Journal of the American Mosquito Control Association, 26(2):183-197, 2010Copyright © 2010 by The American Mosquito Control Association, Inc.

188756

EVALUATION OF ULV AND THERMAL FOG MOSQUITO CONTROL.A.PPLICATIONS IN TEMPERATE AND DESERT ENVIRONMENTS

SETH C. BRITCH,1 KENNETH J. LINTHICUM,! WAYNE W. WYNN,l TODD W. WALKER,2MUHAMMAD FAROOQ,2 VINCENT L. SMITH,2 CATHY A. ROBINSON,2 BRANKA B. LOTHROP,3

MELISSA SNELLING,3 ARTURO GUTIERREZ,3 HUGH D. LOTHROP,4 JERRY D. KERCE,5JAMES J. BECNEL,l ULRICH R. BERNIERl AND JULIA W. PRIDGEONl

ABSTRACT. Ultra-low-volume (ULV) and thermal fog aerosol dispersals of pesticides have been usedagainst mosquitoes and other insects for half a century. Although each spray technology has advantages anddisadvantages, only 7 studies have been identified that directly compare their performance in the field. USmilitary personnel currently operating in hot-arid environments are impacted by perpetual nuisance anddisease vector insect problems, despite adulticide operations using modern pesticide-delivery equipment suchas ULV. None of the identified comparative studies has looked at the relative feasibility and efficacy of ULVand thermal fog equipment against mosquitoes in hot-arid environments. In this study we examine theimpact of ULV and thermal fog applications of malathion against caged sentinel mosquitoes in the field in awarm temperate area of Florida, followed by a similar test in a hot-dry desert area of southern California.Patterns of mortality throughout 150 m X 150 m grids of sentinel mosquitoes indicate greater efficacy fromthe thermal fog application in both environments under suboptimal ambient weather conditions. We discussthe implications of these findings for future military preventive medicine activities and encourage furtherinvestigations into the relative merits of the 2 technologies for force health protection.

KEY WORDS Coachella Valley, Camp Blanding, aerosol pesticide delivery, malathion, Deployed War­Fighter Protection Program

INTRODUCTION

For more than 60 years, spray dispersal ofaerosolized insecticides has been a critical weaponagainst nuisance insects and insect vectors ofdisease. Two technologies have dominated thisapproach and continue to be used throughout theworld: thermal fogging, originating from militarysmoke generators in the early 1940s (LaMer et al.1947), and ultra-low-volume (ULV) "cold mist"spraying, originating from modifications to stan­dard agricultural sprayers in the 1950s (Lofgren1970). Numerous pesticides formulated for use inthermal foggers and ULV sprayers have evolved,as have spray technologies such as those cited inMount (1998) and Hoffmann et al. (2007a, 2007b,2008). The US military depends on aerosoldispersal of adulticides for sand fly and mosquitocontrol in hot, dry, and dusty environments in theMiddle East. However, the dwindling list ofEnvironmental Protection Agency-approved pes­ticides available for military use (Armed Forces

1 United States Department of Agriculture, Agricul­tural Research Service, Center for Medical, Agricul­tural, and Veterinary Entomology, 1600 SW 23rdDrive, Gainesville, FL 32608.

2 Navy Entomology Center of Excellence, Naval AirStation, PO Box 43, Jacksonville, FL 32212.

3 Coachella Valley Mosquito and Vector ControlDistrict, 43-420 Trader Place, Indio, CA 92201.

4 Arbovirus Research Unit, Center for VectorborneDiseases, School of Veterinary Medicine, University ofCalifornia, Davis, CA 95616.

5 Camp Blanding Joint Training Site, 5629 StateRoad 16W, Starke, FL 32091.

Pest Management Board 2010) and failure tocontrol nuisance and vector insect problems inrecent desert military operations underscore theneed to reexamine current control strategies(Linthicum et al. 2007, Cope et al. 2008, Dalton2008).

A survey of the literature between 1940 and2009 reveals an asymmetry between assessmentsof thermal fog and ULV technology. Hundreds ofresearch studies on ULV efficacy against a rangeof insects have been carried out, whereas onlydozens exist for thermal fog technology. Only 7studies have been identified comparing ULV andthermal fog efficacy against mosquitoes in out­door applications (Mount et al. 1968, Mount etal. 1970, Taylor and Schoof 1971, Rathburn andBoike 1972, Phanthumachinda et al. 1976, Wiratet al. 1982, Linley and Jordan 1992). In these 7studies, thermal fogging frequently performedcomparably to ULV, yet current practices in USmosquito abatement agencies and US militarydoctrine emphasize the use of ULV. Contributingfactors for ULV preference by mosquito abate­ment agencies is the lower volume of materialdispersed in the environment, quieter operation,and lack of visible "smoke" compared to thermalfogging, all of which are also important factors inmilitary pest management scenarios. However,renewed interest in optimizing aerosol pesticideefficacy in militarily relevant desert environments,especially given the current difficulty of control­ling sand fly and mosquito populations with ULVin these environments, leads to the comparativeexamination of these 2 technologies in this study.

183

184 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, NO.2

In addition, all previous ULV/thermal fogcomparative studies were conducted in warm,moist temperate or tropical areas such asGeorgia, Florida, and 2 locations in Thailand.Given 17 years of improvements in technologyand chemical formulations since the most recentof the studies, a better understanding of factorsthat contribute to the relative efficacy, feasibility,and limitations of the 2 technologies in hot-aridenvironments is needed.

In this study we assessed the relative efficacy ofULV and thermal fog applications to controlCulex and Aedes adults in temperate and desertenvironments. Multiple factors including envi­ronmental conditions have a significant impactupon the efficacy of aerosols used in adultmosquito control operations; for instance, hot­arid climates are likely to be particularly prob­lematic because of impacts such as rapid evapora­tion. No studies have been identified that look atthe performance of thermal fog in hot, dryconditions; however, ground and aerial ULVaerosol pesticide delivery in hot dry conditionshas been carried out with mixed outcomes againstdesert locusts (Holland and Jepson 1996) andmosquitoes (Lothrop et al. 2007a, 2007b, Lo­throp et al. 2008), in part because of populationdynamics or wind conditions at the time ofdelivery. No studies have been identified thatinvestigate differences in success of ULV orthermal fog (or ULV and thermal fog side byside) applications in hot, dry compared to warm,moist conditions, regardless of wind conditions.We carried out an initial trial of the experiment ina warm, moist temperate area in northern Floridato evaluate the feasibility of the experimentaldesign and obtain a baseline for performance ofthe 2 technologies. The second trial was nearlyidentical in design to the first, but carried out in adesert area in southern California as a proxy forenvironmental conditions in current US militaryoperations in most regions of Iraq and manyregions of Afghanistan.

METHODS AND MATERIALS

We measured the efficacy of ULV and thermalfog spray technologies by way of 2 assays,including spatial patterns of mortality in sentinelmosquito cages placed on test grids in the aerosolcloud, and droplet density and depositioon onpaper ribbons placed within the test grids.

Study sites

We conducted the initial baseline field study ina temperate habitat in Camp Blanding(29.864890N, 81.973534W, 210 ft), a militaryreservation in northern Florida within 60 mintravel distance from the Center for Medical,Agricultural, and Veterinary Entomology labora-

A

B

Fig. 1. (A) Typical view of the habitat at CampBlanding, FL: open ground with sparse forb vegetation;clear, breezy, warm, humid. (B) Typical view of thehabitat in the date palm grove in the Coachella Valley~

CA: rows of planted date palms; clear, breezy, hotand dry.

tory, and conducted the follow-up field study in adate palm desert habitat in the Coachella Valley insouthern California (33.607660N, 116.213721W~

-97 ft), within 60 min travel distance from theCoachella Valley Mosquito and Vector ControlDistrict laboratory. The Florida site was open:level, and clear of any vegetation over """'-'0.5 m tall(Fig. lA). The California site was level withregular rows of date palm trees """'-' 3-5 m tall andlow understory vegetation consisting mostly ofgrass """'-'0.25 m tall (Fig. 1B). In both areas weused identical ULV and thermal fog equipmentand a single pesticide formulation.

Spray materials

In both study areas we used Fyfanon ULV(Cheminova, Wayne, NJ) malathion pesticide at

100% in all ULV applications and in a 5.9%solution in diesel in all thermal fog applications.

JCNE 2010 DESERT ULV AND THERMAL FOG COMPARISON 185

Fig. 2. (A) Upright plastic fence post with 2 sentinelcages attached at 0.3 ill and 1.2 m above ground. (B)The ribbon ladder apparatus, shown deployed alongsidesentinel cage poles in the California date palm habitat.

The label recommends an application rate of 2­4 oz/acre (---148-296 ml/ha). We added anultraviolet (UV) fluorescent tracer dye to thepesticide to permit detection of spray on captureapparatus, described below. For the experimentconducted at the Florida site we mixed Uvitex OBfluorescent dye (Ciba Corporation, Newport,DE) with the Fyfanon ULV at 2,000 ppm forthe ULV application and at 1,000 ppm for thethermal fog application. Subsequent observationsat the Florida site indicated that labeled dropletsfrom the ULV trial were not visible, and labeleddroplets from the thermal fog trial were onlyweakly visible with a UV flashlight. Thus, for theexperiment conducted at the California site, thedye concentrations were increased 100% to4,000 ppm for the ULV application and 50% to1,500 ppm for the thermal fog application.

A B

Sprayers

We used the London Fog 18-20 (London Fog,Long Lake, MN) ULV aerosol generator and theCurtis Dyna-Fog Silver Cloud model 2560(Curtis Dyna-Fog, Westfield, IN) thermal [oggerin this study. Both pieces of equipment werevehicle-mounted, and in both the Florida andCalifornia trials the ULV sprayer was used tospray the west plots, and the thermal fogger wasused to spray the east plots.

The 18-20 ULV aerosol generator is poweredby an 18 hp (13.2 kW) gasoline engine and isequipped with an air-shear nozzle. The ULV hasa net weight of 445lb (202 kg). The air source is arotary, positive-displacement blower that canproduce air flow up to 356 ft3 (10.1 m 3

) perminute, and the nozzle provides 3600 rotationboth horizontally and vertically. The liquid flowrate ranges from 0 to 22 oz (0 to 0.65 liter) perminute, and the tank capacity is 15 gallons(56 liter) with a 0.4 gallon (1.58 liter) flush tank.For this study, we set the sprayer nozzle todischarge horizontally, parallel to and away fromthe direction of travel. We set the flow rate at10.7 oz/min (316 ml/min), and we drove thesprayer mounted on a pickup truck at 10 mph(16 krn/h) to produce an application rate of1.77 oz/acre (0.129 liter/ha) of total solution andan application rate of active ingredient of 1.71 oz/acre (0.125 liter/ha). The 18-20 produces a D vO.1,

Dvo.s, DvO.9 , of 6.1, 14.6, and 26.6 ~m, respec­tively, and 72.70/0 of the spray volume is made upof droplets::::; 20 ~m (Hoffmann et al. 2007a).

The Silver Cloud uses gasoline-powered 88 hp(66 kW) twin pulse jet engines to produce athermal fog. The thermal fogger has a net weightof l06lb (48 kg) and a fuel tank capacity of 6 gal(9.5 liter). The fogger can deliver up to 40 gal/h(151.4 liter/h) and produce over 250,000 ft 3

(6,500 m3) of effective fog per minute. For this

study we set the machine to deliver 85.3 oz/min

(2.5 liter/min). The thermal fogger was mountedon an ATV (Florida trial) or a flat bed truck(California trial) and operated at 5 mph (8 km/h)to produce an application rate of 28.1 oz/acre(2.0 liter/ha) of pesticide diluted with diesel to5.9% in solution, and an active ingredientapplication rate of 1.60 oz/acre (0.114 liter/ha).The fogger produces a D vO .1, D vo.s, D vO.9 , of 1.0,3.1, and 6.0 ~m, respectively, and 1000/0 of the fogvolume is made up of droplets ::::; 20 ~m

(Hoffmann et al. 2008).

Test grids of sentinel mosquito cages

To measure efficacy of ULV and thermal fogapplications we set up 150 m X 150 m grids ofcaged sentinel mosquitoes based on the experi­mental design first outlined by Rogers et al.(1957). We delineated grids using I-m resolution3-band (RGB) natural color Digital OrthophotoQuadrangles (DOQs; available from the USGeological Survey, http://seamless.usgs.gov/) inthe ArcGIS 9.2 (Environmental Systems Re­search Institute, Redlands, CA) geographic in­formation system (GIS). Grids for the _ULVapplication and the thermal fog application wereseparated by at least 125 m to preclude cross­contamination by the insecticide. Each gridconsisted of 25 upright posts (Fig. 2A)-either2-in.-diam PVC pipes (Florida test) or plasticfence poles (California test)-arrayed in a regularpattern of increasing distance from the spray line(Fig. 3), upon which were mounted sentinelmosquito cages at 0.3 m and 1.2 m. The locationsfor the 25 posts were surveyed using coordinatesextracted from the GIS and input into a GeoXT(Trimble, Sunnyvale, CA) handheld geographicposition system (GPS) device. The GPS unitprovided ----3 m precision (uncorrected), whichwas sufficient to quickly stake out the 25 sentinelpoints, followed by a final straightening of rowsby eye based on north-south and east-west

186 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, No.2

Driving Directionand

Spray Line

AS 85 C5 05 E5

A4 84 C4 04 E4

II

IA3 183 C3 103 IE3

IA2 82 C2 02 E2

A1 81 C1 01 E1

15 30 45 60 75 90 105 120 Meters

Fig. 3. Aerosol pesticide application test grid, showing driving direction, spray path, and arrangement of the 25sentinel mosquito poles (AI through E5). The "ribbon ladder" droplet detection apparatus (represented by shortthick black bars) were deployed adjacent to pole positions A3, B3, C3, D3, and E3 (refer to Fig. 2B image).

reference lines measured on the ground using a200-m measuring tape. We set the GPS at theNAD 1983 datum for spatial reference to UTMZone 17N (Florida test) or UTM Zone 11 N(California test) to match the USGS DOQs.

Each sentinel cage contained ---25 colony­reared female Aedes taeniorhynchus Wiedemann(Florida test) or Culex quinquefasciatus Say(California test). The Ae. taeniorhynchus colonyoriginated from a strain sampled in 1952 inOrlando, FL, and has since been maintained onmembranes at the Gainesville laboratories. Thecolony has not been exposed to pesticides duringrearing and is considered a susceptible strain;preliminary tests of oral toxicities with per­methrin are comparable to known susceptiblecolonies of Cx. quinquefasciatus and Ae. aegypti(Allan, personal communication). The Cx. quin­quefasciatus colony originated from a 2004 fieldcollection in Bakersfield (Kern County), Califor­nia. The colony has not been expose@ topesticides during rearing and is considered asusceptible strain; preliminary tests with pyre­thrum and permethrin in bottle bioassays causedeath in ::s 15 min (Wittie, personal communica­tion). We stored sentinel cages containing mos­quitoes in coolers for travel to the field sites andwaited until --- 30 min prior to the applications toattach them to the posts. The sentinel cages werecylindrical paper food containers (8.5 cm wide X4.5 cm deep; Neptune Paper Products, Newark,

NJ) with the paper disk bottom removed, coveredby nylon tulle mesh held fast with a rubber bandon each end (Fig. 2A). We attached sentinel cagesto posts using disposable hook-and-Ioop cableties threaded through one of the rubber bands.The design of the assembly permitted the posts tobe easily rotated to orient the open ends of thecages with wind direction immediately before theapplication and thus maximize exposure of themosquitoes to the aerosol clouds. In both theFlorida and California experiments, a set of 5posts each carrying 2 control mosquito sentinelcages at 0.3 ill and 1.2 ill heights was situated inupwind untreated areas at least 150 m from thetreatment grids.

We applied the insecticide spray in a singleswath along the west side of each grid, and ULVand thermal fog applications commenced simul­taneously during minimum acceptable wind con­ditions just before dark, when adulticide opera­tions are typically conducted to coincide withadult mosquito activity. The ambient conditionsat the Florida test consisted of the spray linemoving downwind (i.e., in the direction the windwas blowing; see Fig. 4); however, at the Cali­fornia test the spray line traveled upwind (i.e.,against the direction of the wind; see Fig. 5). Weadjusted vehicle speeds to the recommended flowrate for each device so that the single pass wouldresult in identical delivery of active ingredient toboth grids. To provide space and time for the

JCNE 2010 DESERT ULV AND THERMAL FOG COMPARISON 187

wind to move the aerosol cloud to potentially:'each all posts in the grids, we began theapplications with a lead of 40 m in the Floridatrial, o\ving to the southwesterly winds, and afollow-on of 40 m in the California trial, owing tothe northwesterly winds. The drive path (sprayline) was a line parallel to and 50 ft (15 m) fromthe first set of sentinel cages. We retrieved allmosquitoes from the field test grids 10 min post­spray and returned them to the laboratory incoolers for transfer to identical pesticide-freecages with a sugar water source (Bunner et al.1989). Mortality of sentinel mosquitoes wastallied during transfer to the new cages, whichwas carried out between 1.5 and 2 h post-spray,and we conducted subsequent observation ofmortality under controlled conditions of tem­perature, light-cycle, and humidity at 12 and 24 hpost-spray. For all transfers of adult mosquitoesinto or between cages we used CO2 (Florida trial)or chilling (California trial) to anaesthetize andimmobilize the insects. .

As an additional measure of efficacy at theCalifornia site, we performed mosquito popula­tion surveys for local wild Psorophora columbiaepopulations one night before and one night afterthe applications using 6 modified EncephalitisVector Surveillance (EVS) traps (Rohe and Fall1979) baited with dry ice (C02) without light setthroughout the 400 m X 400 ill date palmplantation.

Sampling droplet deposition in the test grids

To measure airborne spray passing through avertical plane, we used a specialized apparatusknown as the ribbon ladder (Fig. 2B). The ribbonladder consists of nine 1.0-m lengths of 2.5-cm­wide biodegradable cotton ribbons (Lab SafetySupply, Janesville, WI) stretched horizontally atl-ft (0.3 m) intervals across a 10-ft (3 m) -highframe created from Yz in. PVC pipe in the shapeof an inverted "D." Ribbons were attached withbinder clips to L-hooks screwed into the PVC ateach interval. We placed 5 ribbon ladders alongthe centerline of each test plot at 50, 100, 200,300, and 350 ft (15.3, 30.5, 61.0, 91.5, and106.7 m) from and perpendicular to the sprayline, and adjacent to the corresponding centerlinesentinel mosquito cages (Figs. 2 and 3). Approxi­mately 30 min before aerosol applications wedeployed all ribbons, and ----10 min after comple­tion of the treatment, the ribbons were visuallyexamined for the presence of fluorescent dropsusing a UV flashlight and then were cut near themount and stored separately in previously labeledplastic bags. We discarded ends of the ribbonstouching the L-hook to eliminate contaminationfrom previous applications. We packed thebagged deposition samples in a cooler fortransport to the Navy Entomology Center of

Excellence laboratory in Jacksonville, FL, wheresamples were stored in a refrigerator at -4°C foranalysis.

To prepare the ribbons for analysis, we poured50 mL of denatured ethanol wash liquid into eachbag so that all ribbons were wet, left the ribbonsto soak for 5 min, and then placed the bags on ashaker platform to shake for 4 min. The washsolution was then poured into two 10 ml cuvettesfor analysis with a Model 700 fluorometer(furner Design, Sunnyvale, CA). The rawfluorometer readings were converted to dyeconcentration using calibrations provided froma standard solution. The dye deposition onribbons was determined with the followingformula:

D _ 1,000 CHIS Vw

ep- As '

where

Dep Deposition of dye on ribbon surface(ng/cm2

)

C}VS Concentration of dye in wash solution(ppm or J-lg/ml)

Vli• Wash volume (m!)As Surface area of the ribbon samplers

calculated using dimensions of cutribbon (cm2

).

We then calculated the deposition of activeingredient on the ribbon using the ratio of activeingredient to dye in the tank mix. For DLV andthermal fogger, these ratios were 593.14 and 70.0for Florida trials and 339.4 and 34.0 forCalifornia trials, respectively. The depositiondata in both experiments were comparable,despite the different amounts of Dvitex OBfluorescent dye added, because of the normal­ization inherent in calculating the ratios.

Weather conditions

In both Florida and California trials, wemeasured wind speed, wind direction, and tem­perature with a Model 8100 3-D ultrasonicanemometer (R. M. Young Company, TraverseCity, MI) stationed at a point midway betweenthe 2 spray plots. Wind data were recorded at 10 ft(3 m) at a frequency of 4 Hz for 2.0-2.5 min,synchronized with the spray start, and runningfor the duration of the application. The ultrasonicanemometer measures velocity in 3 dimensionswith the capability to resolve 3 components intoresultant velocity and direction along 2 axes. Thefirst axis (horizontal wind direction) is the normalcompass azimuth in degrees starting from northat 0°; in these experiments, the instrument wasinstalled to read north as the sprayer travel

188 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, NO.2

Fig. 4. Mortality comparisons of ULV (northwest plot at the upper left) and thermal fog (southeast plot at thelower right) applications using malathion on caged Ae. taeniorhynchus under hot-humid conditions conducted inMay 2008 at Camp Blanding, FL, are depicted in a GIS. In the interpolated mortality surfaces, reds symbolize 80­100% mortality, yellows symbolize 70-80% mortality, greens symbolize 30-70% mortality, and blues symbolize 0­30% mortality.

direction, which was approximately north, asopposed to true north. The second axis (verticalwind direction) is an angle relative to the groundsurface, with 0° representing wind parallel to theground surface, positive angles indicating windmoving upward, and negative angles indicatingwind moving downward.

Mapping bioassay efficacy data

We mapped bioassay mortality data on theDOQs of the test areas in the ArcGIS 9.2 GIS byassigning the values of the mean 24-h mortalityfrequency between the upper and lower cages totheir respective pole positions and carrying out aninverse distance-weighted (IDW) interpolationanalysis for each grid. Interpolation by IDWestimates a continuous surface using actual mea­sured values from a defined set of points and isdesigned to constrain the effect of more dfstantpoints on the estimation of a value at any givenpoint. The IDW parameters were set to the defaultexponent of distance of 2, which controlled thesignificance of surrounding points on the inter­polated value, and a fixed search radius of 400 mfrom each interpolated point to take into accountall values in a grid. To enhance visualization of theinterpolated surface, we added null values tocorners of a square 15 m out from each side of thegrid. We color-coded the resulting interpolated

surface using a red to blue color ramp, where redssymbolized 80-100% mortality, yellows symbol­ized 70-80% mortality, greens symbolized 30-70%mortality, and blues symbolized 0-30% mortality.

RESULTS

Weather conditions

Wind speed, horizontal wind direction, andvertical wind direction are given in Fig. 6 for theFlorida and California sites. At the Florida site,temperature and relative humidity were 78°F(25.5°C) and 82%, respectively, during theapplication, and the wind speed ranged from1.1 to 4.4 mph with an average of 2.5 mph. Thehorizontal wind direction ranged between 1890

and 243° with an average of 213°, indicating thatwind was pushing spray into test grids at anangle of 9-63° from behind the sprayer. Thevertical direction ranged from -24° to 22° withan average of -1.3°. At the start of the trialwind speed declined dramatically as the vehicleswith ULV and thermal fog equipment drovenortheast, as indicated by the black arrows inFig. 4.

At the California site, the temperature andrelative humidity were 99°F (37.2°C) and 15%,respectively, during the application, and thewind speed ranged from 0.8 to 9.5 mph withan average of 5.0 mph. The horizontal wind

JUNE 2010 DESERT ULV AND THERMAL FOG COMPARISON 189

Fig. 5. Mortality comparisons of ULV (southwest plot at the lower left) and thermal fog (northeast plot at theupper right of the image) applications using malathion on caged ex. quinquefasciatus conducted in July 2008 underextremely hot-dry conditions in the southern California desert in the Coachella Valley, are depicted in a GIS. Referto Fig. 4 legend for key to interpolated mortality surfaces.

direction ranged between 325° and 411 ° (51 ° toNE; any direction in NE has been plotted as360°+ for clarity) with an average of 351 0,indicating that wind was between 9° NW and51 0 NE. This range of wind directions indicatedthat wind was pushing spray between a range ofangles of 9° into and 51 ° away from the testgrids. The vertical direction ranged from - 54°to 36° with an average of 1.6°. At the start ofthe trial wind speed was variable and changeddirection from the northwest to almost from duenorth (parallel to the northerly drive direction)and then back to a north-northwesterly directionas the vehicles with ULV and thermal fogequipment drove northward, as indicated bythe black arrows in Fig. 5.

Patterns of mortality: Camp Blanding site

Mortality in the north, west, and east sectionsof the thermal fog spray plot (Fig. 4) was high(96-100%) but declined to low values (5-50%) inthe southern portion of the plot; mortality in theULV plot was uniformly low (0-27%). In bothULV and thermal fog plots, at posts wheremortality was observed in sentinel mosquitoes,mortality had occurred in both upper and lowercages. Within a margin of ---.;25%, equivalentmortality in both ULV and thermal fog plots wasobserved in caged mosquitoes positioned at boththe 0.3 ill and 1.2 m heights. Notable exceptionswere observed only in the thermal fog plot, at

posItIons C1, D1, and £4, where mortality was>50% higher in the upper cage for Cl, and ~400/0

higher in the upper cages for Dl and £4. Nofluorescent aerosol droplets were observed onribbons under light from UV flashlights in theULV plot; however, fluorescent droplets wereobserved on ribbons in the thermal fog plot.Mean mortality in the control sentinel mosqui­toes at 24 h averaged between top and bottomcages was 20/0 with a range of 0-50/0.

Patterns of mortality: Coachella Valley site

Mortality was observed throughout the ther­mal fog spray plot but was highest in thesouthwest quadrant and fairly uniformly highthroughout the southern half of this study plot(Fig. 5). Mortality was recorded as much as 350 ft(105 m) into the date palm canopy to the east ofthe spray line in the thermal fog plot. Mortality inthe ULV plot 'was highest along the line of cages50 ft (15 m) from the spray line, ranging from 20/0at positions A5 and A3 to 87-880/0 at positionsAI, A2, and A4. Mortality was observedthroughout the date palm canopy in this plot,but declined rapidly to low levels after 100 [t(30 m). Within a margin of ~25%, equivalentmortality in both ULV and thermal fog plots wasobserved in caged mosquitoes positioned at boththe 0.3 m and 1.2 m heights, and droplets wereclearly visible on the deployed ribbon whenilluminated with UV flashlights at heights ranging

JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, No.2

2.52.0

~Thermal fogging ends

Mean =2.5 mphMax = 4.4 mphMin = 1.1 mph

1.51.0

IHorizontal direction (Opposite to sprayer travel = 360°)

Vertical direction (Horizontal wind = 0°)[Positive angle = upward wind; Negative angle = downward wind]

-100 ULV application ends~

0.0 0.5

190

A

10

'2 80-

S6"'0

CDCD0-

4en"'0c~ 2

0

400

- 300~c0

200""5~is 100"'0C

~ 0

Time after spray start, min

2.5

Mean =5.0 mphMax = 9.5 mphMin =0.8 mph

'"Horizontal direction (Opposite to sprayer travel =360")

Vertical direction (Horizontal wind =0')I [Positive angle =upward wind; Negative angle =downward wind]

-100 ULV application end~

0.0 0.5

B

10

'2 80-

S 6"'0CDCD0- 4en

"'0c~ 2

0400

- 300~

c 2000

U~ 100(5

"'0C a~

~Thermal fogging ends

1.0 1.5 2.0

Time after spray start, min

Fig. 6. Wind speed and direction during spray application: (A) Camp Blanding, FL; (B) Coachella Valley, CA.

from 0.3 to 3 m. Notable exceptions in the ULVplot were positions A3, Bl, and B5, wheremortality was >800/0 higher in upper cages forA3 and Bl, and >50% lower for B5. Similarly, in

the thermal fog plot, mortality was greater bymore than ~60% in upper cages at positions e2,C3, D3, and D4, and greater in the upper cage bynearly 40% at position E3; whereas mortality was

JUNE 2010 DESERT ULV AND THERMAL FOG COMPARISON 191

Comparative deposition from ULV and thermalfogger sprayers: Camp Blanding

On average, the ULV application resulted in163.4 ng/cm2 deposition, and the thermal fogapplication resulted in 173.0 ng/cm2 deposition(Table 1). Deposition data in Table 1 also showthat the deposition from the ULV application,averaged for all ribbon ladder heights, increasedwith increasing distance from spray line. In thecase of the thermal fog application, the trend wasreversed. Figure 7A shows active ingredientdeposition at different heights and distances fromthe spray line, as captured with the ribbonladders. The data indicate that the depositionfrom the ULV application was greater at lowerheights, especially on the 2 ribbon laddersfurthest from the spray line. The deposition fromthermal fog was greater at upper heights on the 3ribbon ladders closest to the spray line.

The active ingredient deposition was convertedto droplet density (number of droplets/cm2

) basedon the active ingredient contained by a dropletequivalent to volume median diameter (Dvo.s)generated by each application system. For thisconversion, the DVo.5 for ULV application andthermal fogger were 14.6 f.lm (Hoffmann et al.2007a) and 3.1 f.lm (Hoffmann et al. 2008),respectively. The resulting droplet densities from2 application systems at different heights anddistances from the spray line are presented inFig. 7B. It should be noted that droplets fromULV number in the 100s while droplets fromthermal fogger number in the 1,000s. Even atlocations of significantly higher deposition fromULV, droplet density from the thermal foggerapplication is significantly higher than the dropletdensity from the ULV application.

Comparative deposition from ULV and thermalfogger sprayers: Coachella Valley site

The overall deposition from the ULV applica­tion was 5.6 ng/cm2

, and the deposition from thethermal fog application was 5.8 ng/cm2 (Table 1).Deposition at the Florida site had been over 30­fold greater; at the California site, much of thematerial appeared to have been lifted away fromthe spray area and not deposited, althoughsufficient material fell and caused some mortality.As shown in Table 1, the deposition as measuredwith the ribbon ladders from the California ULVapplication, unlike the ULV application in

nearly 400/0 higher in the lower cage at positionE2. Pre-spray catch of Ps. columbiae from the 6EVS traps totaled 16,800 females; post-spraycatch from the 6 traps totaled 3,098 females.Mean mortality in the control sentinel mosqui­toes at 24 h averaged between top and bottomcages was 1%, with a range of 0-50/0.

Co)

~6' uo :.ao .S

8oo

.­00oo

\00\001./')'0 .~If) .1./') ,

o 0

N'oo

6'oo

ot-C"'lC"'l'0 'C"'lo ,("f") •

o 0

.-v)' .-0\"001./')('.1

0000"-'''-'-''-'

6' 6'o 0

_O.-~I./') ,,0 "

~~~~-O("f") ......."-' "-'''-' "-'

G' 6'o 0_O.-~t- ,,~ "

~~~~....... 0..-­"-'''-'-''-'

....... t- If) 00

'- . 0\00 . t- .000C"'l0.......

00' 6'o 00

.-0.-00"1 ,,~ "

~~s:~~S~·S~O\ot­

M~M~\Oot-o..- -

192

A

JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, No.2

Distance fromspray line 15 m 30m eo m 90m 105m

2.7 •••••r D••II--]mi----lIr----I=----i

2.1 ••mmm1.8

1.5'-••

1.2

0.9

o 200 400 600 800 0 200 400 600 800 0 200 400 600 800 0 200 400 600 800 0 200 400 600 800 1000

Deposition (ng/cm2)

IF?Z?ld ULV Spraying 111m Thermal fogging I

B

Distanee fromspray line 15m 30m eOm

2.7

2.4

E 2.1"'0C::::J 1.800,Q) 1.5>0

..c 1.2co

.EC> 0.9"w

I0.6

0.3

90 m 105 m

101 1Q2 103 104 105 1Q2 103 104 1Q5 1Q2 103 104 1()5 1Q2 103 104 1Q5 102 103 104 105

Droplet Density (number/cm2)

I~ ULV Spraying mm Thermal fogging IFig. 7. Deposition data from the Camp Blanding, FL, trial. The icon, drawn to scale, to the right of each chart

represents a post with sentinel cages at the 0.3 m and 1.2 m heights. (A) Deposition of active ingredient from 2application systems at heights and distances from·spray line. (B) Droplet densities from 2 application systems atheights and distances from spray line.

Florida, decreased with increasing distance fromthe spray line. However, similar to the Floridatrial, the California thermal fog applicationresulted in the average deposition decreasing withincreasing distance from the spray line. Figure 8Ashows very low deposition at distances fromspray line beyond 50 ft (15 m). At ;::::50 ft (15 m)

from the spray line the deposition was, in general,similar from the 2 application systems. On theother hand, Fig. 8B shows a large difference indroplet density between the 2 application systemsfor the California trial, with the thermal fogproducing significantly more drops at all dis­tances up to 90 m from the spray line.

DESERT ULV AND THERMAL FOG COMPARISON

105m

193

60402060 040

90m

2060 040

60m

2060 040

30m

20

JUNE 2010

A

Distance fromspray line 15m

2.7

2.4

2.1

1.8

1.5

1.2

0.9

0.6

0.3

20 40 60 0

Deposition (nglcm2)

II!i!ZiZI ULV Spraying IIDIII Thermal fOggingl

B

Distance fromspray line 15m 30m 60m gOm 105 m

2.7

2.4

2.1

1.8

1.5

1.2

o.g

0.6

0.3

101 1Q2 10' 1(f 105 1Q2 1(fl 104 10S 1Q2 1(fl 1Q4 10S 1Q2 1(fl 104 10S 102 10 3 104 105

Droplet Density (number/cm2)

I~ ULV Spraying =:II Thermal fogging IFig. 8. Deposition data from the Coachella Valley, CA, trial. The icon, drawn to scale, to the right of each chart

represents a post with sentinel cages at the 0.3 m and 1.2 m heights. (A) Deposition of active ingredient from 2application systems at heights and distances from spray line. (B) Droplet densities from 2 application systems atheights and distances from spray line.

DISCUSSION

With respect to spatial spread and depth ofpenetration, as measured by numbers and loca­:ions of sentinel Ae. taeniorhynchus killed, thermal:'og application of malathion was demonstrated to)e markedly more efficacious, compared to the

ULV application, in a hot-humid, low wind, openhabitat environment at Camp Blanding, FL(Fig. 4). Thermal fog application of malathionwas demonstrated to be moderately more effica­cious, compared to the ULV application, in killingcaged ex. quinquefasciatus mosquitoes, penetrat­ing at least 350 ft (105 m) into a date palm

194 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, No.2

plantation in a hot-dry desert environment in theCoachella Valley in southern California (Fig. 5).In both the Florida and California test areas,mortality in the sentinel control mosquitoes wasextremely low (~5%), indicating that the toxicaerosols had not reached the control mosquitoes,and also indicating that mortality in sentinelmosquitoes was due to the pesticide application.

Deposition was approximately 30-fold greaterat the Florida site when compared to theCalifornia site, presumably because of multiplefactors, including much lower temperatures andhigher humidities, much lower average horizontalwind speed (0.5-fold lower), negative averagevertical wind direction, and only moderate windturbulence. At the California site, much of thematerial in both plots was lifted and moved awayfrom the sentinel grids, although mortality stilloccurred heterogeneously throughout both plots.On the other hand, deposition increased furtherinto the ULV plot at the Florida test, suggestingthat droplets may have lifted for a short time butthen diffused to a wider area that was reflected bypeak mortality for the grid (11-20%) along theCI-C5 and DI-D4 line of cages, althoughsingular instances of 12-20% mortality were alsoseen at positions A4, B4, El, and E4.

A noteworthy outcome of this study is thatactive ingredient deposition was not necessarilypredictive of mortality (Table 1). For example, inspite of instances of equal or lesser amount ofdeposition in both trials, the thermal fog applica­tion consistently resulted in higher levels ofmosquito mortality than from the ULV applica­tion. This difference in deposition was significantin the Florida trial for distances ~ 300 ft (90 m)from the spray line in favor of ULV (Table 1), yetthe thermal fog application resulted in a muchgreater and spatially more extensive mortality.Similarly, the deposition in the California trial at200-300 ft (60-90 m) for ULV was double ortriple that of thermal fog (Table 1), yet themortality at these ribbon ladder positions wassignificantly higher for the thermal fog applica­tion. This incongruity could be attributed to 2factors. First, taking, for example, the Floridadata shown in Fig. 7, the droplet density for thethermal fog application (DVO.5 = 3.1 !-tm) was 104times more than that for the ULV application(DVO.5 = 14.6 f.lm). This means that, even thoughthe average total deposition in ng/cm2 front'eachtechnology is not greatly different (Table 1), thereis at least a 100-fold greater chance that a dropletwill come in contact with a mosquito in thesentinel cage in a thermal fog application versus aULV application. Second, certain modes of airturbulence help spread any aerosol cloud to alarger area: for instance, turbulence of a scalelarger than the spray cloud moves the wholecloud with it, turbulence of a scale smaller thanthe cloud helps to spread the cloud and increase

its size. Thus, under low wind conditions, asfound during applications during, the Floridatrial, small-scale turbulence is prevalent, and theaerosol cloud spreads out and disperses in space.Evidence in support of this scenario is found inthe general equality of mortality in upper andlower cages in the thermal fog plot in the Floridatest, as opposed to more frequent instances ofheterogeneous upper and lower mortality fre­quencies in the thermal fog plot in the Californiatest where wind conditions were more harsh.Given the relatively small size of droplets inthermal fog application, the small magnitude ofturbulent velocity could have propelled thedroplets around and into the cages. On the otherhand, droplets from the ULV application wouldhave required a relatively higher magnitude ofturbulence for dispersal to equal that experiencedby the thermal fog aerosol cloud. Given thesudden drop in wind velocity just after the ULVtreatment began, the majority of the ULVdroplets simply drifted to the ground and didnot reach sentinel cages in great numbers.

The results also indicate that a considerablysmaller amount of active ingredient (and smallerdroplet size) is sufficient to kill a mosquito asopposed to earlier findings (Himel 1969, Haile etal. 1982).

Following the trial at the Camp Blanding sitewe hypothesized factors that may have contrib­uted to the difference in performance between theULV and the thermal fog treatments. One factorwas the possibility that, relative to the capacity ofthe reservoir tank, a small amount of FyfanonULV was used in the 18-20, and the slightlybouncing drive over uneven ground at the test sitemay have caused intermittent delivery of themalathion. We returned to the test site in daylightand ran the 18-20 on the spray line using thesame vehicle and the same volume of pesticide,specifically observing the output from the ULVnozzle, and found that the 18-20 produced asatisfactory, even flow despite the bounces andthe low volume of material in the reservoir.Another factor hypothesized to contribute todifferences in efficacy between the ULV and thethermal fog technology in the Florida test was thepossibility that the diesel present in the aerosolproduced by the Silver Cloud was toxic tomosquitoes or enhanced the toxicity of themalathion in the burned solution. Following thetechnique described in Pridgeon et al. (2008) wecarried out a topical application screening ofunburned diesel, diesel burned through a thermalfogger, technical malathion, unburned 1% Fyfa­non ULV-diesel solution, and 1% FyfanonULV-diesel solution burned through a thermalfogger. These compounds had been collectedduring the mass equipment tests documented inHoffmann et al. (2008) and stored in a refrig­erator. The mosquito species selected for testing

JUNE 2010

100

90

80

70

~ 60

~ 50Ci3t

400~

30

20

10

0

DESERT ULV AND THERMAL FOG COMPARISON

Ae. taeniorhynchus

1\\1 1:1 • I II

195

Concentration (ppm)

ex. quinquefasciatus

100

90

80

70

~ 60

.~CiS 50t

~ 40

30

20

10

O+-.--,....-,......,---..~...IiIiL..r-.-...,..~--.---,-I ....L.rlla.,-L....r--~....--rl--.-.......,...-....,..-...,r---T...-..~.-.....-....r--rIL.r----.-L.rII...L....r--L.r--...,-...,

Concentration (ppm)

D Untreated

D Acetone

Diesel

• Diesel burned

• Fyfanon (1%)

• Fyfanon burned (1%)

• Malathion (technical)Fig. 9. Results of topical application of diesel-formulated Fyfanon before and after burning through a thermal

fogger. Experiments were independently replicated 3 times as completely randomized designs testing 30 insects foreach of 6 concentrations for each replicate according to the methodology of Pridgeon et al. (2008). Bioassays wereheld at 27°C for 24 h prior to scoring.

matched the species used in the Florida andCalifornia trials. The topical application screen­ing consists of applying small droplets ofcompounds directly on the thorax of testmosquitoes and recording mortality over time.The results shown in Fig. 9 indicate that dieselhas essentially no toxicity on the tested mosquitospecies: the lowest concentration of diesel causedclose to zero mortality, whereas a concentrationof Fyfanon ULV-diesel at half that of the rawdiesel caused 70-1000/0 mortality depending onthe species tested.

Overall, these findings echo the results of manyof the seven comparative studies cited earlier:thermal fog applications may result in equal orgreater mortality against sentinel mosquitoeswhen compared to ULV applications. In addi­tion, the present study provides evidence that thesuperiority of thermal fog over ULV with respectto mortality in sentinel mosquitoes holds upacross 2 very different environments. Impor­tantly, given the specific questions of the presentstudy regarding potential efficacy in current USmilitary scenarios, ULV did not function as well

196 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 26, No.2

as thermal fog in the hot-dry desert environment.Another factor should enter into considerationfor military use of one or the other of thesetechnologies: in. a hostile military environmentthe opportunities for an aerosol public healthinsect control operation may be extremely nar­row. Weather events took place in each testedenvironment that demonstrated the flexibility ofthermal fog and the limitations of ULV. Ascompared in the simultaneous application ofULV aerosols, the thermal fog aerosol resultedin high kill throughout more than 50% of thetarget area despite a sudden loss of wind energy inthe Florida trial, and the thermal fog aerosolresulted in kill further and wider through an areawith tall vegetation despite high wind energy andsudden changes in wind direction in the Califor­nia trial. If these 2 events had taken place intactical scenarios and ULV had been the technol­ogy of choice for public health insect controloperations, the opportunities to effectively con­trol disease vectors or nuisance insects could havebeen lost. There is clearly an advantage to using aquiet, invisible technology such as ULV aerosolin a hostile military environment; however, thistechnology should also be effective against thetarget insects, and a new evaluation may beneeded to weigh the costs and benefits of thenoisy clouds of the thermal fog. For instance,ULV does put out much smaller amounts ofmaterial into the environment, but poor perfor­mance of an application could require multiplereapplications and the low volume advantagediminishes, not to mention increased exposure ofpersonnel to potentially hostile contact duringrepeated visits to the same area.

Although these results indicate that more workshould be done to further evaluate thermal fogefficacy in desert environments, future studiesshould also look at efficacy against wild mosquitopopulations as well as sentinel mosquitoes. Thetrap counts for Ps. columbiae in the Californiatrial before and after the applications suggest thatapplications were effective against real-worldpopulations as well as sentinels, but the arrange­ment of traps and aerosol test grids did notpermit us to determine which technology mayhave had a greater effect on reductions in the wildpopulations. Future studies should also look at avariety of ULV and thermal fog equipment, aswell as pesticides, selected from the Depart",-entof Defense lists (Armed Forces Pest ManagementBoard 2010); and future studies should also takeplace in arid-hot areas with wild populations ofOld World mosquitoes and sand flies, such asthose in Kenya or Egypt, and eventually fieldedfor tests in-theater with the US military. Effortsshould be made to perform spray trials throughsentinel grids in a variety of vegetation densitiesand/or profiles in desert environments, and if onespray technology consistently outperforms the

other, efforts should he focused on optimizationof that spray technology. Optimization of a spraytechnology may- include trials with 'a range ofnozzles, pesticide dilutions or formulations, tim­ing and speed of delivery, or machine settings, orit may include trials with experimental pesticidesor spray equipment.

ACKNOWLEDGMENTS

This research was supported by the Depart­ment of Defense (DoD) through the DeployedWar Fighter Protection Program, and the U.S.Department of Agriculture (USDA) AgriculturalResearch Service. The use of equipment andproducts in this study does not constituteendorsement by the USDA, the DoD, or theUS Navy. Technicians from the Coachella ValleyMosquito and Vector Control District andpersonnel from the California Department ofPublic Health and the University of CaliforniaDavis Center for Vectorborne Diseases kindlyprovided expert assistance in the field. .

REFERENCES CITED

Armed Forces Pest Management Board. 2010. DoDstandard pesticides list [Internet]. [accessed April 30,2010]. Available from: http://www.afpmb.org!standardlist.htm.

Bunner BL, Perich MJ, Boobar LR. 1989. Culicidae(Diptera) mortality resulting from insecticide aerosolscompared with mortality from droplets on sentinelcages. J Med Entomol 26:222-225.

Cope SE, Strickman DA, White GB. 2008. TheDeployed Warfighter Protection Research Program:finding new methods to vanquish old foes. Army MedDept J April-June:9-20.

Dalton R. 2008. Battlefield insectica. Nature 454: 18-19.Haile DG, Mount GA, Pierce NW. 1982. Effect of

droplet size of malathion aerosols on kill of cagedadult mosquitoes. Mosq News 42:576-583.

Himel CM. 1969. The optimum size for insecticidespray droplets. J Econ Entomol 62:919-925.

Hoffmann WC, Walker TW, Fritz BK, Gwinn T, SmithVL, Szumlas D, Quinn B, Lan Y, Huang Y, Sykes D.2008. Spray characterization of thermal foggingequipment typically used in vector control. J AmMosq Control Assoc 24:550-559.

Hoffmann we, Walker TW, Martin DE, Barber JAB,Gwinn T, Smith VL, Szumlas D, Lan Y, Fritz BK.2007a. Characterization of truck-mounted atomiza­tion equipment typically used in vector control. J AmMosq Control Assoc 23:321-329.

Hoffmann WC, Walker TW, Smith VL, Martin DE,Fritz BK. 2007b. Droplet-size characterization ofhandheld atomization equipment typically used invector control. J Am Mosq Control Assoc 23:315-320,

Holland JM, Jepson PC. 1996. Droplet dynamics andinitial field tests for microencapsulated pesticideformulations applied at ultra low volume usingrotary atomizers for control of locusts and grass­hoppers. Pesticide Sci 48: 125-134.

JUNE 2010 DESERT ULV AND THERMAL FOG COMPARISON 197

LaMer VK, Hochberg S, Hodges K, Wilson I, Fales lA,Latta R. 1947. The influence of the particle size ofhomogeneous insecticidal aerosols on the mortality ofmosquitoes in confined atmospheres. J Colloid Sci2:539-549.

Linley JR, Jordan S. 1992. Effects of ultra-low volumeand thermal fog malathion, Scourge, and Naledapplied against caged adult Culicoides furens andCulex quinquefasciatus in open and vegetated terrain.J Am Mosq Control Assoc 8:69-76.

Linthicum KJ, Allan S, Barnard D, Becnel J, Bernier U,Britch S, Clark G, Cooperband M, Geden C, HogsetteJ, Kline D, Pereira R, Pridgeon J, Quinn B, Welch C,Zhao L. 2007. Mosquito and fly control research by theUSDA-ARS Center for Medical, Agricultural andVeterinary Entomology (CMAVE) in the DeployedWar-Fighter Protection CDWFP) program. Proc PapMosq Vector Control Assoc Calif75:131-133.

Lofgren CS. 1970. Ultralow volume applications ofconcentrated insecticides in medical and veterinaryentomology. Ann Rev EntomoI15:321-342.

Lothrop H, Lothrop B, Palmer M, Wheeler S, GutierrezA, Gomsi D, Reisen WK. 2007a. Evaluation ofpyrethrin and permethrin ground ultra-low volumeapplications for adult Culex control in rural andurban environments of the Coachella Valley ofCalifornia. J Am Mosq Control Assoc 23:190-207.

Lothrop H, Lothrop B, Palmer M, Wheeler S, GutierrezA, Miller P, Gomsi D, Reisen WK. 2007b. Evalua­tion of pyrethrin aerial ultra-low volume applicationsfor adult Culex tarsalis control in the desert environ­ments of the Coachella Valley, Riverside County,California. J Am Mosq Control Assoc 23:405--419.

Lothrop HD, Lothrop BB, Gomsi DE, Reisen WK.2008. Intensive early season adulticide applicationsdecrease arbovirus transmission throughout theCoachella Valley, Riverside County, California.Vector-Borne Zoonotic Dis 8:475--489.

Mount GA. 1998. A critical review of ultralow-volumeaerosols of insecticide applied with vehicle-mounted

generators for adult mosquito control. J Am MosqControl Assoc 14:305-334.

Mount GA, Lofgren CS, Pierce NW, Husman CN.1968. Ultra-low volume nonthermal aerosols ofmalathion and Naled for adult mosquito control.Mosq News 28:99-103.

Mount GA, Pierce NW, Lofgren CS, Gahan JB. 1970.A new ultra-low volume cold aerosol nozzle fordispersal of insecticides against adult mosquitoes.Mosq Nell'S 30:56-59.

Phanthumachinda B, Samuthrapngse W, Chanthon K.1976. The control offield populations ofAedes aegyptiby malathion, fen itrothion, bioresmethrin and pyre­thrum ULV ground aerosols, and by malathion,DD VP and bioresmethrin thermal foggings. Bangkok,Thailand: Unpublished document WHO/VBC/76.651. 17 p.

Pridgeon IW, Pereira RM, Becnel J1, Allan SA, ClarkGG, Linthicum KI. 2008. Susceptibility of Aedesaegypti, Culex quinquefasciatus Say, and Anophelesquadrimaculatus Say to 19 pesticides with differentmodes of action. J Med Entomol 45:82-87.

Rathburn CB, Boike AH. 1972. Ultra low volume testsof malathion applied by ground equipment forcontrol of adult mosquitoes. Mosq News 32:183-187.

Rogers AJ, Beidler EI, Rathburn CJ lr. 1957. A cagetest for evaluating mosquito adulticides under fieldconditions. Mosq Nelivs 17:194-198.

Rohe D, Fall RP. 1979. A miniature battery poweredCO2 baited light trap for mosquito borne encephalitissurveillance. Bull Soc Vector Ecol 4:24-27.

Taylor RT, Schoof HF. 1971. Relative effectiveness ofmalathion thermal aerosols and ground-applied ULVagainst 3 species of mosquitoes. Mosq NeHlS31:346-349.

Wirat S, Somkiat B, Banyong M, Mongkol G, PoonyosR, Boonluan P. 1982. Tests of insecticides applied byUL V ground and thermal fogging equipment for thecontrol of Aedes aegypti in Thailand. Bangkok,Thailand: Unpublished document WHO/VBC/82.869. 17 p.