BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers,academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.

Dispersal of Adult Western Flower Thrips (Thysanoptera:Thripidae) on Chrysanthemum Plants: Impact of Feeding-InducedSenescence of InflorescencesAuthor(s): Marc Rhainds and Les ShippSource: Environmental Entomology, 32(5):1056-1065. 2003.Published By: Entomological Society of AmericaDOI: http://dx.doi.org/10.1603/0046-225X-32.5.1056URL: http://www.bioone.org/doi/full/10.1603/0046-225X-32.5.1056

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in thebiological, ecological, and environmental sciences. BioOne provides a sustainable onlineplatform for over 170 journals and books published by nonprofit societies, associations,museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated content indicatesyour acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercialuse. Commercial inquiries or rights and permissions requests should be directed to theindividual publisher as copyright holder.

POPULATION ECOLOGY

Dispersal of Adult Western Flower Thrips (Thysanoptera: Thripidae)on Chrysanthemum Plants: Impact of Feeding-Induced

Senescence of Inflorescences

MARC RHAINDS1 AND LES SHIPP2

Agriculture and Agri-Food Canada, Greenhouse and Processing Crops Research Centre, Harrow,Ontario N0R 1G0, Canada

Environ. Entomol. 32(5): 1056Ð1065 (2003)

ABSTRACT The current study investigated the interaction among density, feeding impact, anddispersal of western ßower thrips, Frankliniella occidentalis (Pergande), on potted ßowering chry-santhemum plants. In cage experiments using chrysanthemum plants infested with either 0, 400, 800,or 1,200 thrips, the proportion of senescent inßorescences increased with time and with the numberof thrips releasedonchrysanthemumplants. Positivecorrelationsbetween theproportionof senescentinßorescences and the density of thrips per inßorescence for different time periods indicate that thefeeding activity of thrips causes a premature senescence of inßorescences. On plants infested with 0or 400 thrips, population density slightly increased for 10Ð14 d and then leveled off; on plants infestedwith 800 or 1,200 thrips, in contrast, population density remained high for 7Ð10 d and then steadilydeclined to very low levels. A high proportion of senescent inßorescences was positively correlatedwith the proportion of females that dispersed on blue sticky cards for different time periods, whereasthe rate of dispersal bymaleswas not consistently impacted by the quality of inßorescences. Releasingadult thrips marked with ßuorescent powder in greenhouses indicated that the quality of inßores-cences meditates the dispersal behavior of adult thrips up to a distance of 4 m: females aremore likelyto disperse from senescent than healthy inßorescences and preferentially colonize healthy inßores-cencesover senescent inßorescences.Thedispersal behaviorof adult thripshas important implicationsin terms of sex-speciÞc optimal reproductive strategies, sampling procedures, and population dynam-ics.

KEY WORDS western ßower thrips, Frankliniella occidentalis, chrysanthemum, dispersal

UNDERSTANDING THE CAUSES and consequences of dis-persal is critical to determine the impact of dispersalon the spatial structure of animal populations (Nathan2001). In several phytophageous insects, crowding ofthe host plant is correlated with a high incidence ofdispersal, either as a result of high population densityper se or intense feeding by herbivores on their hostplant (Denno and Peterson 1995, Herzig 1995, Herzigand Root 1996, Dixon 1998, Rhainds et al. 2002). Den-sity-dependent dispersal has a strong inßuence on thepopulation dynamics of phytophageous insects be-cause it stabilizes local populations and generates rel-atively uniform, synchronous distribution of insects,especially when adults preferentially colonize hostswith a low abundance of conspeciÞcs (Ruxton andRohani 1998). Dispersal is of particular adaptive sig-niÞcance for insects that exploit ephemeral resourceswitha lowcarryingcapacity(Kindvall 1999,Travis andDytham 1999, Travis et al. 1999), such as is usually the

case with reproductive structures of plants (Thomp-son 1983). For several insect species, adults are morelikely to disperse when the abundance of ßowers islow (Nealis and Lomic 1994, Kuussaari et al. 1996,Matter and Roland 2002) or when fruits are infestedby conspeciÞcs (Roitberg et al. 1982, 1984, Aluja andBoller 1992).The life history of ßower thrips (Thysanoptera:

Thripidae) provides an opportunity to quantify theimpact of colonization of ßowers on the rate of dis-persal by adults. Although thrips may feed and repro-duce on vegetative plant structures, adults preferen-tially orient toward and land on ßowering plants(Yudin et al. 1988, de Jager et al. 1993, Kumar et al.1995) and are relativelymore abundant on plantswithnumerous, large ßowers (Davidson and Andrewartha1948, de Jager et al. 1995b, Pearsall 2000).Because thripsmay colonize the same host plant for

several generations, the ability to detect shifts in hostquality and disperse contributes greatly to the repro-ductive success of adults (Terry 1997). The incidenceof adult dispersal may increase with either increasingdensity of thrips per ßower (Gopinathan et al. 1981)

1 Current address: College of Tropical Agriculture and HumanResources,University ofHawaii, 7370KuamooRoad,Kapaa,HI 96746.2 E-mail: shippl@agr.gc.ca.

or decreasing availability of pollen (Forbes and Beck1954), and be highest on deteriorating hosts (Lewis1997a). Adultsmay also be least likely to colonize hostplants damaged by conspeciÞcs (Agrawal and Colfer2000). Overall, however, few studies have evaluatedthe proportion of adult thrips that disperse betweenhost plants (Kirk 1997). The current study investi-gated the dispersal behavior of western ßower thrips,Frankliniella occidentalis (Pergande) (Thysanoptera:Thripidae), on ßowering chrysanthemumplants,Den-dranthema grandiflora (Tzelev).

Frankliniella occidentalis is a polyphagous insectthat feeds on the vegetative and reproductive struc-tures of at least 240 host plants (Tommasini andMaini1995), including 34 cultivated species (Lewis et al.1997). Populations of F. occidentalis have the potentialto increase rapidly on suitable hosts, because of theirshort life cycle (�2 wk) and the high fecundity offemales (up to 230 eggs) (Robb 1989). Adults arerelatively weak ßiers, yet they have the capacity todiscriminate between hosts of various quality while inßight (Lewis 1997a). Frankliniella occidentalis canfeed on the foliage of chrysanthemum, resulting inreducedplant growth anddistortion of leaf tissue (vanDijken et al. 1994, de Jager et al. 1997). Chrysanthe-mum plants in their vegetative stage are poor hosts,however, and populations of F. occidentalis cannotgrowwhen fedexclusivelychrysanthemumleaves (in-trinsic rate of increase R � 0) (de Jager et al. 1993,1995a). Adult F. occidentalis preferentially land onßowering rather than nonßowering chrysanthemumplants and achieve higher reproductive output onßowering plants (Kumar et al. 1995, de Jager et al.1993, 1995a, b; see also Gerin et al. 1999). The highpopulationgrowthofF. occidentalisonßoweringchry-santhemum may be caused by the presence of pollen(Kirk 1987; Trichilo and Leigh 1988), although chry-

santhemum cultivars highly susceptible to thrips in-clude ßowering structures with and without pollen(de Jager et al. 1993). Inßorescences of chrysanthe-mum remain fresh for up to 4wk (Nell et al. 1990), butfeeding by F. occidentalis causes a premature necrosisof ßoral tissue (van Dijken et al. 1995). However, therelationship between thrips density and the incidenceof senescent inßorescenceshasnotbeenquantiÞed. Inaddition, no studies have yet investigated the impactof inßorescence senescence on the population dy-namics and dispersal behavior of thrips. Objectives ofthe current study were (1) to quantify the interactionamong density of thrips, senescence of chrysanthe-mum inßorescences and incidence of dispersal byadult F. occidentalis, and (2) to test the hypothesis thatinterplant dispersal by adult thrips is mediated byattributes of inßorescences.

Materials and Methods

Experimental Site and Plant Material.Experimentswere carried out at the Agriculture and Agri-FoodCanada Greenhouse and Processing Crops ResearchCentre (Harrow, Ontario, Canada) betweenMay andSeptember 2002. Flowering chrysanthemum plants(cultivar Eureka) provided by Yoder Canada (Leam-ington, Ontario, Canada) were maintained in groupsof four to Þve rooted cuttings in 15 cm diameter by 10cmhighpots, andat theonsetof experiments, theyhadattained ßowering stages 3Ð4 (50Ð75% of the inßo-rescences open) (Yoder 2001). Potted plants had onaverage 110 inßorescences per pot, with each inßo-rescence consisting of�20 peripheral ray ßorets (pet-als) and 192 disc ßorets in the button center (Fig. 1).No insecticides were applied to the chrysanthemumplants for at least 20 d before conducting experiments.

Fig. 1. Inßorescences of chrysanthemum plants, D. grandiflora (Tzelev) cultivar. Eureka, consisting of �20 peripheralpetal ßorets and 192 disc ßorets in the button center. Inßorescences were classiÞed as either healthy (right) or senescent(left: drooping, discoloured petal ßorets).

October 2003 RHAINDS AND SHIPP: DISPERSAL OF THRIPS ON CHRYSANTHEMUM 1057

Interaction Between Density of Thrips, Senescenceof Inflorescences, and Dispersal of Adults. Chrysan-themum plants were individually enclosed in 50 by 50by 90 cm high screen cagesmaintained in a controlledenvironment chamber (24�C, 30% RH, and a photo-period of 14:10 [L:D] h). A 5 by 7 cm blue trap coatedwith Tanglefoot on both sides was vertically sus-pended 25 cm above the plant canopy in each cage.Each plant was then randomly assigned to one of fourtreatments and infestedwitheither 0, 400, 800, or 1,200thrips. The experiment was replicated seven times,corresponding to a total of 28 chrysanthemum plantsindividually enclosed in screen cages. The incidenceof senescent inßorescences was evaluated 5, 7, 10, 12,14, 16, and 18 d after the release of thrips by classifyinginßorescences as either healthy or senescent (droop-ing, discolored petal ßorets; Fig. 1). On each date, thenumber of males and females trapped on the bluesticky trapswas recorded, and the abundance of thripson chrysanthemum plants estimated by washing in-ßorescences in 250-ml solutions of 70% alcohol for5Ð10 s to extract thrips. Each alcohol solution wasseparately Þltered through a Buchner funnel, and thenumber and life stage (immature [larva, prepupa,pupa], male adult, female adult) of thrips on Þlterpapers were recorded. On each date, between 6 and40 inßorescences were sampled per plant, dependingon the availability of inßorescences, with 15 inßores-cences being sampled on most dates. For each plant,healthy and senescent inßorescences were washed inseparate solutions of alcohol, with the number ofhealthy and sencescent inßorescences sampled eachdate reßecting approximately the relative incidenceofsenescence on whole chrysanthemum plants. After18 d, when all inßorescences had been sampled, thefoliage of each chrysanthemum plant was washed in1,200-ml solutions of alcohol to evaluate the abun-dance and life stage of thrips. Each pot was thentransferred back to its respective cage for 5 d, with ablue sticky card suspended 25 cm above the pot tocapture adults emerging from the soil media.Statistical analyses were conducted using the SAS

statistical package (SAS Institute 1998). For eachchrysanthemum plant and date, the abundance of im-matures and adults per inßorescence or per plant wasestimated using the number of thrips extracted fromhealthy and senescent inßorescences and the relativeincidence of healthy and senescent inßoresences onthe whole plant. The proportion of dispersing adultsper plant (y) was evaluated for males and females ondifferent dates using the ratio y � xtrap/[xtrap� xplant],in which xtrap represents the number of adults cap-tured on blue sticky traps and xplant the estimateddensity of adults per plant. Estimated density of thripsper inßorescence per plant, proportion of senescentinßorescences, and proportion of dispersing adultswere compared for plants infested with 0, 400, 800, or1,200 thrips on different dates using factorial analysisof variance (ANOVA), with date and infestation leveltreated as Þxed factors and replicate as a blockingfactor. The relationship between the number of thripsreleased per plant (x) and the abundance of thrips per

inßorescence per plant (y) was compared for differ-ent time periods (5Ð10, 12Ð14, and 16Ð18 d after re-lease) using linear (y � �o � �1x), quadratic (y � �o� �1x

2), and cubic (y � �o � �1x3) models. The

impact of thrips on inßorescence senescencewas eval-uated by regressing the density of thrips per inßores-cence per plant (independent variable) versus theproportion of senescent inßorescences per plant (de-pendent variable) for different dates. The impact ofinßorescence senescence on dispersal of adults wasevaluated for different dates by regressing the pro-portion of senescent inßorescences (independentvariable) versus the proportion of adults captured onblue sticky cards (dependent variable). Whenevernecessary, heterogeneity of variance was reduced bysubjecting data to square-root (density of thrips perinßorescence or per plant) or arcsine (proportion ofsenescent inßorescences or of dispersing adults)transformations.

Impact of Inflorescence Quality on Rates of Dis-persal by Adult Thrips. Experiments were conductedin large screen cages (4.8 by 2.8 by 2.3 m high) main-tained in 7 by 7 m greenhouses, with the ambientconditions averaging 23�C, 80%RH, andaphotoperiodof 15:9 (L:D) h. In a Þrst experiment, four chrysan-themum plants were placed at the corners of eachcage, 2.8 m from the center, with plants either havinglow (�10%) or high (�90%) incidence of senescentinßorescences. The position of plants with low andhigh proportion of senescent inßorescences was al-ternatedwithin each cage. Four 10 by 25 cmhigh bluetraps coated with Tanglefoot on both sides were ver-tically suspended 1 m above ground halfway betweeneach plant and the center of the cage. Before beingreleased in the cage, 1,200 adults were aspirated intoa plastic container and dusted with ßuorescent pow-der (Dayglo Color, Cleveland, OH) through an aspi-rator. Marked adults were released 75 cm aboveground at the center of the cage; sexing a subsampleof the marked adults indicated a sex ratio of approx-imately three to four females for onemale. On the dayafter the release of marked adults, the abundance ofmarked males and females was assessed on each trap.All inßorescences of each plant were then vigorouslytapped three times inside a white pan, and dislodgedadults were aspirated into a plastic container beforebeing transferred into a petri dish. Adults were frozenfor 1 h, and the number of marked males and femaleswas recorded for each plant. The experiment wasreplicated six times. The abundance of marked thripson plants with low and high incidence of senescentinßorescences, or on traps facing plant with variousinßorescence quality, was compared using factorialANOVA, with sex and inßorescence quality treated asÞxed factors, replicate as a blocking factor, and indi-vidual plants (two plants per treatment per replicate)as repeatedmeasurements. Heterogeneity of variancewas reduced by subjecting data to square-root trans-formations.In a second experiment, chrysanthemum plants

with low and high incidence of senescent inßores-cences were placed 4 m apart inside a screen cage,

1058 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 5

with a blue sticky trap vertically suspended 1m aboveground between the two plants. A total of 600 adultsmarkedwith ßuorescent powderwas released on eachplant, using different colors for different plants. Oneday later, the abundance ofmarkedmales and femaleson sticky traps and chrysanthemum plants was re-corded as described above. Marked adults were clas-siÞed as either “resident” (sampled on the originalplant where they had been released) or “migrant”(sampled on a new plant). The experiment was rep-licated four times, alternating between replicates thepigment color used on plants with high and low inci-dence of senescent inßorescences. The abundance ofmarked thrips on plants with low and high incidenceof sencescent inßorescences was compared using fac-torial ANOVA, with plant quality and resident statusof adults treated as Þxed factors and replicate as ablocking factor. Because few marked males were re-covered, sex was not included in the analysis. Het-erogeneity of variancewas reducedby subjecting datato square-root transformations.

Results

Interaction Between Density of Thrips, Inflores-cence Senescence, and Adult Dispersal.Apreliminarysurvey carried out with seven chrysanthemum plantsshortly before the onset of the experiment indicatedthat all plants had at least some thrips, yet the densitywas relatively low (�1 thrips on average per inßores-cence; range, 0.3Ð2.2). Most (92.4%) thrips were sam-pled on inßorescences, with only 7.6% present on thefoliage. Thrips were recovered from inßorescencesthroughout the experiment, even on plants where nothrips had been released.Comparing the relative abundance of thrips per

inßorescence on chrysanthemum plants infested withdifferent numbers of thrips revealed signiÞcant inter-actions between time and infestation level (imma-tures: F � 5.52, P � 0.0001; males: F � 2.79, P � 0.0003;females: F � 4.97, P � 0.0001; df � 18, 162 for allanalyses). These interactions indicated that variationsin population density as a function of time ßuctuatedaccording to the density of released thrips. On plantsinfested with 0 or 400 thrips, the population densityslightly increased for 10Ð14 d and then leveled off; onplants infestedwith 800 or 1,200 thrips, in contrast, thepopulation density remained high for 7Ð10 d and thensteadily declined to low levels (Fig. 2). The relation-ship between the number of thrips released per plantand the abundance of thrips per inßorescence perplant was more adequately described by linear thanquadratic or cubic models (Fig. 3). The density ofeither immatures, females, ormales increasedwith thenumber of thrips released on the plant within 10 dafter the release of adults, whereas opposite trendswere observed toward the endof the experiment (Fig.3).Comparing the relative abundance of thrips per

inßorescenceonhealthy and senescent inßorescencesindicated signiÞcant interactions between time andinßorescence quality for immatures (F � 14.41, df �

6,166, P � 0.0001) and females (F � 3.83, df � 6,166,P � 0.0013), although the trends were different forimmatures and females. Senescent inßorescences car-ried on average 10Ð15 times more immatures thanhealthy inßorescences within 7 d after the release ofthrips, whereas immatures were slightly more abun-dant on healthy inßorescences thereafter (Fig. 4). Incontrast, the abundance of females on healthy andsenescent inßorescences was similar within 7 d afterthe release of thrips; later, healthy inßorescences car-ried on average two to Þve times more females thansenescent inßorescences (Fig. 4). Variations of maleabundance on healthy and senescent inßorescencesover time exhibited a similar pattern to that of females(Fig. 4), but the interaction among time and inßores-cence quality was not signiÞcant (F � 1.31, df� 6,166,P � 0.257).Relatively few adults emerged from the potted soil

media (7.6 � 1.6 males and 16.9 � 2.4 females), andneither the abundance of males (F � 1.21, df� 3, 15,P � 0.339) nor females (F � 0.92, df� 3, 15, P � 0.455)was affected by the density of thrips released on chry-santhemumplants. The average number of thrips sam-pled on the foliage at the end of the experiment wasnot affected by the density of released thrips (imma-tures: F � 2.15, df � 3, 18, P � 0.129; males: F � 0.81,df� 3, 18, P � 0.504; females: F � 2.97, df� 3, 18, P �0.059). The density of thrips on the foliage was low(60 � 8) compared with the estimated number of

Fig. 2. Abundance of F. occidentalis on chrysanthemumplants infested with either 0, 400, 800, or 1,200 thrips.

October 2003 RHAINDS AND SHIPP: DISPERSAL OF THRIPS ON CHRYSANTHEMUM 1059

inßorescencesperplant perday, evenat theendof theexperiment when most inßorescences had been de-structively sampled (number of thrips per plant after

the onset of the experiment: 5d: 480 � 69; 7 d: 922 �138; 10 d: 1161 � 147; 12 d: 748 � 74; 14 d: 482 � 44;16 d: 246 � 25; 18 d: 97 � 15).The proportion of senescent inßorescences in-

creased asymptotically with time (F � 183.70, df� 6,162,P�0.0001)andwashigheronplants infestedwithnumerous thrips (F � 86.48, df � 3, 162, P � 0.0001;Fig. 5). The signiÞcant interaction between time andinfestation level (F � 5.33, df � 18, 162, P � 0.0001)was caused by the earlier incidence of senescence onplants infested with high number of thrips (Fig. 5).SigniÞcant relationships between density of thrips perinßorescence and the proportion of senescent inßo-rescences 5, 7, 10, and 12 d after the onset of exper-iments (Table 1) indicated that the feeding activity ofthrips contributes to the senescence of inßorescences.The low abundance of thrips on the foliage justiÞed

assessments of population density per plant based ex-clusively on inßorescences. Variations of the propor-tion of females dispersing fromchrysanthemumplantsexhibited the same pattern as for incidence of senes-cent inßorescences, with similar asymptotic incre-ments with time (F � 85.53, df � 6, 162, P � 0.0001),higher incidence on plants infested with numerousthrips (F � 5.96, df � 3, 162, P � 0.0007), and earlyincidence of dispersal on plants with initially highnumber of thrips (interaction between time and in-festation level: F � 2.56, df� 18, 162, P � 0.0010; Fig.5). The incidence of dispersal among males exhibiteda different pattern than for females, with relativelyhigh proportions of males dispersing from chrysan-themum plants at the onset of the experiment (F �18.88, df� 6, 182, P � 0.0001), when the incidence ofsenescent inßorescences was low; neither infestationlevel (F � 1.81, df � 3, 162, P � 0.147) nor theinteraction between time and infestation level (F �0.94, df� 18, 162, P � 0.535) affected the incidence ofmale dispersal (Fig. 5). Comparing the relative pro-

Fig. 3. Relationshipbetween thenumber ofF. occidentalis releasedonchrysanthemumplants and the abundanceof thripson inßorescences for different time intervals after the release of thrips. The lines indicate signiÞcant linear regressions (P �0.05).

Fig. 4. Abundance of F. occidentalis on healthy and se-nescent inßorescences of chrysanthemum plants infestedwith thrips.

1060 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 5

portion of adults dispersing from chrysanthemumplants in relationship with the incidence of senescentinßorescences indicated signiÞcant regression forthree intervals for females and only one interval formales (Table 1).

Impact of Inflorescence Quality on Rates of Dis-persal by Adult Thrips. Of 1,200 marked adults re-leased at the center of a screen cage for each of sixreplicates, 53� 17 were captured on blue sticky traps

and 100 � 22 were recovered on chrysanthemumplants. SigniÞcantly more females than males weretrapped on sticky cards (F � 32.36, df � 1,39, P �0.0001) or colonized chrysanthemum plants (F �98.42, df� 1, 39, P � 0.0001; Fig. 5). Captures of adultson sticky cards facing plants with high or low inci-denceof senescent inßorescences didnot signiÞcantlydiffer (F � 0.00, df� 1, 39, P � 0.959). The quality ofinßorescences greatly affected the number of adultscolonizing chrysanthemum plants (F � 15.39, df � 1,39, P � 0.0003), with the highest number of thripsrecovered on plants with a low incidence of senescentinßorescences (Fig. 5). The signiÞcant interaction be-tween inßorescence quality and sex (F � 11.03, df �1, 39, P � 0.002) indicated that females were muchmore likely than males to discriminate between chry-santhemum plants with high or low incidence of se-nescent inßorescences (Fig. 6).Of 600 marked adults released on each of eight

plants, 74.5 � 22.4 were recovered 24 h later on in-ßorescences, generally on the same plant on whichthey had been released (females: F � 56.58, df� 1, 9,P � 0.0001; males: F � 35.16, df� 1, 9, P � 0.0002; Fig.7). The quality of inßorescences affected the distri-bution of females (F � 38.04, df � 1, 9, P � 0.0002),with the highest density of adults on plants with lowincidence of senescent inßorescences (Fig. 7). A sim-ilar trend was observed for males (Fig. 7), but theimpact of inßorescence quality was not signiÞcant(F � 4.45, df � 1, 9, P � 0.064), Overall, 74 adultsmigrated from plants that had a high incidence ofsenescent inßorescences to plants with a low inci-dence of senescent inßorescences, whereas only 8

Fig. 5. Proportion of senescent inßorescences and inci-dence of dispersal by adult F. occidentalis on chrysanthemumplants infested with either 0, 400, 800, or 1,200 thrips.

Fig. 6. Distribution of adult F. occidentalis marked withßuorescent powder on chrysanthemum plants with low orhigh incidence of senescent inßorescences or on traps facingplants with high or low incidence of senescent inßores-cences.

Table 1. Parameters � SE of regression models describing therelationship between the density of western flower thrips (F. occi-dentalis) per inflorescence (x) and the proportion of senescentinflorescences (y) at various time periods following infestation ofchrysanthemum plants (cv. Eureka) with F. occidentalis

Day after release �0 �1 r2

5 �0.078� 0.046 0.078� 0.022 0.323a

7 �0.241� 0.058 0.199� 0.018 0.822b

10 �0.348� 0.140 0.318� 0.036 0.749b

12 �0.343� 0.220 0.416� 0.062 0.631b

14 0.958� 0.392 0.098� 0.120 0.025

Regressions were not conducted 16 and 18 days after the release ofF. occidentalis, because all inßorescences were senescent on mostplants.Heterogeneity of variance was reduced by subjecting data to

square-root (x � x) or arcine (y � sin�1 y) transformations.a P� 0.01, b P� 0.001.

October 2003 RHAINDS AND SHIPP: DISPERSAL OF THRIPS ON CHRYSANTHEMUM 1061

adults followed the reverse migration route (Fig. 7).Only nine females and two males were captured onblue sticky traps.

Discussion

Results obtained through this study indicated com-plex interaction among density, feeding impact, anddispersal behavior of F. occidentalis on ßowering chry-santhemum.Positive correlations between thedensityof F. occidentalis per inßorescence and the proportionof senescent inßorescences indicate that the feedingactivity of thrips causes a premature necrosis of ßoraltissue (Table 1). Females respond to the necrosis ofßoral tissue by dispersing from chrysanthemumplantswith numerous senescent inßorescences and prefer-entially colonizing plants with healthy inßorescences(Table 2; Figs. 5Ð7). The dispersal behavior of adult F.

occidentalis has important implications in terms ofsex-speciÞc optimal reproductive strategies, samplingprocedures, and population dynamics.

Adaptive Significance of Dispersal Behavior forMale and Female F. occidentalis. The ability to re-spond to host deterioration and disperse fromcrowded resources is highly adaptive for insects thatexploit reproductive structures of plants, because in-ßorescences are typically short-lived and have a lim-ited carrying capacity (Thompson 1983). Parasitismofpollen by ßower thrips (Sakai 2002) provides an op-portunity to evaluate the impact of crowding of in-ßorescences on rates of dispersal by adults, althoughthe evidence was until recently mostly anecdotal. Fe-male F. tritici (Fitch) and F. bispinosa (Morgan) dis-perse from the inßorescences to vegetative structuresof sweet lupine, Lupinus angustifolius (L.), after theyexhausted their source of pollen (Forbes and Beck1954). Adult Microcephalothrips abdominalis (Craw-ford) readilydisperse frominßorescencesofAgeratumconyzoides (L.) infested with numerous conspeciÞcs(Gopinathan et al. 1981). Alarm pheromone releasedby immatures signiÞcantly released the colonizationof chrysanthemum inßorescences by female F. occi-dentalis (MacDonald et al. 2002).Flowering chrysanthemum plants are highly attrac-

tive to thrips and initially allow a rapid populationgrowth (Fig. 2) (de Jager et al. 1993, 1995a, b, Kumaret al. 1995). The feeding activity of F. occidentaliscauses a premature senescence of ßoral tissue (Table1) (van Dijken et al. 1995), however, and senescentinßorescences may have a low nutritional quality, asindirectly suggested by declining density of thripsover time on plants infested with numerous thrips(Figs. 2 and 3; see also Gerin et al. 1999). The lowdensity of immatures on senescent inßorescences 10 dafter the release of thrips on chrysanthemum plants(Fig. 4), inparticular, is consistentwith thehypothesisthat senescent inßorescences have a low quality asfood resource for F. occidentalis, although it remainsunclear whether the low density of immatures wascaused by negative interactions among larvae on se-nescent inßorescences (vanDijken et al. 1995), larvaedispersing from senescent to healthy inßorescences(Kirk 1985), or females preferentially ovipositing on

Fig. 7. Distribution of adult F. occidentalis marked withßuorescent powder on chrysanthemum plants with low orhigh incidence of senescent inßorescences. Residents weresampled on the same plant where they had been released,whereas migrants dispersed on another plant.

Table 2. Parameters � SE of regression models describing the relationship between the proportion of senescent inflorescences (x)and the proportion of dispersing adult thripsa (y) at various time periodsb after infestation of chrysanthemum plants (cv. Eureka) withwestern flower thrips (F. occidentalis)

Day afterrelease

Females Males

�0 �1 r2 �0 �1 r2

5 0.318� 0.026 �0.360� 0.209 0.103 0.751� 0.074 0.894� 587 0.0827 0.263� 0.030 �0.022� 0.070 0.004 0.373� 0.055 0.010� 0.125 0.00010 0.325� 0.083 0.350� 0.087 0.387e 0.419� 0.090 0.218� 0.095 0.169c

12 0.287� 0.088 0.351� 0.074 0.465e 0.526� 0.057 0.069� 0.048 0.07514 0.505� 0.108 0.223� 0.080 0.232d 0.504� 0.091 0.104� 0.067 0.084

a For both males and females, y � ytrap/(ytrap � yplant), in which ytrap represents the number of adults captured in blue sticky traps, andyplant the estimated density of adults per plant.

b Regressions were not conducted 16 and 18 d after the release of F. occidentalis, because all inßorescences were senescent on most plants.Heterogeneity of variance was reduced by subjecting data to square-root (x � x) or arcsine (y � sin�1y) transformations. c P � 0.05,

d P � 0.01, e P � 0.001.

1062 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 5

healthy inßorescences (Fig. 4) (Gerin et al. 1999).Assuming that senescent inßorescences are poor hostsfor larval thrips, enhanced dispersal of female F. oc-cidentalis in response to a deterioration of chrysan-themum inßorescences (Table 2; Figs. 5Ð7) is ex-pected to increase their lifetime reproductive success.This hypothesis remains to be tested experimentally,however, because the nutritional quality of healthyand senescent inßorescences as food resourcewas notassessed in this study. Overall, few studies have eval-uated whether and to what extent crowding of hostplants affects the survivorship and fecundity of indi-vidual F. occidentalis (Wang and Shipp 2001).Comparing the abundance of thrips on sticky cards

and on chrysanthemum inßorescences revealed ahigher proportion of dispersing females with increas-ing incidence of senescent inßorescences (Table 2;Fig. 5). Variations in abundance of immatures andfemales on healthy and senescent inßorescences overtime suggest that females initially emerge on senes-cent inßorescences and progressively disperse tohealthy inßoresences where they may preferentiallyoviposit (Fig. 4). Releases of adult thrips marked withßuorescent powder indicate that females most readilydisperse from chrysanthemum plants with numeroussenescent inßorescenceswhile preferentially coloniz-ing plants with healthy inßorescences (Figs. 6 and 7).The low density of thrips on senescent inßorescences(Fig. 4) indicates that the incidence of dispersal is notdirectly mediated by the relative abundance of thrips.Instead, thedispersal behaviorof femaleF. occidentalisseems to be inßuenced by the deterioration of chry-santhemum inßorescences caused by the feeding ac-tivity of conspeciÞcs (Table 1).Themating systemof F. occidentalis is characterized

by swarms of males aggregating at landmarks, often insites used by females for reproduction (Terry andGardner 1990). IntraspeciÞc competition amongmales for access to females has been documented(Terry and Dyreson 1996), and males enhance theirÞtness by copulating with a large number of females.In contrast, females maximize their reproductive suc-cess by ovipositing on host plantsmost suitable for thedevelopment and survival of their progeny. Sex-spe-ciÞc foragingbehaviorof adult thripsmay thereforeberelated to distinct optimal reproductive strategies formales and females. For example, the higher ratio ofmales to females on sticky cards than on host plants(Higgins 1992, Rhainds and Shipp 2003) and the highproportion of males dispersing from chrysanthemumplants within 5 d after the release of thrips (Fig. 5)suggest that males are more active dispersers thanfemales.Thehigh incidenceofdispersal among femalethrips in response to senescence of chrysanthemuminßorescences (Table 2; Figs. 5Ð7) is consistent withthe hypothesis that the foraging behavior of females islargely inßuenced by the quality of inßorescences asfood resources for their progeny.Thedeterioriationofchrysanthemum inßorescences also enhanced the in-cidence of dispersal among male thrips, yet to a con-siderably smaller extent than for females (Table 2;Figs. 5Ð7). These results suggest that male F. occiden-

talis disperse from senescent inßorescences to trackthe abundance of females in space and time, whileselective pressures favoring dispersal from crowdedresourcesmay bemuchmore stringent for ovipositingfemales than for mate-seeking males.

Dispersal Behavior of Adult F. occidentalis andSampling Procedures. Sticky cards are commonlyused to monitor the foraging movements of adultthrips, and in some crop systems, the number of adultscaptured on sticky cards reßects the abundance ofthrips on the vegetation (Yudin et al. 1987, Shipp andZariffa 1991). The relationship between densities ofthrips on sticky cards and on their host plants is notconsistent (Higgins 1992, Jacobson 1997, Lewis 1997b,Pearsall and Myers 2000), however, which may beattributed inpart to the impactofhostplantattributesonthe dispersal behavior of thrips. Adult chillie thrips, Scir-tothrips dorsalisHood, aremore likely to be captured onsticky cards in plots with few food resources (Shibao etal. 1993). Female F. occidentalis disperse greatest dis-tances in greenhouses with least favored hosts (Robb1989), and captures of adults on sticky cards decreaseafterchrysanthemumplants initiateßowering(Kumaretal. 1995) while increasing with the proportion of senes-cent inßorescences (Figs. 5Ð7; Table 1). Sticky trapsplaced in the vicinity of cotton plants damaged by F.occidentalis attract less adults than traps nearby undam-aged cotton plants (Agrawal andColfer 2000). Capturesof adult thrips on sticky cards may therefore simulta-neously reßect the relative suitability of host plants asfood resources and the abundance of thrips in localpopulations. Moreover, sex ratios of adults captured insticky cards or sampled in the vegetation may consider-ably differ (Higgins 1992, Rhainds and Shipp 2003). Al-together, these results suggest that captures of adults onsticky traps may not provide a consistent, accurate sam-pling procedure to estimate the abundance and sex ratioof adult thrips on crops.

Impact ofDispersal on thePopulationDynamics ofF.occidentalis. As reported for other insect species thatcolonize reproductive structures of plants (Forbesand Beck 1954, Gopinathan et al. 1981, Roitberg et al.1982, 1984, Aluja and Boller 1992), previous coloni-zation of chrysanthemum inßorescences by conspe-ciÞcs affects the dispersal behavior of adult F. occi-dentalis. The high rate of emigration from plants witha high incidence of senescent inßorescences (Figs.5Ð7; Table 1) and preferential colonization of plantswith a low proportion of senescent inßorescences, upto a distance of 4 m (Figs. 6Ð7), may inßuence thepopulation dynamics of F. occidentalis. Density- ordamage-dependent dispersal coupled with the pref-erential colonization of as yet uninfested hosts theo-retically leads to uniform, synchronous distribution ofinsects (Ruxton and Rohani 1998). Although damage-dependent dispersal by females may contribute tostabilize local populations of F. occidentalis, a highlevel of polyphagy (�240 host plants) (Tommasiniand Maini 1995) combined with limited dispersal ca-pacity (Lewis 1997a) suggests that the population dy-namics of F. occidentalis in natural habitats are im-pacted by several other factors, including the relative

October 2003 RHAINDS AND SHIPP: DISPERSAL OF THRIPS ON CHRYSANTHEMUM 1063

abundance and ßowering phenology of host plants(Yudin et al. 1988), the relative lifespan of ßowers inrelationship with the duration of thripsÕ life cycle(Southwood et al. 1974, Kirk 1984, 1985), and thedistance between patches with suitable host plants(Roitberg and Prokopy 1982, Conradt et al. 2000, Cro-nin et al. 2001).In a greenhouse with extensive monoculture of

chrysanthemum, however, the dispersal behavior ofadult F. occidentalis may greatly inßuence their pop-ulation dynamics. For example, in greenhouse com-partments characterized by a continuous croppingsystem (e.g., with all plant developmental stages si-multaneously present), the preference of adult thripsfor ßowering plants (Yudin et al. 1988, de Jager et al.1993, Kumar et al. 1995) likely results in a spatiallyclumped distribution of F. occidentalis. In contrast,chrysanthemum inßorescences could be used as trapplants in greenhouse compartments characterized bymonocultures of chrysanthemum in their vegetativestage, because of the low quality of chrysanthemumleaves as food resource combined with the high at-tractiveness of ßowering plants (de Jager et al.1993,1995a, b, Kumar et al. 1995). Although ßowering chry-santhemum plants would be initially highly attractiveto F. occidentalis, they would need to be replacedperiodically, before a high level of senescent inßores-cences triggers a massive emigration of females. Be-fore a trap plant management strategy can be success-fully implemented, future studies need to assess theimpact of feeding-induced senescence of inßores-cences on the rates of between-plant dispersal byadult thrips at different spatial scales.

Acknowledgments

The authors thank D. DiMilo and M. WhiÞeld for technicalassistance.K.H.WangandS.Evansprovided thephotographofhealthy and senescent inßorescences. Yoder Canada gener-ously provided ßowering chrysanthemum plants. Funding wasprovidedbyFlowersCanada(Ontario)andtheAgricultureandAgri-Food Matching Investment Initiative Program.

References Cited

Agrawal, A. A., and R. G. Colfer. 2000. Consequences ofthrips-infested plants for attraction of conspeciÞcs andparasitoids. Ecol. Entomol. 25: 493Ð496.

Aluja, M., and E. F. Boller. 1992. Host marking pheromoneof Rhagoletis cerasi: foraging behavior in response to syn-thetic pheromonal isomers. J. Chem. Ecol. 18: 1299Ð1311.

Conradt, L., E. J. Bodsworth, T. J. Roper, and C. D. Thomas.2000. Non-random dispersal in the butterßy Maniolajurtina: implications for metapopulation models. Proc.Roy. Soc. Lond. B. 267: 1505Ð1510.

Cronin, J. T., K. Hyland, and W. G. Abrahamson. 2001. Thepattern, rate, and range of within-patch movement of astem-galling ßy. Ecol. Entomol. 26: 16Ð24.

Davidson, J., andH.G.Andrewartha. 1948. Annual trends ina natural population ofThrips imaginis (Thysanoptera). J.Anim. Ecol. 17: 193Ð199.

de Jager,C.M.,R.P.T.Butot, P.G.L.Klinkhamer, andE. vanderMeijden. 1995a. Chemical characteristics of chrysanthe-

mum cause resistance to Frankliniella occidentalis (Thysan-optera: Thripidae). Environ. Entomol. 88: 1746Ð1753.

de Jager, C. M., R.P.T. Butot, P.G.L. Klinkhamer, T. J. deJong, K. Wolff, and E. van der Meijden. 1995b. Geneticvariation in chrysanthemum for resistance to Fran-kliniella occidentalis. Entomol. Exp. Appl. 77: 277Ð287.

de Jager,C.M.,R.P.T.Butot,M.E.C.Uiterdijk, andD.vanderMeijden. 1997. Environmental inßuences on feedingcaused by western ßower thrips (Thysanoptera: Thripi-dae) to chrysanthemum. J. Econ. Entomol. 90: 188Ð194.

Denno, R. F., and M. A. Peterson. 1995. Density-dependentdispersal and its consequences for population dynamics,pp. 113Ð130. In N. Cappuccino and P. W. Price (eds.),Population dynamics: new approaches and synthesis. Ac-ademic, San Diego, CA.

Dixon, A.F.G. 1998. Aphid ecology: an optimization ap-proach. Chapman & Hall, London, United Kingdom.

Forbes, I., and E. W. Beck. 1954. A rapid biological tech-nique for screening blue lupine populations for low-al-kaloid plants. Agron. J. 46: 528Ð529.

Gerin, C., T. Hance, and G. van Impe, G. 1999. Impact ofßowers on thedemographyofwesternßower thripsFran-kliniella occidentalis (Thysan., Thripidae). J. Appl. Ento-mol. 123: 569Ð574.

Gopinathan, K., R. Varatharajan, and T. N. Ananthakrishan.1981. Incidence of Microephalothrips abdominalis(Crawford) (Thysanoptera: Insecta) in relation to thepollinationbiologyof theweedAgeratumconyzoidesLinn(Compositae). Proc. Indian Nat. Sci. Acad. B. 4: 505Ð509.

Herzig, A. L. 1995. Effects of population density on long-distance dispersal in the goldenrod beetle Trirhabda vir-gata. Ecology. 76: 2044Ð2054.

Herzig, A. L., and R. B. Root. 1996. Colonization of hostpatches following long-distance dispersal by a goldenrodbeetle, Trirhabda virgata. Ecol. Entomol. 21: 344Ð351.

Higgins, C. J. 1992. Western ßower thrips (Thysanoptera:Thripidae) in greenhouses: population dynamics, distri-bution on plants, and associationswith predators. J. Econ.Entomol. 85: 1891Ð1903.

Jacobson, R. J. 1997. Integrated pest management (IPM) inglasshouses, pp. 639Ð666. InT.Lewis (ed.),Thrips as croppests. CAB International, Wallingford, United Kingdom.

de Jager, C. M., R.P.T. Butot, T. J. de Jong, P.G.L.Klinkhamer, and E. van der Meijden, E. 1993. Popula-tion growth and survival of western ßower thrips Fran-kliniella occidentalis Pergande (Thysanoptera: Thripi-dae) on different chrysanthemum cultivars. J. Appl.Entomol. 115: 519Ð525.

Kindvall,O. 1999. Dispersal in ametapopulationof thebushcricket, Metrioptera bicolour (Orthoptera: Tettgoni-idae). J. Anim. Ecol. 68: 172Ð185.

Kirk, D. J. 1984. Ecological studies on Thrips imaginis Bag-nall (Thysanoptera) inßowers ofEchiumplantagineumL.in Australia. Aust. J. Ecol. 9: 9Ð18.

Kirk, D. J. 1985. Floral display in Vicia fabia, and the dis-tributionof aßower thrips,Kakothripspisivorus.Entomol.Exp. Appl. 38: 233Ð238.

Kirk, D. J. 1987. Howmuch pollen can thrips destroy? Ecol.Entomol 12: 31Ð40.

Kirk, D. J. 1997. Feeding, pp. 119Ð174. In T. Lewis (ed.),Thrips as crop pests. CAB International, Wallingford,United Kingdom.

Kumar,N.K.,D.E.Ullman, and J. J.Cho. 1995. Frankliniellaoccidentalis (Thysanoptera: Thripidae) landing and re-sistance to tomato spotted wilt tospovirus among Lyco-persicon accessions with additional comments on Thripstabaci (Thysanoptera: Thripidae) and Trialeurodes vapo-

1064 ENVIRONMENTAL ENTOMOLOGY Vol. 32, no. 5

rariorum (Homoptera: Aleyrodidae). Environ. Entomol.24: 513Ð520.

Kuussaari, M., M. Nieminen, and I. Hanski. 1996. An exper-imental study of migration in the Glanville fritillary but-terßy Melitaea cinxia. J. Anim. Ecol. 65: 791Ð801.

Lewis, T. 1997a. Flight and dispersal, pp. 175Ð196. In T.Lewis (ed.), Thrips as crop pests. CAB International,Wallingford, United Kingdom.

Lewis, T. 1997b. Field and laboratory techniques, pp. 435Ð475. In T. Lewis (ed.), Thrips as crop pests. CAB Inter-national, Wallingford, United Kingdom.

Lewis, T., L. A. Mound, S. Nakahara, and C. C. Childers.1997. Major crops infestedby thripswithmain symptomsand predominant injurious species, pp. 675Ð701. In T.Lewis (ed.), Thrips as crop pests. CAB International,Wallingford, United Kingdom.

MacDonald, K. M., J.G.C. Hamilton, R. Jacobson, and W.D.J.Kirk. 2002. Effects of alarm pheromone on landing andtake-off by adult western ßower thrips. Entomol. Exp.Appl. 103: 279Ð282.

Matter, S. F., and J. Roland. 2002. An experimental exami-nationof theeffects of habitat quality on thedispersal andlocal abundance of the butterßy Parnassius smintheus.Ecol. Entomol. 27: 308Ð316.

Nathan, R. 2001. The challenges of studying dispersal.Trends Ecol. Evol. 16: 481Ð483.

Nealis, V.G., andP.V.Lomic. 1994. Host-plant inßuenceonthe population ecology of the jack pine budworm, Cho-ristoneura pinus (Lepidoptera: Tortricidae). Ecol. Ento-mol. 19: 367Ð373.

Nell,T.A.,R.T.Leonard, and J.E.Barrett. 1990. Productionand postproduction irradiance affects acclimatizationand longevity of potted chrysanthemum and poinsettia.J. Am. Soc. Hort. Sci. 115: 262Ð265.

Pearsall, I.A. 2000. Flowerpreferencebehaviourofwesternßower thrips in theSimilkameenvalley,BritishColumbia.Entomol. Exp. Appl. 95: 303Ð313.

Pearsall, I. A., and J. H. Myers. 2000. Evaluation of samplingmethodology for determining the phenology, relativedensity, and dispersion of western ßower thrips (Thys-anoptera: Thripidae) in nectarine orchards. J. Econ. En-tomol. 93: 494Ð502.

Rhainds, M., and Shipp, L. 2003. Short-range dispersalmovements by adult western ßower thrips (Thysan-optera: Thripidae) in greenhouse environment: implica-tion for population dynamics. Can. Entomol. (in press).

Rhainds, M., G. Gries, C. T. Ho, and P. S. Chew. 2002.Dispersal by bagworm larvae, Metisa plana: effects ofpopulation density, larval sex, and host plant attributes.Ecol. Entomol. 27: 204Ð212.

Robb, K. L. 1989. Analysis of Frankliniella occidentalis (Per-gande) as a pest of ßoricultural crops in California green-houses. PhD thesis, University of California, Riverside,CA.

Roitberg, B. D., J. C. van Lenteren, J.J.M. van Alphen, F.Galis, and R. J. Prokopy. 1982. Foraging behaviour ofRhagoletis pomonella, a parasite of hawthorn (Crataegusviridis), in nature. J. Anim. Ecol. 51: 307Ð325.

Roitberg, B. D., and R. J. Prokopy. 1982. Inßuence of inter-tree distance on foraging behaviour of Rhagoletispomonella in the Þeld. Ecol. Entomol. 7: 437Ð442.

Roitberg, B. D., R. S. Cairl, and R. J. Prokopy. 1984. Ovi-position deterring pheromone inßuences dispersal dis-tance in tephritid ßies. Entomol. Exp. Appl. 35: 217Ð220.

Ruxton, G. D., and P. Rohani. 1998. Fitness-dependent dis-persal in metapopulations andits consequences for per-sistence and synchrony. J. Appl. Ecol. 67: 530Ð539.

Sakai, S. 2002. Areviewof brood-site pollinationmutualism:plants providing breeding sites for their pollinators. J.Plant Res. 115: 161Ð168.

SAS Institute. 1998. SAS/STATTM. UserÕs guide, release6.03. SAS Institute, Cary, NC.

Shibao, M., F. Tanaka, K. Fujisaki, and F. Nakasuji. 1993.Effects of lateral shoot cutting on population density ofthe chillie thrips, Scirtothrips dorsalis Hood (Thysan-optera: Thripidae) on grape. Appl. Entomol. Zool. 28:35Ð41.

Shipp, J. L., and N. Zariffa. 1991. Spatial patterns of andsampling methods for western ßower thrips (Thysan-optera: Thripidae) on greenhouse sweet pepper. Can.Entomol. 123: 989Ð1000.

Southwood, T.R.E., R. M. May, M. P. Hassell, and G. R.Conway. 1974. Ecological strategies and population pa-rameters. Am. Nat. 108: 791Ð804.

Terry, I. L. 1997. Host selection, communication and repro-ductive behaviour, pp. 65Ð118. In T. Lewis (ed.), Thripsas crop pests. CAB International, Wallingford, UK.

Terry, L. I., and D. Gardner. 1990. Male mating swarms inFrankliniella occidentalis (Pergande) (Thysanoptera:Thripidae). J. Insect Behav. 3: 133Ð141.

Terry, L. I., and E. Dyreson. 1996. Behavior of Frankliniellaoccidentalis (Thysanoptera: Thripidae) within aggrega-tions, and morphometric correlates of Þghting. Ann. En-tomol. Soc. Am. 89: 589Ð602.

Thompson, J. N. 1983. Selection pressures on phytophagousinsects feeding on small host plants. Oikos. 40: 438Ð444.

Tommasini, M. G., and S. Maini. 1995. Frankliniella occi-dentialis and other thrips harmful to vegetable and or-namental crops in Europe. Biological control of thrips.Wageningen Agricultural University, Wageningen, TheNetherlands.

Travis, J.M.J., and C. Dytham. 1999. Habitat persistence,habitat availability and the evolution of dispersal. Proc.Roy. Soc. Lond. B. 266: 723Ð728.

Travis, J.M.J., D. J. Murrell, and C. Dytham. 1999. The evo-lution of density-dependent dispersal. Proc. Roy. Soc.Lond. B. 266: 1837Ð1842.

Trichilo, P. J., and T. F. Leigh. 1988. Inßuence of resourcequality on the reproductive Þtness of ßower thrips (Thys-anoptera: Thripidae). Ann. Entomol. Soc. Am. 81: 64Ð70.

van Dijken, F. R., M.T.A. Dik, B. Gebala, J. de Jong, and C.Mollema. 1994. Western ßower thrips (Thysanoptera:Thripidae) effects on chrysanthemum cultivars: plantgrowth and leaf scarring in nonßowering plants. J. Econ.Entomol. 87: 1312Ð1317.

van Dijken, F. R., M. Kik, and B. Gebala. 1995. Chrysan-themum ray ßowers are an el dorado for the westernßower thrips (Frankliniella occidentalis). Proc. Sect. Exp.Appl. Entomol. Neth. Entomol. Soc. 6: 173Ð179.

Wang, K., and J. L. Shipp. 2001. Simulation model for pop-ulation dynamics of Frankliniella occidentalis (Thysan-optera: Thripidae) on greenhouse cucumber. Environ.Entomol. 30: 1073Ð1081.

Yoder Brothers’. 2001. Yoder Pot mums. Yoder Brothers,Leamington, Yoder Brothers, Inc., Barberton, OH.

Yudin, L. S., W. C. Mitchell, and J. J. Cho. 1987. Colorpreference of thrips (Thysanoptera: Thripidae) with ref-erence toaphids(Homoptera:Aphididae)and leafminersin Hawaiian lettuce farms. J. Econ. Entomol. 80: 51Ð55.

Yudin, L. S., B. E. Tabashnik, J. J. Cho, and W. C. Mitchell.1988. Colonization ofweeds and lettuce by thrips (Thys-anoptera: Thripidae). Environ. Entomol. 17: 522Ð526.

Received forpublication17December2002; accepted11July2003.

October 2003 RHAINDS AND SHIPP: DISPERSAL OF THRIPS ON CHRYSANTHEMUM 1065