5 — 14 4 89

01–13/02/01 18 — 4 — 14 — 80

27/02/01 8 3 7 4 — 14 99

02–13/03/01 3 1 8 1 23 — 112

Total no. of ‘bay treatments’, a measure of the overall spraying effort

fortnight.

3.7. Insecticide spraying at Site 2

On the 17 August 2000, the treatment recommendationswere amended to take more account of agronomicpractices—and restrictions—in place on the nursery, anda less-intensive spraying programme was instigated. Thisconsisted of three foliar treatments (alternating abamectinwith malathion or deltamethrin), and three ‘space’ treat-ments, every 2 weeks. The most frequently used foliarinsecticides, included—in various combinations—abamec-tin, malathion, deltamethrin, L. muscarium and imidaclo-prid. The typical sequence of applications is shown inTable 3, for a representative period.The T. palmi population was at its peak during 29

August to 11 September, followed by a pronounced ‘crash’(410-fold decrease) in numbers, 12–24 September (Fig. 1).The number of insecticide applications from 19 July 2000to 13 March 2001, per 2-week period, is shown in Table 4.Individual bays received a variable number of treatments,depending on crop growth stage, incidence of the pest andso on. A measure of the intensity of insecticide sprayingcarried out per fortnight, is given in terms of the total ‘baytreatments’, in Table 4. The population crashed after amonth of particularly intensive spraying, i.e. between 30August and 26 September, when a total of 232 (118+114)‘bay treatments’ were made (Table 4). Significantly, foliarapplications of imidacloprid increased markedly duringthis period, with 24 bays receiving one or more treatmentsin the period 30 August–12 September; and 28 bays—all ofthose which were ‘in crop’—during 13–26 September(Table 4). This was the highest ever, site-wide use of thisinsecticide, and the population ‘crashed’ at this time (Fig.1). Applications of other insecticides, such as abamectinand malathion, were also applied more widely during thisperiod. The intensity of insecticide spraying peaked duringthis period (Fig. 4).On average, 29% of bays received no treatment at all in a

given two-week period; these were either fallow (10%) orrecently planted. Some bays—which were ‘hot spots’ forthe pest—received more than the recommended number oftreatments: 34% received four or more applications per 2-week period; 28% received four, and 6% received five(Table 4).

4. Discussion

4.1. An established population

The UK outbreak of T. palmi was first officially notifiedto the Plant Health Service on 26 April 2000, by which timeit was well advanced. The eradication of the outbreak wasmanaged according to an International Standard forPhytosanitary Measures (ISPM): Guideline for PestEradication Programmes (Anon., 1999). A managementteam was established, consisting of officers from the threebranches of the UK Plant Health Service: Central ScienceLaboratory, Defra Plant Health Division and the PHSI.

ARTICLE IN PRESSR.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–1314 1311

This met regularly throughout the campaign—using videolinks with local inspectors—to co-ordinate activities.

Early detection of an outbreak is usually essential to thesuccessful management of incursive, or invasive thrips(Morse and Hoddle, 2006). Their small size and cryptichabits however, make them difficult to detect, and in thisUK outbreak, high numbers of an unidentified leaf-feedingthrips—almost certainly T. palmi—had been present for atleast 6 months, possible longer.

4.2. A comparison of numbers at two sites

Overall, numbers were considerably lower at Site 1 thanat Site 2 (Figs. 1 and 2). This is unlikely to have been as aresult of varietal differences, since only small differences inT. palmi numbers were found between different chry-santhemums varieties in Japan (Miyashita and Soichi,1993). More likely, the ca. 10-fold difference in populationsizes was due to fact that one site (1) was a series of largelyseparate—albeit interconnected—glasshouses, whereas theother (2) was a very large, open space. The largerpopulation may therefore, have developed as a result ofthe greater openness of the larger space, and was sustainedby free movement around the site. The control measuresappear to only have been successful in causing a population‘crash’ when the entire crop in Site 2 was sprayed on thesame day. In other words, there was nowhere left for thepest to persist.

Peaks in numbers generally coincided with bud emer-gence, and numbers were usually higher in bays where the

12493263

1180

3471

1571

37873096

4757

9995

721

209

0

10

20

30

40

50

60

10-26

Apr

15-19

Jun

3-17 Jul 31 Jul -

14 Aug

29 Aug -

11Sep

25 Sep -

9 Oct

Sampl

Avera

ge n

o. o

f T

. p

alm

i p

er

trap

per

day

Fig. 2. T. palmi sticky trap catches at Site 2: average numbers per day (16 tra

labels on columns are the total number caught in each period.

chrysanthemum crops were ca. 40–60 cm tall, compared tobays with a more mature, flowering crop, ca. 70–80 cm.Damage was also most apparent prior to flowering, atabout the 7–8 week stage.Declines in T. palmi numbers within individual bays were

usually associated with crop maturation and harvesting,but numbers trapped were sometimes higher in bays whereflowers had been cut, suggesting that disturbance wasresponsible for certain apparent peaks in the popula-tion. There was also a large increase in the numbersfollowing disturbances in preparation for methyl bromidefumigation.The amount of visible damage caused to the chrysanthe-

mum plants was relatively minor, and no significant flowerproduction was lost. Hot spots of damage were presenthowever, localised in foci along the margins, with severedamage seen on some individual stems along the internaledges of the bays, i.e. on either side of the central walkwayin Site 2. This concrete strip may have been used as adispersal route for aerial thrips displaced from adjacentbays as a result of harvesting. Light levels and microclimatemay also have had localising effects.A degree of synchrony—in terms of overall trapping

patterns—between the two sites is apparent when numbersare plotted on a logarithmic scale (Fig. 3). The T. palmi

populations at both sites reached their peaks at approxi-mately the same time, mid September 2000, allowing forthe overlap in sampling periods. The two T. palmi

populations also declined to a trough at much the sametime, in early July (Figs. 1 and 2). Why the populations

53 84 130 11 3 7

23 Oct -

6 Nov

21 Nov -

4 Dec

18 Dec-3

Jan

15-26

Jan

12-26

Feb

12-26

March

ing period

000004

ps) for different sampling periods: 10 April 2000 to 26 March 2001. Data

ARTICLE IN PRESS

0

0.5

1

1.5

2

2.5

April 1,

2000

May 1,

2000

June 1,

2000

July 1,

2000

August 1,

2000

September

1, 2000

October 1,

2000

November1,

2000

Month

Avera

ge n

os. o

f T

. p

alm

i (l

og

n+

1)

per

trap

per

day

Fig. 3. Numbers of T. palmi (log n+1 mean numbers per yellow trap per day). Light shading: Site 1; dark shading: Site 2.

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–13141312

should be congruent—given (i) the physical separation ofthe sites; (ii) the different growing conditions; (iii) thedifferent varieties; and (iv) the difference in absolutenumbers—requires an explanation. It is possible thatambient external weather factors—such as solar radia-tion—were important drivers, which produced a degree oftemporal synchrony in both populations. There washowever, also a good degree of co-ordination in the timingand application of control measures between the two sites,which could have resulted in similar population trends.

4.3. Sampling rates

The recommended monitoring rate for T. palmi inDendrobium orchid nurseries in Thailand—where T. palmi

is established—is one to two sticky traps every 220m2

(Keinmeesuke, 2005). In the Netherlands, T. palmi infest-ing nurseries growing Ficus benjamina were monitoredusing 10–58 sticky traps, depending on the number ofcompartments, and traps were changed weekly (Vierber-gen, 1996). This is similar to the monitoring rate used in theUK outbreak, which was approximately one trap per1538m2 (Site 1); and one trap per 1139m2 (Site 2).

4.4. A gamut of control measures

The eradication of the UK outbreak of T. palmi relied ona range of different control measures (Table 2). In a non-empirical situation, it is difficult to separate out the relativecontribution of each treatment measure. However, the rate

of application of insecticides was increased as the pestpopulation grew in size during the summer of 2001 andthe demise of the T. palmi population was probably as a

result of both an intensification of the frequency ofinsecticide applications, and an increase in the proportionof the glasshouses which were sprayed on a given occasion(Table 4 and Fig. 4). The introduction of methyl bromidefumigation treatments, on 14 August 2000 at Site 2, wasanother contributing factor.A major problem in attempting to eradicate T. palmi in

all-year-round crops is the absence of a significant cropbreak (Stansly, 2000). The situation was particularly acuteat the UK outbreak, where the short, 7–10 day fallowperiod, meant that empty beds only occupied a smallproportion (ca. 10%) of the glass houses at any onetime.At Site 1 however, it was possible to have a completelyempty house, although the maintenance of all-year-roundproduction was the same overall. The absence of animmediate source of the pest, i.e. to re-infest maturingchrysanthemum plants, was probably the reason why T.

palmi populations were so much lower at this site.The difficulties experienced by the nursery in trying to

meet the requirements of the statutory eradication pro-gramme, reflected practical problems in attempting tomaintain normal flower production whilst at the same timegreatly increasing spray applications. Problems of staffingand efficiency of spraying were however, graduallyimproved, but on no occasion did the sites ever receivethe recommended number of treatments in any 2-weekperiod, and there was only one period (30 August–12September) when all of the crop bays in Site 2 received atleast one treatment (Fig. 4). Significantly, there followed amarked decline in the population (Fig. 2).The insecticides, which appeared to have contributed

most to the decrease in population, were those whichwere applied more frequently during the population peak

ARTICLE IN PRESS

0

10

20

30

40

50

60

70

80

90

19 July-

01

Aug

02 A

ug- 1

5 Aug

16 A

ug- 2

9 Aug

30 A

ug-1

2Sep

t

13 S

ept-

26 S

ept

27 S

ept-

10 O

ct

11Oct- 2

4 Oct

25 O

ct- 0

7 Nov

08 N

ov- 2

1 Nov

22 N

ov- 0

5 Dec

06 D

ec- 1

9 Dec

20 D

ec- 0

2 Ja

n

03 Jan

- 16

Jan

17Ja

n- 3

0 Ja

n

31Ja

n- 1

3 Feb

14 F

eb- 2

7 Feb

28 F

eb- 1

3 M

ar

Perc

en

tag

e

Fig. 4. Intensity of spraying at Site 2: percentage of bays treated with insecticide applications on three or more occasions, within 2-week periods, 19 July

2000–13 March 2001.

R.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–1314 1313

(30 August to 12 September 2000). They included,imidacloprid in particular, followed by abamectin andmalathion. The use of L. muscarium may also havecontributed. Sequential applications of this biologicalagent with imidacloprid have subsequently been shown toresult in higher mortalities of T. palmi than when thecompounds were used individually (Cuthbertson et al.,2005).

The relative contribution of the white plastic mulching,used in Site 2, is difficult to judge. Although onlyintroduced late in the outbreak (October 2000), by whichtime T. palmi numbers had already declined substantially(Fig. 2), it may nonetheless have contributed to theeradication of the pest. In Japan, mulching ridges withblack plastic reduced emergence of T. palmi (Tsuchida,1997). A plastic covering would have reduced the surfacearea of the soil available for pupation, although somelarvae would probably have been able to emerge via gapsaround the chrysanthemum stems. Both ingress (of larvaeseeking pupation sites) and egress (of emerging adults)into/from the soil would however, be expected to decreaseas a result of a plastic covering.

4.5. Declaring eradication

The average temperature on the nursery over a given 24-h period was 21–22 1C. In the laboratory, T. palmi takes16.6 days on average to develop from egg to adulthood at aconstant temperature of 21 1C (McDonald et al., 1999).However, this is not the same as one full life cycle; for anestimate for the generation time, it is necessary to add onthe pre-oviposition period, i.e. the time between adultemergence and the onset of egg-laying. In Taiwan, the pre-oviposition period was 1–3 days for virgin females and 1–5days for mated ones (Wang et al., 1989). Using thesefigures, T. palmi should complete one full generation in17.6–21.6 days at 21 1C. Extrapolating from this, a 70-dayperiod free of the pest represents between 3.2 and 4.0

generations. In order to declare the sites eradicated of thepest, a minimum of one complete cycle of crop production(ca. 13 weeks) without any T. palmi being trapped, wasconsidered necessary. As an additional safety factor, aminimum of two such crop cycles with pest freedom wererequired before withdrawal of statutory notices. In theevent, this was achieved—together with the absence of anyT. palmi on specially introduced ‘bait plants’—and theoutbreak was formally declared as eradicated on 17 July2001. Official restrictions were removed some 15 monthsafter confirmation of the pest.One reason for the relative lack of commercial loss at the

nursery, i.e. despite the presence of T. palmi and costly theeradication programme which ensued (MacLeod et al.,2004), was because T. palmi, unlike F. occidentalis, feedsmostly on leaves rather than flowers of chrysanthemums.

4.6. Lack of spread

Despite the presence of very large populations ofT. palmi on the two sites for a considerable period—probably at least 18 months (April 1999–October 2000)—no evidence for spread on to neighbouring nurseries insouthern England was obtained. Repeated sampling andinspection of 68 nurseries within a 5 km radius of theoutbreak did not reveal a single individual. Although twoT. palmi were found 10m outside site 2, remarkably fewindividuals were found on the outdoor traps, given the veryhigh populations inside. It appeared that the thrips werenot inclined to leave the glasshouses, presumably becausethe conditions (humidity, temperature and so on) inside theprotected areas were more favourable.

T. palmi is not thought to use airborne chemical cues ofplant origin in long-distance orientation towards hostplants (Hirano et al., 1993); adult invade greenhouseswithout prior discrimination of the vegetables grownwithin (Hirano et al., 1990). Although there is evidenceof movement from in-door, heated facilities, to out-door

ARTICLE IN PRESSR.J.C. Cannon et al. / Crop Protection 26 (2007) 1303–13141314

surrounding crops (Kawai, 2001), such spread is opportu-nistic, rather than directed migration, and may depend onthe specific (i.e. high temperature) conditions prevailingoutside. In the UK, survival out-of-doors would beexpected in the summer, but the potential of adults andlarvae to over-winter is very limited (McDonald et al.,2000). This suggests that spread between neighbouringpremises will not be facilitated by host-plant attraction.Separation distances, together with plant movementprohibitions, probably prevented any additional outbreaksin the region.

5. Conclusions

T. palmi is a notoriously difficult pest to control, buteradication has been achieved in both the Netherlands andthe UK, using largely insecticide-based programmes.Furthermore, the eradication of the pest in the UK wascarried out whilst maintaining all-year-round flowerproduction. Although it was 15 months before the removalof official restrictions, the outbreak itself was under controlwithin 6 months. This was only possible because of the fullco-operation of the growers and their concerted efforts tointroduce more intensive control measures whilst main-taining production. In the event, it proved to be a win-winsituation, in that the pest was eradicated and theproduction methods were improved, albeit at a consider-able cost (MacLeod et al., 2004).

Since the time of this outbreak, an effective newinsecticide (spinosad) has been registered in the UK(Cannon et al., 2006) and has been incorporated intocontingency plans for eradicating any future outbreaks onT. palmi in the UK.

Acknowledgements

The authors are grateful to the growers and workersinvolved in the UK nursery where the UK outbreakoccurred, which remains anonymous for reasons ofcommercial confidence.

References

Anon, 1999. International Standards for Phytosanitary Measures (ISPM)

No. 8.

Anon, 2001. Diagnostic protocols for regulated pests. Thrips palmi. EPPO

Bull. 31, 53–60.

Anon, 2005. The Plant Health (England) Order 2005. Statutory

Instrument 2005 No. 2530. The Stationary Office, London.

Cannon, R.J.C., Matthews, L., Collins, D.W., 2006. A review of the pest

status and control options for Thrips palmi. Crop Prot., in press,

doi:10.1016/j.cropro.2006.10.03.

Cox, P.D., Matthews, L., Jacobson, R.J., Cannon, R., MacLeod, A.,

Walters K.F.A., 2006. Potential for the use of biological agents for the

control of Thrips palmi outbreaks. Biocont. Sci. Technol., in press.

Cuthbertson, A.G.S., North, J.P., Walters, K.F.A., 2005. Effect of

temperature and host plant leaf morphology on the efficacy of two

entomopathogenic biocontrol agents of Thrips palmi (Thysanoptera:

Thripidae). Bull. Ent. Res. 95, 321–327.

Gress, A., 2003. Putting science into practice. Grower, 20–22.

Heinze, K., 1952. Polyvinylakohol-Lactophenol-Gemisch als Einbettungs-

mittel fur Blattlause. Naturwissenschaften 39, 285–286.

Hirano, C., Jinbo, Y., Yasumi, K., Horiike, M., 1990. Thrips palmi:

indiscriminate invasion into green-house with host and non-host

plants. Proc. Assoc. Pl. Protec. Shikoku 25, 53–55.

Hirano, C., Itoh, E., Yasumi, K., Horiike, M., 1993. Short-distance

walking responses of Thrips palmi Karny (Thysanoptera: Thripidae) to

tomato leaves. Appl. Entomol. Zool. 28, 233–234.

Jones, D.R., 2005. Plant viruses transmitted by thrips. Eur. J. Plant

Pathol. 113, 119–157.

Kawai, A., 1990. Control of Thrips palmi Karny in Japan. JARQ 24,

43–48.

Kawai, A., 2001. Population management of Thrips palmi Karny. Jpn. J.

Appl. Entomol. Zool. 45, 39–59.

Keinmeesuke, S.T.P., 2005. Coloured sticky traps for catching thrips in

orchid crops. Food and Fertilizer Technology Center (FFTC). /http://

www.fftc.agnet.org/library/abstract/pt2002007.htmlSLayland, J.R., Upton, M., Brown, H.H., 1994. Monitoring and

identification of Thrips palmi Karny (Thysanoptera: Thripidae). J.

Aust. Ent. Soc. 33, 169–173.

McDonald, J.R., Bale, J.S., Walters, K.F.A., 1999. Temperature,

development and establishment potential of Thrips palmi in the UK.

Eur. J. Entomol. 96, 169–173.

McDonald, J.R., Head, J., Bale, J.S., Walters, K.F.A., 2000. Cold

tolerance, overwintering and establishment potential of Thrips palmi.

Physiol. Entomol. 25, 159–166.

MacLeod, A., Head, J., Gaunt, A., 2004. An assessment of the potential

economic impact of Thrips palmi on horticulture in England and the

significance of a successful eradication campaign. Crop Protect. 23,

601–610.

Miyashita, T., Soichi, N., 1993. Varietal differences in injury to

chrysanthemums by Thrips palmi. III. Relation between infested part,

density and injury. Jpn. J. Appl. Entomol. Zool. 37, 227–233.

Moritz, G., Kumm, S., Mound, L., 2004. Tospovirus transmission

depends on thrips ontogeny. Virus Res. 100, 143–149.

Morse, J.G., Hoddle, M.S., 2006. Invasion biology of thrips. Annu. Rev.

Entomol. 51, 67–89.

Mound, L.A., Morison, G.D., Pitkin, B.R., Palmer, J.M., 1976.

Thysanoptera. Handbooks for the Identification of British Insects 1

(11), 1–79.

Sakimura, K., Nakahara, L.M., Denmark, H.A., 1986. A thrips, Thrips

palmi Karny (Thysanoptera; Thripidae). Entomology Circular, Divi-

sion of Plant Industry, Florida Department of Agriculture and

Consumer Services (No. 280).

Smith, I.M., McNamara, D.G., Scott, P.R., Holderness, M. (Eds.), 1997.

Quarantine Pests for Europe, Second ed. Wallingford, Oxon, UK,

CAB International.

Stansly, P., 2000. A crop-free period is considered a necessity for Thrips

palmi on vegetable. Leadership In Vegetables: Cultural Control for

Management of Vegetable Pests in Florida. University of Florida,

Immokalee /http://www.imok.ufl.edu/LIV/groups/cultural/pests/in-

sects.htm#S.

zur Strassen, R., 1989. Was ist Thrips palmi? Ein neuer Quarantane-

Schadling in Europa. Gesunde Pflanzen 41, 63–67.

Tsuchida, K., 1997. Adult emerging site of T. palmi (Thysanoptera:

Thripidae) in an eggplant field with mulching sheet. Appl. Entomol.

Zool. 32, 246–249.

Vierbergen, G., 1996. After introduction of Frankliniella occidentalis in

Europe: prevention of establishment of Thrips palmi (Thysanoptera:

Thripidae). Acta Phytopathol. Entomol. Hungar. 31, 267–273.

Wang, C.L., Chu, Y.I., Lo, K.C., 1989. The reproductive mechanism of

Thrips palmi Karny. 1. The female ovipositional behaviour. Chin. J.

Entomol. 9, 251–261.