Iategrated Pest Management of

Thrips tabaci Lindeman [Thysanoptera: Thripidael in

Greenhouse Cucumber Production

Celia K. Boone

Submitted in partial fulf'ilment of the requirements for the degree of Master of Science in Agriculture

Nova Scotia Agricuiturai Collzge Truro, Nova Scotia

in cooperation with

Dalhousie University Halifax, Nova Scotia

March 1999

Q Copyright by Celia K. Boone 1999

National Libraiy 1*1 of Canada Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395. nie Welligtari OnawaON KlAON4 OrtawaON K i A ON4 Canada Camda

The author has granted a non- exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sel1 copies of this thesis in microform, paper or electronic formats.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission.

L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/film, de reproduction sur papier ou sur format électronique.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

DEDICATION

To my parents, Helen and Terry, for aiways emphasizing the importance of education and

encouraging my pursuit of knowledge in a multitude of directions in science and in Me;

without whose wisdom and support 1 could never have made it this far.

TABLE OF CONTENTS

Dedication ..............................................................................................................

.......................................................................................................... List of Tables

......................................................................................................... List of Figures

Abstract ..................................................................................................................

List of Abbreviations and Symbols ........................................................................

............,...... ............ ............................................. Ac knowledgements .... ....

1 . Literature Review

1.1 Introduction ....................................................................................

1 -2 Thrips tabaci Lindeman 1 .2.I Taxonomy and synonymy .................... .. ............................... 1.2.2 Origin and geographic distribution .......................................... 1 .2.3 Host plants ............................................................................. 1 .2.4 Life cycle and behaviour ......................................................... 1.2.5 Economic importance .............................................................

1 -3 Integrated Pest Management (PM) ...................................................... 1.3.1 Monitoring ............................................................................. 1.3.2 Control measures

1.3.2.1 Cultural and physical control .................................. 1.3.2.2 Biologicai control .................... .... .....................

.................................................... 1.3 -2.3 Chernical control

2 . Predatory Mites

........................... ................................ Abstract .. . - .

2.1 Background ...................... ... ........................................................... 2.2 Materials and Methods

2.2.1 Thrips source ........................................................................ 2.2.2 Plant material propagation ...................... .... .................... 2.2.3 Experimental arenas ................................. .. ......................... 2.2.4 Predatory mites source .......................................................... 2.2.5 Predation experiments .......................................................... 2.2.6 Statistical analysis ................................... .. .........................

vii

LIST OF TABLES

Table 1. Predation rate of predatory mites on T. tabaci first and second l a n d 32 instars on cucumber leaf discs.

Table 2. The effect of scent on captures of T. tabaci in a commercial 47 cucumber greenhouse.

Table 3. The effect of trap colour on captures of T. tabaci in a commercial 47 cucumber greenhouse.

Table 4. The interactive effect of scent x colour on captures of T. tabaci in 48 research cucumber greenhouses.

Table 5. The effect of trap colour on Orius captures on sticky traps in 53 greenhouse cucumber.

Table 6. Statistics for the regression of mean crowding (5) on mean density 65 (2) for captures of T. tabaci larvae and adults, and the predatory mite, A. czrmmeris, in greenhouse cucumbers.

Table 7. T-test of the index of basic contagion (ar) for significance greater 67 than O for 7: tabaci larvae and adults and A. cucumeris in greenhouse cucumbers.

Table 8. T-test of the density contagious coefficient (b,) for significance greater than 1 for T. tubaci lawae and adults and A. cucumeris in greenhouse cucumbers.

Table 9. Statistics for the regression of log,, variance on log,, mean for captures of T. tabaci larvae and adults, and the predatory mite, A. cucumeris, in greenhouse cucumbers.

Table 10. T-test of the index of aggregation (b) for significance greater than 1 for T. tubaci larvae and adults and A. cucumeris in greenhouse cucumbers.

v i i

LIST OF FIGURES

Fig. 1. Relationship betweea adult thrips captures on clear sticky trap captures 52 and standard leaf samples in greenhouse cucumbers.

Fig. 2. Correlation between adults thrips captures on clear sticky traps and 52 standard leaf simples in greenhouse cucumbers ('significance at a=O. OS).

Fig. 3. Iwao's patchiness regression of meao crowding (2) on mean density 66 (2) of i? tabaci larvae (A) and adults (B) and A. c u ~ ~ m e r i s (C) in greenhouse cucumbers.

Fig. 4. Iwao's sequential sampling plan for T. tabuci lawae (A) and adults (B) 70 in greenhouse cucumbers.

Fig. 5. Taylor's power law regression of log,, variance on log,, rnean of T. 72 tabaci larvae (A) and adults (B) and A. cucumeris (C) in greenhouse cucumbers.

Fig. 6. The relationship between proportion of leaves infested with 2 1 and s 5 thrips larvae (A) and thrips adults (B) in greenhouse cucumbers.

Fig. 7. Binomial sequential sampling plan for T. rabaci larvae (A) and adults (B) in greenhouse cucumbers.

Fig. A.1. Mean captures of T. tabaci lawae and adults per cucumber le& in Harlow research greenhouse 1 (A) and 2 (B) - NSAC, Fa11 1996.

Fig. A.2. Mean captures of T. tabaci larvae and adults per cucumber leaf in Harlow reseorch greenhouse 1 (A) and 2 (B) - MAC, Fa11 1997.

Fig. A.3. Mean captures of T. tabaci larvae and adults per cucumber leaf in Stokdijk Greenhouses crop 1 (A) and 2 (B) - Beaverbrook, Fa11 1997.

viii

Various aspects of integrated pest management (PM) for onion thrips, Thrips tabaci Lindeman [Thysanoptera: Thripidael, in greenhouse cucwnber (Cucumis sativus) were studied. An experiment was conducted to determine and compare the predation rate of the gravid fernales of two predatory mites, Amblyseius cucumeris (Oudemans) and Iphiseius degenerans (Berlese) [Acarina: Phytoseiidae], on first and second instar onion thrips larvae on cucumber cv 'Jessica' leafdisc arenas. Significantly more first instar thrips (3.6 I 1-13) than second instar thrips (3 -4 * O. 8 1) were preyed upon by the gravid f d e mites. There was no significant difference between the predation rate of A. cucumeris or I. degenerans for d e r first or second thrips larval instars suggesting that these mites have an equal potahal to control onion thrips in greenhouse cucumbers.

The effect of the addition of the volatile chernical am-actants, p-anisaldehyde (4- methoxybddehyde) and ethyl nicotinate (3-pyridinecarboxylic acid), to commerciaiiy available blue and yellow sticky traps was examinai for adult onion thrips in greenhouse cucumber production. Two studies were conducted at separate locations: a commercial cucumber greenhouse and two research greenhouses. There was no interaction between scent and colour at the commercial site. Anisaldehyde significantly increased captures by a fàctor of 1.7 over ethyl nicotinaie and 1.2 times more than the control. Ethyl nicotinate did not increase captures over the control. Blue sticky traps caught mice as many thrips as yeliow traps, and yellow caught 5 .O times more than clear. This indicates that blue is the colour of choice for capturing ad& onion thrips in greenhouse cucumber production. in the research greenhouses, there was a simiificant interaction between ethyl n i c o ~ a t e and yellow traps. This combination caught 1 -6 times as many thrips as the next most attractive combination, non-baited blue traps. The factors and thrips behavioural mechanisms for host-fin- cues affecàng these differential results between locations are discussed. Captures on the non-baited traps of the commercial site data were correlated with leafsamples taken within one metre of the trap to determine if stic- trap captures were indicative of infestation levels on the leaves. There was no correlation between either blue or yellow sticlcy traps, and a moderate correlation between clear traps and adult thrips captures. Captures of Orius predators on scented sticky traps were also examined. Scent did not increase captures of these predators on sticky traps. As reflected in the thrips r&, blue sticky traps were 1.4 tirnes more attractive than yellow and 15.0 tirnes more than clear, indicating captures were not incidental and bio1ogica.I control agents rnay also be susceptible to mes similar to those of their prey.

The spatial distributions of Thrips tabaci larvae and adults and the predatory mite, Amblyseius cucumeris, on greenhouse cucumbers were caiculated using w o variancemean models: Iwao's patchiness regression (PR) and Taylor's power law (TPL). Both models determineci thrips larvae and adults to be contagiously distnbuted in greenhouse cucurnbers Mth a density contagious coefficient, b, (IPR) and index of aggregation, 6, (T'PL) significantly p a t e r than one. The predatos. mites population studies yielded ambiguous results for b, according to IPR they were not contagiously distributeci tvhile T'PL revealed they were. The index of basic contagio~ a, (PR) reflected that aggregates were the basic component of thrips larvae populations only. Based on these parameters, iwao and binomial sequential sampling plans were developed for thrips larvae and adults on greenhouse cucumbers. The economic threshold was estimated at 75 percent of a working economic injury level of 9.5 lamie and 1 -7 aduhs and calculated to be 7.1 and 1.3, respectively . ï h e ensuing maximum sample number to be taken was 66 for larvae and 46 for adults for Iwao's plan. The maximum sample number for the binomial sequential sarnpling plan was 3 9 for larvae and 67 for adults. These plans need to be validated before they can be applied to commercial production.

d

EIL

IPR

m

R'

cl2

S!

SD50

TSWV

TPL

VPD

2

List of Abbreviations and Symbols

scaling factor related to sample size (Taylor's power law)

index of basic contagion @-intercept of Iwao's patchiness regression)

active ingredient

index of aggregation (Taylor's power law)

density coatagious coefficient (slope of Iwao's patchiness regression)

cultivar

confidence interval of estimated mean density of Iwao's sequential sampling

economic injuq level

economic threshold

precision level of maximum sample number of binomial sequential sampling

integrated pest management

Iwao's patchiness regression

population mean

coefficient of determination

population variance

sample variance

saturation deficit

tomato spotted wilt virus

Taylor's power law

vapour pressure deficit

Lloyd's mean crowding index

ACKNOWLEDGEMENTS

I would Iike to thank the foiiowing orgaaizations and individuals for their support and

assistance during my research and the preparation ofthis thesis: Nova Scotia Department of

Agriculture and Marketing for hanciai support; Production Technology and Peter Stokdijk

for use of their greenhouses; Brian Toms and Arthur Haskins for their technical advice;

Wendy Hoiiis, Karen Dickie, and Pam Sandeson for data collection; Sawler Gardens for use

of their fields for thrips collection; Dr. Tessema Astatkie for statistical assistance; the faculty

and technical staff of the Biology/Environrnentai Science Department and Plant Science

Department of the Nova Scotia Agriculturai Coiiege for their technicai advice and assistance;

my employer, Dr. Gary A t h for allowing me the time to complete my thesis.

1 gratefLIly acknowledge my supe~sory cornmittee: Supervisor Dr. Jean-Pierre R. Le Blanc,

and cornmittee members Dr. A. Bruce Gray and Prof Lloyd R Mapplebeck for their guidance

and encouragement throughout the course of this project.

Finally, 1 would like to thank my fnend Doug, for his support and encouragement.

1.1 Introduction

Greenhouses cover approximately 150,000 hectares worldwide. They offer an

excellent opportunity to grow high quality products in large quantities within a srnail surface

area (van Lenteren 1990). In Canada, there is an ùicreasing trend in total greenhouse area,

product volume, and subsequentiy, revenue. Greenhouse cucumber production alone

increased fiom 73 5 tomes of cucumbers in 1995 to 1059 tonnes in 1996 resulting in a $10.20

million increase in revenue in that time period. This trend is reflected in Nova Scotia

greenhouse cucumber production where total revenue increased fiom $1.40 million in 1 995

to $1.98 million in 1996 (Statistics Canada 1998).

Unfortunately, greenhouses also provide optimal conditions for development of pest

insect and mite infestations due to maintenance of warm temperatures, high humidity,

conEïned growing are& and monoculture. Foliage and fiuit feeding arthropods are among the

most diverse and destructive pests occurring in vegetables (Trumble 1994). Abundant

international trade has resulted in the unintentional import of numerous pests (van Lenteren

1990). One of these pests imported fiom Europe was the onion thrips, M p s ~abaci

Lindeman [Thysanoptera: Thripidael. In North America, the onion thrips has a host range

of over 300 plant species and is a recognized economic Pest of many crops. Thrips initially

invaded greenhouses fiom nearby fields, however they have successtùlly adapted to

greenhouse conditions (van Rijn et al. 1995). Damage due to thrips is caused by feeding on

the leaves and fruits or through transmission of virus particles. Due to its high capacity for

population growth, broad host range, and potential for pesticide resistance, the Pest status of

2

the onion thrips remains a concem. The importance of onion thrips in greenhouses is

enbanced where insecticidal pressure applied against other pests is relaxed because of

biological controi programmes (Binns et al. 1 982). Research and development have

successfully identified biological control agents for multiple classes of greenhouse pests such

as whiteflies, spider mites, fungus gnats, and aphids. Natural enemies developed for thrips

have been incoosistent, and onion thrips are frequentiy reported to escape these naturai

enemies and reach epidemic levels. This ultimately requires the use of pesticides which are

seldom effective or compatible with biological control agents, thereby disrupting the natural

enemy complex established in the greenhouse.

The overaii objective of this study was to develop various aspects of an integrated

pest management (IPM) programme for onion thrips in greenhouse cucumber production.

1.2 Thrips tabaci Lindeman

1.2.1 Taxonomy and synonymy

The onion thrips, nrzps tabaci Lindeman [Thysanoptera: Thripidae], has had a stable

taxonomie position since it was described by Lindeman (Daniel 1904). The onion thrips

features two subspecies, T. tabaci tabaci and T. tabaci cornmunis (Schliephake and Klimt

1979).

1.2.2 Origin and geographic distribution

The onion thrips originate Grom Central Asia (O'Neil1960) and are now known world

wide. These thrips were introduced to North America fiom Europe in the early 1900's (Anon.

3

1968) and have since become established as serious pests (Edwards and Heath 1961).

1.2.3 Host plants

7hr@s tabaci has a number of ecotypes eacb of which is polyphagous and can be

hosted by a wide range of plants (Tommasini and Maini 1995). They infest several species

of various families such as the foiiowing reported by Sakimura (1932):

Amaranthaceae Compositae Gramineae Maivaceae Rubiaceae Araliaceae Convulvulaceae Labiatae Nyctaginaceae Solanaceae B oraginaceae Cruciferae Leguminosae Oxalidaceae S terculiaceae Bromeliceae Cucurbitaceae Liliaceae Phytolaccaceae Umbelliferae Caryophyllaceae Euphorbiaceae Lythraceae Portulacaceae Verbenaceae

Practicaüy all plants in the greenhouse are attacked by the onion thrips. Amongst

those that suffer rnost severely are cucumbers, roses, carnations, and chrysantbemums

(Metcalf and Metcalf 1 993).

1.2.4 Life cycle and behaviour

The onion thrips undergo a six stage life cycle: the egg, two larval instars, prepupa,

pupa, and adult (Palmer et al. 1989). The delicate, tramlucent white, reaiform (Bailey 1933)

egg is laid just under the surface of the leaf and is large (about O. 1 1 mm wide and 0.22 mm

in length) in cornparison to the femaie body (Salas 1994). As the embryo develops the outline

of the lama, in particular the eye spots, is visible through the thin chorion and le& surface.

Upon hatching, the shell splits lengthwise and the lama pulls itself out of the shell. Feeding

commences immediately (Bailey 193 3).

4

The newly emerged lama is translucent white in colour with comparatively large head,

antennae, and legs. It measures 0.29 mm (Bailey 1933) to 0.34 mm (Salas 1994) in length.

The body of the second instar is yellowish with variable crimson blotches. This instar

averages 0.96 mm in length (Bailey 1993). Both 1arva.l instars have two compound eyes, no

ocelli, short eight segmented antennae, three pairs of legs, and no wing pads (Salas 1994).

The onion thrips larvae are thigmotactic, resting in veins or other distinct grooves close to

their emergence sites formllig a colony (Salas 1994). The lawae drop fiom the leafto pupate

on the soil surface or in natural crevices in the soil (B im et al. 1982).

The prepupal stage is considered an intermediate stage between the lama and the true

pupa where feeding ceases, respiration is retarded, but no cocoon is formed. Its shape

resembles that ofthe mature larva with a general body colour that is yellowish to orange with

translucent white legs and antennae and an average size of 1.01 mm. The eyes are dark

orange with a few visible facets. The wing pads are very short, translucent white, with a few

scattered setae.

The pupal stage can be divided into two penods: early and late. The pupae are

sensitive to light and are also thigmotactic, remaining in enclosed surfaces such as cracks or

crevices, during pupation (Salas 1994). In the early stage the colour is orange and the prior

red markings are reduced. The body appears shorter and stouter than the prepupa averaging

0.86 mm in length. The antennae are folded back over the head. The eyes are much larger

and darker red and the three orange coloured oceUi are visible between the eyes, their location

forming a triangle. The wing pads are translucent white, have scattered setae and are

extended to the sixth and seventh abdominal segment. The ovipositor is clearly demarked in

5

the female so the sexes can be readiiy distinguished (Bailey 1933). In the late stage the

antennae corne forward, straighten out, and darken. The abdomen begins to exhibit

considerable pigmentation, the wings within their cases darken, followed by the legs, then the

head, and lastly, the thorax. This reticulation can be seen on the surface of the body.

Moulting and emergence of the adult occur at this point (Bailey 1933).

Newly emerged addts are inactive for about twenty-four hours. Feeding begins on

the second day and mating occurs after feeding. Of the adults, the male is smaller than the

female and has a tapering abdomen (Bailey 1933). Copulation begins on the second day after

emergence (males) whiie oviposition begins on the third day (females) (Bailey 1933). Peak

oviposition is within the first ten days (Salas 1994) and may occur anytime throughout the

day. Mean fecundity is 37 eggs per female, or 1.9 eggs oday-'-fernale-'.

Reproduction may be sexual or panhenogenic. In the wild mating is required to

maintain some percentage of males. A large number of males is not required because they are

very promiscuous and fertilize a large number of females (Bailey 1933). Although T. tabaci

originated in the outdoors, van Rijn et al. (1 995) suggests a specific greenhouse population

has evolved. in European and British greenhouses, males are virtually unknown (Morison

1957). Parthenogenesis is the most common reproductive strategy employed by imported

species as it is the easiest and most direct method of reproduction (Bournier 1983). Onion

thrips exhibit thelytokous parthenogenesis in which unmated females produce only female

offspring (Lewis 1973). The advaotage to this is two fold: first, the population growth is

promoted by all female offspring; secondly, at low densities the population growth is not

limited by the availability of males (van Rijn et al. 1995). The onion thrips hibernate as adults

6

in several types of vegetation and residues outside the greenhouse. in late October to early

November egg laying ceases. folfowed by copulation, and ka l ly feeding. There is significant

mortality during the five montb duration of hibernation (North and Shelton 1986). When

females emerge in the spring they are capable of oviposition and can reenter the greenhouse.

Duration of the We cycle is iduenced by several factors including temperature,

humidity, sometimes expressed as vapour pressure deficit, and suitability ofthe plant cuitivars

grown (Men 1995). In a life history study of onion thrips on cucumber leaf discs at 2S°C,

T. tabaci bad an intfinsic growth rate of 0.17 thrips daily (van Rijn et d. 1995). The life

stadia (in days) were as follows: egg (3.9), first instar (2. l), second instar (3.2), prepupa

(1.3), and pupa (2.4). From egg to adult, a total of 12.9 days is required. Including a

preoviposition penod of 1.9 days, the period fiom egg to egg was 14.8 days. Atmospheric

humidity is of great importance to the survival of the larva. At high temperatures, a relatively

high atmospheric humidity (RH > 70%) is necessary if the insects are to reach maturity.

The median adult life stadium of the onion thrips is reported to be approximately 12

days. Total mortality fiom egg to adult in these trials was nineteen percent with more than

half occumng during the second larval instar (van Eüjn et al. 1995). Temperature was the

primacy factor affecting mortality - at either end of the temperature range mortaiity increased.

Below 1 2 OC mortality increased due to the extended developmental time. Maximum survival

occurred at 23°C (van Rijn et al. 1995). During hot months with air temperatures

approximateiy 30" C, populations in the greenhouse may increase steadily for six weeks (Allen

1995). Consequently, i7 tabaci can have up to ten generations a year (Lewis 1973); however,

five to eight are generaiiy encountered (Metcalf and Metcalf 1993).

7

The flight ability in the thrips is not weli developed @ d e y 1933). It is assisted by the

curling of the abdomen upwards to separate the wings and hopping (Allen 1995). Thrips'

rhythm of flight tends to be random in the greenhouse and is affected by air currents and

structures, thus resulting in the development of areas of high densities commoniy referred to

as hot spots (Men 1995). Thrips undergo two major dispersai periods a day, one in early

morning and the other in late aftemoon. These activity periods are naturd daily ones

(circadian rhythms) and are not inûuenced by the presence or location of the crop nor

cropping practices. Flight activity of thrips may also be irifluenced by environmental factors.

Shipp and Zhang (1997) studied the effect of temperature and vapour pressure deficit on

flight activity of western flower thrips, Frankliniella occidentalis (Pergande) (Thysanoptera:

Thripidael. Humidity expressed as vapour pressure deficit (VPD) is defined as a measure of

the difference in air moisture between the saturated and acnid water vapour pressure at a

specific temperature (Clarke et al. 1994). M e r three hours, thrips exposed to 23 "C and

VPD of 1.42 kPa showed the greatest flight activity.

Thrips populations are not independent Erom each other (Suman et al. 1980). This

behaviour, called contagion, is a departure fiom a random to a clumping distribution where

the presence of one insect in a unit increases the probability of the occurrence of other insects

in that unit. These factors may explain the rapid reinfestation observed when thrips are

removed or new plants are introduced at any time (Men 1995).

1.2.5 Economic importance

Feeding by thrips can cause direct and indirect damage. Indirect damage arises fiom

feeding on parenchyma of leaves and subsequent reduction in photosynthetic ability of the

plant. Ln cucumbers, direct damage results fkom feeding on the b i t causing the fruit to curl

and rendering it unmarketable. Darnage fkom oviposition and hat ching is minimal (B ailey

1933).

The economic importance of thrips is dependent on their unique, asymmetrical,

piercing-sucking mouthparts (Lewis 1973). Both feeding damage and virus transmission are

facilitated by the feeding apparatus. Two types of feeding are identified in the thrips: shallow

feeding which involves applying the mouth cone directiy to the le& flower, or b i t surface;

and, deep feeding, which involves sucking Sap through the stylets (Mound 1971). When

feeding is initiated, the mouth cone is applied to the leaf surface. The single lefi mandibular

stylet pierces the membrane, then it is withdrawn rapidly and replaced by the paired maxillary

stylets which form a tube, i.e. 'tongue and groove' system, which penetrates the deeper ce11

Iayers. This tube is supported and protected inside the mouth cone through which fluid passes

into the food canal. Liquid is drawn up the food canal by the pumping action of the cibarium

which has a valve at both ends to regulate food intake. To aid in feediag, saliva is produced

by the labial glands to partially digest food prior to cibarial pumping (Chisholm and Lewis

1984). Saliva could also serve to lubricate the mandible and stylets by flooding the area

around the point of insertion, preventing leakage of plant ce11 contents, and providing

adhesion of the mouthcone to the surface and between the protracted stylets (Heming 1 978).

Probing and feeding removes surface wax fiom the leaves, exposing the cuticle; thus,

9

mouthparts were formerly referred to as rasping-sucking (Lewis 1 973). With cibarid

pumping at a rate of two to six pulsations-second-', thrips cm consume sap at a rate of

8.5x10-' p~minute- ' or 12.5 percent of their body weight per hour (Chisholm and Lewis

1984).

Feeding of thrips is not conhed to leaves alone, but to al1 vegetative parts of the

plant. In cucumbers, they concentrate on rapidly growing tissues of the leaves, flowers and

h i t s (Ananthaknskoao 1979). Feeding in the growing tips and flowers causes defomed

leaves and t'niits, while feeding on the leaves and fniits causes tissue scarring. Superficial

appearance of feeding damage is silvering or streaking and distortion. The cells beneath the

pierced epidermal ceiis are usually completely emptied of their contents. Under intensive

feeding many mesophyll cells are totally destroyed with whole chioroplasts, measuring

approximately 6 p m in diameter, ingested through the 1 pm diameter mouth cone. Other cells

suffer extreme plasmolysis due to contorted grana sacs and the formation of starch grains

caused by desiccation (Chisholm and Lewis 1984). The epidermal cells above this interna1

damage are collapsed, the out er cuticle is wrinkled and appears silvery before turning yellow

and brown. This superficiai scarring, distortion, and silvering is attributed to shrinkage of the

epidemal and mesophyll cells and the penetration of air following the removal of sap.

Necrosis and discolouration spread as punctured areas coalesce (Chisholm and Lewis 1984).

Cucumbers c m tolerate a leaf area loss up to thirty percent before yield is &ected (Hussey

and Parr 1 989). If lefi untreated, thrips cm cause a leaf area loss of twenty percent within

nine weeks due to thrips feeding (Jacobsen 1995). Aithough it is necessary to gain control

of thrips quickly, thrips populations need only to be maintained at levels sufficient to avoid

10

economic losses. Damage is accentuated by stress which retards growth (Rossiter 1980).

Disease transmission is another f o m of indirect damage. Feeding wounds caused by

thrips enhance entry and development of diseases by providing alternative penetration sites

(McKenzie et al. 1993). Fungi c m either invade the damaged tissue or grow on the outer

surface nourished by thrips faecal deposits (Chisholm and Lewis 1984). Onion thrips are

known vectors of virai and bacteriai diseases (Lewis 1973). This is particufarly dangerous

because they have an extensive host range from which diseases can be retrieved. Viral

diseases, such as tomato spotted wiit virus (TSWV), are transmitted intemaily, i.e. feeding

larvae acquire the v h s and transmit it through the saliva of the ensuing adult. Of onion

thrips, only the subspecies T. tabaci tabaci is reported to transmit TSWV (Zawirska et a/.

1979) and there have been no reported incidence of transmission of TSWV by onion thrips

in Canada (Paliwal 1974). Transmission of virus particles is also achieved extemally via

infected pollen, where the thrips cany the pollen to susceptible plants and transmit the disease

through pollination. Continuous redistribution of thrips in the greenhouse accounts for rapid

dissemination of the virus they transmit (Allen 1995).

1.3 Integrated Pest Management (TFM)

There is a need for a more holistic approach to Pest management policy which will

effectively and safely respond to the need for a wholesome and adequate food supply (Wilde

198 1). One control strategy used singly is not likely to disadvantage thrips sufficiently to

adequately manage their number and virus spread in greenhouses throughout the year (Allen

1995). Thus, an integrated approach with less dependence on toxic pesticides is necessary

11

and becoming accepted practice. Integrated pest management (IPM) is a complex approach

to pest control which incorporates aii available tactics to reduce pest populations to an

acceptable level in a cost effective and environmental1y rational manner (Murp hy et al. 1 99 5 ) .

Biological control techniques play a crucial part in IPM programmes, but, these strategies also

draw on physical and cultural measures. If these tactics fail, pest control can still be

accomplished through effective and seleetive use of target-specific insecticides to maintain

fluctuating pest popuiations below economic injury levels. A 1994 survey indicated that P M

is ofien used in greenhouse edible crops in Canada and north-west Europe, but is much less

cornrnon elsewhere (Jacobson 1997). Limiting factors preventing the adoption IPM practices

in other areas, particularly hotter climates in the northern hernisphere, include difficulties in

combatting sudden and huge pest invasions, speed of development of crops, and the fear of

virus infections spread by insects. In Austraiia and Japan, the limiting factor is the

unavailability of biologicai control agents (Jacobson 1997).

1.3.1 Monitoring

Due to their cryptic behaviour, thrips are difficult to detect in the initial stages of an

infestation, thus timing of control measures is difficult. This is hrther complicated in

biological control situations by the slow response time of beneficials to become established

and achieve control (Frey et al. 1994). Systematic monitoring of the pest and determination

of its damaging densities is one of the most important factors determining the long term

success of plant protection in greenhouses (Frey et al. 1994). Monitoring provides an early

means of detecting the pest, enabling prompt intervention with appropriate control measures

12

before crop yield and quality are affecteci. It is therefore important to develop and optimize

monitoring tools suited for plant producers. The effectiveness ofthese monitoring tools relies

on protocols for deciding on the need for some management intervention based on an

assessrnent of the state of the Pest population, and ideab its natural enemies (Binns and

Nyrop 1992). These protocols, or contra1 decision rules, consist of at least two components,

and possibly a thid.

The first component is a procedure for assessing the density of a population. In

greenhouse cucumber, the recommended sarnpling methods for monitoring the population

densities of tbrips are blue sticky traps (Gillespie and Vernon 1990) and Ieaf samples (Steiner

1 990). Coloured sticky traps are mainly used to detect potentially h a d l populations early

enough to allow preventive action. Yellow sticky traps are ofien used in greenhouses because

a wide range of pests is attracted to this colour (Shipp 1995). Blue sticky traps are specific

to thrips species in greenhouse cucumbers and are well suited for long term monitoring (Frey

et al. 1994). The number of thrips on such traps is not a reliable indicator of other thrips

populations on nearby plants or the amount of darnage that has occurred (Allen 1995). It

mainly indicates relative population level and overall trends in the population, whether it is

increasing, decreasing, or stable. This information helps growers determine action thresholds,

which varies not o d y within crops but within varieties growa. At this point the grower can

establish the base upon which to build aa efficient control. Management benefits include

compatibility with biological control techniques and the ability to assess the effectiveness of

control treatrnents. Yearly trap sarnpling records provide an indication ofwhen to expect Pest

invasions in subsequent seasons and Wtely peaks in activity (Jackson and Scopes 1993).

13

Because management tactics are directed towards the injurious insect, assessrnent of

Pest infestation usudy requires obtaining actual counts of the pest. Density (a mean number

of insects inhabithg a plant) or intensity (a proponion of plants infested with insects) are

typically used as the injury unit (Brewer et al. 1994). Plant age does not s e c t thrips

reproduction, whereas Leaf age is strongly correlated with resistance of mature plants in the

greenhouse (de Kogel et ai. 199%). In dual choice assays, thrips preferred younger leaves

over older leaves for oviposition, indicating that differences in leaf suitability are an important

factor in determinu>g thrips distribution on cucumber plaats (de Kogel et ai- 1997a). Steiner

(1 990) studied distribution characteristics and established samphg procedures for western

flower thrips and its predatory mites on greenhouse cucumbers. Mature, middle leaves were

determined to be the most consistent reliable sarnpling unit for determining thrips densities.

The second component is an economic threshold which is derived from an

understanding of the economic injury level (EL), which is a fùnction of yield and market

value of the cornmodity per production unit, cost of managing the insect per production unit,

the injury potential per insect infestation unit, management tactic effectiveness, and the

environmental cost-benefit associated with the management tactic (Pedigo and Higley 1 992).

Steiner (1 990) determined a working E L for cumulative numbers of western flower thrips

in greenhouse cucumber of 1.7 adults or 9.5 larvae per le& sample. Finaliy, a phenologicd

forecast is often necessary to determine the appropnate tirne to assess population densities

(Logan et al. 1979).

1.3.2 Control Measures

1.3.2.1 Cultural and physicat control

Management strategies and conditions within the greenhouse must be such that they

enhance crop productivity, foster the activity of biological control agents, and hinder the

development of pests and diseases (Clarke et al. 1994). These methods are considered

preventive rather than curative. Vigorous crops are able to tolerate much higher populations

of thrips (Rossiter 1980). Sanitation measures begin in and around the perimeter of the

greenhouse. Crop residues should be rernoved well away and downwind fiom the

greenhouse. Weed control inside and outside the greenhouse is essential to eliminate altemate

bosts and overwinteriag sites of onion thrips (Anon. 1993). To reduce the chance of Pest

carryover, thoroughiy clean and disinfect empty structures, aüow sufficient tirne between crop

removal and replanting, inspect new stock before introduction in the greenhouse, and avoid

ovenvintering garden or house plants in the greenhouse. Sowing or replanting when another

crop is in pronmity should also be avoided. Optimum nutrition produces sturdier plants that

are better able to withstand attack. Monitor potassium, nitrogen ratios, and calcium levels

because they are associated with plant resistance.

Environmental factors can alter host susceptibility to attack, pest population levels,

host or pest phenologies, or pest occurrence in a host (Higley and Peterson 1994).

Manipulation of the greenhouse environment is one of the most undemtilised methods of

discouraging thrips establishment in the greenhouse- Microenvironments exert a significant

influence upon the behaviour and population dynarnics of arthropods (Willmer 1982) and the

greenhouse represents a plant production system where critical control of environmental

15

parameters is feasible (Shipp and Gillespie 1993). Temperature and humidity or vapour

pressure deficit have been s h o w to affect thrips development and flight activities (Shipp and

Zhang 1997) and influence effectiveness ofbiological control agents in the greenhouse (Shipp

el al. 1996). Growth, reproduction, and s u ~ v a l of insect pests can be positively correlated

with nutrients in their food. Some plant nutrients are components of antifeedants and toxins

or are present in forms that are not readily utilizable by the inseas (Hunt et a[- 1998). In

modem hydroponic systems, control of plant nutrition and crop growth is a promising method

for controlling greenhouse pests.

Many consider plant resistance a separate category in control measures, but the choice

of crop variety grown is a management decision. Host-plant resistance offers the most

potential for future advancements in Pest management (Anon. 1993). Varietal resistance to

thrips has been found in peppers, letnice, and tomatoes (Mollema and Cole 1996), cabbage

(S helton et al. 1 983), chrysanthemums (De Jager et al. 1993), and roses (Gaum et al. 1 994).

In greenhouse cucumber cultivars, resistance linked to amino acid content has been

recognized. In greenhouse trials, these cultivars lived five weeks longer than registered

cultivars before succumbing to attack. Uofortunately, these cultivars are unable to compete

with market leaders in yield or quality (Long 1995). These cultivars provide genetic material

for future breeding programmes.

Screening the vents is a physical control that may be employed. However, the mesh

size required to exclude thrips would be so fine, it would become expensive and decrease air

flow (Anon. 1993). Another solution is to suspend polyfilm several inches above the crop in

strategic locations (Allen 1995). When placed between benches and dong walkways it can

16

alter thrips dispersal fiom areas of hi& populations or virus infected sources. This is a quick

setup and allows tirne until the main treatment c m be appIied or infected plants are removed.

1.3.2.2 Biologicd control

Acceptance of biologicai control as a serious control strategy depends on good public

relations and education. The beliefs that the use of natural enemies creates new pests and

biological control agents are unreliable arose mainly as a result of strong pressure to market

natural enemies before they were fiiily tested. The methods of researching biological control

is quite different from that of pesticides but not as expensive as commonly believed. Cost

benefit analysis of research and development indicates a 30: 1 ratio for biological control and

a 5: 1 for pesticides (van Lenteren 1990). Many growers have yet to adopt nonchernical

approaches to greenhouse crop management because biological and environmental control

methods are thought to be more d8icdt to manage and possibly more expensive (Clarke et

al. 1994). Incentive for use by the grower is that the cost of the recommended rate of

inundative releases of predators for thrips in greenhouse cucumbers is siniilar to pesticides

(Gilkeson 1990).

Biological control agents may be predators, parasites, or pathogens. A predator feeds

on its prey by catching it but othenvise lives independent of it. Use of the predatory mites has

become an accepted method of controllhg thrips (Ferguson 1995). Amblyseius nrnrmeris

(Oudemans) [Acarina: Phytoseiidae] is the most commonly used predator in greenhouse

cucumber production. With the development of a controiled release system which consists

of a bran based culture in a waxed paper sachet ofspecific porosity, A. cucumeris establishes

17

rapidly with better distribution, and continues to breed and emerge fiom packs for at least six

weeks (Bennison and Jacobson 1991). A. niamteris enters diapause at a critical daylength

of 1 2.5 hours with 22 C day temperature and 1 7 O C night temperature (Gilkeson et al. 1 990).

Diapause affects only female predators by baiting production of eggs (Rodriguez-Reina et al.

1994). This not only causes a loss of reproduction, but also reduces the fùnctiooal response

since gravid females are the most effective predatory stage (Shipp and Whitfield 1991). A

nondiapause strain of A. cucumeris was selected by van Houten et al. (1995) in which

diapause was reduced fiom 41 to O percent in ten generations without a reduction in

predatory performance or oviposition rate. Over an eighteen month period, the

nondiapausing strain of A. cucumeris did not enter diapause under diapause inducing

conditions which indicates that it was a stable trait for this predator.

/phise jus (=Amblyseius) degenerm (Berlese) [Acarina: Phytoseiidae] has also shown

strong predaceous capabilities in the laboratory (Eveleigh and Chant 198 1) and its eggs are

also more resistant to drought than those of A. cucumeris (van Houten and van Stratum

1993). When A. cucumeris was tested in the greenhouse against the nondiapausing, drought

resistant strain of 1. degenerans, the latter one was found to have a greater impact on the Pest

population than A. cucumeris (van Houten and van Stratum 1993). 1. degenerans is

considered one of the most promising candidates for biological control of thrips under

conditions of low humidity and short day length (van Houten et al. 1995). The drawback is

that these mites are only predacious on larvae.

Imundative applications of the minute pirate bug Orius insidiosus (Say) and 0.

tristicolor w t e ) weteroptera: Anthocoridae] can be used when thrips populations exceed

18

the control capacity of resident predatory mites. Both nymphs and adults are voracious

predators on adult thrips as well as larvae. Adult minute pirate bugs are known to consume

one lama per hour, and nymphs are even more voracious (Bailey 1933). Orius species are

cannibalistic in crowded conditions and until 1989 they were difficuit to mass rear (Gilkeson

et ai. 1990). Orius spp. are aiso generalist predators and it was initially thought that thrips

control may be impaired ifpirate bugs prey on the predatory mites in lieu of thrips. On single

leaves in srnall cages, mortality of A. cucumeris due to 0. tristicolor was inverse1y

proportional to thrips density. Thrips mortality due to O. tristicolor was unaff'ected by A.

czicz~rneris; therefore, it may be concluded that these biocontrol agents are compatible

(Gillespie and Quiring 1992). In wee t pepper crops, &er 1. degenerms displaced A.

mcumeris, Orius spp. appeared in the crop simultaneously with 1. degenerms without

compromise to thrips control or predatory performance (van Houten and van Stratum 1993)-

Soil-dwelling predatory mites such as Hypoaspis spp. (Bailey 1933) and Geoiaelaps sp. nr.

aceieijèr (Canestrini) (Gillespie and Q u i ~ g 1990) contribute to biological control of thrips

by preying on pupating thrips in the soi1 thereby reducing emergence of adults.

Parasites kill their host by living in or on the host body and therefore depending on

it for completion of their development (Anon. 1993). Thrips have s h o w variable responses

in their susceptibility to hymenopterous parasites particulariy of the Ceransius species

&oomans and van Lenteren 1990). Eulophid, or minute wasps, lay their eggs in the young

thrips larvae. If the second larval instar is parasitized, the wasp cannot develop successfully

but will shorten the adult lie of thrips (Lewis 1973). Other natural enernies such as

nematodes, mirid bugs, spiders, syrphid wasps, pseudoscorpions ail prey on thrips in nature

19

but have not been developed as applied biological control agents.

Few serious attempts have been made to develop fungi as biological control agents

for onion thnps (Gillespie 1986). A pathogen must be ingested to be effective; therefore,

method and timing of application are crucial (Splittstoesser 198 1). T. tabaci has been sbown

to be susceptible to several species of fùngi. In laboratory experiments, Beauvaria bmsiurm

and Melarhizium anisupiiae kiiied a population of thrips witbin four days. in that same time

period, Veriicifiiurn iecmii killed 85 percent of the population. In greenhouse experiments,

?? leca~tii reduced thnps populations on cucumbers (Gillespie 1986). Entomophrhora

sphaerosperma and E. tarichium have also caused epizootics in thrips (Binns et al. 1982).

Disadvantages to many of these entomopathogens are that they are generalists and will also

kill natural enemies used to control other pests in the greenhouse (Binas et al. l982), and they

are inhibited by fiingicides (Shipp et a/. 1998).

Availability and quality of biological control agents are controlled by companies

producing and ~ p p l y i n g growers. Packaging and transport to the growers can adversely

affect survival or efficacy of the natural enemies. Healthy control agents in adequate slow

release pacs should contain a developing culture and be fiee of mould or ammonia odours

(Ferguson 1995).

Detailed information on thnps Me bistory is important to understand biological control

by natural enemies. The time spent in stages vulnerable to natural enemies can be decisive

for predator impact (van Rijn et al. 1995). Biologcal control involves more than just

releasing natural enemies. It also requires strict monitoring to determine how much natural

enemies are contributing to control (Anon. 1993). Hence, in biological control situations,

two crops are being produced - the plant and the beneficial insects (Murphy et al. 1995).

1.3.2.3 Chemical control

Adults are the preferred target when using insecticides because they are easier to bit

than larvae with the mist sprayers and are also generally more sensitive to the products.

Chemical control is seldom effective because thrips are protected in the growing tips of leaves

and flowers and it is difncult to expose thrips to pesticides (Shipp et al. 1990). Greenhouse

cucumbers are more commonly grown as cordons rather than lateral training systems and

therefore produce a dense mass of foliage which is difficult to spray, while coverage by

fogging and ultra low volume methods is poor (Hussey and Scopes 1985). Timing of

applications is essential in maximizing the effectiveness of the pesticide. if an insecticide

application is required, the temperature and VPD in the thrips vicinity should be 25 O to 30°C

and greater than 1.2 kPa for a period of up to three hours f i e r the insecticide is applied

(S hipp and Zhang 1998). When the greenhouse climate is modified for ody three hours, crop

production is not afEected and the insecticide is more effective by the immediate contact with

the pesticide during fiight and when thrips land on the upper leaf surface where rnost of the

deposit occurs. Stornach poisons distributed on leaf surfaces are not effective because thrips

only feed on superficial plant tissues. Systemic insecticides are even slower because thrips

do not feed on phloem tissues where the insecticide is present (Shipp and Whitfield 199 1).

If a spray is applied in response to an increase in thrips detection during the previous week,

spraying occurs after most of the egg Iaying by new adults is completed (Allen 1995).

2 1

Chernicals as a sole approach to thrips control is becoming either less effective or less

acceptable environmentally or by the public (Lewis 1997a). In an IPM programme, the use

of pesticides remains necessary for two reasons. First, there are a number of pests and

diseases that camot yet be controlied by their natural enemies. Second, pesticides mppon

biological control in case the predators or parasites are temporarily unable to control the Pest

(van der Staay 199 1).

Pesticide appfication in the greenhouse results in residual effects that are different

tiom those observed in the field and therefore subrnitted to more stringent pesticide

regulations. Temperature, relative humidity, and air transfer in an enclosed environment may

affect pesticide movement and breakdown rates (Matteoni et al. 1993). In addition to the

obvious consideration of environmental and worker safety, and the development of resistant

Pest populations, a major concern of pesticide use in greenhouses is the effect on biological

control agents. It is important to know the life cycle and feeding habits ofthe natural enemies

because pesticides af5ect predator and parasite populations dinerently. Predators are most

susceptible because they consume prey, accumulating residues over time and t hereby

becoming less efficient due to reduction of their populations. Parasites feed on only one host;

however, as they emerge fiom their cases, the pupae may be poisoned as they eat their way

out of the host. Soi1 dwellers are more susceptible to soil drenches, while leaf dwellers are

more vulnerable to sprays (Matteoni et ai. 1993).

The effect of chernical pesticides on naturai enemies also depends on the type of

pesticide used. Broadspectrum insecticides adversely afTect more naturai enemies than

selective ones. Systemic insecticides have a longer residual effect than nonsysternic. Most

22

fungicides are compatible with biological control programmes, except benzimidazoles (e.g.

benomyl, cardendazim, and thiophanate methyl) which adversely affect the egg laying ability

of predatory mites (Matteoni et al. 1993), and pyrazofos and triforine because their LD, are

at the same level as the current recommended dosage (van der Staay 1991). All wetting

agents, excluding Epsom salts, are toxic. Enstar (kinoprene), an insect growth regulator

recommended for control of whitefiies and aphids, and Vendex (fenbutatin oxide), a miticide,

are the only chernical pesticides safie for use in coordination with onion thrips natural enemies

(Amblyseius, Orius, and Hypoe~pis spp.). Others have an intermediate effect or a harmful

period associated with their use (Matteoni et al. 1993).

Currently there are no selective insecticides against thrips. The firture of chemicai

involvement in thrips control lies in a change from invertebrate poisons targeting al1

arthropods to synthetic analogues of pheromones, allomones and kairamones. The

development and implementation of biorational agents insecticides requires thorough

knowledge of the tritrophic relationship between the plants, the herbivores, and the

beneficials. Antif'eedants or invertebrate a l m pheromones offer the most potential (Jacobson

1997).

2. Predatory Mites

ABSTRACT

An experiment was conducted to determine and compare the predation rate of the

gravid females of two predatory mit es, Ambiyseius nrmmeris (Oudemans) and Iphiseim

(=Ambfyseius) degeneruns (Berlese) [Acarina: Phytoseüdae] on first aad second instar onion

thrips, Thrips tabaci Lindeman [Thysanoptera: Thripidael, larvae on leaf disc arenas. Over

3 days, mites were removed from their food source and placed individually on an arena

containing a 20 mm cucumber leafdisc cucumber cv ' Jessica' . Gravid fernaies were identified

by the presence of an egg d e r the 24 hour period. For each mite species, one gravid femaie

was placed on an arena with either six teneral first instar larvae less than 24 hours old, or six

second instar larvae less than 24 hours after eclosion. Expenments were arranged in a

completely randomïzed design with 15 replications and placed in an controlied environment

chamber maintained at 26OC * 1°C with a 16L:8D photophase:scotophase. The number of

larvae consumed over a 24 hour period was recorded and analysed using MOVA.

Significantly more first instar thrips (3.6 * 1.13) than second instar thrips (3 -4 + 0.8 1 ) were

preyed upon by the gravid female mites. There was no significant difference between the

predation rate of A. cucumeris or I. degenerans for either first or second thrips larval instars

suggesting that these mites have an equal potential to control onion thrips in greenhouse

cucumbers.

24

2.1 Background

There have been successes and failures with the introduction of predatory mites in

greenhouses in different countries. The rate of success has been lower on cucumbers than on

peppers, lower for Frmkiiniefla occidentafis than 77wips tabaci, and soilless cultures (e.g.

hydroponics, nutrient film technique) have more problems than traditional or natural soi1 ones

(Ruidavets 1995). In Canada, Giikeson et ai. ( 1 990) reported that A. cucumeris suçcessfidly

controlled T. tahaci on greenhouse cucumbers, sweet peppers, and chrysanthemums. The

onion thrips, however, have been reported to escape these predators and reach epidemic

levels in greenhouse cucumber production in Nova Scotia (Stokdijk', pers. comm.), England,

and British Columbia (Elliott', pers. comm.).

Many factors, partinilady environmental factors, affect predator efficiency.

Temperature has an obvious influence because mites are poïkilotherms. Consequently,

environmental temperatwe determines the kinetics of biochemical reactions in the mites'

physiology (Sabelis 1985). The effects of temperature and relative humidity on the

development ofA. cucumeris were studied by van Houten and van Lier (1 995). At a constant

relative humidity of 80 percent, mean development time ofA. cuainzeris eggs decreased fiom

4.0, 2.4 and 1.8 days at temperatures of 20 O , 25", and 30°C, respectively. The saturation

deficit at which only 50 percent of eggs hatch (SD,) increased as temperature increased. The

cntical relative humidity for egg sunival was approximately the same at d i temperatures

tested. In a greenhouse expenment alrnost al1 eggs of A. cucumeris survived at a relative

' Stokdijk, P. : Stokdijk Greenhouses, Beaverbrook, Nova Scotia.

'Elliot, D.: Appiied Bionomics, Sydney, British Columbia.

25

humidity of 45 percent.

The influence of VPD (vapour pressure deficit) on the rate of predation by A.

arcirmeris on Franklinieilu occidentalis was also determined in the laboratory by Shipp et

al. (1 996). At a constant temperature, the rate of predation decreased with increasing VPD

between 1.24 to 1.44 kPa. Above these, the rate of predation started to increase again. At

constant VPD, the rate of predation slightly decreased at Iower temperatures and iocreased

at higher temperatures. The optimal conditions for maximum predation for A. crm~meris on

first instar nymphs of F. occidentafis was 0.75 kPa. For greenhouse cucumber crops, the

recornmended production temperatures and VPDs are 17 O to 25 "C and 1.5 kPa.

On parthenocarp cucumbers success has been more variable due to the limited

availability of pollen in the crop as an alternative food source for predatory mites during

periods of low prey densities (Gilkeson et al. 1990). There is conflicting evidence about the

effect of pollen on predation by A. cuc11meri.s. van Rijn and Sabelis (1993) studied pollen as

an alternative food source for A. ctrctlmeris on cucumber leaves. Development and

reproduction of the predator was enhanced but predation rate was decreased. Experiments

by van Houten and van Stratum (1995) indicated when both pollen and thrips larvae were

available, A. cucumeris and 1. degenet-uns feed on thrips larvae in lieu of pollen . If alternative

food sources adversely affect predation, they may, however, prevent predators emigrating

during penods of low prey density. Spraying suspensions of bee-collected pollen on

cucumber plants facilitated the establishment ofA. degeneruns and ailowed it to reproduce

in the absence of prey maintainhg isolated predator colonies for about one month with 0.5

g pollen*~olony-~ (Ramakers et al. 1989). Thrips larvae also feed on pollen and this may

26

shorten the developmental phase in which they are Milnerable to attack by predatory mites

(van Rijn and Sabelis 1997). Avdability of suitable prey is dependent on the dynamics of the

age structure of the prey population. Upon contact with a starved predator, second stage

onion thrips larvae incurred a lower death risk than first stage larvae (Bakker and Sabelis

1989). The lanrae of onion thrips reduced the attack success of their predators by jerking the

abdomen and by producing a drop of rectal fluid. M e r contact with the sticky faecal

droplets, individuals of A. cucumeris usually withdrew and cleaned themselves. When

contacted on the thorax region, the jerking and faecal droplet defence was ineffective (Bakker

and Sabelis 1989). In another study with A. mnrmeris, in which both thrips first and second

larval stages were anaesthetized, the difference in capture:success ratio, i-e. when prey is in

fact consumed, was unchanged (Bakker md Sabelis 1989). This may be due to physical

characteristics such as cuticle toughness or food quality.

ln laboratory studies, predation rate and oviposition rate for 1. degenerans were

lower than A. cucumeris on a diet of western flower thrips larvae. At 25 OC and 70 percent

humidity, A. cucumeris killed 6.0 first instar larvae per day and had a mean daily oviposition

rate of 2.2 eggs, while 1. degenerans had a predation rate of 4.4 larvae per day and laid only

1.4 eggs daily (van Houten et ai. 1995). I n sweet pepper crops, i. degenerans displaced A-

czcclcmeris within 13 weeks and was found on leaves and in flowers located on plants in rows

where it was not reieased indicating better searching efficiency due to higher locomotory

activity (van Houten and van Stratum 1 995). In cornparison to A. mcumeris, the eggs of /.

degenerans were shown to be least sensitive to low air humidity. The females of 1.

degenerans showed total absence of diapause and intermediate levels of predation and

27

oviposition on both thrips larvae and pollen. L degenerans is considered one o f the most

promising predators for biological control of thrips in greenhouses, particularly under

conditions of low humidity and short day length (van Houten et al. 1995).

The objective of this experiment was to determine and compare the predation rate of

the gravid females of these two predatocy mites, Ambfyseius cucumeris and Iphiseius

degenerans, on first and second instar onion thrips larvae on greenhouse cucumbers.

2.2 Materials and Methods

2.2.1 Thrips source: Onion thrips were collected Erom several commercial onion (Alliurn

c e p ) fields in the Annapolis Valley area. Samples were sent to Agriculture and Agri-Food

Canada3 for species identification and were confirmed to be 771rip.s tabaci Lindeman. The

insects were reared on potted onions, leeks (Allium pomm), and cabbage (Brclssica

oleracea) in a controlled environment chamber! The plants were maintained at 25*1 OC

under supplemental fluorescent and incandescent lighting ( 125 pmol-m-'~s-') at 16L: 8D

photophase:scotophase and high humidity. The plants were fertilized once a week with a 200

ppm solution of chelated 20-20-20 plant growth fertilker. Aphids on the cabbage were

controlled with the specialist parasitic wasp, Aphidtus matricariae m e n o p t e r a :

Aphidiidae] .

'Foottit, R: Eastern Cereal & Oiiseed Research Centre, Agriculure & Agi-Food Canada, Ottawa, Ontario.

'Conviron CMP 3023: Controlled Environments Ltd., Winnipeg, Manitoba.

28

Thrips used in the experiment were d o w e d to complete a minimum of two

geoerations on cucumber leaf discs prior to experirnentation. This was accomplished by

placing five adult thrips on rearing arenas (described below) and allowed to oviposit for two

days. The Live thrips were then transferred to new arenas and dead ones were replaced with

fiesh ones fiom the rearing source. This procedure was repeated on alternate days over a

period of several weeks prior to and during the experiments to maintain the availability of

necessary populations. Those larvae not used in experimentation were aüowed to reach

adulthood to maintain the rearing process.

2.2.2 Plant material propagation: Parthenocarp cucumber plants, Cuaimis sarivz~s cv.

Jessica, were grown under similar conditions as the rearing plants. They were seeded in

rockwool propagation blocks and maintained in a 200 ppm 20-20-20 plant growth fertilizer

nutrient solution. Micronutrient deficiencies were corrected with two foliar sprays ten days

apart of 30 g- 100 L-' chelated micronutrient mix. The plants remained Pest and disease fiee

for the duration of the experiment and required no fiirther treatments.

2.2.3 Experimental arenas: A modified leaf disc method developed by Brodeur and

Cloutier (1 992) was used for the expenmental arenas. The arenas were constructed fiom 28

mL clear plastic cups with lids. A 23 mm circular piece was bumed fiom the base of the cup

and the holes were covered with nitex nylon screen (188 prn mesh size) stuck with double

sided tape. The seam between the cup and the screen was sealed with melted paraffin wax.

The cups were inverted so that the Lids served as the base. The shape of the lid provided a

raised well which was surface sterilized under ultraviolet Iight for 20 minutes and then filled

29

with 3.5 mL fertiiized agar. This agar was made by adding 8 g-L" agaroseS to a solution of

20-20-20 plant growth fertiluer formulated at a rate of 2 g*L-' distilled water. The mixture

was autoclaved6 for 20 minutes at 12 1 OC and 1 O 6 kPa. The agar was incubated7 at 45 O C

until used. Immediately prior to use, a h g i c i d e solution was added at a rate of 1 mL-L-'

fertilized agar. The fungicide solution consisted of 0.3 g a.i. benomy18 [methyl 1-

@utylcarbamoyl)-2-benzimidazole carbarnate], 7 mL 100 percent ethanof, and 3 mL distilled

water. To accommodate thrips' thigmotactic behaviour, leaf discs measuring 20 mm in

diameter were cut dong the main veins of mature, middle aged leaves. Prior to being placed

on the agar, the leaf discs were dipped for approximately three seconds in a 3 mL-L" bleach1°

(commercial 5.25 percent hypochlorite concentrate) solution and 3 seconds in distilled water.

The moat surroundhg the raised well was Wed with distilied water. This water was multi-

purpose: it provided a water source for the mites, prevented complete escape of mites and

larvae fiom the arenas, and maintained high hurnidity within the arenas. The arenas were

assembled within a larninar flow cabinet1' and placed in clear containers measuring 32.0 cm

x 25.5 cm x 9.5 cm which were sprayed with 70 percent ethanol prior to use. The bonom

- - - -

'Agarose: Becton Dickenson & Co., Cockeysviile, Maryland.

"teromaster Mk II SSR-3A: Consolidated Stills & Sterilizers, Boston, Massachusetts.

'Imperid II: Lab-Line instruments Lnc. Designers & Manufacturers, Melrose Park, Illinois.

"Benlate@ 50WP: Du Pont Inc. Agrîcultural Products, Mississauga, Ontario.

'Commercial Alcohols Inc., Boucherville, Quebec.

'~Javex@: Colgate-Palmolive Canada Inc., Toronto, Ontario.

:'Laminar flow cabinet CF72: Western Scientific Services Ltd. Richmond, British Columbia.

3 O

was h e d with paper towel moisteaed with distiiîed water to prevent the arenas Eom

desiccating.

2.2.4 Predatory mites source: Two species of predatory mites, Amblyseius cucumeris and

Iphiseius degenerans, were acquired from a Canadian supplier". Over three days, 30 adult

mites were removed fiom their container and food source daily and placed individually on an

arena and starved for 24 hours. Only gravid females were used in the experiments because

they are the most effective predator stage. Gravid females were identified by the presence of

an egg on the arena after 24 hours.

2.2.5 Predation experiments: For each species of mite, one gravid female was placed

individually on an arena with either six teneral first instar larvae less than 24 hours old, or six

second instar larvae less than 24 hours &er eclosion. Arenas were arranged in a completely

randomized design, with treatments replicated five times each over three days (a total of 15

replications). Containers holding the arenas were placed in an environmental chamber" and

maintained at 26* 1 O C with flourescent and incandescent supplemental lighting 95 pm~l-rn-~s- '

with 16L:8D photophase:scotophase. The number of larvae consumed over a 24 hour period

was recorded. This was determined by considering both the number of shrivelled thrips

masses and the number of remaining whole live larvae.

"Koppen Canada Ltd., Scarborough, Ontario.

13Biotronette Mark Ill: Lab-Line Instruments Inc. Designers & Manufacturers, Melrose Park, Illinois.

3 1

2.2.6 Statistical analysis: Final data were subjected to analysis of variance (ANOVA) using

the general linear mode1 procedure of SAS'? Examination of the residuals showed them to

be normaüy distributed with constant variance. Significant differences between treatment

means were determined using p-values from analysis of variance. P-values c0.0 1 were

considered highly si@cant, r 0.0 1 and s 0.05 were considered marginally significant, and

>O.OS were not significant.

2.3 Results and Discussion

There was no significant difference between the rate of predation of the two predatory

mite species tested (p=0.2482). This corresponds with findings by Bakker and Sabelis ( 1 989)

in which A. cumrneris and 1. degenerans had equal abilities to successfùlly attack T. tabaci.

In studies performed by van Houten et al. (1995) A. cucumeris had a higher predation rate

than I. degenerms on first instar western flower thrips larvae. When preying on larvae in the

laboratory, attack success depended on size of the prey, feeding state of the predator, and diet

of the predator (van der Hoeven and van Rijn 1990). Caution must be employed, however,

when interpreting differential attack success of predatory mites. A. cucumeris reared on

broad bean pollen had a bigher capture:success ratio on thrips larvae when first removed from

its food source, but decreased after successive feeds (van der Hoeven and van Rijn 1990).

A. cuctcrneris are transported in controlled release system packages with a bran mite as their

food source, while I. degeneruns are often transported with pollen as their food source. By

removing the mites fiom their respective food sources and not rearing them on onion thrips

'4Statistical Analysis System, SAS Institute, Cary, North Caroha.

32

larvae prior to experimentation, the capture:success ratio rnay be iduenced by an attempt of

the mites to counterbalance a nutritionai deficiency acquired f?om feeding on an alternative

food source (van der Hoeven and van Rijn 1990). These results do indicate, however, the

state of the predator d e r transport and the initial effect the predator may have on the thrips

population in the greenhouse.

Significaatly more fist instar thrips were preyed upon by the gravid females than

second instar @=0.0001) (Table 1). There was no significant interaction between predatory

mite species and l a r d instar (p0.6456). Sabelis (1992) suggested that relative body size

is the key in predator-prey dynamics, however, Bakker and Sabelis (1986) suggested an

alternative explmation. A. cummerzs and A. b k e r i were more successfùl with first instar

thrips larvae than second instar larvae. When both larval stages were anaesthetized, the

difference in capture:success ratio of the two larval instars was not changed. Physical

characteristics such as cuticle toughness or food quality may be a serious contnbuting factor

(Bakker and Sabelis 1986).

- -- -- - - -

Table 1. Predation rate of predatory mites on T. tabaci first and second larval instars on cucumber leaf discs.

Life Stage Mean predation * S.E.M

First larval instar 3.60 0.207

Second l a r d instar 3.37 * O. 148

2.4 Recomrnendations and Future Research

This study suggests there is no difference between the potential predatory

effectiveness ofA. cucumeris and 1. degenerms on greenhouse cucumbers in the laboratory.

33

The true predatory effectiveness remains to be proven by testing both predators in the

environment in which they are to control the Pest - parthenocarp greenhouse cucumbers. An

attempt was made to test these species in the research greenhouses; however, a sudden and

unexpected decrease in the thrips population occurred which nulliiied the attempt.

In the greenhouse environment, the effect of the predator on the Pest population takes

into account important factors which idenw effective biocontrol agents. The predator must

have the searching capacity or dispersai powers (distribution in relation to prey) to locate prey

even when they are rare. The searching ability of 1. degenerans in sweet peppers was also

found to be greater than A. cucumeris (van Houten and van S tratum 1 99 5). This has yet to

be proven in greenhouses cucumbers. Ramakers et al. (1989) maintained 1. degenerarls

colonies in parthenocarp greenhouse cucurnbers with pollen but did not determine their

dispersal or predatory abilities when thrips larvae became available for prey.

The specificity of the predator to the prey is important for two obvious reasons. First,

the introduced predator must not become a pest itself on the crop in which it is to exhibit

control of other pests. Second, the predators must show minimal prey tendencies on other

beneficials present in the naturd enemy complex developed in the greenhouse or intraguitd

predation. This type of predation occurs when different predators are introduced to control

the same Pest (prey) and compete for the same food source and may prey on each other

(Sabelis and van Rijn 1997). It is not necessary for the predator to be highly specific, but it

must show a distinct preference for their prey, yet be able to sunive on alternative food

sources when the prey is rare, usuaiiy at the order or family level (e.g. onion thrips or western

flower thrips - Thysanoptera: Thripidae). The predation rate of 1. degenerans has also been

34

tested on spider mites (Yao and Chant 1989) and polien ( Ramakers 1995; van Houten and

van Stratum 1995). It cm survive and reproduce on both food sources. As mentioned above,

the true effect of pollen on the overail performance of either A. cucumeris or degerreruns

remains in question.

The predator must have the power to increase in number, Le. the intrinsic rate of

natural increase ( rd must be at least as rapid as that of the pest. ïhis represents important

economic information for greenhouse producers. Current recommendatiom comprise of

regular applications ofA. cucumeris in slow release packages to maintain predator levels. In

recent years, the research has shiffed fkom testing predator effectiveness on T. tabac1 to F.

occidentulis because the latter has surpassed the former in Pest st atus in greenhouse vegetable

production. Consequently, recommendations are currently based on this information. Bakker

and Sabeiis (1 989) found that A. clrcumerisand 1. degenems had equal ability to successfûily

attack T. tabaci. Lower oviposition rates have been found for I. degenerans than A.

ct~cz~meris on western fiower thrips larvae in the laboratory (van Houten and van Stratum

1 995). Castagnoii et al. ( 1 990) found that A. cucumeris not only consumed more larvae per

day on a diet of T. tabaci (5.41) than F: occidentalis (3.24), but also laid more eggs, 2.00 and

1 -88 eggs per day, respectively. Such information indicates a possible distinction in versatility

of the predators, e.g. a particular predator may be more effective on one particular thrips

species than the other. This variation in population dynamics of the predator may also suggest

a reassessment of the effects of both I. degeneram and A. cucumeris on the control of T.

tubaci in parthenocarp cucumber production.

3 5

This study has shown that both predators have similar potential to control T. tubaci

on cucumber le& discs. More studies must be conducted to determine if 1. degenerans is

more efficacious against T. tabaci in a greenhouse cucumber production environment than

A. cucumeris. The degree of mortaiity produced should be designed to prevent peak

populations that wiii cause econornic loss, e.g. 80 percent mortality as opposed 98 percent

may be sufficient to achieve this (De Bach and Rosen 1991). Meanwhile, producers can

continue to reiease either predator depending on availability andlor cost. Ultimately,

producers could rear 1. dkgenerans on pollen, e.g castor bean plants (Ricinus cornmunis)

within their greenhouses and distribute it as necessary (StokdijklS, pers. comm.)

"Peter Stokdîjk: Stokdijk Greenhouses, Beaverbrook Nova Scotia.

3. Scented Tnps

ABSTRACT

The effect of the addition of the volatile chernical attractaiits, p-anisaldehyde (4-

methoxybeddehyde) and ethyl nicotinate (3-pyridinecarboxylic acid), to commercially

available blue and yellow sticky traps was examined for adult onion thrips, I;hrips tabuci

Lindeman [Thysanoptera: Thripidae], in greenhouse cucumber production. Blue and yellow

sticky traps were baited with 50 pL of each scent in a slow release dispenser and placed 10

cm above the crop canopy perpendicular to the rows and their daily captures were recorded.

Two studies were conducted at separate locations: a commercial cucumber greenhouse and

two research greenhouses.

There was no interaction between scent and colour at the commercial site.

Anisaldehyde sigdïcantly increased captures by a factor of 1.7 over ethyl nicotinate and 1.2

tirnes more than the control. Ethyl nicotinate did not increase captures over the control. Blue

sticky traps caught twice as many thrips than yeliow traps, and yellow caught 5.0 times more

than clear. This indicates that blue is the colour of choice for capturing adult onion thrips in

greenhouse cucumber production. In the research greenhouses, there was a significant

interaction between ethyl nicotinate and yellow traps. This combination caught 1.6 times as

many thrips as the next most attractive combination, non-baited blue traps. The factors and

thrips behavioural mechanisms for host-fhding cues aEiecting these different results between

locations are discussed.

Correlations between captures by the non-baited traps at the commercial site and the

numbers of thrips per leaves were investigated to determine ifadult thrips captures on sticky

37

traps were indicative of infestation levels on the leaves. There was no correlation between

either blue or yeliow sticky traps, and a moderate correlation between clear trap captures and

captures on leaf sarnptes. Consequently coloured traps seem to concentrate thrips in their

vicinity. Captures of Orius predators on sticky traps were also examined. Scent did not

increase captures of these predators on nicky traps. As retlected in the thrips results, blue

sticky traps were 1.4 times more attractive than yellow and 1 5 .O times more than clear,

indicating captures were not incidental, however biological control agents may also be

susceptible to cues similar to those of their prey, especidy to yellow hues in general. In this

study a specific predator (Orius sp.) was more attracted to a blue hue.

3.1 Background

Although coloured sticky traps catch and id adult thrips, the overall reduction in the

population is not signiticant (Teulon and Ramakers 1990). Sticky traps may not be effective

during flowering or fniit ripening because thrips rnay be more attracted to these plant

structures than the traps (Frey er crl. 1994). %sion and olfaction are the primary mes used

by insects to orient to plant and floral hosts and sometimes these two cues work in concert

(Dobson 1994). Consequently, control using sticky traps could be enhanced by the addition

of an attractant. Addition of volatile chemicals can be useful in two ways: 1) it may aid in the

determination of Uiitial occurrence ofthrips; 2) traps with volatile chemicais may be useful for

control by trapping thrips where wind or air movement does not preclude the movement of

thrips towards traps and does not quickly disperse volaules (Teulon et al. 1993).

38

Several volatile chernicals have been tested to improve sticky trap captures. Aldehydes

were the tùst floral scents shown to attract thrips (Howlett 1914) and currently show the

most potential for attracting thrips (Teulon and Ramakers 1990). .bisaldehyde (4-

methoxybenzaldehyde), a flower scent and a metabolite of certain wood rotting fungi, has

been shown to be a strong thrips attractant (Krk 1985; Brnrdsgaard 1990, Teuion and

Ramakers 1990). Teulon et ai- (1993) found that anisaldehyde baited traps attracted one

hundred times more X tabaci than non-baited traps. Other thrips species are associated with

ripe fniits or vegetative hosts, tberefore a non-floral olfactory cue may be used to detect them

(Teny 1997). Ethyl nicotinate (3-pyridinecarboxylic acid) has been found in Iow

concentrations in the t'niit of carambola (Averrhoa carambola), but has not been reported in

other foods (Wilson et aL 1985) and does not appear to be a constituent ia fiord fiagrances

(Teulon et al. 1993). Ethyl nicotinate was shown to increase captures of ïhrips obscuratzcs

by one hundred times as weîi as capninng thrips two weeks earlier than non-baited traps.

Attraction of T. tabaci was increased by a factor two to five times (Penman et al. 1982).

The behavioural mechanism responsible for these renilts remains controversial.

Observations in flight chamber studies by Hollister et al. (1995) suggest a cbemokinetic

(undirected) response of western flower thrips to p-anisaldehyde. Thrips use scents more

efficiently as an arrestant or to stimulate a visual response than for anemotaxes (Kirk 1985).

Several researchers support odour-induced visual response as the most probable orientation

response involved (Kirk 1987; Briidsgaard 1990; Teulon md Ramakers 1990). It is unclear

whether host finding cues of flower and âuit dweiiers, such as Frankiinieiia species, are the

same as leaf dwellers such as the onion thrips. When detennined, it could be possible to

39

exploit this response to prevent infestation or spread (Kîrk 1985).

The success of volatile chernicals in uapping depends on several factors. Odour

attractiveness usuaiiy shows a concentration related optimum above which compounds that

usually attract insects exert repeliency (Frey et al. 1994). Kirk (1987) showed that the

presence of scent represented a smaiier factor of increase of the trap catches for large traps

compared to smai i ones, thus a high visual apparency could reduce the effect of scent.

Bredsgaard (1 990) suggested that ifa strong visual apparency decreases the effect of scent,

a smaller trap or sub-optimal colour rnay improve effectiveness. Appropriate dispensing of

the attractant is also critical. in another study, painting the chemical directly on the trap

resulted in higher catches than using an impregnated filter paper wick (Teulon et al. 1993).

Trapping efficiency is a consequence of the release rate of the chemical; therefore, a slow

release dispenser may offer an effective alternative. The purpose of this study was to

determine if volatile attractmts can significantly increase thrips captures on commercially

available sticky traps and assess their potential for use for mass trapping.

3.2 Materials and Methods

3.2.1 Greenhouse facilities: Commercial greenhouses (Stokdijk): There were two crops

considered in 1997 in this greenhouse. The first was established in one half of the greenhouse

in mid-Apd and the second one was located in the other halfby mid-May. Total plant density

was approxhately 5000 plants per 3500 m2 (1.43 plantsm*?. A section meamring 22.2 m

x 19.2 m was used for experimentation in each crop. The plants were grown hydroponically

4 0

according to Grodania16 recommendations with greenhouse specific adjustments for the crop.

The plants were trained in a modified v-system umbrella and maintained under standard

production regimes of nutrient and controiied climate. Thrips were controlled using regular

releases of the predatory mite A. cucumeris. During periods of peak densities, these

predators were supplemented witb degenerans and the minute pirate bug 0. insidioszs

Wemiptera: Anthocoridae]. Other pests were controlled as necessary with their respective

biological control agents: whitetlies - Encwsia fomosa wymenopt wa: Aphehidae], h g u s

gnats - Hypoaspis miles [Acari: Laelapidae], and spider mites - Phytoseiulus persimilis

[Acari: Phytoseiidae].

Research greenhouses (Harlow): Two research greenhouses were used, each

measuring 15.2 m x 9.6 m. Both were under plastic and were attached to a common header

house. Plants were grown in a modified nutrient film technique developed by Cooper" and

modified by Toms", and maintained at standard production regimes of nutrition and

controlled c S ï t e and trained in a modified umbrella system.

The summer 1996 cucumber crop was not considered because thrips populations were

extremely low in the greenhouses at this time. Regular inundations of several hundred thrips

did not establish. The 1996 fa11 crop was seeded 30 July in a propagation house and moved

into the greenhouse 9 August. There were six cucumber cultivars grown: Discover, E24 13,

Flamingo, Kalunga, LM8 19, and Tyria. These cultivars were arranged in plots of four plants

'"odania Ltd., Milton, Ontario.

"Cooper, A.: Giasshouse Crops Research Institute, Littiehampton, United Kingdom.

'"Toms, B.: Greenhouse Specialist (Ret.), Truro, Nova Scotia.

4 1

in a randomkd complete block design with five replications Guard plaats were placed at

the ends of each row to avoid edge effeas. The total number of plants was 130 plants

resuiting in a spacing of 0.89 plantsm? To estabiish an onion thrips population in the crop,

plauts were infested with one thrips per plant at the two leaf stage.

The 1997 summer crop was also not considered this time due to crop failure. Plants

were seeded 27 March and placed in the greenhouse 9 April and were subjected to cold

shock. A heavy infestation of the greenhouse whitefly, Triafeurodes vaprariorium

womoptera: Aleyrodidae] M e r weakened the plant. Warming temperatures encouraged

Pyfhiurn species, which, in combination with the other stresses, caused the ultimate demise

of the crop. Plants were removed from the greenhouses 12 and 18 June. The whiteflies were

eradicated with a nicotine sulfate application 19 June. The greenhouses were thoroughly

cleaned by rinsing the lines and tubs with a 10 percent virucidal disinfectant19 (50 percent

potassium monopersulphate, potassium bisulphate, potassium sulphate) solution followed by

a 3.0 percent bleach20 (commercial 5.25 percent sodium hypochlorite concentrate) solution

and finished with two naal flushes of water. This is important to note as our goal was to limit

as much as possible, residual thrips populations (of any species) and to have a command over

the initial ï?arip.s tabaci population for the subsequent attempt.

A second crop was seeded 18 June and moved into the greenhouse 7 July. There

were five cucumber cultivars represented: Corona, Exacta, Pinnacle, Pyralis, and Titleist.

These were placed raadomly in the greenhouse with al1 cultivars represented equally in each

"virkonm: Noraid Laboratones hc., Joliette, Quebec.

'OJavexo: Colgate-Palmolive Canada hc., Toronto, Ontario.

42

house. Rows were divided into five plots of five plants each with guards at the ends of each

row. The total number of plants in each house was 13 5 with a final plant spacing of 0.93

plant~-rn-~. T tabaci was acquired fiom a local culture2' and applied at a rate of

approximately two per plant one week after placement in the greenhouse and aiiowed to

establish. The greenhouse whitefly was controiled with its specific biological control agent,

Encarsia fonnosa. Powdery d d e w w u veated with a solution of 0.5 g-L" benamyi'

[methyl 1 -(butylcarbamoyl)-2-benzïmidazole carbarnate] and 0.25 go L-' etradiazoie [ 1, 2,

4-thiadiazole, 5-ethoxy - 3 (trichloromethyl) J on 12 August aud 1 .O g * ~ - l sulfûr on 28 August.

3.2.2 Volatile chemicals: p-hsaldehyde (4-methoxybenzaldehyde) and ethyl nicotinate

(nicotinic acid ethyl ester, 3-pyridinecarboxyfic acid) were purchased directly? Subsamples

were diluted to a concentration of 1 : 1000 scent:ethanol? Ethanol was used in the trials as

a control. A volume of 50 pL of the scents were pipetted into a nalgene slow release

dispenser and attached to the trap with a metailic and plastic twist tie. Paper twist ties were

not used to avoid possible absorption of attractants.

3.2.3 Sticky traps: Blue and yellow sticky uaps measuring 10 cm x 26 cm were acquired

from a Canadian distributo? Clear sticky traps were cut fiom clear acetate to the same

"Nova Scotia Agricultural College, Truro, Nova Scotia.

"B enlateo SOWP : Du Pont Inc., Agriculturai Produaq Mississauga, Ontario.

"Trubanf 3 OWP: Scotts-Sierra Crop Protection Co., Marysvilie, Ohio.

"Sigma Chernicd Co., St. Louis, Missouri.

"Commercial Alcohols Inc., Boucherville, Quebec.

'"ero-kure International, Sherbrooke, Quebec.

43

dimensions as the commerciaiiy available uaps and coated on both sides with commercial

insect trapping adhesiven.

3.2.4 Volatile Chemical Attractant Experiments

3.2.4.1 Trial 1: commercial greenhouse: In the commercial greenhouse (Stokdijk), the

traps were instded with the base approhate ly 10 cm above the crop canopy perpendicular

to the row and attached to supporting chains with wooden clothes pins. Spacing between

traps was 6.4 m within rows and 7.4 m between rows. The trials were designed as a

randomized complete block, spiit plot with two replications (one replication in each section),

and was repeated four times throughout the cropping season. To accommodate air currents

within the greenhouse, the scent variable was designated as the main plot and placed within

the same row to avoid mixing of scents. The subplots were trap colour, Le. yellow, blue, and

clear. Thrips captures were counted and both traps and scents were replaced daily for ten

days. A time lapse of at least one week was diowed between tnds to allow dissipation of

remaining scent within the area.

3.2.4.1.1 Correlation trial: Leaf samples as per the method developed by Steiner (1 WO),

detailed in Section 4.2.2, were taken within one metre of the traps to correlate trap captures

to standard leaf counts to determine if trap captures were indicative of populations of adult

thrips on leaves.

3.2.4.1.2 O d s captures: Traps with volatile chernicals would be usehl as a supplement for

other biocontrol methods and it is therefore important to examine the effect of traps with

2ÏTangle Trap: The Tanglefoot Co., Grand Rapids, Michigan.

4 4

volatile chernicals have on other controls (Teulon et al. 1993). To that end, the number of

minute pirate bug predators, Orius spp., captured on the traps was also recorded.

3.2.4.2 Trial 2: research greenhouses: In the research greenhouses (Harlow) d u ~ g the

second crop of 1997, traps were hung with string 60m the supporting frames approximately

10 cm above the crop canopy perpendidar to the row. Spacing between traps was 2.5 rn

within row and 3.0 m between rows. Only coloured s t i ch traps were used in these trials

because it had been determined in the previous experiment that the clear traps did not

significantly interaa with scent to increase thrips capture. Air currents were not a signiticant

factor in these houses; therefore, the expenmental design was modified to a randomized

complete block with two repiications (one in each house). This was repeated three times

throughout the cropping season and thrips captures were recorded daily for five successive

days and only scent was replaced on a d d y basis.

3.2.5 Statistical analysis: Recorded data were subjected to anaiysis of variance (ANOVA)

using the general iinear mode1 (GLM) procedure of SAS^*. Significance dierences between

treatment means were determined using p-vaiues tiom ANOVA. P-values <0.01 were

considered higbly significant, 2 0.0 1 and s 0.05 were considered marginaüy significant, and

>0.05 were not significant. The population of thrips in the vicinity of the traps may also

influence trap capture (Parker and Skinner 1997). Immediately pnor to trial establishment,

the mean of two leaf counts of adult thrips in the area of the trap placement was used as a

covariate in the initial statistical anaîyses of the commercial greenhouse data. The covariate

did not improve the R2 value or the mean squares signincantly and was therefore removed.

-- - - -

2'Statistical Analysis System, SAS Institute, Cary, North Carolina.

4 5

Individual analyses were performed on the commercial data replicated over ten days.

Inconsistent results prompted another anaiysis to distinguish between the effects during pre-

fruit and post-f i t production. A final combined andysis was then performed on the four

triais. in ail cases, examination of the residuals reveaied that the data were not n o d l y

distnbuted and did not exhibit constant variance. A naniral logarithmic transformation of t he

data restored nomality and constant variance and improved the R' value. In trials containing

counts of 0, the transformation h(x+l) was used. A second analysis was performed on the

transformed data. Results in tables are presented in original units.

For the commercial greenhouse experiments, correlation analyses were performed

between the mean counts of adult thrips captured on leaf samples and the mean number of

thrips captured on the three trap colours tested. The data for the Oritrs study were

transformed using the natural log and was anaiysed in the same marner as adult thrips

captures were. These results are also presented in original units.

3.3 Results and Discussion

3.3.1 Trial 1: commercial greenhouse: The effect of scent within the individual replications

was not consistent. This may be a consequence of chernical attractants exhibiting differential

effectiveness during ditrerent growth stages of the plant. Attractants may not be cornpetitive

with flowers or ripening h i t . A second analysis was perfonned which separated the

replications into pre-miit and post-fniit production categones. The effect of scent was not

significant in the pre-hit analysis @=0.0797), nor the post-fruit analysis @=0.3946). The

interactive effect of scent by colour was also not significant for the pre-fiuit @=0.2785) nor

4 6

the post-nuit (p=0.4926) production. It was concluded that growth stage did not affect the

attractiveness of these chernicals alone or in interaction with colour and these effects were

therefore, only discussed in the combined analysis.

In all the analyses, the effect of colour was consistent. Blue traps caught significantly

more thrips than yeiiow, and yellow caught si@cantiy more thrips than the clear traps.

Some authors suggest that thrips respond to dserent colour hues and Levets of ultraviolet

reflectance in difTerent crops or surrounâiigs. In grasses, T. tabaci prefer white water traps

over shades of yeliow and blue, and in field crops, yellow was the colour of choice. It has

been constant in the literature that western flower thrips captures are highest on a specific

blue hue over yeflow or white in the greenhouse environment (Bradsgaard 1993, Le Blanc

1993). The results f?om the current trials are concurring for T. tabaci.

In the combined analysis, anisaldehyde baited traps captured significantly more thrips

(p=0.0459) than either ethyl nicotinate or the control (Table 2). Ethyl nicotinate did not

increase thrips captures over non-baited traps. Again, the effect of colour was highly

significant (p=0.0001) with the same trend as in the individual analysis (Table 3). The

interaction between scent and colour was not significant (p=0.9405) at this location.

Table 2. The effect of scent on captures of T. ta- in a commercial cucumber greenhouse.

Scent Mean captures

Anisaldehyde 35.45'

Ethyl nicotinate 21.33b Values foliowed by the same letter are not significantly Merent (u=0.05), LSMeans, pdiK

Table 3. The effect of trap colour on captures of T. tabaci in a commercial cucumber greenhouse.

Trap colour Mean captures

BIue 56-72'

Yellow 25-83b

Clear 5. OZC

Values followed by the same letter are not signincantly dserent (u=0.05), LSMeans, pdiK

3.3.2 Trial 2: researcb greenhouses: Only a combined analysis was performed at this

location. There was a marginal signifïcant interaction between scent and colour @=O. 049 1 ).

The combination ofyeilow traps baited with ethyl nicotinate captured 1 -6 times as many thrips

as the unscented blue traps which produced the next highest level of captures (Table 4).

Bradsgaard (1989) supports Kirk (1987) in the statement that a strong visual appearance may

reduce the effect of scent. In the present study, the effect of a sub-optimal colour for thrips

attraction, yellow, in combination with a non-floral volatile attractant, ethyI nicotinate,

resulted in a heightened attraction for T. tabaci.

When interpretiag of the efficiency of traps containing attractants, caution must be

employed. Composition and concentration of volatiles are affecteci by temporal and spatial

Table 4. The interactive effect of scent x colour on captures of T. tabaci in research cucumber greenhouses.

Scent Colour Mean captures

Ethyl nicotinate Yello w 33-93'

Control Blue 2 1.30*

Anisddehyde Blue 20.53~

Et hyl nico tinate Blue 1 7 . 8 0 ~

Anisaldehyde Yellow 1 7.47bcd

Control Yeiiow 12.10'

Values followed by the same letter are not sigoifcantly different (a=0.05), LSMeans, pdiff.

factors (Dobsoo 1994). If the traps are in proximity to each other, baited traps may draw

thrips away fiom nearby non-baited traps or Erom traps baited with chernicals that induce a

weak response. The contrary may also occur, i.e. traps baited with chernicals that induce a

strong response may draw thrips towards an area containing a non-baited trap or a trap

containing a chernical that induces a weak response, resulting in these less attractive traps

capturing more thrips than these would in isolation (Teulon et ai. 1993). The distance over

which such a response occurs is unknown; however, it will be dected by wind speed and

direction (Lewis 1997~). The possible effects of air movement, speed and direction, and the

proxirnity of the baited traps was considered when designing the commercial greenhouse

trial, therefore, it is udikely that these effects occurred at this location. Teulon et a!. (1 993)

suggested that the slightest breeze is likely to influence the direction of thrips movement.

Thrips capability of exerting a degree ofchoice as to where to alight may be reflected in their

distribution patterns (Lewis 199%). This occurs with the selection between trap colours;

consequently, when mixing of the scents within a trial is not a factor, thrips should be able to

4 9

exhibit this same level of choice when exposed to chernical attractants. Within the Harlow

trials, traps were in closer proximity; however, there was very Little air movement to d i f i se

the scent. The use of dental roll and filter paper wicks and painting the attractant directly on

the trap is reported to leave a strong, lingering scent in the greenhouse (Teulon et al. 1993).

This was not the case with the slow release dispenser used in these triais. Only in close

proxlmity (< 1 m) could the scent be detected by human means and the scent did not linger

when the baited traps were removed.

The variation in results between those of the commercial greenhouse and the ones of

the research greenhouses could be a reflection of the type of response mechanism involved

in thrips host selection. Teulon et aL (1993) found that thrips responded to ethyl nicotinate

in both windy (open fields) and caim (greenhouse) conditions, therefore, host-finding

behaviour of thrips is unlikely to involve directional host-finding responses, e.g. anemotaxis

or odour-induced visual response. Conversely, Visser and Piron (1995) stated that in insects7

search for host plants, their perception of host-plant odour triggers positive anemotaxis, thus

increasing the probability of encountering host plants. insect responses in flight are often

anemotactic and would be impossible in completely still air (Kirk 1985). It seems probable

that thrips could use scent more effectively as an arrestant or to stimulate a visual response,

because then the cue could also be used when the air is completely still. The commercial

greenhouses had fans to promote air and carbon dioxide circulation in sections of the

greenhouse causing wind patterns within rows above the crop, while the Harlow greenhouses

had minimal air movement witbin the house. This could suggest that the host finding cues

stimulated by anisaldehyde and ethyl nicotinate are different. If the host-hding response

50

mechanism were the same for both volatile chemicals, the results should be sirnilar for both

houses.

Lfthe proximity of the scented traps in the research greenhouses resulted in mùring of

the scents, perhaps a 'cocktail' of scents may be most effective in attracting thrips. Most

polyphagous insects are attracted to a number of scents and colours or combination and

chernical mediation of host finding is thou@ to occur with a mixture of chemicals involving

synergism (Miller and Strickler 1984). A physiological 'green odour', in which there is a

combination of two or more volatile chemicals exuding fiom the host, is fiequently involved

in host finding by leaf-feeding insects (Vkser 1986).

3.3.3 Correlation analysis: Correlation analysis between the means ofboth blue (p=0.4865)

and yellow (p=0.5940) sticky trap captures and adults thrips counts on leaves were not

significant. There was clearly no relationship between captures on coloured sticky traps and

adult thrips counts on cucumber leaves. Steiner (1990) also found that there was no

correlation between blue sticky trap captures and leafcounts for F. occzdPntalis. Sticky traps

are designed to attract a high number of insects fiom a large area. This is done by capturing

adult thrips in flight as they disperse above the crop canopy. Thrips numbers on traps are

ultimately gohg to be much higher than the population indicated by leaf samples in the

vicinity of the trap. This cm be extrapolated to the general population in the greenhouse.

The number of thrips caugbt on traps wouid reflect an overestimate of the infestation which

is further complicated by the contagious distribution ofthrips in the greenhouse. Traps placed

within or near thrips' aggregations wiil capture more thrips than in sparsely populated areas.

51

There was a moderately significant correlation between the number of adult thrips

captured on clear sticky traps and those adults counted on leaf samples (p=0.03 18). This

suggests that there is a relationship between aerial thrips populations and thcse found on

leaves when there is no influence of colow (Fig. 1 and Fig. 2). This hding strengthens my

eariier staternent that colour distorts the number of thrips captures and c m o t be a reliable

indicator of the infestation level in the greenhouse. These colowed traps may stiil indicate

the trend in the population, whether it is increasing, decreasing, or stable, or may be used in

areas of high populations to assist other control measures in place.

3.3.4 Orius captures: The effect of scent @=0.6287) and scent by colour interaction

(p=0.9570) was not signifiant for Orius predators. The colour of the sticky trap was highly

significant (p=0.000 1) corresponding with the effects on T. fabaci. Blue sticky traps caught

significantly more Orzus than yellow, and yellow caught significantly more than clear traps

(Table 5).

The behavioural mechanisms aiiowing predatory arthropods to detect their prey are

Iargely unexplored (Sabelis and van Rijn 1997). Effective predators have a variety of

characteristics that aid them in their detedon of suitable prey . Domatia, areas known to be

preferred by insects (insect-domatia association), may also be inhabited by their predators,

e.g. Amblyseius cucumeris inhabiting cracks and crevices where onion thrips larvae are also

known to reside. The evidence of anthocorids inhabiting domatia is scarce (Sabelis and van

Rijn 1 997); however, these predators have shown differential amestment to host plants species

without any association with the prey population on these plants (Beekman et al. 1991).

Adult Encmsia formosa are known to be attracted to yeliow traps when they are dispersing

- . . . . . . .

leaf

Sampling occasion (approx. one week)

Fig. 1. Relationship between adults thrips captures on clear sticky traps and standard leaf samples in greenhouse cucumbers.

O 2 4 6 8 10 J 12

Clear trap captures

Fig. 2. Correlation between adults thrips captures on clear sticky traps and standard leaf samples in greenhouse cucumbers ('significant at a = 0.05).

Table 5. The effect of trap colour on Oriirs captures on sticky traps in greenhouse cucumbers.

Trap colour Mean captures

Blue 3.02'

Yellow

Clear 0.23' Values followed by the same letter are not significantly dinerent (a=0.05), LSMeans. pdiE

in search of whitefly hosts (van de Veire and Vacante 1984). It is probable that anthoconds

may also have similar host finding cues, such as attraction to specific colour hues, as their

thrips prey.

The alarm pheromone of thrips is present in their anal droplets and is believed ta act

as a prey finding kairomone for predators. It has been shown that 0. tristicolor increased its

rate of turning and spent 23 percent of its time within 5 mm of a pheromone source on bean

leaf discs (Teerling et al. 1993). Traps that accumulate large numbers of thrips, such as blue

traps, may have levels of this pheromone that are detectable by Orius predators in flight

(Sabelis 1992).

3.4 Recommendations and Future Research

These results indicate that the volatile chemical attractants, anisaldehyde and ethyl

nicotinate, do not enhance sticky trap captures sufficiently to control thrips by mass trapping

in greenbouse cucumber crops. These resuhs may indicate that volatile chemical attractants

may enhance early detection of iower densities of tbrips in the greenhouse to allow the

expedient employment ofother control measures. The observed variations in results corn the

54

two locations stroogly support the need for continued research in several areas of volatile

chemical attractants. The behavioural mechanism involved in host-hding cues is the most

important factor to be determined if this control tactic is to be exploited. The dserential

attractiveness of scents and their interaction with colour at the two locations imply that the

two scents tested rnay induce different behaviourai responses or rnay be iduenced by air

currents within the greenhouse.

Many researchers state that volatile chemical attractants naturally occur in

combination rather than singly. In this context, i fa mixture is to be tested, the appropriate

concentrations of each volatile must be determined. High concentrations of some attractants

rnay be necessary to induce a response, whiie others rnay require minute amounts. The

hypothesis is that ethyl nicotinate exists in small quantities in nature but induces a strong

result (Teulon et al. 1993). Anisaldehyde rnay not induce the same response. The use of a

slow release dispenser reduced, if not eliminated, the problem of blending and carryover of

scents in the greenhouse for these trials. The slow release dispensers rnay have adversely

affected the trials by not ailowing sufficient release of the chemical necessary to induce the

expected response.

The relationship between trap catches and infestation levels is complex and the

evidence for trap catches being reliable indicators of the size of the crop infestation or the

amount of damage is mixed (Lewis 1997~). This study indicated that clear trap captures were

correlated with adult thrips populations on nearby leaves while coloured traps distorted this

relationship. If traps are to be used to estimate the thrips population in the greenhouse, a

reliable relationship mode1 needs to be developed. Non-baited bi-colour traps have had

55

variable results in attracting western flower thrips. Le Blanc (1993) tested a number of

coiours with contrasting background combinations, none ofwhich attracted signrficantly more

WFT than single colour traps. Vernon and Gillespie (1995) found that sticky traps with

contrasting background colours increased thrips captures. These traps could be tested in

combination with volatile chernical attractants.

Finally, with the pressure to adopt environmentaliy sound control measures, it is

important to assess the effects of the various control measures on each other. Insect

responses to =cezts and colours are known to occur in flight (Lewis 1997b). If an appropriate

scent and trap colour combination is found, it wiil be necessary to determine its effect on

aerially mobile predators, such as Orius spp.

It is recommended that sticky traps remain in use for selective trapping in highly

populated areas to determine the initial occurrence of thrips in the greenhouse or to monitor

the population trends thereafter. ifmass trapping by s t i cb traps baited with volatile chemicaf

attractants is to be recommended as a control measure, insect captures need to be increased

dramatically to jus* the expense and extra caution required to work with these devices.

4. Sequential Sampbg

Abstract

The spatial distribution of Thrips tabaci Lindeman [Thysano ptera: Thripidael larvae

and adults and the predatory mite, Amblyseius cucumeris (Oudemans) [ A d : Phytoseiidae],

on greenhouse cucumbers was calculated using two variance-mean models: Iwao' s patchiness

regression P R ) and Taylor's power Law (TPL). Both models determined thrips larvae and

adults to be contagiously distributed in greenhouse cucumbers with a density contagious

coefficient, b, @PR) and index of aggregatioa, b, (PL) significantly greater than 1. These

vaiues for the predatory mites resulted in ambiguities, e-g. the IPR determined that these

predators were not contagiously distributed while the TPL revealed they were. The index of

basic contagion, a, (PR) indicted that aggregates were the basic component of thrips larvae

population.

Based on these parameters, Iwao and binomial sequential sampiing plans were

developed for thrips larvae and adults on greenhouse cucumbers. The economic threshold

used was estimated at 75 percent of a working economic injury level of 9.5 larvae and 1.7

adults and calculated to be 7.1 and 1 -3, respectively. The maximum sample number to be

taken were calculated to be 66 for larvae and 46 for adults for Iwao's plan. The maximum

sample number for the binomial sequential sampling plan was 39 for larvae and 67 for adults.

57

4.1 Background

Sampling populations to determine âiversity and estirnated numbers of living species

are the most fiindamental research activities in ecology because ecological questions focus

on distribution and abundance of organisms as influenced by biological and physical aspects

of the environment. Sampling provides a foundation for research programmes generating

idonnation on density, dispersion, age structure, reproduction, and migration. A synthesis

of the data ultimately yields an understanding of the population dynamics of the species

(Pedigo 1994). The synchrony between injurious life stages o f the Pest and susceptible stages

of the host is important (Higley and Peterson 1994). The development of sampling

programmes is based upon clear understanding of insect bionomics, host interaction, and

management goals (Hutchinson 1994) and these samphg plans serve as the basis of

integrated Pest management (Pedigo and Bentin 1994).

When a population is sampled, three basic pieces of information are calculated: (i) the

estimate (2) of the true mean (m); (ü) the estimate (.?)of the true variance (a2); and (iii) the

size (unit). The indices used for the description of animal populations are derived tiom

various arrangements of these pieces of information (Southwood 1978). Ideaily, sequential

sampling plans should be calculated for those stages for which there is spatial dispersion data

(Boivin et al. 1 99 1). The mean-variance relationships of Taylor (1 96 1) and Iwao (1 968) have

been used eEectively as foundations for many protocols (B~Ms and Nyrop 1992). These

relationships are extremely important because they permit the prediction of variances for

estimated mean, which in tum d o w s for the development of sequential sampling procedures.

Both models describe this relationship weli, although the Taylor's power law (TPL) seems

58

to be more common (Binns and Nyrop 1992).

Taylor (1 961) showed the mean-variance relationship to follow a power law which

approximates an index of aggregation, 6, describing an intrinsic property of the organism

under study. This index of aggregation is a true population statistic, with a continuous

graduation fiom near unifonn (b - 1; variance < mean) tbrough random, (b = 1 ; variance =

rnean) to highly aggregated (b - =; variance > mean) disuibution (Taylor 196 1).

Lloyd (1967) determined the mean crowding index, < which approximates the

intensity of interaction between individuals as they express a level of crowding in a given unit

of habitat. Iwao (1968) demonstrated that the mathematical relationship between mean

density and mean crowding descnbes certain characteristics of the spatial distribution that are

inherent to each species in a given habitat and can be described by simple linear regression,

called Iwao's patchiness regression (PR). The y-intercept of the regression line is termed the

'index of basic contagion' (Iwao 1970) and indicates the insect's tendency to crowding

(positive) or repulsion (negative) and is a property of the species. The dope of the regression

iine is the 'density contagious coefficient' and is related to the pattern in which the organism

utilizes its habitat and p d l e l s that ofthe index of aggregation of the TPL. Both models have

received critical attention and both remain important for the development of sampling

procedures (Binns and Nyrop 1992).

The simplest way to avoid Pest populations reaching unacceptable levels involves

taking samples firom an ongoing process, infer nom them what the process is doing, and make

a decision either to not take action or reset the process (Binas 1994). It is essential to know

how to gather sufficient information about Pest abundance to be able to make correct

59

decisions without incurring expensive costs (Binas and Nyrop 1992). Only the degree of

effort necessary to start sampling before pest injury becomes unacceptable should be

employed (Higley and Peterson 1994). Eaumerative counts of a Pest is the most accurate

method to assess the pest population (B~Ms and Bostonian 1990); however, when the Pest

is smaü or abundant, coilecting and processing a large number of samples is very time

consumuig. In sequential sampling, the total number of samples taken is variable and depends

on whether or not the results so far obtained give a definite answer to the question posed

about the fiequency of occurrence of an event, Le. abundance of an insect (Southwood 1978).

Pests for which a certain level of infestation c m be tolerated and tbat can change in numbers

rapidly are suited for sequential sampling (Fournier et ai. 1995). In addition to being faster

and therefore less costly per sample unit basis, it is also the most feasible field sampling

method for many organisms (Binns and Nyrop 1992).

When t h e constraints are important, a binomial, or presence-absence, sampling plan

in which decision making is based on intensity, or proportion of plants infested, is prefened

(Shipp and ZariEa 1991). Decision-making based on intensity oflen reduces sampling time

because the sample size necessary to make a decision is based in part on the information

gathered as sample units are inspected and the practitioner terminates inspection once an

infestation is found (Brewer et al. 1994). There are only two possible outcornes in binomial

sampling - the orgaaisrn is either present or absent.

The compromise with binomial sampling is the increased uncertainty with respect to

estimated densities or sample classification decisions (Binns and Nyrop 1992). Estimates of

a mean using binomial sampling have coasiderably more variation associated with them than

6 O

one based on complete enurneration methods. In agricuihiral systems, the increased variation

and subsequent increase in sample size needed to predict the population mean can be offset

by the ease with which the sample units are classified as infested or not infested. As the

degree of clumping increases, this advantage becomes more apparent. An advantage of

binomial sampling that is seldom mentioned is that these methods are impervious to the

effeas of one to a few unusual observations in a sample. This is important because the goal

of a sampling plan is to estimate the population density and use the resulting value as an

estimate of the 'typical' damage that a 'typical' plant experiences. A given level of pests is

required before damage occurs, therefore the estimate of the typical damage is critical. The

use of different t d y thresholds, the minimum number of organisms present to categorize the

plant as infested, can extend and irnprove binomial sampling plans (Jones 1994). This does

not significantly compromise the advantage of binomial sampling of not requiring ac tud

counts of the pest. The use of a taüy threshold with a specific sequential sampiing approach

may allow decisions based on the proportion of leaves having darnaging population levels

present rather than a mean population density Ievel. This would eliminate the variance

associated with the required conversion of the proportion of plants infested to the mean, and

result in fewer samples being required (Jones 1994).

Such plans have been developed for minute and abundant pests such as aphids on

bmssel sprouts (Wilson et al. 1983), mites in garden seed beans (Bechinski and Stoltz i 985),

and onion thrips on field onions (Fournier et al. 1995). The objective of this sîudy was to

adapt existing sequential sampling plans, lwao and binomial, developed for onion thrips on

field onions to this Pest in greenhouse cucumber production.

61

4.2 Materials and Methods

4.2.1 Greenhouse facilities: The same crops used in the volatile chemicai attractant trials

were used in the development of sequential samplig programmes. In the research

greenhouses 1996 fa11 crop, 18 samples were taken from each house. In the second crop of

1997, 1 5 samples were also taken. Lady, in the commercial greenhouses, 9 samples were

taken from the designated sections of each crop.

4.2.2 Sample preparation: The sampling unit was a maNe, middle leaf (Steiner 1990)

collected by cutting the leaf, including its petiole, over a large container wbich was sealed

immediately. Samples were obtained fiom each location on a weekly basis. Samples were

not taken fkom the outside rows or fiom the 6rst or the last plants in a plot to avoid edge

effects. Sampling from the same plant was also avoided to prevent weakening the plant.

In the laboratory, the leaf sample was held over a large fiinne1 and washed twice under

a strong water Stream (Steiner 1990). The insects and mites were caught in a nitex nylon

mesh screen (mesh size = 188 pn) attached at the base of the fiinne1 with an elastic. The

container, leaf and fùmel were rinsed with 70 percent ethanol to dislodge any remaining

thrips and anaesthetize the entire catch to facilitate counting. Counting was performed under

a stereomicroscopetg with fibre optics illumination usually at 40x rnagnification. The number

of first and second land instar thrips, addt thrips, and predators (commercial greenhouse

data) were recorded. The data fiom Harlow research greenhouses and Stokdijk greenhouses

were pooled to give one data set for the development of the sequential sampling plans.

"Wiid Heerbrugg stereomicroscope: W~ld Leitz Canada Ltd., Willowdale, Ontario.

62

4.2.3 Iwao sequentid sampling plan: For each sampling period, the mean and variance of

larval and addt thrips were determined and used to calculate the mean crowding index

according to the equation by Lloyd (1967):

i i= x + ( S / x - 1)

in which 2 is the mean crowding of the sample, 2 is the mean density of thrips of each age

class per leaf and 2 is the variance of the number of thrips of each age class per l e d The

mean crowding indices describes the intensity of interactions between individuds as they

express a level of crowding in a given unit of habitat.

The aggregation of the individuals was quantified by the patchiness regression (PR)

represented by the linear regression of 2 on X in the equation:

X = brX +ar

in which the intercept (ar) is the index of basic contagion and the slope (4 is the density

contagious coefficient (Iwao 1968). When a, = O, a single individual is the basic component

of the population. When a, > O or q < O, there is a positive (clumping) or negative

(repulsion) association between individuals, respectively. When br = 1 : the basic components

of the population are randomly distributed in space, and when 6, > 1, the distribution is

contagious. The patchiness regressions were determined for thrips larvae and adults for dl

locations and for the predatory mites, A. cucumeris, at the commercial site. The resulting

values for a, and br were tested for sigaincant departure fkom O and 1 respectively, using f-

tests.

63

The sequential sampling procedure proposed by Iwao (1 975) proposes the curves for

the upper and lower acceptance Limits using the foliowing equation:

T,, = n ET + t [n ((a, + 1 ) ET + (6,- 1 ) ET*)]"

Th = n ET - t [n ((ar + 1) ET + (b, - 1) ET?]"

where T is the total number of thrips captures (larvae or adults), n is the number of samples

taken, ET is the economic threshold, t is the student's value of t at O . 1 signrficance level for

a two-sided t-test, a, is the index of basic contagion, and b, is the density-contagious

coefficient. The economic thresholds were based on 75 percent of the preliminary economic

injury level (EL) estimates of Steiner (1990) for larval and adults thrips.

The maximum number of samples to be taken in order to determine if the population

level is equal to the economic threshold was determhed by the following fonnula:

4~ = ?/cf [(a, + 1 ) ET + (b, - I)ET']

where d is the confidence intervai of the estimated mean density.

4.2.4 Binomial sequential sampling plan: In binomial sequential sampiing, the relationship

between the proportion of plants infested and the mean number of thrips per sample is

estimated by the formula of Wilson and Room (1983):

P = 1 - e [(- >( In(a R (&'))/(a 3 - 1 ) ] - 1

where P is the proportion of plants infested, and R is the mean density of thrips per plant.

Parameters a and b are estimated fiom (T'PL) (Taylor 196 1 ) :

? = a X~

where a is a scaling factor related to sample size and b is an index of aggregation

characteristic of the species. To determine the value for a, the values of R and 3 were

64

plotted on a log,,,/log,, scale. From the regression equation, the value of a was calculated at

the point where X = 1. The value of b was then determined fiom the equation:

log,, 2 = log,, a + b log,, 2

The index of aggregation, b, can be used to classify dispersion patterns. If b = 1, thrips are

evenly distributed in space (variance < mean). Ifb < 1, then thrips exhibit random distribution

(variance = mean). if b >1 . then thrips exhibit contagious distribution (variance > mean).

The value of b was tested for significance greater than 1 using t-tests.

Acceptance limits were estimated by cdculating a confidence interval for the

proportion of plants infested corresponding to the economic threshold rnultiplied by the

number of plants inspected (Bechinski and Stoltz 1985):

PP * t,. ,,-, n [((P.( 1 -P,))/n)lV'

where P, is the economic threshold (expressed as a proportion of plants infested), n is the

number of plants inspected, t is the student value of t for two sided test at a level of O. 1 with

infinite degrees of fieedom.

The maximum sarnple number to check in the eventuaiity of a density equal to the

economic threshold is determined by the equation of Karandinos (1976):

%a = (~(0.05,-~P.( 1 - P M 2

where h is the precision desired.

4.3 Results and Discussion

The mean captures of thrips Life stages are shown in Fig. A. 1, A.2, and predatory

mites in the Appendix, Fig. A.3 of the Appendix. Populations of thrips were established in

the greenhouses prior to sampling. In the commercial greenhouse, thrips were controlled with

predatory mites.

4.3.1 Iwao sequential sampling plan: Iwao's patchiness regressions of the mean crowdine

(2) on mean density (2 ) obtained for the pooled results were significantly linear for both

onion thrips larvae and adults, as well as the predatory mites, A. cucumeris (Table 6 and Fig.

Table 6. Statistics of the regression of mean crowding (2) on mean density (2 ) for captures of T. tabaci larvae and adults, and the predatory mite, A. czîcurneris, in greenhouse cucumbers.

Life htercept Slope ~djusted-R2 Correlation n Stages a, br coefficient'

T. tabaci: Lawae 1.761 1.3 14 96.3 0.982* * 43

Aduits 0.29 1 1.415 90.3 0.95 1 ** 43

' Correlation coefficient, all of these were significant at a = 0.01.

The values for the index of basic contagion (a,) are listed in Table 7. The index of

basic contagion for thrips larvae was significantly p a t e r than O indicating that the basic

components of these populations were aggregates. For thrips adults and the predatory mites,

the index of basic contagion was not significantly greater than O indicating that the basic

component of the population was an individual.

. - 7

O ra so roo lai rra 160 tao

(BI

p9 / y =-O291 + 1.416 z - = I adj - R' = 90.3 2 * l O

O 10 20 30 40 50

Mean density

Fig. 3. Iwao's patchiness regression of mean crowding (2) on mean density ( R ) of T. tabaci larvae (A) and adults (B) and A. cucumeris (C) in greenhouse cucumbers.

Table 7. T-test of the index of basic contagion (ar) for significance greater than O for T. tabaci larvae and adults and A. cunrmeris in greenhouse cucumbers.

Life stages Mean + S.E.M- f P 111 > O

T. tubaci: Larvae 3.09 * 1.520 2.04 0.050

Adults 0.15 *0.567 0.27 0.400

A. cucumeris 2.70 * 1.120 2.42 O. 160

The values for density contagious coefficient (b,) are listed in Table 8. The density

contagious coefficient for thrips larvae and adults was significantly greater than I indicating

that the basic components of the thrips Iarvae and adults are contagiously distributed in

greenhouse cucumbers. For the predatory mites, the density contagious coefficient was not

significantly greater thanl and therefore, mites do not exhibit contagious distribution, but

exhibit randoxn distribution within greenhouse cucumbers. This could be due to the

application of these predators in slow release packages at regular intervals, thereby inducing

a less aggregated popdation in the greenhouse.

Table 8. T-test of the density contagious coefficient (b,) for sigmficance greater than 1 for T. tabaci larvae and adults and A. cucumeris in greenhouse cucumbers.

Life stages Mean * S.E.M. t P Ir! > 1

T. tabaci: Larvae 1.30 * 0.074 4.11 0.007

Adults 1.36 * 0.087 4.13 0.007

In commercial monitoring programmes, it is more practical to count the total larvae

or adults rather than the individual larval instars because it is faster and requires less training

for the sampier. Two sequential sampling programmes were calculated; one for combined

68

first and second instar thrips, and a second plan for adult thrips. The thresholds employed

were based on 75 percent of the preliminary estimate of the E L for cumulative western

flower thrips larvae (9.5) and adults (1.7) on the middle leaves of greenhouse cucumbers

(Steiner 1990). This estimate can be adopted for onion thrips on greenhouses cucumbers

because these species cause similar leaf damage. The economic thresholds were calculated

to be cumulative counts of 7.1 larval and 1.3 adult thrips on mature, middle leaves of

greenhouse cucumbers.

Precision is important in IPM because it assures the sampler that the estimated Pest

density is relatively close to the true pest density (Legg and Moon 1994). An error level of

0.1 was used in the calcuiation of the upper and lower acceptance lirnits of the sequential

sampling plans (Fig. 4). m e r each sample, the cumulative number of thrips larvae or adults

is cornpared with the appropriate sequential sampling plan. A decision is based on whether

this number is smalier or larger than the acceptance limits. Lf the number is smaller than the

lower limit, sampling is stopped and the grower is advised not to treat. On the contrary, if

the cumulative number is larger than the upper lirnit, sampling is stopped and the grower is

advised to treat. Sampling continues until one of the acceptance limits is crossed or the

maximum number of samples is reached (Fig. 4). Lf no decision c m be made when the

maximum number of samples is reached, the sampler can either arrive at the decision

according to the ciosest limit, or return in a few days to repeat the process and determine if

the population has changed to a level upon which a clear decision can be made.

The maximum number of samples required before sampling can be terrninated varies

with the economic threshold, the size of d, and the error level associated with d. The changes

69

in d are necessary to maintain the sampling sue manageable for commercial monitoring

programmes (Boivin et al. 1991). The maximum sample number (n-) to be taken before

terminating sampling for thrips larvae at a d level of 1.2 and u level of O. 1 was calculated to

be 66 samples. Lf, &er this many samples are taken, the cumulative number of thrips larvae

has not crossed the acceptance lirnits, the population estimate is then ET i d, i.e. 7.1 k 1.2

larvae per leaf with an error level of O. 1. The 4, calculated for addt thrips when the d value

was reduced to 0.3 but the error level remained at O. 1 . was calculated to be 46. In ali cases,

it is recommended that at least 1 O samples are examined before a decision is made (Nyrop and

Simmons 1984).

When issuing recommendations, caution must be employed when interpreting the

results obtained with higher error levels (Boivin et al. 199 1). The producer's notion of a pest

threat cm be quite different than that of a researcher or IPM advisor (Bechinski 1994). The

information must be presented in such a way that the user can understand and balance the risk

of making a Type 1 or a Type II error. If the population is overestimated, a Type 1 error

occurs which results in an umecessary pesticide application that reduces economic efficiency

and potentially has some environmental impact. Ifthe population density is uaderestimated,

a Type II error is made which results in missing a necessaxy pesticide application which

obviously results in an increased pest population and possibly higher yield losses. To the

producer, this may be economically catastrophic (Cupems and Berberet 1994).

20 40 60

Number of sarnples

Fig. 4. Iwao's sequential sampling plan for T. tabaci larvae (A) and adults (B) in greenhouse cucumbers.

4.3.2 Binomial sequential sampling plan: The relationship between the log,,(variance) and

log,, (mean) were significantly linear for aU life stages of onion thrips and the predatory mites

(Table 9 and Fig.5). For thrips Me stages and the predatory mites, the index of aggregation,

6, from TPL was determined to be significantly greater than 1 (Table 1 O), indicating that ail

were contagiously distributed in greenhouse cucumbers. The thrips larvae and adults results

are in accordance with Iwao's density contagious coefficient. The discrepancy between the

aggregations of the predatory mites may be due to P R using each visit per site to develop a

regression line for that site (therefore two b, data points) while b of the TPL can be caiculated

for each visit (20 data points). In this situation, the aggregation based on TPL would be more

robust than PR When more sites are sampled, P R may show sirnilar results.

Table 9. Statistics of the regression of log,, variance on log,, mean for captures of T. tabaci larvae and adults, and the predatory mite, A. cucumeris, in greenhouse cucumbers.

Life Lntercept Slope Adjusted-It2 Correlation n Stages coefficient'

T. tabaci: Larvae 0.265 1.554 96.1 0.98 1 43

Addts -0,060 1.670 87.2 0.936 43

A. criaimeris 0.32 1 1 -446 97.1 0.986 20 'Correlation coefficient significant at a = 0.0 1.

The estimates of the a and b parameters of T'PL were used to establish the relationship

between thrips mean density and the proportion of plants infested in the formula by Wilson

and Room (1983). These relationships appear in Fig. 6.

The sequentiai sarnpling programmes were based on the same economic thresholds

as Iwao's sequentid samphg programmes, 7.1 larvae and 1.3 adults perle& The proportion

(BI 3 -

a - = -0.060 + 1.670 log,,a 3 u adj - R~ = 87.2

i tu I I

(Cl 3

loglo s2 = 0.321 + 1.446 log,, ?

adj - R' = 97.1 2 -

1 -

I O 1

-1

Fig. 5. Taylor's power law regression of log,, (variance) on log,, (mean) of I: tubaci larvae (A) and adults (B) and A. nrcumeris (C) in greenhouse cucumbers.

Table 10. T-test of the index of aggregation, b, greater than 1 for T. tabaci larvae and adults and A. cucumeris in greenhouse cucumbers.

Life stages Mean * S.E.M. z P 111 > 1

Adults 1.36 * 0.082 4-38 0.000

of plants infested corresponding to these economic tbresholds were determined corn the

equation of Wilson and Room (1983) to be 0.17 for larvae and 0.47 for adults. To ensure

accuracy of the Pest density estimate, the corresponding proportion of infested samples

shouId not exceed 80 percent (Southwood 1978) because the maximum detectable densities

tend to be below the economic threshold. These proportions are incorporated into the

formula for the acceptance limits based on the formula of Bechinski and Stoltz (1 985) and are

presented in Fig. 7. The y-axis indicates the maximum aliowable number of plants infested

with one or more thrips corresponding to the number of samples taken. Parallel to Iwao's

sequential sampling plans, if either of these lines are crossed before the maximum sample

xmber is reached, sampling is terminated and a decision to either treat (upper limit) or not

to treat (lower lirnit) is made.

The proportion of infested leaves corresponding to the economic threshold is also

incorporated into the formula for maximum sample number by Karandinos (l976), and was

determined to be 39 samples for larvae and 67 samples for adults. The conditions of the enor

ternis @=O. 1) and minimum sample number suggested by Nyrop and Simmons (1984) also

applies to the binomial sequential sampling.

2 1 thrips larvae c l

Q O an O

a 63 a9 O

O r 1 thrips adults O 0 r 5 thrips adults

a 1 I I

15 20

Mean densrty

Fig. 6. The relationship between proportion of leaves infested with 2 1 and 2 5 thrips larvae (A) and thrips adults (B) in greenhouse cucumbers.

Treat

Io 1 ( Continue sampling \ /-'

40

Number of samples

Fig. 7. Binomial sequential sampling plan for T. tubaci larvae (A) and adults (B) in greenhouse cucumbers.

76

4.4 Recommendations and Future Research

There is considerable criticism of the variance-mean relationship models of Iwao and

Taylor. Iwao's patchiness regression is based on deductive reasoning of the parameter

est imates and t hese parameters were originally derived with close reference to theoretical

distribution models, thus allowing ecological implications of the parameters to be estimated

(Kuno 199 1). Taylor's index of aggregation is purely an empirical model with no definite

theoretical background. Taylor et al. (1 978) defended the descriptive capabiiity of this

parameter by cornparing multiple sets of extensive data and concluded that this model was

supenor to the patchiness regression parameters. Kuno (1991) noticed that in these

comparisons, the R2 value is sufficiently high in both models for the practical purpose of

describing the populations. Many researchers comparing IPR and TPL have found that TPL

fits data better than the PR A disadvantage to P R is that it is not sensitive to density

dependent changes in spatial patterns that may induce some non-linearity to the regression

(Kuno 1991). Taylor (1978) insisted that the index of aggregation is a defrnite 'species'

specific characteristic that reflects the mode of density dependent dispersal of that species,

being entirely independent of other factors such as sampling scale and quadrat size. These

studies have shown that these models have equally high adjusted-R2 values and correlation

coefficients (Tables 7 and 10) to be equally vdid in their descriptive ability of onion thrips in

greenhouse cucumbers.

Development of sequential sampling plans is a varied and sometimes complicated

procedure. Producers often make decisions afFecting crop and livestock commodities wonh

thousands of dollars without adequate iaformation, seemingly because they are unaware of

77

sampling protocols, or these protocols are perceived as being too cornplex and having

unacceptable risk of an erroneous decision (Cuperus and Berberet 1994). The adoption of

sampling protocols must demonstrate to the producer a relative advantage over the existing

practices to the producer and the protocols must be compatible and readiiy integrated with

the overall operation. Effective communication and cooperation arnong researchers,

extension specialists, and consultants are essentid to the task of information transfer and

technologies implementation. Both of these sequential sampling plans have the advantage of

not requiring a theoretical mathematical mode1 approaching the tnie spatial distribution of the

insect. An advantage to the binomial sequential sampling plan is the ease and efficiency with

which sample units can be classified and its robustness in spite of a few unusual observations.

Fournier et al. (1 995) found that binomial sequential sampiing was as reiiable as Iwao's for

this species in field onions.

Validation of these sequential sampling plans will determine which type of sequential

sampling plan is best suited for onion thrips in greenhouse cucumbers. It is of the utmost

importance that these plans be validated before any attempt to apply them to a commercial

situation is made. This may be accomplished in the field, i e . cornmerciai greenhouses, or by

simulation. Field validation is best because there are multiple sources of uncertainty inherent

in the development and implementation of a sampling procedures.. AIthough field validation

is more expensive than simulation, it best represents this uncertainty (Hutchinson 1994).

Sarnpling a large number of greenhouses allows the representation of a variety of growing

conditions and may provide more robustness to the estimates of the pest number (Higley and

Pedigo 1997) and the pest's behaviour in that system. This validation could be achieved by

78

using the plans in several greenhouses and computing the number of t h e s the plan provided

early valid recornmendations over the total number of trials. Boivin et ai. (1 99 1) tested their

plans for tarnished plant bugs in 30 commercial celery fields and obtained satisfactory results

in 26 of these fields. This is an example that validation of a sequential sampling plan can put

to test a large network of resources. Failure to validate a proposed sampling plan in large

scale commercial operations is probably the single most important reason that such

programmes are not adopted (Tnimble 1994).

Accurate quantification of the relationship between Pest number and crop damage is

crucial to the development of economic injury ievels and has been a major preoccupation of

agricultural entomologists (Higley and Pedigo 1997). For greenhouse crops, there is almost

a complete lack of economic injury level (EL) or economic thresholds (ET) for greenhouse

pests (Shipp et al. 1997). The ET is aiways set below the E L (also called action threshold)

to allow lag time for the chosen control measure to operate and prevent the E L ffom being

reached andlor surpassed (Metcalf and Luckman 1982). This lack of quantitative yield loss-

feeding damage tùnctions provides an inherent source of variability or uncertainty because

the sequential sampling plan based on these d u e s are only as accurate as the values

themselves. The concept of EIL is very straightforward. EU represents the point where

costs equal benefits. Costs are the losses associated with pest action and the cost of

managing the pest. Benefits are the losses prevented by management. Hidden in this simple

statement is considerable biological complexity and the application of E L to specific Pest

problems adds further complexity. Consequently, E L is an expression of both economic and

biological parameters (Higley and Pedigo 1997). Development of these values are complex

79

and variances associated with these fùnctions di probably always be higher than those for

statistical sampling plans and much less predictable due to the complex interaction of the plant

and its environment (Jones 1994). While development of E L has been one of the most

important and usefiil concepts in Pest management, these limitations prevent their cornplete

development and implementation (Higiey and Pedigo 1 997).

5. Conclusion

These studies have examhed specific aspects of an integrated pest management

programme for a single pest on greenhouses cucumbers. However, the host, Pest and

environment create an interplay, representing a moving surface upon which timing and

monitoring approaches for sampling must contend (Higley and Peterson 1994). There is a

compelling need for reexamination and recommitment of the basic tenets of pest management

because sequential sampling for IPM purposes has the potential to enhance such programmes

in a multitude of directions.

Time sequential sarnpling requires a mathematical mode1 of data over time,

classification limits of outbreak and non-outbreak populations, and levels of acceptable risk

in m a h g classification errors (Pedigo 1994). If knowledge of the population growth is

available, sample information can be combined with this knowledge to forecast tùture density.

Such monitoring protocols determine whether the pest density exceeds an intervention

threshold and ifdensity is less than the threshold, the protocols determine how long one can

wait before sampiing the pest population again and be reasonably sure that the density will

not have grown to exceed an intervention threshold (Nyrop and van der Werf 1994).

Biological control is a key factor in many current IPM programmes. The key to using

biological control agents in pest management is to identi& when naturd enemies are

sufficiently abundant to prevent herbivores f?om inflicting economic loss (Nyrop 1994). By

incorporating the prediction or monitoring biological control into the sampling plan, the ratio

of pest-natural enemy can sometimes be used to determine whether natural enernies are

sufficiently abundant to control Pest populations. This same ratio may not always be used as

8 1

an index of the îikelihood of biological conaol because other factors may influence the index,

more than one natural enemy may be involved in the interaction and the relative species mix

may influence the ratio, or it may not be practical to measure the natural enemy density

(Nyrop and van der Werf 1994).

Ultimately, there must be a change in management philosophy for systems which have

more than one herbivore that simultaneously feed upon a crop. Guilds of arthropods,

established on their physiological mode of injury to the host, wili be managed better

collectively rather than as a single species (Hutchins 1994). One can rarely develop a

sampling plan which does more than estimate the abundance each herbivore and each

entomophage. Basic field data cm be incorporated into crop-herbivore-natural enemy models

that estimate and forecast the interaction between these trophic levels (Wilson 1994). The

most widely adopted IPM programmes are those practised in commercial sweet pepper and

cucumber crops throughout north-west Europe and Canada. Ln Ontario, a Harrow

Greenhouse Crop Manage? software has been developed for greenhouse cucumbers which

provides advice on identification and control of pests and physiological disorders and the

latest information on production practices and management strategies. This system aids

producers in reducing operational expenses such as fenilizer and pesticides and allows the

user to maintain a database on al1 aspects of their operation. A communication interface has

been added that ultimately makes it possible to develop the forecast module for predicting

potential Pest and disease outbreaks and advises measures to be taken to avoid them (Shipp

et al. 1997). This system provides the producer with the necessary categones of a complete

P M programme so that an informed and accurate decision cm be made.

T 1 adub

4 6

Sampling occasion (weeks)

Fig. A. 1. Mean captures of T. tabaci larvae and adults per cucumber leaf at Harlow research greenhouse 1 (A) and 2 (B) - NSAC, Fa11 1996.

Sampling occasion (weeks)

Fig. A.2. Mean captures of T. tabac1 larvae and adults per cucumber leafin Harlow research greenhouse I(A) and 2 (B) - NSAC, Fa11 1997.

+ thrips iarvae + thrips aduits -F predatory mites

+ thrips larvae -0- thrips a d u k + predatory mites

4 6 8

Sampling occasion (weeks)

Fig. A.3. Mean captures of T. tubaci larvae and adults and A. cucumeris per cucumber leaf in Stokdijk Greenhouses, crops 1 (A) and 2 (B) - Beaverbrook, Fall, 1997.

References

Allen, W.R 1995. Understanding thrips behaviour. Canadian Greenhouse Conference. University of Guelph. 15 p.

Ananthakrishnan, T.N. 1 979- Biosystematics o f the Thysanoptera. AM, Rev. Entomol. 24: 159- 183.

Anonymous. 1968. Van Nostrand's Scientific Encyclopedia. 4' Edition. D. Van Nostrand's Company hc. Princeton.

Anonymous. 1993. Pest management recommendations for Ontario greenhouse crops. Ontario Ministry of Agriculture and Food. Publ. 365: 64 p.

Bailey, S.F. 1933. The biology of the bean thrips. Hilgardia. 7: 407-522.

Bakker, F.M., and M. W. Sabelis. 1986. Attack success ofAmbfyseiusmckentiei and the stage related defensive capacity of thrips larvae. Med- Fac. Landbouww. Rijksuniv. Gent. 51 : 1041-1044.

Bakker, F.M., and M.W. Sabelis. 1989. How larvae of Thrips fabuci reduce the attack success of phytoseiid predators. Entomol. Exp. Appl. 50: 47-5 1.

Bechinski, E. J. 1 994. Designhg and delivering in-the-field scouting programmes. pp. 6 83 - 706. In: Pedigo, L.P,. and G.D. Buntin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Bechinski, E.J., and R.L. Stoltz. 1985. Presence-absence sequentid decision plans for Tetranychus rrrticae (Acari: Tetranychidae) in garden-seed beans Phasedus wtIgarzs. J. Econ. Entomol. 78: 1475-1480.

Beekman, M., J. J. Fransen, RD. Oetting, and M. W. Sabelis. 1 99 1. Differential arrestment ofthe minute pirate bug, Orius insidiosus (Say) (Hemiptera: Anthocoridae), on two plant species. Med. Fac. Landbouww. Rijksuniv. Gent. 56: 273-276. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press, Cambridge. 740 p.

Bennison, J. A., and R Jacobson. 1 99 1. Integrated control of Frmkfinieffa occidentafis (Pergande) in UK cucumber crops - evaluation of a controiled release system of introducing Amblyseius cucumeris. Med. Fac. Landbouww. Rijksuniv. Gent. 56: 25 1 -258.

Bims, M.R 1994. Sequentiai samplhg for classifjmg Pest status. pp. 13 7- 1 74. In: Pedigo, L.P., and G.D. Buatin (Eds.). 1994. Haadbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

B k , M.R, and N.J. Bostonian. 1990. Robustoess in empincdy based binomial decision mles for integrated pest management. J. Econ. Entomol. 83: 420427.

Bims, M.R., and J.P. Nyrop. 1992. Sampiing insect populations for the purpose of P M decision making. Ann. Rev. Entomol. 37: 427-453.

Bims, E. S., R.A Hd, and RJ.J. Pickford. 1982. 7?1rips labaci Lind. (Thysanoptera: Thripidae) - distribution anâ behaviour on glasshouse cucumbers in relation to chernical and integrated control. Entomol. Mon. Mag. 118: 55-68.

Boivin, G., J.-P-R Le Blanc, and I.A. Adams. 1991. Spatial dispersion and sequential sampling plan for the tamished plant bug (Hemiptera: Miridae) on celery. J. Econ. Entomol. 84: 158- 164.

Bournier, A. 1983. Les Thrips. Biologie, Importance Agronomique. Paris. 128 p.

Brewer, M.J., D.E. Legg, and J.E. Kaltenbach. 1994. Cornparison of three sequential sampling plans using binomial counts to class~fy insect infestation with respect to decision thresholds. Environ. Entomol. 23: 8 12-826.

Brodeur, J., and C. Cloutier. 1992. A modified le& disk method for rearing predacious mites [Acarina: Phytoseiidae]. Phytoprotection. 73: 69-72.

Brradsgaard, H. F. 1989. Coloured sticky traps for Frankliniella occidentalis (Pergande) (Thysanoptera: Thripidae) in glasshouses. J. Appl. Entomol. 107: 136- 140.

Brdsgaard, H. F. 1990. The effect of anisaldehyde as an attractant for Frankiinieffa occidenialis (Thysanoptera:Thripidae) and the response mechanism involved. SROPNPRS Bull. 8: 37-38.

Br~dsgaard, H. F. 1993. Coloured sticky traps for thrips (Thysanoptera: Thripidae) monitoring on glasshouse cucumbers. SROP/WPRS Bulletin. 16: 19-22.

Castagnoli, M-, G. del Bene, E. Gardani, and S. Simoni. 1990. Possibilita di controllo biologico di l%rips tabuci Lind. e Franklïnielh occidentalis Pergande (Thy sanoptera: Thripidae) con Amblyseius cucumeris (Oud.) (Acarina: Phytoseiidae). Redia. 73: 53-6 1. In: Chishoh, I.F., and T. Lewis. 1984. A new look at thrips (Thysanoptera) mouthparts, their action and effects of feeding on plant tissue. Bull. Entomol. Res. 74: 663-675.

Clarke, N.D., J.L. Shipp, W.R Jarvis, A.P. Papadopoulos, and T. J. Jewett. 1994. Integrated Pest management of greenhouse crops - A conceptual and potentially practical model. HortScience. 29: 846-849.

Cuperus, G.W., and RC. Berberet. 1994. Training specialists in sampling procedures. pp. 669-68 1. In: Pedigo, L.P., and G.D. Buntin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Daniel, S.M. 1904. New California Thysanoptera. Entomol. News. 15: 293-297.

De Bach, P., and D. Rosen. 1991. Biologicai Coatrol by Natural Enemies. Cambridge University Press, Cambridge. 440 p.

De Jager,C.M., RP.T. Butot, T. J. De Jong, and J.T. Andora. 1993. Population growth and survival of westem flower thrips, Franklinielh occidentalis ('ergande) (Thysanoptera: Thripidae) on dinerent chrysanthemum cultivars. J. Appl. Ent. 115: 5 19-525.

de Kogel W.J., M. van der Hoek, and C. Mollema. 1997a. Oviposition preference of westem flower thrips for cucumber leaves from dinerent positions dong the plant stem. Entomol. Exp. Appl. 82: 283-288.

de Kogel, W.J., A. Bailcema-Boomstra, M. van der Hoek, S. Zijlstra, and C. Mollema. 1997b. Resistance to western flower thrips in greenhouse cucumber: effect ofleafposition and plant age on thrips reproduction. Euphytica. 94: 63-67.

Dobson, H.E. 1994. Floral volatiles in insect biology. In: Bemays, E.A. (Ed.) uisect Plant Interactions. Vol. S. CRC Press hc., Boca Raton. pp. 47-8 1. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p.

Edwards, C.A., and G. W. Heath. 1964. The Principles of Agricultural Entomology. Chapman and Hall Ltd., London. 41 8 p-

Eveleigh, E S , and D.A. Chant. 198 l . Experimental studies on acarine predator prey interactions numerical response of immature and adult predators (Acarina: Phytoseiidae). Can. 3. Zool. 59: 1407-1418.

Frey, J.E., RV. Cortada, and H. Helbling. 1994. The potential of flower odours for use in population monitoring of westem flower thrips Frankiinîella occidentalis (Perg.) (Thysanoptera: Thripidae). Biocont. Sci. Technol. 4: 177- 186.

Ferguson, G. 1995. Controlling thrips the Dutch way. Greenhouse Canada. Oct: 58-83.

Fournier, F., G. Boivin, and RK. Stewart. 1995. Sequential sampling for Thrips tubaci on onions, pp. 557-562. In: Parker, B-L, M. Skinner, and T. Lewis (Eds.). 1995. Thrips Biology and Management. Plenum Press, New York. 636 p.

Gaum, W.G., J.H. Giliomee, and K.L. Pringle. 1994. Resistance of some rose cultivars to western flower thrips, Frankliniella occidentulis (Thysanoptera: Thripidae) . Bull. Entomol. Res. 84: 487-492.

Gilkeson, L.A. 1990. New crops, new pests, new predators. Growertallcs Aug: 102- 104. In: van Lenteren, J-C., A.J.M. Loomans, M.G. Tommasini, S. Maiai, and J. Riudavets (Eds.). 1995. Biological Control of Thrips Pests. Veeman Drukkers, Wageningen. 20 1 p.

Gilkeson, L.A., W.D. Morewood, and D.E. Eiiiott. 1990. Current statu of biological control of thrips in Canadian greenhouses with Amblyseius cucumeris and Orius frisficolor. SROP/WPRS Bull. 8: 7 1-75.

Gillespie, A.T. 1986. The potential of entomophagous kngi as control agents for onion thrips, Thrips tubuci. pp. 237-243. BCPC Mono. No.34 Biotechnology and Crop Lmprovement and Protection.

Gillespie, D-R, and D.M.J. Quiring. 1990. Biological control of hngus gnats, Bru&sia spp. (Diptera: Sciaridae), and western flower thrips, F r d n i e l l a occidentuiis (Pergande) (Thysanoptera: Thripidae), in greenhouses using a soil-dwelling predatory mite, Geolalqs sp. nr. amlefer (Canestrini) (Acari: Laelapidae). Can. Entomol. 122 : 975- 983.

Gillespie, D.R, and R S. Vernon. 1990. Trap catch of western flower thrips (Thysanoptera: Thripidae) as afCected by colour and height of sticky traps in mature greenhouse cucumber crops. J. Econ. Entomol. 83: 97 1-975.

Heming, B-S. 1978. Structure and fùnction of the mouthparts in larvae of Hapioh-pis verbasci (Osborn) (Thysanoptera: Tubulifera, Phlaeothripidae). J. Morphol. 156: 1-3 7. In: T. Lewis (Ed.), Thrips as Crop Pests. University Press, Cambridge. 740 p.

Higley, L.G., and R.K.D. Peterson. 1994. initiating sampling programs. pp. 1 19- 13 6. In: Pedigo, L.P., and G.D. Butin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 7 14 p.

Higley, L.G., and L.P. Pedigo, 1997. Economic Thresholds for Integrated Pest Management. University of Nebraska Press, Lincoln. 3 27 p.

Hollister, B., E-A Cameron, and D.A.J. Teulon. 1995. Effect of p-anisaldehyde and a yellow colour on behaviour and capture of western flower thrips. pp. 571-574. In: Parker, BAL, M. Skinner, and T. Lewis (Eds.). 1995. Thrips Biology and Management. Plenum Press, New York. 636 p.

Howlett, F.M. 19 14. A trap for thrips. J. Econ. Biol. 9: 2 1-23. In: Lewis, T. 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p.

Hunt, D., D. Gagner, and C. Drury. 1998. The influence of greenhouse plant nutrution and crop phenology on host plant tesistauce and susceptibiiy to insect pests [Online]. Greenhouse and Processing Crops Research Centre Annual Report. Harrow. Available ffom: URL: htt~://res.aer.c~arrow/~atvlfe~98/ii-entt [Accessed 1 9 May 1 9981.

Hussey, N.W., and N.E.A. Scopes. 1985. Greenhouse vegetables (Britain). pp 285-297. In: Helle, W., and M. W. Sabelis (Eds.). 1985. Spider Mites: Their Biology, Natural Enemies, and Control. Vol 1B. Elsevier Science Publishing Company Inc., New York. 458 p.

Hussey, N. W. and Parr, W. J. 1 989. Effect of the greenhouse red spider mite ( Tetranychus rrrticae Koch) on the yield of cucumber. J. Hortic. Sci. 38: 225-263.

Hutchins, S.H. 1994. Techniques for sampliog arthropods in integrated Pest management. pp. 73-97. In: Pedigo, L.P., and G.D. Buntin (Eds.). 1994. Handbook of Sampiing Methods for Arthropods in Agriculture. CRC Press Lnc., Boca Raton. 714 p.

Hutchinson, W.D. 1994. Sequential sampling to determine population density. pp. 207-243. In: Pedigo, L.P., and G.D. Buntin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Iwao, S. 1968. A new regression method for analyshg the aggregation pattern of animal populations. Res. Popul. Ecol. 10: 1-20.

Iwao, S. 1970. Analysis of contagiousness in the action of mortality factors on the western tent caterpiilar population . Res. Popul. Ecol. IO: 1-20.

Iwao, S. 1975. A new method of sequential sampling to classi@ populations relative to cntical density. Res. Popul. Ecol. 16: 281-288.

Jackson, A., and N. Scopes. 1993. Sticky solutions. Greenhouse Grower. Oct: 8-9.

Jacobson, R J. 1995. lntegrated Pest management in cucumbers - prevention of establishment of Frankliniella occidenfalis (Pergande). Med. Facult. Landbouww. Rij ksuniv. 60: 8 5 7- 863.

Jacobson, RJ. 1997. Integrated pest management (PM) in glasshouses. In: Lewis, T (Ed.). 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p.

Jones, V.P. 1994. Sequential estimation and classification procedures for binomial counts. pp. 175-205. In: Pedigo, L.P., and GD. Buntin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Karandinos, M.G. 1976. Optimal sarnple size and comments on some published formulae. Bd. Entomol. Soc. Am. 22: 41 7 4 2 1.

Kirk, W.D.J. 1985. Effect of some floral scents on host finding by thrips (Insecta: Thysanoptera) J. Chem. Ecol. 11: 35-43.

Krk, W.D.J. 1987. Effects of trap size and scent on catches of 7hr@s imaginis Bagnell (Thysanoptera: Thripidae). J. Aust. Entomol. Soc. 26: 299-302.

Kuno, E. 1991. Sampiing and analysis of insect populations. An.. Rev. Entomol. 36: 285- 304.

Le Blanc, J.-P. R 1993. Western flower thrips in Nova Scotia greenhouses. Canada/Nova Scotia AFDA Report 1993-10. 42 p.

Legg, D.E., and RD. Moon. 1994. Bias and variability in statistical estimates. pp 55-69. In: Pedigo, L.P., and G.D. Buntin. 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Lewis, T. 1973. Thrips: Their Biology, Ecology and Economic Importance. Academic Press New York. 349 p.

Lewis, T. 1997a. Thrips as Crop Pests. University Press. Cambridge. 740 p.

Lewis, T. 1997b. Flight md dispersal. pp 175-196. In: T. Lewis (Ed.). 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p-

Lewis, T. 1997c . Field and laboratory techniques. pp 43 5-475. In: T. Lewis (Ed.) . Thrips as Crop Pests. 1997. University Press. Cambridge. 740 p.

Lloyd, M. 1967. Mean crowding. J. Anim. Ecol. 36: 1-30.

Logan, J.A., RE. Steiner, RL. Rabb, and J.S. Bacheler. 1979. A descriptive mode1 for predicting spring emergence of Heliothis zea populations in North Carolina. Environ. Entomol. 8: 141- 146.

Long, E. 1995. Enter AQ 10, The fnendly fimgus. Grower. Nov: 22-23.

Loomans, A.J.M., and J.C. van Lenteren. 1990. Hymenopterous parasites as biological control agents of Franklineilu occidenlalis (Pergande). SROPiWPRS Bull. 8 : 1 09- 1 1 4.

Matteoni, J., D. Eliot, D. Con, W.J. Ravesberg, and M. Dissevelt. 1993. Chernical effects on greenhouse biological control agents. Grower Talks. Aug: 8 1-85.

McKenzie, CL., B. Cartwright, M.E. Miller, and I.V. Edelson. 1993. Injury to onions by Thrips tabaci (Thysanoptera: Thripidae) and its role in the development of purple blotch. Environ. Entornol. 22: 1266- 1277.

Metcalf, RL., and W.H. Luckman. 1982. Introduction to Insect Pest management. Zn* Ed. John Wdey and Sons, New York. 577 p.

Metcalf, RL., and RA. MetcaK 1993. Destructive and Usefùl Insects: Their Habits and Control. Sm Ed. McGraw-Hill hc., New York. 1 100 p.

Miller, J-R, and K.L. Strickler. 1984. Finding and accepthg host plants. pp 127-1 57. In: Bell W. J., and RT. Cardé (Eds.). 1984. Chemical Ecology of Insects. Chapman and Hall, New York. 524 p,

Moilema, C . and RA. Cole. 1996. Low aromatic amino acid concentrations in Ieaf protein determines resistance to Frankfiniella occidentalis in four vegetable crops. Entomol. Exp. Appl. 78: 325-333.

Morison, G.D. 1957. A review of British glasshouse Thysanoptera. Tram. R Entomol. Soc. Lond. 109: 467-534. In: Lewis, T. 1973. Thrips: Their Biology, Ecology and Econornic Importance. Academic Press London and New York, London. 349 p.

Mound, L.A. 1971. The feeding apparatus of thrips. Buil. Entomol. Res. 60: 547-548.

Murphy, G., B. Broadbent, and W. Brown. 1995. Using integrated Pest management in the greenhouse. Hort East. 22 p.

North, RC., and A.M. Shetton. 1986. Overwintering of the onion thrips, Thrips tabaci (Thysanoptera: Thripidae), in New York. Environ. Entornol. 15: 695-699.

Nyrop, J.P. 1994. Sampling to predict or monitor biological control. pp. 245-336. In: Pedigo, L.P., and G.D. Buntin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Nyrop, J.P., and G.A. Simrnons. 1984. Errors occurred when using Iwao's sequentid decision rule in insect sampling. Environ. Entomol. 13: 1459-1465.

Nyrop, J.P., and van der Werf. 1994. Sampling to predict or monitor biological control. pp 245-336. In: Pedigo, L.P., and G.D. Buntin (Eds.) 1994. Handbook of Sarnpling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 7 14 p.

ONeil, K. 1960. Identification of newly introduced phlaeothripid HapIothrips cclrisetis Priesner (Thysanoptera). Am. Entomol. Soc. Am. 53: 507-5 10.

Paiiwai, Y.C. 1974. Some properties and thrips transmission of tomato spotted wilt virus in Canada. Can. J. Bot. 52: 1 177- 1 182.

Palmer, J.M., L.A. Mound, and G. J. du Heaume. 1989. CE Guide to Insects of importance to Man: 2. Thysanoptera. Betts, C.R(Ed.), CAB Intemationai,Oxon, UK. 73 p.

Parker, B-L, and M. Skimer. 1997. Lntegrated Pest management (IPM) in tree crops. pp. 6 15-63 5. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p.

Pedigo, L.P. 1994. Time sequential sampling for taking tacticai action. pp. 337-3 53. In: Pedigo, L.P., and G.D. Buntin (Eds.). 1994. Handbook of Sampliog Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 7 14 p

Pedigo, L.P., and L.G. Higley 1992. The economic injury level concept and environmental qudity: a new perspective. Am. Entomol. 38: 12-2 1.

Pedigo, L.P., and G.D. Buatin. 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 7 14 p.

Penman, DR, G.O. Osborne, S.P. Womer, RB. Chapman, and G.F. McLaren. 1982. Ethyl nicotinate: a chernical attractant for Tnrips obscuratus (Thysanoptera: Thripidae) in Stone fruit in New Zealand. J. Chem. Ecol. 8: 1299-1303.

Ramakers, P.M. J. 1995. Biological control using oligophagous predators. In: Parker, B .L., M. Skinner, and T. Lewis (Eds.). 1995. Thrips Biology and Management. Plenum Press, New York. 636 p.

Rarnakers, P.M.J., M. Dissevelt, and K. Peeters. 1989. Large scale introductions of phytoseiid predators to control thrips on cucumber. Med. Facult. Landbouww. Rijksuniv. Gent. 54: 923-929.

Riudavets, J. 1995. Predators of Frankliniella occidenialis (Pergande) and Thrips tabac; Lindeman: A review. pp. 43-87. In: van Lenteren, J-C., A.J.M. Loomans. M.G. Tommasini, S. Maini, and J. Riudavets (Eds.). 1995. Biologicai Control of Thrips Pests. Veeman Drukkers, Wageningen. 201 p.

Rodriguez-Reina, J.M., F. Ferragut, A. Camero, and M.A. Pena. 1994. Diapause in the predaceous mite A. cucumeris (Oudemans) and Amblyseius barReri (Hughes): Consequences of use in integrated control programmes. J. Appl. Entomol. 118: 44-50.

Rossiter, P.D. 1980. Onion Thrips. Queensl. Agric. J. Jan-Feb: 67-68.

Sabelis, M.W. 1985. Development. pp. 43-53. In Helie, W., and M.W. Sabelis (Eds.) 1985. Spider Mites: Their Biology, Natural Enemies and Control. Elsevier Science Publishing Company Lnc. New York. 458 p.

Sabelis, M.W. 1992. Arthropod predators. pp. 225-264. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press, Cambridge. 740 p.

Sabelis, M.W., and P.C.J. van Rijn. 1997. Predation by insects and mites. pp. 259-354. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press. Cambndge. 740 p.

Sakimura, K. 1932. Life history of Thrips tabacï L. on Emzlia sagifata and its host plant range in Hawaii. J. Econ. Entomol. 25: 884-89 1.

Salas, J. 1994. Biology and Me habits of the onion thrips (ïhrips tabaci Lindeman). Acta Hortic. 358: 3 83-3 87.

Schliephake, G., and K. Klimt. 1979. Thysanoptera, Thrips. Gustav Fischer Verlag, Jena. 477 p.

Shelton, A.M., RF. Becker, and J.T. Andalora. 1983. Varietal resistance to onion thrips (Thysanoptera: Thripidae) in processing cabbage. J. Econ. Entomol. 76: 85-86.

Shipp, J.L. 1995. Monitoring of western flower thrips on glasshouse or vegetable crops. In: Parker, B.L., M. Skinner, and T. Lewis (Eds.). 1995. Thrips Biology and Management. Plenum Press, New York. 636 p.

Shipp, J.L., aad G.H. Whitfield. 1 99 1. Functional response of the predatory mite, Amblyseirrs cucumeris, (Acari: Phytoseiidae) on western flower thrips, Frankiiniella occidentafis (Thysanoptera: Thripidae). Environ. Entomol. 20: 694-699.

Shipp, J.L., and N. ZarEa. 1991. Spatial patterns of and sarnpling methods for western flower thrips (Thysanoptera: Thripidae) on greenhouse sweet pepper. Cm. Entomol. 123: 989- 1000.

Shipp, J.L., and T.J. Gillespie. 1993. Lnfluence of temperature and water vapour pressure deficit on survival of Frank1inieih occidentalis (Thysanoptera: Thripidae). Environ. Entomol. 22: 726-732.

Shipp, J.L., and Y. Zhang. 1997. Fiight activity of the western flower thrips under different temperature and VPD conditions. pp. 43-44. In : Papadopoulos, A.P. (Ed.). 1 997. Greenhouse and Processing Crops Research Centre Annual Report. 90 p.

Shipp, J.L., and Y. Zhang. 1998. Using the greenhouse climate to improve the efficacy of insecticide applications for western flower thrips [Online]. Greenhouse and Processing Crops Research Centre Annual Report. Harrow. Available fiom: URL: http://res. a~ .ca /hmow/p:Wannre~98/ i i . [Accessed 19 May 1 9981.

Shipp, J.L., G.H. Whitfield, and N. Z d a . 1990. To develop strategies for control of thrips on greenhouse peppers. Agriculture Canada Research Station, Harrow, Ontario. 43 p.

Shipp, J.L., K.I. Ward, and T.J. Gillespie. 1996. Influence of temperature and vapour pressure deficit on the rate of predation by the predatory mite, Ambiyseitcs ammeris, on F~mkZinze~a occidentalis. Entomol. Exp. Appl. 78: 3 1 -3 8.

Shipp, J.L., D. Hunt, W.J. Jarvis, S. Barrie, and G. Ferguson. 1998. Entomopathogens for greenhouse Pest control [Online]. Greenhouse and Processing Crops research Centre Annual Report. Harrow. Available fiom: URL: httu://res.aer. calharrow/~hlannrep98/ii- ent-htm. [Accessed 19 May 19981.

Shipp, J.L., N D . Clarke, A.P. Papadopoulos, W.R. Jarvis, T.J. Jewett, S. Khosla, and G. Ferguson. 1997. Update on Harrow Greenhouse Crop Manager: A decision-support sy stem for integrated management of greenhouse cucumber and tomato. pp.45-47. Greenhouse and Processing Crops research Centre Annual Report. 90 p.

Southwood, T.RE. 1978. Ecological Methods. 2°6 Ed. Chapman and Hail, London. 524 p.

Splittstoesser, C.M. 198 1. The use of pathogens in insect control. pp. 285-298. In: Pimentel, D. (Ed.). 1 98 1 . Handbook of Pest Management in Agriculture. Vol 3. CRC Press Inc., Boca Raton. 656 p.

Statistics Canada. 1998. Fruit and Vegetable Production. Catalogue No. 22-0035XPB.

Steiner, M.Y. 1990. Determining population characteristics and sampling procedures for the western flower thrips (Thysanoptera: Thnpidae) and the predatory mite Amblyseius cucumeris (Acari: Phytoseüdae) on greenhouse cucumber. Environ. Entomol. 19: 1605- 1613.

Suman, C.L., S.D. Wahi, and N. J. Mohan. 1980. Distribution pattern of onion thrips (771rip.s tabaci Lind.). Current Sci. 49: 28-29.

Taylor, L.R 196 1. Aggregation, variance, and the mean. Nature. 189: 732-735.

Taylor, L.R, I.P. Woiwoid, and J.N. Perxy. 1978. The density dependence of spatial behaviour and the rarity of randomness. J. Anim. Ecol. 47: 383-406.

Teerling, CR, D.R Gillespie, and J.H. Borden. 1993. Wtilization of western flower thrips alami pheromone as a prey hding kairomone by predators. Can. Entomol. 125: 43 1437.

Terry, L.I. 1997. Host selectioa, communication and reproductive behaviour. pp 65-1 18. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p.

Teulon, D. A. J., and P.M. J. Ramakers. 1990. A review of attractants for trapping thrips with particular reference to glasshouses. SROP/WPRS Bull. 8: 2 1 2-2 14.

Teulon, D.A.J., D.R Penman, and P.M.J. Ramakers. 1993. Volatile chemicals for thnps (Thysanoptera: Thripidae) host-fkding and applications for thrips pest management. 1. Econ. Entomol. 86: 1405- 14 15.

Tommasini, M.G., and S. Maini. 1995. Frmkhiel la occidentalis and other thrips h d l to vegetable and ornamental crops in Europe, pp. 1-42. In: van Lenteren, J.C., A. J.M. Loomans, M.G. Tommasini, S. Maini, and J. Riudavets (Eds.). 1995. Biologicai Control of Thrips Pests. Veeman Dmkkers, Wageningen. 201 p.

Trumble, J.T. 1994. Sampling arthropod pests in vegetables. pp. 603-626. In: Pedigo, L.P. and G.D. Buntin (Eds.). 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

van de Veire, M., and V. Vacante. 1984. Greenhouse whitefly control through the combined use of the colour attraction system with the parasite wasp Encarsia formosa (Hym.: Aphelinidae). Entomophaga. 29: 303-3 10.

van der Hoeven, W.A.D., and P.C.J. van Rijn. 1990. Factors af5ecting the attack success of predatory mites on thrips larvae. Proc. Sect. Exp. Appl. Entomol. Netherlands. Entomol. SOC. 1: 25-30.

van der Staay. 199 1. Side effects of pesticides on predatory mites. Med. Fac. Landbouww. Riksuniv. Gent. 56: 3 55-3 58.

van Houten, Y., and P. van Stratum. 1993. Biological control of western flower thrips in greenhouse sweet peppers using non-diapausing predatory mites. Proc. Exp. Appl. Entomol. 4: 229-234.

van Houten, Y.M., and A-H.M. van Lier. 1995. Influence of temperature and humidity on the survival of eggs of the thrips predator, Amblyseius cucumeris. Med. Fac. Landbouww. Rijksuniv. Gent. 60: 879-884.

van Houten, Y., and P. van Stratum. 1995. Control of western flower thrips on sweet pepper in winter with Amblyseius cucumeris (Oudemans) and A. degenerans (Berlese), pp. 245- 248. In: Parker, B.L., M. Skinner, and T. Lewis (Eds.). 1995. Thrips Biology and Management. Plenum Press, New York. 636 p.

van Houten, Y.M., P. van Stratum, J. Bruin, and A. Veerman. 1995. Selection for non- diapause in Amblyseius ctrntmeris and Ambiyseius barkeri and exploration of the effectiveness of selected strains for thrips control. Entomol. Exp. Appl. 77: 289-295.

van Houten, Y.M., P.C.J. van Rijn, L.K. Tanigoshi, P. van Stratum, and J. Bruin. 1995. Preselection of predatory mites to improve year-round biological control of western flower thnps in greenhouse crops. Entomol. Exp. Appl. 74: 225-234.

van Lenterea, J-C. 1990. Integrated pest and disease management in protected crops: the inescapable fiiture. SROP/WPRS Bull. 8: 9 1-99.

van Rijn, PCJ . , and M.W. Sabelis. 1993. Does alternative food always enhance biological control? The effect of pollen on the interaction between western flower thrips and its predators. SROPiWPRS Bull. 8: 123- 125.

van Rijn, P.C. J., and M. W. Sabelis. 1997. Should plants provide food for predators when it also benefits herbivores? Effect of pollen on a thrips-predatory mite system. In: Mitchell, R., D.Z. Hom, G.R Meedham, and W.C. Weibourn (Eds.). Acarology IX : Vol. 2. Proceedings, Ohio Biological Survey, Columbus, Ohio. In: Lewis, T. (Ed.). 1997. Thrips as Crop Pests. University Press. Cambridge. 740 p.

van Rijn, PCJ . , C. Mollema, and G.M. Steenhuis-Broers. 1995. Comparative life histones of FrankZiniela occidentaiis and Thrp tabaci on cucumber. Buii. Entomol. Res. 85: 285-297.

Vernon, R.S., and D.R. Gillespie. 1995. Influence oftrap shape, size, and background colour on capture of Frankiiniella occidentalis (Thysanoptera: Thripidae) in a cucumber greenhouse. J. Econ. Entomol. 88: 288-293.

Visser, J.H. 1986. Host odour perception in phytophagous insects. Am. Rev. Environ. Entornol. 3: 12 1 - 144.

Visser, J.H., and P.G.M. Piron. 1995. OLfactory responses to plant volatiles in apterous virginoparae of the vetch aphid Megoura viciae. Entomol. Exp. Appl. 77: 37-46.

Wilde, G. 1 98 1. Wheat arthropods - pest management in agriculture. pp. 3 1 7-325. In: Pimentel, D. (Ed.). 198 1. Handbook of Pest Management in Agriculture. Vol 3. CRC Press Inc., Boca Raton. 656 p.

Wilmer, P.G. 1982. Microclimate and environmental physiology of insects. Adv. Insect Physiol. 16: 1-57. In Shipp, J.L., and T.J. Gillespie. 1993. lnnuence of temperature and water vapour pressure deficit on survival of Franklinieliu occidentalis (Thysanoptera: Thripidae). Environ. Entomol. 22: 726-732.

Wilson, L.T. 1994. Estimahg abundance, impact, and interactions among arthropods. pp. 475-5 14. In: Pedigo, L.P., and G.D. Buntin. 1994. Handbook of Sampling Methods for Arthropods in Agriculture. CRC Press Inc., Boca Raton. 714 p.

Wilson, L.T., and P.M. Room. 1983. Clumping patterns of fniit and arthropods in Cotton, with implications for binomial sarnplïng. Euviron. Entomol. f 2: 50-54.

Wilson, L.T., C. Pickel, RC. Mount, and F.G. Zalorn. 1983. Presence-absence sampling for cabbage aphid and green peach aphid (Homoptera: Aphidae) on brussel sprouts. J. Econ. Entomol. 76: 476-479.

Wilson, C.W., P.E. Shaw, RJ. Knight, S. Nagy, and M. Kiimt. 1985. Volatile constituents of carambola (Averrhoa carambola L.) J. Agric. Food Chem. 33: 199-201. In: Teulon, D.A.J., D.R Penman, and P.M.J. Ramakers. 1993. Volatile chernicals for thrips (Thysanoptera: Thripidae) host-finding and applications for thrips pest management. J. Econ. Entomol. 86: 1405- 14 15.

Yao, D. S ., and D.A. Chant. 1 989. Population growth and predation interference between two species of predatory phytoseiid mites (Acarina: Phytoseiidae) in interactive systems. Oecologia. 80: 443-455.

Zawirska, I., W. Wegorek, and J. J. Lipa. 1979. Studies on the tobacco thrips (7ïirips tabuci Lind.) and its role in the transmission of tomato spotîed wiit virus (TSWV) on tobacco. Proceedings of the MX Conference of the Scientific Institute of Plant Protection. Panstwowe Wydawiadczalna Rolnicze i Lesne Oddzial w Poznaniu, Poznan. pp. 367-278.