Journal oJ the American Mosquito Control Association, 17(3):196-2O5,2OO1Copyright @ 2001 by the American Mosquito Control Association, Inc.

LIGHT CARBON DIOXIDE, AND OCTENOL-BAITED MOSQUITOTRAP AND HOST-SEEKING ACTIVITY EVALUATIONS FOR

MOSQUITOES IN A MALARIOUS AREA OF THEREPUBLIC OF KOREA

DOUGLAS A. BURKETTI WON J. LEE,' KWAN W. LEE,] HEUNG C. KIM,3 HEE I. LEE,' JONG S. LEE,'E. H. SHIN,' ROBERT A. WIRTZ,4 HAE W. CHO.' DAVID M. CLABORN.5 RUSSEL E, COLEMAN6

nNo TERRY A. KLEIN?

ABSTRACT. Two field trials for commercially available and experimental mosquito traps variously baitedwith light, carbon dioxide, octenol, or combinations of these were evaluated in a malarious area at Paekyeon-Ri near Tongil-Chon (village) and Camp Greaves, Paju County, Kyonggi Province, Republic of Korea. The host-seeking activity for common mosquito species was determined using hourly aspirator collections from a human-and propane lantern-baited Shannon trap. The total number of mosquitoes and number of each species capturedduring the test were compared using 8 x 8 and 5 X 5 Latin square designs based on trap location. Significantdifferences were observed fbr the total number of mosquitoes collected in the 8 X 8 test, such that counterflowgeometry (CFG) with CO, > CFG with CO, and octenol > Shannon trap > Mosquito Magnet@ with octenol> American Biophysics Corporation (ABC) light trap with light, CO, (500 ml/min), and octenol > ABC lighttrap with light and dry ice > ABC light ffap with light and CO, > ABC light trap with light only. A concurrent5 X 5 test found significant differences in trap catch, where Mosquito Magnet with octenol > New Jersey lighttrap > EPAR@ Mosquito Killer with CO, > ABC light trap with light and dry ice ) Centers for Disease Control(CDC) light trap (manufactured by John W Hock) with light and octenol. Significant differences in trap catchwere noted for several species including: Aedes vexans, Anopheles sinensis, An. yatsushiroensis, An, Iesteri,Culex pipiens, and Cx. orientalis. Traps baited with octenol captured significantly fewer Cx. pipiens than thosenot baited with octenol. Likewise, no Cx. orientalls were captured in octenol-baited traps. Host-seeking activityshowed a similar bimodal pattern for all species captured. Results from these field trap evaluations can signifi-cantly enhance surveillance efforts. Significantly greater numbers of mosquitoes were captured with mosquitotraps using counterflow technology (e.g., Mosquito Magnet and CFG traps) when compared to standard lightand carbon dioxide-baited traps. Additionally, field evaluations demonstrate that various traps can be utilizedfor isolation and detection of arboviruses and other pathogens.

KEY WORDS Korea, light traps, Anopheles sinensis, mosquito surveillance, attractants

INTRODUCTION

The effectiveness of several commercially avail-able and experimental mosquito traps using variouscombinations of attractants were evaluated in a ricefield habitat in a malarious area near a U.S. militaryinstallation (Camp Greaves) and at Paekyeon-Ri,near Tongil-Chon (village), Paju County, in North-ern Kyonggi Province, Republic of Korea (ROK).Other mosquito trapping studies have been con-ducted in the ROK (Self et aI. 1973: Lee et al.

rDetachment 3, U.S. Air Force Institute for Environ-ment, Safety, and Occupational Health Risk Analysis (AF-IERA), Okinawa, Japan, APO AP 96368.

2 Korean National Institute of Health, Department of Vi-ral Diseases, 5 Nokbun-Dong, Eunpyung-Gu, Seoul 122-701, Republic of Korea.

3U.S. Army, 5th Medical Detachment, 18th MedicalCommand. Unit 14247. APO AP 96205-0020.

a Entomology Branch, Centers for Disease Control andPrevention, F-22, 47OO Buford Highway, NE, Atlanta, GA30341-3724.

5 Uniformed Services University of Health Sciences,4301 Jones Bridge Road, Bethesda, MD 20814-4799.

6 Department of Entomology, U.S. Army Medical,Armed Forces Research Institute of Medical Sciences.APO AP 96546.

? Preventive Services Directorate, 18 Medical Com-mand. Unit 15281. APO AP 96205-0054.

1984; Joo and Wada 1985; Shim et al. 1990: Baikand Joo 1991;' Lee and Ree l99l; Joo and Kang1992; Yang and Yu 1992; Shim et al. 1997:' Kim etal. 1999; 2000; Strickman et al. 2000). Most ofthese investigators used New Jersey (NJ) lighttraps, black light traps, livestock-baited windowtraps, and human and livestock landing collections.The Korean Ministry of Health and Social Affairsbegan a peninsula-wide vector surveillance pro-gram in 1969 (Lee et al. 1971, Ree et al. 1973).Likewise, the U.S. Army has used NJ light traps tomonitor mosquito populations at fixed U.S. militaryinstallations since the end of the Korean War, butdid not institute an organized ROK-wide vector sur-veillance program until 1974 (Lee et al. 1984). Mis-cellaneous and inconsistent mosquito surveillanceoperations using a variety of commercially avail-able light trap designs and attractants are also con-ducted throughout Korea (and throughout most ofthe rest of the world) by different branches of themilitary, either at the installation level or duringfield exercises and contingency operations. Sur-veillance results from these efforts that use an as-sortment of mosquito traps and attractants cannotbe compared and the results are of questionable val-ue.

The poor condition of specimens collected fromNJ or other nonselective light traps often results in

196

Ssprrt'aspn 2001 Mosqurro Tnep EveluerroNs

Table 1. Description of mosquito traps and attractant combinations

Trap type Light COt Octenolr Trial

Shannon trapABC light trapABC light trapABC light trapABC light trapABC counterflow geometryABC counterflow geometryMosquito Magnet@CDC light trapNew Jersey light trapEPAR@ Mosquito Killer

NoNoNoNoYesNoYesYesYesNoNo

YesYesYesYesYesNoNoNoYesYesNo

YesNoNoNoNoNoNoNoNoNoNo

YesNoYes2Yes3Yes3YeslYes3YesaNoNoYes5

8 x 88 x 8Both8 x 88 x 88 x 88 x 8Both

5 x 5J X )

5 x 5'Octenol cartridge.'�Source of CO., dry ice.I Source of COr, gas cylinder (500 ml/min).' Source of COr, propane gas.5 Source of COr, gas cylinder (750 ml/min).

inaccurate identification of morphologically similarmosquitoes, which may result in errors when in-criminating malaria or arbovirus vectors. Addition-ally, specimens are generally collected dead and inrelatively small numbers, which precludes adequatesample sizes for determining malaria and arbovirusinfection rates. This is the 1st published report thatevaluates the attractiveness of commonly usedmodern mosquito traps and attractants for medical-ly important mosquitoes in the ROK. Moreove!few of the previously reported trapping studies inthe ROK have differentiated among the morpho-logically similar Anopheles malaria vectors (An. si-nensis Wiedemann, An. lesteri Baisas and f{lu, An.yatsushiroensis Miyazaki), all of which are impli-cated in the transmission of malaria in the ROK.Hourly host-seeking activity of the commonly cap-tured mosquitoes also was recorded in this study toidentify periods of greatest human exposure.

MATERIALS AND METHODS

Two mosquito trap evaluation field trials wereconducted during June 19-30, 2000, from 1900 to0600 h. Because of time limitations and the elec-trical requirements for some traps, both an 8 X 8and a 5 X 5 Latin square design were employed toevaluate the effectiveness of various traps and at-tractants for mosquitoes. Trap data were not usedfor days when it rained during the test period. Trapdata were transfbrmed to log (x * l) before anal-ysis. Trap, day, and position effects were evaluatedusing a 3-way analysis of variance (SAS Institute1995). Mean comparisons were made using theRyan-Einot-Gabriel-Welsh multiple range test (o.: 0.05). The positions of each trap were changednightly so that each trap would occupy every po-sition during each of the test periods. After eachtrap night, mosquito collections were placed inshipping containers over dry ice and transported tothe 5th Medical Detachment Entomology Labora-tory, where they were identified using keys specific

to Korean mosquitoes (Lee 1998) and counted.Anopheles were separated by species and date ofcollection and sent to the Armed Forces ResearchInstitute of Medical Sciences, Bangkok, Thailand,to determine malaria infection rates (by enzyme-linked immunosorbent assay). Culicine mosquitoeswere separated by species, placed in cryovials (30/vial), and then maintained on dry ice. Specimenslater were sent to the U.S. Army Medical ResearchInstitute of Infectious Disease for virus isolation.Both of the latter studies will be reported separate-ly.

8 x 8 rice .field tapping study. Trials were con-ducted in approximately 4 ha of terraced rice fields(37'54'N, 126"43'E) adjacent to and northeast ofCamp Greaves (U.S. Army installation) in theROK. Traps were positioned on elevated walkwaysthat separated the terraced rice fields. One hundredmeters or more separated the traps. Several of thetrap locations were bordered by primary woodlandhabitat (mixed deciduous and coniferous forest withunderbrush). The 8 x 8 trial trap and attractantcombinations are shown in Table l.

Human landing/biting collections were not madeduring this investigation. As a substitute, humanhost-seeking/attractant collections were conductedusing mouth aspirators from a 6 x 4 x 6-ft whitecotton Shannon-type trap (Service 1993) baitedwith 2 human collectors and a propane lantern(Model 5I52D7OOT, Coleman Company Inc.,Wichita, KS) placed about 40 cm from the ground.A 20-cm gap at the bottom of the trap allowed mos-quitoes to enter the trap. Two groups of 2 collectorseach manned the trap. Groups were rotatedthroughout the test to reduce collection bias. Mos-quitoes landing on the outside or inside of theShannon trap were aspirated continuously through-out the collection period. Captured mosquitoeswere placed in a screen-topped pint carton at hourlyintervals. At the termination of each hour/collectionperiod, mosquitoes were placed in a cooler, At theend of the daily collection, each carton was placed

1 9 8 JounNaL oF THE AMERTcaN Mosqurro CoNrnol Assocr,qrloN Vol. 17, No. 3

on dry ice and transported to the 5th Medical De-tachment, Yongsan, Korea.

Two carbon dioxide (Cor)-baited Centers forDisease Control (CDC)-type traps (TiapkitDl andTiapkitl, American Biophysics Corp., East Green-wich, RI) were evaluated; 1 with 1.5 kg of dry iceand the other with CO, dispensed from a 9-kg com-pressed gas cylinder. Although the CO, flow ratefor the dry ice was not measured, Reisen et al.(2000) reported an average CO, release rate of 500mUmin for a 1.5-kg block of dry ice in a similartrap. Locally obtained compressed gas cylinderswere used as the source of carbon dioxide for theother ABC light traps. The CO, flow rate (500 ml/min) was controlled using regulators, restrictioncouplings, and filters (Flowkitl, American Bio-physics Corp.). Battery power was provided usingPowersonic@ (PowerSonic Corp., San Diego, CA)6-V, 10-amp-h rechargeable gel cell batteries to runthe fan motor and incandescent light. Tiaps werehung from tripods constructed from aluminum tentpoles so that the light sources were approximately60 cm from the ground. The ABC traps were usedas received from the manufacturer and included astandard C]N4-47 bulb, and a 3-in.-diameter,4-blad-ed fan inserted into a plastic housing and coveredwith a rain guard. The dry ice-baited ABC lighttrap had an insulated container above the rain guardto hold the dry ice. To save battery power, bothtraps were operated with light set to flicker (32.5Hz). Studies show that the ABC light traps are rep-resentative of other typical CDC-type traps. For ex-ample, Vaidyanathan and Edman (1997) compared11 trapping methods and found no significant dif-ference in trap collections between the John W.Hock CDC-type traps (John W. Hock Company,Gainesville, FL) and similarly equipped ABC lighttraps. They also observed no significant differencesin trap collection numbers for ABC traps with bulbsset to steady or flicker.

Two counterflow geometry (CFG) traps wereevaluated, I with an octenol cartridge (OCTI,American Biophysics Corp.) and the other without.Neither of the traps used a light source, and bothwere baited with CO, from a compressed gas cyl-inder dispensed as described above for the ABClight traps. Both were operated using 6-V recharge-able batteries as described above. The CFG trapsuse 2 fans simultaneously that move air in oppositedirections. A smaller, 40-mm fan sends a COr-en-riched attractant plume down and out of a central5-cm pipe extending medially and slightly beneaththe bottom of the trap. Simultaneously, a strongerSo-mm fan creates an updraft through the trap en-trance that surrounds the central downdraft pipe,forcing attracted insects into the trap's internal col-lection chamber. As of this writing, CFG traps, asdescribed by Kline (1999), are not commerciallyavailable.

As with the CFG traps, the Mosquito Magnet@(Pro Model, American Biophysics Corp.) uses a

similar counterflow technology to capture insects.Propane gas supplied by 20Jb tanks powered thefan motors, produced heat, and generated CO, (aby-product of combustion). Otherwise, the trap wasoperated (including the OCT1 octenol cartridge)per manufacturer's instructions. The trap hung froma wheeled stand that placed the opening 60 cmabove the ground. As with the other traps, the Mos-quito Magnet was operated nightly from 1900 to0600 h and shut off during the day. Collection netsfrom all the traps were emptied and replaced daily.

5 X 5 trapping study. Trials also were conductedon Camp Greaves (37"53'N, 126"43'E) about 0.5km from the 8 X 8 study site. Traps were placedadjacent to areas where soldiers congregate (i.e.,dining halls, guard gates, recreational areas, andbarracks). The 5 X 5 trial evaluated the trap andattractant combinations shown in Table l. In gen-eral, trap sites were not ideal (competing lightsources) because they were limited by the electricaloutlet requirements for the NJ light trap andEPAR@ Mosquito Killer trap (Model MKS-H, En-vironmental Products and Research, Blytheville,AR).

The Mosquito Magnet and the ABC light trapwith dry ice were operated and specimens wereprocessed as described in the preceding 8 x 8methods. Similar to the ABC light traps, the CDClight trap (Model 1012, John W Hock Company)consisted of a CM-47 bulb (steady light) and a 3-in.-diameter, 4-bladed fan inserted into a plastichousing covered with a black plastic rain guard.Insects were collected alive in a mesh collection netand otherwise processed and powered as with theABC light traps. An octenol cartridge (OCTI) wasfastened with a wire to the plastic trap housing ofthe CDC trap at the same height as the light. Nocarbon dioxide was used for the CDC light traptreatment.

Mosquitoes and other insects collected in the NJlight trap (John W. Hock, Model I I 12) were cap-tured in a pint polypropylene jar containing a 6 X6-cm piece of dichlorvos-impregnated vinyl stripused as a killing agent. The NJ light trap used a25-W incandescent lamp and was positioned so thelight soirrce was 1.5 m above the ground.

The Mosquito Killer trap also required an elec-trical outlet to operate the fan, heating element, andelectric grid. Light is not used as an attractant, butthe trap is otherwise similar in size, shape (19 X24 1n., 19 lb), and color to that of a standard NJlight trap. An internal fan forces mosquitoes andother insects landing on the side of the trap (oralighting close to the trap opening located beneaththe rain shield) into and through an electrical gridwhere the insects are killed and blown into a net atthe trap bottom. A f,ne-mesh nylon net, fastenedwith an elastic cord around the trap bottom, is notnormally used for the EPAR Mosquito Killer, butwas supplied by the manufacturer for these trials.Carbon dioxide was supplied using a 9-kg com-

SeprpNrern 2001 Mosqurro Tnep EveluerroNs t99

pressed gas cylinder coupled to a manufacturer-sup-plied single-stage CO, regulator and a 100 X 1.5-cm rubber hose. The hose was attached to the sideof the trap using Velcro@ tape so that the end of thehose was just beneath the rain shield. A perforatedplastic disk supplied by the manufacturer was in-serted into the rubber hose to control the CO, flowrate, which was estimated to be about 750 ml/min.Before the start of the 5 X 5 trials, preliminarytrials were conducted using the Mosquito Killertrap operated per manufacturers instructions with-out using CO, for 4 consecutive nights. The trapwas turned on at 1745 h and turned off at 0615 hthe following morning. Nightly collections weretransferred to petri dishes and stored on dry ice,processed, and identified as described above for the8 X 8 test. Mosquitoes captured in this trap or theNJ light traps were not captured alive and subse-quently were not processed for arbovirus studies.

RESULTS

8 x 8 study

A total of 4O,764 mosquitoes was collectedamong rice fields during the trial (8 trap nights).Back-transformed means, P values, and significantdifferences for the common species collected areshown in 'fable 2. As noted in the table, significanttrap-location and day effects were found for somespecies. Significant differences in the total numberof mosquitoes captured were observed among dif-ferent traps and attractant combinations (P < 0.01)and among species (An. lesteri [P < 0.01], An. si-nensis fP < 0.Oll, An. yatsushiroensis lP < 0.011,Aedes vexans (Meigen) [P < 0.01], Culex pipiensL. [P < 0.01], and Cx. orientalis Edwards [P :O.O2l). Overall, the greatest numbers of anophelinemosquitoes were captured with the Shannon andMosquito Magnet traps. Conversely, significantlygreater numbers of Ae. vexans were captured in the2 CFG traps, followed by the Shannon and Mos-quito Magnet traps. The ABC traps baited with CO,(compressed gas and dry ice) and the CFG trapwith no octenol captured significantly more Cx. pi-piens than did the other traps. Interestingly, thosetraps baited with octenol as 1 of the attractants cap-tured signiflcantly fewer Cx. pipiens. Likewise, noCx. orientalis were collected in any of the octenol-baited traps. Ochlerotatus dorsalis (Meigen) (l),Anopheles sineroides Yamada (7), An. lindesayii ja-ponicus Yamada (5), Culex tritaeniorhynchus Clles(2), Cx. bitaeniorhynchus Giles (1), and Cx. vagansWiedemann (2) were collected in insignificant num-bers for analysis (totals are given in parentheses).Small numbers of at least 3 species of tabanid flies(species not determined) were captured in the Mos-quito Magnet and Shannon traps. Ceratopogonidmidges, although not routinely counted, also werecaptured in some traps.

o

o

q

> t l> , i t l7 ^ t l

E E I Ia . e - l l

* ' l l

o F l l+ > l t9 2 l lE F t l

N 6 I I

o ! t lF , o l l

: 1 E l l; € l lr ! i lo ̂ l l

! i l ld r . l l

* 5 Q l ld b 0 : l l

6 - , , t l> ! , , 1 1o . 9 a l li E ; l lu F oo l l

E i : o o l lg 6 . " "i e s l lo o e t l

E 3 6 l l- o d : t l

a e ! t lO b O > \ l l

6 Sq tr ll> - 6 t l

h 9 € l lo E o 0 l lE ! = t l

o - = t lt r . e I t l> o o l lE B H I Ib 0 | ! l l

E E g I Ie S e l lE , E o l lF V t r I Io , i d i li ; 9 - l l

' j E 10l lI X = t ln d e t l

€ ; l lo d t l

a E l l

e 8 t lF q l lo . = t l

u = t lU E I Io t la c l l

= 9 l lr H t l> s l lq t l

. o t lN o l le ' 6 l l? > t l. c d l l

o t l

U

T

m . =

+ 3

k -< !

, !.1f 6

9 O

-U+U

u)

oX d

v r )+

+r lr f

U

U+r a

U

+

p.q

* c { * i * F

E . " - €3 o o ? o o d +c o \ o c . t $ f . . o

cl \oc{

F c i N o o C ) c o O

C ) - € S € 6 Of-

P ( . ) € c ! o c i O

c . l * $ c . l € o \ i

d

* E - o! v F ! w ! v * F U

r o - c ' . r - r o o i do t € N o \r F O

I

a q q v l a n c ?i c - . 1 O t - - : a t O

o : @ r )O\ \O

-o

d q o g \ e o g q* \ O . O r ) \ O t - - Oo o \ c { ! +n o

€ * - o E o6 q , . l 9 , i n q

O C n O \ O \ O \ t OO\ al cn \O

. k

i t S : i . *i ' i 3 3 i S EE i P R F i Pr s ' a x i \ -- 6 s s s ! i . { . } iF < < < < U U

t r A

r"i '

, l 3 i lv - i o. . - p

* " o$ a c

\i e U< 1 1 9

r ) 3

< 5 ! !J 0 0 0 I )

< 6 A

200 JounNaL op rua AtragnrclN Moseurro ColrRor AssocrATIoN Vor-. 17, No. 3

Table 3. Hourly mean (ISEM) number of mosquitoes, by species, collected for Shannon trap collection in a ncefield habitat (n : 8 nights).

Speciesr l 900 2100 2300

Total mosquitoesAn. lesteriAn. sinensisAn. yatsushiroensisAe. vexansCx. pipiens

0.4 + 0.30.0 + 0.00.0 + 0.00.0 + 0.00.3 + 0.30.0 + 0.0

196 + 53.33 . 3 ! 2 . 7

32.6 + 13.839.5 + 14.6l t 9 + 4 0 . 11 . 8 + 0 . 6

283.8 + tO3;76.4 + 3.'l

45.6 + '19.9

49.8 + r7.O182 + 75.90 .1 + 0 . .1b

145 + 50.50.8 + 0.4

1 3 . 8 + 5 . 022.O + 7.2109 + 40-30.0 + 0.0

106 + 36.71 .5 + 0 .79 .9 + 5 .3

1 3 . 1 + 4 . 581.6 + 27.80.0 + 0.0

tAn., Anopheles; Ae., Aedes; Cx., Culex.

Shannon trap

A total of 8,653 mosquitoes was collected in theShannon trap. Arithmetic means with standard er-rors for each of the common species are listed byhourly collections from 1900 ro 0600 h (Table 3).No mosquitoes were captured between 1800 and1900 h. Overall, Ae. vexans (7OO + 268) was themost abundant species, followed by roughly equalnumbers of An. sinensis (176 + 68) and An. yat-sushiroensis (183 * 49; Thble 3). More (althoughnot significant) An. lesteri (18 t 7) were capturedwith the Shannon trap than with the other traps.Conversely, few C"r. pipiens or Cx. oientalis werecollected in the Shannon trap when compared tothose traps not baited with octenol as I of the at-tractants. All anophelines and Ae. vexans showedsimilar bimodal flight activities, which peaked dur-ing the periods 2OOO-22OO and 0300-0400 h. Rel-atively small numbers (mean : 2.9 per nighQ ofCx. pipiens were collected, but more than 5O7o ofthose collected were active between 2000 and 2100h. The mean (n : 8 nights) relative percent com-position of diel host-seeking activity for the mostcommon species captured at the Shannon trap isshown in Fig. l .

5 x 5 study

A total of 5,443 mosquitoes was collected atCamp Greaves during the 5 trap nights. Back-trans-formed means, P values, and significant differencesfor the common species collected in the 5 X 5 trialsare shown in Table 4. No day effects were found,although significance (cr : 0.05) due to trap loca-tion was found for An. yatsu.shiroensis and Cx. pi-piens. Overall, the Mosquito Magnet captured 3times as many mosquitoes as the next most attrac-tive trap (NJ) and more than 13 times as many asthe dry ice-baited ABC light trap and MosquitoKiller trap. With the exception of An. yatsushiroen-sls, the NJ light trap did not capture significantlymore mosquitoes (cr : 0.05) than the dry ice/COr-baited ABC trap and Mosquito Killer. Analysis us-ing multiple comparison procedures was not donefor An. lesteri (P : 0.08) or Cx. orientalis (P :0.44) because of the small numbers that were col-lected and the lack of significant (ct : O.O5) differ-ences among trap collections. Numbers too small

to analyze were collected for An. sineroides (3) andCx. tritaeniorhynchus (2). Nearly equal numbers(532 and 547, respectively) of mosquitoes werecaptured in the Mosquito Magnet for both the 5 X5 and 8 X 8 studies. However, perhaps because ofcompeting light sources, the dry ice-baited ABClight trap captured 77Vo fewer mosquitoes than thedry ice-baited ABC light trap in the 8 X 8 study.More An. sinensis were collected in the MosquitoMagnet, NJ light trap, and the Mosquito Killer thanin the ABC light trap with dry ice. The MosquitoMagnet and NJ light trap captured more An. yat-sushiroensis than the other traps. The dry ice-bait-ed ABC light trap, Mosquito Killer, and MosquitoMagnet captured the most Cx. pipiens. When op-erated with COr, the Mosquito Killer captured num-bers similar to the dry ice-baited ABC light trap.However, during a 3-night preliminary evaluationconducted before starting the 5 X 5 trial, the Mos-quito Killer trap was used per manufacturer's in-structions without CO, as a supplementary attrac-tant. During this 3-night period, no mosquitoes andvery few other flying insects were captured.

DISCUSSION

Few field studies in the ROK have evaluatedmosquito trap collections comparing different trapdesigns and attractants. A paucity of human casesof mosquito-borne diseases (malaria and Japaneseencephalitis UEI) in the ROK from the 1970s tothe mid-1990s provided little impetus for U.S.Forces Korea or the Korean civilian governmentand research community to reevaluate mosquitotrapping methods that remain an important com-ponent of their respective vector surveillance pro-grams. As a result, both military and civilian mos-quito surveillance programs have remained largelyunchanged for decades.

Starting in 1993-94, autochthonous Plasmodiumvivtu malaia reestablished itself in the civilian andmilitary communities in northwestern Kangwonand northern Kyonggi provinces bordering theROK Demilitarized Zone (DMZ) (Lee et al. 1998,Kho et al. 1999, Ree 2000). The reemergence ofmalaria, as well as the presence of JE, emphasizesthe need to develop efficient and effective mosquitosurveillance methods. For malaria, this is especially

SppmMsen 2001 Mosquno TRAP EVALUATToNS 201

0000 0t00

Table 3. Extended

0300 0400 Total

60.8 + 23.60.8 + 0.36.4 + 2.4

10.3 + 4 .843.1 + t1.O0.3 + o.2

56.9 + 29.10.9 + 0.53 .9 + 0 .68 .5 + 3 .6

43.6 + 25.20.0 + 0.0

68.0 + 35.21 .3 + 0 .5

12.6 + 7.68.9 + 2.7

44.9 + 25.60.3 + o.2

1 I 1 + 4 0 . 32 . 1 + l . O

33.3 + 16.922.1 + 6 .153.3 + 25.70.3 + o.2

52.O + r9.3l . l + 0 . 5

l 8 . l + 9 . 18 .5 + 3 .4

23 .9 + l t - 1o.3 + 0.2

1,082 + 359.718.0 + 6 .8176 + 68.7183 + 49.4700 + 267.72.9 + O.8

0.8 + 0.30.0 + 0.00 . 1 + 0 . 10 . 1 + 0 - t0.5 + 0.30.0 + 0.0

critical because the ROK had been largely malariafree since the late 1970s; malaria results in highmorbidity among civilian and military personneland adversely affects the economy and wartimereadiness; delayed onset of symptoms of 6-18months for the endemic P. vivax strain expedites itsspread into malaria-free areas, especially via U.S.soldiers returning to the United States; and im-provements in pathogen identification from mos-quito pools and trapping technology expands vectorsurveillance to include disease surveillance andprovides useful data for the medical communitywhen applying and prioritizing limited vector con-trol and medical resources.

Mosquito trap data from the ROK during the last25-30 years provide a useful historical perspectiveof vector populations, changes in species compo-sition, and control effectiveness. Unfortunately, thenonselective light-baited traps capture specimens inpoor and often unidentifiable condition or in num-bers insufficient for accurately estimating mosquitopopulations or evah,rating infection rates for arbo-virus or malaria parasites. With the exception of theMosquito Killer and NJ light trap used in the 5 X5 study, the other traps in this evaluation generallycaptured specimens alive and in good condition.Therefore, specimens collected in these traps weremore easily and rapidly identified and were usedfor arbovirus isolations and determining malaria in-fection rates. This is especially important for theidentification of emerging or reemerging and un-known arboviruses, because polymerase chain re-

Time

Fig. l. Relative percent composition of hourly aspi-rator collections for common mosquito species capturedat a human and propane lampbaited Shannon trap (n :8 nights).

action techniques would only capture those patho-gens for which tests are specifically conducted.

Inconsistent reporting for Anophele.s species cap-tured in older trapping studies makes direct com-parison of trapping studies difficult. Relativelylarge quantities of living, undamaged anophelinespecimens preserving key identification character-istics were collected in the traps using counterflowtechnology (Mosquito Magnet and CFG traps),making this technology more efficient for both vec-tor and nuisance mosquito identification.

Various review articles report that both An. si-nen.sis and An. yatsushiroensis have been found tobe naturally infected with sporozoites and to trans-mit malaria parasites to humans (Chow 1973, Palket al. 1988, Chai 1999, Ree 2000). Anopheles les-teri also has been collected throughout the Koreanpeninsula and is often misidentified as the morpho-logically similar An. sinensis. In China, eventhough An. lesteri is less common, it is considereda more important malaria vector than An. sinensis(Liu et al. 1986). Some confusion remains in Koreaover distinguishing An. sinensis and An. lesteri.Preliminary taxonomic examination of individualprogeny broods showed that the primary morpho-logicaf characters used to distinguish An. sinensisand An. lesteri are not always transferred to sub-sequent progeny. Because characters used to sepa-rate the adults of these 2 species are very similaqpopulation densities of An. lesteri may be greaterand their subsequent importance in malaria trans-mission more important than estimated by eitherlight trap or human-bait collections. It is criticalthat the malaria vectors are characterized and theirdistribution and populations determined as a basicpart of understanding the epidemiology of malariaand developing effective malaria vector and diseasecontrol strategies in the ROK.

In our field trials, both trap design and attractantwere found to be important deterrninants for cap-turing total and species-specific numbers of medi-cally important and nuisance species of Anophelesand culicine mosquitoes. Historically, large num-bers of the primary JE vector, Cx. tritaeniorhyn-chus, do not appear along the Korean DMZ untilmid-July. Subsequently, an additional trap studywas conducted in September 2O0O (published sep-arately) to evaluate many of the same commerciallyavailable traps. In other regions of the world, stud-

60

g 5 0EEo +o

.E

E r oos 1 0

01900 2000 2100 2200 2300 0000 0100 0200 0300 0400 0500

JounN,ql op rnr AvnnrcaN Moseuno CoNrnor- Assocra.rroN VoL. 17, No. 3

Table 4. Mosquito species composition and back-transformed nightly means for variously baited commerciallyavailable mosquito traps. Three-way analysis of variance and multiple comparison (Ryan-Einot-Gabriel-Welsh

multiple range test) were perfbrmed after 1og (x + l) transformations. Means within each row having the same letterare not significantly diff'erent (n : 5 nights; a : 0.05).

SpeciesrMagnet +

Octenol New JerseyABC +dry ice

Killer +CO.

CDC +Octenol P value

Total mosquitoesAn. lesteriAn. sinensisAn- yatsushiroensiszAe. vexansCx. pipiens2Cx. orientalis

547 a2.648.4 a13.2 a4O7 a3.8 ab0.0

1 5 6 bo.623.3 a15.7 al 0 l b1 . 1 b0.0

50.1 b0.01 . 8 b1 . 2 b29.O b10.4 a0.6

42.0 b0.08.5 ab2.6 b1 3 . 0 b3.9 ab0.0

7 . 1 c0.01 . 2 b1 . 1 b2.8 c0.4 b0.0

0.00010.080.0020.0010.00020.006o.44

'An., Anopheles; Ae., Aedes; Cx., Culex.'�Significant trap location effect (P < 0.05).

ies have compared and evaluated modern trap de-signs against mosquitoes and differences were ob-served in mosquito col lect ions using variousmodern trap designs and attractant combinationswhen compared to older established methods (Vai-dyanathan and Edman 1997, Kline 1999, Mboeraet al. 2000, Reisen et al. 2000). As with those stud-ies, we found that the addition of CO, increased thesize of trap collections. However, unlike the studyof Reisen et al. (2000), our study did not capturesignificantly fewer mosquitoes in dry ice-baitedtraps over those baited with CO, from compressedgas cylinders. For our study, mosquito representa-tion differed significantly among trap design andattractant combinations, ranging from (1 mosquitoper night for the ABC trap using only light as anattractant to >900 per night for the CFG traps. Alltrap designs in both trials consistently showed Ae.vexans as the most abundant mosquito capturedduring the trapping period. With few exceptions(notably An. lesteri and An. yatsushiroensis), otherlight trap studies in the same Korean province (Reeet al. 1973; Frommer et al. 1979; Lee et al. 1984;Kim et al. 1995; Shim et al. 1997; Kim et al. 1999:Kim 2000; Strickman et al. 2000) during a similartime of year found mosquito species compositionand relative abundance consistent with those forour study.

With the exception of the Shannon trap, thosetraps employing counterflow technology (MosquitoMagnet and CFG traps) captured significantlygreater numbers of mosquitoes than the variouslybaited ABC or CDC light traps. This agrees withthe studies of Kline (1999) and Mboera et al.(2000), which also found CFG traps to capture sig-nificantly more mosquitoes than Cor-baited CDCtraps. Unfortunately, the CFG traps are not yet (andmay not become) commercially available.

Some of the results from octenol-baited trap weresurprising. Kline (1994) reported that traps baitedwith octenol and CO, are synergistic for many spe-cies, but attractive to only a few when octenol isused alone. In the 8 X 8 study, when octenol-baitedtraps were compared to similarly baited traps with

no octenol, no synergistic effects were observed.Indeed, the opposite occurred for selected species.Although other authors (Becker et al. 1995, Mboeraet al. 2000) found no significant difference in Cx.pipiens in octenol- versus non-octenol-baited traps,we observed that all octenol-baited traps were re-pellent and captured signiflcantly fewer Cx. pipiensand no Cx. orientalis than unbaited traps. No pre-vious report has documented species-specific de-creases in trap collections for octenol-baited traps.Although not compared directly, the octenol-baitedCDC light trap (no COr) in the 5 X 5 trial capturedmore (7ltrap/night) mosquitoes than the unbaitedABC light trap in the 8 X 8 trial, which captured(1 mosquito per night. In another study (Reisen etal. 1999), NJ light traps were found to collect rep-resentative samples of several, but not all, medi-cally important mosquitoes. Simultaneous use ofCOr-baited CDC traps was required to capture mos-quito species not adequately represented in the NJtraps. In our 5 X 5 study, the unbaited NJ light trapcaptured signiflcantly more mosquitoes than the dryice-baited ABC light trap. This is inconsistent withother studies (Acuff 1976, Slaff et al. 1983), whichalmost invariably found 4 or more times as manymosquitoes (variable by species) collected in COr-baited CDC traps.

Although human landing/biting collections werenot used in our study per se, human bait was aprimary attractant to gather hourly flight informa-tion for host-seeking mosquitoes collected at theShannon trap. Dry ice-baited bed nets or Shannon-like traps have previously been used in Japan andKorea. Sasa and Sabin (1950) and Ree et al. (1969)captured similar species compositions to those inthis study, including Ae. yexans, An. sinensis, Cx.pipiens, and Cx. tritaeniorhynchus. Although notsignificantly different, the mean number of Anoph-e/es species captured by the Shannon trap wasgreater than for other traps. Assuming that human-bait and Shannon trap collections are similar, thisis not consistent with results of Davis et al. (1995),who captured 1.2 times as many Anopheles usinglight traps when compared to human-bait collec-

SoprBN.reen 2001 Moseutro Tn,rp EvllulrroNs 203

tions. The Shannon trap was not effective for allspecies. Although useful forAnopheles species, thehuman-baited Shannon trap did not effectively at-ttract Cx. pipiens or Cx. orientalis.

Time of year and weather conditions, among oth-ers, are obvious and well-documented factors thatplay significant roles in mosquito diel flight andhost-seeking activities. In Korea, location, evenwithin the same province, also seems to affect flightand biting activity. As an example, hourly human-landing collections reported by Strickman et al.(2OOO) found a high degree of variability for bitingactivity in An. sinensis at different locationsthroughout Korea. Interestingly, our study, con-ducted near Taesung-Dong, showed a similar bi-modal biting pattern to that observed by Strickmanet al. (2000). The results of an hourly light-trapcollection made in Kyungbook Province in July byJoo and Kang (1992) also demonstrated An. sinen-sis host-seeking activity similar to that found in ourstudy. However, results of the study of Joo andWada (1985) were very different from ours, withpeak activity from 1200 to 0400 l:.for An. sinensisfor cow-baited light trap collections. Depending onthe mosquito species, diel host-seeking and flightvariability for vector or nuisance species can com-plicate control efforts when trying to optimize thetiming of ultra-low-volume adulticide applicationsor other management efforts to correspond withthose of peak mosquito activity and susceptibilityto insecticidal treatments.

Based on either the total number of mosquitoes,the number of a vector species (Cx. tritaeniorhyn-chus), or both collected in NJ light traps, Yi et al.(1988) developed trap indices for U.S. military in-stallations in the ROK. A modified trap index (25females/night/trap [May I to July 31], and 10 1'e-males/trap/night [August I to October l5]) basedon this study is used by U.S. Army pest manage-ment and preventive medicine personnel as a meansto justify and implement control measures. Shouldthe U.S. military decide to change trapping proce-dures based on the results presented in our studyand use more efficient Mosquito Magnets or otherCFG traps for mosquito and vector surveillance atall or select installations (especially in malariousareas along the DMZ) the trap indices will have tobe adjusted. Additional research will be needed toestablish realistic and effective trap indices for se-lected traps. Changing traps to I of the more effi-cient CFG-type traps will not solve the present ma-laria situation, but would almost certainly improveand the ability to monitor vector populations andpossibly provide a significant reduction of mosquitopopulations in limited areas. The Mosquito Magnetin particular can be placed where needed and is notsubject to the restrictions and logistical problemsimposed by using NJ light traps or dry ice-baitedCDC-type traps. An additional benefit of switchingto I of the counterflow technology traps includesexpanding the current mosquito and vector surveil-

Iance program into a more comprehensive diseasesurveillance program where larger numbers of mos-quitoes are collected and where mosquito pools canbe tested for arbovirus and malaria infection rates.Taking statistical subsamples from the larger trapcollections can attenuate the additional burdenplaced on identification personal.

The vector surveillance options currently avail-able for use during military exercises and human-itarian, contingency, or other remote operations arelimited. For those surveillance options available,landing/biting collections are controversial and mayexpose personnel to life-threatening diseases. Me-chanical mosquito traps (i.e., NJ light traps) are im-practical because ofelectricity requirements and areotherwise bulky and difficult to pack. Likewise, thewide range of collapsible, and portable battery-op-erated CDC light or CFG traps are virtually uselessin the absence of carbon dioxide as an attractant.Dry ice or compressed gas cylinders are generallynot readily available during remote operations. Thecurrently available counterflow technology traps(i.e., Mosquito Magnet), although practical for usein cantonment and nearby training sites, are largelyimpractical for deployments because of their sizeand poor shipping ability. Ideally, a portable andeffective mosquito trap needs to be developed thatincorporates durability, increased portability, andeasily obtained and used attractants that collect asmany mosquitoes as possible. A collapsible CFG-type trap with the ability to use propane gas as anattractant and power source would be ideal. In gen-eral, propane gas and tanks are used worldwide forheating, refrigeration, and other uses and tend to beavailable worldwide and are less of a logisticalproblem to transport than compressed CO, gas cyl-inders or dry ice.

This study illustrates both similarities and differ-ences in adult mosquito collections based on bothold and new technologies, as well as results fromother parts of the world. The unique mosquito fau-na of the ROK makes it critical that tests are con-ducted for local species and not extrapolated forspecies elsewhere. The benefits of these new tech-nologies are great, including producing live collec-tions for arbovirus isolation, collecting specimensin good condition fbr increasing ease and reliabilityof identification, allowing uniformity of collectiontechnique, and collecting large numbers of mos-quitoes that may effectively reduce biting popula-tions when traps are employed near human activi-ties. These new technologies need to be employedin highly portable, effective, and efficient traps forinclusion as part of the U.S. military mosquito sur-veillance and control system. Evaluating new de-signs and technologies where U.S. soldiers are de-ployed will increase our knowledge of vectors andmethods for implementing vector and disease con-trol stratesies.

204 Joununl or rue AlrenIcaN Mosquno CoNrnol AssocIATIoN Vor-. 17. No. 3

ACKNOWLEDGMENTS

We are extremely thankful to James Kirkpatrick,Commande! 18th Medical Command, for his sup-port in evaluation of mosquito traps that may serveto enhance the U.S. Forces Korea Malaria Preven-tion Plan. We thank Wan Y. Kim from the U.S.Army Center for Health Promotion and PreventiveMedicine-Pacific (CHPPM-PAC), whose help inlogistics and as an interpreter were vital for com-pleting these joint agency field evaluations. We alsothank Kotu Phull, Zia Mehr, Anthony Schuster, andothers from CHPPM-PAC for their support and crit-ical review. Sincere thanks to Patrick T. Stackpole,Commander, and the other soldiers of CampGreaves, who provided facilities necessary forthese studies. Thanks to Brian Blaylock and VelvanWebb fiom Public Health Flight at the 5lst Aero-space Medical Squadron, Osan Air Base, ROK, andRichard Johnson, Armed Forces Pest ManagementBoard, MD, who kindly provided us with severaltraps. We would also like to acknowledge HenryMcKeithen, USDA-ARS, Gainesvi l le, FL, andChad McHugh, Air Force Institute for Environ-ment, Safety, and Occupational Health Risk Anal-ysis, who helped run SAS data sets. We are gratefulfor the assistance of Tae Kyu Kang, Edwin Huertas,Hee Choon Lee, Myong Wa Yi, and other membersof the Preventive Services Directorate, l8th Medi-cal Command, for their logistical and technical sup-port. Also, we thank Sonya Schleich, Commander,5th Medical Detachment, l68th Medical Battalion(Area Support), for providing much needed person-nel to perform the study. We greatly appreciate theassistance of Seung Ho Kang and Yong Bum Yi,Laboratory of Medical Zoology, Department of Vi-ral Disease, National Institute of Health, Ministryof Health and Welfare, Seoul, Republic of Korea.Finally, we wish to acknowledge trap donationsfrom American Biophysics Corporation and Envi-ronmental Products and Research. Funding for thisproject was provided by the Department of DefenseGlobal Emerging Infections Surveillance and Re-sponse System (DoD-GEIS), Walter Reed Army In-stitute of Research. and the 18th Medical Com-mand, 8th U.S. Army, Republic of Korea. Themention of trade names or commercial productsdoes not constitute endorsement or recommenda-tion for use by the Department of Defense, Centersfor Disease Control and Prevention, or the KoreanNational Institute of Health.

REFERENCES CITED

Acuff VR. 7976. Trap biases influencing mosquito col-lecting. Mosq News 36:173-176.

Baik DH, Joo CY. 1991. Epidemio-entomological surveyof Japanese encephalitis in Korea. Korean J Parasitol29:67-85.

Becker N, Zgomba M, Petric D, Ludwig M. 1995. Com-parison of carbon dioxide. octenol and a host-odour as

mosquito attractants in the Upper Rhine Valley, Ger-many. Med Vet Entomol 9:377-380.

Chai JY. 1999. Re-emergrng Plasmodium vivax malaria inthe Republic of Korea. Korean J Parasitol3T:129-143.

Chow CY. 1973. Arthropods of public health importancein Korea. Korean J Entomol 3:31-54.

Davis JR, Hall T, Chee EM, Majala A, Minjas J, Shiff CJ.1995. Comparison of sampling anopheline mosquitoesby light-trap and human-bait collections indoors at Ba-gamoyo, Tanzania. Med Vet Entomol 9:249-255.

Frommer RL, Pae CM, Lee TK. 1979. The distribution ofadult mosquitoes collected from light traps in the Re-public of Korea during 1977. J Korean Med Assoc 22:3'73-38r.

Joo CY, Kang GT, 1992. Epidemio-entomological surveyon malarial vector mosquitoes in Kyongbuk, Korea. Ko-rean J Parasitol 3O:329-34O.

Joo CY, Wada Y. 1985. Seasonal prevalence of the vectormosquitoes of Japanese encephalitis virus in Kyung-pook Province, Korea. Korean J Parasitol 23:139-15O.

Kho WG, Jang JY Hong STI Lee HW, Lee WJ, Lee JS.1999. Border malaria characters of reemerging vluarmalaria in the Republic of Korea. Korean J Parasitol37 :7 l - 76 .

Kim HC 2000. January 2000. Light trap mosquito sur-veillance on U.S. Army installations in the Republic ofKorea, May-October 1999. Available from Departmentof the Army, 5th Medical Detachment, 168th MedicalBattalion, 18th Medical Command, APO AP96205-0020. 39 p.

Kim HC, Lee KW Klein TA, Strickman DA. 1999. Sea-sonal prevalence of mosquitoes collected from lighttraps in Korea (1995-1996). Korean J Entomol29:181-' t 87 .

Kim HC, Lee KW Robert LL, Sardelis MR, Chase FE.1995. Seasonal prevalence of mosquitoes collectedfrom light traps in Korea (1991-1992). Korean J En-tomol 25:225-234.

Kim HC, Strickman DA, Lee KW. 2000. Seasonal prev-alence and feeding activity of Anopheles sinensis (Dip-tera: Culicidae) in the northwestern part of KyonggiProvince, Republic of Korea. Korean J Entomol 3O:.193-197.

Kline DL. 1994. Olfactory attractants for mosquito sur-veiflance and control: 1-octen-3-ol. J Am Mosq ControlAssoc lO:28O-287.

Kline DL. 1999. Comparison of two American Biophysicsmosquito traps: the professional and a new counterflowgeometry trap. J Am Mosq Control Assoc 15:'276-282.

Lee KW 1998. A revision of the illustrated taxonomic keysto genera and species of female mosquitoes of Korea(Diptera: Culicidae) Available from Department of theArmy, 5th Medical Detachment, 168th Medical Battalion,18th Medical Command, APO AP96205-0020. 40 p.

Lee KW Gupta RK, Wildie JR. 1984. Collection of adultand larval mosquitoes in U.S. Army compounds in theRepublic of Korea during 1979-1983. Korean J Par-asitol 22:102-108.

Lee JS, Kho WG, Lee HW, Seo M, Lee WJ. 1998- Currentstatus of vivax malaria among civilians in Korea. Ka-rean J Parasitol 36:241-248.

Lee KW Shin HK, Yoon HS, Tonn RJ, SeIf LS, Cho YS,Ahn SK (World Health Organization). 1971. Japaneseencephalitis vector studies in Korea: light trap surveysWHO/VBCi71.324. 10 p.

Lee SK, Ree HI. 1991. Studies on mosquito populationdynamics in Cholla-bugdo, Korea (1985-1990). I. Sea-

Seprpvnen 2001 Moseurro TRAe EvALUATToNs 205

sonal and annual fluctuations in population size. KoreanJ Entomol 21:141-155.

Liu C, Qian H, Gu Z, Pan J, Zheng X. 1986. Quantirativestudy on the role of Anopheles lesteri anthropophapiusin malaria transmission. J Parasitol Parasitic Dis 4:16t-164.

Mboera LE, Takken W, Sambu EZ..2OOO. The responseof Culex quinquefasciatrzs (Diptera: Culicidae) to trapsbaited with carbon dioxide, 1-octen-3-ol, acetone. bu-tyric acid and human foot odour in Tanzania. Bult En-tomol Res 90:155-159

Paik YH, Ree HI, Shim JC. 1988. Malaria inKorea. JnnJ Exp Med 58:55-66.

Ree HI. 2000. Unstable vivax malaia in Korea. Korean JParasi to l 38:1 l9-138.

Ree HI, Chen YH, Chow CY. 1969. Methods of samplingpopulations of the Japanese encephalitis vector mos-quitoes-a preliminary report. Med J Malaya 23:293-295.

Ree HI, Self LS, Hong HK, Lee KW. 1973. Mosquito lighttrap surveys in Korea, 1969-1971. SE Asian J TropMed Public Health 4:328-386.

Reisen WK, Boyce K, Cummings RC, Delgado O, Gu-tierJ�ez A, Meyer RB Scott TW. 1999. Comparative ef-fectiveness of three adult mosquito sampling methodsin habitats representative of four different biomes ofCalifornia. J Am Mosq Contol Assoc 15:24-37.

Reisen WK, Meyer RP, Cummings RII Delgado O. 2000.Effects of trap design and CO, presentation on the mea-surement of adult mosquito abundance using Centersfor Disease Control-style miniature light traps. J AmMosq Control Assoc 16:13-18.

Sasa M, Sabin AB. 1950. Ecological studies on the mos-quitoes of Okayama in relation to the epidemiology ofJapanese B encephalitis. Am J Hyg 5l:21-35.

SAS Institute. 1995. SAS/STAT user's manual, version6.03. Cary, NC: SAS Instirute.

Self LS, Shin HK, Kim KH, Lee KW, Chow CY, HongHK. 1973. Ecological studies on Culex *itaeniorhyn-chus as a vector of Japanese encephalitis. BulI WHO49:.41-47.

Service MW. 1993. Mosquito field ecology, field samplingmethods London: Chapman and Hall.

Shim JC, Shin EH, Yang DS, Lee WK. 1997. Seasonalprevelance and feeding time of mosquitoes (Diptera:Culicidae) at outbreak regions of domestic malaria (P.vivax) in Korea. Korean J Entomol 27:267-274.

Shim JC, Yoon YH, Kim CL, Lee WJ, Shin EH, Cho YB.1990. Population densities of the vector of Japaneseencephalitis Culex tritaeniorhynchus in Korea. NIH RptKorea 27:165-172.

Slaff M, Crans WJ, McCuiston LJ. 1983. A comparisonof three mosquito sampling techniques in northwesternNew Jersey. Mosq News 43:287--290.

Strickman D, Miller ME, Kim HC, Lee KW. 2000. Mos-quito surveillance in the Demilitarized Zone, Republicof Korea, during an outbreak of Plasmodium vivix ma-laria in 1996 and 1997. J Am Mosq Control Assoc 16:100 -1 13 .

Vaidyanathan R, Edman JD. 1997. Sampling methods forpotential epidemic vectors of eastern equine encepha-Iomyelitis virus in Massachusetts. J Am Mosq ConftolAssoc 13:342-347.

Yang KH, Yu HS. 1992. Seasonal abundance and breedinghabitats of mosquitoes (Culicidae; Diptera) from Su-won, Korea. Entomol Res Bull (Korea) 18:15-24.

Yi SC, Zorka TJ, Edgecomb RS. 1988. Comparative eval-uation of trap index limit to control population. KoreanJ Entomol 18:197-2O3.