INSECTICIDE RESISTANCE AND RESISTANCE MANAGEMENT

Effect of Application Technology and Bacillus thuringiensisSubspecies on Management of B thuringiensis subsp

kurstaki-resistant Diamondback Moth(Lepidoptera Plutellidae)

CARLOS J PEREZl ANTHONY M SHELTONl AND RICHARD C DERKSEN2

Dtpartmtnt of Entomology New York Agricultural Experiment Station Cornell University Geneva NY 14456

J Econ Entomol 88(5) 1113-1119 (1995)ABSTRACT Field and laboratory tests were done to determine effects of application tech-nology plant age Bacillus thuringiensis (Berliner) subspecies and rate of application on mor-tality of 2 populations of diamondback moth Plutella xylostella (L) One population wassusceptible (Geneva 88) and the other (Loxa b) was resistant to Bacillus tlwringiensis subspkurstaki In the field a knapsack a drop nozzle and an electrostatic sprayer were used toapply Javelin WG (64 [AI] (B thuringiensis subsp kurstaki) and XenTari WG (32 [AI](8 thllringiensis subsp aiZllwai) Laboratory bioassays were done with sprayed leaves Methodof application significantly affected efficacy of B thuringiensis Compared with the other twosprayers the electrostatic technique showed significantly lower variation between plant sec-tions Efficacy of XenTari increased 2-fold when applied with the electrostatic sprayer Mor-tality of Geneva 88 with either formulation was gt90 even at the lowest rates tested butJavelin was significantly more effective compared with XenTari Mortality of Loxa b withJavelin was lt60 even at double the recommended field rate Susceptibility of Loxa b toXenTari remained constant across generations whereas resistance of Loxa b to Javelin de-creased from 624-fold at F2 to 41-fold at F6 in absence of selection Javelin and XenTari didnot show cross-resistance although they share some of the S-endotoxins Studies with individualtoxins are needed for a better understanding of the use of B thuringiensis subspp kurstakiand aiawaiagainst resistant P xylostella

KEY WORDS Bacillus tlmringiensis subspecies Plutella xylostella resistance

DIAMONDBACK MOTH Plutella xylostella (L) is aworldwide pest of cmcifers Its cosmopolitan dis-tribution and occurrence in economically damag-ing numbers results from its high reproductive ca-pacity adaptation to a wide range of climatic con-ditions and wide host range (Talekar and Shelton1993) In most countries insecticide applicationsconstitute primary control (Sun et aI 1986 An-drews et al 1992 Beck and Cameron 1992) butfrequent use of synthetic insecticides has causedinsecticide-resistant populations to evolve Resis-tance to at least 46 formulations of organophos-phates carbamates pyrethroids and DDT hasheen documented (Sun et aI 1986 Miyata et aI1992 Sun 1992)

Resistance to lS-endotoxins of Bacillus thurln-gicllsis subsp kurstaki has been documented in Pxylostdla in the continental United States (Sheltonpt al 1993) Hawaii (Tabashnik et aI 1990) andMalaysia (Syed 1992) Poor control of P xylostellawith a B thurlllgiensis subsp kurstaki formulation

I Dtpartnltnt of Entomology Ntw York State Agricultural Ex-periment Station Conwll University Gtneva NY 14456

2 Dtpartnltnt of Agricultural and Biological Enginetring Cor-Itll University Ithaea NY 14853

in a field experiment with cabbage was also re-ported in the Philippines the authors suspectedresistance (Kirsch and Schmutterer 1988) With 1exception described by Shelton et aI (1993) resis-tance studies have relied on laboratory results andno data from the field other than general obser-vations of the growers have been reported How-ever several factors may be involved when a farm-er experiences lack of control Without concomi-tant documentation of resistance by laboratory andfield experiments both growers and scientists mayonly speculate about resistance when they observecontrol failures Other factors such as applicationtechnology timing rate of application and weath-er conditions can affect control of an insect pop-ulation

The efficacy of foliar sprays of insecticides de-pends in part on the spatial distribution of the ac-tive ingredient throughout the surface of the plant(Law 1982) In cabbage spray coverage is criticalbecause P xylostella larvae have a tendency to feedon the underside of leaves or in hidden places ofthe plant (Harcourt 1957) Smith et aI (1977a) re-ported differences in mortality for Trlchoplusia ni(Hiibner) exposed to soybean leaves treated with

0022-0493951113-1119$02000 copy 1995 Entnmological Socicty of America

1114 JOURNAL OF ECONOMIC ENTOMOLOCY Vol 88 no 5

B thuringiensis subsp kurstaki with use of differ-ent types of hydraulic nozzles Similar studiesshowed that application rate was more importantthan droplet size and density (number of dropletsper cm2) but the latter 2 factors were also posi-tively and significantly correlated with T ni mor-tality (Smitll et al 1977b)

The application devices preferred for use bycabbage growers have not been documented but90 of pesticides are sprayed with conventionalhydraulic nozzle sprayers (Law 1983) Dependingon the resources available to the farmer sprayequipment varies from knapsack sprayers with asingle hydraulic nozzle to tractor-mounted equip-ment (Kirsch and Schmutterer 1988 Adams 1992Andrews et al 1992 Talekar and Shelton 1993)When B thuringiensis (Berliner) formulations aresprayed with conventional terrestrial applicationequipment the volume of water used for P xylo-stella control ranges from 300 to 900 literslha (Ad-ams 1992)

With the advent of electrostatic application im-provement of the efficacy of microbial insecticidesmay be possible because electrically charged drop-let deposition on the plant is greater than that ob-tained with conventional hydraulic nozzles (Law1982) More detailed description of the physicalprinciples involved in electrostatic spraying was re-ported by Law (1983) When B thuringiensissubsp kurstaki was applied in field experiments tocontrol T ni in broccoli (Law 1980) better controlwas achieved when the microbial insecticide wasapplied with an electrostatic sprayer and only 12-1h the recommended dose per hectare was re-quired compared with the dose applied with a con-ventional sprayer The volume of total mixture perhectare applied with the electrostatic sprayer was8-fold lower than that with conventional sprayers

From the cmcifer growers perspective few op-tions are available to enhance the efficacy of Pxylostella management by using B thuringiensisformulations in the field Assuming no cross-resis-tance between B thuringiensis formulations thefarmer can either switch from one B thuringiensisformulation to another with different B-endotoxincomposition increase the dose of active ingredientper hectare or modify the application method

The objective of our study was to explore on-farm management practices that cmcifer growerscould implement against P xylostella populationswhen using B thuringiensis toxins We did fieldand laboratory experiments to determine the ef-fects of application technology on the efficacy of 2B thuringiensis formulations against susceptibleand resistant P xylostella We also conducted teststo compare the dose-response lines of susceptibleand resistant populations to 2 commercial formu-lations of B thuringiensis

Materials and MethodsEffects of Application Technology and B tlm-

ringiensis Subspecies on P xylosteUa Popula-

tions Thirty-six Bravo cabbage plots containing50 plants each were planted at the vegetable farmof the New York State Agricultural ExperilllentStation Geneva NY in the summer and fall of1992 Three application techniques and 2 COIll-mercial formulations of B thuringiensis (JavelinWG [64 (AI) (32000 IU per mg) (B thurin-giensis subsp kurstaki) lot no 0720349 SandozDes Plaines IL and XenTari we [32 (AI)(15000 IV per ml) (B thuringiensis subsp aiza-wail lot no 63-063-PG Abbott LaboratoriesNorth Chicago IL) were combined for furtherfield application Three plots were treated onlywith water on each of 4 replicates to be used ascontrols All treatments were assigned randomly tothe plots and replicated 4 times

After each application the spray was allowed todry for 1-2 h Five leaves (located on the uppermiddle and lower part of the plant) were collectedfrom the treated plots The leaves were taken tothe laboratory for insecticide residue bioassaysTwo disks (6 cm diameter each) were obtainedfrom each leaf and placed in individual plastic Pe-tri dishes (100 X 15 mm) A filter paper wettedwith distilled water was located at the bottom ofeach dish to provide moisture Five to 10 larvae (5d old) of 2 populations of P xylostella were placedin a petri dish containing a leaf disk One appli-cation was done at each of 3 cabbage growingstages precupping early head formation and headfilling (Andaloro et al 1983) In all cases the larvaewere allowed to feed for 72 h at 26 2degC 50 10 RH and a photoperiod of 168 (LD) h afterwhich mortality was recorded A larva was consid-ered dead if it did not move when prodded

Sprayers Three types of application equipmentwere used in this experiment A tractor-mollntedair-assisted electrostatic sprayer (ElectrostaticSpraying Systems Watkinsville GA) was calibratedto deliver 47literslha Four nozzles per row of cab-bage delivered =50 em above the plant canopyThe 2 nozzles in the center were in a vertical po-sition the 2 on the sides were at a slight angletoward the inside of the row An assessment of theelectrostatic charge was done before each appli-cation all nozzles delivered droplets negativelycharged with the current measuring 5-7 microam-peres in the spray cloud The amollnt of fluid andatomization delivered per nozzle resulted from 2types of pressure the liquid pressure given by tbetractor-driven hydraulic pump (14 kgcm2) and thepressure (175 kgcm2) given by the air pump atthe tip of the nozzles The air-flow was 011 m3

min per nozzleA 12-liter knapsack sprayer (SOLO Model 425

Newport News VA)with a single hoIlow-cone noz-zle was used to simulate a typical pesticide appli-cation device used by smaIl-scale cabbage growersin developing countries Applications with theknapsack sprayer were done at 315 kgcm2 and200 litersha When the cabbage plants weresprayed the nozzle was held at 45-50 em above

Octolwr 1995 PEREZ ET AI IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1115

till plants TIlt 3rd type of equipment was a CO2-assisted drop nozzle sprayer that was calibrated todeliver 360 Iitersha at 315 kgcm2 with 3 hollow-cone nozzls per row (1 in the center and 1 oneach side) A spreader-sticker (Bond Loveland In-dustritS Loveland CO) was added to all treat-ments at 02 volvol to enhance spray deposition

Bacillus tllllringiensis Formulations Javelin(B thurilliclISis subsp kurstaki) was applied at112 k~ formulated material per hectare We alsoapplied XenTari (B thuringiensis subsp aizawai)at 056 kg formulated material per hectare Theserates were the recommended field rates of bothformulations and were kept constant throughoutthe experiment

Plutella xylostella Populations We used 2 dia-mondback moth populations in these studies Ge-neva 88 previously characterized as susceptible toB thurillgiellsis subsp klrstaki formulations(Slwlton et al 1993) had been reared in thc lab-oratory (New York State Agricultural ExperimentStation Department of Entomology Geneva NY)for gt92 ~enerations without exposure to any pes-ticide The other population Loxa b was charac-terized previously as resistant to B thuringiensissubsp kurstaki formulations (Javelin Dipel andDipel 2X) Loxa b was collected in April of 1992from a commercial field in a major cabbage-grow-ing ngion of central F]orida (Loxahatchee PalmBeach County) Leaf-dip bioassays done by Shel-tOil et al (1993) revealed that F2 Loxa b had anLCoto Javelin that was 624-fo]d higher than thecorrespondin~ LCso of Geneva 88 In the samestudy the Loxa b population was tested for resis-tance to XenTari and had an LCso 28-fold higherthan the Leo of Geneva 88 A similar bioassaydone with F6 Loxa b in the absence of selectionshowed an LCso of Javelin 41-fold higher than theLCoof Gemva 88 (CJP unpublished data) Thelarvae of the 2 PxyZostella populations were reared011 rape seedlings Brassica napus L as describedby Slwlton et aJ (1991) before exposure to the Bthurilliellsis residues on leaf disks

DOlle-Response Field Test Six rates of com-mercial formulations of both Javelin (0 028 056084 112 168 and 224 kgha) and XenTari (0028 042 056 084 and 112 kgha) were appliedto plots of cabbage at the head-filling stage TheCO2-assisted drop nozzle sprayer described abovewas calibrated to deliver 300 litersha of spray mix-tun at 315 kgcm2 A spreader-sticker was addedas in previous tests A typical cabbage plot con-tained 46-50 plants arranged in 6 rows (each 4 mlong) In the field each treatment was replicatedtwice The spray was allowed to dry for 1 h afterwhich 5 Itaves per treatment (numbers 14-16counted from tIlt bottom part of the plant) werecollected from phmts sdected at random Theleaves were taktn to the laboratory for an insecti-cide residue bioassay as described above Leafdisks (6 cm diameter) were cut from the treatedleaVtsand placed in petri dishes Eight to 10 2nd-

Table 1 Number of 2nd-instar diamondback mothsused to study interactions betweell application tehnolo-g) and B thllringiensis suhspecies 011 control of P xylo-steHa at 3 cabhage stages

Crop stagePopulation Pre- Early head Headfilling Totalcupping formation

Geneva 88 437 755 798 1990Loxa b 402 780 763 1945Total 839 1535 ]56] 3935

instar P xylostella of either Geneva 88 or Loxa bpopulations were placed into each petri dish andallowed to feed for 72 h after which mortality wasrecorded A larva was considered dead if it did notmove when prodded In the laboratory each treat-ment was replicated 5 times

Analysis We lIsed analysisof variance (ANOVAGLM procedure [SYSTAT1992]) to test for effectsand interactions of application technology B thu-ringiensis subspecies P xylostella population andwithin-plant mortality variation at 3 sections of theplants ANOVAfollowed a 4-factor factorial modeland was performed for data obtained at each of 3crop stages Eventually overall data were pooledto perform ANOVA following a 5-factor factorialmodeL The purpose of the latter procedure was todetermine persistence of factor effects throughoutthe experiment

The dependent variable was the arcsine-trans-formed square root of the proportion of dead ]ar-vae (Snedecor and Cochran 1989) recorded fromeach petri dish The Abbott (1925) correction for-mula was used to correct for mortality in the con-trol groups When a treatment or interaction wassignificant their means were separated by the Tu-key highly significant difference procedure (P lt005 [SYSTAT1992]) We report backtransformedmeans of the treatments or their interactions

We used the probit regression (POLO [Russellet aL 1977]) to estimate the LDooand correspond-ing 95 CL of Loxa b and Geneva 88 to Javelinand XenTari We used POLO to conduct tests ofparallelism and equality of the responses of 2 pop-ulations treated with the same formulation and theresponse of a population treated with 2 formu]a-tions Two slopes were considered significantlydif-ferent when P lt 005 Relative toxicities were alsoestimated to compare efficacyof 2 formulations onthe same population

Results and Discussion

At precupping stage 839 P xylostella larvaewere exposed to the different treatments In thenext 2 crop stages the number of larvae tested wasincreased 2-fold We used =4000 larvae for ourexperiments (Table 1)

Effects of Application Technology Bacillusthuringiensis Subspecies and P xylostella Pop-ulation The overall effect of application technol-

1116 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 no 5

Table 2 Sununary of ANOVA for the arcsine-trans-formed percentages of P xylostella mortality across 3cabbage stages

Factor df F P value

Main effect~Application technology (A) 2 1414 0000middotmiddotB tlmringiensis ~ub~pecies (B) 1 10]65 0000middotmiddotLeaf position in the plant (C) 2 350 0031middotP xylostella population (D) 1 163606 0000middotmiddotCmp ~tage (E) 2 1006 0000middotmiddot

InteractionsAxB 2 1166 0000middotmiddotAXC 4 416 0003middotmiddotAXE 4 454 0001middotmiddotBxD 1 26276 0000middotmiddotAXBXE 4 585 0000middotmiddotBXDXE 2 1886 0000middotmiddot

Error 432

Signillcant effect 001 lt P lt 005 bullbull highly ~ignillcant effeet P lt 001

ogy on P xylostella mortality was highly significantand interacted significantly with B thuringiensissubspecies and leaf position within the plant (Table2) ANOVA done at each crop stage revealed thatthe interaction between application technologyand B thuringiensis subspecies was highly signifi-cant at early head formation and head-filling stages(P lt 00001 and P lt 0001 respectively Table 3)Variations in mortality associated with applicationtechniques were not significant (P gt 005) at pre-cupping stage when fewer leaves were on theplants and each leaf was more exposed As theplants developed toward head formation and head-filling stages however leaf coverage became crit-ical and the performance of the electrostatic spray-er was significantly better than that of the other 2application techniques (Table 3) The significantinteraction between application technology andcrop stage (Table 2) confirmed the significant vari-ations caused by application technology across the3 crop stages No significant differences were ob-served in the overall performance of the knapsackand drop nozzle sprayers (Table 3)

Table 4 Percentage of mortality plusmn SD of U~CI)tibleand resistant P xylostella at early head formation as afunction of application technology and B thuringiemusubspecies

B thuringmiddot P xylostella populationSprayer iensis

subspecies Geneva 88 100 Loxa b Fr

Electrostatic Javelin 968 ~ 004a 97 2ThXenTari 969 ~ 003a 914 35a

Knap~ack Javelin 965 009a 64 20bXenTari 878 37a 548 KOd

Drop nozzle Javelin 969 004a 174 44cXenTari 865 80a 529 ]84d

Numbers within columns or rows followed by the same letterare not ~ignillcantly different by the Tukey tpound~t (SYSTAT 1992)w = 331 r v = 12 ]44 n = 15 P = 005

Application technology affected tlle efficacy ofJavelin and XenTari on the control of susceptibleand resistant P xylostella At early head formationfor instance mortality of Loxa b (resistant to Bthuringiensis subsp kurstaki) was ~90 onlywhen XenTari was applied with the electrostaticsprayer (Table 4 Fig 1) Mortality of Geneva 88(susceptible to B thuringiensis subsp kurstaki)was ~90 regardless of the application techniqueand B thuringiensis subspecies applied (Tables 4and 5) B thuringiensis subspecies and P xylostellapopulations interacted significantly throughout theexperiments (Tables 2 and 5) Although mortalityof Loxa b from Javelin and XenTari differed sig-nificantly at precupping they had little practicalimportance because tllOse mortality levels wereconsistently low Mortality of Geneva 88 larVae atearly head formation and head-filling stages wassignificantly higher when exposed to Javelin incomparison with XenTari (Table 5)

The relatively better performance of Javelinagainst a population that is assumed susceptible toboth B thuringiensis subspecies may be a result ofdifferences in application rates and concentrationof active ingredient (see above) Javelin was ap-plied at 112 kglha whereas XenTari was appliedat 056 kgha Compared with Javelin mortality ofLoxa b from XenTari was significantly higher (Ta-

Table 3 Average percentage of P xylostella larvae mortality plusmn SD expressed us a fWlction of the interaltiollbetween application technology and B thuringiensis subspecies

B thuringiensis subspecie~Crop stage

Precupping(NS P gt 005)

Early head formation(I = 169 df = 2 144 P lt 00001)

Head filling(I = 69 df = 2 144 P = 0001)

Spmyer

Eleetro~taticKllap~ackDrop nozzleElectrostaticKnap~ackDrop nozzleElectro~taticKllap~ackDrop nozzle

Javelin

64 21a59 22a63 21a57 28a53 30a63 24a65 22a62 25a63 22a

XellTari

64 20a70t 7a67t 15a95t 2b73t 9c71 16ac92 5b75 lac73 9ac

Numbers within columns or row~ of any crop stage followed by the same letter are not signillcantly different by Tukey te~t (SYSTAT1992) (J) = 288 r v = 6 144 n = 30 P = 005) Crop stages are followed by the eorre~pollding F df and P value obtained fromthe ANOVA at that stage

October 1995 PEREZ ET AL IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1117

Fig 1 Mortality of Plutella xylostella (Loxa b) as afunction of th~ interaction between application technol-oro and B tllUringiellsis subspecies Mortality was eval-uat~d on 3 sections of the plants (upper middle andlower) ESS electrostatic spraying systems KN knapsacksprayer DN drop nozzle sprayer JAV Javelin XENXenTari

ble 5) The difference in perrormance of the 2 Bthllringicnsis subspecies on the Loxa b populationwas the result of development of resistance to BthllringiensL~ subsp kurstaki

A significant interaction of application technol-ogy B thuringiensis subspecies and leaf positionwithin the cabbage plant at early head formationis illustrated in Fig 1 Loxa b mortality showedlittle variation among leaves when exposed to res-idues of XenTari applied witll the electrostaticsprayer However when applied with the knapsackor drop nozzle sprayers mortality caused byXenTari was usually low except on tlle lowerleaves mortality of Loxa b varied widely across leafpositions (Fig 1)

Genetic Effects The differences in responsesof Loxa b and Geneva 88 to the 2 B thuringiensissubspecies accounted for most of the variations inmortality (Table 2) The overall average mortality SEM of Geneva 88 and Loxa b was 94 22and 36 13 respectively Resistance of Loxa bto B thllringiensis subsp kllrstaki had been doc-umented previously and our results are consistentwith those reported by Shelton et al (1993) Thegenetic differences are confirmed by the significant

100

80

5 60ltCI-ao 40~~

20

o

bull Upper~ Middle~ Lower

ESSlJAY ESSIXEN HfWAY HtmN OWJAY OWXEN

Sprayer by B thurlnglensls subspecies

Table 6 Dose-mortality regression data of suscepti-ble and resistant P xyloslella in sinmlated field applica-tions of B lhuringiensis

Population n Slope SEM LD90 (kWha)(95 CL)

Geneva 88 (susceptible)Javelin 285 036 060 026

(01-05)XenTari 280 111 027 05

(03--08)Loxab (resistant)

Javelin 290 230 041 63(42-126)

XenTari 259 390 085 16(09-23)

95 CL (POLO [Russellet aI 1977]) significantby t ratiotest Slope SEM gt 196

differences in mortality of the 2 populations fromJavelin whereas the application rate (112 kglha)remained constant (Table 5) We expected to ob-serve similar responses of the 2 populations toXenTari because resistance to B thuringiensissubsp aizawai had not been documented in Flor-ida where Loxa b was collected Except for appli-cations with the electrostatic sprayer the averageefficacy of XenTari against Loxa b was 560 Fail-ure to control Loxa b witll conventional applicationdevices raises concern about tlle appropriate rateof XenTari and we recommend a change in itscurrent field application rate

Field Tests There were significant dose re-sponses as determined by significant slopes for allbut Javelin when used against tlle Geneva 88 pop-ulation (Table 6) In that case the shallow slope ofthe response resulted from mortality gt90 at allrates tested The calculated LDgo of tlle Geneva88 population to Javelin (026 kglha) was lowerthan the recommended rate indicating that suscep-tible populations may be controlled by Javelin atrates lt112 kglha The LDgo of ~enTari againstGeneva 88 (05 kgha) was close to the recom-mended rate of application in the field (056 kgha) The LDgos of Loxa b from Javelin and XenTariwere 24- and 32-fold higher respectively than thecorresponding LDgos of Geneva 88 from the samematerials

Table 5 Av(ra~e percentage of mortality t SD of susceptible and resistant larvae of P yloslella exposed to 2B tlutringiensis ~uh~pccics

P xylostella populationCrop stage

Prpcupping(F ~ 175df = I 144 P lt 00001)

Early Iwad formation(F ~ 1846df = 1 144 P lt 00001)

lltad filling(F = U97 df = I 144 P lt 00001)

B thuringiensissubspecies

JavelinXenTariJavelinXenTariJavelinXenTari

Geneva 88

943 Ola890 30a967 OOla911 46b970 Ola924 30b

Loxab

218 34b388 92c1O8 34c684 139d188 34c661 92d

Numlwrs within columns or rows of any crop stage followedby the same letter are not significantlydifferent by the Tukey test(SYSTAT1992) w = 26 r v = 12 144 II = 45 P = 005) Crop stages are followedby the corresponding F df and P valuesobtailltd from tIlt ANOVAat that stage

1118 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 110 5

In these studies we identified on-farm manage-ment practices that are essential for successful con-trol of susceptible P xylostella and populations thathave developed resistance to B thuringiensissubsp kurstaki These results along with those re-ported by Shelton et al (1993) and Tabashnik etal (1990) strongly suggest that a resistance mon-itoring program is needed for B thuringiensis sub-species and that such programs should include ju-dicious use of this material to delay the onset ofresistance Because only operational factors can bemanipulated at the level of the farm to delay theonset of resistance (Georghiou and Taylor 1977)we encourage farmers and researchers to use moreefficient pesticide application methods and con-stantly to review the application rates of B thurin-giensis formulations Our results indicate that theuse of an improved application technology such asthe electrostatic application technique is equiva-lent to a low dose resistance management strat-egy (Tabashnik et al 1991 McGaughey and Wha-Ion 1992) The high efficacy obtained with theelectrostatic application device at low rates will re-duce the number of treatments per season thusincreasing the time of onset of resistance

Appropriate choice of B thuringiensis subspe-cies is of great concern in the field Despite thesimilarity of slopes Javelin seems to be a betterchoice than XenTari to manage susceptible P xylo-stella Although cross-resistance to Javelin andXenTari was not apparent a rotational scheme formanagement of resistant P xylostella and B thu-ringiensis subsp kurstaki does not seem appropri-ate for 2 reasons First this approach assumes thatresistance will decline over time Tabashnik (1994)noted that four to 6 generations are required toachieve a 10-fold reduction in resistance to B thu-ringiensis toxins in P xylostella We do not knowhowever how many generations without exposureare necessary for complete restoration of suscep-tibility In at least one population resistance hasbeen shown to be relatively stable over severalgenerations in the absence of selection pressure(AMS unpublished) Second XenTari containsCrylA toxins which are also present in Javelin andother commercial formulations containing toxins ofB thuringiensis subsp kurstaki (for example Di-pel Dipel 2X Biobit Thuricide) Such common-ality of toxin composition suggests that the use ofXenTari to control resistant populations will tendto select for resistance to the above toxins

Integrated management of P xylostella will de-pend on the conservation of its susceptibility to Bthuringiensis More effort should be invested indeveloping application techniques that result inhigh control efficacy at low application rates andat low cost Additional research on the middotefficacy ofindividual toxins is needed to design possible Bthuringiensis resistance management by use of tox-in rotations

Acknowledgments

We are grateful to R T Roush and M Hoffmann (De-partment of Entomology Cornell University) for theirvaluable advice and comments The statistical advice ofJohn Barnard (Computer Technology New York StatlAgricultural Experiment Station Gemva) is highly ap-preciated J Tang and laboratory personnel from the De-partment of Entomology (New York State AgriculturalExperiment Station Geneva) and field personnel of theDepartment of Agricultural and Biological Enginelring(Cornell University) provided valuable contributions tothis research Financial support for these investigationswas provided by different institutions including the Ful-bright Scholarship Program the Graduate School of Cor-nell University and Encyclopaedia Britannica Financialsupport for personnel and materials was supplied by sev-eral grants to AMS including the BT Managementlorking group and the New York State Cabbage Grow-ers Association

References Cited

Abbott W S 1925 A method of computing the ef-fectiveness of an insecticide J Econ Entomol 18265-267

Adams D B 1992 On-farm components of dia-mondback moth management in Georgia USA pp499-501 In N S Talekar [ed] Management of dia-mondback moth and other cmcifer pests AVRDCShanhua Taiwan

Andaloro J T K B Rose A M Sh(lton C WHoy and R F Becker 1983 Cabbage growthstages NY Food Life Sci Bull 101

Andrews K L R J Sanchez and R Cav 1992Management of diamondback moth in Central Amer-ica pp 487-497 III N S Talekar [Id] Managementof diamondback moth and other crucifer P(middotstsAVRDC Shanhua Taiwan

Beck N G and P J Cameron 1992 Developinga reduced-spray program for brassicas in New Zea-land pp 341-350 III N S Talekar [cd] Manage-ment of diamondback moth and other crucifer P(stsAVRDC Shanhua Taiwan

Georghiou G P and C E Taylor 1977 Opera-tional influences in the evolution of insecticide resis-tance J Econ Entomol 70 653-658

Harcourt D G 1957 Biology of the diamondbackmoth Plutella maculipennis (Curt) (Lepidoptera Plu-tellidae) in eastern Ontario II Life-history behaviourand host relationships Can Entomol 89 554-564

Kirsch K and H Schmutterer 1988 Low efficacyof a Bacillus thurillgiensis (Berl) formnlation in con-trolling the diamondback moth Plutella xylostella(L) in the Philippines J Appl Entomol 105 249-255

Law S E 1980 Electrostatic application of low-vol-ume microbial insecticide spray on broccoli plants JAm Soc Hortic Sci 105(6) 774-777

1982 Spatial distribution of electrostatically depositedsprays on living plants J Econ Entomol 75 542-544

1983 Electrostatic pesticide spraying concepts andpractice IEEE (Inst Electr Electron Eng) TransInd Appl Vol IA-19 No2

McGaughey W H and M E Wltalon 1992 Man-aging resistance to Bacillus thurillgiensis toxins Sci-ence (Washington DC) 258 1451-1455

1114 JOURNAL OF ECONOMIC ENTOMOLOCY Vol 88 no 5

B thuringiensis subsp kurstaki with use of differ-ent types of hydraulic nozzles Similar studiesshowed that application rate was more importantthan droplet size and density (number of dropletsper cm2) but the latter 2 factors were also posi-tively and significantly correlated with T ni mor-tality (Smitll et al 1977b)

The application devices preferred for use bycabbage growers have not been documented but90 of pesticides are sprayed with conventionalhydraulic nozzle sprayers (Law 1983) Dependingon the resources available to the farmer sprayequipment varies from knapsack sprayers with asingle hydraulic nozzle to tractor-mounted equip-ment (Kirsch and Schmutterer 1988 Adams 1992Andrews et al 1992 Talekar and Shelton 1993)When B thuringiensis (Berliner) formulations aresprayed with conventional terrestrial applicationequipment the volume of water used for P xylo-stella control ranges from 300 to 900 literslha (Ad-ams 1992)

With the advent of electrostatic application im-provement of the efficacy of microbial insecticidesmay be possible because electrically charged drop-let deposition on the plant is greater than that ob-tained with conventional hydraulic nozzles (Law1982) More detailed description of the physicalprinciples involved in electrostatic spraying was re-ported by Law (1983) When B thuringiensissubsp kurstaki was applied in field experiments tocontrol T ni in broccoli (Law 1980) better controlwas achieved when the microbial insecticide wasapplied with an electrostatic sprayer and only 12-1h the recommended dose per hectare was re-quired compared with the dose applied with a con-ventional sprayer The volume of total mixture perhectare applied with the electrostatic sprayer was8-fold lower than that with conventional sprayers

From the cmcifer growers perspective few op-tions are available to enhance the efficacy of Pxylostella management by using B thuringiensisformulations in the field Assuming no cross-resis-tance between B thuringiensis formulations thefarmer can either switch from one B thuringiensisformulation to another with different B-endotoxincomposition increase the dose of active ingredientper hectare or modify the application method

The objective of our study was to explore on-farm management practices that cmcifer growerscould implement against P xylostella populationswhen using B thuringiensis toxins We did fieldand laboratory experiments to determine the ef-fects of application technology on the efficacy of 2B thuringiensis formulations against susceptibleand resistant P xylostella We also conducted teststo compare the dose-response lines of susceptibleand resistant populations to 2 commercial formu-lations of B thuringiensis

Materials and MethodsEffects of Application Technology and B tlm-

ringiensis Subspecies on P xylosteUa Popula-

tions Thirty-six Bravo cabbage plots containing50 plants each were planted at the vegetable farmof the New York State Agricultural ExperilllentStation Geneva NY in the summer and fall of1992 Three application techniques and 2 COIll-mercial formulations of B thuringiensis (JavelinWG [64 (AI) (32000 IU per mg) (B thurin-giensis subsp kurstaki) lot no 0720349 SandozDes Plaines IL and XenTari we [32 (AI)(15000 IV per ml) (B thuringiensis subsp aiza-wail lot no 63-063-PG Abbott LaboratoriesNorth Chicago IL) were combined for furtherfield application Three plots were treated onlywith water on each of 4 replicates to be used ascontrols All treatments were assigned randomly tothe plots and replicated 4 times

After each application the spray was allowed todry for 1-2 h Five leaves (located on the uppermiddle and lower part of the plant) were collectedfrom the treated plots The leaves were taken tothe laboratory for insecticide residue bioassaysTwo disks (6 cm diameter each) were obtainedfrom each leaf and placed in individual plastic Pe-tri dishes (100 X 15 mm) A filter paper wettedwith distilled water was located at the bottom ofeach dish to provide moisture Five to 10 larvae (5d old) of 2 populations of P xylostella were placedin a petri dish containing a leaf disk One appli-cation was done at each of 3 cabbage growingstages precupping early head formation and headfilling (Andaloro et al 1983) In all cases the larvaewere allowed to feed for 72 h at 26 2degC 50 10 RH and a photoperiod of 168 (LD) h afterwhich mortality was recorded A larva was consid-ered dead if it did not move when prodded

Sprayers Three types of application equipmentwere used in this experiment A tractor-mollntedair-assisted electrostatic sprayer (ElectrostaticSpraying Systems Watkinsville GA) was calibratedto deliver 47literslha Four nozzles per row of cab-bage delivered =50 em above the plant canopyThe 2 nozzles in the center were in a vertical po-sition the 2 on the sides were at a slight angletoward the inside of the row An assessment of theelectrostatic charge was done before each appli-cation all nozzles delivered droplets negativelycharged with the current measuring 5-7 microam-peres in the spray cloud The amollnt of fluid andatomization delivered per nozzle resulted from 2types of pressure the liquid pressure given by tbetractor-driven hydraulic pump (14 kgcm2) and thepressure (175 kgcm2) given by the air pump atthe tip of the nozzles The air-flow was 011 m3

min per nozzleA 12-liter knapsack sprayer (SOLO Model 425

Newport News VA)with a single hoIlow-cone noz-zle was used to simulate a typical pesticide appli-cation device used by smaIl-scale cabbage growersin developing countries Applications with theknapsack sprayer were done at 315 kgcm2 and200 litersha When the cabbage plants weresprayed the nozzle was held at 45-50 em above

Octolwr 1995 PEREZ ET AI IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1115

till plants TIlt 3rd type of equipment was a CO2-assisted drop nozzle sprayer that was calibrated todeliver 360 Iitersha at 315 kgcm2 with 3 hollow-cone nozzls per row (1 in the center and 1 oneach side) A spreader-sticker (Bond Loveland In-dustritS Loveland CO) was added to all treat-ments at 02 volvol to enhance spray deposition

Bacillus tllllringiensis Formulations Javelin(B thurilliclISis subsp kurstaki) was applied at112 k~ formulated material per hectare We alsoapplied XenTari (B thuringiensis subsp aizawai)at 056 kg formulated material per hectare Theserates were the recommended field rates of bothformulations and were kept constant throughoutthe experiment

Plutella xylostella Populations We used 2 dia-mondback moth populations in these studies Ge-neva 88 previously characterized as susceptible toB thurillgiellsis subsp klrstaki formulations(Slwlton et al 1993) had been reared in thc lab-oratory (New York State Agricultural ExperimentStation Department of Entomology Geneva NY)for gt92 ~enerations without exposure to any pes-ticide The other population Loxa b was charac-terized previously as resistant to B thuringiensissubsp kurstaki formulations (Javelin Dipel andDipel 2X) Loxa b was collected in April of 1992from a commercial field in a major cabbage-grow-ing ngion of central F]orida (Loxahatchee PalmBeach County) Leaf-dip bioassays done by Shel-tOil et al (1993) revealed that F2 Loxa b had anLCoto Javelin that was 624-fo]d higher than thecorrespondin~ LCso of Geneva 88 In the samestudy the Loxa b population was tested for resis-tance to XenTari and had an LCso 28-fold higherthan the Leo of Geneva 88 A similar bioassaydone with F6 Loxa b in the absence of selectionshowed an LCso of Javelin 41-fold higher than theLCoof Gemva 88 (CJP unpublished data) Thelarvae of the 2 PxyZostella populations were reared011 rape seedlings Brassica napus L as describedby Slwlton et aJ (1991) before exposure to the Bthurilliellsis residues on leaf disks

DOlle-Response Field Test Six rates of com-mercial formulations of both Javelin (0 028 056084 112 168 and 224 kgha) and XenTari (0028 042 056 084 and 112 kgha) were appliedto plots of cabbage at the head-filling stage TheCO2-assisted drop nozzle sprayer described abovewas calibrated to deliver 300 litersha of spray mix-tun at 315 kgcm2 A spreader-sticker was addedas in previous tests A typical cabbage plot con-tained 46-50 plants arranged in 6 rows (each 4 mlong) In the field each treatment was replicatedtwice The spray was allowed to dry for 1 h afterwhich 5 Itaves per treatment (numbers 14-16counted from tIlt bottom part of the plant) werecollected from phmts sdected at random Theleaves were taktn to the laboratory for an insecti-cide residue bioassay as described above Leafdisks (6 cm diameter) were cut from the treatedleaVtsand placed in petri dishes Eight to 10 2nd-

Table 1 Number of 2nd-instar diamondback mothsused to study interactions betweell application tehnolo-g) and B thllringiensis suhspecies 011 control of P xylo-steHa at 3 cabhage stages

Crop stagePopulation Pre- Early head Headfilling Totalcupping formation

Geneva 88 437 755 798 1990Loxa b 402 780 763 1945Total 839 1535 ]56] 3935

instar P xylostella of either Geneva 88 or Loxa bpopulations were placed into each petri dish andallowed to feed for 72 h after which mortality wasrecorded A larva was considered dead if it did notmove when prodded In the laboratory each treat-ment was replicated 5 times

Analysis We lIsed analysisof variance (ANOVAGLM procedure [SYSTAT1992]) to test for effectsand interactions of application technology B thu-ringiensis subspecies P xylostella population andwithin-plant mortality variation at 3 sections of theplants ANOVAfollowed a 4-factor factorial modeland was performed for data obtained at each of 3crop stages Eventually overall data were pooledto perform ANOVA following a 5-factor factorialmodeL The purpose of the latter procedure was todetermine persistence of factor effects throughoutthe experiment

The dependent variable was the arcsine-trans-formed square root of the proportion of dead ]ar-vae (Snedecor and Cochran 1989) recorded fromeach petri dish The Abbott (1925) correction for-mula was used to correct for mortality in the con-trol groups When a treatment or interaction wassignificant their means were separated by the Tu-key highly significant difference procedure (P lt005 [SYSTAT1992]) We report backtransformedmeans of the treatments or their interactions

We used the probit regression (POLO [Russellet aL 1977]) to estimate the LDooand correspond-ing 95 CL of Loxa b and Geneva 88 to Javelinand XenTari We used POLO to conduct tests ofparallelism and equality of the responses of 2 pop-ulations treated with the same formulation and theresponse of a population treated with 2 formu]a-tions Two slopes were considered significantlydif-ferent when P lt 005 Relative toxicities were alsoestimated to compare efficacyof 2 formulations onthe same population

Results and Discussion

At precupping stage 839 P xylostella larvaewere exposed to the different treatments In thenext 2 crop stages the number of larvae tested wasincreased 2-fold We used =4000 larvae for ourexperiments (Table 1)

Effects of Application Technology Bacillusthuringiensis Subspecies and P xylostella Pop-ulation The overall effect of application technol-

1116 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 no 5

Table 2 Sununary of ANOVA for the arcsine-trans-formed percentages of P xylostella mortality across 3cabbage stages

Factor df F P value

Main effect~Application technology (A) 2 1414 0000middotmiddotB tlmringiensis ~ub~pecies (B) 1 10]65 0000middotmiddotLeaf position in the plant (C) 2 350 0031middotP xylostella population (D) 1 163606 0000middotmiddotCmp ~tage (E) 2 1006 0000middotmiddot

InteractionsAxB 2 1166 0000middotmiddotAXC 4 416 0003middotmiddotAXE 4 454 0001middotmiddotBxD 1 26276 0000middotmiddotAXBXE 4 585 0000middotmiddotBXDXE 2 1886 0000middotmiddot

Error 432

Signillcant effect 001 lt P lt 005 bullbull highly ~ignillcant effeet P lt 001

ogy on P xylostella mortality was highly significantand interacted significantly with B thuringiensissubspecies and leaf position within the plant (Table2) ANOVA done at each crop stage revealed thatthe interaction between application technologyand B thuringiensis subspecies was highly signifi-cant at early head formation and head-filling stages(P lt 00001 and P lt 0001 respectively Table 3)Variations in mortality associated with applicationtechniques were not significant (P gt 005) at pre-cupping stage when fewer leaves were on theplants and each leaf was more exposed As theplants developed toward head formation and head-filling stages however leaf coverage became crit-ical and the performance of the electrostatic spray-er was significantly better than that of the other 2application techniques (Table 3) The significantinteraction between application technology andcrop stage (Table 2) confirmed the significant vari-ations caused by application technology across the3 crop stages No significant differences were ob-served in the overall performance of the knapsackand drop nozzle sprayers (Table 3)

Table 4 Percentage of mortality plusmn SD of U~CI)tibleand resistant P xylostella at early head formation as afunction of application technology and B thuringiemusubspecies

B thuringmiddot P xylostella populationSprayer iensis

subspecies Geneva 88 100 Loxa b Fr

Electrostatic Javelin 968 ~ 004a 97 2ThXenTari 969 ~ 003a 914 35a

Knap~ack Javelin 965 009a 64 20bXenTari 878 37a 548 KOd

Drop nozzle Javelin 969 004a 174 44cXenTari 865 80a 529 ]84d

Numbers within columns or rows followed by the same letterare not ~ignillcantly different by the Tukey tpound~t (SYSTAT 1992)w = 331 r v = 12 ]44 n = 15 P = 005

Application technology affected tlle efficacy ofJavelin and XenTari on the control of susceptibleand resistant P xylostella At early head formationfor instance mortality of Loxa b (resistant to Bthuringiensis subsp kurstaki) was ~90 onlywhen XenTari was applied with the electrostaticsprayer (Table 4 Fig 1) Mortality of Geneva 88(susceptible to B thuringiensis subsp kurstaki)was ~90 regardless of the application techniqueand B thuringiensis subspecies applied (Tables 4and 5) B thuringiensis subspecies and P xylostellapopulations interacted significantly throughout theexperiments (Tables 2 and 5) Although mortalityof Loxa b from Javelin and XenTari differed sig-nificantly at precupping they had little practicalimportance because tllOse mortality levels wereconsistently low Mortality of Geneva 88 larVae atearly head formation and head-filling stages wassignificantly higher when exposed to Javelin incomparison with XenTari (Table 5)

The relatively better performance of Javelinagainst a population that is assumed susceptible toboth B thuringiensis subspecies may be a result ofdifferences in application rates and concentrationof active ingredient (see above) Javelin was ap-plied at 112 kglha whereas XenTari was appliedat 056 kgha Compared with Javelin mortality ofLoxa b from XenTari was significantly higher (Ta-

Table 3 Average percentage of P xylostella larvae mortality plusmn SD expressed us a fWlction of the interaltiollbetween application technology and B thuringiensis subspecies

B thuringiensis subspecie~Crop stage

Precupping(NS P gt 005)

Early head formation(I = 169 df = 2 144 P lt 00001)

Head filling(I = 69 df = 2 144 P = 0001)

Spmyer

Eleetro~taticKllap~ackDrop nozzleElectrostaticKnap~ackDrop nozzleElectro~taticKllap~ackDrop nozzle

Javelin

64 21a59 22a63 21a57 28a53 30a63 24a65 22a62 25a63 22a

XellTari

64 20a70t 7a67t 15a95t 2b73t 9c71 16ac92 5b75 lac73 9ac

Numbers within columns or row~ of any crop stage followed by the same letter are not signillcantly different by Tukey te~t (SYSTAT1992) (J) = 288 r v = 6 144 n = 30 P = 005) Crop stages are followed by the eorre~pollding F df and P value obtained fromthe ANOVA at that stage

October 1995 PEREZ ET AL IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1117

Fig 1 Mortality of Plutella xylostella (Loxa b) as afunction of th~ interaction between application technol-oro and B tllUringiellsis subspecies Mortality was eval-uat~d on 3 sections of the plants (upper middle andlower) ESS electrostatic spraying systems KN knapsacksprayer DN drop nozzle sprayer JAV Javelin XENXenTari

ble 5) The difference in perrormance of the 2 Bthllringicnsis subspecies on the Loxa b populationwas the result of development of resistance to BthllringiensL~ subsp kurstaki

A significant interaction of application technol-ogy B thuringiensis subspecies and leaf positionwithin the cabbage plant at early head formationis illustrated in Fig 1 Loxa b mortality showedlittle variation among leaves when exposed to res-idues of XenTari applied witll the electrostaticsprayer However when applied with the knapsackor drop nozzle sprayers mortality caused byXenTari was usually low except on tlle lowerleaves mortality of Loxa b varied widely across leafpositions (Fig 1)

Genetic Effects The differences in responsesof Loxa b and Geneva 88 to the 2 B thuringiensissubspecies accounted for most of the variations inmortality (Table 2) The overall average mortality SEM of Geneva 88 and Loxa b was 94 22and 36 13 respectively Resistance of Loxa bto B thllringiensis subsp kllrstaki had been doc-umented previously and our results are consistentwith those reported by Shelton et al (1993) Thegenetic differences are confirmed by the significant

100

80

5 60ltCI-ao 40~~

20

o

bull Upper~ Middle~ Lower

ESSlJAY ESSIXEN HfWAY HtmN OWJAY OWXEN

Sprayer by B thurlnglensls subspecies

Table 6 Dose-mortality regression data of suscepti-ble and resistant P xyloslella in sinmlated field applica-tions of B lhuringiensis

Population n Slope SEM LD90 (kWha)(95 CL)

Geneva 88 (susceptible)Javelin 285 036 060 026

(01-05)XenTari 280 111 027 05

(03--08)Loxab (resistant)

Javelin 290 230 041 63(42-126)

XenTari 259 390 085 16(09-23)

95 CL (POLO [Russellet aI 1977]) significantby t ratiotest Slope SEM gt 196

differences in mortality of the 2 populations fromJavelin whereas the application rate (112 kglha)remained constant (Table 5) We expected to ob-serve similar responses of the 2 populations toXenTari because resistance to B thuringiensissubsp aizawai had not been documented in Flor-ida where Loxa b was collected Except for appli-cations with the electrostatic sprayer the averageefficacy of XenTari against Loxa b was 560 Fail-ure to control Loxa b witll conventional applicationdevices raises concern about tlle appropriate rateof XenTari and we recommend a change in itscurrent field application rate

Field Tests There were significant dose re-sponses as determined by significant slopes for allbut Javelin when used against tlle Geneva 88 pop-ulation (Table 6) In that case the shallow slope ofthe response resulted from mortality gt90 at allrates tested The calculated LDgo of tlle Geneva88 population to Javelin (026 kglha) was lowerthan the recommended rate indicating that suscep-tible populations may be controlled by Javelin atrates lt112 kglha The LDgo of ~enTari againstGeneva 88 (05 kgha) was close to the recom-mended rate of application in the field (056 kgha) The LDgos of Loxa b from Javelin and XenTariwere 24- and 32-fold higher respectively than thecorresponding LDgos of Geneva 88 from the samematerials

Table 5 Av(ra~e percentage of mortality t SD of susceptible and resistant larvae of P yloslella exposed to 2B tlutringiensis ~uh~pccics

P xylostella populationCrop stage

Prpcupping(F ~ 175df = I 144 P lt 00001)

Early Iwad formation(F ~ 1846df = 1 144 P lt 00001)

lltad filling(F = U97 df = I 144 P lt 00001)

B thuringiensissubspecies

JavelinXenTariJavelinXenTariJavelinXenTari

Geneva 88

943 Ola890 30a967 OOla911 46b970 Ola924 30b

Loxab

218 34b388 92c1O8 34c684 139d188 34c661 92d

Numlwrs within columns or rows of any crop stage followedby the same letter are not significantlydifferent by the Tukey test(SYSTAT1992) w = 26 r v = 12 144 II = 45 P = 005) Crop stages are followedby the corresponding F df and P valuesobtailltd from tIlt ANOVAat that stage

1118 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 110 5

In these studies we identified on-farm manage-ment practices that are essential for successful con-trol of susceptible P xylostella and populations thathave developed resistance to B thuringiensissubsp kurstaki These results along with those re-ported by Shelton et al (1993) and Tabashnik etal (1990) strongly suggest that a resistance mon-itoring program is needed for B thuringiensis sub-species and that such programs should include ju-dicious use of this material to delay the onset ofresistance Because only operational factors can bemanipulated at the level of the farm to delay theonset of resistance (Georghiou and Taylor 1977)we encourage farmers and researchers to use moreefficient pesticide application methods and con-stantly to review the application rates of B thurin-giensis formulations Our results indicate that theuse of an improved application technology such asthe electrostatic application technique is equiva-lent to a low dose resistance management strat-egy (Tabashnik et al 1991 McGaughey and Wha-Ion 1992) The high efficacy obtained with theelectrostatic application device at low rates will re-duce the number of treatments per season thusincreasing the time of onset of resistance

Appropriate choice of B thuringiensis subspe-cies is of great concern in the field Despite thesimilarity of slopes Javelin seems to be a betterchoice than XenTari to manage susceptible P xylo-stella Although cross-resistance to Javelin andXenTari was not apparent a rotational scheme formanagement of resistant P xylostella and B thu-ringiensis subsp kurstaki does not seem appropri-ate for 2 reasons First this approach assumes thatresistance will decline over time Tabashnik (1994)noted that four to 6 generations are required toachieve a 10-fold reduction in resistance to B thu-ringiensis toxins in P xylostella We do not knowhowever how many generations without exposureare necessary for complete restoration of suscep-tibility In at least one population resistance hasbeen shown to be relatively stable over severalgenerations in the absence of selection pressure(AMS unpublished) Second XenTari containsCrylA toxins which are also present in Javelin andother commercial formulations containing toxins ofB thuringiensis subsp kurstaki (for example Di-pel Dipel 2X Biobit Thuricide) Such common-ality of toxin composition suggests that the use ofXenTari to control resistant populations will tendto select for resistance to the above toxins

Integrated management of P xylostella will de-pend on the conservation of its susceptibility to Bthuringiensis More effort should be invested indeveloping application techniques that result inhigh control efficacy at low application rates andat low cost Additional research on the middotefficacy ofindividual toxins is needed to design possible Bthuringiensis resistance management by use of tox-in rotations

Acknowledgments

We are grateful to R T Roush and M Hoffmann (De-partment of Entomology Cornell University) for theirvaluable advice and comments The statistical advice ofJohn Barnard (Computer Technology New York StatlAgricultural Experiment Station Gemva) is highly ap-preciated J Tang and laboratory personnel from the De-partment of Entomology (New York State AgriculturalExperiment Station Geneva) and field personnel of theDepartment of Agricultural and Biological Enginelring(Cornell University) provided valuable contributions tothis research Financial support for these investigationswas provided by different institutions including the Ful-bright Scholarship Program the Graduate School of Cor-nell University and Encyclopaedia Britannica Financialsupport for personnel and materials was supplied by sev-eral grants to AMS including the BT Managementlorking group and the New York State Cabbage Grow-ers Association

References Cited

Abbott W S 1925 A method of computing the ef-fectiveness of an insecticide J Econ Entomol 18265-267

Adams D B 1992 On-farm components of dia-mondback moth management in Georgia USA pp499-501 In N S Talekar [ed] Management of dia-mondback moth and other cmcifer pests AVRDCShanhua Taiwan

Andaloro J T K B Rose A M Sh(lton C WHoy and R F Becker 1983 Cabbage growthstages NY Food Life Sci Bull 101

Andrews K L R J Sanchez and R Cav 1992Management of diamondback moth in Central Amer-ica pp 487-497 III N S Talekar [Id] Managementof diamondback moth and other crucifer P(middotstsAVRDC Shanhua Taiwan

Beck N G and P J Cameron 1992 Developinga reduced-spray program for brassicas in New Zea-land pp 341-350 III N S Talekar [cd] Manage-ment of diamondback moth and other crucifer P(stsAVRDC Shanhua Taiwan

Georghiou G P and C E Taylor 1977 Opera-tional influences in the evolution of insecticide resis-tance J Econ Entomol 70 653-658

Harcourt D G 1957 Biology of the diamondbackmoth Plutella maculipennis (Curt) (Lepidoptera Plu-tellidae) in eastern Ontario II Life-history behaviourand host relationships Can Entomol 89 554-564

Kirsch K and H Schmutterer 1988 Low efficacyof a Bacillus thurillgiensis (Berl) formnlation in con-trolling the diamondback moth Plutella xylostella(L) in the Philippines J Appl Entomol 105 249-255

Law S E 1980 Electrostatic application of low-vol-ume microbial insecticide spray on broccoli plants JAm Soc Hortic Sci 105(6) 774-777

1982 Spatial distribution of electrostatically depositedsprays on living plants J Econ Entomol 75 542-544

1983 Electrostatic pesticide spraying concepts andpractice IEEE (Inst Electr Electron Eng) TransInd Appl Vol IA-19 No2

McGaughey W H and M E Wltalon 1992 Man-aging resistance to Bacillus thurillgiensis toxins Sci-ence (Washington DC) 258 1451-1455

Octolwr 1995 PEREZ ET AI IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1115

till plants TIlt 3rd type of equipment was a CO2-assisted drop nozzle sprayer that was calibrated todeliver 360 Iitersha at 315 kgcm2 with 3 hollow-cone nozzls per row (1 in the center and 1 oneach side) A spreader-sticker (Bond Loveland In-dustritS Loveland CO) was added to all treat-ments at 02 volvol to enhance spray deposition

Bacillus tllllringiensis Formulations Javelin(B thurilliclISis subsp kurstaki) was applied at112 k~ formulated material per hectare We alsoapplied XenTari (B thuringiensis subsp aizawai)at 056 kg formulated material per hectare Theserates were the recommended field rates of bothformulations and were kept constant throughoutthe experiment

Plutella xylostella Populations We used 2 dia-mondback moth populations in these studies Ge-neva 88 previously characterized as susceptible toB thurillgiellsis subsp klrstaki formulations(Slwlton et al 1993) had been reared in thc lab-oratory (New York State Agricultural ExperimentStation Department of Entomology Geneva NY)for gt92 ~enerations without exposure to any pes-ticide The other population Loxa b was charac-terized previously as resistant to B thuringiensissubsp kurstaki formulations (Javelin Dipel andDipel 2X) Loxa b was collected in April of 1992from a commercial field in a major cabbage-grow-ing ngion of central F]orida (Loxahatchee PalmBeach County) Leaf-dip bioassays done by Shel-tOil et al (1993) revealed that F2 Loxa b had anLCoto Javelin that was 624-fo]d higher than thecorrespondin~ LCso of Geneva 88 In the samestudy the Loxa b population was tested for resis-tance to XenTari and had an LCso 28-fold higherthan the Leo of Geneva 88 A similar bioassaydone with F6 Loxa b in the absence of selectionshowed an LCso of Javelin 41-fold higher than theLCoof Gemva 88 (CJP unpublished data) Thelarvae of the 2 PxyZostella populations were reared011 rape seedlings Brassica napus L as describedby Slwlton et aJ (1991) before exposure to the Bthurilliellsis residues on leaf disks

DOlle-Response Field Test Six rates of com-mercial formulations of both Javelin (0 028 056084 112 168 and 224 kgha) and XenTari (0028 042 056 084 and 112 kgha) were appliedto plots of cabbage at the head-filling stage TheCO2-assisted drop nozzle sprayer described abovewas calibrated to deliver 300 litersha of spray mix-tun at 315 kgcm2 A spreader-sticker was addedas in previous tests A typical cabbage plot con-tained 46-50 plants arranged in 6 rows (each 4 mlong) In the field each treatment was replicatedtwice The spray was allowed to dry for 1 h afterwhich 5 Itaves per treatment (numbers 14-16counted from tIlt bottom part of the plant) werecollected from phmts sdected at random Theleaves were taktn to the laboratory for an insecti-cide residue bioassay as described above Leafdisks (6 cm diameter) were cut from the treatedleaVtsand placed in petri dishes Eight to 10 2nd-

Table 1 Number of 2nd-instar diamondback mothsused to study interactions betweell application tehnolo-g) and B thllringiensis suhspecies 011 control of P xylo-steHa at 3 cabhage stages

Crop stagePopulation Pre- Early head Headfilling Totalcupping formation

Geneva 88 437 755 798 1990Loxa b 402 780 763 1945Total 839 1535 ]56] 3935

instar P xylostella of either Geneva 88 or Loxa bpopulations were placed into each petri dish andallowed to feed for 72 h after which mortality wasrecorded A larva was considered dead if it did notmove when prodded In the laboratory each treat-ment was replicated 5 times

Analysis We lIsed analysisof variance (ANOVAGLM procedure [SYSTAT1992]) to test for effectsand interactions of application technology B thu-ringiensis subspecies P xylostella population andwithin-plant mortality variation at 3 sections of theplants ANOVAfollowed a 4-factor factorial modeland was performed for data obtained at each of 3crop stages Eventually overall data were pooledto perform ANOVA following a 5-factor factorialmodeL The purpose of the latter procedure was todetermine persistence of factor effects throughoutthe experiment

The dependent variable was the arcsine-trans-formed square root of the proportion of dead ]ar-vae (Snedecor and Cochran 1989) recorded fromeach petri dish The Abbott (1925) correction for-mula was used to correct for mortality in the con-trol groups When a treatment or interaction wassignificant their means were separated by the Tu-key highly significant difference procedure (P lt005 [SYSTAT1992]) We report backtransformedmeans of the treatments or their interactions

We used the probit regression (POLO [Russellet aL 1977]) to estimate the LDooand correspond-ing 95 CL of Loxa b and Geneva 88 to Javelinand XenTari We used POLO to conduct tests ofparallelism and equality of the responses of 2 pop-ulations treated with the same formulation and theresponse of a population treated with 2 formu]a-tions Two slopes were considered significantlydif-ferent when P lt 005 Relative toxicities were alsoestimated to compare efficacyof 2 formulations onthe same population

Results and Discussion

At precupping stage 839 P xylostella larvaewere exposed to the different treatments In thenext 2 crop stages the number of larvae tested wasincreased 2-fold We used =4000 larvae for ourexperiments (Table 1)

Effects of Application Technology Bacillusthuringiensis Subspecies and P xylostella Pop-ulation The overall effect of application technol-

1116 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 no 5

Table 2 Sununary of ANOVA for the arcsine-trans-formed percentages of P xylostella mortality across 3cabbage stages

Factor df F P value

Main effect~Application technology (A) 2 1414 0000middotmiddotB tlmringiensis ~ub~pecies (B) 1 10]65 0000middotmiddotLeaf position in the plant (C) 2 350 0031middotP xylostella population (D) 1 163606 0000middotmiddotCmp ~tage (E) 2 1006 0000middotmiddot

InteractionsAxB 2 1166 0000middotmiddotAXC 4 416 0003middotmiddotAXE 4 454 0001middotmiddotBxD 1 26276 0000middotmiddotAXBXE 4 585 0000middotmiddotBXDXE 2 1886 0000middotmiddot

Error 432

Signillcant effect 001 lt P lt 005 bullbull highly ~ignillcant effeet P lt 001

ogy on P xylostella mortality was highly significantand interacted significantly with B thuringiensissubspecies and leaf position within the plant (Table2) ANOVA done at each crop stage revealed thatthe interaction between application technologyand B thuringiensis subspecies was highly signifi-cant at early head formation and head-filling stages(P lt 00001 and P lt 0001 respectively Table 3)Variations in mortality associated with applicationtechniques were not significant (P gt 005) at pre-cupping stage when fewer leaves were on theplants and each leaf was more exposed As theplants developed toward head formation and head-filling stages however leaf coverage became crit-ical and the performance of the electrostatic spray-er was significantly better than that of the other 2application techniques (Table 3) The significantinteraction between application technology andcrop stage (Table 2) confirmed the significant vari-ations caused by application technology across the3 crop stages No significant differences were ob-served in the overall performance of the knapsackand drop nozzle sprayers (Table 3)

Table 4 Percentage of mortality plusmn SD of U~CI)tibleand resistant P xylostella at early head formation as afunction of application technology and B thuringiemusubspecies

B thuringmiddot P xylostella populationSprayer iensis

subspecies Geneva 88 100 Loxa b Fr

Electrostatic Javelin 968 ~ 004a 97 2ThXenTari 969 ~ 003a 914 35a

Knap~ack Javelin 965 009a 64 20bXenTari 878 37a 548 KOd

Drop nozzle Javelin 969 004a 174 44cXenTari 865 80a 529 ]84d

Numbers within columns or rows followed by the same letterare not ~ignillcantly different by the Tukey tpound~t (SYSTAT 1992)w = 331 r v = 12 ]44 n = 15 P = 005

Application technology affected tlle efficacy ofJavelin and XenTari on the control of susceptibleand resistant P xylostella At early head formationfor instance mortality of Loxa b (resistant to Bthuringiensis subsp kurstaki) was ~90 onlywhen XenTari was applied with the electrostaticsprayer (Table 4 Fig 1) Mortality of Geneva 88(susceptible to B thuringiensis subsp kurstaki)was ~90 regardless of the application techniqueand B thuringiensis subspecies applied (Tables 4and 5) B thuringiensis subspecies and P xylostellapopulations interacted significantly throughout theexperiments (Tables 2 and 5) Although mortalityof Loxa b from Javelin and XenTari differed sig-nificantly at precupping they had little practicalimportance because tllOse mortality levels wereconsistently low Mortality of Geneva 88 larVae atearly head formation and head-filling stages wassignificantly higher when exposed to Javelin incomparison with XenTari (Table 5)

The relatively better performance of Javelinagainst a population that is assumed susceptible toboth B thuringiensis subspecies may be a result ofdifferences in application rates and concentrationof active ingredient (see above) Javelin was ap-plied at 112 kglha whereas XenTari was appliedat 056 kgha Compared with Javelin mortality ofLoxa b from XenTari was significantly higher (Ta-

Table 3 Average percentage of P xylostella larvae mortality plusmn SD expressed us a fWlction of the interaltiollbetween application technology and B thuringiensis subspecies

B thuringiensis subspecie~Crop stage

Precupping(NS P gt 005)

Early head formation(I = 169 df = 2 144 P lt 00001)

Head filling(I = 69 df = 2 144 P = 0001)

Spmyer

Eleetro~taticKllap~ackDrop nozzleElectrostaticKnap~ackDrop nozzleElectro~taticKllap~ackDrop nozzle

Javelin

64 21a59 22a63 21a57 28a53 30a63 24a65 22a62 25a63 22a

XellTari

64 20a70t 7a67t 15a95t 2b73t 9c71 16ac92 5b75 lac73 9ac

Numbers within columns or row~ of any crop stage followed by the same letter are not signillcantly different by Tukey te~t (SYSTAT1992) (J) = 288 r v = 6 144 n = 30 P = 005) Crop stages are followed by the eorre~pollding F df and P value obtained fromthe ANOVA at that stage

October 1995 PEREZ ET AL IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1117

Fig 1 Mortality of Plutella xylostella (Loxa b) as afunction of th~ interaction between application technol-oro and B tllUringiellsis subspecies Mortality was eval-uat~d on 3 sections of the plants (upper middle andlower) ESS electrostatic spraying systems KN knapsacksprayer DN drop nozzle sprayer JAV Javelin XENXenTari

ble 5) The difference in perrormance of the 2 Bthllringicnsis subspecies on the Loxa b populationwas the result of development of resistance to BthllringiensL~ subsp kurstaki

A significant interaction of application technol-ogy B thuringiensis subspecies and leaf positionwithin the cabbage plant at early head formationis illustrated in Fig 1 Loxa b mortality showedlittle variation among leaves when exposed to res-idues of XenTari applied witll the electrostaticsprayer However when applied with the knapsackor drop nozzle sprayers mortality caused byXenTari was usually low except on tlle lowerleaves mortality of Loxa b varied widely across leafpositions (Fig 1)

Genetic Effects The differences in responsesof Loxa b and Geneva 88 to the 2 B thuringiensissubspecies accounted for most of the variations inmortality (Table 2) The overall average mortality SEM of Geneva 88 and Loxa b was 94 22and 36 13 respectively Resistance of Loxa bto B thllringiensis subsp kllrstaki had been doc-umented previously and our results are consistentwith those reported by Shelton et al (1993) Thegenetic differences are confirmed by the significant

100

80

5 60ltCI-ao 40~~

20

o

bull Upper~ Middle~ Lower

ESSlJAY ESSIXEN HfWAY HtmN OWJAY OWXEN

Sprayer by B thurlnglensls subspecies

Table 6 Dose-mortality regression data of suscepti-ble and resistant P xyloslella in sinmlated field applica-tions of B lhuringiensis

Population n Slope SEM LD90 (kWha)(95 CL)

Geneva 88 (susceptible)Javelin 285 036 060 026

(01-05)XenTari 280 111 027 05

(03--08)Loxab (resistant)

Javelin 290 230 041 63(42-126)

XenTari 259 390 085 16(09-23)

95 CL (POLO [Russellet aI 1977]) significantby t ratiotest Slope SEM gt 196

differences in mortality of the 2 populations fromJavelin whereas the application rate (112 kglha)remained constant (Table 5) We expected to ob-serve similar responses of the 2 populations toXenTari because resistance to B thuringiensissubsp aizawai had not been documented in Flor-ida where Loxa b was collected Except for appli-cations with the electrostatic sprayer the averageefficacy of XenTari against Loxa b was 560 Fail-ure to control Loxa b witll conventional applicationdevices raises concern about tlle appropriate rateof XenTari and we recommend a change in itscurrent field application rate

Field Tests There were significant dose re-sponses as determined by significant slopes for allbut Javelin when used against tlle Geneva 88 pop-ulation (Table 6) In that case the shallow slope ofthe response resulted from mortality gt90 at allrates tested The calculated LDgo of tlle Geneva88 population to Javelin (026 kglha) was lowerthan the recommended rate indicating that suscep-tible populations may be controlled by Javelin atrates lt112 kglha The LDgo of ~enTari againstGeneva 88 (05 kgha) was close to the recom-mended rate of application in the field (056 kgha) The LDgos of Loxa b from Javelin and XenTariwere 24- and 32-fold higher respectively than thecorresponding LDgos of Geneva 88 from the samematerials

Table 5 Av(ra~e percentage of mortality t SD of susceptible and resistant larvae of P yloslella exposed to 2B tlutringiensis ~uh~pccics

P xylostella populationCrop stage

Prpcupping(F ~ 175df = I 144 P lt 00001)

Early Iwad formation(F ~ 1846df = 1 144 P lt 00001)

lltad filling(F = U97 df = I 144 P lt 00001)

B thuringiensissubspecies

JavelinXenTariJavelinXenTariJavelinXenTari

Geneva 88

943 Ola890 30a967 OOla911 46b970 Ola924 30b

Loxab

218 34b388 92c1O8 34c684 139d188 34c661 92d

Numlwrs within columns or rows of any crop stage followedby the same letter are not significantlydifferent by the Tukey test(SYSTAT1992) w = 26 r v = 12 144 II = 45 P = 005) Crop stages are followedby the corresponding F df and P valuesobtailltd from tIlt ANOVAat that stage

1118 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 110 5

In these studies we identified on-farm manage-ment practices that are essential for successful con-trol of susceptible P xylostella and populations thathave developed resistance to B thuringiensissubsp kurstaki These results along with those re-ported by Shelton et al (1993) and Tabashnik etal (1990) strongly suggest that a resistance mon-itoring program is needed for B thuringiensis sub-species and that such programs should include ju-dicious use of this material to delay the onset ofresistance Because only operational factors can bemanipulated at the level of the farm to delay theonset of resistance (Georghiou and Taylor 1977)we encourage farmers and researchers to use moreefficient pesticide application methods and con-stantly to review the application rates of B thurin-giensis formulations Our results indicate that theuse of an improved application technology such asthe electrostatic application technique is equiva-lent to a low dose resistance management strat-egy (Tabashnik et al 1991 McGaughey and Wha-Ion 1992) The high efficacy obtained with theelectrostatic application device at low rates will re-duce the number of treatments per season thusincreasing the time of onset of resistance

Appropriate choice of B thuringiensis subspe-cies is of great concern in the field Despite thesimilarity of slopes Javelin seems to be a betterchoice than XenTari to manage susceptible P xylo-stella Although cross-resistance to Javelin andXenTari was not apparent a rotational scheme formanagement of resistant P xylostella and B thu-ringiensis subsp kurstaki does not seem appropri-ate for 2 reasons First this approach assumes thatresistance will decline over time Tabashnik (1994)noted that four to 6 generations are required toachieve a 10-fold reduction in resistance to B thu-ringiensis toxins in P xylostella We do not knowhowever how many generations without exposureare necessary for complete restoration of suscep-tibility In at least one population resistance hasbeen shown to be relatively stable over severalgenerations in the absence of selection pressure(AMS unpublished) Second XenTari containsCrylA toxins which are also present in Javelin andother commercial formulations containing toxins ofB thuringiensis subsp kurstaki (for example Di-pel Dipel 2X Biobit Thuricide) Such common-ality of toxin composition suggests that the use ofXenTari to control resistant populations will tendto select for resistance to the above toxins

Integrated management of P xylostella will de-pend on the conservation of its susceptibility to Bthuringiensis More effort should be invested indeveloping application techniques that result inhigh control efficacy at low application rates andat low cost Additional research on the middotefficacy ofindividual toxins is needed to design possible Bthuringiensis resistance management by use of tox-in rotations

Acknowledgments

We are grateful to R T Roush and M Hoffmann (De-partment of Entomology Cornell University) for theirvaluable advice and comments The statistical advice ofJohn Barnard (Computer Technology New York StatlAgricultural Experiment Station Gemva) is highly ap-preciated J Tang and laboratory personnel from the De-partment of Entomology (New York State AgriculturalExperiment Station Geneva) and field personnel of theDepartment of Agricultural and Biological Enginelring(Cornell University) provided valuable contributions tothis research Financial support for these investigationswas provided by different institutions including the Ful-bright Scholarship Program the Graduate School of Cor-nell University and Encyclopaedia Britannica Financialsupport for personnel and materials was supplied by sev-eral grants to AMS including the BT Managementlorking group and the New York State Cabbage Grow-ers Association

References Cited

Abbott W S 1925 A method of computing the ef-fectiveness of an insecticide J Econ Entomol 18265-267

Adams D B 1992 On-farm components of dia-mondback moth management in Georgia USA pp499-501 In N S Talekar [ed] Management of dia-mondback moth and other cmcifer pests AVRDCShanhua Taiwan

Andaloro J T K B Rose A M Sh(lton C WHoy and R F Becker 1983 Cabbage growthstages NY Food Life Sci Bull 101

Andrews K L R J Sanchez and R Cav 1992Management of diamondback moth in Central Amer-ica pp 487-497 III N S Talekar [Id] Managementof diamondback moth and other crucifer P(middotstsAVRDC Shanhua Taiwan

Beck N G and P J Cameron 1992 Developinga reduced-spray program for brassicas in New Zea-land pp 341-350 III N S Talekar [cd] Manage-ment of diamondback moth and other crucifer P(stsAVRDC Shanhua Taiwan

Georghiou G P and C E Taylor 1977 Opera-tional influences in the evolution of insecticide resis-tance J Econ Entomol 70 653-658

Harcourt D G 1957 Biology of the diamondbackmoth Plutella maculipennis (Curt) (Lepidoptera Plu-tellidae) in eastern Ontario II Life-history behaviourand host relationships Can Entomol 89 554-564

Kirsch K and H Schmutterer 1988 Low efficacyof a Bacillus thurillgiensis (Berl) formnlation in con-trolling the diamondback moth Plutella xylostella(L) in the Philippines J Appl Entomol 105 249-255

Law S E 1980 Electrostatic application of low-vol-ume microbial insecticide spray on broccoli plants JAm Soc Hortic Sci 105(6) 774-777

1982 Spatial distribution of electrostatically depositedsprays on living plants J Econ Entomol 75 542-544

1983 Electrostatic pesticide spraying concepts andpractice IEEE (Inst Electr Electron Eng) TransInd Appl Vol IA-19 No2

McGaughey W H and M E Wltalon 1992 Man-aging resistance to Bacillus thurillgiensis toxins Sci-ence (Washington DC) 258 1451-1455

1116 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 no 5

Table 2 Sununary of ANOVA for the arcsine-trans-formed percentages of P xylostella mortality across 3cabbage stages

Factor df F P value

Main effect~Application technology (A) 2 1414 0000middotmiddotB tlmringiensis ~ub~pecies (B) 1 10]65 0000middotmiddotLeaf position in the plant (C) 2 350 0031middotP xylostella population (D) 1 163606 0000middotmiddotCmp ~tage (E) 2 1006 0000middotmiddot

InteractionsAxB 2 1166 0000middotmiddotAXC 4 416 0003middotmiddotAXE 4 454 0001middotmiddotBxD 1 26276 0000middotmiddotAXBXE 4 585 0000middotmiddotBXDXE 2 1886 0000middotmiddot

Error 432

Signillcant effect 001 lt P lt 005 bullbull highly ~ignillcant effeet P lt 001

ogy on P xylostella mortality was highly significantand interacted significantly with B thuringiensissubspecies and leaf position within the plant (Table2) ANOVA done at each crop stage revealed thatthe interaction between application technologyand B thuringiensis subspecies was highly signifi-cant at early head formation and head-filling stages(P lt 00001 and P lt 0001 respectively Table 3)Variations in mortality associated with applicationtechniques were not significant (P gt 005) at pre-cupping stage when fewer leaves were on theplants and each leaf was more exposed As theplants developed toward head formation and head-filling stages however leaf coverage became crit-ical and the performance of the electrostatic spray-er was significantly better than that of the other 2application techniques (Table 3) The significantinteraction between application technology andcrop stage (Table 2) confirmed the significant vari-ations caused by application technology across the3 crop stages No significant differences were ob-served in the overall performance of the knapsackand drop nozzle sprayers (Table 3)

Table 4 Percentage of mortality plusmn SD of U~CI)tibleand resistant P xylostella at early head formation as afunction of application technology and B thuringiemusubspecies

B thuringmiddot P xylostella populationSprayer iensis

subspecies Geneva 88 100 Loxa b Fr

Electrostatic Javelin 968 ~ 004a 97 2ThXenTari 969 ~ 003a 914 35a

Knap~ack Javelin 965 009a 64 20bXenTari 878 37a 548 KOd

Drop nozzle Javelin 969 004a 174 44cXenTari 865 80a 529 ]84d

Numbers within columns or rows followed by the same letterare not ~ignillcantly different by the Tukey tpound~t (SYSTAT 1992)w = 331 r v = 12 ]44 n = 15 P = 005

Application technology affected tlle efficacy ofJavelin and XenTari on the control of susceptibleand resistant P xylostella At early head formationfor instance mortality of Loxa b (resistant to Bthuringiensis subsp kurstaki) was ~90 onlywhen XenTari was applied with the electrostaticsprayer (Table 4 Fig 1) Mortality of Geneva 88(susceptible to B thuringiensis subsp kurstaki)was ~90 regardless of the application techniqueand B thuringiensis subspecies applied (Tables 4and 5) B thuringiensis subspecies and P xylostellapopulations interacted significantly throughout theexperiments (Tables 2 and 5) Although mortalityof Loxa b from Javelin and XenTari differed sig-nificantly at precupping they had little practicalimportance because tllOse mortality levels wereconsistently low Mortality of Geneva 88 larVae atearly head formation and head-filling stages wassignificantly higher when exposed to Javelin incomparison with XenTari (Table 5)

The relatively better performance of Javelinagainst a population that is assumed susceptible toboth B thuringiensis subspecies may be a result ofdifferences in application rates and concentrationof active ingredient (see above) Javelin was ap-plied at 112 kglha whereas XenTari was appliedat 056 kgha Compared with Javelin mortality ofLoxa b from XenTari was significantly higher (Ta-

Table 3 Average percentage of P xylostella larvae mortality plusmn SD expressed us a fWlction of the interaltiollbetween application technology and B thuringiensis subspecies

B thuringiensis subspecie~Crop stage

Precupping(NS P gt 005)

Early head formation(I = 169 df = 2 144 P lt 00001)

Head filling(I = 69 df = 2 144 P = 0001)

Spmyer

Eleetro~taticKllap~ackDrop nozzleElectrostaticKnap~ackDrop nozzleElectro~taticKllap~ackDrop nozzle

Javelin

64 21a59 22a63 21a57 28a53 30a63 24a65 22a62 25a63 22a

XellTari

64 20a70t 7a67t 15a95t 2b73t 9c71 16ac92 5b75 lac73 9ac

Numbers within columns or row~ of any crop stage followed by the same letter are not signillcantly different by Tukey te~t (SYSTAT1992) (J) = 288 r v = 6 144 n = 30 P = 005) Crop stages are followed by the eorre~pollding F df and P value obtained fromthe ANOVA at that stage

October 1995 PEREZ ET AL IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1117

Fig 1 Mortality of Plutella xylostella (Loxa b) as afunction of th~ interaction between application technol-oro and B tllUringiellsis subspecies Mortality was eval-uat~d on 3 sections of the plants (upper middle andlower) ESS electrostatic spraying systems KN knapsacksprayer DN drop nozzle sprayer JAV Javelin XENXenTari

ble 5) The difference in perrormance of the 2 Bthllringicnsis subspecies on the Loxa b populationwas the result of development of resistance to BthllringiensL~ subsp kurstaki

A significant interaction of application technol-ogy B thuringiensis subspecies and leaf positionwithin the cabbage plant at early head formationis illustrated in Fig 1 Loxa b mortality showedlittle variation among leaves when exposed to res-idues of XenTari applied witll the electrostaticsprayer However when applied with the knapsackor drop nozzle sprayers mortality caused byXenTari was usually low except on tlle lowerleaves mortality of Loxa b varied widely across leafpositions (Fig 1)

Genetic Effects The differences in responsesof Loxa b and Geneva 88 to the 2 B thuringiensissubspecies accounted for most of the variations inmortality (Table 2) The overall average mortality SEM of Geneva 88 and Loxa b was 94 22and 36 13 respectively Resistance of Loxa bto B thllringiensis subsp kllrstaki had been doc-umented previously and our results are consistentwith those reported by Shelton et al (1993) Thegenetic differences are confirmed by the significant

100

80

5 60ltCI-ao 40~~

20

o

bull Upper~ Middle~ Lower

ESSlJAY ESSIXEN HfWAY HtmN OWJAY OWXEN

Sprayer by B thurlnglensls subspecies

Table 6 Dose-mortality regression data of suscepti-ble and resistant P xyloslella in sinmlated field applica-tions of B lhuringiensis

Population n Slope SEM LD90 (kWha)(95 CL)

Geneva 88 (susceptible)Javelin 285 036 060 026

(01-05)XenTari 280 111 027 05

(03--08)Loxab (resistant)

Javelin 290 230 041 63(42-126)

XenTari 259 390 085 16(09-23)

95 CL (POLO [Russellet aI 1977]) significantby t ratiotest Slope SEM gt 196

differences in mortality of the 2 populations fromJavelin whereas the application rate (112 kglha)remained constant (Table 5) We expected to ob-serve similar responses of the 2 populations toXenTari because resistance to B thuringiensissubsp aizawai had not been documented in Flor-ida where Loxa b was collected Except for appli-cations with the electrostatic sprayer the averageefficacy of XenTari against Loxa b was 560 Fail-ure to control Loxa b witll conventional applicationdevices raises concern about tlle appropriate rateof XenTari and we recommend a change in itscurrent field application rate

Field Tests There were significant dose re-sponses as determined by significant slopes for allbut Javelin when used against tlle Geneva 88 pop-ulation (Table 6) In that case the shallow slope ofthe response resulted from mortality gt90 at allrates tested The calculated LDgo of tlle Geneva88 population to Javelin (026 kglha) was lowerthan the recommended rate indicating that suscep-tible populations may be controlled by Javelin atrates lt112 kglha The LDgo of ~enTari againstGeneva 88 (05 kgha) was close to the recom-mended rate of application in the field (056 kgha) The LDgos of Loxa b from Javelin and XenTariwere 24- and 32-fold higher respectively than thecorresponding LDgos of Geneva 88 from the samematerials

Table 5 Av(ra~e percentage of mortality t SD of susceptible and resistant larvae of P yloslella exposed to 2B tlutringiensis ~uh~pccics

P xylostella populationCrop stage

Prpcupping(F ~ 175df = I 144 P lt 00001)

Early Iwad formation(F ~ 1846df = 1 144 P lt 00001)

lltad filling(F = U97 df = I 144 P lt 00001)

B thuringiensissubspecies

JavelinXenTariJavelinXenTariJavelinXenTari

Geneva 88

943 Ola890 30a967 OOla911 46b970 Ola924 30b

Loxab

218 34b388 92c1O8 34c684 139d188 34c661 92d

Numlwrs within columns or rows of any crop stage followedby the same letter are not significantlydifferent by the Tukey test(SYSTAT1992) w = 26 r v = 12 144 II = 45 P = 005) Crop stages are followedby the corresponding F df and P valuesobtailltd from tIlt ANOVAat that stage

1118 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 110 5

In these studies we identified on-farm manage-ment practices that are essential for successful con-trol of susceptible P xylostella and populations thathave developed resistance to B thuringiensissubsp kurstaki These results along with those re-ported by Shelton et al (1993) and Tabashnik etal (1990) strongly suggest that a resistance mon-itoring program is needed for B thuringiensis sub-species and that such programs should include ju-dicious use of this material to delay the onset ofresistance Because only operational factors can bemanipulated at the level of the farm to delay theonset of resistance (Georghiou and Taylor 1977)we encourage farmers and researchers to use moreefficient pesticide application methods and con-stantly to review the application rates of B thurin-giensis formulations Our results indicate that theuse of an improved application technology such asthe electrostatic application technique is equiva-lent to a low dose resistance management strat-egy (Tabashnik et al 1991 McGaughey and Wha-Ion 1992) The high efficacy obtained with theelectrostatic application device at low rates will re-duce the number of treatments per season thusincreasing the time of onset of resistance

Appropriate choice of B thuringiensis subspe-cies is of great concern in the field Despite thesimilarity of slopes Javelin seems to be a betterchoice than XenTari to manage susceptible P xylo-stella Although cross-resistance to Javelin andXenTari was not apparent a rotational scheme formanagement of resistant P xylostella and B thu-ringiensis subsp kurstaki does not seem appropri-ate for 2 reasons First this approach assumes thatresistance will decline over time Tabashnik (1994)noted that four to 6 generations are required toachieve a 10-fold reduction in resistance to B thu-ringiensis toxins in P xylostella We do not knowhowever how many generations without exposureare necessary for complete restoration of suscep-tibility In at least one population resistance hasbeen shown to be relatively stable over severalgenerations in the absence of selection pressure(AMS unpublished) Second XenTari containsCrylA toxins which are also present in Javelin andother commercial formulations containing toxins ofB thuringiensis subsp kurstaki (for example Di-pel Dipel 2X Biobit Thuricide) Such common-ality of toxin composition suggests that the use ofXenTari to control resistant populations will tendto select for resistance to the above toxins

Integrated management of P xylostella will de-pend on the conservation of its susceptibility to Bthuringiensis More effort should be invested indeveloping application techniques that result inhigh control efficacy at low application rates andat low cost Additional research on the middotefficacy ofindividual toxins is needed to design possible Bthuringiensis resistance management by use of tox-in rotations

Acknowledgments

We are grateful to R T Roush and M Hoffmann (De-partment of Entomology Cornell University) for theirvaluable advice and comments The statistical advice ofJohn Barnard (Computer Technology New York StatlAgricultural Experiment Station Gemva) is highly ap-preciated J Tang and laboratory personnel from the De-partment of Entomology (New York State AgriculturalExperiment Station Geneva) and field personnel of theDepartment of Agricultural and Biological Enginelring(Cornell University) provided valuable contributions tothis research Financial support for these investigationswas provided by different institutions including the Ful-bright Scholarship Program the Graduate School of Cor-nell University and Encyclopaedia Britannica Financialsupport for personnel and materials was supplied by sev-eral grants to AMS including the BT Managementlorking group and the New York State Cabbage Grow-ers Association

References Cited

Abbott W S 1925 A method of computing the ef-fectiveness of an insecticide J Econ Entomol 18265-267

Adams D B 1992 On-farm components of dia-mondback moth management in Georgia USA pp499-501 In N S Talekar [ed] Management of dia-mondback moth and other cmcifer pests AVRDCShanhua Taiwan

Andaloro J T K B Rose A M Sh(lton C WHoy and R F Becker 1983 Cabbage growthstages NY Food Life Sci Bull 101

Andrews K L R J Sanchez and R Cav 1992Management of diamondback moth in Central Amer-ica pp 487-497 III N S Talekar [Id] Managementof diamondback moth and other crucifer P(middotstsAVRDC Shanhua Taiwan

Beck N G and P J Cameron 1992 Developinga reduced-spray program for brassicas in New Zea-land pp 341-350 III N S Talekar [cd] Manage-ment of diamondback moth and other crucifer P(stsAVRDC Shanhua Taiwan

Georghiou G P and C E Taylor 1977 Opera-tional influences in the evolution of insecticide resis-tance J Econ Entomol 70 653-658

Harcourt D G 1957 Biology of the diamondbackmoth Plutella maculipennis (Curt) (Lepidoptera Plu-tellidae) in eastern Ontario II Life-history behaviourand host relationships Can Entomol 89 554-564

Kirsch K and H Schmutterer 1988 Low efficacyof a Bacillus thurillgiensis (Berl) formnlation in con-trolling the diamondback moth Plutella xylostella(L) in the Philippines J Appl Entomol 105 249-255

Law S E 1980 Electrostatic application of low-vol-ume microbial insecticide spray on broccoli plants JAm Soc Hortic Sci 105(6) 774-777

1982 Spatial distribution of electrostatically depositedsprays on living plants J Econ Entomol 75 542-544

1983 Electrostatic pesticide spraying concepts andpractice IEEE (Inst Electr Electron Eng) TransInd Appl Vol IA-19 No2

McGaughey W H and M E Wltalon 1992 Man-aging resistance to Bacillus thurillgiensis toxins Sci-ence (Washington DC) 258 1451-1455

October 1995 PEREZ ET AL IMPROVED TECHNIQUE FOR CONTROL OF P xylostella 1117

Fig 1 Mortality of Plutella xylostella (Loxa b) as afunction of th~ interaction between application technol-oro and B tllUringiellsis subspecies Mortality was eval-uat~d on 3 sections of the plants (upper middle andlower) ESS electrostatic spraying systems KN knapsacksprayer DN drop nozzle sprayer JAV Javelin XENXenTari

ble 5) The difference in perrormance of the 2 Bthllringicnsis subspecies on the Loxa b populationwas the result of development of resistance to BthllringiensL~ subsp kurstaki

A significant interaction of application technol-ogy B thuringiensis subspecies and leaf positionwithin the cabbage plant at early head formationis illustrated in Fig 1 Loxa b mortality showedlittle variation among leaves when exposed to res-idues of XenTari applied witll the electrostaticsprayer However when applied with the knapsackor drop nozzle sprayers mortality caused byXenTari was usually low except on tlle lowerleaves mortality of Loxa b varied widely across leafpositions (Fig 1)

Genetic Effects The differences in responsesof Loxa b and Geneva 88 to the 2 B thuringiensissubspecies accounted for most of the variations inmortality (Table 2) The overall average mortality SEM of Geneva 88 and Loxa b was 94 22and 36 13 respectively Resistance of Loxa bto B thllringiensis subsp kllrstaki had been doc-umented previously and our results are consistentwith those reported by Shelton et al (1993) Thegenetic differences are confirmed by the significant

100

80

5 60ltCI-ao 40~~

20

o

bull Upper~ Middle~ Lower

ESSlJAY ESSIXEN HfWAY HtmN OWJAY OWXEN

Sprayer by B thurlnglensls subspecies

Table 6 Dose-mortality regression data of suscepti-ble and resistant P xyloslella in sinmlated field applica-tions of B lhuringiensis

Population n Slope SEM LD90 (kWha)(95 CL)

Geneva 88 (susceptible)Javelin 285 036 060 026

(01-05)XenTari 280 111 027 05

(03--08)Loxab (resistant)

Javelin 290 230 041 63(42-126)

XenTari 259 390 085 16(09-23)

95 CL (POLO [Russellet aI 1977]) significantby t ratiotest Slope SEM gt 196

differences in mortality of the 2 populations fromJavelin whereas the application rate (112 kglha)remained constant (Table 5) We expected to ob-serve similar responses of the 2 populations toXenTari because resistance to B thuringiensissubsp aizawai had not been documented in Flor-ida where Loxa b was collected Except for appli-cations with the electrostatic sprayer the averageefficacy of XenTari against Loxa b was 560 Fail-ure to control Loxa b witll conventional applicationdevices raises concern about tlle appropriate rateof XenTari and we recommend a change in itscurrent field application rate

Field Tests There were significant dose re-sponses as determined by significant slopes for allbut Javelin when used against tlle Geneva 88 pop-ulation (Table 6) In that case the shallow slope ofthe response resulted from mortality gt90 at allrates tested The calculated LDgo of tlle Geneva88 population to Javelin (026 kglha) was lowerthan the recommended rate indicating that suscep-tible populations may be controlled by Javelin atrates lt112 kglha The LDgo of ~enTari againstGeneva 88 (05 kgha) was close to the recom-mended rate of application in the field (056 kgha) The LDgos of Loxa b from Javelin and XenTariwere 24- and 32-fold higher respectively than thecorresponding LDgos of Geneva 88 from the samematerials

Table 5 Av(ra~e percentage of mortality t SD of susceptible and resistant larvae of P yloslella exposed to 2B tlutringiensis ~uh~pccics

P xylostella populationCrop stage

Prpcupping(F ~ 175df = I 144 P lt 00001)

Early Iwad formation(F ~ 1846df = 1 144 P lt 00001)

lltad filling(F = U97 df = I 144 P lt 00001)

B thuringiensissubspecies

JavelinXenTariJavelinXenTariJavelinXenTari

Geneva 88

943 Ola890 30a967 OOla911 46b970 Ola924 30b

Loxab

218 34b388 92c1O8 34c684 139d188 34c661 92d

Numlwrs within columns or rows of any crop stage followedby the same letter are not significantlydifferent by the Tukey test(SYSTAT1992) w = 26 r v = 12 144 II = 45 P = 005) Crop stages are followedby the corresponding F df and P valuesobtailltd from tIlt ANOVAat that stage

1118 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 110 5

In these studies we identified on-farm manage-ment practices that are essential for successful con-trol of susceptible P xylostella and populations thathave developed resistance to B thuringiensissubsp kurstaki These results along with those re-ported by Shelton et al (1993) and Tabashnik etal (1990) strongly suggest that a resistance mon-itoring program is needed for B thuringiensis sub-species and that such programs should include ju-dicious use of this material to delay the onset ofresistance Because only operational factors can bemanipulated at the level of the farm to delay theonset of resistance (Georghiou and Taylor 1977)we encourage farmers and researchers to use moreefficient pesticide application methods and con-stantly to review the application rates of B thurin-giensis formulations Our results indicate that theuse of an improved application technology such asthe electrostatic application technique is equiva-lent to a low dose resistance management strat-egy (Tabashnik et al 1991 McGaughey and Wha-Ion 1992) The high efficacy obtained with theelectrostatic application device at low rates will re-duce the number of treatments per season thusincreasing the time of onset of resistance

Appropriate choice of B thuringiensis subspe-cies is of great concern in the field Despite thesimilarity of slopes Javelin seems to be a betterchoice than XenTari to manage susceptible P xylo-stella Although cross-resistance to Javelin andXenTari was not apparent a rotational scheme formanagement of resistant P xylostella and B thu-ringiensis subsp kurstaki does not seem appropri-ate for 2 reasons First this approach assumes thatresistance will decline over time Tabashnik (1994)noted that four to 6 generations are required toachieve a 10-fold reduction in resistance to B thu-ringiensis toxins in P xylostella We do not knowhowever how many generations without exposureare necessary for complete restoration of suscep-tibility In at least one population resistance hasbeen shown to be relatively stable over severalgenerations in the absence of selection pressure(AMS unpublished) Second XenTari containsCrylA toxins which are also present in Javelin andother commercial formulations containing toxins ofB thuringiensis subsp kurstaki (for example Di-pel Dipel 2X Biobit Thuricide) Such common-ality of toxin composition suggests that the use ofXenTari to control resistant populations will tendto select for resistance to the above toxins

Integrated management of P xylostella will de-pend on the conservation of its susceptibility to Bthuringiensis More effort should be invested indeveloping application techniques that result inhigh control efficacy at low application rates andat low cost Additional research on the middotefficacy ofindividual toxins is needed to design possible Bthuringiensis resistance management by use of tox-in rotations

Acknowledgments

We are grateful to R T Roush and M Hoffmann (De-partment of Entomology Cornell University) for theirvaluable advice and comments The statistical advice ofJohn Barnard (Computer Technology New York StatlAgricultural Experiment Station Gemva) is highly ap-preciated J Tang and laboratory personnel from the De-partment of Entomology (New York State AgriculturalExperiment Station Geneva) and field personnel of theDepartment of Agricultural and Biological Enginelring(Cornell University) provided valuable contributions tothis research Financial support for these investigationswas provided by different institutions including the Ful-bright Scholarship Program the Graduate School of Cor-nell University and Encyclopaedia Britannica Financialsupport for personnel and materials was supplied by sev-eral grants to AMS including the BT Managementlorking group and the New York State Cabbage Grow-ers Association

References Cited

Abbott W S 1925 A method of computing the ef-fectiveness of an insecticide J Econ Entomol 18265-267

Adams D B 1992 On-farm components of dia-mondback moth management in Georgia USA pp499-501 In N S Talekar [ed] Management of dia-mondback moth and other cmcifer pests AVRDCShanhua Taiwan

Andaloro J T K B Rose A M Sh(lton C WHoy and R F Becker 1983 Cabbage growthstages NY Food Life Sci Bull 101

Andrews K L R J Sanchez and R Cav 1992Management of diamondback moth in Central Amer-ica pp 487-497 III N S Talekar [Id] Managementof diamondback moth and other crucifer P(middotstsAVRDC Shanhua Taiwan

Beck N G and P J Cameron 1992 Developinga reduced-spray program for brassicas in New Zea-land pp 341-350 III N S Talekar [cd] Manage-ment of diamondback moth and other crucifer P(stsAVRDC Shanhua Taiwan

Georghiou G P and C E Taylor 1977 Opera-tional influences in the evolution of insecticide resis-tance J Econ Entomol 70 653-658

Harcourt D G 1957 Biology of the diamondbackmoth Plutella maculipennis (Curt) (Lepidoptera Plu-tellidae) in eastern Ontario II Life-history behaviourand host relationships Can Entomol 89 554-564

Kirsch K and H Schmutterer 1988 Low efficacyof a Bacillus thurillgiensis (Berl) formnlation in con-trolling the diamondback moth Plutella xylostella(L) in the Philippines J Appl Entomol 105 249-255

Law S E 1980 Electrostatic application of low-vol-ume microbial insecticide spray on broccoli plants JAm Soc Hortic Sci 105(6) 774-777

1982 Spatial distribution of electrostatically depositedsprays on living plants J Econ Entomol 75 542-544

1983 Electrostatic pesticide spraying concepts andpractice IEEE (Inst Electr Electron Eng) TransInd Appl Vol IA-19 No2

McGaughey W H and M E Wltalon 1992 Man-aging resistance to Bacillus thurillgiensis toxins Sci-ence (Washington DC) 258 1451-1455

1118 JOURNAL OF ECONOMIC ENTOMOLOGY Vol 88 110 5

In these studies we identified on-farm manage-ment practices that are essential for successful con-trol of susceptible P xylostella and populations thathave developed resistance to B thuringiensissubsp kurstaki These results along with those re-ported by Shelton et al (1993) and Tabashnik etal (1990) strongly suggest that a resistance mon-itoring program is needed for B thuringiensis sub-species and that such programs should include ju-dicious use of this material to delay the onset ofresistance Because only operational factors can bemanipulated at the level of the farm to delay theonset of resistance (Georghiou and Taylor 1977)we encourage farmers and researchers to use moreefficient pesticide application methods and con-stantly to review the application rates of B thurin-giensis formulations Our results indicate that theuse of an improved application technology such asthe electrostatic application technique is equiva-lent to a low dose resistance management strat-egy (Tabashnik et al 1991 McGaughey and Wha-Ion 1992) The high efficacy obtained with theelectrostatic application device at low rates will re-duce the number of treatments per season thusincreasing the time of onset of resistance

Appropriate choice of B thuringiensis subspe-cies is of great concern in the field Despite thesimilarity of slopes Javelin seems to be a betterchoice than XenTari to manage susceptible P xylo-stella Although cross-resistance to Javelin andXenTari was not apparent a rotational scheme formanagement of resistant P xylostella and B thu-ringiensis subsp kurstaki does not seem appropri-ate for 2 reasons First this approach assumes thatresistance will decline over time Tabashnik (1994)noted that four to 6 generations are required toachieve a 10-fold reduction in resistance to B thu-ringiensis toxins in P xylostella We do not knowhowever how many generations without exposureare necessary for complete restoration of suscep-tibility In at least one population resistance hasbeen shown to be relatively stable over severalgenerations in the absence of selection pressure(AMS unpublished) Second XenTari containsCrylA toxins which are also present in Javelin andother commercial formulations containing toxins ofB thuringiensis subsp kurstaki (for example Di-pel Dipel 2X Biobit Thuricide) Such common-ality of toxin composition suggests that the use ofXenTari to control resistant populations will tendto select for resistance to the above toxins

Integrated management of P xylostella will de-pend on the conservation of its susceptibility to Bthuringiensis More effort should be invested indeveloping application techniques that result inhigh control efficacy at low application rates andat low cost Additional research on the middotefficacy ofindividual toxins is needed to design possible Bthuringiensis resistance management by use of tox-in rotations

Acknowledgments

We are grateful to R T Roush and M Hoffmann (De-partment of Entomology Cornell University) for theirvaluable advice and comments The statistical advice ofJohn Barnard (Computer Technology New York StatlAgricultural Experiment Station Gemva) is highly ap-preciated J Tang and laboratory personnel from the De-partment of Entomology (New York State AgriculturalExperiment Station Geneva) and field personnel of theDepartment of Agricultural and Biological Enginelring(Cornell University) provided valuable contributions tothis research Financial support for these investigationswas provided by different institutions including the Ful-bright Scholarship Program the Graduate School of Cor-nell University and Encyclopaedia Britannica Financialsupport for personnel and materials was supplied by sev-eral grants to AMS including the BT Managementlorking group and the New York State Cabbage Grow-ers Association

References Cited

Abbott W S 1925 A method of computing the ef-fectiveness of an insecticide J Econ Entomol 18265-267

Adams D B 1992 On-farm components of dia-mondback moth management in Georgia USA pp499-501 In N S Talekar [ed] Management of dia-mondback moth and other cmcifer pests AVRDCShanhua Taiwan

Andaloro J T K B Rose A M Sh(lton C WHoy and R F Becker 1983 Cabbage growthstages NY Food Life Sci Bull 101

Andrews K L R J Sanchez and R Cav 1992Management of diamondback moth in Central Amer-ica pp 487-497 III N S Talekar [Id] Managementof diamondback moth and other crucifer P(middotstsAVRDC Shanhua Taiwan

Beck N G and P J Cameron 1992 Developinga reduced-spray program for brassicas in New Zea-land pp 341-350 III N S Talekar [cd] Manage-ment of diamondback moth and other crucifer P(stsAVRDC Shanhua Taiwan

Georghiou G P and C E Taylor 1977 Opera-tional influences in the evolution of insecticide resis-tance J Econ Entomol 70 653-658

Harcourt D G 1957 Biology of the diamondbackmoth Plutella maculipennis (Curt) (Lepidoptera Plu-tellidae) in eastern Ontario II Life-history behaviourand host relationships Can Entomol 89 554-564

Kirsch K and H Schmutterer 1988 Low efficacyof a Bacillus thurillgiensis (Berl) formnlation in con-trolling the diamondback moth Plutella xylostella(L) in the Philippines J Appl Entomol 105 249-255

Law S E 1980 Electrostatic application of low-vol-ume microbial insecticide spray on broccoli plants JAm Soc Hortic Sci 105(6) 774-777

1982 Spatial distribution of electrostatically depositedsprays on living plants J Econ Entomol 75 542-544

1983 Electrostatic pesticide spraying concepts andpractice IEEE (Inst Electr Electron Eng) TransInd Appl Vol IA-19 No2

McGaughey W H and M E Wltalon 1992 Man-aging resistance to Bacillus thurillgiensis toxins Sci-ence (Washington DC) 258 1451-1455

October 1995 PEREZ ET AL IMPHOVED TECHNIQUE FOR CONTROL OF P xylostella 1119

Miyntn T V Noppum and T Snito 1992 Inheri-tan(e of resistance to phenthoate and fenvalerate indiamondba(k moth and management of insecticideresistance pp 477-484 In N S Talekar [ed] Man-ag(l11(ntof diamondback moth and other cruciferIJPstsAHDC Shanhua Taiwan

Russl R M J L Robertson and N E Savin1977 POLO a new computer program for probitanalysis Bull Entomol Soc Am 23 209-213

Sh-hOl A M R J Cooley M K Kroening W TWilsey nnd S D Eigenbrode 1991 Compara-tingt analysis of two rearing procedures for diamond-back moth (Ltpidoptera PlutelIidae) J EntomoI Sci26 17-26

Slwhon A M J L Robertson J D Tang C PerezS n Eig-nbrodO H K Preisler W T Wilseyttl R J Cbullbullbullbullley 1993 Resistance of diamondbackl110th (Lepidoptera Plutellidae) to Bacillus thurin-gilllsb subslwcies in the field J Econ Entomol 86697-705

Smith n B n L Hostetter and C M Ignoffo1977a Ground spray equipment for applying Bacil-lus thuriflgicflsis suspension on soybeans J Econ En-tOl11ol70 633-637

1977b Laboratory performance specifications for ahactlrial (BllCilus tlwrillgiensis) and a viral (Baculo-tims hcliothis) insecticide J Econ Entomol 70437-441

SmmiddotItmiddotbullbullbullc G W nnd W G Cochcnn 1989 Statis-timl Ilwthods 8th ed Iowa State University PressAmes

SWl C N 1992 Insecticide resistance in diamond-back moth pp 419-426 In N S Talekar [ed] Man-agement of diamondback moth and other cruciferpests AVRDC Shanhna Taiwan

SWl C N T K Wu J S Chen and W T LOe1986 Insecticide resistance in diamondback mothpp 359-371 In N S Talekar [ed] Diamondbackmoth management AVHDC Shanhua Taiwan

Syed A R 1992 Insecticide resistance in diamond-back moth in Malaysia pp 437-442 In N S Talekar[ed] Management of diamondback moth and othercrucifer pests AVRDC Shanhua Taiwan

SYSTAT 1992 SYSTATfor windows statistics version5 ed SYSTATEvanston II

Tabashnik B E 1994 Evolution of resistance to Ba-cillus tlwringiensis Annu Rev Entomol 39 47-79

Tabashnik B E N L Cushing N Finson andM W Johnson 1990 Field development of resis-tance to Bacillus thuringiensis in diamondback moth(Lepidoptera Plutellidae) J Econ Entomol 831671-1676

Tabashnik B E N Finson nnd M W Johnson1991 Managing resistance to Bacillus thuringiensL~lessons from the diamondback moth (LepidopteraPlutelIidae) J Econ EntomoI 84 49-55

Taleknc N S amI A M Shehon 1993 Biologyecology and management of the diamondback mothAnnu Rev EntomoI 38 275-301

Received for publication 22 Jull 1993 accepted 8March 1995