volume 23, issue 2, december 1998

Journal

of

Vector EcologyJournal of the Society for Vector Ecology

Volume 23, No. 2 December 1998 4/

JVE 23( 2): 107- 201 ISSN 1081- 1710

Printed on Acid-Free Paper

Journal of Vector Ecology

Volume 23- Number 2- December 1998

Published by the Society for Vector Ecology

Marc J. Klowden, EditorDivision of Entomology

University of IdahoMoscow, ID 83844- 2339

Phone: ( 208) 885- 7546

Fax: ( 208) 885- 7760

E-mail: mklowden @uidaho.edu

EDITORIAL BOARD

M. S. Mulla, Chair( 2000) C. I. Dahl( 2001) J. W. Beehler( 2001)

University of California University of Uppsala Northwest Mosquito and Vector

Riverside, CA, USA Uppsala, Sweden Control District, Corona, CA,

USA

R. C. Axtell( 1998) F. R. Holbrook( 1998)

North Carolina State University USDA-ARS M. W. Service( 1998)

Raleigh, NC, USA Laramie, WY, USA Liverpool School of Tropical

Medicine

N. Becker( 2001) R. S. Lane( 1999) Liverpool, England

KABS University of CaliforniaWaldsee, Germany Berkeley, CA, USA E. D. Walker( 1999)

Michigan State UniversityD. R. Barnard( 1998) G. G. Marten( 1998) East Lansing, MI, USAUSDA-ARS Kwansei Gakuin UniversityGainesville, FL, USA Sanda, Hyogo, Japan S. K. Wikel( 1998)

Oklahoma State UniversityH. Briegel( 1999) L. S. Mian( 2000) Stillwater, OK, USA

University of Zurich County Vector ControlZurich, Switzerland San Bernardino, CA, USA S. C. Williams( 1998)

San Francisco State UniversityW. J. Crans( 1998) R. S. Nasci( 200(3) San Francisco, CA, USA

Rutgers University CDC-DVBID

New Brunswick, NJ, USA Ft. Collins, CO, USA

The Journal of Vector Ecology is published biannually in June and December. Authors agree to transfer thecopyright for their article to the publisher when the article is accepted for publication. Authorization to photocopyarticles is granted by the Society for Vector Ecology provided the indicated fee is sent to the Copyright ClearanceCenter, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. Individuals may make single copies of articleswithout charge. Communications relating to editorial matters and manuscripts should be addressed to the Editor.Communications relating to galley proofs, reprints, subscriptions, SOVE membership, change of address, andother matters should be addressed to the Business Office.

Publications and Business Office: Society for Vector Ecology, 1966 Compton Avenue, Corona, CA 91719 USA.

Subscription Rates: Membership, including the Journal ofVector Ecology,$ 50.00, Student membership$ 25.00,Institutional subscription$ 50.00, Sustaining membership $ 100.00

SOCIETY FOR VECTOR ECOLOGY

1998 BOARD OF DIRECTORS

OFFICERS

President Rex E. ThomasPresident-Elect David A. DameVice-President John D. EdmanPast- President Michael W. Service

Secretary- Treasurer Major S. Dhillon

REGIONAL DIRECTORS

Southwestern Minoo B. MadonNorthwestern Sammie DicksonNorth Central Thomas R. WilmotSouth Central Cluff E. HoplaNortheastern Wayne J. Crans

Southeastern Jonathan F. DayEuropean Isabelle ThierySOVE Journal Editor Marc J. Klowden

JOURNAL OF VECTOR ECOLOGY

Volume 23 DECEMBER, 1998 Number 2

CONTENTS

Board of Directors iiGuidelines for Contributors ivWilliam Donald Murray Memoriam v

Howard R. Greenfield Memoriam vii

Submitted Papers

Olfactory Responses of Female Culex quinquefasciatus Say ( Diptera: Culicidae) in a Dual-Choice Olfactometer L. E. G. Mboera, B. G. J. Knols, W. Takken, and P. W. T. Huisman 107

Antifeedancy of Neem Products Containing Azadirachtin against Culex tarsalis and Culexquinquefasciatus( Diptera: Culicidae) Tianyun Su and Mir S. Mulla 114

A New Enzyme Immunoassay to Detect Antibodies to Arboviruses in the Blood of Wild BirdsR. E. Chiles and W.K. Reisen 123

Surveillance Studies of Orthopodomyia signifera with Comparisons to Aedes sierrensis

David L. Woodward, Arthur E. Colwell, and Norman L. Anderson 136

Effectiveness of Control Measures against Mosquitoes at a Constructed Wetland in SouthernCalifornia William

E. Walton, Parker D. Workman, Louie A. Randall, Joshua A. Jiannino, and Yvonne A. Offill 149

Effect of Two Rice Culture Methods on the Seasonal Occurrence of Mosquito Larvae and Other

Aquatic Animals in Rice Fields of Southwestern Korea Dong- Kyu Lee 161

Evaluation of Attractant-Baited Traps/Targets for Mosquito Management on Key Island,Florida, USA Daniel L. Kline and Gene F. Lemire 171

Olfactory Responses and Field Attraction of Mosquitoes to Volatiles from Limburger Cheese andHuman Foot Odor Daniel L. Kline 186

Egg Maturation in Neotropical Malaria Vectors: One Blood Meal is Usually EnoughL. P. Lounibos, D. Couto Lima, R. Lourenco-de- Oliveira, R. L. Escher, and N. Nishimura 195

JOURNAL OF VECTOR ECOLOGY

Guidelines for Contributors

The Journal of Vector Ecology is an international journal published by the Society forVector Ecology. It is concerned with all aspects of the biology, ecology, and control of arthropodvectors and the interrelationships between the vectors and the disease agents they transmit. Thejournal publishes original research articles and research notes, as well as comprehensive reviews

of vector biology based on presentations at Society meetings. All papers are reviewed by at leasttwo referees who are qualified scientists and who recommend their suitability for publication.Acceptance of manuscripts is based on their scientific merit and is the final decision of the editor,

but these decisions may be appealed to the editorial board.Manuscripts intended for publication should be sent to Dr. Marc J. Klowden, Editor,

Division of Entomology, University of Idaho, Moscow, Idaho 83844- 2339, U.S. A. Manuscriptsmust be double spaced on a single side of bond paper with 25 mm margins. An original and two

clear copies are required. Draft mode dot matrix type should not be used. Submission of text on

a 3- 1/ 2" computer diskette formatted in MS- DOS is encouraged. Microsoft Word, Word Perfect,

or Wordstar formats are acceptable, as well as unformatted text files. Please indicate the type of

format on the diskette label. Papers must be organized under the following headings, each on aseparate page, in order: Title page, abstract, text, acknowledgments ( if appropriate), references

cited, tables, figure legends, and figures. The title page should contain the names of all authors,

their affiliations and the identification and address of the corresponding author. It should alsoinclude a keyword index containing no more than five words that best describe the paper. Pagesshould be numbered consecutively starting with the title page. References should conform to thestyle in recent volumes. Illustrations that are submitted must be ofhigh quality and remain legibleafter reduction.

Page charges, which partially defray the cost of publication, are $ 35 per printed page.

SOVE members who are unable to pay page charges may apply for a limited number of waivers.Reprint charges are shown in the table below.

Pages 1- 4 5- 8 9- 12 13- 16 17- 20

50 copies 70.00 115. 00 160.00 205. 00 255. 00

or less

Each add' 1 30.00 48.00 66.00 84.00 102.00

50 copies

Same order

Communications relating to editorial matters and manuscripts should be addressed to theEditor. Communications concerning galley proofs reprints, subscriptions, SOVEmembership, and change of address should be addressed to the Business Office.

PUBLICATIONS AND BUSINESS OFFICE

Society for Vector Ecology1966 Compton Avenue

Corona, CA 91719

Phone: ( 909) 340-9792; Fax: ( 909) 340-2515; E-Mail: nwmvcd @pe.net

NUMBER 2 DECEMBER, 1998 VOLUME 23

IN MEMORIAM

WILLIAM DONALD MURRAY

1913- 1998

r4p, , .,,,,;,, :". ,, ,,, , ,,

a. a

l .' i1F It

ji * {t i i it

1

libl iqi 4 tad

r+ ra ss#i

hF

a s .. o s c s• :

rtk:

ail +

11. t .o s • rsa ar

ai i, icy

iti. R 4}

iif..} .S ; t ill: i#

1.

fig ` it„it*. 0r. z.

Dr. W. Donald Murray, a dedicated leader in the Minnesota. While there he was awarded a masters

early development of mosquito control in California, degreeandlateraPh.D. Hisgraduateworkwasprimarilypassed away on March 18, 1998, at his home in Visalia, in the research of wasps. He married his wife Frances

California.. Dr. Murray was born on August 9, 1913, in during the time he was in graduate school.Shayside, Ohio, and was known by friends and colleagues This was the time of the depression years and full-

asDon. His early interest in natural history was spawned time employment was difficult to obtain. Don spent

by his experience in summer YMCA camp sessions near summers evaluating grasshopperpopulations and controlColumbus, Ohio. There he met Dr. Donald Borror and operations in the midwestern prairie states. This is

Dr. Delong, both professors at entomology from Ohio where he first developed the pants leg count that wasStation University, who served as his mentors and later used to count Aedes mosquitoes in California.influenced him to attend Ohio State and major in After graduate school he taught botany and zoologyentomology, where he received his BS degree. Don then at Eveleth Junior College in Eveleth, Minnesota. Hereceived a scholarship to attend the University of then went on to teach biology at Bemidji State Teachers

v

VOLUME 23 DECEMBER, 1998 NUMBER 2

College in Bemidji, Minnesota from 1942- 1943. developed methods to eliminate larval sources ofMuscaIn November 1943, Don received a commission in domestica using a rotary scraper to spread and dry the

the U.S. Navy as a Lieutenant and sent to the South feed-manure mixture behind the feed bunkers on largePacific on the island of Efate to head- up a malaria dairy ranches. To control the green blow fly, Phoeniciacontrol unit. In the early phases of the operation, diesel sp. in the urban areas, John Combs( who later becameoil was used as a larvicide; DDT powder mixed with District Manager and ExecutiveDirectoroftheCaliforniadiesel oil gave excellent control. His idea of mosquito Mosquito and Vector Control Association) was hired to

source reduction ( Don was a strong early advocate of direct the fly control program. John' s effort was verythis approach in California) was started on Efate, where effective in developing a community fly controlcoconut hulas and debris cans were disposed of in a education program and working with city officials tolandfill, and the axil points of the breadfruit trees were employ better methods of refuse storage and collection.drained to eliminate standing water that served as Don was Past-President( 1956), elected Honorarysources of the malaria vector. Member, and served as Secretary- Treasurer for 16 years

Upon his return to the States he talked to Professor for the California Mosquito and Vector Control

Herms at U.C. Berkeley who suggested that he contact Association. He was awarded the" Medal ofHonor" for

Dick Peters of the California Department of Public his service as Treasurer ( 8 years) by the AmericanHealth where they were establishing a new program unit Mosquito Control Association. He was a foundingin vector control. Don worked for the newly formed member of the Society for Vector Ecology and a long-Bureau of Vector Control during 1946. In the spring of time member of the Entomological Society of America.1947, there was a job opening for Manager of the Delta He continued to be active as an officer and consultant toMosquito Abatement District. He applied and was hired the U.S. Navy Medical Service Corps., obtaining thefor the position in the spring of 1947. rank of Captain before his retirement from the

In his 31 years ( 1947- 1978) as Manager of the Commission Officers Corps.

District, Don developed many innovative approaches to Don was very active in his community. He servedmosquito and fly control. He originated the concept of on the Tulare County Grand Jury, and was a member ofthe section survey procedure that involved the mapping the Visalia Racquet Club( he was still playing tennis inof each square mile in each control zone, delineating the his 80s). He was active in the Visalia Ys Men Club, andfield-crop boundaries, type of crop, irrigation and served as its President for several years. He was listeddrainage patterns, and plotting areas of standing water in American Men and Women of Science-Medical andon a frequent schedule. This allowed the zone operator Health Sciences.

to become more effective in locating the mosquito He is survived by his wife, Frances Murray ofsource within the proper time sequence. he was an early Visalia, son David Murray of Arroyo Grande, daughteradopter of an effective source reduction program that Lois Feleay ofVisalia, brothers Allen and Robert Murrayemployed personnel and equipment to establish drainage of Columbus, Ohio, and grandchildren Kevin Murray,channels, return drain water systems, and dairy drain Aaron Feleay, and Shara Hendren.recycle disposal facilities.

In response to community pressure for a fly control Earl W. Mortenson, Assistant Chief, Retiredprogram during the early 1960s, the District, along with Vector Surveillance and Control Branchpersonnel from the State Bureau of Vector Control, California Department of Health Services

vi

NUMBER 2 DECEMBER, 1998 VOLUME 23

IN MEMORIAM

HOWARD R. GREENFIELD

1918- 1998

its,,,

s

dim \

k

s EItt

4It was a sad time to hear that our good friend and town of Prescott. He later moved back to California and

colleague, Howard Greenfield, passed away on August lived in Los Altos. Howard attended San Jose State

31, 1998, after a year- long struggle with cancer. I first College for two years before joining the U.S. Navy.had an opportunity to meet Howard in 1950, when he Most of his Navy experience was in the South Pacificwas entomologist for the Merced Mosquito Abatement where he participated in the Leyte, Guadalcanal, and

District. Deed Thurman and I had been working on other campaigns, where his unit was awarded five

collecting soil samples for Aedes eggs from irrigated combat battle stars. After his five year Navy service, hepastures. Howard was helping us to obtain samples returned to San Jose State College to complete his BS

from pastures in Merced County that were heavily degree.

infested with Aedes nigromaculus. Howard' s interest in biology was soon channeledHoward was born in 1918 in Santa Cruz, California. into courses in entomology. Dr. Carl Duncan, Professor

He spent a portion of his youth in Arizona where his of Entomology and Dr. Ed Ross of the Californiafather and grandparents operated a gold mine near the Academy of Science served as his mentors and in the

vii

VOLUME 23 DECEMBER, 1998 NUMBER 2

senior year, Howard was awarded the Outstanding Control Association Meritorious Service Award in 1973.Science Student of the Year. He was very active and made major contribution to the

His first professional position was with the Bureau California Mosquito and Vector Control Association,of Vector Control, California Department of Public serving as President in 1957 and was on key committees;health, designated to work with Dr. Basil Markos in Ways and Means, Legislative, and Bylaws. He was one1948 to evaluate the effectiveness of DDT as a of the original members of the 1948 Entomologypretreatment application on irrigated pastures in Committee and was elected Honorary Member in 1983.Merced County. Ed Smith, Manager of the Merced Howard retired as manager of the North Salinas ValleyMosquito Abatement District hired Howard as the Mosquito Abatement District in 1982. Howard was

District Entomologist. In 1951, Howard was appointed active in his community, serving as a member of thethe Manager-Entomologist of the newly formed North School Board, Past- President of the Salinas CountySalinas Valley Mosquito Abatement District. The Club( a very good golfer), and member of many otherlegendary Harold Gray, Manager- Engineer of the community organizations.Alameda Mosquito Abatement District was another His wife, Betty Greenfield, son Brice Greenfield ofmentor to Howard in providing consultation in Pensacola, Florida, daughter Diane Lott, Rio Vista,designing the source reduction program that was grandchildren Nicole Greenfield, Samantha Greenfield,required for this new district to help solve the drainage and Jennifer Lott survive him.

problems inherent to the noi. h part ofMonterey County.Howard was a Charter Member of the Society for Earl W. Mortenson, Assistant Chief, Retired

Vector Ecology and served as the fourth President of the Vector Surveillance and Control Branch

Society in 1971. He received the American Mosquito California Department of Health Services

viii

Journal of Vector Ecology 23( 2): 107- 113

Olfactory Responses of Female Culex quinquefasciatus SayDiptera: Culicidae) in a Dual-Choice Olfactometer

L. E. G. Mboera1"2, B. G. J. Kno1s13, W. Takken', and P. W. T. Huisman'

Laboratory ofEntomology, Wageningen Agricultural

University, Wageningen, THE NETHERLANDS.2 National Institute for Medical Research,

Ubwari Field Station, Muheza, TANZANIA.

3International Centre of Insect Physiologyand Ecology, Nairobi, KENYA.

Received 24 February 1998; Accepted 13 July 1998

ABSTRACT: Olfactory responses of individual female ( n= 1010) Culex quinquefasciatus Say ( Diptera:Culicidae) to various odor stimuli were studied in a dual- choice olfactometer. Responses( i.e., the number of

mosquitoes entering either of both olfactometer ports) were studied towards clean conditioned air( control),human foot skin emanations( collected on worn stockings), carbon dioxide( 4.5% in clean air), moistened air,

and various combinations thereof. Skin emanations were significantly more attractive( x2= 23.0, p< 0.001)than clean stockings( control). The mosquito was also significantly more attracted( x2= 7. 7, p< 0.01) to skinemanations than to a clean stocking to which water( an equivalent of that absorbed by a worn stocking) wasadded. A moistened( 1 g H2O) clean stocking, however, was slightly more attractive than a dry stocking(

x2=

6, p< 0.025). Carbon dioxide( 4.5%) did not elicit higher responses than clean air, and no synergistic effect

was observed in combination with skin emanations. With the aim of developing an odor-baited trap, ourresults indicate that Cx. quinquefasciatus responds well to human body odors which can be collected onpolyamide materials.

Keyword Index: Culex quinquefasciatus, host- seeking behavior, human skin emanations, carbon dioxide,olfactometer.

INTRODUCTION 1996) showed a random distribution of bites on a naked

motionless volunteer.

Culex quinquefasciatus Say( Diptera: Culicidae) is Host-seeking mosquitoes use visual, physical, andthe most important vector of bancroftian filariasis in the olfactory cues produced by their host to orient themselvesworld. The mosquito is widely distributed in tropical and to that host( Takken 1991). Of the available stimuli, host

subtropical regions of the world where it has established odor is considered to be the most important, especially foritself in towns and areas where human habitation has led nocturnally active species that seek their host when visualto the creation offavorable breeding sites, mainly habitats cues are at a minimum or even nonexistent. Several

containing highly polluted water rich in organic matter studies have examined organochemical compounds for

that larvae use for nourishment( Subra 1981). their role in mosquito host location. Of these, carbon

Host-seeking and biting behavior of Cx. quinque- dioxide is generally considered to be an importantfasciatus constitute important aspects of bancroftian mosquito attractant( reviews by Clements 1963, Gilliesfilariasis epidemiology. The mosquito is markedly 1980, Mboera and Takken 1997), although its specific

endophagic, anthropophilic, and nocturnal( White 1971, role in the host-seeking behavior of different mosquitoBeier et al. 1990), which are behavioral traits favoring the species is poorly understood. In addition to carbon dioxide,transmission of Wuchereria bancrofti microfilariae from other compounds such as L(+)-lactic acid( Acree et al.

the peripheral blood system to the vector. Self et al. 1968), 1- octen- 3- ol( Takken and Kline 1989), butanone,

1969) reported a preference for biting the lower limbs of phenols ( Kline et al. 1990), and carboxylic fatty acidshumans but recent experiments by De Jong and Knols ( Knols et al. 1997) have been shown to attract mosquitoes.

108 Journal of Vector Ecology December, 1998

The majority of these studies, however, focused on the Pumped air ( 1400 ml/min) was cleaned by activatedyellow fever mosquito Aedes aegypti( Acree et al. 1968, charcoal, humidified, and regulated by volume beforeCarlson et al. 1973) and very little is known for other entering the olfactometer. Wind speed, temperature, andspecies of medical and veterinary importance. The odor- relative humidity of the air passing the tunnel ports( seemediated host-seeking behavior of Cx. quinquefasciatus Fig. 1) were measured before and after each test, andhas received remarkably little attention despite its medical averaged 20±4 cm/ sec., 26.9± 0.3° C, and 93.9± 0.4% RH,importance. In this preliminary study we sought to respectively.investigate the response of Cx. quinquefasciatus toward The Cx. quinquefasciatus strain used originated fromhuman foot skin emanations, carbon dioxide, and moisture Colombo, Sri Lanka( courtesy Prof.C. F. Curtis) and hasas a first st,;p in the development ofan effective odor-bait been maintained under laboratory conditions for morefor this filariasis vector. than ten years. Mosquitoes were kept in a climate-

controlled room at 27± 1° C, 80±5% RH and a light regimeMATERIALS AND METHODS ofLD 12: 12. Adult mosquitoes were kept in 30 cm3 cages

and offered a 6% glucose solution. Females were offered

Dual-Choice Olfactometer, Mosquitoes, and Experi- blood from a human arm twice weekly for egg production.mental Procedures Wet filter paper on glass petri dishes was provided for

The dual-choice olfactometer( Fig. 1) was made of oviposition. LarvaewererearedinplastictrayscontainingPerspex and was placed in a climate-controlled room at tap water and fed Tetramin®fish food. Females used in27± 1° C and 80± 5% RH with a light intensity of 110 lux. the experiments were four to eight days old and had not

E- 8 -›<- 8

34

92.5 g

OTh3 , 4 rmir

k

e ; ; e

16J

c

a

Figure 1. The dual-choice olfactometer setup( dimensions in cm). Room air was pumped( a) at 1400 ml/min, cleanedby passing activated charcoal( b), moistened by passing distilled water( c), passed a flow meter( e) for eachport, and entered the olfactometer( f).Baits( worn/clean stockings,( g) were positioned in compartments andair that passed over them entered the tunnel through the ports( h). Mosquitoes were released from glass vials

j) at the downwind end of the tunnel, and an imaginary flight path is shown( k). When needed, carbon dioxidewas pumped into the circuit from a pressurized gas cylinder at( d).

December, 1998 Journal of Vector Ecology 109

received a blood meal. They were deprived of sugar water the test port before the experiment started.18 h before testing. In a last series of experiments we examined the

Experiments were carried out during the last six possibility of synergism between foot odors and carbonhours of the dark period. Test mosquitoes were randomly dioxide. Carbon dioxide ( 4.5% at 300 ml/ min) passed

selected from a colony cage and transferred individually over foot skin emanations on polyamide stocking andinto glass vials( 5 x 1. 5 cm) using a suction tube, which was tested against either foot skin emanations on

were then sealed with a water-moistened plug of cotton polyamide stocking or carbon dioxide ( at a similarwool. Individuals were released at the downwind end of concentration).

the olfactometer and the response( i.e., entering of eitherof the two olfactometer ports within the 3- min. RESULTS

experimental period) recorded. Surgical gloves were

worn at all times in order to avoid contamination. All Figure 2A shows that when clean conditioned air

parts of the olfactometer were cleaned with hexane and was pumped into both ports of the tunnel, no significantethanol( 96%) between test series. preference( x2= 0.9, p> 0.05) for either port was observed.

Similar results were obtained when both ports were

Experimental Odor Stimuli baited with clean stockings( x2= 0. 8, p> 0.05). MosquitoesBefore running tests in which odor stimuli were

responded significantly more to skin emanations onadded to a clean air stream, a series of routine blank tests polyamide stockings than to clean stockings ( x2= 23. 0,clean conditioned air from both tunnel ports) was run. p< 0.001). Of those mosquitoes entering the port with the

Similar tests were also conducted with clean polyamide worn stocking, 87.7% landed and probed on it as comparedstockings or carbon dioxide released from both ports in to 12. 3% of those entering the port with the clean stocking.order to test the symmetry of the system. Thereafter, to Worn stockings were also more attractive than moistenedcounteract the effect ofany biased directional response of clean stockings (

x2= 7. 7, p< 0.01), confirming theunknown origin, the positions of the treatment and control attractiveness of body odors other than water.were alternated between ports after every 10 mosquitoes. Nevertheless, moistened stockings were significantly

Skin emanations were collected from a human foot more attractive than dry ones( x2= 6, p< 0.025). The totalon polyamide stockings by wearing them for 4- 5 days percentage of mosquitoes responding to either or both testbetween 06. 00 hr and 18. 00 hr at which time they were ports was significantly higher( p< 0.01) in those treatmentskept in a tightly stoppered bottle until being used for an in which foot odors were pumped in the tunnel than in theexperiment the following morning. Throughout the treatment pumping clean air from both test ports. However,experiments the same volunteer ( LEGM) provided the no difference was observed with the treatments usingskin emanations. The rolled up stocking was placed on a clean stockings in both ports or those that had water addedglass petri dish 17 cm from the air entrance in the tunnel to clean stockings.

ports, and clean air was pumped over it. The control odor When the response of Cx. quinquefasciatus to airconsisted of a clean stocking. The number of mosquitoes with added carbon dioxide( 4.5%) was compared to that

that responded to and probed the polyamide stocking in of atmospheric air( c 0.03%, Fig. 2B), the percentages ofeach treatment was recorded. mosquitoes entering the ports did not differ between the

A clean polyamide stocking onto which water equal treatments( x2= 0. 9, p> 0.05). These results indicate thatto the increase in weight( c 1. 0 g) of a stocking worn for carbon dioxide at human breath concentration does notfour days was pipetted, was tested against a worn( 4 days) attract this strain of Cx. quinquefasciatus. When bothpolyamide stocking in order to examine the effect of ports oftheolfactometerwerebaitedwithcarbondioxide,water on mosquito responses. In another series of the mosquitoes did not show any preference ( x2= 0. 1,experiments, clean stockings moistened with 1 g of water p> 0.05) for either of the ports showing maintenance ofwere tested against clean dry polyamide stockings.

symmetry of the olfactometer.Carbon dioxide( 4.5%, the concentration in human When carbon dioxide plus a worn polyamide

breath[ Gillies 19801) discharged from a cylinder at 300 stocking was tested versus carbon dioxide only,ml/ min was initially pumped into both ports of the tunnel. significantly more( x2= 24.6, p< 0. 001) mosquitoes wereIn the second experiment, carbon dioxide was compared attracted to the carbon dioxide/worn stocking combinationto clean air. The control was conditioned air alone at a than to carbon dioxide alone( Fig 2B). On comparing theflow rate equal to the combined rates of flow of air and response to carbon dioxide/worn stocking versus worncarbon dioxide in the treatment port. When running tests stocking alone, it was found that there was no significantwith carbon dioxide, the gas was allowed to flow freely difference between the treatments ( x2=

2.9, p> 0.05),through the conduit for five minutes to displace all air in indicating that the response was more to the worn stocking


Treatment 1 Treatment 2 N % resp.

A I I 160 67. 5 a

46 80.4 ab

141 83. 7 b

141 83. 0 b

1m1 H2O) 96 77. 1 ab

1 ml H20)iiiii4iiiiiiiiiiiiiiiiiiiii

I 1 1 1 1 1 1 1 1 1 1 1 1 150 30 10 10 30 50 70

choice

N % resp.

B 181 63. 0 a

54 70.4 ab

163 83. 4 b

68 73. 5 ab

w r >

1 1 1 1 1 1 1 1 1 1 1 1 1 150 30 10 10 30 50 70

choice

clean air VA CO2( 4.5%)

clean stocking ah, worn stocking+ CO2( 4. 5%)

worn stocking

Figure 2. Results of olfactometer tests with various odor stimuli. Treatments 1 and 2 were tested against each other,and bars indicate the percentage of responding mosquitoes flying into either test port(% choice). *: p< 0.025;

p< 0.01,***: p<0.001. N: total number tested.% resp.: mosquitoes that entered either of both test ports.

Percentages with no letters in common are significantly different at p< 0.005.

than to carbon dioxide, and that no synergistic effect of DISCUSSION

skin emanations with carbon dioxide was present. Theoverall percentage ofmosquitoes responding to any of the Our results show that Cx. quinquefasciatus is attractedtest ports in this second series ofexperiments was highest to body odors collected on polyamide stockings, that thisfor those treatments in which foot odor and/ or carbon effect is partially caused by the uptake of water by thedioxide was present in the tunnel. stocking, that carbon dioxide at a human breath


concentration is not more attractive than clean air, and with the present finding that moisture is a cofoundingthat the combination of body odors with carbon dioxide factor in attraction to skin odor. Moisture is naturallydoes not result in increased responses compared with present in the total emanations of the human body and itbody odor alone. Responses of Cx. quinquefasciatus to seems that Cx. quinquefasciatus uses water vapor as anskin emanations have also been observed in wind tunnel additional cue to respond to( human) hosts from a distance.studies by Omer( 1979), who found that air passing over As humans have a far greater eccrine sweat gland densitya hand was attractive to the mosquitoes. However, these than other mammals, moisture may be a behavioralstudies did not compensate for the presence ofwater in the stimulus for anthropophilic mosquitoes such as Cx.emanations that may well have influenced the results. As quinquefasciatus and An. gambiae sensu stricto( Takkenfor human feet, hands contain high densities(> 300/ cm2)

et al. 1997). Nevertheless, our studies have indicated thatofeccrinesweatglands, excretinglargeamountsofwater skin emanations elicit stronger attraction to Cx.Marples 1969). Other workers have also shown that skin quinquefasciatus than variations in the moisture gradient,

emanations are attractive to a variety of mosquito species and that these skin emanations, left as residues in theThompson and Brown 1955, Khan and Maibach 1966, stockings, are probably a source of human odor that plays

Price et al. 1979), and it has been found that odors an important role in the host-seeking behavior of thiscollected on worn clothing attract more female Anopheles species. It should be borne in mind, however, that due togambiae and Anopheles funestus to an unoccupied hut the size ofthe olfactometer, responses were only observedthan clean clothing( Haddow 1942). Recently, Geier et over a small distance near the source. It may well be thatal. ( 1996) collected human skin odors on an ethanol- over longer distances other kairomones also influencesoaked pad, and demonstrated that this extract was as this behavior. So far, we have failed to attract wild Cx.attractive to Ae. aegypti as a human hand and that the quinquefasciatus in Tanzania to unlighted CDC trapsextract maintained its activity for months after storage at baited with worn stockings( L. E. G. Mboera, unpublishedlow temperatures. Other hematophagous insects have data), although they are readily attracted to humanalso been shown to be attracted to worn materials and emanations released in-vivo from a tent or a bed neteven exhibited differential attractiveness to clothing worn ( Costantini et al. 1996, Dekker and Takken 1998).on specific parts of the body. Thompson ( 1976) for Carbon dioxide is considered to be a universalinstance, in Cameroon, attracted more blackflies belonging attractant for many hematophagous Diptera, but ourto the Simulium damnosum complex with worn trousers. results show that Cx. quinquefasciatus is poorly attractedFrom these and other studies ( e.g., De Jong and Knols to this compound at a concentration equivalent to that in1996) it can be concluded that mosquitoes may respond human breath. However, the response of the mosquito toto kairomones originating from some body regions, these different concentrations of carbon dioxide was not testedodors can be transferred onto an adsorbent cloth( cotton to determine optimum response levels for this species.wool pads, polyamide material, etc.) and still retain their Although Costantini et al. ( 1996) found that Cx.attractiveness. And due to a lasting effect of the quinquefasciatus in Burkina Faso responded at dosesattractiveness, these specific kairomones are not very above that normally released by one human, it is likelyvolatile. that at short range carbon dioxide plays a minor role in the

It is interesting to note that the addition of small attractiveness of a host. Other workers have made similaramounts of water to clean stockings provides a sufficient observations on different mosquito species( Crumb 1922,stimulus to attract Cx. quinquefasciatus. It was observed Healy and Copland 1995, De Jong and Knols 1995).that during all experiments the stockings increased in Moreover, field studies in Tanzania failed to attract An.weight by approximately 0.4 g, presumably as a result of gambiae s. l. to electric nets baited with human breath orwater uptake from the humid airstream. This may have carbon dioxide although they were readily caught on netsresulted in intermittent variations of the relative humidity baited with complete human odor from an occupiedof the air passing the tunnel ports, which may have been bedroom ( Knols et al. 1998). Recent work in Tanzaniaresponsible for the overall increase of responsiveness of showed that tents baited with carbon dioxide( pure, at 300the mosquitoes. Like An. gambiae( Takken et al. 1997), ml/min) attracted 9% as many An. gambiae s. l.( Mboerait thus appears that Cx. quinquefasciatus is also extremely et al. 1997) and 28% as many Cx. quinquefasciatus as asensitive to small changes in the relative humidity gradient. tent baited with a human( L. E. G. Mboera, unpublishedEarlier work showed that Cx. quinquefasciatus bit data). As carbon dioxide is not signifying the identity ofsignificantly more on the drier areas of the body than on an upwind host, it is likely that specialized mosquitoesareas with a high eccrine sweat gland density( De Jong like Cx. quinquefasciatus show less dependency on thisand Knols 1996), suggesting that drier skin is preferred compound in their host- seeking behavior than generalistabove a moister surface. These results are not in contrast species; and that, when used as a kairomone on its own,


itaccounts for onlyaminorpartoftheoverallattractiveness Carlson, D. A., N. Smith, H. K. Gouck, and D. R.

of a human( Mboera and Takken 1997). Godwin. 1973. Yellow fever mosquitoes:

No interaction between carbon dioxide and skin Compounds related to lactic acid that attract females.

emanations was observed. This is contrary to earlier J. Econ. Entomol. 66: 329- 331.

observations by other workers who reported the Clements, A. N. 1963. The Physiology of Mosquitoes.presence of a synergistic effect of carbon dioxide and Pergamon Press, Oxford, 393 pp.other compounds for various species of mosquitoes. Costantini, C., G. Gibson, N. Sagnon, A. della Torre, J.

For example, Vickery et al. ( 1966) demonstrated Brady, and M. Coluzzi. 1996. Mosquito response tosynergism of carbon dioxide and chicks by showing carbon dioxide in a West African Sudan savanna

that the combination of the two could increase the village. Med. Vet. Entomol. 10: 220-227.

catch of mosquitoes. Lactic acid has been shown to Crumb, S. E. 1922. A mosquito attractant. Science 126:

activate and elicit oriented flight behavior in mosquitoes 446-447.

only in the presence of carbon dioxide ( Smith et al. De Jong, R. and B. G. J. Knols. 1995. Olfactory responses1970, Price et al. 1979). Takken and Kline ( 1989) and of host- seeking Anopheles gambiae s. s. Diptera:Kline et al.( 1990, 1991) have shown that the presence Culicidae). Acta Trop. 59: 333- 335.of carbon dioxide in conjunction with other mammalian De Jong, R. and B. G. J. Knols. 1996. Selection of bitingodors ( namely 1- octen- 3- ol) resulted in synergism for sites by mosquitoes. Olfaction in mosquito-host

several mosquito species, in particularAe. taeniorhynchus. interactions ( G. Cardew, ed.). CIBA symposium

Since in our experiments the combination of carbon 200: 89- 99.

dioxide and skin emanations could not significantly Dekker, T. and W. Takken. 1998. Field studies on the

increase the number of mosquitoes responding compared responses of Anopheles arabiensis Patton and An.

to carbon dioxide only, it is likely that there is no quadriannulatus Theobald to carbon dioxide, a

synergistic effect for Cx. quinquefasciatus by the man and a calf. Med. Vet. Entomol. 12: 136- 140.

combination of the two. Geier, M., H. Sass, and J. Boeckh. 1996. A search for

The overall conclusion of this work is that the human components in human body odor that attract femalesskin produces compounds to which Cx. quinquefasciatus of Aedes aegypti. Olfaction in mosquito-host

is attracted over a short distance, and that this effect may interactions ( G. Cardew, ed.). CIBA symposium

be enhanced by slight variations in relative humidity. The 200: 132- 144.

recently described foot odor- Limburger cheese analogy Gillies, M. T. 1980. The role of carbon dioxide in host-

and the subsequent identification of attractive carboxylic finding by mosquitoes ( Diptera: Culicidae): a

acids ( Knols et al. 1997) may also apply to Cx. review. Bull. Ent. Res. 70: 525- 532.

quinquefasciatus and studies to determine the Haddow, A. J. 1942. The mosquito fauna and climate of

attractiveness of these kairomones are currently being native huts at Kisumu, Kenya. Bull. Ent. Res. 33:

undertaken in Tanzania. 91- 142.

Healy, T. P. and M. J. W. Copland. 1995. Activation ofAnopheles gambiae mosquitoes by carbon dioxide

Acknowledgments and human breath. Med.Vet. Entomol. 9: 331- 336.

Khan, A. A., H. I. Maibach, W. G. Strauss, and W. R.

We are grateful to Prof. C. F. Curtis for helpful Fenley. 1966. Quantitationofeffect ofseveral stimulicomments on an early draft of the manuscript. BGJK on landing and probing by Aedes aegypti. J. Econ.gratefully acknowledges financial support from the Niels Entomol. 59: 690-694.

Stensen Foundation, The Netherlands. Kline, D. L., W. Takken, J. R. Wood, and D. A. Carlson.

1990. Field studies on the potential of butanone,

REFERENCES CITED carbon dioxide, honey extract, 1- octen- 3- ol, L-lacticacid and phenols as attractants for mosquitoes. Med.

Acree, F. Jr., R. B. Turner, H. K. Gouck, M. Beroza, and Vet. Entomol. 4: 383- 391.

N. Smith. 1968. L-lactic acid: A mosquito attractant Kline, D. L., J. R. Wood, and J. A. Cornell. 1991.isolated from humans. Science 161: 1346- 1347. Interactive effects of 1- octen- 3- ol and carbon dioxide

Beier, J. C., W. 0. Odago, F. K. Onyango, C. M. Asiago, on mosquito( Diptera: Culicidae) surveillance andD. K. Koech, and C. R. Roberts. 1990. Relative control. J. Med. Entomol. 28: 254- 258.

abundance and blood feeding behaviour of Knols, B. G. J., J. J. A. Van Loon, A. Cork, R. D.

nocturnally active culicine mosquitoes in western Robinson, W. Adam, J. Meijerink, R. De Jong, andKenya. J. Am. Mosq. Contr. Assoc. 6: 207- 212. W. Takken. 1997. Behavioral and electrophysio-


logical responses of the female malaria mosquito H. Gilbert, M. S. Mayer, B. J. Smittle, and A.Anopheles gambiae ( Diptera: Culicidae) to Hofbauer. 1970. L-lactic acid as a factor in the

Limburger cheese volatiles. Bull. Ent.Res. 87: 151- attraction of Aedes aegypti to human hosts. Ann.

159. Ent. Soc. Am. 63: 760-770.

Knols, B. G. J., L. E. G. Mboera, and W. Takken. 1998. Subra, R. 1981. Biology and control of Culex pipiensElectric nets for studying odour-mediated host- quinquefasciatus Say, 1823 ( Diptera: Culicidae)seeking behaviour of mosquitoes ( Diptera: with special reference to Africa. Insect Sci. Applic.

Culicidae). Med. Vet. Entomol. 12: 116- 120. 1: 319- 338.

Marples, J. J. 1969. Life on the human skin. Sci. Am. 220: Takken, W. 1991. The role of olfaction in host-seeking108- 115. of mosquitoes: A Review. Insect Sci. Applic. 12:

Mboera, L. E. G., B. G. J. Knols, W. Takken, and A. della 287- 295.

Tone. 1997. The response of Anopheles gambiae Takken, W. and D. L. Kline. 1989. Carbon dioxide and

s. l. and An. funestus ( Diptera: Culicidae) to tents 1- octen- 3- ol as mosquito attractant. J. Am. Mosq.baited with human odour or carbon dioxide in south- Control Assoc. 5: 311- 316.

east Tanzania. Bull. Ent. Res. 87: 173- 178. Takken, W., B. G. J. Knols, and H. Otten. 1997.

Mboera, L. E. G. and W. Takken. 1997. Carbon dioxide Interactions between physical and olfactory cueschemotropism in mosquitoes ( Diptera: Culicidae) in the host- seeking behaviour of mosquitoes: theand its potential in vector surveillance and role ofrelative humidity. Ann. Trop. Med. Parasitol.management programmes. Rev. Med. Vet.Entomol. 91( Supp1. 1): S119- S120.

85: 355- 368. Thompson, B. H. 1976. Studies on the attraction of

Omer, S. M. 1979. Responses of females of Anopheles Simulium damnosum to its hosts. II. The nature of

arabiensis and Culex pipiensfatigans to air current, substances of the human skin responsible for

carbon dioxide and human hands in a flight tunnel. attractant olfactory stimuli. Tropenmed. Parasitol.Entomol. Exp. Applic. 26: 142- 151. 27: 83- 90.

Price, G. D., N. Smith, and D. A. Carlson. 1979. The Thompson, R. P. and A. W. A. Brown. 1955. The

attraction of female mosquitoes ( Anopheles attractiveness of human sweat to mosquitoes and

quadrimaculatus Say) to stored human emanations the role of carbon dioxide. Mosq. News 15: 80- 84.in conjunction with adjusted levels of relative Vickery, C. A. Jr., K. E. Meadows, and I. E. Baughman.humidity, temperature and carbon dioxide. J. Chem. 1966. Synergism of carbon dioxide and chick bait

Ecol. 5: 383- 395. for Culex nigripalpus. Mosq. News 26: 507- 508.Self, L. S., M. H. M. Abdulcader, and M. M. Tun. 1969. White, G. B. 1971. The present importance of domestic

Preferred biting sites of Culex pipiens fatigans on mosquitoes( Culex pipiens fatigans Wiedmann) in

adult Burmese males. Bull. WHO 40: 324- 327. East Africa and recent steps towards their control.

Smith, C. N., N. Smith, H. K. Gouck, D. E. Weidhaas, I. East Afr. Med. J. 48: 266- 274.


Antifeedancy of Neem Products Containing Azadirachtin againstCulex tarsalis and Culex quinquefasciatus ( Diptera: Culicidae) 1

Tianyun Su and Mir S. Mulla

Department of Entomology, University of California,Riverside, CA 92521- 0314, U.S.A.

Received 26 March 1998; Accepted 1 June 1998

ABSTRACT: In order to develop and utilize new botanical insecticides originating from neemAzadirachta indica A. Juss) in mosquito control programs as a potential larvicide, the antifeedant activity

of three formulations of neem against the larvae of Culex tarsalis Coquillett and Culex quinquefasciatus

Say was investigated under laboratory conditions. A significant antifeedancy was indicated at 5 ppm and10 ppm azadirachtin( AZ) for all formulations and both species. Within the test concentration range of AZ

1- 10 ppm), 5 ppm was the minimum effective concentration for antifeedancy in most cases. Somedifferences in larval susceptibility in terms of antifeedancy to the test formulations were noted between thetwo species. The Cx. tarsalis larvae were more susceptible than Cx. quinquefasciatus larvae to Azad EC

4. 5 at 1 ppm during the first 15- min exposure period and at 5 and 10 ppm during the whole test period. TheCx. quinquefasciatus larvae were more susceptible than Cx. tarsalis larvae to Azatin WP 4.5 at 1, 5, and

10 ppm during the first 15- min exposure period or to Neemix EC 4.5 at 1 ppm during the first 45- minexposure period. The formulation- related differences in antifeedant activity appeared when the concentrationincreased from 1 to 10 ppm. In Cx. tarsalis, the Azad EC 4. 5 and Neemix EC 4.5 were more effective than

Azatin WP 4.5 at 5 ppm during the first 15- min exposure period and at 10 ppm during the whole test period.In Cx. quinquefasciatus, the Azatin WP 4. 5 and Neemix EC 4. 5 were more effective than Azad EC 4. 5 at1 ppm during the first 30-min exposure period, and at 5 ppm during the first 60-min exposure period, andat 10 ppm during the whole test period.

Keyword Index: Neem, azadirachtin, Culex tarsalis, Culex quinquefasciatus, antifeedancy.

INTRODUCTION that about 2,000 plant species possess pest control

properties( Ahmed et al. 1984); and among these, 344Chemical control of insect pests and disease vectors species show some activities against mosquitoes

using synthetic insecticides has played an important ( Sukumar et al. 1991). In these plants producing bioactiverole in protecting the food sources and the health of materials, the neem tree, Azadirachta indica A. Juss,mankind during the past decades. However, concerns and closely related species A. excelsa Jack, A. siamensregarding safety and the environmental impact of Valeton, Melia azedirach L., and M. volkensii Giirke insynthetic pesticides has resulted in an intensified search the family Meliaceae have been well investigated andfor alternative strategies for pest control. In recent years, proven to possess considerable insecticidal properties.biopesticides having attributes of activity and efficacy, At least 35 components exhibiting some insecticidalspecificity and selectivity, low mammalian toxicity, activity have been identified from the neem tree ( Raoenvironmental acceptability, safety to nontarget and and Parmar 1984), among which is the triterpennoid,beneficial biota, and economic viability ( Mulla 1997) azadirachtin ( AZ). The AZ moiety has structuralhave been actively investigated in the fields of similarity to the molting hormone of insects. It is quitemicrobiology, phytochemistry, and entomology. Among unique that AZ can act and affect multiple systems inbiopesticides, the botanical insecticides constitute a insects and exhibit various modes ofaction. The activitiesdiverse group ofbioactive compounds. It has been noted of AZ against a variety of phytophagous insects and a

These studies were supported by University of California Special Funds for Mosquito Research.


few hematophagous insects have been studied. These for 1st, 2nd, 3rd, and 4th- instar larvae, respectively. Theactivities include antifeedancy, growth regulation, larval food consisted of powdered rabbit pellets

reproduction suppression, biological fitness alteration, ( Brookhurst®, Brookhurst Mill, Riverside, CA) and

and so forth. In mosquitoes, the crude preparations of Brewer' s yeast in the ratio of 3: 1. Water was added to the

neem or experimental AZ formulations primarily act as pans every other day to replenish loss due to evaporation.larvicides by growth regulation ( Mulla et al. 1997, Pupae were removed from the pans in a cup containingNaqvi et al. 1991, Rao et al. 1995, Zebitz 1986). The water and placed in screened cages ( 23 x 23 x 32 cm)

ovicidal activity ( Su and Mulla 1998a), oviposition where the adults emerged. Adults were provided

attractancy or repellency ( Su and Mulla 1998b), and continuously with 10% sucrose solution in ajar provided

host-seeking/blood- feeding repellency ( Sharma and with a cotton wick. On day 5 postemergence, the adultsAnsari 1994, Sharma et al. 1993 a, b) were also studied were deprived of sugar feeding for 12 h, then providedin a few species. with a restrained one-week-old chick overnight for

The larvae ofmost mosquito species are considered blood feeding( Animal Use Protocol No. A-S 9509052-to be omnivorous, which filter-feed on microorganisms 3, University of California, Riverside). The early 4th-and particulate materials floating on, suspended in, or instar larvae from both colonies were used in the feedingsedimented to the bottom of aquatic habitats. The mouth test.

brushes set up the water currents, which carry theparticles toward the mouth. The particles, both nutritive Test Materials and Preparation of Stock Pre-

and inert, with suitable size are indiscriminately ingested suspensions

Clements 1992). However, under experimental The test materials were three experimental

conditions, the ingestion ofthe broadly nutritive materials formulations of neem products, one wettable powder

such as yeast, wheat flour, dry blood, fishmeal, or algae ( Azatin WP 4. 5) and two emulsifiable concentrates

was faster than that of inert materials such as charcoal, ( Azad EC 4. 5 and Neemix EC 4. 5), all supplied bykaolin, brick dust, talc, chalk, or dye particles ( Aly Thermo-Trilogy Co.( Salt Lake City, UT, USA). Stock1985, Dadd 1968, 1970a,b; Dadd et al. 1982, Rashed suspensions of the formulated products were prepared

and Mulla 1989). The larval feeding rate varies among in distilled water at a concentration of 0. 1% ( w/v) or

species, instars, and the extent of starvation, and is 1, 000 ppm of AZ. These stock preparations were

affected by factors such as water temperature, employed in antifeedancy tests, where the neededphagostimulants, phagodeterrents, and types of the food aliquots were added to 200 ml of 2% ( WN) yeast

particulates. As to the effects of the neem products on suspension in 250 ml disposable paper cups for feedingthe feeding behavior in insects, antifeedancy has been ( see below). For each test, stock suspensions were

indicated in many species of phytophagous insects freshly prepared.Ascher 1993, Mordue and Blackwell 1993, Schmutterer

1988, 1990). To date, there are no reports on the effects Antifeedancy Testsof neem products on the feeding behavior of mosquito The antifeedancy tests were conducted in a holdinglarvae. In order to gain information on antifeedancy of room maintained at 28± 1° C, 35- 45% RH illuminated

neem products in mosquitoes, the current research was by two 40W fluorescent and one 15W incandescentundertaken to investigate the effects of experimental lamp. The feeding rate was determined according to theneem products on the feeding behavior of larvae of method of Dadd( 1968). Briefly, the larvae were allowedCulex tarsalis Coquillett and Culex quinquefasciatus to feed on contrasting color inert particulates such asSay in the laboratory. kaolin or charcoal for some time to fill the gut completely.

The inert material acted as color indicator distinguishable

MATERIALS AND METHODS visibly from the color of the test material. Then theglutted larvae were removed to the test material. The

Mosquito Colony Rearing newly ingested test material displaced rearward theThe test species, Cx. tarsalis and Cx. quinque- previously ingested inert material column, which was

fasciatus were maintained in two separate culture rooms excreted in the form of small pellets. Because the gut

under conditions of26± 1° C, 50-60% RH and a 14L: 10D peristalsis did not mix the gut content( Jones 1960) and

h photoperiod with 1 h dawn and dusk periods. To rear the sharp boundary between indicator particle mass andthe colony, 5 to 7 egg rafts of each species were placed the test material mass was maintained, the ingestion rate

in an enamel pan( 40 x 24 x 6 cm) containing 2, 500 ml was quantified by counting the segments filled ordistilled water. The larvae were fed dry powder food at displaced by test material within specified time intervals.the doses of 80, 160, 320, and 240 mg per pan every day If the boundary had reached the posterior margin of the


thorax, a value of 1 was assigned. Values of 1. 5, 2, 2.5, determined by unpaired, two tailed t-test.and 3- 6 were given when the boundary was at themidpoint or posterior margin of the 1st to 5th abdominal RESULTS

segments, respectively. If the indicator column wascompletely displaced but still remained as pellets in the Comparisons Among Different Concentrationshindgut, a value of6.5 was assigned. When no indicator The antifeedant activity of the neem formulationsmass was seen in the whole gut, a value of 7 was given. was tested at 0( control), 1, 5, and 10 ppm AZ at variousValues of0 and 0.5 were applied if no displacement had formulations against the two test species. Significantoccurred or if test material had only advanced half way antifeedancy was indicated at 5 and 10 ppm AZ for allthrough the thorax( Dadd 1968). formulations and both species. In the cases of Azad EC

In the current research, the indicator suspension 4.5 against Cx. tarsalis or Neemix EC 4.5 against Cx.was made by adding 2 g kaolin( Fisher Scientific, Fair tarsalis and Cx. quinquefasciatus, 10 ppm AZ was moreLawn, NJ) to 1, 000 ml distilled water( 0.2%, WN) and effective than 5 ppm AZ in terms of antifeedancy. Fiveshaking vigorously to insure the homogeneity of the ppm AZ was the minimum effective concentration forprepared suspension. The whitish, amorphous kaolin antifeedancy in most cases. The concentration 1 ppmparticles were suspended singly or aggregated in clusters AZ did not affect the feeding rate, except in Azad EC 4.5with the size of 2. 5- 37. 5 gm in water( Rodcharoen and and Neemix EC 4.5 against Cx. tarsalis( Fig. 1).Mulla 1995). For feeding, approximately 100 unstarvedearly 4th- instars from the rearing pans were transferred Comparisons Between Two Speciesto 200 ml 0.2% kaolin suspension in 250 ml disposable Some differences in susceptibility to neem productspaper cups and allowed to feed on this inert material between the two test species, regardless of formulationsovernight. The larvae engorged with kaolin were then and concentrations, were indicated. The feeding rates oftransferred to 200 ml 0. 2% yeast ( United States 4th- instar Cx. tarsalis and Cx. quinquefasciatus onBiochemical Corp., Cleveland, OH) suspension in 250 yeast suspension under neem- free condition wasml disposable paper cups, which were treated with essentially the same, with the exception that the formerdifferent neem formulations ofAzatin WP 4. 5, Azad EC fed slightly faster than the latter during the first 15- min4.5, or Neemix EC 4. 5 at the concentrations of 0( control), feeding period after being transferred to yeast suspension.1, 5, and 10 ppm. The yeast suspension was prepared The Cx. tarsalis larvae were more susceptible than Cx.previously by adding 2 g yeast powder to 1, 000 ml quin-quefasciatus larvae to Azad EC 4.5 at 1 ppmdistilled water and shaking vigorously. The light brown, during the first 15- min exposure period and at 5 and 10oval or spherical yeast particles were suspended singly ppm during the whole test period. The Cx. quinque-or aggregated in clusters( 5- 25 pm) in water( Rodcharoen fasciatus larvae were more susceptible than Cx. tarsalisand Mulla 1995). At a series of time intervals of 15, 30, larvae to Azatin WP 4.5 at 1, 5, and 10 ppm during the45, 60, and 90-min, 10 larvae were screened out with first 15- min of exposure period or to Neemix EC 4.5 atplastic screener and killed by putting them in a 5 ml 1 ppm during the first 45- min exposure period( Fig. 2).plastic dish containing 3 ml ofhot water at 75- 85° C. Thenumbers of the segments filled or displaced by yeast Comparisons Among Various Formulationsparticles were scored within 2 h under dissecting The formulation- related differences in antifeedancymicroscope( 8x), by noting the different color of kaolin activity occurred when the concentration increased fromand yeast mass in the gut. The experiment using three 1 to 10 ppm. In Cx. tarsalis, the Azad EC 4. 5 andformulations at four concentrations against two species Neemix EC 4. 5 were more effective than Azatin WP 4.5was repeated three times. The mean values of segment at 5 ppm during the first 15- min exposure period and atfilled at different intervals were determined. 10 ppm during the whole test period. In Cx.

quinquefasciatus, the Azatin WP 4.5 and Neemix ECData Analysis 4.5 were more effective than Azad EC 4.5 at 1 ppm

The parameter for determining feeding rate on during the first 30- min exposure period and at 5 ppmyeast in the aquatic medium treated with neem products during the first 60-min exposure period, and at 10 ppmwas mean numbers of segments filled or displaced by during the whole test period( Fig. 3).yeast particles at various time intervals. At each interval

of feeding process, the comparisons among various DISCUSSION

concentrations or formulations were conducted by 1-factor analysis of variance ( ANOVA Scheffe F test), The extracts from the seeds, leaves, and bark of thewhile the difference between two test species was neem tree, where the major pesticidal component is AZ,


III Control 0 1 ppm ® 5 ppm :. 10 ppm

Azatin WP4.5 vs. Cx. tarsalis Azatin WP4.5 vs. Cx. quinquefasciatus

as as a as7-

a7-

a ab6- a

sa

b

b6- bb

5- a

ia b

5- aa b bb

b 4a

ab 4'b bf 4

b2- 05 2

b a

Azad EC 4.5 vs. Cx. tarsalis Azad EC 4.5 vs. Cx. quinquefasciatus

a as asa a

a aaab aaaa

7- a aa 7- ab

ab b

6- a a

Eaabb

b

t„,:„„

K. 2_ ::: Ii:

i: b :: iji:''..:b ......:•°

O -

L ....,... 4 i. iiii h+ ft;iii

I I f I I I I I I I

Neemix EC 4.5 vs. Cx. tarsalis Neemix EC 4.5 vs. Cx. quinquefasciatus

7-a as a s as

7- aas a s

aa

b aa

6- 6- b

5- b b 0 5-

b b

4_b

b 4-

bc

3- cc

3- a cLa:*

c b2- C

Mme, 2

4 4•

11; i: *.:: iii fi. 1_ . iiil: b tie i:: i: +++ i; i: t... :!:1. 4

1 I I I I I I I

15 30 45 60 90 15 30 45 60 90

Feeding time( min)

Figure 1. Antifeedancy of neem formulations, Azatin WP 4.5, Azad EC 4.5, and Neemix EC 4.5 against 4th- instarlarvae of Culex tarsalis and Culex quinquefasciatus at various concentrations. Significant differences

of the mean numbers of segments filled among concentrations and control at each feeding time areindicated by different letters at 0.05 level using 1- factor ANOVA.


IllCx. tarsalis Ej Cx. quinquefasciatus

Control Azatin WP4.5 at 1 ppm

aa

a a a a a a a aa

6- aa.,

111a,..

54: ri iii1; 1. 101 : 112- :::: ',] :::::::: i:]: iV •::;: i::;1- ::: :: ::::::::: 1:::::::::::: . M:0 MI

I

Azatin WP4.5 at 5 ppm Azatin WP4.5 at 10 ppm7- a a 7-

a a

3-

2- b

Azad EC4.5 at 1 ppm Azad EC4.5 at 5 ppm

a a a a a a b a b

s. a Eli iHQ : ii!lpi •; R: s.

a

F.,il z:I., 3_

7.7a

iitil Illi!I TI

Azad EC4.5 at 10 ppm Neemix EC4.5 at 1 ppmb a a a a a

7- b 7- ab

a'ii:,1, 6- a5- b ..... x., !ip:!

a ! 5-

blhil in mi. 3-

1111 ;:::i:

li11.:• i••:::'':1:':.: 1: : 1.::::'''1•..

i:

iii

2' ,

2271RIN a 1g1111111. • IliCH 11: 11 N:1: i1:1 111;1 Ilk1- " :•':":".':

i; m::::::::: : I ::::: : :. i:

I I

Neemix EC4.5 at 5 ppm Neemix EC4.5 at 10 ppm7- a 7-

5-

iri 111s-a

a a a3- i.•..,.: i.••

2- a2 ! kali IT1 Ili NI:ii: 2-

a aa a a

1- iii.15-; :;:::::;:: ii: ::::::::::: i.„, 'Tin . al

15 30 45 60 15 30 45 60 90

Feeding time( min)

Figure 2. Species- dependent difference in antifeedancy by neem formulations against Culex tarsalis and Culexquinquefasciatus. Significant differences of the mean numbers of segments filled between these speciesat each feeding time are indicated by different letters at 0.05 level using unpaired 2- tailed t test.


Azatin WP4.5 E.1 Azad EC4.5 ® Neemix EC4.5

Cx. tarsalis at 1 ppm Cx. quinquefasciatus at 1 ppm

7- a

asaaa aaa a

aaa aaa

bab

6-

flh1HHiiflH6- b

5- aa

a•

a

2

1IO 1 I 1 1

Cx. tarsalis at 5 ppm Cx. quinquefasciatus at 5 ppm7- a b a

11776- b

5- aba 6- b

a

a a s5

ab

4- b aa

a aa

a3- aa

34:i

b aa

2_ 2- a

bb a

1- ig gi i::: 1- iill., .:,:i:,:

I I I 1 IM

I 1 I I I

Cx. tarsalis at 10 ppm Cx. quinquefasciatus at 10 ppm7- 7 b

a6

b6-

ab

a

5- 5a b

b a4- a

4

3_ tib

3a

a

2- a

Ilkb 2

ba a

b1

bb 1 a aa

JeFeeding time( min)

Figure 3. Antifeedancy of various neem formulations against 4th-instar larvae of Culex tarsalis and Culexquinquefasciatus. Significant differences of the mean numbers of segments filled among variousformulations at each feeding time are indicated by different letters at 0. 05 level using 1- factor ANOVA.


can induce multiple effects in a variety of insect species early as 15- min after the test larvae were transferred toby its unconventional pesticidal properties. Some of the yeast suspension containing AZ.Therefore, primarythese effects are antifeedancy, growth regulation, antifeedancy could be putatively a major form of thisfecundity suppression and sterility, oviposition activity, where the AZ either blocked the input fromrepellency or attractancy, changes in biological fitness, phagostimulant chemoreceptors or stimulated specific

or even blocking the development of vector-borne " deterrent" cells, or both. Furthermore, even though nopathogen in the arthropods ( Mordue and Blackwell mortality was indicated during antifeedant test period,1993). Most of the pesticidal neem products used in the the detrimental effects through the larvicidal activityearlier investigations against mosquitoes were crude ( Mulla et al. 1997) of the consumed AZ could occur atextracts frcm different parts of the neem tree rather than these high AZ concentrations of 5 and 10 ppm, which

commercial formulations ofAZ. In the current research, would result in the secondary antifeedancy during thethe experimental formulations ofwettable powder Azatin process of intoxication. The yeast extracts contain someWP 4. 5 and emulsifiable concentrate Azad EC 4. 5 and components such as nucleic acids, adenylic acid,

Neemix EC 4. 5 developed for agricultural pest control nucleotides, nucleosides, and nucleic acid bases, whichwere employed to investigate the antifeedant activity are powerful phagostimulants to mosquito larvae( Dadd

against larvae of Cx. tarsalis and Cx. quinquefasciatus. 1970b, Dadd et al. 1982). The test larvae ingested inertChemical inhibition of feeding behavior has been materials much faster in the presence of yeast extracts

well investigated in some phytophagous insects. The than in suspensions without yeast extract( Dadd 1970b,primary antifeedancy may function by blocking the Dadd et al. 1982). The present work used 2% yeast

input from chemoreceptors normally responding to suspension as test material for feeding rate. Therefore,phagostimulants, or by stimulating the specialized the observed antifeedancy exhibited by neem productsdeterrentreceptors( Chapman 1974, Schoonhoven 1982). against the test species is probably the result ofThe former may have a general effect on all insects, but overcoming the phagostimulant activity of yeastthe chemicals in the latter category will only be effective components by AZ.if the insects have neurons capable of responding to The species- and formulation-dependent differencesthem( Chapman 1974). The antifeedancy resulting from in the magnitudes of antifeedant activity was indicatedtoxic effects, if consumed, is called secondary antifee- in the present research. Similar trends were also

dancy ( Ascher 1993, Mordue and Blackwell 1993). demonstrated in the larvicidal( Mulla et al. 1997), ovicidalWhen applied in many different ways, AZ and AZ con- ( Su and Mulla 1998a), and ovipositional( Su and Mullatainingextractsfromtheneemtreeshowdistinctprimary 1998b) tests of these formulations against the sameantifeedancy, as well as secondary antifeedancy against species. In larvicidal tests, Azad WP 10 was 10 timesa variety of insects( Mordue and Blackwell 1993). more effective than Azad EC 4. 5 against Cx.

We initiated the current studies on antifeedancy quinquefasciatus( Mulla et al. 1997). In ovicidal tests,

activity of neem products against mosquito larvae by the eggs of Cx, quinquefasciatus were more susceptible

using the assessment method for larval feeding rate than those of Cx. tarsalis to the same neem formulationsproposed by Dadd( 1968), which has been proven as a at equal AZ concentrations. The Azad WP 10 was morepractical and reliable approach for evaluating ingestion effective and more persistent than the Azad EC 4. 5 as anrates of particulates ( Rodcharoen and Mulla 1995). It ovicide against Cx. tarsalis and Cx. quinquefasciatus atwas found in the current research that 5 ppm of AZ was the same AZ concentration ( Su and Mulla 1998a). Inthe minimum effective concentration as an antifeedant ovipositional tests, the Azad WP 10 acted as anin most cases. This effective concentration was higher ovipositional attractant at AZ concentration? 0. 5 ppmthan that for IGR activity by neem products against test against Cx. tarsalis, but acted as a repellent at AZinsect species. For instance, in Cx. quinquefasciatus, concentration 10 ppm against Cx. quinquefasciatus( Sualmost 100% and more than 80% adult emergence and Mulla 1998b).

inhibition was attained at 1 ppm and 0. 1 ppm AZ, By applying the neem products to crops, withinrespectively, in the formulations of Azad WP 10, Azad artificial diets, or on simplified feeding discs, or to theEC 4. 5, and Neemix EC 0.25 ( Mulla et al. 1997). In insects as sprays, topical applications, or by injection orother test insect species, AZ concentrations of 0.2- 4. 1 cannulation, both primary and secondary antifeedantppm by injection, 1. 5- 11. 3 ppm by oral application, and activity of neem products have been indicated in the0. 14- 4. 5 ppm by topical application yielded 15- 50%, members of the following orders: Orthoptera, Isoptera,15- 100%, and 5- 50% molt inhibition, respectively, Hemiptera, Coleoptera, Lepidoptera, and Dipteradepending on the test species ( Mordue and Blackwell ( Mordue and Blackwell 1993). There are clear1993). In our tests the antifeedancy activity was noted as differences, however, regarding the magnitude of this


effect, depending on the concentration and formulation Dadd, R. H. 1970a. Relationship between filteringof the active principles, the ways in which the neem activity and ingestion of solids by larvae of theproducts are applied, and species of test insects. Different mosquito Culex pipiens: a method for assessingfrom phytophagous insects, mosquito larvae inhabit an phagostimulant factor. J. Med. Entomol. 7: 708-aquatic medium. Therefore, the antifeedancy activity or 712.

toxicity can only be achieved by treatment of the whole Dadd, R. H. 1970b. Comparison of rates of ingestion of

breeding medium, rather than larval food materials particulate solids by Culex pipiens larvae: Phago-only. Therefore, the stability of insecticidal principles stimulant effect of water-soluble yeast extract.

of neem products in an aquatic environment, as well as Entomol. Exp. Appl. 13: 407- 419.the safety to nontarget aquatic organisms cohabitating Dadd, R. H., J. E. Kleinjan, and L. D. Merritt. 1982.with mosquito immatures should be considered. Phagostimulant effects of simple nutrients on larval

Neem products degrade faster and easier than Culex pipiens( Diptera: Culicidae). Ann. Entomol.synthetic organic insecticides and in general are less Soc. Am. 75: 605- 612.

hazardous to the nontarget organisms, even though Jones, J. C. 1960. The anatomy and rhythmical activitysome adverse effects of neem components have been of the alimentary canal of Anopheles larvae. Ann.noted in some nontarget aquatic invertebrates ( Stark Entomol. Soc. Am. 53: 459- 474.

1997) and fish ( Tangtong and Wattanasirmkit 1997). Koul, O., M. B. Isman, and C. M. Ketkar. 1990. PropertiesNeem trees have been introduced into vast areas of and uses of neem, Azadirachta indica. Can. J. Bot.tropical and subtropical countries, where it is used as the 68: 1- 11.

source of pharmaceuticals, the manufacture of Mordue, A. J., and A. Blackwell. 1993. Azadirachtin: anagricultural implements and furniture, cattle and poultry update. J. Insect Physiol. 39: 903- 924.

feeds, nitrification ofsoils for various agricultural crops, Mulla, M. S. 1997. Nature and scope of biopesticides.and pest control ( Koul et al. 1990). The pesticidal Proc. 1st Int. Symp. Biopesticides( Rodcharoen, J.,efficacy, environmental safety, and public acceptability S. Wongsiri, and M. S. Mulla, eds.) 1: 5- 9.

of neem products for the control ofcrop pests during the Mulla, M. S., J. D. Chaney, and J. Rodcharoen. 1997.past several years ensure the potential introduction into Activity and efficacy of neem products againstmosquito control programs. mosquito larvae. Proc. 1st Int. Symp. Biopesticides.

Rodcharoen, J., S. Wongsiri, and M. S. Mulla,eds.). 1: 149- 156.

REFERENCES CITED Naqvi, S. N. H., S. O. Ahmed, and F. A. Mohammad.

1991. Toxicity and IGR effect ofnew neem productsAhmed, S., M. Grainge, J. W. Hylin, W. C. Mitchell, and against Aedes aegypti( PCSIR strain). Pakistan J.

J. A. Litsinger. 1984. Investigating the feasibility of Pharma. Sci. 4: 71- 76.

using botanical materials for pest control under Rao, D. R., R. Reuben, and B. A. Nagasampagi. 1995.traditional farming system: A suggested approach. Development ofcombined useofneem( AzadirachtaProc. 2nd Int. Neem Conf. ( Rauischholzhausen, indica) and water management for the control of1983). 2: 545- 550. culicine mosquitoes in rice fields. Med. Vet.

Aly, C. 1985. Feeding rate of larval Aedes vexans Entomol. 9: 25- 33.

stimulated by food substances. J. Am. Mosq. Contr. Rao, K. N., and B. S. Parmar. 1984. A compendium ofAssoc. 1: 406-410. chemical constituents of neem. Neem Newsletter.

Ascher, K. R. S. 1993. Nonconventional insecticidal 1: 39- 46.

effects of pesticides available from the neem tree, Rashed, S. S. and M. S. Mulla. 1989. Factors influencingAzadirachta indica. Archs. Insect Biochem. ingestion ofparticulate materials by mosquito larvaePhysiol. 22: 433- 449. Diptera: Culicidae). J. Med. Entomol. 26: 210-

Chapman, R. F. 1974. The chemical inhibition of feeding 216.

by phytophagous insects: A review. Bull. Entomol. Rodcharoen, J. and M. S. Mulla. 1995. ComparativeRes. 64: 339- 363. ingestion rates of Culexquinquefasciatus( Diptera:

Clements, A. N. 1992. The Physiology of Mosquitoes: Culicidae) susceptible and resistant to BacillusVol. 1. Development, nutrition and reproduction. sphaericus. J. Invert. Pathol. 66: 242- 248.Chapman & Hall. London, Glasgow, New York, Schmutterer, H. 1988. Potential of azadirachtin-Tokyo, Melbourne, Madras 1: 73- 80. containing pesticides for integrated pest control in

Dadd, R. H. 1968. A method for comparing feeding developing and industrialized countries. J. Insectrates in mosquito larvae. Mosq. News. 28: 226- 230. Physiol. 34: 713- 719.


Schmutterer, H. 1990. Properties and potential ofnatural Su, T. and M. S. Mulla. 1998a. Ovicidal activity ofneempesticides from the neem tree, Azadirachta indica. products( azadirachtin) on Culex tarsalis and CulexAnn. Rev. Entomol. 35: 271- 297. quinquefasciatus( Diptera: Culicidae). J. Am. Mosq.

Schoonhoven, L. M. 1982. Biological aspects of Contr. Assoc. 14: 204- 209.

antifeedants. Entomol. Exp. Appl. 31: 57- 69. Su, T. and M. S. Mulla. 1998b. Effects of neem productsSharma, V. P. and M. A. Ansari. 1994. Personal containing azadirachtin on oviposition activity of

protection from mosquitoes ( Diptera: Culicidae) the mosquitoes Culex tarsalis and Culex

by burning neem oil in kerosene. J. Med. Entomol. quinquefasciatus ( Diptera: Culicidae). Entomol.

31: 505- 507. Exp. Appl. (Submitted).Sharma, V. P., M. A. Ansari, and R. K. Razdan. 1993a. Sukumar, K., M. J. Perich, and L. R. Boombar. 1991.

Mosquito repellent action of neem ( Azadirachta Botanical derivatives in mosquito control: areview.indica) oil. J. Am. Mosq. Contr. Assoc. 9: 359- 360. J. Am. Mosq. Contr. Assoc. 7: 210-237.

Sharma, V. P., B. N. Nagpal, and A. Srivastava. 1993b. Tangtong, B. and K. Wattanasirmkit. 1997. AcuteEffectiveness of neem oil mats in repelling toxicity of neem extract on certain blood parametersmosquitoes. Trans. R. Soc. Trop. Med. Hyg. 87: of tilapia, Oreochromis niloticus. Proc. 1st Int.626. Symp. Biopesticides.( Rodcharoen, J., S. Wongsiri,

Stark, J. D. 1997. Risk assessment of neem insecticides: and M. S. Mulla, eds.). ( Phitsanulok, Thailand

Persistence in the environment and potential impact 1996). 1: 94- 103.

on nontarget organisms. Proc. 1st Int. Symp. Zebitz, C. P. W. 1986. Effects of three different neem

Biopesticides. ( Rodcharoen, J., S. Wongsiri, and seed kernel extracts and azadirachtin on larvae of

M. S. Mulla, eds.).( Phitsanulok, Thailand 1996). 1: different mosquito species. J. Appl. Entomol. 102:69- 74. 455- 463.


A New Enzyme Immunoassay to Detect Antibodiesto Arboviruses in the Blood of Wild Birds

R. E. Chiles and W.K. Reisen

Arbovirus Research Unit, Centerfor Vector-borne Disease Research,School of Veterinary Medicine, University of California, Davis, CA 95616

Reprint address: Arbovirus Field Station, 4705 Allen Rd, Bakersfield, CA 93312

Received 23 March 1998; Accepted 24 August 1998

ABSTRACT: A new indirect enzyme immunoassay ( EIA) was developed to screen wild bird sera forantibodies against western equine encephalomyelitis( WEE) and St. Louis encephalitis( SLE) viruses. The

detector antibody was made by immunizing rabbits with serum proteins pooled from single speciesrepresentatives of four bird orders and was conjugated with horseradish peroxidase to allow visualizationwith the ABTS substrate in an EIA plate reader set at 405 nm. The detector antibody recognized a widerange of bird species and was more accurate, sensitive, and specific than a hemaglutination inhibition testwhen compared to a plaque reduction neutralization test( PRNT). EIA positive sera frequently could notbe confirmed by PRNT; however, practically all sera positive by PRNT also were positive by EIA. Thenew EIA has been incorporated into our field research program and has been used to economically screenover 10, 000 wild bird sera from 124 species for antibodies against WEE and SLE.

Keyword Index: Enzyme immuno assay, birds, arboviruses, antibodies, WEE, SLE.

INTRODUCTION they detect antibody in a variety of bird species and donot require laboratory containment procedures.

Wild birds are the principal vertebrate hosts for However, the HI test is labor intensive, tedious toseveral North American encephalitis viruses, including perform, and notorious for producing nonspecific resultswestern equine encephalomyelitis( WEE, Togaviridae) at low antibody titers ( e. g., Hall et al. 1995).and St. Louis encephalitis ( SLE, Flaviviridae) viruses Neutralization tests, such as the plaque reductionReisen and Monath 1989, Tsai and Mitchell 1989). neutralization test ( PRNT), do not require species-

Species in a variety ofavian taxa( mostly Passeriformes specific reagents and are considered to be the" standard"and Columbiformes) become infected and are important for serological testing. However, PRNTs are expensive,in virus enzootic maintenance and amplification require up to two weeks to perform, and must beMcLean and Bowen 1980, Hardy and Reeves 1990, conducted within adequate containment facilitiesMilby and Reeves 1990). A key component in because of the test requirement for live virus culture.encephalitis virus surveillance programs has been the Several enzyme immunoassays ( EIA) have beenmonitoring ofenzootic transmission by measuring either developed. A blocking or competitive enzymeseroconversion rates in sentinels, such as chickens immunoassay( EIA) was developed for Murray ValleyReeves et al. 1990) or seroprevalence rates among encephalitis and several other flaviviruses ( Hall et al.

field-collected birds, especially house sparrows and 1995). One of the monoclonal antibodies( 3H6) crosshouse finches( McLean et al. 1988, Gruwell et al. 1988). reacts with several of the Flaviviridae, including SLE;

Monitoring temporal changes in seroprevalence however, similar competitive monoclonals currentlyrates in wild birds in arbovirus surveillance programs are not available for the Togaviridae. Indirect EIAshas been complicated by the absence of a rapid, specific, have been developed to detect antibodies in specific birdsensitive, and cost effective assay that can detect species, such as sentinel chickens( Oprandy et al. 1988,antibodies against a variety of viruses in different bird Reisen et al. 1994); however, these EIAs rely on hostspecies ( Chiles and Reisen 1997). Hemagglutination specific reagents and are not suited for a surveillanceinhibition assays ( HI) have been used widely because program that tests a variety of bird species. Calisher et


al. ( 1986) reported broad cross reactivity against both were used as negative controls. Both viral strains

IgG and IgM antibodies from several bird species in an previously elicited a strong antibody response in aEIA using goat anti- chicken reagents. However, in our variety of bird species ( Hardy and Reeves 1990) andlaboratory commercial goat anti- chicken antibodies have been used as antigens in EIAs to detect antibodies

failed to react with known positive birds from several to WEE and SLE in humans ( Reisen et al. 1996) and

different orders including Passeriformes ( Chiles and chickens( Reisen et al. 1994). Both viral antigens and

Reisen 1997). the control antigen were diluted in carbonate- bicarbonate

Renewed interest in using temporal changes in coating buffer( pH 9. 6)avian seroprevalence as an early warning of arbovirus Procedures and test parameters are summarized in

activity by California mosquito control agencies Figure 1. Briefly, viral antigen was allowed to bind tostimulated research to develop an assay that was cost Immulon 1 ( Dynatek Laboratories Inc., Winooski, VT)

effective, able to detect antibodies to both WEE and 96 well plates overnight at 4° C. Nonspecific bindingSLE in a variety of bird species, and could be used was blocked with a 2% casein/ PBS- T ( phosphate

without containment. The current paper describes the buffered saline— Tween detergent) solution for 30 min

rationale, reagent development, and laboratory and field at 37°C. After blocking, 0. 1 ml of unknown bird seraevaluations of an indirect EIA that detects antibodies ( diluted 1: 50; e. i., 1: 10 field dilution x 1: 5 laboratoryagainst WEE and SLE in a variety of California bird dilution in physiological saline) was added to each well

species. With slight modification, this EIA may be able and allowed to react with the antigen coated surface. A

to screen sera from most bird species for pathogens that positive reaction was detected by the subsequent bindinginduce antibodies in birds. of the rabbit anti- bird detector antibody that was

conjugated with horseradish peroxidase to the unknown

MATERIALS AND METHODS bird sera, now bound to the virus antigen. The latter

reaction was detected by color development after the

Assay Rationale and Procedures addition of the ABTS substrate( Kirkegaard and Perry,Our assay is based upon the hypothesis that a Gaithersberg, MD) with plates quantitatively assessed

mammalian host immunized with the IgG antibody at 405 nm using a standard EIA plate reader( Dynatekfraction from normal sera pooled from bird species Instruments, Winooski, VT). In practice, sera from 30

representative of several different orders will produce birds were tested per plate for each virus using the mapantibodies that will " recognize" serum proteins from shown in Figure 2. Controls on each plate included

most bird taxa. The degree ofcross- reactivity presumably mock antigen( uninfected Vero cell extract, columns 1,

is related to the diversity of proteins within the 4, 7, and 10) and known positive ( B 1- 3) and negative

immunogen. Our antibody was produced by Bethyl ( A1- 3) bird sera. Although it is best to use known sera

Laboratories, Montgomery, Texas, in domestic rabbits from the same bird taxa that is being tested, it wasthat were immunized with the IgG fraction pooled from difficult to get large amounts of high titered positive

sera collected from white- crowned sparrows sera from small birds such as sparrows, finches, and

Zonotrichia leucophrys, order Passeriformes), ringed warblers. Therefore, known positive and negative

turtle-doves ( Streptopelia risoria, order Columbi- chicken sera were utilized at dilutions to simulate the

formes), domestic chickens ( Gallus gallus, order bird taxa being tested. In practice, wild bird sera

Galliformes), and Muscovy ducks( Cairina moschata, presumptively positive with a mean optical density> 2Xorder Anseriformes). Our choice of bird species was the negative controls were retested by a PRNT. If usedbased on availability and the historical involvement of as a stand alone assay, we recommend using the meanthese bird orders with enzootic virus transmission. The + 3 STD of the negative control wells per plate as the

degree of cross- reactivity may be enhanced by the cutoff for positive sera.

addition of sera from additional orders. Alternatively,antisera production could be directed specifically against Evaluation

a single taxon such as the order Passeriformes to provide The specificity and sensitivity of the new detectora more focussed reagent. antibody conjugate was evaluated in five ways.

Inactivated viral antigens were derived from Vero

cell cultures infected with the BFS 1703 strain ofWEE 1. Species Recognition

or the BFS 1750 strain of SLE. Both viruses originally The ability of the rabbit anti-bird sera to detectwere isolated from Culex tarsalis collected near proteins from a variety of avian taxa was evaluated byBakersfield in 1953 and were passaged twice in suckling a sandwich EIA at Bethyl Laboratories and the

mice prior to Vero cell infection. Uninfected Vero cells Arbovirus Research Laboratory ( ARL). In the


ABTS SUBSTRATE READ AT 405 NM

WASH

CONJ POLYCLONAL HORSERADISH PEROXIDASEANTI-BIRD ABY LABELED ANTI- BIRD ANTIBODY

WASH

BIRD SERAUNKNOWN SERA DILUTED 1: 50

WASH & BLOCK BLOCK FOR 30 MIN WITH 2% CASEIN

ilvERITT ---- ormsremINACTIVATED VIRAL ANTIGENGE NFROM VERO CELL CULTURE

IMMULON® 1 96 WELL PLATE

Figure 1. A diagram showing the components of the indirect EIA.

1 2 3 4 5 6 7 8 9 10 11 12

A - - - 7 15 23

B + + + 8 16 24

C 1 9 17 25

D 2 10 18 26

E 3 11 19 27

F 4 12 20 28

G 5 13 21 29

H 6 14 22 30

Figure 2. Protocol for specimen positioning on the 96 well plates. A1- 3 negative chicken sera. B1- 3 positivechicken sera. Columns 1, 4, 7, and 10 are negative antigen control wells and 2, 3, 5, 6, 8, 9, 11, and 12contain positive antigen. Numbers refer to sera tested in duplicate adjacent wells.

sandwich EIA, affinity purified rabbit anti- bird IgG dilution that produced an absorbance of 1. 0 was> 1: 100.was used as a capture antibody and was adsor-bedto 96 well plates, rinsed, and blocked. Ten- fold 2. Experimental Infectionsdilutions of bird sera from different taxa then were Two groups of six adult ( after hatching year byadded, incubated, and rinsed, after which the plumage) house finches ( Carpodacus mexicanus) ofconjugated rabbit anti-bird sera was added. ABTS mixed sexes each were inoculated subcutaneously in thesubstrate then was added and absorbance read at 405 cervical region with either 4. 0log1oPFU( plaque formingnm. The reaction was considered positive when the units)/ 0. 1 ml of the BFS 1703 strain of WEE or 3. 0 log10


PFU/0. 1 ml of the BFS 1750 strain of SLE. A single encephalitis virus surveillance program. Selected

after hatching year male house sparrow ( Passer positive and negative sera then were shipped to the ARL

domesticus) and two after hatching year mourning doves where they were tested blindly by EIA and PRNT( atZenaida macroura) also were inoculated with SLE. dilutions of 1: 20 and 1: 40).

Birds were bled by jugular puncture( 0.2 cc diluted in0. 8 cc physiological saline) on days 2 and 3 post infection RESULTS

PI) for acute antibody samples for WEE and SLE, andthen on days 13, 17, 20, 22, 24, and 27 PI. Sera were Species Recognition

tested for antibody by the new EIA and by PRNTs in Sera from 13 of 16 bird species in 7 of 10 orders

Vero cell cultures. Confirmatory PRNTs were done at cross- reacted with the rabbit anti- bird antibody anda titer of 1: 20 except on day 27 when end point titers produced an absorbance value of 1 at a dilution of

were determined. 1: 100 in the sandwich EIA (TABLE 1). Strongest

reactions were observed for sera from species that

3. Comparative Sensitivity of the EIA and PRNT comprised the immunogen, but strong reactions alsoSera collected from experimentally infected birds were obtained for domestic turkey, emu, and ostrich.

on day 27 were diluted in a doubling sequence and then Species with end points< 100 included the least bittern

tested by EIA. EIA formula values then were plotted as ( Ixobrychus exilis, order Ardeiformes), sora( Porzana

a function of calculated PRNT titers to determine EIA carolina, order Ralliformes), and least sandpiper

test results on sera of known but varying PRNT titers. ( Calidris minutilla, order Charadriiformes).

Experimental Infections

4. Comparison Between EIA and PRNT in Wild EIA and PRNT tests on house finch sera were

Birds negative for antibody to WEE and SLE on days 2 and 3Wild birds were collected by mist nets or ground post- infection( PI), respectively( TABLE 2, Fig. 3). On

traps in Riverside, Kern, and Sacra- mento counties, day 13 PI, antibody to WEE was detected at low levelsCalifornia, and immediately identified to species, age, in three of five house finches, but results for the two

and sex, bled by jugular puncture( 0. 1 cc blood in 0.9 cc assays were only in agreement for sera from two birds.saline), and released. Blood samples were allowed to Antibody to SLE was detected first on day 17 PI in fiveseparate at ambient temperature, centrifuged, and the of five house finches and two of two mourning doves;sera stored at- 70° C. Sera from Riverside( September the single house sparrow remained negative. By day 22and October 1996) and Sacramento ( 1997) counties PI all birds were positive for antibodies against both

were tested for antibody to either WEE or SLE using WEE and SLE viruses by both tests. Variable test

both the new EIA and a PRNT in Vero cells. Positive results during days 13 to 20 reflected the low concen-EIAs exhibited absorbance values of > 2 times the trations of antibodies present before the rise in titer at 22

negative controls. Positive PRNTs inhibited the days post- infection( Fig. 3). On day 27 PI, the reciprocalformation of> 80% of plaques developed in negative of the geometric mean titer by PRNT was 101 for housecontrols at a serum dilution of>_1: 20. These data were finches infected with WEE, and 60 for house finches

analyzed to estimate the accuracy, sensitivity, and and 631 for mourning doves infected with SLE.specificity of the EIA in comparison to the PRNT.Additional sera from Riverside County( March 1996- Comparative Sensitivity of the EIA and PRNTSeptember 1997) and Kern County ( 1997) were EIA test results ( expressed by formula as meanscreened by EIA and presumptive positives confirmed optical density of the two test wells divided by theby PRNT. These demonstrated the extent of cross negative control well) were plotted as a function of the

reactivity by the detector antibody and the effectiveness doubling dilution of sera collected from experimentallyof the EIA as a screening assay. infected birds on day 27 PI( Fig. 4). EIA formula values

did not vary proportionally with PRNT titers, even for5. Comparison Between HI and EIA in Wild Birds the undiluted sera. For example, house finch 2113 with

Birds ( house finches and house sparrows) were a PRNT titer of 1: 320 had an EIA value of 3. 5 that was

collected in crow traps in Orange County, bled by lower than the remaining birds with a titer of 1: 80.jugular puncture( 0. 1 cc in 0.9 cc diluent), and released. However, there was complete congruence of test

Blood samples were centrifuged, and the sera stored at results; i.e., all undiluted sera( points on Fig. 4 furthest4°C. Sera were tested for antibody against either WEE to the right) had PRNT titers >_1: 20 and EIA formulaor SLE by HI( Clarke and Casals 1958) at the Orange values > 2. The EIA for WEE ( Fig. 4a) seemed lessCounty Vector Control District as part of their sensitive than the EIA for SLE( Fig. 4b), because EIA


values at PRNT titers of 1: 10 were all < 2 for WEE 9), and specificity was 0.997 ( 886/ 889). For SLE,but> 2 for PRNT titers of 1: 5 and 1: 10 for SLE. In fact, accuracy was 0.996( 894/ 898), sensitivity was 0.714( 5/only birds 2121 and 2118 gave negative values for SLE 7), and specificity was 0.997( 889/ 891).at PRNT titers of 1: 20 and 1: 40, respectively( Fig. 4b). Overall, 10,077 sera from 124 bird species collected

in Riverside and Kern counties were screened by EIA,Comparison Between EIA and PRNT in Wild Birds of which 136 and 69 birds were positive by EIA for

A total of 898 birds comprising 70 species was antibodies against WEE and/ or SLE ( TABLE 5). Ofcollected in Riverside County during September and these, 47 ( 35%) and 14 ( 20%) were confirmed byOctober 1996 and Sacramento County during 1997 and PRNT, respectively. Confirmed positives included 15were tested for antibodies against WEE and SLE by both species in four orders ( Gallliformes, Ardeiformes,EIA and PRNT( TABLE 3). Twelve birds were positive Columbiformes, and Passeriformes). Higher confir-for WEE by EIA, of which eight were confirmed by mation rates for WEE than SLE reflected the greaterPRNT, including a black-crowned night heron ( order number ofbirds with WEE titers? 1: 40 collected inKernArdeiformes). There were no false negatives by EIA; County.i. e., sera negative for antibodies to WEE by EIA alsowere negative by PRNT. Four house finches and one Comparisons Among HI, EIA, and PRNT Assayssong sparrow positive for SLE antibodies by EIA were A total of 141 sera was collected in Orange County,confirmed by PRNT, but two Gambel' s quail negative tested for WEE and SLE antibody by HI, and thenby EIA were positive by PRNT. Repeat PRNTs on shipped to the ARL where they were retested by EIAthese quail were again positive, with titers of 1: 20 and PRNT. Overall, 47 were positive for SLE by HI; 3and>_1: 40. had titers of 1: 40, 15 of 1: 20, and an additional 29 were

ElAdataforbothviruseswerecomparedstatistica lly listed as " borderline." By EIA, 17 were positive byagainst the PRNT standard to determine accuracy, formula ( i. e., mean of positive wells divided by thesensitivity and specificity ( TABLE 4). For WEE, negative well was> 2) for SLE and six were nonspecificaccuracy was 0.997( 895/ 898), sensitivity was 1. 000( 9/ ( i. e., noticeable color development in both positive and

TABLE 1. Ability of rabbit anti- bird sera to recognize wild bird sera in asandwich EIA at Bethyl Laboratories and the Arbovirus Research

Laboratory ( ARL). Values are the reciprocal titers from dilution

series starting at 1: 100 that gave an optical density of 1. 0 at 405 nm.

Bird species Bethyl AR

Domestic chicken* 100,000 100,000

Ringed turtle dove* 100,000 1, 600

White-crowned sparrow* 65,000 nd

Muscovy duck* 100,000 nd

Ostrich 6,000 nd

Emu 10,000 nd

Domestic turkey 100,000 nd

Gambel' s quail nd 100

Marsh wren nd 100< 200

Audobon' s warbler nd 100

Northern rough-winged swallow nd 100< 800Bullock' s oriole nd 100Bonaparte' s gull nd 100Least bittern nd 100

Sora nd 100

Least sandpiper nd 100

Species sera used in the rabbit immunogen.

Nd, not done


TABLE 2. Ability of the new EIA( rabbit anti- bird conjugatedantibody) and PRNT to detect antibodies to WEEand SLE in experimentally infected after hatchingyear house finches( n= 6 per virus).

Virus( strain, innoculum) No. positive*

Days Post

infection n EIA PRNT

WEE( BFS 1703, 4.O log PFU/0. 1 ml)10

2 6 0 nd

13 5 3 3

17 5 2 2

20 5 1 5

22 5 5 5

24 5 5 5

27 5 5 5

SLE( BFS 1750, 3. O log PFU/0. 1 ml)io

3 6 0 nd

13 6 0 0

17 5 4 3

20 5 4 3

22 5 5 5

24 5 5 5

27 5 5 5

n, sample size ( 1 bird died in each group); EIA, positive byformula where mean optical density of antigen positive wells/control well> 2; PRNT, titer>_ 1: 20.

negative antigen wells, but value< 2). By PRNT, 19 missed and 16 PRNT negative birds would have been

were positive for SLE; 12 at a titer of 1: 20 and 7 at?. 1: 40. reported as positive. In contrast, the EIA detected all

Two birds strongly positive for WEE by PRNT ( titer seven PRNT positive birds, with formula values

1: 40) and EIA( formula>_4.0) were reported as being As indicated earlier, the 10 false EIA positives were not

negative by HI. considered to be a problem because these would be

When all positives were included and compared discarded after the confirmatory PRNT.against the PRNT results, the HI was less accurate,

sensitive, and specific than the EIA in detecting SLE DISCUSSION

antibody in bird sera ( TABLE 6). Considering thatbirds with a PRNT titer of 1: 20 may represent old or Rabbits immunized with sera pooled from singleborderline positive infections, the analyses were repeated species representatives of four orders of birds produced

using more conservative criteria of PRNT titer? 1: 40, antibodies that reacted with serum proteins from 29

HI titer>_1: 20, and EIA>_2. 0 by formula. This analysis species of birds in seven orders. This extensive cross

markedly improved the agreement between the EIA and reactivity was exploited in a simple three- step indirectPRNT, but decreased agreement between the HI and EIA that detected antibodies in wild birds against WEEPRNT. Reporting only HI titers? 1: 20 as positive would and SLE viruses. When compared against the PRNT,have resulted in only two tests in agreement with PRNT the new EIA was acceptably accurate and sensitive toresults; five PRNT positive birds would have been function as a rapid, inexpensive, screening assay for


OD2. 5-

A. WEE

2 -

HOUSE FINCH

1. 5-

1 —

0. 5r9 8 CUTOFF

02 13 17 20 22 24 27

DAYS POST INFECTION

OD2. 5

B. SLE

2 HOUSE FINCH

HOUSE SPARROW1. 5- + MOURNING DOVE

1 —

0. 5

cfi

0CUTOFF

3 13 17 20 22 24 27

DAYS POST INFECTION

Figure 3. EIA test results expressed as mean optical density of two test wells per serum sample plotted as a functionofdays after infection of adult birds with either( A) WEE or( B) SLE. Cut off values above which values

are positive were expressed as the mean optical density plus two standard deviations( solid line) or twotimes the mean of the negative control wells.

processing the large numbers of wild bird sera needed vigilance may be necessary to ensure that false negativesfor surveillance programs and epidemiological studies. are not recorded for taxa not adequately cross reacting

Although currently in use by our research program, with our detector antibody. Future studies will evaluateminor anomalies with the new EIA were bothersome. the reactivity ofdetector antibody prepared by combiningFirst, the sandwich EIA used to screen for cross reactivity sera from rabbits immunized with serum from individualfailed to react strongly( titer< 1: 100 at an optical density bird taxa.

of 1. 0) with sera from a least bittern, sora, and least Secondly, two Gambel' s quail and two house finchessandpiper. However, the final conjugated detector were positive for SLE antibodies by PRNT, but negativeantibody reacted positively at a 1: 50 dilution to sera by EIA. The quail were collected on the same day andfrom least bitterns and black crowned night herons were tested initially in adjacent wells. Repeat tests onconfirmed as positive for WEE by PRNT. Although the original sera again were positive by PRNT, but ne-some Chadriiformes and Ralliformes were positive by gative by EIA. Interestingly, the two finches missed byEIA, these were not confirmed by PRNT. Continued the EIA were marginally positive by HI. At present, our


12A. WEE

211010r

1— 2111

A 8 42112

F2113

O 6 2115RMU

4n

LA 2 L

0 _ - 1

10 20 40 80 160 320

PRNT TITER

14 rB. SLE

19. 312i - 2116

2117

A102118

F 8- 2120

O 2121R 6-M

4

A2

r

05 10 20 40 80 160 320

PRNT TITER

Figure 4. EIA results expressed by formula( mean optical density of two test wells divided by the control well) andplotted as a function of PRNT titer of sera collected on day 27 post- infection from adult house finchesexperimentally infected with( A) WEE or( B) SLE. Sera from each bird were doubly diluted and thentested by EIA; initial concentrations furthest to the right.

only explanation was that these birds had either recent Thirdly, when used to screen field samples, onlyor very old infections and did not have sufficient anti- 35% of WEE and 20% of SLE EIA positives were

body titers to produce positive reactions in both tests. confirmed by PRNT. Hall et al. ( 1995) attributedSimilar problems of assay agreement were observed comparable assay disparity for several flaviviruses tofor birds infected experimentally with SLE and tested reduced sensitivity by a comparable neutralizationearlier than 17 d PI. Overall, testing California and procedure. In contrast, Calisher et al.( 1986) and OlsonGambel' s quail sera by EIA detected 10 WEE and 12 et al.( 1991) documented good agreement among theirSLE positives, of which 3 and 4 were confirmed by EIAs and HI and neutralization assays. Comparing thePRNT, respectively. Similarly, 29 and 5 house finches performance of our new EIA on sera from experi-were positive by both EIA and PRNT for WEE and SLE, mentally infected birds indicated that EIA formularespectively. Therefore, the detector antibody readily values did not vary quantitatively as a function ofdetects the serum proteins of these two bird species, and PRNT titer, although there was congruence amongwe currently have no explanation for these negative EIA test results among undiluted samples. The WEE EIAresults. seemed less sensitive than the SLE EIA, especially at


TABLE 3. Birds collected in Riverside and Sacramento counties and tested for antibodies to WEE and SLEviruses by both EIA and PRNT assays.

No. WEE SLE

COMMON NAME SCIENTIFIC NAME Tested EIA PRNT EIA PRNT

GAMBEL' S QUAIL Callipepla gambelii 30 0 0 0 2

LEAST BI1"1'ERN Ixobyrchus exilis 3 0 0 0 0

GREAT EGRET Casmerodius albus 1 0 0 0 0

GREEN-BACKED HERON Butorides striatus 4 0 0 0 0

BLACK-CROWNED NIGHT HERON Nycticorax nycticorax 3 1 1 0 0

SORA Porzana carolina 2 0 0 0 0

COMMON MOORHEN Gallinula chloropus 1 0 0 0 0

AMERICAN COOT Fulica americana 1 0 0 0 0

MOURNING DOVE Zenaida macroura 11 0 0 0 0

COMMON GROUND DOVE Columbina passerina 34 0 0 0 0

RED-NAPED SAPSUCKER Sphyrapicus nuchalis 1 0 0 0 0

NUTTALL' S WOODPECKER Picoides nuttallii 6 1 0 0 0

SAY' S PHOEBE Sayornis saya 1 0 0 0 0

BLACK PHOEBE Sayornis nigricans 11 0 0 0 0

UNIDENTIFIED FLYCATCHERS Empidonax spp. 34 0 0 0 0

ASH-THROATED FLYCATCHER Myiarchus cinerascens 1 0 0 0 0

BROWN-HEADED COWBIRD Molothrus ater 1 0 0 1 0

RED-WINGED BLACKBIRD Agelaius phoeniceus 4 0 0 0 0

GREAT-TAILED GRACKLE Quiscalus mexicanus 1 0 0 0 0

CASSIN' S FINCH Carpodacus cassinii 1 0 0 0 0

HOUSE FINCH Carpodacus mexicanus 93 5 5 5 4

LESSER GOLDFINCH Carduelis psaltria 6 0 0 0 0

LAWRENCE' S GOLDFINCH Carduelis lawrencei 1 0 0 0 0

AMERICAN GOLDFINCH Carduelis tristis 24 1 0 0 0

VESPER SPARROW Pooecetes gramineus 3 0 0 0 0

SAVANNAH SPARROW Passericulus sandwichensis 13 0 0 0 0

GRASSHOPPER SPARROW Ammodramus savannarum 1 0 0 0 0

LARK SPARROW Chondestes grammacus 6 0 0 0 0

GOLDEN-CROWNED SPARROW Zonotrichia atricapilla 14 0 0 0 0

WHITE-CROWNED SPARROW Zonotrichia leucophrys 133 1 0 0 0

WHITE-THROATED SPARROW Zonotrichia albicollis 2 0 0 0 0

BREWER' S SPARROW Spezella breweri 6 0 0 0 0

SAGE SPARROW Amphispiza belli 5 0 0 0 0

SONG SPARROW Melospiza melodia 62 1 1 1 1

LINCOLN' S SPARROW Melospiza lincolnii 89 0 0 0 0

FOX SPARROW Passerella iliaca 8 0 0 0 0SPOTTED TOWHEE Pipilo erythrophthalmus 20 0 0 0 0

CALIFORNIA TOWHEE Pipilo fuscus 3 0 0 0 0GREEN-TAILED TOWHEE Pipilo chlorurus 1 0 0 0 0

LAZULI BUNTING Passerina amoena 5 0 0 0 0

BLACK-HEADED GROSBEAK Pheucticus melanocephalus 4 0 0 0 0

WESTERN TANAGER Piranga ludoviciana 2 0 0 0 0

BARN SWALLOW Hirundo rustica 13 0 0 0 0

NO. ROUGH-WINGED SWALLOW Stelgidopteryx serripennis 1 0 0 0 0

TABLE 3 continued on next page


TABLE 3 continued.

No. WEE SLE

COMMON NAME SCIENTIFIC NAME Tested EIA PRNT EIA PRNT

LOGGERHEAD SHRIKE Lanius ludovicianus 1 0 0 0 0

WARBLING VIREO Vireo gilvus 6 0 0 0 0

ORANGE-CROWNED WARBLER Vermivora celata 19 0 0 0 0

NASH\ ILLE WARBLER Vermivora ruficapilla 3 0 0 0 0

YELLOW WARBLER Dendroica petechia 28 0 0 0 0

YELLOW- RUMPED WARBLER Dendroica coronata 10 0 0 0 0

BLACK-THROATED GRAY W. Dendroica nigrescens 3 0 0 0 0

HERMIT WARBLER Dendroica occidentalis 1 0 0 0 0

NORTHERN WATERTHRUSH Seiurus noveboracensis 1 0 0 0 0

COMMON YELLOWTHROAT Geothlypis trichas 27 0 0 0 0

WILSON' S WARBLER Wilsonia pusilla 19 0 0 0 0

MACGILLIVRAY' S WARBLER Oporornis tolmiei 4 0 0 0 0

AMERICAN REDSTART Setophaga ruticilla 1 0 0 0 0

HOUSE SPARROW Passer domesticus 56 0 0 0 0

NORTHERN MOCKING BIRD Mimus polyglottos 3 0 0 0 0

SCRUB JAY Aphelocoma coerulescens 4 0 0 0 0

BEWICK' S WREN Thryomanes bewickii 7 1 1 0 0

MARSH WREN Cistothorus palustris 15 0 0 0 0

HOUSE WREN Troglodytes aedon 5 0 0 0 0

WHITE-BREASTED NUTHATCH Sitta carolinensis 2 0 0 0 0

OAK TITMOUSE Parus inornatus 2 1 1 0 0

WRENTIT Chamea fasciata 2 0 0 0 0

VERDIN Auriparus flaviceps 2 0 0 0 0

BLUE-GRAY GNATCATCHER Polioptila caerulea 2 0 0 0 0

HERMIT THRUSH Catharus guttatus 8 0 0 0 0

AMERICAN ROBIN Turdus migratorius 1 0 0 0 0

TOTAL 898 12 9 7 7

TABLE 4. Comparison of the EIA with the PRNT standard for detectingWEE and SLE antibodies in wild birds collected in Riverside and

Sacramento counties.

PRNT

Virus EIA Positive Negative Total

WEE Positive 9 3 12

Negative 0 886 886

Total 9 889 898

SLE Positive 5 2 7

Negative 2 889 891

Total 7 891 89


TABLE 5. Birds collected in Riverside( March 1996—September 1997) and Kern( March—October1997) counties, screened by EIA, and confirmed by PRNT assays for WEE and SLEantibodies.

NO. WEE SLECOMMON NAME TESTED EIA PRNT EIA PRNT

GAMBEL' S QUAIL 612 9 3 12 4CALIFORNIA QUAIL 69 1 0 0 0EARED GREBE 8 1 0 0 0MALLARD 1 1 0 0 0LEAST BI'1- FERN 68 5 1 3 0SNOWY EGRET 5 1 0 0 0GREEN-BACKED HERON 17 0 0 1 0BLACK-CROWNED NIGHT HERON 8 1 1 0 0WESTERN SANDPIPER 123 2 0 0 0KILLDEER 9 0 0 1 0MOURNING DOVE 45 0 0 1 0COMMON GROUND DOVE 203 1 1 4 2LESSER NIGHTHAWK 21 1 0 0 0BLACK PHOEBE 72 0 0 1 0EUROPEAN STARLING 84 1 0 1 0BROWN-HEADED COWBIRD 145 2 1 1 0RED-WINGED BLACKBIRD 263 10 3 7 0

NORTHERN ORIOLE 19 2 0 1 0BULLOCK' S ORIOLE 151 4 2 0 0GREAT-TAILED GRACKLE 17 2 0 0 0HOUSE FINCH 1, 256 65 29 10 5WHITE-CROWNED SPARROW 1, 577 7 0 7 0GOLDEN-CROWNED SPARROW 51 2 0 0 0SONG SPARROW 1, 133 1 0 3 0ABERT' S TOWHEE 55 1 0 2 0BLACK-HEADED GROSBEAK 18 1 1 0 0VIOLET-GREEN SWALLOW 1 1 0 0 0NORTHERN ROUGH-WINGED SWALLOW 66 0 0 1 0

LOGGERHEAD SHRIKE 25 1 0 0 0WARBLING VIREO 79 1 1 0 0NASHVILLE WARBLER 27 0 0 1 0YELLOW- RUMPED WARBLER 693 1 0 1 0MACGILLIVRAY' S WARBLER 27 0 0 1 0COMMON YELLOWTHROAT 186 1 0 2 0YELLOW-BREASTED CHAT 14 0 0 1 0HOUSE SPARROW 291 1 0 3 1NORTHERN MOCKING BIRD 39 0 0 1 0CALIFORNIA THRASHER 19 2 2 0 0CACTUS WREN 7 0 0 1 1MARSH WREN 211 1 0 2 1AMERICAN ROBIN 39 6 2 0 0

TOTAL POSITIVE( 41 SPECIES) 7,754 136 47 69 14TOTAL NEGATIVE( 83 SPECIES) 2, 323 0 0 0 0TOTAL( 124 SPECIES) 10,077 136 47 69 14


TABLE. 6. Comparison of HI, EIA, and PRNT on bird sera collected in Orange County.

Standard

Comparison' Pos. Neg. Total Statistic'

PRNT(? 1: 20)

HI Positive 15 32 47 ACC 0.745

all) Negative 4 90 94 SEN 0.789

Total 19 122 141 SPE 0. 738

EIA Positive 17 6 23 ACC 0.943

all) Negative 2 116 118 SEN 0.895

Total 19 122 141 SPE 0.951

PRNT(>_1: 40)

HI Positive 2 16 18 ACC 0.851

1: 20) Negative 5 118 123 SEN 0.286

Total 7 134 141 SPE 0.881

EIA Positive 7 10 17 ACC 0.929

2.0) Negative 0 124 124 SEN 1. 000

Total 7 134 141 SPE 0.925

Comparison: HI, all= 1: 20, 1: 40 and borderline pooled; EIA, all=> 2.0 and

nonspecific

Statistic: ACC, accuracy, SEN, sensitivity, SPE, specificity

low PRNT titers, and this may have accounted for the Lothrop, and Vincent Martinez, University ofhigher confirmation rate among sera from field-collected California, Davis, and Stan Wright, Sacramento/Yolo

birds. Mosquito and Vector Control District for collecting theThe new EIA has several advantages. First, the birds in Kern, Riverside, and Sacramento counties.

detector antibody was relatively inexpensive to produce David Gutierrez, University of California, Davis,in large volumes. Immunization, bleeding, purification, assisted in screening the bird sera by EIA. Carrie

and conjugation ofrabbit sera by the Bethyl Corporation Fogarty, Orange County Vector Control District,required approximately three months and<$ 3,000 for performed HI tests on the Orange County birds. Weenough antibody to test thousands of birds when diluted thank Jame P. Webb, Jr. for permission to cite this data.

1: 1500. Production would be more cost-effective using Laura D. Kramer, University of California, Davis,a goat as the host animal to produce the antisera. critically read the manuscript. We especially thank twoSecondly, the three step EIA is rapid and inexpensive to anonymous reviewers of the Journal of Vector Ecologyperform. We routinely screen 240 bird sera for WEE whose comments markedly improved an earlier versionand SLE by this EIA in a single day at a cost of about of this manuscript.

1. 50 per sera. Third, the detector antibody This research was funded, in part, by grants fromtheoretically can be used in a variety of similar assays to the Mosquito and Vector Control Association of

test any bird taxa for infection with any pathogen California, the Coachella Valley Mosquito Abatementproviding a suitable antigen is available and will bind to District, the Mosquito Research Program of the

Immulon®plates. Division of Agriculture and Natural Resources,

University of California, and the US NIH ( Grant R21Acknowledgments. AI39483- 01A1). Logistical support was provided by

the Kern Mosquito and Vector Control District and the

We especially thank Robert Cusack, Hugh Coachella Valley Mosquito Abatement District.


REFERENCES CITED Reeves, ed.). Calif. Mosq. Vector Contr. Assoc.,Sacramento, CA, 508 pp.

Calisher, C.H., H.N. Freemount, W.L. Vesley, A.O. El- Olson, J. G., T. W. Scott, L. H. Lorenz, and J. L.Kafrawi, and M.I. Al- Deen Mahmud. 1986. Hubbard. 1991. Enzyme immunoassay for detectionRelevance ofdetection of immunoglobulin class M of antibodies against eastern equine encepha-

antibody response in birds used in arbovirus lomyelitis virus in sentinel chickens. J. Clin.surveillance. J. Clin. Microbiol. 24: 770-774. Microbiol. 29: 1457- 1461.

Chiles, R. E. and W. K. Reisen. 1997. Research toward Oprandy, J. J., J. G. Olson, and T. W. Scott. 1988. Athe development of an enzyme immunoassay to rapid dot immunoassay for the detection of serumdetect arbovirus antibodies in wild birds. Proc. antibodies to eastern equine encephalomyelitis and

Mosq. Vector Contr. Assoc. Calif. 65: 24- 27. St. Louis encephalitis viruses in sentinel chickens.Clarke, D. H. and J. Casals. 1958. Techniques for Am. J. Trop. Med. Hyg. 38: 181- 186.

hemagglutination and hemagglutination inhibition Reeves, W. C., M. M. Milby, and W. K. Reisen. 1990.with arthropod- borne viruses. Am. J. Trop. Med. Development of a statewide arbovirus surveillance

Hyg. 7: 561- 573. program and models of vector populations and

Gruwell, J. A., B. L. Brown, and J. P. Webb Jr. 1988. virus transmission. Pp. 431- 458 in EpidemiologyPasseriform birds as a surveillance method for and Control of Mosquito- Borne Arboviruses inarbovirus activity in rural and suburban sites in California, 1983- 1987 ( W. C. Reeves, ed.). Calif.

Orange County, California 1987. Proc. Calif.Mosq. Mosq. Vector Contr. Assoc., Sacramento, CA, 508Vector Contr. Assoc. 56: 58- 68. pp.

Hall, R. A., A. K. Broom, A. C. Hartnett, M. J. Howard, Reisen, W. K. and T. P. Monath. 1989. Western equineand J. S. Mackenzie. 1995. Immunodominant encephalomyelitis. Pp. 89- 138 in The Arboviruses:epitopes on the NS 1 protein of MVE and KUN Epidemiology and Ecology, Vol. V( T. P. Monath,viruses serve as targets for a blocking ELISA to ed.). CRC Press, Boca Raton, FL, 241 pp.detect virus-specific antibodies in sentinel animal Reisen, W. K., R. E. Chiles, H. D. Lothrop, S. B.serum. J. Virol. Meth. 51: 201- 210. Presser, and J. L. Hardy. 1996. Prevalence of

Hardy, J. L. and W. C. Reeves. 1990. Experimental antibodies to mosquito- borne encephalitis virusesstudies on infection in vertebrate hosts. Pp. 66- 127 in residents of the Coachella Valley. Am. J. Trop.in Epidemiology and Control of Mosquito-Borne Med. Hyg. 55: 667- 671.Arboviruses in California, 1943- 1987 ( W. C. Reisen, W. K., R. P. Meyer, S. B. Presser, and J. L.Reeves, ed.). Calif. Mosq. Vector Contr. Assoc., Hardy. 1993. Effect of temperature on the trans-Sacramento, CA, 508 pp. mission of western equine encephalomyelitis and

McLean, R.G. and G. S. Bowen. 1980. Vertebrate hosts. St. Louis encephalitis viruses by Culex tarsalisPp. 381- 450 in St. Louis Encephalitis ( T. P. Diptera: Culicidae). J. Med. Entomol. 30: 151-

Monath, ed). American Public Health Association, 160.

Washington, DC, 508 pp. Reisen, W. K., S. B. Presser, J. Lin, B. Enge, J. L.McLean, R. G., J. P. Webb, E. G. Campos, J. Gruwell, Hardy, and R. W. Emmons. 1994. Viremia and

D. B. Francy, D. Womeldorf, C. M. Myers, T. H. serological responses in adult chickens infectedWork, and M. Jozan. 1988. Antibody prevalence of with western equine encephalomyelitis and St. LouisSt. Louis encephalitis virus in avian hosts in Los encephalitis viruses. J. Am. Mosq. Contr. Assoc.Angeles, California, 1986. J. Am. Mosq. Contr. 10: 549- 555.

Assoc. 4: 524- 528. Tsai, T. F. and C. J. Mitchell. 1989. St. Louis

Milby, M. M. and W. C. Reeves. 1990. Natural infection encephalitis. Pp. 431- 458 in The Arboviruses:in vertebrate hosts other than man. Pp. 26- 65. in Epidemiology and Ecology, Vol. IV ( T. P.

Epidemiology and Control of Mosquito- Borne Monath, ed.). CRC Press, Boca Raton, FL, 243Arboviruses in California, 1943- 1987 ( W. C. pp.


Surveillance Studies of Orthopodomyia signifera

with Comparisons to Aedes sierrensis

David L. Woodward, Arthur E. Colwell, and Norman L. Anderson

Lake County Vector Control DistrictP. O. Box 310, Lakeport, CA 95453

Received 26 March 1998; Accepted 13 May 1998

ABSTRACT: The oviposition behaviors of Orthopodomyia signifera and Aedes sierrensis were

examined with ovitraps and larval surveys during a five-year field study in northern California.Ovipositional periodicity was found to be an important factor affecting the segregation of the species intotree holes that were temporarily (Ae. sierrensis only) or permanently ( both species) filled with water.Orthopodomyia signifera females also used cues associated with the permanence of the habitat when

selecting oviposition sites. The distributions of the eggs of each species along horizontal and verticaltransects did not indicate that interspecific competition affected oviposition site selection. The data

indicate that temporal partitioning of larval development may reduce interspecific competition. Ovitrapswere determined to be more sensitive tools for surveillance of Or. signifera than larval surveys, CO2- baitedFay traps and CDC light traps.

Keyword Index: Tree hole, mosquito, Orthopodomyia, oviposition.

INTRODUCTION because females are economically important bitingpests of humans( Papineau 1984) and known vectors of

Seasonal flooding places an important environ- Dirofilaria immitis ( Leidy), the canine heartworm

mental constraint on mosquitoes utilizing tree holes as ( Weinmann and Garcia 1974, Walters and Lavoipierre

breeding sites in northern California oak woodlands. 1982).

This region is characterized by a Mediterranean climate Comparatively, much less is known about theSteinhauser 1979) with mild, wet winters and hot, dry adaptive traits which allow Or. signifera populations to

summers. More than 90% of tree holes dry completely survive the annual perturbation of summer drought.

each summer( Washburn and Hartmann 1992), making Previous studies ( Chapman 1964, Baerg 1968) havemost of these habitats seasonally unsuitable for larval shown Or. signifera eggs hatch upon completion of

development. As a result, the mosquito fauna lacks embryonic development, a period of four days at 30° C.

diversity relative to other regions of North America. The lack of a resting state in the egg stage coupled withOrthopodomyia signifera ( Coquillett) and Aedes an unusually long larval development period ( Bohartsierrensis ( Ludlow) are the only northern California 1950) restricts successful populations to the lowCulicidae largely restricted to tree holes ( Bohart and percentage of tree holes that hold water year- round.Washino 1978). Grant ( 1953) and Baerg ( 1968) observed that Or.

Aedes sierrensis, the western tree hole mosquito, is signifera larvae do not pupate in California until high

highly adapted for survival through the dry- summer outdoor temperatures have been reached. This trait mayconditions that characterize most California forests. delay adult activity until temporary rain- filled tree holesFemales oviposit drought-resistant eggs above the have dried, thereby restricting oviposition to permanentwater in tree holes during spring and summer. The eggs habitats. Due to a lack of effective surveillancedo not hatch until flooded by rain, usually in fall and techniques, this hypothesis has not been directlywinter( Reisen and Reeves 1990). The life history( e. g., investigated.

Peyton 1956, Hawley 1985) and seasonal abundance Further study of the ecology and vector potential ofGarcia et al. 1989, Washburn et al. 1989, Woodward et Or. signifera is needed. This species is widely

al. 1996) ofAe. sierrensis have been extensively studied distributed throughout temperate regions of eastern


North America. The known distribution in the west Weather Dataincludes the lower elevation mountains and valleys of Daily precipitation and maximum temperature datacentral California and southern Oregon ( Darsie and for each year of the seasonal study were obtained fromWard 1981). Although females have been collected a United States National Weather Service stationby human sentinels( J. O. Washburn, pers. commun.), located 10 km north of the study site.they are not known to bite people ( Zavortink 1968).

They may, however, indirectly affect public health Surveillance of Natural Tree Holes for Water andbecause they readily take blood meals from birds Immature MosquitoesChapman 1964, Baerg 1965, Zavortink 1968). For Prior to the study, the volumes of the 10 tree holes

example, Chamberlain et al. ( 1954) determined Or. in the woodland were determined when they weresignifera females were effective vectors of both EEE naturally full of water. All standing water was removed,eastern equine encephalitis virus) and WEE( western measured, and replaced to determine maximum

equine encephalitis virus) in laboratory transmission volumes. The distance from the water surface to thestudies, and Vargas ( 1960) isolated EEE from field- ground was measured for each tree hole. Polyethylenecollected females in eastern Mexico. Further evaluation rulers were mounted in each tree hole and during eachof the role of Or. signifera in maintenance of medi- year of the seasonal study, water depths were monitoredcally important arbovirus cycles in sylvan bird weekly ( April to November), or one to four timespopulations has been hampered by a lack of ecological monthly ( December to March), with the exceptions ofdata. January and December 1991 ( no data were collected).

In the present study, ovitraps were utilized to On each sampling date, the water in each tree holemonitor the seasonal distribution of oviposition by Or. was gently agitated prior to removal of each of threesignifera females. The investigation was designed to water samples( ca. 45 ml each) by pipette( 5 mm mouthanalyze the relationships among climate, seasonal diameter). Immature mosquitoes collected by theflooding of natural tree holes, and the timing of pipette were identified ( Bohart and Washino 1978) inoviposition. The temporal and spatial relationships of the field ( Ae. sierrensis) or the laboratory ( all otheroviposition by Or.signifera and Ae. sierrensis were also species) and returned to the tree holes. Tree holes werecompared. Finally, larval surveys, CO2- baited Fay considered to be dry when no water could be removedtraps( Fay and Prince 1970), CDC light traps( Sudia and with the pipette. All ten tree holes were surveyed forChamberlain 1962), and ovitraps were compared as water and immature mosquitoes on a total of 36 dates inmethods for surveillance of Or. signifera. 1991, 43 dates in 1993, 42 dates in 1995, and 50 dates in

1997. In addition, on July 14, 1993, a water sample( 50METHODS AND MATERIALS ml) was removed from each of the five deepest tree

holes and replaced with deionized water. The pH ofStudy Site each sample was measured in the laboratory with a

The study was conducted in a 1. 2 ha densely meter. The samples were then centrifuged for 8 minutesforested northern oak woodland ( Munz 1965) near at 2200 rpm to remove suspended particles. The truePotter Valley, Mendocino County, California( Latitude color( platinum-cobalt units) of the supernatant of each39° 14' N, Longitude 123° 06' W, elevation 332 m) sample was determined by spectrophotometerdominated by interior live oak ( Quercus wislizenii according to Greenberg ( 1985).Candolle) and Pacific madrone ( Arbutus menziessiiPursh). Maximum height of the tree canopy was ca. 16 Ovitrap Designm. Winter searches of the woodland for water- filled tree A translucent polyethylene cup( 473 ml) was filledholes( 0-5 m above ground) were conducted each year with 380 ml ofattractant water. Each cup was lined withof the study on foot and with an extension ladder. Ten a Teri- wiper® towel strip( 10 x 27 cm), the ovipositiontree holes were located on interior live oaks( n= 8) and substrate, and placed in a plywood box( 29. 2 cm height,Pacific madrones ( n= 2), all of which held water each 15. 2 cm width, and 17. 8 cm depth) that was paintedyear of the study. The woodland was bordered by a gloss black on all surfaces. An 11. 4 x 11. 4 cm verticalhighway to the south and in the other directions by less entrance hole ( backed with 2. 5 cm mesh hardwaredensely forested foothill woodland dominated by blue cloth) was centered on the front ofeach box at a point 7. 7oak( Q. douglassii Hooker and Arnott). Study periods cm below the top. Each lid was hinged and hasped toincluded the years 1991, 1993, 1995, and 1997 allow access( see Woodward et al. 1996).seasonal surveillance of oviposition) and 1994 Blue oak tree hole water was used in the ovitraps asvertical distribution of oviposition). an ovipositional attractant from 1991 to 1995. During


these years attractant water was collected from the same Response of Or. signifera Adults to Carbon Dioxide-

tree hole ( volume ca. 85 liters) which was kept full Baited Fay Traps and CDC Light Trapsduring study periods by replacing removed or Two stations, one near the middle of each ovitrapevaporated water with deionized water. Similar transect, were established as locations for the operation

methods were used during 1997 except the attractant of Fay traps. These suction traps present a black-and-water used was a mixture of 25% blue oak tree hole white visual pattern without light and were baited with

water( all collected from a second tree hole, volume ca. 2. 3 kg of dry ice( as per Garcia et al. 1989). Traps were

16 liters) and 75% blue oak leaf infusion. The infusion placed in the field at ca. 1400 hr and retrieved the

was prepared on March 19, 1997 and stored according following day at ca. 1000 hr. Each was operated onceto Woodward et al. ( 1996). The pH of the attractant per week on 32 dates from April 3 to November 18,

water( measured on at least two dates during each year 1991, on 34 dates between March 25 and November 16,

of the study) ranged from 7. 56 to 7. 83 during 1991 to 1995, and on 34 dates between March 4 and October 21,

1995. In 1997 the pH ranged from 7. 78 to 7.93. The true 1997. Live mosquitoes were anesthetized with carbon

color of the attractant water measured 5773 Pt-Co units dioxide, counted ( 10X magnification), identified

on July 14, 1993 and 4140 Pt-Co units on July 24, 1997. ( Bohart and Washino 1978), and released in the field.

Dead adults were returned to the laboratory forSeasonal Distribution of Or. signifera Oviposition enumeration and identification. In addition, three

Seasonal oviposition was studied using a total of standard CDC light traps were operated at stations

eight ovitraps located on two east-west transects of the transecting the center of the woodland east to west inwoodland. Stations 1- 4 transected the center of the 1997. These traps were operated without carbon

woodland and stations 5- 8 were < 15 m from the dioxide once per week on 15 dates between June 26 and

transition to foothill woodland at the northern edge of September 14, 1997. Other methods conformed to those

the study site. Each ovitrap was staked to the ground on described for operation of the Fay traps.the north side of a tree trunk( 25- 50 m apart). Ovitrapentrances were 0.2 m above ground and faced north. Statistical Analyses

Using the same locations each year, the ovitraps were Statistical analyses were performed according tooperated continuously from April 2 to December 4, methods described by Zar ( 1980). Variables used in

1991, from April 5 to December 16, 1993, from April 11 statistical analyses were tested for normality with theto December 11, 1995, and from March 26 to December Shapiro-Wilks' W test. Analyses which included

1, 1997. Tree hole water and liners were replaced once variables with a significant W( P< 0.05) were completed

per week. Mosquito eggs oviposited onto liners were with nonparametric methods.

identified using methods described by Woodward et al. Daily precipitation from January 1 to June 30 and1996). Species determinations were confirmed by daily maximum temperatures from June 1 to September

identification ( Bohart and Washino 1978) of reared 30 were compared between years with a Kruskal- Wallis

fourth- instar larvae and adults. Eggs were counted in ANOVA by ranks followed by a multiple range test.the laboratory with a dissecting microscope ( 10X Using a Spearman rank order correlation ( R), total

magnification). Rarely, Or. signifera eggs hatched monthly collections of Or. signifera eggs werebefore the liners were returned to the laboratory. All Or. regressed with corresponding mean monthly maximumsignifera eggs, hatched or unhatched, were included in temperatures, total monthly precipitation, the numberthe counts. of natural tree holes holding water and mean tree hole

water depth. Seasonal totals of Or. signifera and Ae.

Vertical Distribution of Or. signifera Oviposition sierrensis oviposition at each of the eight ovitrapThe vertical pattern of oviposition was examined stations were compared for each year of the study with

with six ovitraps mounted on the north side of the a Spearman rank order correlation analysis.

primary stem of an interior live oak with the entrances Oviposition site selection by Or. signifera and Ae.0.2, 1. 0, 3. 0, 5. 0, 7. 5, and 10.0 m above ground. The tree sierrensis females was further investigated by Chi-was near the center of the study site and corresponded to square analysis. The frequencies of ovitrap liners withstation 2 of the seasonal study. Higher level ovitraps and without the eggs of each species were analyzed forwere serviced with a ladder( 3. 0 and 5. 0 m) or a rope and independence using a 2 x 2 contingency table. The

pulley system ( 7. 5 and 10.0 m). These ovitraps were vertical ovipositional distributions of each species wereoperated continuously from May 9 to November 2, 1994 analyzed using Kruskal- Wallis ANOVA by ranksusing the same methods described for other years of the followed by multiple range tests. The numbers of eggsstudy. collected at each height were compared for the entire


seasonal ovipositional periods of each species. during the entire year( Fig. 2). Throughout the study, thefive deepest tree holes ( mean depth = 23. 9 cm) held

RESULTS water longest into the summer months. Each of these

tree holes had an exposed, horizontal opening. On JulyWeather Data 14, 1993 the pH of the water in these deep tree holes

Rainfall and maximum temperature data are shown ranged from 6. 19 to 7.26( mean± std. dev.= 6. 84± 0.42)

in Figure lA ( 1991, 1993, and 1995) and Figure 2A and true color measured between 1144 and 2356 Pt- Co

1997). Daily precipitation totals from January 1 to June units( mean± std. dev.= 1732± 454 Pt-Co units).

30 were significantly( P< 0.05) lower in 1991 ( 57.9 cm)than in 1993 ( 88. 9 cm), 1995 ( 136. 8 cm), and 1997 Temporal Distribution of Or. signifera Oviposition

106. 7 cm). Maximum air temperatures peaked during Orthopodomyia signifera females oviposited totals

the dry summer months each year, but were highest of 4,247, 6, 083, and 4, 135 eggs into the eight ovitraps

P< 0.05) in 1991. From June 1 to September 30, 1991 during 1991, 1993, and 1995, respectively( Fig. 1B). Andaily maximum temperatures averaged 32. 1° C. Means additional 750 Or. signifera eggs were found in the

for the same months in 1993, 1995, and 1997 were 30. 1, ovitraps during 1997( Fig. 2C). A single, peak period of30. 1, and 29. 9° C, respectively. The months of August ovipositional activity occurred during August orand September lacked rain each year except 1997. September of each year. Orthopodomyia signifera

oviposition was detected in the woodland from June 19

Surveillance of Natural Tree Holes for Water to October 9, 1991. Oviposition was more restricted in

Maximum volumes of the ten tree holes ranged duration in 1993( July 8 to October 5), 1995( July 26 tofrom0.4to 12. 21iters( mean± std. dev.= 4. 5± 4.3 liters) October 5), and 1997( July 17 to September 4).and maximum water depths were between 6. 4 and 30.0 The oviposition periods of Or. signifera coincided

cm( mean± std. dev.= 17. 1± 8. 2 cm). The height of the with the hottest and driest months during each yeartree holes above ground ranged from 0.07 to 1. 31 m ( Figs. 1 and 2). More than 97% of the oviposition

mean± std. dev.= 0. 34± 0.38 m). coincided with trap periods that completely lackedThe data shown in Figures 1 and 2 indicate the rainfall and 77% occurred during periods with mean

presence of water in the tree holes was rainfall- maximum temperatures > 30° C. The cessation of

dependent. Winter rain kept water levels high in all of oviposition also showed close coincidence with the

the tree holes, but water depths declined with the annual onset of cooler weather. The last detectable oviposition

onset ofsummer drought. All of the tree holes were dry of each year occurred within two weeks of the last trapfrom July 24 to October 31, 1991, from August 30 to periods with mean maximum temperatures > 30° C.

November 29, 1993, and from August 24 to December Regression analysis( TABLE 1) showed mean monthly11, 1995( Fig. 1B). Following unusual rains in August maximum temperatures were positively correlated withand September of 1997, three tree holes held water total monthly Or. signifera oviposition during each year

TABLE 1. Spearman rank correlation coefficients ( R) for the total monthly ovitrap collections ofOrthopodomyia signifera eggs and the mean monthly maximum temperature, total monthlyrainfall, mean number of tree holes holding water, and mean monthly tree hole water depth.Data for the months from April through November are included for each year.

Spearman R

Eggs Mean Monthly Total Number of Mean Monthlyfor Maximum Monthly Tree Holes Tree Hole

Year Temperature Rainfall Holding Water Water Depth

1991 0. 88 ** 0.76 * 0.90 ** 0.90 **

1993 0. 80 * 0.91 ** 0.76 * 0.74 *

1995 0. 86 ** 0.63 0.54 0.53

1997 0. 85 ** 0.46 0.82 * 0.85 **

P< 0.05 ** P< 0. 01


0 ..o 0Do) • dutas xul l uuays g1daU' alum ' xuwJO% puu papoo[ d% 3

a E wr) OM N NNO '

n OO

000 Tt ON O p

lA Q

w

P . 8m"-"" T E N

cNtrcz

0 QO > E T

Q a0

7 r' ro .(

2 e. >,

n111 O t)

Ellin.

111E

0

M

3O :

3 w

b a O

nC:om „ '^

mmnn

ii!3

Q e• kr) > Mo rn o

CI.

I.

03 0

4. j---

z. ....„ix,

r,i

a) D

v XX.

0 6.bqi o

o

A ason c

Q b .00 el.)

r ., ' Ft' ,—' W

aa\ i v c

11COQa Lm a ,

6

g

N O v, O , O v, 00 v O N 00 '' t O ¢ criM M N N •-+ —, N N N ^'

r-i

ma) Ilujuiug XEG iad s253 viaf}u2ts ..to 6on

w


35 -

MI Rainfall A3630 -

Max. Temp.

25 - U

30 °

E g.20 •

cu

24 F"15 -

10 18

5 -12

1 r II .1 •10 - 1.k.*..1*.. k k.A•* k..A..,.*.. frk..,..+..,.A.

B ' 18

ctk•+ k....k.'' +•• 1 -

163 8 . •• kr .'

14

36 6

12szA

o1* + t l

i l n

it'

A°- Treeholes with Water 6c

1 2 • 0— Mean Water Depth

4

Z2

0 •0

60

kC -

5

Ae. sierrensis50

0— Or. signifera4 c,

i 40 ' Q

a i ion

0.cn

W

L. 20 ' i

2 ti

410 ' 1

A i

0 h'.+•i A +.+..

A 0Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure 2. A. Mean maximum temperatures and total precipitation recorded at a United States Weather service atPotter Valley, CA during weekly sampling periods in 1997.

B. Numbers of natural tree holes holding water and mean tree hole water depth during 1997.C. Mean numbers of Or. signifera and Ae. sierrensis eggs per ovitrap day during weekly trap periods from

March 26 to December 1, 1997.


of the study. Oviposition was negatively correlated However, when yearly ovipositional totals of the twowith total monthly rainfall in 1991 and 1993. species were compared for all of the stations there was

Orthopodomyia signifera oviposition was also not a significant correlation for 1991 ( R= - 0.43, P=

largely restricted to periods when most, or all, of the 0.29), 1993 ( R= 0. 14, P= 0.74), 1995 ( R= - 0. 52, P=

monitored tree holes were dry. In 1991, 1993, and 1995, 0. 18), or 1997( R= 0. 00, P= 1. 00), indicating females ofoviposition was first detected after 30% of the tree holes each species chose oviposition sites independently ofhad dried, but most of the oviposition( 78%) occurred the other species. This conclusion was supported by anafter all of the tree holes had dried completely( Fig. 1). analysis of the distributions of eggs on individual

In 1997 oviposition peaked on August 14, coinciding ovitrap liners. During the four yearly Or. signiferawith the date when the fewest number of tree holes in the oviposition periods 394 liners were examined for the

woodland held water( Fig. 2). Mean monthly tree hole presence of mosquito eggs. A total of 36. 3% of the

water depth and the mean number of tree holes holding liners was positive for Or. signifera eggs and 87. 5%

water were negatively correlated with oviposition in were positive for Ae. sierrensis eggs. The frequencies

1991, 1993, and 1997( TABLE 1). of liners with Or. signifera eggs were not significantlyThere were marked differences in the temporal different(

x2=

1. 83, P> 0. 17) on liners with( 38. 8%) or

patterns and numbers of eggs oviposited into the without( 28. 6%) the presence of Ae. sierrensis eggs.

ovitraps by Or. signifera and Ae. sierrensis females In the 1994 vertical distribution study, Or.signifera

during each year of the study. Yearly Ae. sierrensis females laid a total of 1, 293 eggs into the ovitraps

ovipositional totals from the eight ovitraps ( with between June 6 and October 4. All of the eggs were

inclusive dates of occurrence in parenthesis) were oviposited between 3. 0 and 10.0 m above ground( Fig.38, 725( May 17 to October 23) in 1991, 25, 358( May 4 4), and most( 74%) were oviposited in the 5. 0 and 7.5 m

to October 20) in 1993, and 65, 054 ( April 25 to ovitraps. The frequencies of positive liners were more

November 9) in 1995. During these years all ten of the evenly distributed among the higher level ovitraps. Themonitored tree holes were partially filled with water 3. 0, 5. 0, 7. 5, and 10.0 m ovitraps had six, seven, six, and

when Ae. sierrensis oviposition was first detected and five Or. signifera positive liners, respectively, over the79% occurred before all of the tree holes had dried. In course of the season. Although the number of Or.

1997 all Ae. sierrensis oviposition occurred between signifera eggs found at middle( 3. 0 and 5. 0 m) and high

April 16 and September 24, totaling 27, 761 eggs( Fig. ( 7. 5 and 10.0 m) heights above ground were not

2C). More than 97% of theAe. sierrensis eggs were laid significantly different from each other, all of thebefore the tree holes reached their mean minimum water elevated ovitraps collected significantly( P< 0.05) moredepth on August 14. Conversely,> 84% of Or. signifera eggs than ovitraps at low ( 0.2 and 1. 0 m) levels above

oviposition occurred after August 7. During the entire ground. In the same study, Ae. sierrensis femalesfour year study period, Ae. sierrensis oviposition was oviposited a total of 21, 538 eggs into the ovitraps

detected an average of 73 days before and 21 days after between May 24 and October 25. There was relativelythe yearly oviposition periods of Or. signifera. The little variation in the percentage of Ae. sierrensis

magnitude ofAe. sierrensis oviposition into the ovitraps oviposition that occurred at each height. Ovitraps at 0.2,

was> 10 times that of Or. signifera. No other species of 1. 0, 3. 0, 7. 5, and 10.0 m above ground each had between

mosquito oviposited into the ovitraps during the study. 16% and 20% of the total oviposition. The lowest

percentage( 10%) occurred at 5. 0 m above ground. A

Spatial Distributions of Oviposition by Or. signifera Kruskal-Wallis ANOVA by ranks indicated there wasAll of the Or. signifera eggs found in the ovitraps no significant difference in the number ofAe. sierrensis

were above the waterline adhering to the ovitrap liners. eggs oviposited at any height( P> 0. 65).The external membranous sheath( Chapman 1964) was

nearly always oriented away from the oviposition Larval Surveys of Natural Tree Holes

substrate. Positive liners ( n= 168) held between 1 and A total of 28, 129 immature mosquitoes was

758 eggs. Most( 54%) had totals ofbetween 30 and 150 collected from the ten natural tree holes surveyed and

eggs. More than 27%, however, had< 30 eggs and 4 held identified to species during the four year seasonal study.only 1 Or. signifera egg. The vast majority of the larvae (> 96%) were Ae.

While all of the ovitraps were utilized by Or. sierrensis, the only species identified during surveyssignifera during each year of the study( Fig. 3), females conducted from August until April. Eggs of this species

laid more eggs at station 8( 35% of the total) than at any hatched in all of the tree holes as soon as they wereother location. Station 8 also had the lowest flooded by rain each fall. Nearly all overwintering Ae.ovipositional total of any location for Ae. sierrensis. sierrensis immatures had emerged by the end of May


oo

oNch1r;

e•'

pa

NC.'

S: r,-,-

BCT C7, CT CT Q1 1 .... g I

4

Ell I •::::'.:::::::;:.:1 - •'..........:. . . ....-.. ....... ..:.. .-.::..... ...:

064a'

o E.5

i.o

vsN 1

7)

o

51)

o;

c)

c)

11

4:s,Ri

0 VD N 00 d-0 0 0 0 0 0N 0 00 VD ytt N Cfi

i

2a

uosuas lad s353 olplbsow 4


40

Or. signifera

A Ae. sierrensis

30

0

6a00 -ao 20

to

0

0.2 1. 0 3. 0 5. 0 7. 5 10.0

Height above Ground( m)

Figure 4. Percentages of Or.signifera ( n= 1, 293 eggs) and Ae. sierrensis( n= 21, 538 eggs) oviposition that occurred

at each of six heights above ground during weekly trap periods between May 16 and November 2, 1994.

each year but small summer generations occurred northeast of the study site. This tree hole had a large

following rainfall in seven tree holes in 1991 ( June 30- volume( 9. 3 liters) and maximum depth( 52. 1 cm) with

July 18), three tree holes in 1995 ( July 20- August 9), a restricted, vertical opening. Intermittent observationsand five tree holes in 1997( May 28- July 4 and August indicated rain filled the tree hole each winter, but the

17- September 4). Culiseta incidens Thomson small opening reduced evaporation, which prevented

immatures were found in five tree holes in 1993 ( May the habitat from drying completely during summer. The19- July 14), four tree holes in 1995( May 23- July 5), and tree hole held opaque, black water at each observation

five tree holes in 1997 ( April 23 to July 16). Culex ( pH= 9. 03, true color= 288,000 Pt-Co units on July 24,stigmatosoma Dyar immatures occurred in four tree 1997). Larvae of Or. signifera and Ae. sierrensis were

holes in both 1993( July 1- July 21) and 1997( June 12- identified from the tree hole during each year of theJuly 31). These data indicate that three species of study.

mosquitoes utilized the tree holes as breeding sitesduring summer months that overlapped the oviposition Response of Or. signifera Adults to Carbon Dioxide-

periods of Or.signifera. However, Or. signifera larvae Baited Fay Traps and CDC Light Trapswere never found in any of the tree holes during the Carbon dioxide- baited Fay traps have beenstudy. determined to be an accurate method for surveillance of

The closest natural tree hole known to hold Or. adult Ae. sierrensis( Garcia et al. 1989, Washburn et al.

signifera immatures was located 1. 58 m above ground 1992). Although they collected large numbers of Ae.on a Pacific madrone in the foothill woodland ca. 90 m sierrensis, they were not attractive to Or. signifera


adults during three years of operation ( TABLE 2). possibility that females may have used photoperiod as aStandard CDC light traps failed to detect the presence of cue affecting the cessation of oviposition cannot beadult Or. signifera or Ae. sierrensis in the woodland ruled out since eggs were last found during earlyduring 1997. October in four of the five studied years. It is less likely

that minimal rainfall or the absence of rain directlyDISCUSSION affected ovipositional periodicity. Although signifi-

cantly lower rainfall totals preceded the earlyUnlike Ae. sierrensis, the eggs of Or. signifera lack oviposition period in 1991, the ends of the oviposition

a resting state for survival through dry- summer periods did not coincide with the onset ofrain in 1995 or

conditions. Female Or. signifera that oviposit into tree 1997.

holes temporarily filled with water may increase the Although most Or. signifera oviposition occurred

chances that their progeny will not complete after all of the monitored tree holes had dried > 20%

development before the habitat dries out. The data occurred during periods when one to seven of theshown in Figures 1 and 2 largely support Zavortink' s monitored tree holes still held water. Despite this

1968) conclusion that in California the activity periods overlap Or. signifera females avoided ovipositing intoof Or. signifera females occur after most rain- filled tree the temporary tree holes; conversely, Culiseta incidensholes have dried, thereby excluding these temporary and Culex stigmatosoma females utilized many of thehabitats as oviposition sites. This periodicity is an temporary tree holes for oviposition. These generalistsadaptive trait for Or.signifera females in California that are attracted to a wide variety of habitats for breedingdoes not occur in some other areas of North America. ( Colwell et al. 1995), but both species completelyFor example, Bradshaw and Holzapfel ( 1984) found avoided use of the ovitraps for deposition of egg rafts.that Or. signifera oviposition began in late April in These data suggest that cues associated with the ovitraps

Leon County, Florida, a location where rain caused the and tree holes were used by female mosquitoes selectingmajority of the tree holes examined to hold water oviposition sites. Bradshaw and Holzapfel ( 1988)

through the summer months. Under the dry- summer previously reported that water from permanent rain-conditions present in northern California oviposition filled tree holes at their Florida study site tended to havewas not detected until June or July, six to twelve weeks higher optical density, pH, conductivity, and tannin-after the observed onset of ovipostion in Florida. Grant lignin concentration than water from intermittently1953) and Baerg( 1968) concluded that Or. signifera flooded tree holes. They demonstrated that Or.

larvae do not pupate in California until high outdoor signifera females avoided oviposition into jars holdingtemperatures have been reached. Among the physical water from tree holes fated to dry out, but they did layfactors reported in TABLE 1, maximum air temperature eggs in jars with water from permanent tree holes. In the

was also the most likely to have had a cause- and-effect present study, Or. signifera females exhibited similarrelationship with the periodicity of oviposition. This behavior. They consistently avoided ovipositing intofactor was positively correlated with oviposition during temporary tree holes which had exposed, horizontaleach year of the study; and following significantly openings and held light-colored water with acidic to

higher temperatures in 1991, the oviposition period slightly basic pH. However they readily laid eggs in theoccurred ca. one month earlier than in other years. The ovitraps apparently after receiving cues which

TABLE 2. Seasonal totals of tree hole mosquitoes captured in two carbon dioxide-baited Fay traps and three CDClight traps operated at the study site.

Total Total Number of Adults Captured

Trap Or. signifera Ae. sierrensis

Year Trap Type Trap Period Days 9 d' 9 d'

1991 CO2-baited Fay 4/ 03- 11/ 18 64 0 0 1, 543 2, 081

1995 CO2-baited Fay 3/ 25- 11/ 16 68 0 0 4, 554 5, 695

1997 CO2- baited Fay 4/ 04- 10/ 21 68 0 0 2, 132 5, 242

1997 CDC light trap 6/ 26- 10/ 08 45 0 0 0 0


mimicked those presented by permanent rain- filled tree sierrensis larval development does not begin until fallholes. Each ovitrap provided a darkened, sheltered rains inundate resting eggs. Data from the present studycavity and held dark-colored, basic water. A similar indicate Or. signifera larvae had a one to four monthassociation of Or. signifera with tree holes holding window for development each year during periodshighly basic water has been reported many times( e. g., when Ae. sierrensis larval populations were at theirBohart 1950, Chapman 1964, Mitchell and Rockett minimum.

1981, Woodward et al. 1988). Several researchers ( Scholl and DeFoliart 1977,In contrast, Ae. sierrensis females readily Sinsko and Grimstad 1977) have reported that Ae.

oviposited both into tree holes that dried completely triseriatus showed preference for oviposition nearlater in the summer and into the permanent tree hole ground level, while Ae. hendersoni preferred to ovipositlocated near the study site. Most of the oviposition into at above- ground locations in the same easternthe ovitraps occurred as tree hole water levels declined woodlands. Figure 4 indicates that Or. signiferaafter the end of the rainy season ( Fig. 2). During this females may prefer to oviposit at locations aboveperiod females have access to the maximum number of ground. However since Ae. sierrensis females showedavailable developmental sites and they can place their no ovipositional preference relative to height aboveeggs at locations above the waterline that will be flooded ground, Or. signifera females would not avoidby future rains. Since water levels fluctuate in rain- interspecific competition by depositing their eggs atfilled tree holes regardless of whether they dry specific heights. Although host-seeking female Or.completely during summer, Ae. sierrensis eggs hatch signifera might find greater numbers ofbirds well aboveand develop in both temporary and permanent habitats the ground, it is less apparent why females seekingZavortink 1985). During the late summer overlap of oviposition sites might also prefer locations high in

the oviposition periods of Ae. sierrensis and Or. trees. It is possible that most permanent rain- filled treesignifera, most of the eggs of both species were holes in the region occur well above ground. Copelandprobably laid into tree holes that held water all year. and Craig ( 1992) positively associated Or. signifera

In the more diverse tree hole mosquito populations with " elevated deep rotholes" at their

communities of eastern North America, the permanence Indiana study site and concluded these types of treeof the aquatic habitat( Bradshaw and Holzapfel 1988), holes only occurred above ground at the sites of woundspredation ( Bradshaw and Holzapfel 1984) and to the bases of tree branches. In a northern Californiainterspecific competition ( Copeland and Craig 1990) survey, all of the rain- filled tree holes positive for Or.are considered to be important factors affecting the signifera occurred at above ground locationsculicid complexes that occur within individual tree ( Woodward et al. 1988).holes. For example, Copeland and Craig ( 1992) In summary, this analysis indicates the ovipositionconcluded that Ae. hendersoni ( Cockerell) used periods ofboth northern California tree hole mosquitoesdifferential oviposition based on tree hole water quality are adapted to weather patterns associated with theto avoid competition with Ae. triseriatus ( Say). Mediterranean climate. Peak oviposition periods of Or.Northern California tree holes lack a predatory trophic signifera occurred during August and September,level ( Woodward et al. 1988) but intraspecific temporally excluding most of the eggs from drought-competition for limited food resources is considered to susceptible habitats. Most Ae. sierrensis ovipositionbe a major force regulating larval populations of Ae. occurred from late spring to mid- summer when thesierrensis ( Hawley 1985, Washburn et al. 1991, largest numbers of tree holes with declining water levelsColwell et al. 1995). Despite the known resource were available as oviposition sites. The segregation oflimitations in regional tree holes, data from the present the species into temporary ( Ae. sierrensis only) andstudy did not indicate that Or.signifera females avoided permanent ( both species) habitats results from theinterspecific competition by ovipositing their eggs at periodicity of the oviposition of each species, as well aslocations that were unattractive to Ae. sierrensis. cues used by Or. signifera females for oviposition siteInstead, the ovipositional periodicity of each species, selection. The study did not produce any evidence thatcoupled with the hatching characteristics of the eggs, interspecific competition affected the spatialmay result in modulation of larval competition between distribution of oviposition by Or. signifera females, butthe species. Barrera ( 1996) has shown that temporal interspecific larval competition may be modulatedhabitat partitioning can reduce interspecific compe- temporally. The data also indicate that Californiatition between tree hole mosquitoes. Orthopodomyia populations of Or. signifera possess a number of traitssignifera eggs oviposited during summer hatch and that may cause their distributions and densities to bebegin larval development within a few days; Aedes underestimated. Adults are nocturnal and not known to


bite humans ( Zavortink 1968). Although adults are Culicidae) and Chaoborids in relations to weather

occasionally collected in light traps ( Grant 1953: and predation. J. Med. Entomol. 21: 366- 378.

Zavortink 1968), in the present study they failed to Bradshaw, W. E. and C. M. Holzapfel. 1988. Drought

respond to traps using either light or CO2 as attractants. and the organization of treehole mosquito

The nature of the larval habitat in oak woodlands( deep communities. Oecologia 74: 507- 514.

tree holes with restricted openings) often causes them to Chamberlain, R. W., R. K. Sikes, D. B. Nelson, and W.

occur at above ground, cryptic locations. For example, D. Sudia. 1954. Studies of the North American

larval surveys of four isolated oak woodlands in Lake arthropod-borne encephalides. VI. Quantitative

County, California failed to locate any tree holes determination of virus- vector relationships. Amer.

holding Or. signifera larvae. The ovitraps described in J. Hyg. 63: 278- 85.this paper later detected the presence of Or. signifera at Chapman, H. C. 1964. Observations on the biology andall four locations ( D.L.W. and A.E.C., unpublished). ecology of Orthopodomyia californica Bohart.Similar ovitraps have detected the presence of Or. Mosq. News 24: 432- 439.signifera in woodlands in southern Mendocino County Colwell, A. E., D. L. Woodward, and N. L. Anderson.

J. O. Washburn, pers. comm.) and in Fresno County, 1995. Environmental factors affecting the westernCalifornia( Lewis and Tucker 1978). Since" tree hole" treehole mosquito ( Aedes sierrensis). Northwest

ovitraps are not widely used in California, these reports Sci. 69: 151- 162.

indicate Or. signifera populations may be more Copeland, R. S. and G. B. Craig, Jr. 1990. Habitat

widespread and abundant than is widely believed. The segregation among treehole mosquitoes ( Diptera:difficulties associated with locating permanent rain- Culicidae) in the Great Lakes region of the United

filled tree holes also have implications for the control of States. Ann. Entomol. Soc. Am. 83: 1063- 1073.

Ae. sierrensis. Tree hole surveys which do not locate Copeland, R. S. and G. B. Craig, Jr. 1992. DifferentialOr. signifera may be missing large numbers of Ae. oviposition by Aedes hendersoni and Aedessierrensis that co-habit permanent tree holes with Or. triseriatus ( Diptera: Culicidae) in response to

signifera. chemical cues associated with treehole water. J.

Med. Entomol. 29: 33- 36.

Acknowledgments Darsie, R. F., Jr. and R. A. Ward. 1981. Identification

and geographical distribution of the mosquitoes of

We thank Dr. Jan Washburn and Dr. Robert North America, north ofMexico. Am. Mosq. Cont.Colwell for reviewing an earlier version of this Assoc., Fresno, CA., 313 pp.manuscript. The technical assistance of Joann Fay, R. W. and W. H. Prince. 1970. A modified visualHaberthur and Terry Sanderson is gratefully trap for Aedes aegypti. Mosq. News 30: 20- 23.acknowledged. Garcia, R., A. E. Colwell, W. G. Voigt, and D. L.

Woodward. 1989. Fay- Prince trap baited with CO2REFERENCES CITED for monitoring adult abundance ofA edes sierrensis

Diptera: Culicidae). J. Med. Entomol. 26: 327-

Baerg, D. C. 1965. A study of the biology and ecology 331.

of the treehole mosquitoes of California. MS Grant, C. D. 1953. Notes on the occurrence of

Thesis, University of California, Davis, 66 pp. Orthopodomyia californica Bohart. Proc. Calif.

Baerg, D. C. 1968. Overwintering habits of Mosq. Cont. Assoc. 22: 73.Orthopodomyia californica Bohart ( Diptera: Greenberg, A. E.( ed.). 1985. Standard Methods for the

Culicidae). Mosq. News 28: 234- 235. Examination of Water and Wastewater. American

Barrera, R. 1996. Species concurrence and the structure Public Health Assoc., Washington, D.C., 1268 pp.of a community of aquatic insects in tree holes. J. Hawley, W. A. 1985. Population dynamics of AedesVector Ecol. 21: 66- 80. sierrensis. Pp. 167- 184 in Ecology of Mosquitoes:

Bohart, R. M. 1950. A new species of Orthopodomyia Proceedings of a Workshop( L. P. Lounibos, J. R.from California ( Diptera: Culicidae). Ann. Rey and J. H. Frank, eds.) Florida Medical

Entomol. Soc. Am. 43: 399-404. Entomology Laboratory, Vero Beach, FL., 579 pp.Bohart, R. M. and R. K. Washino. 1978. Mosquitoes of Lewis, L. F. and T. W. Tucker. 1978. Fabrication of

California. University ofCalifornia, Berkeley, 153 artificial tree holes and their performance in field

pp. tests with Aedes sierrensis and Orthopodomyia

Bradshaw, W. E. and C. M. Holzapfel. 1984. Seasonal signifera. Mosq. News 38: 132- 135.development of treehole mosquitoes ( Diptera: Mitchell, L. and C. L. Rockett. 1981. An investigation


on the larval habitat of five species of treehole sierrensis( Diptera: Culicidae) from California

breeding mosquitoes ( Diptera: Culicidae). Great tree holes with particular reference to parasiteLakes Entomologist 14: 123- 129. loads. J. Med. Entomol. 26: 173- 182.

Munz, P. A. 1965. A California Flora. Univ. Calif. Washburn, J. O. and E. U. Hartmann. 1992. CouldPress, Berkeley, 1, 681 pp. Aedes albopictus (Diptera: Culicidae) become

Papineau, E. A. 1984. Controlling Aedes sierrensis in established in California tree holes? J. Med.

Jackson County, a year-round program. Proc. Entomol. 29: 995- 1005.

Calif. Mosq. Contr. Assoc. 52: 63- 64. Washburn, J. 0., D. R. Mercer, and J. R. Anderson.Peyton, E. L. 1956. Biology of the Pacific Coast 1991. Regulatory role of parasites: Impact on

treet.ole mosquito Aedes varipalpus( Coq.). Mosq. host population shifts with resource availability.News 16: 220-224. Science 253: 185- 188.

Reisen, W. K. and W. C. Reeves. 1990. Bionomics and Washburn, J. 0.,D. L. Woodward, A. E. Colwell, andecology ofCulex tarsalis and other potential vector J. R. Anderson. 1992. Correlations of Aedesspecies. Pp 324- 325 in Epidemiology and Control sierrensis captures at human sentinels with CO2-of Mosquito-borne Arboviruses in California, baited Fay-Prince and duplex cone traps. J.

1943- 1987 by W. C. Reeves, Calif. Mosq. Vect. Amer. Mosq. Contr. Assoc. 8: 389- 393.Contr. Assoc., Sacramento, 508 pp. Weinmann, C. J. and R. Garcia. 1974. Canine


predominant host- seeking mosquito, Culex erythro- californicus ( Meyer) Sojak and S. acutus ( Muhl. ex.thorax Dyar, collected during the first two years( 1995- Bigel.) Love and Love. In 1997, the wetland was in the1996) after flooding the wetland is thought not to play third year of operation and the surface area wasa significant role in St. Louis encephalitis ( SLE) approximately 70% covered by vegetation ( Thullen ettransmission in southern California( Reisen et al. 1992b). al. 1998). The inlet and outlet marshes containedLarge populations of host-seeking Culex tarsalis bands of vegetation separated by open water. TheCoquillett were collected at the wetland during the water depth within the inlet and outlet marshes wassummer( approximately 1, 000 females per trap night in maintained between approximately 0.5 and 0. 6 m during1996: WEW, unpublished data). Culex tarsalis is the the summer. Water temperatures at the middle of theprimary vector of SLE and western equine encephalo- water column of the open water regions in the inlet andmyelitis( WEE) to humans in rural areas of southern and outlet marshes ranged between 21. 5° C and 26.5° Ccentral California( Reeves and Hammon 1962, Reeves during the summer.1990).

We report here on the efficacy of control efforts METHODS AND MATERIALSagainst mosquitoes associated with a 10 ha multipurpose,constructed wetland in southern California. Larval Treatment of the wetland was carried out using asurveys, trapping ofemerging adults, and collections of helicopter by a local pest control company under contracthost-seeking females by carbon dioxide- baited traps with the water district. Pelletized Bacillus thuringiensiswere used to evaluate the effectiveness of larvicides and var. israelensis ( Bti, Bactimos® pellets) or granularadulticides applied to the entire wetland. Bacillus sphaericus( Vectolex®CG) formulations were

applied using a hopper-spreader apparatus suspendedSITE DESCRIPTION below the helicopter. The two Bti treatments were made

in conjunction with an adulticide ( Pyrenone®; 6.0%The 10 ha demonstration wetland is located at pyrethrins, 60% piperonyl butoxide; application rate: 3

Eastern Municipal Water District' s Hemet- San Jacinto quarts/ 30 acres= 0.014 kg AI/ha) during mid-August.Regional Wastewater Reclamation Facility( HSJRWRF) The adulticide was applied by cold aspiration throughin San Jacinto, California. The demonstration wetland an hydraulic spray rig suspended below the helicopter.consisted of five treatment wetlands( inlet marshes 1- 5), These treatments were followed by approximatelya central pond, and two outlet wetlands( outlet marshes biweekly applications ofB.sphaericus until mid-OctoberA and B)( USBR, NB S, and EMWD 1994, Walton et al. and a final application in early November( TABLE 1).1997). The wetland received approximately one million Treatment rate for the bacterial larvicides was either 19gallons daily of secondary effluent from the HSJRWRF. kg/ha or 23. 6 kg/ ha. Because of restricted after-hours

The wetland was planted in autumn 1994 with two access to the site and safety concerns, treatments werespecies of bulrush, Schoenoplectus (= Scirpus} carried out during daylight hours between approximately


Effectiveness of Control Measures against Mosquitoes at aConstructed Wetland in Southern California

William E. Walton, Parker D. Workman, Louie A. Randall,Joshua A. Jiannino, and Yvonne A. Offill

Department ofEntomology, University of California, Riverside CA 92521 USA

Received 20 April 1998; Accepted 18 May 1998

ABSTRACT: The effectiveness of larvicide and adulticide treatments against mosquitoes at a constructedwetland in San Jacinto, California was assessed with larval surveys, trapping of emerging adults, andcollections ofhost- seeking females by carbon dioxide-baited traps. Bacillus thuringiensis var. israelensisBti, Bactimos®pellets) applied at a rate of 19 kg/ ha did not demonstrably affect Culex larval and emergent

adult populations. Larval populations in the seven marshes of the wetland decreased from approximatelyone third- fourth instar larva/ dip to undetectable levels following two applications of Bacillus sphaericusVectolex®CG) at a rate of either 19 or 23. 6 kg/ ha. The largest decline in the number of adult mosquitoes

emerging per day from vegetated regions of the wetland occurred after B. sphaericus treatments. The Culex

erythrothorax host- seeking population declined about 80- fold during September beginning three weeksafter the first treatment with B. sphaericus; however, the Culex tarsalis host-seeking population did notdecline abruptly until mid-October 1997. This result suggests that immigration of females from otherdevelopmental sites might be an important factor influencing the Cx. tarsalis host- seeking population atthe wetlands. Safety concerns required that insecticide applications were carried out during daylight hours,and two daytime applications ofadulticide( Pyrenone®) in early August were ineffective against mosquitoesresting in the thick vegetation.

Keyword Index: Culex, constructed wetlands, Bacillus, bacterial larvicides, bulrush.

INTRODUCTION Rapid growth of vegetation in constructed wetlands

quickly creates ideal conditions for mosquitoThe incorporation of alternative water reclamation development( Walton et al. 1996, Walton and Workman

facilities, such as constructed wetlands, into water 1998). Mosquito production enhanced by constructedresource management programs is likely to affect vector wetlands is likely to require increased mosquitocontrol efforts. In order to fulfill the increasing demand abatement efforts, particularly where human populationsfor water and wastewater treatment in densely populated continue to encroach on wetlands situated in previouslyarid regions, such as southern California, alternative agriculture- dominated regions.

water management strategies will be needed. California In 1997, encephalitis virus activity in westernhas approximately 200 water reclamation facilities that Riverside and San Bernardino counties and large host-recycle about 450,000 acre- feet annually ( McCarthy seeking adult mosquito populations at a multipurpose,1997). The number of water reclamation facilities is constructed wetland in San Jacinto, California promptedprojected to approximately double by 2000( McCarthy the undertaking of control efforts against Culex1997); most of the increase is projected to occur in populations. Seroconversion to SLE in sentinel chickensouthern California. Multipurpose, constructed wetlands flocks in western San Bernardino and Riverside countiesare expected to play an important role in water in early summer 1997 indicated a comparatively earlyreclamation programs. In addition to processing onset of annual virus activity in the region( cf. Emmonssecondary- or tertiary- treated effluent, these wetlands et al. 1990, Kramer et al. 1996). Marked increases in theprovide habitat for wildlife and a site for public education host- seeking and larval mosquito populations were alsoon issues related to water and wildlife conservation. observed at the wetland during 1997. Although the


predominant host-seeking mosquito, Culex erythro- californicus ( Meyer) Sojak and S. acutus ( Muhl. ex.

thorax Dyar, collected during the first two years( 1995- Bigel.) Love and Love. In 1997, the wetland was in the

1996) after flooding the wetland is thought not to play third year of operation and the surface area was

a significant role in St. Louis encephalitis ( SLE) approximately 70% covered by vegetation( Thullen ettransmission in southern California( Reisen et al. 1992b). al. 1998). The inlet and outlet marshes contained

Large populations of host- seeking Culex tarsalis bands of vegetation separated by open water. TheCoquillett were collected at the wetland during the water depth within the inlet and outlet marshes was

summer( approximately 1, 000 females per trap night in maintained between approximately 0. 5 and 0.6 m during1996: WEW, unpublished data). Culex tarsalis is the the summer. Water temperatures at the middle of the

primary vector of SLE and western equine encephalo- water column of the open water regions in the inlet and

myelitis( WEE) to humans in rural areas of southern and outlet marshes ranged between 21. 5° C and 26. 5° C

central California( Reeves and Hammon 1962, Reeves during the summer.1990).

We report here on the efficacy of control efforts METHODS AND MATERIALS

against mosquitoes associated with a 10 ha multipurpose,

constructed wetland in southern California. Larval Treatment of the wetland was carried out using asurveys, trapping ofemerging adults, and collections of helicopter by alocal pest control company under contracthost-seeking females by carbon dioxide-baited traps with the water district. Pelletized Bacillus thuringiensis

were used to evaluate the effectiveness of larvicides and var. israelensis ( Bti, Bactimos® pellets) or granular

adulticides applied to the entire wetland. Bacillus sphaericus( Vectolex®CG) formulations were

applied using a hopper- spreader apparatus suspendedSITE DESCRIPTION below the helicopter. The two Bti treatments were made

in conjunction with an adulticide ( Pyrenone®; 6.0%

The 10 ha demonstration wetland is located at pyrethrins, 60% piperonyl butoxide; application rate: 3

Eastern Municipal Water District' s Hemet- San Jacinto quarts/ 30 acres= 0.014 kg AI/ha) during mid-August.Regional Wastewater Reclamation Facility( HSJRWRF) The adulticide was applied by cold aspiration throughin San Jacinto, California. The demonstration wetland an hydraulic spray rig suspended below the helicopter.consisted of five treatment wetlands( inlet marshes 1- 5), These treatments were followed by approximatelya central pond, and two outlet wetlands( outlet marshes biweekly applications ofB. sphaericus until mid-OctoberA and B)( USBR, NB S, and EMWD 1994, Walton et al. and a final application in early November( TABLE 1).1997). The wetland received approximately one million Treatment rate for the bacterial larvicides was either 19

gallons daily ofsecondary effluent from the HSJRWRF. kg/ ha or 23. 6 kg/ ha. Because of restricted after-hoursThe wetland was planted in autumn 1994 with two access to the site and safety concerns, treatments were

species of bulrush, Schoenoplectus {= Scirpus} carried out during daylight hours between approximately

TABLE 1. Larvicide and adulticide applications to the 10 ha demonstration wetlands

at the Hemet-San Jacinto Regional Water Reclamation Facility during1997.

Quantity QuantityDate Larvicide ( kg/ ha) Adulticide ( kg AI/ha)

August 13 Bactimos 19. 0 Pyrenone 0.014

August 20 Bactimos 19. 0 Pyrenone 0.014

August 29 Vectolex CG 19. 0

September 12 Vectolex CG 23.6

September 26 Vectolex CG 19.0

October 10 Vectolex CG 23. 6

November 3 Vectolex CG 19.0


10: 00 A.M. and noon. populations was indicated by a marked reduction in theIn order to assess whether larvicide and adulticide abundance, or by an abrupt change in population growth,

treatments had a demonstrable effect on mosquito of a particular stage in the life cycle. The rate of

populations, larval abundance in dip samples, population population change( per day) for larvae and for emergenttrends for emerging adult mosquitoes, and host-seeking adults was calculated as: ln( N+1)- ln( N)/t, where N isadult abundance in CO2- baited traps were examined. the average number of larvae per dip or the averageImmature mosquitoes were sampled by dipping along number of emergent adults/ m2/ week, x is sample date,

three transects in each marsh of the wetland. The three and t is the interval between samples in days. Larvae

transects were positioned in each marsh at successive were divided into two subpopulations: early instar larvaeopen water-vegetation interfaces from either the inlet ( stages I and II) and late instar larvae( stages III and IV).

weir or the southernmost outlet weirs. Samples were The abundance of larvae and emergent adults was

taken biweekly at five equally spaced stations along averaged for each of the seven marshes in the wetland byeach transect. At each station, three 400 ml dips were sample date. The predominant host- seeking mosquito,taken, combined in a concentrator cup( mesh opening: Cx. erythrothorax, found at the wetlands was typically200µ m), and preserved with 95% ethanol. Develop- under- represented in larval surveys ( Walton and

mental stage and abundance of immature mosquitoes Workman 1998) and, therefore, the efficacy of controlwere determined at 25X to 50X magnification using a measures against this species was assessed using adultsstereo dissecting microscope. Stage III and IV larvae collected in emergence traps. Because the entire marsh

were identified to species using Loomis ( 1959) and was treated, both untreated areas and replicated

Bohart and Washino( 1978). experimental units were lacking.The number of adults emerging from the wetlands Chemical and physical parameters related to water

was determined from collections using eighty- four, quality were routinely measured at the wetland.

0.25 m2

emergence traps. Collapsible emergence traps However, two parameters are particularly relevant towere constructed by affixing fiberglass window screen the results presented here: residual chlorine concentration

to a hinged wooden frame. The gently sloping sides of and bacterial density. Residual chlorine in the waterthe trap concentrated emerging insects into a wide- entering the wetland was measured using the iodometricmouth( 16 oz.) Mason jar fitted with a removable plastic method( APHA 1995). Coliform bacteria density in thefunnel. Quadrats ( mean number of Schoenoplectus influent water was measured using the fermentationshoots/

m2± SD: 265± 98) were established within the tube test and calculated as the Most Probable Number

seven marshes of the wetland and the bulrush was cut ( MPN) using the number of positive reactions in thejust above the water surface on June 26- 27. Emergence dilution series( APHA 1995).

traps were placed into the wetlands on July 1. Collectionswere made weekly from July 8 through September 26, RESULTS

except for the period between July 15 and July 22 whenno jars were placed on the traps. At collection, the jar Larval Populations

was removed from the apex of the trap and the funnel in Larval samples contained predominantly Cx.each jar was plugged with cotton. The jars were returned tarsalis. Late instar larvae of Cx. quinquefasciatus Say,to the laboratory where the collections were killed by Cx. stigmatosoma Dyar, and Cx. erythrothorax were

freezing and then enumerated under a stereo dissecting less frequent in dip samples. The abundance of third andmicroscope. Because Schoenoplectus grows rapidly fourth instar larvae in dip samples increased from fewerand could upset emergence traps, clipping of emergent than one larva per dip in April to an average of fourshoots was carried out twice each week. larvae per dip in May ( Fig. 1). The abundance of late

Host-seeking female mosquitoes at the San Jacinto instar larvae in the inlet marshes decreased in mid-June;

wetlands complex were collected using three carbon in four of the five inlet marshes, larval abundance

dioxide-baited traps. Mosquitoes were collected weekly declined by nearly two orders of magnitude. Larval

from May 2 through October 30 and biweekly from abundance in four of the inlet marshes on the next

February through April and from November through sampling date increased to levels observed in late May.December 1997. Traps were run overnight from For most dates during the summer, the average numberapproximately 15: 00 until 08: 00. Mosquitoes were of late instar larvae per marsh ranged between one and

identified to species ( Bohart and Washino 1978) and eight larvae per dip. In contrast to older larvae, the

counted using a stereo dissecting microscope at 12X abundance of young larvae was comparatively stablemagnification. from April through July, averaging between three and

Effectiveness of the treatments against mosquito seven larvae per dip( Fig. 1).


100 . Early instar larvae ( L1 & L2) 4/ 4 4 r/ j0—Late instar larvae ( L3 & L4)

Q.

1 0 -

0. 1 w

ii•Z

co<

0.01 I I I I I I I I I I I •

31- Mar 30- Apr 30-May 29- Jun 29- Jul 28- Aug 27- Sep

Date

0.2 — 11— Early instar larvae ( L1 & L2)

0 —Late instar larvae ( L3 & L4)

0. 15 —

d0. 1 —

A31.- --- mi\

c0.05 —

Icd

U0

ao

vs - 0.05 —

n.0

c -0.1 —

0

111.0. 15 —

0.2 —

0.25 —

0.3 I I I I I I I I I I 1

31- Mar 30- Apr 30- May 29- Jun 29-Jul 28- Aug 27- SepDate

Figure 1. Population trends of Culex larval subpopulations in dip samples from the HSJRWRF demonstrationwetland during 1997. Upper panel: abundance of larvae ( mean ± SE) in dip samples. The arrowsindicate treatments of bacterial insecticides. Open arrows are Bacillus thuringiensis var. israelensistreatments. Closed arrows are Bacillus sphaericus treatments. Lower panel: the rate of populationchange( mean± SE) for two larval mosquito subpopulations.


The abundance of both larval subpopulations per quadrat was about ten times greater than Cx. tarsalis

declined during August and late instar larvae were not and more than one hundred times greater than Cx.

collected from all marshes in late September( Fig. 1). In quinquefasciatus. Culex stigmatosoma adults were

mid-August, larval abundance declined by more than rarely encountered in samples. Culex erythrothorax

ten- fold and averaged only 0. 3 larva per dip. Larval emergence declined approximately 10- fold betweencounts in the inlet marshes were lower than in the outlet July 14 and July 28. Production for the second Cx.

marshes at the end of August. By mid-September, the erythrothorax generation in 1997 was between 200 to

abundance of late instar larvae was below detectable 400 individuals/m2/ week from late July until earlylevels in inlet marshes 2 and 5 and remained at this level September, and then declined about 10- fold duringin the last set of samples from late September. On September( Fig. 3).September 26, late instar larvae were absent from all The numberofCx. tarsalis females produced weeklyseven marshes. from vegetated quadrats in the wetland declined at a rate

Three periods of comparatively large negative of 7%/ d during the summer( Fig. 3). An abrupt decline

population change(<- 0. 1/ d) were observed in the late in adult emergence was not observed in early September;instar larval subpopulation during 1997. These large however, adult production was already very low at thedecreases in growth rate of the late instar larval end of August(< 10 females/m2/ week).

population occurred during early June(- 0. 122/ d), earlyAugust(- 0. 176/ d), and during September(<- 0. 118/ d: Host-Seeking Populations

Fig. 2). The early instar larval subpopulation exhibited In 1997, host- seeking populations of Cx.

comparatively large declines in growth rate in early erythrothorax and Cx. tarsalis ( Fig. 4) increasedAugust(- 0.229/d) and during September(- 0.084/d, - throughout April and May and reached annual maxima0.211/ d). However, the negative growth rate of the in June. The maximum number of host- seeking Cx.older larvae observed in early June was not observed for erythrothorax collected was nearly 33, 000 individuals/the young larval subpopulation; the subpopulation of trap night. The maximum number of host- seeking Cx.1st and 2nd larval instars declined at 3%/ d during the tarsalis collected was 4, 560 individuals/ trap night, abouttwo week period in early June when the abundance of one- tenth of Cx. erythrothorax.

older instar larvae declined appreciably. The Cx. erythrothorax host- seeking populationdeclined approximately 80- fold during September( Fig.

Adult Emergence 4). Two periods ofstasis in population numbers occurred

The number of Culex spp. emerging per square in 1997 ( Fig. 4). Catches of Cx. erythrothorax host-

meter of vegetated surface declined gradually in all seeking females fluctuated around a mean abundance ofmarshes between July 8 and August 15 ( Fig. 2). On 23, 700 individuals/ trap night from June 12 until the lastaverage, the number of mosquitoes emerging per unit week of August. After declining for a three- weekarea declined from approximately 1, 500 adults/

m2/

week period, the host- seeking population stabilized at anin early July to 300 adults/

m2/

week in early August. average of 475 females/ trap night for four weeksThe number of Culex emerging from the inlet ( September 18 through October 16).

marshes declined during the third week of August( Fig. The Cx. tarsalis host- seeking population attained a2). Mosquito production from the inlet marshes declined maximum in June and declined slowly throughout thefrom 250 to 300 adults/

m2/

week to 80 adults/m2/

week summer and early autumn( Fig. 4). The rate of decline

during the third week of August. Mosquito emergence for the host- seeking population in 1997 was 2. 6%/ d.

from inlet marshes 1, 2, 4, and 5 increased during early Unlike the Cx. erythrothorax host-seeking populationSeptember. Unlike the inlet marshes, adult production which declined abruptly in early September, Cx. tarsalisfrom the outlet marshes was nearly constant at 500 collections in carbon dioxide- baited traps fluctuated

mosquitoes/m2/

week during August. between 100 and 300 females/ trap night duringThe most marked declines in adult mosquito September and then declined abruptly in mid- October.

emergence occurred after mid- September. Adult Host- seeking females of both species disappeared inmosquito production decreased to about 25 mosquitoes/ early November.m2/ week between the second and third week of

September. This was the largest negative change in Water Qualitypopulation size (- 0.24/ d), nearly twice the largest Residual chlorine concentration was above the

average decrease observed earlier in the summer(- 0. 14/ limit of detection( 0.2 mg/ L) for the iodometric methodd: Fig 2). on five dates between April and September 1997.

The number of adult Cx. erythrothorax produced Residual chlorine concentrations were greater than 1


100001

III-- 2

6- 3

11000 ia0-5

o art—r-.

i_i A

100 B

Q

in-

E 10

w

1 I I 1 I 1 I I I 1 I I I

1- Jul 16- Jul 31- Jul 15- Aug 30-Aug 14- Sep 29- Sep

Date

b0. 15 —

0. 1 —

0.05 —et

A0 0

o- 0.05 — 1

0. 1 —a

a-0 - 0. 15 —

0.2 — N

cl - 0.25 -

0.3 i I i I i I I I I I i I

1- Jul 16- Jul 31- Jul 15- Aug 30-Aug 14- Sep 29- Sep

Date

Figure 2. Number of Culex spp. emerging from the five inlet marshes( 1- 5) and two outlet marshes( A and B) inthe HSJRWRF demonstration wetland during summer 1997. Upper panel: number of individualsemerging per week. Lower panel: the rate of population change ( mean ± SE) for the emergingmosquitoes.

mg/L on three dates: the first week of May, a two week DISCUSSION

period in June- July, and at the end of August( Fig. 5).Increases in residual chlorine concentration were Wetlands constructed for wastewater treatmentinversely related to bacterial abundance. Coliform present a significant challenge for vectorcontrol agenciesbacteria densities declined by three or more orders of in the arid southwestern United States. The nutrientsmagnitude when residual chlorine concentration was supplied by wastewater and the nearly year-roundgreater than 1 mg/L( Fig. 5). favorable growing conditions create dense stands of


10000 yerythrothorax

X-- Cx. tarsalis1000 —

a1 100 —

EW

10 —

1 =

0. 1

1- Jul 16-Jul 31- Jul 15- Aug 30- Aug 14- Sep 29- Sep

Date

Figure 3. The average number(± SE) of Culex erythrothorax and Culex tarsalis females emerging from the sevenmarshes in the demonstration wetland, San Jacinto, California during 1997.

100000 X Cx. tarsalis yerythrothorax

10000

oq

l000

Ippploo

to

1 i I I

1- Mar 31- Mar 30-Apr 30-May 29-Jun 29-Jul 28- Aug 27- Sep 27- Oct

Date

Figure 4. Culex tarsalis and Culex erythrothorax host- seeking populations collected by carbon dioxide-baitedtrapping at the demonstration wetland, San Jacinto, California during 1997. The arrows indicate

treatments of bacterial insecticides. Open arrows are Bacillus thuringiensis var. israelensis treatments.

Closed arrows are Bacillus sphaericus treatments.


4 100, 000,000chlorine

Q— bacteriat 10,000,000

a3 1, 000,000 0

100,000 a

0 2 10,000 yN

U •

1, 000

i10

0 1 1 i 1 1 1 1 I 1 i 1 1

1- Apr 1- May 31- May 30- Jun 30-Jul 29- Aug 28- Sep

Figure 5. Residual chlorine concentration and coliform bacterial density in influent water to the HSJRWRFdemonstration wetland, San Jacinto, CA during 1997.

wetland vegetation. Mosquito populations increased mounted application devices. Environmental factors

concomitantly( Walton, unpublished data) as vegetation such as temperature, salinity, and suspended solidsdensities increased and lateral growth of bulrush, influence the effectiveness and persistence of bacterial

particularly S. californicus, reduced the proportion of toxins ( reviewed in Walton and Mulla 1992). For

open water habitat at the San Jacinto wetland during the constructed wetlands receiving secondary- treatedinitial three years of operation( Thullen et al. 1998). In effluent, Bti is less effective in organically enriched1997, the adult host-seeking Culex spp. population water than is B. sphaericus( Mulla et al. 1990).

averaged nearly 30,000 females/ trap night during the Bti did not have a demonstrable effect on larval andsummer. The large populations of resident and migratory emergent adult populations at the San Jacinto wetland.birds that utilize constructed wetlands serve as potential Although larval and emergent adult populations declinedreservoirs of arboviruses( Reeves 1990). Rapid human appreciably in the inlet marshes after application ofBti,development in regions surrounding constructed larval and emergent adult mosquito populations in thewetlands can create an important public health concern outlet marshes were not affected by the treatments.especially when mosquitoes capable of vectoring disease, Because the entire wetland was treated with Bti, somesuch as Cx. tarsalis, can readily move from develop- other factor was presumably associated with the declinemental sites into human neighborhoods. in the larval and emergent adult populations in the inlet

Bacterial larvicides provide an alternative to marshes observed in August. The comparatively highchemical insecticides ormosquitocidaloilsformosquito levels of residual chlorine (> 1 mg/L) observed in thecontrol at constructed wetlands where effluent water influent water during late August were directly relatedquality is an important consideration. However, thick to marked declines in both the older larval subpopulationstands of vegetation and other environmental factors and the number ofadults emerging from the inlet marshes.limit the effectiveness of bacterial larvicides ( Walton A similar concurrent decline of mosquito populations inand Mulla 1992). Thick stands of bulrush limit the the inlet marshes with high levels of residual chlorine ineffective application of pelletized and granular the influent water was observed in June.formulations of the bacterial larvicides to constructed The declines in larval abundance and larvalwetlands. The width of vegetated regions in the San population growth rate during September were associatedJacinto wetland( e. g., nearly 83 m in places) precludes with B. sphaericus treatments. Larval abundancethe application of larvicides by backpack and truck- declined by 80% in samples taken two weeks after


treatment of the wetland with Vectolex. Two weeks failed to disperse farther than 2 km from the wetland,later, third and fourth instar larvae were below detectable and were concentrated only at the wetland and not atlevels throughout the wetland. The greatest negative other developmental sites within a 3 km radius of ourpopulation change for mosquito larvae occurred after study site( Walton et al. 1998).the B. sphaericus treatments. Even though residual The difference in the dispersal tendencies of the

chlorine was measured only once after August 29, high two dominant Culex may have a very importantcoliform bacterial abundance in the influent water during consequence for control programs that utilize B.

September indicated that residual chlorine did not cause sphaericus. Unlike Bti which contains multiple toxins

the decline in mosquito populations. Larval surveys and does not readily promote the evolution of resistanceindicated that B. sphaericus was effective against the in mosquitoes ( Georghiou and Wirth 1997), B.

prevalent species in dipper samples, Cx. tarsalis. sphaericus has a pair ofmosquitocidal toxins( Baumann

Bacillus sphaericus was also effective against the et al. 1991) which have been demonstrated in laboratorymost prevalent host- seeking mosquito in the wetland, studies( Georghiou et al. 1992, Rodcharoen and Mulla

Cx. erythrothorax. Culex erythrothorax larvae were 1994) to increase the risk for resistance. A recent studyunder- represented in larval surveys from adjacent ( Nielsen-LeRoux et al. 1997) indicated the involvement

experimental wetlands ( Walton and Workman 1998) ofmultiple mechanisms in low- level( e. g., approximatelyand in dip samples from the demonstration wetland. 30- fold resistance after 30 or more generations of

Emergent adults provided the best indirect measure of laboratory selection) versus rapid (< 8 generations)

the effectiveness of bacterial larvicides on the Cx. evolution of high levels(> 10, 000- fold) ofB. sphaericus

erythrothorax larval population. The largest declines in resistance in the Cx. pipiens complex. Different levels

the emergent adult population followed the Vectolex of resistance in Cx. quinquefasciatus to B. sphaericus

treatments. Furthermore, the Cx. erythrothorax adult have also been demonstrated in several field studies

host-seeking population declined nearly two orders of ( Sinegre et al. 1994, Rao et al. 1995, Silva-Filha et al.

magnitude after treatment of the wetland with B. 1995). Because host-seeking Cx. erythrothorax femalessphaericus. do not move very far from their natal developmental

The lack of a marked decline in the Cx. tarsalis site, and if males exhibit similar dispersal tendencies,

host-seeking population after treatment of the wetland this mosquito has a greater potential to evolve resistance

with bacterial larvicides suggests that immigration of to B. sphaericus toxins than does Cx. tarsalis. Out-

females from peripheral sources might be an important crossing of Cx. erythrothorax individuals from treatedfactor contributing to the maintenance of host- seeking areas with susceptible individuals from untreated sites

populations at the wetlands. Culex tarsalis adult is less likely than for Cx. tarsalis. Because of their

production from vegetated quadrats was already low greater tendencies for dispersal, Cx. tarsalis populations

approximately 5 females/m2/

week) in late August, and from Bacillus-treated wetlands are likely to mix to aeven though the number of Cx. tarsalis adults emerging greater extent with susceptible populations from other

from the wetland declined after the larvicide treatments, developmental sites that either are untreated or where an

the host- seeking population did not decline abruptly alternative larval control is utilized. Cross resistance to

until October. Culex tarsalis adult female populations other compounds used for mosquito abatement ( e. g.,decline naturally during autumn. The timing of the Bti, methoprene, etc.) is unlikely( Nielsen- LeRoux et al.decline for female abundance at the San Jacinto wetland 1997).

was intermediate to the phenologies observed in New Our analyses suggest that B. sphaericus was

Jersey light trap samples from the cooler, more northern effective against Culex at this wetland; however, the

Central Valley and from the hotter, agricultural valleys effectiveness of the larvicide treatments should be

in southern California( Reisen and Reeves 1990). Mark- interpreted cautiously. The timing of the declines forrecapture studies carried out during 1995 indicated that the larval populations, for emergent adult Cx.

Cx. tarsalis host-seeking females could move two or erythrothorax, and for host- seeking adult Cx.more kilometers in one night and were also produced in erythrothorax are suggestive that bacterial larvicides

large numbers at three sites surrounding the wetland affected the mosquitoes. The results might be viewed as

WEW, unpublished data). The maximum distances equivocal because the declines in the mosquito

moved by marked female Cx. tarsalis have been populations occurred towards the end of the season and

estimated at between approximately 10 and 40 km there were no untreated populations against which to

Bailey et al. 1965, Dow et al. 1965, Reisen et al. I992a). compare the effects of the larvicides. Clearly, replicatedIn contrast to Cx. tarsalis, Cx. erythrothorax host- plots, untreated controls, and posttreatment samples

seeking females moved an average of 0.5 km per night, closer to application dates( e. g., within 48 hours for B.


sphaericus treatments) would facilitate a less ambiguous in wetlands( Levy 1997). There is a need to reevaluatestatement of the efficacy of bacterial larvicides at the wetlands operations procedures( e.g., changing accesswetland. Biweekly larval surveys, intended for policies so that adulticiding can take place after dusk,interannual comparisons of larval populations, might establishment of effective vegetation management

not detect a short- lived ( i.e., < 2 weeks) reduction in strategies) so that the public health and water needs oflarval abundance by the bacterial larvicides. Short- humans residing near man-made wetlands are fulfilled.lived control of mosquito larvae by bacterial larvicides Our results suggest that B. sphaericus may providehas been observed in other organically enriched habitats. some measure of mosquito control in large- scaleBacillus sphaericus reduced Culex larval populations in multipurpose constructed wetlands that receive

treated plots relative to untreated plots for one to three organically enriched wastewater.weeks in Schoenoplectus{= Scirpus} - Typha wetlands

Yoshimura et al. 1996) and in dairy wastewater lagoons AcknowledgmentsMulla et al. 1988, Binding et al. 1996). Bacillus

sphaericus was however effective for three weeks or This study was supported by Special Funds forlonger against mosquitoes in catch basins ( Siegel and Mosquito Research from the Division of AgricultureNovak 1997) and in polluted water( Mulla et al. 1997). and Natural Resources of the University of CaliforniaYet, environmental factors ( e.g., precipitation) had a and by a grant from the Academic Senate at U.C.-marked effect on longevity of control in highly polluted Riverside. We appreciate the continued cooperation ofmosquito developmental sites in Thailand( Mulla et al. Eastern Municipal Water District and the U. S.1997). Bti and B. sphaericus have proven only partially Geological Survey. We thank Dr. G. Smith( Universityeffective against Culex at treatment wetlands in Arizona of Wisconsin- Whitewater) for clarification of bulrushbecause thick vegetation inhibited penetration of the classification and Stella Denison ( EMWD) forlarvicides ( Levy 1997). Nevertheless, trends for the providing additional information regarding mosquitoemergent adult populations at the San Jacinto wetland abatement operations and water quality data collectedindicated that significant, short- lived reductions in larval by the EMWD. We benefited from discussions with Dr.abundance between sampling dates did not occur. M. C. Wirth and P. DeChant. We thank Dr. P. Walton

The projected increase in the use of constructed for review of the manuscript.wetlands in regions of rapid human population growthis likely to cause a greater need for intensified vector REFERENCES CITEDcontrol. A combination of integrated pest management

strategies was needed to reduce both the larval APHA. 1995. Standard Methods for the Examination

populations and the large host-seeking populations at of Water and Wastewater. 19th ed. Am. Pub. Hlth.

the San Jacinto wetland. The inability to effectively Assoc., Washington, D.C.

control the adult mosquito population, and continued Bailey, S. F., D. A. Eliason, and B. L. Hoffmann. 1965.recruitment into aquatic stages of the life cycle from the Flight and dispersal of the mosquito Culex tarsalis

resident adult population, slowed mosquito abatement. Coquillett in the Sacramento Valley of California.In order to reduce mosquito populations, larviciding Hilgardia 37: 73- 113.

and effective adulticiding needed to be carried out Baumann, P., M. A. Clark, L. Baumann, and A. H.

concurrently. Daytime application of adulticides was Broadwell. 1991. Bacillus sphaericus as a mosquitoineffective; adulticiding would have been more effective pathogen: properties of the organism and its toxins.

during the two to three hours after dusk when the Microbiol. Rev. 55: 425- 436.

mosquito adults were most active. Binding, P. L., A. R. Pfuntner, and M. S. Dhillon. 1996.The addition of larvivorous fish to the wetland and Field trials of Vectolex granules ( Bacillus

the establishment ofan effective vegetation management sphaericus) in southern California dairy wastewaterprogram will greatly aid vector control efforts. Whereas, lagoons. Proc. Calif. Mosq. Vector Contr. Assoc.integrated control measures using mosquitofish and 64: 118- 122.

source reduction were effective in an urban wetland in Bohart, R. M. and R. K. Washino. 1978. Mosquitoes ofsouthern California (Pelsue 1986), wetland managers California. Div. Agric. Sci. Univ. Calif. Berkeley,have been reluctant to stock mosquitofish into CA, 153 pp.constructed wetlands because of concerns related to Dow, R. P., W. C. Reeves, and R. E. Bellamy. 1965.clogging of pipelines moving reclaimed water or, after Dispersal of female Culex tarsalis into a larvicidedseveral months of effective mosquito control, area. Am. J. Trop. Med. Hyg. 14: 656- 670.mosquitofish were ineffective after the vegetation filled Emmons, R. W., D. V. Dondero, M. M Milby, L. T. Hui,


R. A. Murray, J. E. Tueller, F. Ennik, B. A. Wilson, larvae. J. Med. Entomol. 34: 321- 327.

L. Pitstick, J. L. Hardy, S. B. Presser, W. C. Reeves, Pelsue, F. W. 1986. Integrated pest management of an

L. Barrett, and M. S. Ascher. 1990. Surveillance urban swamp. Proc. Calif. Mosq. Vector Contr.for arthropod-borne viral activity and disease in Assoc. 52: 60- 61.

California during 1989. Proc. Calif. Mosq. Vector Rao, D. R., T. R. Mani, R. Rajendran, A. S. Joseph, A.

Contr. Assoc. 58: 1- 11. Gajanana, and R. Reuben. 1995. Development of

Georghiou, G. P., J. I. Malik, M. Wirth, and K. Sainato. a high level of resistance to Bacillus sphaericus in

1992. Characterization of resistance of Culex a field population of Culex quinquefasciatus from

quinquefasciatus to insecticidal toxins of Bacillus Kochi, India. J. Am. Mosq. Contr. Assoc. 11: 1- 5.sphaericus( strain 2362). Pp. 34- 35 in Univ. Calif. Reeves, W. C. ( ed.) 1990. Epidemiology and ControlMosquito Control Research Annual Report. Univ. of Mosquito-borne Arboviruses in California,

Calif., Berkeley, CA. 1943- 1987. Calif. Mosq. Vector Contr. Assoc.,Georghiou, G. P. and M. C. Wirth. 1997. Influence of Sacramento, CA, 508 pp.

exposure to single versus multiple toxinsofBacillus Reeves, W. C. and W. McD. Hammon. 1962.

thuringiensis subsp. israelensis on development of Epidemiology of the arthropod- borne viral activityresistance in the mosquito Culex quinquefasciatus encephalides in Kern County. Univ. Calif.

Diptera: Culicidae). Appl. Environ. Microbiol. Publications in Public Health 4, Univ. Calif. Press.

63: 1095- 1101. Berkeley, CA.Kramer, V. L., M. S. Ascher, B. Enge, L. T. Hui, S. Ball, Reisen, W. K., M. M. Milby, and R. P. Meyer. 1992a.

F. Ennik, K. Reilly, W. K. Reisen, D. A. Eliason, J. Population dynamics of adult Culex mosquitoes

Lin, R. A. Murray, M. M. Milby, J. L. Hardy, S. B. Diptera: Culicidae) along the Kern River, KernPresser, R. E. Chiles, and W. C. Reeves. 1996. County, California, in 1990. J. Med. Entomol. 29:Surveillance for mosquito-borne encephalitis virus 531- 543.

activity and human disease in California, 1995. Reisen, W. K., M. M. Milby, S. B. Presser, and J. L.Proc. Mosq. Vector Contr. Assoc. 64: 3- 7. Hardy. 1992b. Ecology of mosquitoes and St.

Levy, C. 1997. Mosquitoes and wetlands. SOVE Louis encephalitis virus in the Los Angeles Basin

Newsletter 28( 3): 13- 14. of California, 1987- 1990. J. Med. Entomol. 29:

Loomis, E. C.( ed.) 1959. A Field Guide to Common 582- 598.

Mosquitoes in California. Calif. Mosq. Control Reisen, W. K. and W. C. Reeves. 1990. Bionomics and

Assoc. Sacramento, CA, 27 pp. ecology of Culex tarsalis and other potentialMcCarthy, E. 1997. California Water Issues Briefing. mosquito vector species. Pp. 254- 329 in

Water Education Foundation. Sacramento, CA, 12 Epidemiology and Control of Mosquito-bornepp. Arboviruses in California, 1943- 1987 ( W. C.

Mulla, M. S., H. Axelrod, H. A. Darwazeh, and B. A. Reeves, ed.). Calif. Mosq. Vector Control Assoc.,Matamni. 1988. Efficacy and longevity ofBacillus Sacramento, CA.

sphaericus 2362 formulations for control of Rodcharoen, J. and M. S. Mulla. 1994. Resistance

mosquito larvae in dairy wastewater lagoons. J. development in Culex quinquefasciatus ( Diptera;

Am. Mosq. Contr. Assoc. 4: 448- 452. Culicidae) to the microbial agent Bacillus

Mulla, M. S., H. A. Darwazeh, and M. Zgomba 1990. sphaericus. J. Econ. Entomol. 87: 1133- 1140.

Effect ofsome environmental factors of the efficacy Siegel, J. P. and R. J. Novak. 1997. Field trials of

of Bacillus sphaericus 2362 and Bacillus Vectolex CG, a Bacillus sphaericus lorvicide, in

thuringiensis ( H- 14) against mosquitoes. Bull. Illinois waste tires and catch basins. J. Am. Mosq.Soc. Vector Ecol. 15: 166- 175. Contr. Assoc. 13: 305- 310.

Mulla, M. S., J. Rodcharoen, W. Ngamsuk, A. Tawatsin, Silva-Filha, M.-H., L. Regis, C. Neilsen-LeRoux and J.-

P. Pan- Urai, and U. Thavara. 1997. Field trials F. Charles. 1995. Low- level resistance to Bacillus

with Bacillus sphaericus formulations against sphaericus in a field- treated population of Culex

polluted water mosquitoes in a suburban area of quinquefasciatus ( Diptera: Culicidae). J. Econ.

Bangkok, Thailand. J. Am. Mosq. Contr. Assoc. Entomol. 88: 525- 530.

13: 297- 304. Sinegre, G., M. Babinot, J.- M. Quermel, and B. Gaven.

Nielsen- LeRoux, C., F. Pasquier, J.- F. Charles, G. 1994. First field occurrence of Culex pipiens

Sinegre, B. Gaven, and N. Pasteur. 1997. Resistance resistance to Bacillus sphaericus in southern France.

to Bacillus sphaericus involves different Proceedings 8th European Meeting Soc. Vectormechanisms in Culex pipiens( Diptera: Culicidae) Ecol., 5- 8 Sept. 1994, Barcelona, Spain. 8: 7.


Thullen, J. S., J. J. Sartoris, Jr., L. B. Barber, and D. E. and Natural Resources, Berkeley, CA.Salas. 1998. Expansion of transplanted bulrush Walton, W. E., M. C. Wirth, P. D. Workman, and L. A.and related water quality in the Hemet/ San Jacinto, Randall. 1997. Survival of two larvivorous fishesCA constructed wastewater treatment wetlands. in a multipurpose constructed wetland in southernEcol. Engineering. In press. California. Proc. Mosq. VectorContr. Assoc. Calif.

U. S. Bureau ofReclamation, National Biological Survey 65: 51- 57.

and Eastern Municipal Water District. 1994. Walton, W. E. and P. D. Workman. 1998. Effect ofMultipurpose Wetlands: Phase II/III Report. Final marsh design on the abundance of mosquitoes inDesign and Ongoing Research Investigations. San experimental constructed wetlands in southernJacinto, CA. Sept. 1994. California. J. Amer. Mosq. Contr. Assoc. 14: 95-

Walton, W. E. and M. S. Mulla. 1992. Impacts and fates 107.

of microbial pest-control agents in the aquatic Walton, W. E., P. D. Workman, and C. H. Tempelis.environment. Pp. 205- 237 in Dispersal of Living 1998. Dispersal, survirorship, and host selectionOrganisms into Aquatic Ecosystems( A. Rosenfield of Culex erythrothorax ( Diptera: Culicidae)and R. Mann( eds.). Maryland Sea Grant College, associated with a constructed wetland in southernUniv. of Maryland, College Park, MD. California. J. Med. Entomol. In press.

Walton, W. E., M. C. Wirth, P. D. Workman and L. A. Yoshimura, G., S. Wright, and J. Townzen. 1996.Randall. 1996. Ecology and biological control of Efficacy of Bacillus sphaericus at the Sacramentomosquitoes in a constructed wetland. Pp. 46- 51 in Regional Wastewater Treatment Plant demon-Mosquito Control Research Annual Report 1996. stration wetlands. Proc. Mosq. VectorContr. Assoc.University of California, Division of Agriculture Calif. 64: 124- 129.


Effect of Two Rice Culture Methods on the Seasonal Occurrenceof Mosquito Larvae and Other Aquatic Animals in

Rice Fields of Southwestern Korea

Dong- Kyu Lee

Department of Biological Sciences,Kosin University, Pusan 606- 701, KOREA

Received 19 May 1998; Accepted 14 September 1998

ABSTRACT: An ecological study has been performed on the community structures ofaquatic animals andon abiotic factors in organically and conventionally- farmed rice fields at Bulkyo, Bosong- gun, Chollanam-do during the rice growing periods between May 1995 and October 1996. The total numbers of aquaticinsect taxa in these fields were 25 species, 22 families in 10 orders. In 1995 and 1996, three and then fourmore species ofaquatic insects were collected in the organically- farmed rice fields than in conventionally-fanned rice fields. The common dominant species in both rice fields during the rice cultivation period was

Chironomus sp. The abundance of two vector mosquitoes, Anopheles sinensis and Culex tritaeniorhynchus,was lower in the organically- farmed rice fields as compared to the conventionally- farmed rice fields. Thepopulations of mosquito larvae in the organically- farmed rice fields might have been influenced by Chinesemuddy loaches, Misgurnus mizolepis. Coefficients of correlation between Chinese muddy loaches andabundance ofmosquito larvae showed negative correlations in An. sinensis(- 0.66) and Cx. tritaeniorhynchus

0.47). The average of the species diversity index for organically- farmed rice fields was almost twice0.62) as much as is of conventionally- farmed rice fields ( 0. 35). The average community similarity

between both types of fields was very low( 0. 33).

Keyword Index: Culex, organic farming, rice field, Anopheles, larvivorous fish.

INTRODUCTION enemies and various abiotic factors in ecosystems, and

the ecology of these mosquitoes.

Culex tritaeniorhynchus Dyar and Anopheles Traditional mosquito control strategies in

sinensis Wiedemann are distributed widely in Korea conventionally- farmed rice fields used mainlyand are vectors of Japanese encephalitis and of both adulticides which include fogging, aerosol sprays, andmalaria and inland filariasis, respectively. These larval control from agro- pesticides in Korea( Ree et al.

mosquitoes are particularly abundant in riceland 1981, Shim et al. 1995a and b). Each of these methods

agroecosystems where they breed in irrigated rice fields involves the application of relatively large amounts ofand associated lowland areas. Anopheles sinensis and insecticides into the environment of the rice ecosystems.

Cx. tritaeniorhynchus are the predominant blood- seeking These methods also may deposit insecticidal residues inmosquitoes in July and August, respectively( Shim et al. rice that can be taken inadvertently by consumers as1987, 1990, 1997); and, when present in large numbers, well as farmers. Therefore, environmental concerns

these species present a serious threat to human and have stimulated other farming methods to reduce the useanimal health through annoyance and as vectors of of insecticides for controlling pests in rice fields. Onedisease. Because of the threat that An. sinensis and Cx. approach that has recently gained popularity is organictritaeniorhynchus pose to human and animal health, farming. In Korea, 0. 08% of farmers cultivate several

there is a need to develop management programs that crops using organic farming, which utilizes organicwill maintain populations of these species at acceptably fertilizers instead of chemical fertilizers and which does

low levels. The development and implementation of not use pesticides( Paek 1992). The objectives of this

such programs require a detailed study of the relationship study were to establish the population abundance patterns

between environmental factors, including natural of vector mosquitoes and other aquatic animals and to


evaluate the associations between several environmental with an UVNIS spectrophotometer ( Model U 3210,factors and the vector mosquitoes in organically- farmed Hitachi) and an atomic absorption spectrophotometerrice versus conventionally- farmed rice. Model Smith-Hieftje 12, Thermo Jarrell Ash).

MATERIALS AND METHODS Sampling

Four paddies each from organically- farmed andStudy Areas

conventionally- farmed rice fields were randomlyThe study was carried out in the rice field( 150 ha) selected as study plots. The survey was carried out once

near an estuary, which was less than 2m above sea level, or twice a month from June to September in both 1995located in Bulkyo, Chullanam-do, southwestern Korea. and 1996 when the rice fields were flooded withIn this area, most farmers have used a conventional agricultural water. The collections were not carried outculture method employing chemical fertilizers and in other months because the water of the rice paddiesinsecticides. The insecticides were each applied one was completely drained out in late September. Thetime in June and July, and three times in August of 1995;

aquatic animals were randomly collected in each ricebut three times in early July, and each one time in late

paddy. Larval mosquito abundance was monitored byJuly, August, and September of 1996. The chemical taking 16 samples of water with a 355 ml dipper. Othernames of the insecticides used in the field were cartab aquatic animals and benthic samplings were also takenw.s. p. 50%, fenitrothion E.C. 50%, BPMC G. 2%, and by 32 sweeps with an aquatic net of 32 mesh/ cm and 33Quratel G. 3%. However, one farmer has cultivated rice cm in diameter. Fish samplings in the rice paddies weresince 1980 in the rice field of 3 ha, using an organic performed by using an aquarium- type vinyl fish trapculturing method. Each of the four organically- and ( 16cm dia. x 20cm long)( Yu et al. 1981), placed at twoconventionally- farmed rice paddies for the study were sites in each rice paddy.randomly selected from 16 and 850 rice paddies, and The samples of three different types from the riceeach paddy was a 1, 860 m2

and 1, 730 m2

rectangle, paddies were put into a small plastic concentrator withrespectively. The conventionally- farmed rice fields a fine mesh on the bottom. They were then transferredwere separated from the organically- farmed rice fields to 500 ml plastic bottles with water and transported backby a road and an irrigation ditch connected to a main to the laboratory in two ice chests. To minimizewatering canal and a draining canal. There was no physical damage and prevent predators from killing orevidence of habitat disruption in either one, which eating other animals during transportation, the chestsmight have created an effective ecological change in were chilled with ice packs. The aquatic animals werethe two different farming conditions. Water was supplied identified under a dissecting microscope and a stereofrom a reservoir and an underwater spring near the rice microscope, according to characters described in eachfields. A network of irrigation and drainage channels specific key of the corresponding taxa ( Merritt andallowed individual watercontrol for each rice paddy and Cummins 1984, Stehr 1987a and b, Usinger 1956,permitted quick flooding and draining by means of an Peterson 1982, Fitzpatrick 1983, Yoo 1986, Yoon 1988,efficient network of parallel ditches connected to them. Yoon and Ahn 1988a and b, Yoon and Kong 1990).Analysis of Water and Soil Data Analysis

The analytical methods given here were based on The physical and chemical characteristics of thethose given in Field and Laboratory Methods for General soils exposed to the two farming techniques wereEcology( Brower and Zar 1977). The salinity and pH of statistically compared by a Student' s t- test. In thewater in the two rice field types were measured using a present study, dominant species, dominance indices,portable salinity meter( NS- 3P, Merbabu Trading. Co.) and species diversity indices of the organically- farmedand a portable pH meter ( Model TS- 1, Suntex Ins.). rice field were compared with those of theOnce or twice every month, water depths of the rice conventionally- farmed rice field in order to analyze thepaddies were measured at 10 to 16 different spots. The aquatic animal community and its dynamics. Thesoil organic matter was analyzed using a method of loss dominance index is a method measuring the simplicityof organic carbon on ignition analysis( Cox 1976). The of each community and was here calculated withproportion of soil particles( e.g., clay and silt< 50 pm) McNaughton' s dominance index ( DI) ( McNaughtonwas determined by the USDA method. Also, standard 1967). This index is as follows: DI=(n1+ n2)/ N( N: themethods were used for examinations of ions in the total number of individuals in all the species; n1 and n2:water. To analyze water contained main ions, the the numbers of individuals of the 1St and 2nd dominantsampled water with soils in each rice field were examined species, respectively). The species diversity was


calculated using the Shannon-Wiener index( H')( Pielou depending on rice conditions, pesticides were sprayed1969) such as H'=- Sigma pi log pi, where, pi=n/ N. The five to six times during the growing season from June topi is the proportion of the total number of individuals September. The pesticides used were cartap andoccurring in species i. Finally, the community similarity carbofuran for the control of agricultural pests such aswas calculated using the Sorensen Coefficient ( CCs) the rice stemborers, the brown planthoppers, the riceBrower and Zar 1977). green leafhoppers, the smaller brown leafhoppers, and

the grass leafrollers. In the organically- farmed riceRESULTS AND DISCUSSION field, an organic fertilizer was supplied once in late May

before transplanting rice seedlings. The organic fertilizerRice culturing practices affected the seasonal occur- was made mainly from a mixture of chicken dung, rice

rence ofAn. sinensis and Cx. tritaeniorhynchus as well bran, and sawdust at the ratio of 6: 3: 1, respectively.as other aquatic animals. The observed schedule for rice However, no pesticide was applied in the rice fieldcultivation in the conventional and organic rice fields during the year.changed little during the study period, except for minor The physical characteristics of the soil and the rice-adjustments caused by the making of a road near the rice field water of two rice fields is shown in TABLE 1.fields in May, 1996. In both 1995 and 1996, plowing Some abiotic factors, such as water pH, water salinity,began at both fields in early May, and water was intro- depth of water, various water ions, organic matter, andduced between the end of May and the beginning of texture of soil were not significantly different betweenJune. The rice plants were grown from transplanting the organically- farmed and the conventionally- farmedrice seedlings. The seedling transplantation was corn- rice fields, except for the concentration of P2O5 in thepleted within a week in early June of both 1995 and water. The water salinity of both the organically-1996. farmed and the conventionally- farmed rice fields

In the conventionally- farmed rice field, chemical appeared somewhat higher than those of other rice

fertilizers were distributed in late May of both years to fields: 0.4° I and 1. 1%. , respectively. The reason

supply phosphate, potassium, and nitrogen with a might be that the rice fields are located near an estuary.herbicide before transplanting rice seedlings. Also, The concentration of phosphoric acid ( P2O5) in the

TABLE 1. Physical and chemical factors of soil and water in the

conventionally- and organically- farmed rice fields, 4- 16 replicates.

Rice Fields( Mean± S. D.)

Factor Conventional Organic

Water

pH 5. 6a'± 0. 10 5. 4a ± 0. 10

Salinity M.) 1. 1a ± 1. 20 0. 4a ± 0. 10

Depth( cm) 7. 2a ± 2. 10 6. 8a ± 1. 43

Soil

Soil Texture2(%) 59.9a - 73. 0a -

Organic Matter(%) 4. 1a ± 0.20 4. 4a ± 0. 20

pH 6.0a ± 0.60 6. 3a ± 1. 00

Na+ (ppm) 47. 5a ± 16. 57 52. 4a ± 17. 62

K+ (ppm) 27.0a ± 23. 37 27. 8a ± 25. 97Mgt,(

ppm) 3. 6a ± 1. 49 8. 1a ± 13. 63Cat+(

ppm) 0. la ± 0. 10 0. 1a ± 0. 08S042-(

ppm) 35.7a ± 6.46 32. 6a ± 7. 93P042-(

ppm) 1. 9a ± 0.41 2. 3a ± 0. 37

P2O5( ppm) 65. 5b ± 10.61 173. 0a ± 32. 53

Means within a row followed by the same letters were not significantlydifferent( P> 0.05; t- test).

2 Soil percentage indicates particles< 50 pm in diameter.


organically- farmed rice field water ( 173.0 ppm) was such as fish and insect predators, however, it wassignificantly higher than that of the conventionally- observed that in the conventionally- farmed rice field,farmed rice fields( 65. 5 ppm), due to a farmer' s use of the pesticide application acted as one of the mainchicken dung as a natural fertilizer. Chemical mortality factors. Other reports showed similar resultscomponents of the environment importantly affect the on the population dynamics of aquatic invertebratesabundance and distribution of species( Brower and Zar other than mosquitoes, as particularly related to1977). However, there was no evidence that the insecticide application in rice fields. Service ( 1977)

phosphoric acid affected the abundance and distribution studied mortalities of An. gambiae larvae and otherofmosquito larvae or other aquatic organisms. Ikemoto aquatic insects in rice fields in Kenya before and afterand Sakai(: ( 1979) reported that there was a positive spraying Dimecron [ 0, 0-dimethyl-0-( diethylanido- 1-correlation between the number of larvae and the chloro-crotnyl) phosphate] applied for the control ofconcentration of NH4- H in the water among the rice stem borer. Whereas before spraying there was acharacteristics of water such as temperature, water very rich and numerous aquatic fauna, exceedingly fewquantity, pH, DO content, and NH4 H concentration. live invertebrates remained after spraying. The larval

There are few reports on the population dynamics densities of An. gambiae 14 days after spraying wereof aquatic invertebrates other than mosquitoes and the significantly larger than pre- spray densities, whereasrelationships among them in Korea, as particularly other aquatic invertebrates were not. In spite of the lackrelated to the various aquatic animals in organically- of quantitative data, it was clear that spraying withfarmed rice fields, as well as the insecticide application Dimecron drastically reduced the numbers of aquaticin conventionally- farmed rice fields. TABLE 2 shows insects whereas recolonization with An. gambiae wasthe number of the aquatic animal taxa from organically- rapid.

and conventionally- farmed rice fields in 1995 and 1996. Simpson ( 1949) and McNaughton ( 1967) con-The aquatic animal taxa showed a total of 25 species in sidered not only the number of species and the total22 Families and 10 Orders or Subclasses, such as number of individuals, but also the proportion of theEphemeroptera, Odonata, Hemiptera, Coleoptera, total that occurs in each species. A collection of speciesDiptera, Branchiopoda, Ostracoda, Copepoda, with high diversity will have low dominance. The

Mesogastropoda, and Pisces, from the organically- or dominance indices of the rice fields showed a little

conventionally- farmed rice fields of Bulkyo from June fluctuation in 1996( TABLE 3). The values ofdominanceto September, 1996( TABLE 2). The monthly occurrence index ranged between 0.48 and 0.96 in the organic riceof species numbers at the rice fields ranged from 14 to field, and between 0.52 and 0.93 in the conventional rice24 species during this period. The vector mosquito field that the range of dominance was not so differentspecies in those rice fields were An. sinensis and Cx. between the rice fields. Chironomus sp. was frequentlytritaeniorhynchus. More species( 24) ofaquatic animals the dominant species in both rice fields during thewere collected from the organically- farmed rice field period ofrice cultivation. In the conventionally- farmedthan( 20) from the conventionally- farmed rice field. rice field, individual numbers ofAn. sinensis in July and

Shim et al.( 1995 a, b) concluded that both species, Cx. tritaeniorhynchus in September were very large.An. sinensis and Cx. tritaeniorhynchus, had developed Anopheles sinensis and Chironomus sp. appeared to bea high resistance to most of the pesticides that had been dominant species at both rice fields, and were especiallyapplied to conventionally- farmed rice fields so that their abundant in early July, 1996. Also, An. sinensis was aprevalence was generally not influenced by the pesti- dominant mosquito species in the conventionally- farmedcides. In the case of the predators of mosquito larvae, rice field during July. The populations of Cx. tritaenior-

TABLE 2. Number of the aquatic animal taxa from the organically- farmed and conventionally- farmed ricefields of Bulkyo from June through September, 1995 and 1996.

1995 1996

Rice fields Order Family Genus Species Order Family Genus Species

Organic 7 12 14 14 10 22 23 24Conventional 6 9 10 11 10 17 19 20Total 7 12 15 16 10 22 24 25


hynchus increased at both fields in September although caused by a heavy rainy season. The average of thethese mosquito larvae were not the dominant species. species diversity at the organic farm( 0.62) was almostThe abundance of Chinese muddy loaches, Misgurnus twice as much as that of the conventional farm ( 0.35).mizolepis( Gunther)( Cypriniformes: Cobitidae) showed The majority of living aquatic animals probably lesssmaller fluctuations than did the mosquitoes in the tolerate the chemicals than mosquito larvae in rice fields

organically- farmed rice field. Yet, this fish was a ( Ree et al. 1981, Shim et al. 1985, 1995a,b). Furthermore,dominant species only in June of both years. Sigara sp. the effect of pesticides might have been evident on thewas frequently a dominant species in the organically- predator populations because development rates of thefarmed rice fields in 1996, but after early July they did predators are slower and life histories differ appreciablynot appear to be a dominant species in the conventionally- from the comparatively rapidly developing and fastfarmed rice field. Also, Cercion hieroglyphicum and colonizing mosquitoes ( Service 1977). In 1996, theHebrus nipponicus were dominant species at both rice community similarities were the highest ( 0. 42) infields in August and September. The aquatic insects September before harvest but the lowest( 0.28) in earlymight have been impacted by the predators in the July. These results might be caused by polyphagousorganic farm and insecticides in the conventional farm. Chinese muddy loaches and King mud snails in the

Aquatic animal communities in the organically- organically farmed rice paddies. The average of thefarmed rice fields were more diverse than those in the similarities ( 0. 33) showed dissimilar communitiesconventionally- farmed rice fields, possibly because of between the rice field ecosystems during the year.various insecticide applications in the conventional rice The results ofthe monthly or biweekly prevalenciesfield. The values of species diversity index ranged from of An. sinensis, Cx. tritaeniorhynchus, and Chinese0. 15 in late July to 0.86 in September in the organic rice muddy loaches are summarized in Figures 1, 2, and 3,field and from 0.24 in late July to 0.39 in June and respectively. The population of An. sinensis larvaeAugust in the conventional rice field during 1996 appeared from early July, which was much earlier thanTABLE 4). In late July, the species diversity indices of that of Cx. tritaeniorhynchus, and kept a rather constant

both fields were the lowest, which might have been seasonal prevalence, showing an unusual peak in the

TABLE 3. Dominant species and their dominance indicies( DI) of the aquatic animal communities

at the organically- and conventionally- farmed rice fields in Bulkyo from June throughSeptember, 1996.

Date Rice Field Dominance species DI

Jun. 19 Organic Sigara sp. ( Hemiptera: Corixidae) 0. 51

Misgurnus mizolepis( Cypriniformes: Cobitidae)

Convent. Chironomus sp. ( Diptera: Chironomidae) 0. 85

Sigara sp. ( Hemiptera: Corixidae)Jul. 10 Organic Chironomus sp. ( Diptera: Chironomidae) 0. 83

Anopheles sinensis( Diptera: Culicidae)Convent. Chironomus sp. ( Diptera: Chironomidae) 0. 52

Anopheles sinensis( Diptera: Culicidae)

Jul. 25 Organic Chironomus sp. ( Diptera: Chironomidae) 0.96

Sigara sp. ( Hemiptera: Corixidae)Convent. Chironomus sp. ( Diptera: Chironomidae) 0.93

Anopheles sinensis( Diptera: Culicidae)

Aug. 21 Organic Cercion hieroglyphicum( Odonata: Coenagrionidae) 0.57

Chironomus sp. ( Diptera: Chironomidae)Convent. Chironomus sp. ( Diptera: Chironomidae) 0.87

Hebrus nipponicus( Hemiptera: Hebridae)

Sep. 12 Organic Hebrus nipponicus( Hemiptera: Hebridae) 0.48

Chironomus sp. ( Diptera: Chironomidae)Convent. Hebrus nipponicus( Hemiptera: Hebridae) 0.87

Chironomus sp. ( Diptera: Chironomidae)


TABLE 4. Species diversity( H') and community similarity( CCs)of the aquatic animal communities in the organically-farmed and conventionally- farmed rice fields in Bulkyo,1996.

Month Species diversity( H') CommunityDate Organic Conventional Similarity( CCs)

Jun. 19 0.74 0.39 0.29

Jul. 10 0.53 0.36 0.28

Jul. 25 0. 15 0.24 0.32

Aug. 21 0.81 0.39 0.35

Sep. 12 0.86 0.37 0.42

Average 0.62 0.35 0.33

TABLE 5. Association between the aquatic predators and the population of mosquito larvae in the organically-farmed rice fields of Bulkyo.

Correlation

Species An. sinensis Cx. tritaenior.

n r P n r P

Cercion hieroglyphicum( Odonata: Coenagrionidae) 8 - 0.25 0.274 4 0.04 0.478

Symprtrum darwinianum( Odonata: Libellulidae) 8 0.21 0.305 4 0.63 0. 186

Sigara sp. ( Hemiptera: Corixidae) 8 0.01 0.490 4 0.84 0.082

Hydaticus grammicus( Coleoptera: Dytiscidae) 8 - 0.38 0. 177 4 - 0.44 0.279

Helochares striatus( Coleoptera: Hydrophilidae) 8 - 0. 10 0.404 4 0.88 0.059

Ampullarius insularus( Mesogastropoda) 8 - 0.26 0.267 4 0. 13 0.437

Misgurnus mizolepis( Cypriniformes: Cobitidae) 5 - 0.66 0. 111 5 - 0.47 0.211

middle of September ( Fig. 1). The larvae of Cx. for the relative density for the field comparison becausetritaeniorhynchus were found from the middle ofAugust the fish are active in night time, yet they were collecteduntil the middle of September. This resulted from the using fish traps in daytime. No mortality was observedcomplete drainage of the water from both rice fields. As for the Chinese muddy loaches even during the drainingshown in the Figure 2, the population of this species out of the organically- farmed rice paddies and over theincreased in the rice fields during September. The winter. On the other hand, some mortality was observedmosquito larval populations of An. sinensis and Cx. for both young and adult Chinese muddy loaches in thetritaeniorhynchus in the organically- farmed rice field conventionally- farmed rice paddies just after insecticideswere generally much lower than those in the conventional were applied. The population of this species in therice field throughout the period. The Chinese muddy conventional rice field seemed to be deeply affected byloaches, Misgurnus mizolepis, were collected mostly in insecticide applications showing only 0. 3, 0.0, 0. 0, 0. 0,the organically- farmed rice paddies during the survey and 1. 0, compared to 3. 3, 3. 0, 4. 5, 4.0, and 4.0 in theperiod( Fig. 3). The population dynamics of those fish, organically- farmed rice field in June, early July, latewho were predators of the mosquitoes, were difficult to July, August, and September of 1996, respectively.assess with accuracy. It is not clear whether the density Four insecticides were used in the conventionally- farmeditself is low or the collection method is inadequate for rice field to control mainly the Pyralid moth larvae andfish populations. The number of this species appeared the Delphacid planthoppers. The results showed thatmuch smaller than the actual density, and is only reliable the monthly populations of Misgurnus mizolepis in the


55

Organic F45

c— Conven. F.

35

a)a)

n 25E

z

g 15a)

5t•16.

0

Jun., 95 Jul. Aug. Sep. Jun., 96 E. Jul. L. Jul. Aug.' Sep.'Month

Figure 1. Seasonal prevalence ofAnopheles sinensis larvae in the organically- and conventionally- farmed rice fieldsin 1995 and 1996.

34

0— Organic F28 Conven. F.

CS.

22

I)

S 16

l0

4

0

Jun., 95 Jul. Aug. Sep. Jun., 96 E. Jul. L. Jul. Aug.' Sep.'Month

Figure 2. Seasonal prevalence of Culex tritaeniorhynchus larvae in the organically- and conventionally- farmed ricefields in 1995 and 1996.


6. 5

5. 5 0– Organic F

a– Conven. F.

c 4.5clF. H

t..Na. 3. 5

E

2 2. 501

1. 5

0. 5n a-----___

e A

Jun., 95 Jul. Aug. Sep. ' Jun., 96 E. Jul. L. Jul. Aug.' Sep.'Month

Figure 3. Collection of Chinese muddy loaches, Misgurnus mizolepis in the organically- and conventionally-farmed rice fields in 1995 and 1996.

conventional rice fields were seriously suppressed by Sympet: um darwinianum against the vector mosquitoinsecticide application, and their recovery was not species were 0.21 and 0. 63, respectively. Also, Sigaraobserved. sp. ( Hemiptera: Corixidae), Helochares striatus

Hydaticusgrammicus( Coleoptera: Dytiscidae) was ( Coleoptera: Hydrophilidae), and King snails,collected in the rice fields during the study period. Both Ampullarius insularus did not significantly reduce theadults and larvae of this species seemed very susceptible mosquito populations. On the other hand, Hydaticusto the insecticide application, as abundance markedly grammicus ( Coleoptera: Dytiscidae) was relateddecreased whenever the insecticides were applied to the inversely to the mosquito populations, as the correlationconventional rice fields. The populations of other coefficients were- 0.38 with An. sinensis and- 0.44 withaquatic animals were not significantly different between Cx. tritaeniorhynchus. The population of mosquitothe organic and the conventional farming of rice fields, larvae in the rice fields might have been influencedexcept for Cloeon dipterum( Ephemeroptera: Baetidae) largely by oligophagous Chinese muddy loaches,and Cercion hieroglyphicum ( Odonata: Coenagrio- Misgurnus mizolepis. The coefficients of correlationnidae). However, it is not clear whether the fluctuation between Chinese muddy loaches and the population ofwas due to insecticide pressure or due to inadequate mosquito larvae showed the highest correlation: - 0.66sampling methods. It probably resulted from the in An. sinensis and- 0.47 in Cx. tritaeniorhynchus.insecticide treatments in the conventionally- farmed rice It has shown that there are not many effectivefield. Most predatory insect species were not affected control methods against adult mosquitoes because ofby the decrease ofAn. sinensis and Cx. tritaeniorhynchus their outdoor feeding, resting habits, and the socio-larval populations in the rice fields (TABLE 5). The geographical condition ( Ree et al. 1981, Shim et al.populations of Odonata nymphs ( Cercion 1987). Therefore, studies on population dynamics ofhieroglyphicum and Sympetrum darwinianum) were immature stages of the vectors are urgently needed andnot closely related to the mosquito larval populations. considered to be essential for effective control. It isThe correlation coefficients of Cercion hieroglyphicum important to understand the factors regulating naturalagainst An. sinensis and Cx. tritaeniorhynchus were - populations, particularly agricultural pesticides and0.25 and 0. 04, respectively. The coefficients of predators such as fish, adults, and/ or nymphs of


Coleoptera, Hemiptera, and Odonata and many others, Central Meteorological Office. 1996. Dailywhich are found in rice fields in Korea. It is apparent Meteorological Data. Apr.-Dec., No. 256.

that some natural enemies play an important role in Collins, F. H. and R. K.Washino. 1985. Insect predators.

regulating the population densities of the mosquito Pp. 25- 42. in Biological Control of Mosquitoes( H.larvae. Wada( 1975) carried out the experimental study, C. Chapman, ed.). Am. Mosq. Contr. Assoc. Bull.

showing that either Odonata nymphs or Notonectidae 6, 218 pp.Coleoptera) nymphs, or both, are important in reducing Cox, G. W. 1976. Laboratory manual ofGeneral ecology.

the Cx. tritaeniorhynchus larvae. Service ( 1977) 3rd Ed. Wm. C. Brown. Dubuque, 232 pp.collected large numbers of potential aquatic predators Fitzpatrick Jr., J. F. 1983. How to know the freshwater

from rice fields and identified by serological techniques crustacea. Wm. C. Brown. Dubuque, 227 pp.their predatory feeding habits. A broad diversity of Ikemoto, T. and I. Sakaki. 1979. Physico-chemical

invertebrate predators ofmosquitolarvaeintheirhabitats characters of the water in rice fields in relation to

has been reported in the literature (Jenkins 1964, Bay their suitability for breeding of the mosquito larvae1974, Collins and Washino 1985). Among them, the A. sinensis. Jap. J. Saint. Zool. 30: 87- 92.most abundant predators were Coleoptera and Hem iptera. Jenkins, D. W. 1 964. Pathogens, parasites and predators

Larvivorous fish have been a biological tool for mosquito of medically important arthropods. Bull. W.H.O.abatement for nearly 100 years in the U.S. ( Meisch 30( Suppl.): 1- 150.

1985). Yu et al. ( 1981, 1982, 1983) and Yu and Lee Kim, H. C., M. S. Kim, and H. S. Yu. 1994. Biological

1985) proved experimentally that the native fish control of vector mosquitoes by the use of fishAphyocypris chinensis and Aplocheilus latipes, which predators, Moroco oxycephalus and Misgurnus

had been very common in rice fields in Korea until 1985, anguillicaudatus in the laboratory and semi- fieldare effective predators of mosquito larvae and potential rice paddy. Korean J. Entomol. 24( 4): 269- 284.biological control agents. According to the report of McNaughton, S. J. 1967. Relationship among functionalKim et al. ( 1994), the introduction of other species of properties of California grassland. Nature. 216:

muddy loaches, Misgurnus anguillicaudatus, at 2 fish/ 168- 169.m2,

produced a 47. 7% reduction ofAn. sinensis larvae in Meisch, M. V. 1985. Gambusia affinis affinis. Pp. 3-a simulated rice paddy over five weeks. A marked 17 in Biological Control of Mosquitoes ( H. C.

reduction of mosquito abundance was observed at all Chapman, ed.). Am. Mosq. Contr. Assoc. Bull. 6,organically- farmed rice fields and might have been 218 pp.caused by the Chinese muddy loaches. These results Merritt, R. W. and K. W. Cummins. 1984. An

indicated that Misgurnus mizolepis is a promising introduction to the aquatic insects ofNorth America.

biological control agent for use against several mosquito 2nd Ed., Kendal/ Hunt Pub. Dubuque, 722 pp.larvae species breeding in the rice fields of Korea. Paek, I. K. 1992. A survey on the production and

circulation of low polluted crops. J. Environ.

Acknowledgments Pollut. Contr. 23( 8): 45- 49.

Peterson, A. 1982. Larvae of insects, Coleoptera, Diptera,

The author would like to thank Ms. Sun- hee Choi Neuroptera, Siphonaptera, Mecoptera, Trichoptera.

and Ms. Hye-kyung Hong for assistance in the field, and Part II, Edwards Brothers. Columbus, 416 pp.the farmer, Mr. Dae In Kang, for information on the Pielou, E. C. 1969. An introduction to mathematical

organically- farmed rice cultivation of his rice field. ecology. Willey- Interscience. New York, 384 pp.This study was financially supported by the Korea Ree, H. I., H. K. Hong, J. C. Shim, J. S. Lee, H. W. ChoScience and Engineering Foundation( 94- 0402-03- 01- and C. L. Kim. 1981. A study on seasonal prevalence3), 1994- 1997. of the populations of the mosquito larvae and other

aquatic invertebrates in rice fields in Korea. Korean

REFERENCES CITED J. Zool. 24( 3): 151- 161.

Service, M. W. 1977. Mortalities of the immature

Bay, E. C. 1974. Predator-prey relationships among stages of species B. of the Anopheles gambiae

aquatic insects. Ann. Rev. Entomol. 19: 441- 453. complex in Kenya: Comparison between rice fields

Brower, J. E. and J. H. Zar. 1977. Field and laboratory and temporary pools, identification of predators,methods for general ecology. Wm. C. Brown. and effects of insecticide spraying. J. Med. Ent.Dubuque, 194 pp. 13( 4- 5): 535- 545.

Central Meteorological Office. 1995. Daily Shim, J. C., H. K. Hong, and D. K. Lee. 1995a.Meteorological Data. Apr.-Dec., No. 256. Susceptibilities of Culex tritaeniorhynchus larvae


Culicidae, Diptera) to insecticides. Korean J. eds.). Plenum Pub., 138 pp.Entomol. 25( 1): 13- 20. Yoo, J. S. 1986. Korean shells in colour. Ilji-Sa, 196 pp.

Shim, J. C., H. K. Hong, S. H. Koo, and D. K. Lee. Yoon, I. B. 1988. Illustrated encyclopedia of fauna and1995b. Susceptibilities ofAnopheles sinensis larvae flora of Korea. Vol. 30 Aquatic insects. Ministry ofCulicidae, Diptera) to various insecticides. Korean Education, 840 pp.

J. Entomol. 25( 1): 69- 76. Yoon, I. B. and K. J. Ahn. 1988a. A systematic study ofShim, J. C., C. L. Kim, W. J. Lee, and E. H. Shin. 1990. Korean Dytiscidae II, Laccophilinae. Korean J.

Population densities of Culex tritaeniorhynchus Entomol. 18( 3): 191- 195.for surveillance of Japanese encephalitis in Korea. Yoon, I. B. and K. J. Ahn. 1988b. A systematic study ofKorean J. Entomol. 20(4): 213- 222. Korean Dytiscidae III, Colymbetinae and

Shim, J. C., E. H. Shin, D. S. Yang, and W. K. Lee. 1997. Dytiscinae. Korean J. of Entomol. 18( 4): 251- 268.Seasonal prevalence and feeding time ofmosquitoes Yoon, I. B. and D. S. Kong. 1990. Systematic study ofDiptera: Cul icidae) atoutbreakregionsofdomestic the dragonfly ( Odonata) larva from Korea ( I).

malaria (P. vivax) in Korea. Korean J. Entomol. Superfamily Aeshnoidea. Korean J. Entomol.27( 4): 265- 277. 20( 2): 55- 81.

Shim, J. C., Y. H. Yoon, C. L. Kim, W. J. Lee, and S. B. Yu, H. S. and D. K. Lee. 1985. Biological control ofLim. 1985. Studies on the effects of pesticides to mosquito larvae in South Korea using indigenousthe vector of J. E. and aquatic animals in rice fields. larvivorous fish. Proc. Texas Mosq. Contr. Assoc.Report of NIH Korea 22: 255- 266. 29: 7- 9.

Shim, J. C., Y. H. Yoon, C. L. Kim, W. J. Lee, B. I. Lee, Yu, H. S., D. K. Lee, and W. J. Lee. 1982. Mosquitoand S. C. Kim. 1987. Integrated control of vector control by the release of fish predator, Aphyocyprismosquitoes in rice fields. Korean J. Entomol. 17( 2): chinensis in natural mosquito breeding habitats of83- 91. rice paddies and stream seepage in South Korea.

Simpson, E. H. 1949. Measurement ofdiversity. Nature Korean J. Entomol. 12( 1): 61- 7.163: 688. Yu, H. S., D. K. Lee, and J. O. Na. 1983. Biological

Stehr, F. W. 1987a. Immature insects. Vol. 1, Kendall/ control ofmosquitoes by the release of larvivorousHunt Pub. Dubuque, 754 pp. fish, Aplocheilus latipes in confined natural rice

Stehr, F. W. 1987b. Immature insects. Vol. 2, Kendall/ paddies. Korean J. Entomol. 13( 1): 84.Hunt Pub. Dubuque, 975 pp. Yu, H. S., Y. H. Yun, D. K. Lee, and W. J. Lee. 1981.

Usinger, R. L. 1956. Aquatic insects of California. Biological control of mosquito larvae breeding inUniv. of Calif. Press, 508 pp. rice paddies in the presence of fish predator,

Wada, Y. 1975. Culex tritaeniorhynchus. Pp. 105- 118 Aphyocypris chinensis in Korea. Korean J. Entomol.in Control of Arthropods( R. Pal and H. Wharton, 6( 2): 29- 37.


Evaluation of Attractant-Baited Traps/Targets for MosquitoManagement on Key Island, Florida, USA

Daniel L. Kline' and Gene F. Lemire2

Centerfor Medical, Agricultural and Veterinary Entomology, Agricultural ResearchService, U.S. Department ofAgriculture, Gainesville, Florida 32604 USA.

2Collier MCD, 600 North Road, Naples, FL 34104- 3464 USA.

Received 24 August 1998; Accepted 28 September 1998

ABSTRACT: A three- year research project was conducted on Key Island, Collier County, Florida, USA,to evaluate an innovative attractant-based mosquito management technique. In the first year, speciescomposition, relative abundance, and spatial distribution were determined. Although 16 species werecollected, the dominant species was the black salt marsh mosquito, Aedes taeniorhynchus, which was theprimary pest species. Efficacy of a single line barrier, consisting of 52 carbon dioxide ( 200 cc/ min) +octenol( ca. 4 mg/ h) baited traps( 1994) or insecticide( lambdacyhalothrin) impregnated shade cloth targets1995), spaced ca. 16. 5 m apart, to reduce mosquito abundance in a resort area, was evaluated on the

northern end of the island. Success was evaluated by means of baited surveillance traps located on bothsides of the barrier. Traps and targets performed equally well. Though not statistically significant( p>0.05), there was a reduction in mosquito abundance in the resort area when the barrier was functional. Thesedata indicate that with refinement this mosquito management technique may be practical in certainsituations.

Keyword Index: Aedes taeniorhynchus, Anopheles atropos, Culex nigripalpus, mosquito, attractant,carbon dioxide, octenol, barrier, traps, target, control.

INTRODUCTION impact of a perimeter barrier of attractant- baited trapson mosquito abundance in a resort area( 1994), and the

Development of new mosquito management impact of a similar barrier of insecticide- impregnatedstrategies that reduce reliance on chemical pesticides cloth targets on mosquito abundance, i. e., how wouldare needed to counter the problems posed by targets perform compared to traps to reduce mosquitoenvironmental concerns, a dwindling selection of abundance in a designated area ( 1995). This paperapproved chemical insecticides for mosquito control, summarizes the data obtained for the first phase of thisand an increasing number of mosquito species that have project.

developed resistance to these approved insecticides.One possible alternative is to use strategically placed MATERIALS AND METHODSattractant- baited traps and/ or targets.

Research towards the development of such a strategy Study Site. Key Island, located in Collier Co., FL,for the management of the black salt marsh mosquito, immediately south of Naples, was selected because ofAedes taeniorhynchus( Wiedemann), and other species its abundant salt marsh mosquito populations and limitedassociated with estuarine ecosystems in south Florida, access ( either by boat or helicopter), that reduced thewas initiated in 1993. The ultimate goal of this research likelihood of vandalism. This ca. 14. 5 km long island,project is the development of a cost-effective, environ- the largest remaining undeveloped barrier island inmentally friendly, attractant-based operational mosquito Florida, combines with a number of smaller islands tomanagement program. Specific goals for the first phase provide an essential physical barrier between the 3644of this research program were the determination of: ha Rookery Bay National Estuarine Research Preservespecies composition and spatial distribution( 1993), the and the Gulf of Mexico. These islands contain critical


wildlife habitats, including nesting areas for endangered found on the pressure regulators. This combination of

sea turtles and shorebirds. The vegetation on Key Island attractants and release rates was based on studies

is a mixed` jungle" ofred and blackmangroves, palmetto previously conducted with Everglades mosquito speciesthickets, arboreal orchids," slash" pine, and Australian ( Kline et al. 1990, Kline et al. 1991b, Takken and Kline

pine. Fauna includes bobcats, armadillos, gopher 1989). The traps were hung from metal poles so that thetortoises, raccoons, deer, as well as nesting of the Great top of the trap was approximately 1. 8 m above groundHorned Owls, Osprey, Pileated Woodpeckers, and level.

Towhees. In 1993, 42 traps were placed throughout the north

Because of its location and relatively pristine state, end of Key Island within the resort area and along thethis island has been the center of conflict between various nature trails established by the proprietors of theenvironmentalists and land developers for several resort( Fig. 1). All traps were baited as described above

decades. In 1993, all but the northernmost 24. 3 ha of the and operated continuously from June 21 until July 16. Inisland was purchased by the state of Florida and addition to providing data on species composition,incorporated into the Rookery Bay Preserve. A resort seasonal incidence and relative abundance, these traps

with small rustic guest cottages was established on the also provided information on spatial distribution. From

24.3 ha, which remained in the possession of a private July 16 through August 18 only six of the traps, i.e., 1, 5,corporation. The land purchase and subsequent 6, 33, 35, and 43 were used to obtain additional baseline

incorporation into the preserve has had an impact on data. Data obtained in 1993 determined placement of

mosquito control options that might be utilized. Collier ten surveillance traps in 1994. These 10 surveillance

Mosquito Control District( CMCD) personnel consider traps were positioned so that five were located on each

this island as an important resting stop for salt marsh side of a protective barrier, i. e., outside and within the

mosquitoes believed to be involved in massive migrations area designated for protection from mosquito invasions.

from Ten Thousand Islands and the Everglades into These surveillance traps were operated from May 19Naples. To prevent their invasion of Naples, CMCD has through July 20. In 1995, surveillance trap locationsrelied on aerial sprays of Baytex ( fenthion). Since remained the same as 1994 and were operated

Rookery Bay Preserve allows no aerial sprays ofchemical continuously from May 15 through July 6.pesticides over its lands, alternative mosquito control Trapping intervals for surveillance traps weremethods need to be developed. Thus, the resort owners approximately the same each day and were servicedand estuarine preserve personnel have become very daily in the same sequence between 8: 00- 10: 00 AMreceptive to the proposed research project. EDT. The only major change to this routine was that in

Surveillance: species composition, relative abun- 1995 an interval rotator device with timer ( Collector

dance, seasonal, and spatial distribution studies. The Bottle Rotator-Model 1512, John W. Hock Company,model 512 CDC type trap( John Hock Co., Gainesville, Gainesville, FL) was used with each surveillance trap.FL) was used without a light source. Each trap was This provided the same defined time interval for all

baited with 1- octen- 3- ol ( octenol) and carbon dioxide locations and also provided flexibility in the daily ser-

CO2). Octenol( Aldrich, Milwaukee, WI) was released vicing of traps. This also helped to reduce the need forfrom microreaction vials ( 5 ml; Supelco, Bellefonte, weekend work and servicing of traps during inclementPA) fitted with plastic lids and neoprene septa using a weather.

wick( Dills 15- cm pipe cleaner) folded into a V-shape Collections were made into 568 ml ( pint) Mason

with its point in contact with the neoprene septum. With jars containing a small piece ( ca. 2. 5 x 5. 0 cm) ofthis technology octenol was released at ca 4 mg/ h over Revenge®Bug Strip""( Roxide International, Inc., New

a 24 h period. Octenol vials were affixed near the trap Rochelle, NY) with dichlorvos ( 18. 6%) as the killingentrance adjacent to the CO2 release point( illustrated in agent. All collections were weighed and then stored in

Kline et al. 1991a). CO2 was metered from a 9kg a freezer until further processing. In 1993, the entirecompressed gas cylinder at 200 ml/min using a double- sample was frozen after weighing. Therefore, uponstage pressure regulator ( Victor Model VTS 453B; removal from the freezer, the entire sample was

Victoria Equipment Company, Denton, TX). Gas flow reweighed and a 0.5 gm subsample was taken, identified

was monitored using a compact# 12 flowmeter( Gilmont to species, and an estimate of the total trap collectionInstruments, Barrington, IL) and delivered to its release extrapolated from this data. In 1994 and 1995, the total

point ca. 5 cm from the top trap entrance via polyethylene sample was weighed and a 1 gm subsample removed

tubing. To assure accurate flow calibration, micro- immediately before placement into the freezer. Onlymetering valves ( series M, Nupro, Willoughby, OH) the subsample was identified to species. An estimate of

were used in place of the standard open- shut valves the total collection size was made by extrapolation. If


Naples

r__,42 l n 7

Ci0rdo.' 41 26•_

2724

a\.-23

t 28limmit

22

X740

29

21

3 1• 3020

1 •

32• 18

I • 161 3 •

41 .

12

34• I•

C l(_

1

10

Q 36• g 1J `36•

31• •7

1J

3•

5. 2

50 0I

N

A EW- • •

Scale I in. = 725 ft. S2. 54 cm. = 22 I m.)

Figure 1. 1993 distribution of Model 512 surveillance traps( black dots), baited with CO2( 200 cc/ min)+ octenol

ca. 4 mg/h), on the northern end of Key Island, Naples, Collier Co., FL

the total sample was 1 gm or less the entire sample was ofattraction for some species ofWest Africa mosquitoescounted and identified to species. to CO2 or calf bait( 15 m for CO2 alone and 9- 18 m for

Barrier studies—baited traps. Based on 1993 data, one calf) and partially to logistical restraints, i. e., thea single line protective barrier of 52 Model 512 traps, fact that this spacing yielded the maximum number ofbaited the same as the surveillance traps, was placed traps that we felt we could supply with CO2. Barrieralong New Trail and its extensions( extended northeast traps were operated 24 h/ day from June 14 through Julyto a small cove [ Fig. 2] and southwest to the Gulf of 11. The ten surveillance traps were operated as a pre-Mexico). The trap barrier was strategically placed so treatment from May 19 through June 13, and as a post-the entire northern resort area with cottages would be treatment from July 12 through July 20.protected. These barrier traps were placed ca. 16. 5 m Barrier studies—baited targets. Since our ultimateapart. This spacing was based partially on the data objective was to kill the mosquitoes rather than retainreported by Gillies and Wilkes( 1972) on effective range them for surveillance purposes, it was decided that a


21\\15

i20

IIMININ i AM

C 5• 25 • 10

30

91

r h 35

40

cfi45

Z

50

5 8

60 \

7•

Figure 2. Illustration of placement of barrier of traps( 1994)/ targets( 1995) on Key Island and distribution of thesurveillance traps. The resort area( area to be protected) is cross hatched.

barrier of targets, which require less maintenance, would June 4 through 28. The ten surveillance traps werebe more economical than traps. Therefore, in 1995 the operated during this time and as a pretreatment frombarrier traps were replaced with cloth targets impregnated May 16 through June 3; and as a posttreatment fromwith an insecticide ( lambda-cyhalothrin), baited with June 29 through July 6.the same combination and release rates of attractants as The collapsible targets consisted of cylinders ( ca.the traps. Each trapping site remained the same. The 60.25 cm length x 53. 21 cm diameter)( Fig. 3), of 60%objective was to determine if targets would have the polypropylene black shade cloth ( DeWitt Company,same impact as baited traps on mosquito abundance Sikeston, MO) treated with an EC formulation( 120 gni/

within the resort area. Targets were placed on site from 1) of lambda-cyhalothrin at 0.2 g A.I./m2. The targets


M! w'' 4. p114.. d+. YYIIiiiLE+.M*. y.M

F-'x'"§+

9• y`....

E oka.

iii94e7...4„, -..„ . 3.,. ,

m 1 3" 41

4104 . . A§`.e t SRC 41 ' '".... .,°

a

04,,'t: 1M, j, fir,

ro.. ;. ire

4at

F a y

4._

f '.. r

x'

1' 1' 1

i

ai,4 1

4;-),,,...:

4 , msµ„

4,0(4)4 v

y^,t-

r 4f

big4, r T

Figure 3. Photograph of cylindrical shade cloth target with stake holding CO2 release tube and octenol vial.

were fabricated by attaching the insecticide impregnated of the target' s upper ring. Each target was suspended socloth to upper and lower 5. 13 cm wide bands of lexan the bottom was ca. 15 cm above ground level. The

1. 56 mm thickness) by means of six sheet metal screws. octenol vial and CO2 release tube were taped( duct tape)The targets were constructed to allow for a 10. 26 cm to a short stake located centrally under each target sooverlap of cloth, which could be removed for bioassay odors were dispensed ca. 67 cm above ground level.purposes. The upper surface of the cylinder was also Our goal for 1995 was to utilize a target impregnatedcovered with insecticide impregnated shade cloth. The with an insecticide formulation that would be effectivebottom was left open to allow mosquitoes to enter and for at least 30 days before retreatment was necessary.rest on the inner surface of the target. Targets were Lambdacyhalothrin was selected as the insecticide ofsuspended from poles that had been used for the barrier choice based on preliminary screening of seven candidatetraps in 1994 with string tied to the sheet metal screws insecticides ( unpublished data) against a laboratory


colony ofAe. tacniorhynchus. The amountof insecticide Psorophora ciliata( Fabricius), Aedes infirmatus Dyar

mixture required to treat the shade cloth was determined and Knab, Aedes atlanticus Dyar and Knab, and Aedes

by first determining the amount of water required to mitchellae ( Dyar). Each of the three most abundant

saturate the fabric. The desired treatment rate( 0.2 g AU species had a different temporal( Figs. 4- 6) and spatialm2)

was obtained by mixing2.4mlofLambdacyhalothrin distribution patterns ( Figs. 7- 9). Relative abundance

E.C.( 119 g AI/1) with 195 ml of water. This mixture and varied from week to week( TABLE 1). Although Ae.

the cloth required to make the target were placed into an taeniorhynchus was always the dominant species

approximately 8 liter resealable freezer bag. The bag collected, its degree ofdominance fluctuated from 99. 1%

was sealed and inverted daily for four days. The cloth in week 1 to 57. 6% in week 7. In week 8 it rebounded

was removed one week after treatment, hung from a to 94. 2%. Culex nigripalpus started out in week 1 as

fence with clothespins, and allowed to dry. This treat- < 1. 0% of the total collection, and gradually peaked atment combination met the 30 day effective specifica- 37. 7% in week 5; weeks 6 and 7 remained relativelytion when exposed to natural environmental conditions. stable at 32.0 and 29. 5%, respectively. In week 8 it

Effective field longevity was monitored during the declined to 4.4%. Anopheles atropos composed< 1%

current study by bioassaying pieces of cloth removed of the collections during weeks 1 through 3, and graduallyfrom five randomly selected targets at the beginning of increased to 12. 9% in week 7 at which time it declined

each sampling week. Ten field-collected Ae. taenior- to 1. 4% in week 8.

hynchus were exposed to each piece of cloth for one In 1994, an estimated 3, 045, 312 mosquitoes were

minute. Two checks consisting of exposing mosquitoes collected, with a similar relative abundance pattern for

to untreated pieces of shade cloth were run with each the three major species as was found in 1993, i. e., Ae.

bioassay. Mortality was determined at 30 min. and 1 taeniorhynchus ( 89.6%), Cx. nigripalpus ( 8. 4%), and

hour post- exposure. All targets were operated An. atropos( 1. 8%). Aedes taeniorhynchus had several

continuously( i.e., 24 hours/ day) for the entire time they peaks of abundance, the largest occurring June 27- 30were in the field. Fig. 4). After this major peak, the population crashed

Data Analysis. In 1994 and 1995, the efficacy of and never rebounded during our evaluation period.the barrier traps/ targets was determined by comparisons Culex nigripalpus ( Fig. 5) populations were very lowbetween the five outside and five inside traps. A ratio of until the third week in June. The largest peak occurred

mean outside/ mean inside surveillance trap collections on June 28. Several smaller peaks occurred after the

was determined daily, each year, for pre- and post- barrier of traps was turned off. Anopheles atropos( Fig.treatment( barrier traps/ targets were functional) periods. 6) populations had two spikes, June 24 and June 28. For

Data were analyzed by the SAS GLM and means the most part of 1994 this species was relatively low incomparison procedures( SAS 1988). This ratio ofmean the trap collections. The only differences in speciescatch for outside/ mean catch inside traps pre- trial, trial, composition were that some An. quadrimaculatus Sayand post- trial would give an indication of the and no Ps. ciliata were caught.

effectiveness of the barrier traps/ targets. A ratio that In 1995, the same three species again dominated

increased during the treatment period was an indication collections ( TABLE 1), but their relative abundance

that mosquitoes were being" filtered out" by the perimeter was different compared with the previous two years.

barrier. Aedes taeniorhynchus comprised only slightly> 54% of

the collections. Aedes taeniorhynchus populations were

RESULTS highest in May ( Fig. 4), after which they graduallydeclined for the remainder of this study. Culex

Surveillance: species composition and relative nigripalpus comprised slightly> 16% of the collections.

abundance. In 1993, an estimated 6,592, 164 mosquitoes Culex nigripalpus followed a similar pattern as in 1994

were captured, consisting of 16 species/ species groups. ( Fig. 5), except that after reaching a spike on July 6, thisTrap collections were dominated by three species: Aedes species virtually disappeared from the trap collectionstaeniorhynchus( 84.7%), Culex nigripalpus Theobald for the remainder of the study. Anopheles atropos was

13. 8%), and Anopheles atropos Dyar and Knab( 1. 4%). the second most abundant species collected, composingOther species in order of decreasing abundance were: ca. 28% of the collections compared to only 1- 2% the

Psorophora columbiae ( Dyar and Knab), Aedes previous two years. The An. atropos population was

sollicitans ( Walker), Anopheles crucians Walker, relatively high throughout most of the evaluation periodMansonia spp., Psorophora howardii Coquillett, in 1995( Fig. 6) until the end of June when it crashed andPsorophora ferox ( Von Humboldt), Coquillettidia never rebounded. No Ae. mitchellae were collected.

perturbans ( Walker), Culex ( Melanoconium) spp., Psorophora ciliata specimens were again found in


INa.i

I:,

a91. ca

1C..)

1 nC ZI. ti E

a

o9

1.)

9 ob

ct

GE

b

LZ 0ct

AM

CZ6)

3 C

E 0

E 6I.a.

I a bSI - s;

E- chM

L0

EINE. owI

1

U, a

CC))C))

eAA VAS

T. @, 6Z

tn

t M,v, cE 1Z b

y

S 0t `

0

8z b1

bZm o. Z.

1I I

i

E .s9- OZ a> O.

inN N u Oo w G N 0 N 0 v M N " 0 4

spuesnoyl spuesnoylspuesnoyl

a

o9w


act

c.o

9l aa

1 a\\ ZG 15 0

8 aa

p to

C4£ o

LZaa

Z

11 klawKW1

a a

ENI

6l4... o

SG

a

1\ ll

E

I L rn

l j E 1E cr.

ii of 1 nE-£ oco

T T T

6Za N

k. - b

a.

SZ

E a.

C GZa

M a

3 Ql n,

E to

EU y

U

js

oa

G oa

08Z v,

70t, m

2a

E .

OZ y

E a

9L

M N . O M N ,- 0 N O tri

spuesnoyl spuesnoyl spuesnoyl cuaoq

w


Ia)ab

a.)6.

a9L as

ap

E ZL 6 E

m aa a

I9 Eccs

Ea

0b 0

LE o

a\\ v` v- I Io

LZc9

E T

CZ

a

E o

C 6Lo.o.

M, SL .) aE rn

LL ov

H E

L o,

E o

E c.x\, E b

co E e u) Eo

0 E 0) r<\\\\\; \\\\ 0 E- 6Z a

I ON

SZ

a ` 4Z a

E b

La) a

OE- IL\ E L O

L . h

E~ ES w

0

aoi.

9Z0

17Z >, aZto

CCI

a,OZ a) a

9LM r N O 7- N O CV 1- O p

spuesnoyl woyspuesnoyl spuesnl 6"

we


42

041 260 250

027024

23028

I 22

029

040 021

043 31• 0 30020

19•

I 320 018

17016

330 15413 0 12

34• 10 ,/

Average #

9• Aedes taeniorhynchus

35• 08 per trap/night20,000

30 • 37 7

m > 10, 000 - 20,00002 36

5, 000 - 10,000

61> 0 > 1, 000 - 5,000

I 1 < 1, 001

5

Figure 7. Relative spatial distribution of Aedes taeniorhynchus on the northern end of Key Island, June 22- July16, 1993.

some collections. This was the first year that some between 5, 000 and 10, 000 Ae. taeniorhynchus per trapWyeomyia spp. were found in the collections. night. Culex nigripalpus was generally most abundant

Spatial distribution. Mean trap collection and along New Trail, the eventual location of the perimeterrelative spatial distribution data for the three major barrier. Anopheles atropos was most abundant in trapsspecies are illustrated in Figures 7- 9. Asa generalization located south and east of the resort area. Very fewforAe. taeniorhynchus( Fig. 7), the further southeast the specimens of this species were collected in the traps

trap was placed, the greater the mean collection size. located along New Trail. Trap# 38 and Trap# 2 locatedThe majority of traps located along New Trail averaged within and just outside the resort area, respectively,


42

041 260 25

J027 024

2328

22

029 /

40 021

013 31• X30

020190

32• I8

170

15016

331> El 1413•

I2

34• I 1O 10 Average #

9• oCulex nigripalpus35 o s i per trap/night

30 36 ,/- 3740 / 7 j • > 20002 1500- 2000

1000 - 1500

600 > 500 - 1000

N i < 501

4

5•

0

Figure 8. Relative spatial distribution of Culex nigripalpus on the northern end of Key Island, June 22- July 16,1993.

caught a moderate number of this species. was functional. Similar trends were not obtained forAn.Effectiveness of trap/ target barrier. No statistical atropos for both years. No filtering was indicated in

differences ( p > . 05) among the outside/ inside ratios 1994 and only slight filtering in 1995. Results of

during any period were observed either year. However, bioassays from targets indicated 100% mortality 30 minthe trends ( TABLE 2) for Ae. taeniorhynchus and Cx. post-exposure for the first three weeks. By the beginningnigripalpus were similar for both years. The trend of the fourth week some of the targets were showing ca.indicated for these species was that the barrier was 90% mortality 30 min. post- exposure, but by one hourfiltering out" some of both species during the times it post-exposure all the target species tested resulted in


042

041 26 250

24

127

1

i028

23

1 El 22i 029

E4 02

3 0320

19

32I8

17

IMINIIIIMINIIMMINIMIN 16 1WIIIIIIIIIIIMINIIIMINIIIIIMI__

337 1

11014

1301234

10.

10 Average #INIIMI

9 o Anopheles atropos35 08 per trap/ night

7 K • > 3003 ' v 370

m > 200 - 30036 1 100 - 200

o > 50 - 1006•

11 < 51

4 /

N\

r l5m

Figure 9. Relative spatial distribution of Anopheles atropos on the northern end of Key Island, June 22-July 16,1993.

100% mortality. No check mortality was observed at insecticide applied, allows the choice of where and

either post-exposure period any week. when the spot treatments will be made, and greatlyreduces the impact of insecticides on nontarget

DISCUSSION organisms.

This study, to the best of our knowledge, is the firstThe use of attractant baited traps/ targets for removal attempt to use a barrier of baited traps/ targets for

trapping of mosquitoes offers an alternative to the population management of any species of mosquito. Inconventional approach of broadcasting chemical Zimbabwe( Torr 1994), tsetse flies( Glossina spp.) areinsecticides over large tracts of land. This approach has controlled using insecticide- impregnated targets baitedthe advantage of greatly reducing the amount of with a blend of acetone, octenol, and several phenolic


v v o compounds. While we are disappointed that our barrier

00 4 v o of traps/ targets did not result in a more dramatic reduction

d rn ° v in mosquito abundance in the resort area, we remain

N o -. encouraged by the fact that there was any reduction. In

TN `° '" both years when the barrier was present, the trend was

O definitely toward reduction of mosquito abundancea in rn o within the area designated for protection.0

Ncr From the outset we knew that adjustments would

00 o CA r probably have to be made in trap/ target type, spacing,00 4 strategic placement, as well as the quality and quantity

M of attractants used. Among the refinements underN

consideration are reduction in distance between traps/

R ° R °°5

targets( e. g. utilize a spacing of 6- 8 m), increased CO2U o levels, a double and staggered line of traps/ targets, or

o M O 4 complete encirclement of the area designated for

v:s

oM ,o protection. Therefore, further studies on mosquito

o00

N behavior need to be conducted, including movemento v o patterns( horizontal, vertical and diel) and their responsesoN a 06 r- 4 o to different kinds and heights of traps/ targets. Do these

n 3 "' n M ' r three major species respond to different combinations

0 o N o ofchemical and physical cues? Do they fly low or high?o 06

s co en One of us( DLK) observed early morning movements ofo

14An. atropos along the sandy beach parallel to the island' s

6 'r o = coastline. This type of movement pattern allowed atN o

least a portion of this species to completely circumvent3 D N 00 00 the barrier. Indeed the data( TABLE 2) indicate that theO VD

a- rn m - protective barrier was least effective for this species.cis

cNV Another important reason why the barrier may havevsa. i o M i o been least effective for this species may be its spatialb N o o distribution. Assuming that spatial distribution for thisQz "'

o o species was the same in 1994 and 1995, as it was in 1993

aN Fig. 9), then very few An. atropos would have been

M ° attracted to the traps/ targets making up the barrier alongNew Trail. Perhaps that is why the barrier was not as

7 N r - effective in" filtering out" this species as it was the otherG o o

two major species. In contrast, Cx. nigripalpus shouldaNN rn a- en have been the most impacted of the three major species

o en r N v since its spatial distribution pattern indicated that its

o °` 0

ppulations were most concentrated along New Trail.oo Indeed the barrier' s impact was greatest on this species

orr,

00 kn en .--,

o 00 0 0 o both years.

E 4 n o Therefore, we remain optimistic that this technologyQ

N 00

N N will be developed where it can be used to effectivelya _

7:). — — reduce mosquito annoyance in areas designated for

r''protection. This optimism is based partly on theconcurrent cooperative research projects that we initiated

0withacademia, government, and private industryC VI o personnel at the outset of phase one of this project.Ct 6,

o o These cooperative research projects have focused on theGs ° ... z development of better traps/ targets, attractants, and theU 0 4.. h

a t 0° use of spatial analysis to strategically locate the traps/•

a f' y o targets for maximum impact on the target species.

C.) Q Q These efforts have already resulted in the development


TABLE 2. Effectiveness of a single line barrier of traps/ targets to prevent mosquito invasions intoa designated area; z =mean number mosquitoes captured/ trap night and standarderror of mean( S. E.)

1994 1995

Area Period T (S. E.) 0/I z (S. E.) 0/I

Aedes taeniorhynchus

Inside Pretreatment 2964.86 ( 458. 01) 2.57 752.64 ( 138. 58) 2.54

Outside 7634.50 ( 919. 00) 1910.01 ( 312.73)

Inside Treatment 2572.48 ( 539. 79) 3. 21 325. 87 ( 37. 27) 3. 19

Outside 8249.58 ( 122. 30) 1040.08 ( 225. 47)

Inside Posttreatment 466. 47 ( 88. 22) 1. 69 141. 04 ( 28.48) 4. 11

Outside 789.73 ( 122. 3) 580. 16 ( 75.69)

Anopheles atropos

Inside Pretreatment 21. 88 ( 9. 19) 2. 02 392.91 ( 88. 72) 1. 58

Outside 44.24 ( 16. 35) 621. 54 ( 144. 87)

Inside Treatment 100.22 ( 28. 89) 1. 60 374.40 ( 51. 96) 1. 83

Outside 160.79 ( 42. 72) 687.02 ( 93. 18)

Inside Posttreatment 55. 73 ( 16. 07) 1. 05 19. 46 ( 5. 37) 2. 16

Outside 58. 39 ( 22. 13) 42. 12 ( 9.73)

Culex nigripalpus

Inside Pretreatment 2. 64 ( 0.56) 4.06 3. 01 ( 0. 64) 1. 21

Outside 10.70 ( 2. 69) 3. 64 ( 0.97)


Outside 1007. 41 ( 280.51) 353.99 ( 58. 83)

Inside Posttreatment 566.04 ( 258. 44) 3. 61 238.08 ( 84.93) 2. 35

Outside 2045.92 ( 586. 31) 560. 19 ( 130. 33)

All Mosquito Species Combined

Inside Pretreatment 2990.72 ( 462.65) 2. 57 1149.03 ( 207. 69) 2. 21

Outside 7698.23 ( 925. 49) 2536. 88 ( 427. 85)


Outside 9437.29( 1650.46) 2134. 12 ( 263. 65)

Inside Posttreatment 1089.68 ( 303. 89) 2. 67 400.41 ( 107. 85) 3. 04

Outside 2907.03 ( 620.04) 1218. 18 ( 162. 69)

of several new trap types, which in small scale field systems are now commercially available. New host

studies have resulted in captures which are 5- 10 times odors have been isolated, identified, and successfullygreater than those obtained by the traps/ targets with the tested in the laboratory ( unpublished data). Another

same attractant combination used in phase one reason for optimism is that the control of tsetse inunpublished data). One new trapping technology has Zimbabwe, which is the best example of the use of

been developed, which even produces its own attractants kairomones to control a medical/ veterinary entomologyCO2, heat, and water vapor) and generates its own pest, was not an overnight success ( Vale 1993, Torr

electricity to operate the trap. This trap is already 1994). Success was achieved as the result of a long,commercially available. Progress has also been made dynamic process over many decades through closein the attractants area. Several different octenol delivery interdisciplinary collaboration of many dedicated


scientists. Our three-year research project is simply the Africa. Bull. Entomol. Res. 61: 389- 404.first step towards the development of attractants/ traps/ Kline, D. L., W. Takken, J. R. Wood, and D. A. Carlson.targets for mosquito management. 1990. Field studies on the potential of butanone,

carbon dioxide, honey extract, 1- octen- 3- ol, L-Acknowledgments lactic acid and phenols as attractants for mosquitoes.

Med.Vet. Entomol. 4: 383- 391.This project was a cooperative effort by personnel Kline, D. L., D. A. Dame, and M. V. Meisch. 1991a.

of the United States Department ofAgriculture( USDA), Evaluation of 1- octen- 3- ol and carbon dioxide asAgricultural Research Service ( ARS), Center for attractants for mosquitoes associated with irrigatedMedical, Agricultural and Veterinary Entomology rice fields in Arkansas. J. Am. Mosq. Contr. Assoc.CMAVE), Gainesville, FL, and Collier Mosquito 7: 165- 169.

Control District ( CMCD), Naples, FL. The authors Kline, D. L., J. R. Wood, and J. A. Cornell. 1991b.would like to thank H.T. McKeithen ( USDA, ARS, Interactive effects of 1- octen- 3- ol and carbonCMAVE) for his technical help in setting up and servicing dioxide on mosquito ( Diptera: Culicidae)the field traps/ targets, identification of mosquitoes, surveillance and control. J. Med. Entomol. 28:conducting the statistical analyses of the data, and 254- 258.

preparing the graphics used to illustrate the data. The SAS. 1988. SAS/ STAT user' s guide release 6. 03. SASfield assistance of E. Lavagnino( USDA, CMAVE), A. Institute, Cary, NC, 1, 028 pp.Salinas and C. Laidlaw-Bell ( CMCD) was also much Takken, W. and D. L. Kline. 1989. Carbon dioxide andappreciated. E. Lavagnino also assisted in the 1- octen- 3- ol as mosquito attractants. J. Am. Mosq.identification of mosquitoes. Contr. Assoc. 5: 311- 316.

Torr, S. J. 1994. The tsetse ( Diptera: Glossinidae)REFERENCES CITED story: implications for mosquitoes. J. Am. Mosq.

Contr. Assoc. 10: 258- 265.Gillies, M. T. and T. J. Wilkes. 1972. The range of Vale, G. A. 1993. Development of baits for tsetse flies

attraction of animal baits and carbon dioxide for Diptera: Glossinidae) in Zimbabwe. J. Med.mosquitoes. Studies in a freshwater area of West Entomol. 30: 831- 842.


Olfactory Responses and Field Attraction of Mosquitoes toVolatiles from Limburger Cheese and Human Foot Odor

Daniel L. Kline

USDA, ARS, CMAVE, P. 0. Box 14565, Gainesville, FL 32604 USA

Received 26 August 1998; Accepted 24 September 1998

ABSTRACT: Olfactory responses of femaleAedes aegypti( Linnaeus) to various odor stimuli were studiedin a dual- port olfactometer. Responses( i. e., the percent of ca. 75 available female mosquitoes in flight

chamber entering each olfactometer port) were studied toward clean conditioned air( control), human footskin emanations( collected on socks by wearing them for three days), human hand, and Limburger cheese.Mean percent response was greatest to the human hand ( 80. 1%), followed by the human worn sock66. 1%), Limburger cheese( 6.4%), and control(< 0. 1%). In field studies the worn sock alone attracted

very few mosquitoes but a synergistic response occurred to the sock+ carbon dioxide baited traps for mostspecies of mosquitoes in six genera( Aedes, Anopheles, Coquillettidia, Culex, Culiseta, and Psorophora).

This synergistic effect persisted even when the socks were exposed to environmental conditions for eight

consecutive days. Limburger cheese alone did not attract mosquitoes to traps compared to unbaited traps,

and there appeared to be a slight repellent effect for most mosquito species when used in combination with

carbon dioxide.

Keyword Index: Foot odor, olfactometer, Limburger cheese, mosquitoes, kairomones, host-seeking.

INTRODUCTION between octenol and CO2 has been reported for severalmosquito species( Takken and Kline 1989).

A major emphasis of current pest management Recently, Limburger cheese has been identified asresearch is the development of selective, environmen- an effective attractant for Anopheles gambiae s. s. Giles

tally friendly methods of control. Mosquito control ( Knols and De Jong 1996, Knols et al. 1997). Knols

research is no exception. One such mosquito manage- ( 1996) implied that Limburger cheese is an effective

ment concept currently under scrutiny is the use of attractant for this species because the blend of volatiles

attractant-baited traps/ targets to capture/ kill nuisance from this cheese is very similar to foot odor. The use ofmosquitoes( Kline 1994). The success of this approach analytical chemistry confirmed the similarity of fattydepends upon the development of efficient trapping acids found emanating from Limburger cheese and foottechnology, the discovery of new, effective attractants, scrapings ( Knols and Meijerink 1997). The foot is a

and subsequently, strategic placement of these baited preferred biting site for An. gambiae s. s. ( De Jong andtraps/ targets for maximum impact on the target Knols 1995). The research reported herein was conducted

population. to evaluate the potential use of Limburger cheese and

The present work was undertaken as part of a larger human foot volatiles as attractants for Aedes aegypti

study on the behavioral responses of mosquitoes to ( Linnaeus)( olfactometer studies) and natural populations

semiochemicals. The ultimate goal of this phase of the of mosquitoes associated with wooded wetlands( field

research program is to isolate and identify new attractants. studies).

To date, few chemicals other than carbon dioxide have

been isolated and identified as behaviorally active for MATERIALS AND METHODS

host- seeking mosquitoes. 1- Octen- 3- ol ( octenol),originally identified from ox emanations ( Hall et al. Attractants

1984) as an attractant for tsetse flies has been shown to Limburger cheese. Amish Country Limburgerelicit positive behavioral responses from mosquitoes cheese( Amish Country Cheese, Linwood, MI 48634)Takken and Kline 1989, Kline 1994). Synergism was purchased from a local supermarket. It was used for


both laboratory and field studies. filters and then heated and humidified, if necessary.Socks. Two types ofsocks were used in these trials. One hour before initiation of tests, ca. 75 femaleAe.

White crew socks (AthletechTM,

Troy, MI) with fiber aegypti were aspirated into the olfactometer chamber

content consisting of 82% cotton and 18% polyester and allowed to acclimatize for ca. one hour before

were used in both olfactometer and field studies. Black testing. Treatments were placed into two test ports

Farah® casual crew socks ( Kayser-Roth Corp., upwind of the traps and olfactometer chamber. A

Greensboro, NC) with fiber content of 75% Hi-Bulk completely randomized design was used for cage positionacrylic, 24% nylon, and 1% Lycra®spandex were used ( 1= top; 2= middle cage; 3= bottom chamber) and port

only in field studies. For each study a human subject ( A or B). For each test, one port contained a candidate

wore the socks from 0600- 1800 for three consecutive attractant ( human arm, sock worn by human, ordays. When the socks were removed at 1800, they were Limburger cheese) and the other remained emptyallowed to lay on the floor by the shoes, i. e., they were ( control). Treatments( worn folded sock or Limburger

not placed in any special bag until the evening of the cheese) were presented in glass Petri dishes that were

third day when they were placed in a resealable plastic cleaned and sterilized in a vacuum oven and then handled

bag for transport to the laboratory on the following only with gloves to minimize any chance of

morning. Two subjects, the author of this paper( DLK) contamination. A small block( ca. 12.4 gm) ofLimburger

and his son( JDK), contributed worn socks for the field cheese was placed on the sterilized glass petri dish. A

studies. White socks were worn by both subjects and fresh piece of cheese was used for each test. Treatments

were designated as DLKWS and JDKWS for these were alternated between ports. After a three minute

studies. Black socks were worn only by DLK(designated exposure period, the relative numbers caught in the

as DLKBS) in the same manner as the white socks. treatment and control traps upwind of each port were

Human hand. The left hand of the author was also recorded.

used as a treatment in the olfactometer studies.

Substances obtained from human hands have been shown Field Trapping Studiesto be attractive to several mosquito species in previous Field trapping experiments were conducted fromolfactometer studies( Schreck et al. 1981, 1990). May through June 1997 against natural populations of

Carbon dioxide. Carbon dioxide was supplied woodland mosquito species associated with wooded

from 9 kg( 20 lb) compressed gas cylinders. The flow wetlands located near a water management area in

rate used for all trap- bait treatments that included CO2 Gainesville, FL. ABC' s Pickle Jar( PJ) trap was used forwas 500 ml/min. Control ofCO2 flow rate was achieved all the field trials. This new trap design, which has notwith American Biophysics( ABC, East Greenwich, RI) yet been made commercially, utilizes a novel, patentFLOWSET1, which is comprised ofa pressure regulator pending, counterflow concept for capturingwith output fixed at 15 prig, a 10 micron line filter, a 500 mosquitoes. The trap is constructed from a clear PVCml/min flow control orifice, and quick-connect luer container( ca. 11. 4 liter volume). The trap design andfittings. operating principle have been described in detail by

Kline( 1998). The trap requires 12 VDC; in these fieldOlfactometer Studies studies two Panasonic 6 volt, 10 amp- hr gel- cells,

To determine if volatiles emanating from a human connected in series were used. Traps were placed at wellworn sock and Limburger cheese elicited an upwind

defined sampling stations. At each station the trap wasorientation response, six to eight day old, laboratory- hung from a pole so that the bottom tube, through whichreared, host- seeking Ae. aegypti were tested in a dual- the attractant is released, was ca. 50 cm( 19. 69 in) aboveport olfactometer, illustrated and described in detail by ground level.

Posey et al( 1998). This system allows mosquitoes to Paired comparisons were made between CO2- baitedchoose between two different stimuli. The olfactometer PJ traps with and without the addition of a human worn

is constructed of clear acrylic, comprises three test sock. Treatments were alternated on a daily basis betweenchambers in a tiered configuration, has paired removable two sites. Two studies were conducted between Maysleeves, and mosquito traps on each chamber, and is 14- 22, 1997. In the first study a sock worn by the authorequipped with a filtered external air supply system that was taped near the bottom of one of the PJ traps. In the

allows precise temperature (± 0.5 ° C) and relative second study a white crew sock worn by the author' s sonhumidity (± 2%) control. Only one chamber at a time ( JDK) was taped in the same way to another PJ trap.was used for assays. Outside air was conditioned prior Each day the mosquitoes were identified and counted.to entry through the choice ports, the mosquito trap, and Statistical analysis of the trap collection data consistedtheolfactometerbypassingthroughaseriesofcharcoal of the paired difference t- test ( SAS PROC


UNIVARIATE [ SAS Institute 1985]). The data from compared to 0. 1% for the control; and the human handeach study were analyzed separately. resulted in an 80. 1% mean response compared to 0. 1%

Three additional field studies were conducted for the control( TABLE 1).

using a 4 x 4 Latin square experimental design. Fourtraps were located ca. 18 m apart. Each trap was Field Studies. Species Compositionprovided with a different bait daily for ca. 16 hours Twenty- two species of mosquitoes were collected1800- 0600 Eastern Daylight Savings Time[ EDST]). during these experiments. In order of decreasing

In each experiment the four bait combinations were abundance they were: Aedes atlanticus Dyar and Knab,alternated so that each bait occupied each trap station for Anopheles crucians Wiedemann, Aedes infirmatus Dyara single nigh.. In the first Latin square experiment( May and Knab, Psorophora ferox( Von Humboldt), Aedes26- 30, 1997) the treatments were no bait, DLK black canadensis ( Theobald), Aedes dupreei ( Coquillett),sock, CO2 only, and CO2+ DLK black sock. In the Coquillettidia perturbans( Walker), Culiseta melanurasecond experiment ( May 30- June 02, 1997) the ( Coquillett), Aedes vexans( Meigen), Culexnigripalpustreatments were no bait, Limburger cheese, CO2 only, Theobald, Culex salinarius Coquillett, Culexand CO2 + Limburger cheese. A small fresh block( ca. ( Melanoconion) spp. Theobald, Aedes fulvus-pallens12. 4 gm) ofLimburger cheese was used each day. It was ( Wiedemann), Psorophora howardii Coquillett,placed in a small sandwich bag, which was taped to the Psorophora ciliata ( Fabricius), Anopheles quadri-bottom of the PJ trap so that the volatiles could escape. maculatus Say, Aedes triseriatus( Say), Aedes sticticusIn the third experiment ( June 08- 12, 1997) the ( Meigen), Culex quinquefasciatus Say, Psorophoratreatments were no bait, DLK white crew sock only, columbiae ( Dyar and Knab), Anopheles punctipennis

CO2 only, and CO2 + DLK white crew sock. Each day ( Say), and Uranotaenia sapphirina ( Osten Sacken)the mosquitoes were identified and counted. Each ( TABLES 2- 6).

experiment was statistically analyzed separately. In allthese Latin square experiments the data were Paired Tests

transformed to log( n+ 1) for analysis of variance. The DLKWS+ CO2 vs CO2 only. In this experiment 18transformed data were analysed with Statistical of the species mentioned above were collected( TABLE

Analysis System ( SAS) PROC ANOVA and Means/ 2). No Cx. quinquefasciatus or Ur. sapphirina were

REGWQ for the analysis of variance and means collected in either trap. Overall, 2.09x more mosquitoescomparisons( SAS Institute 1985). were collected when the trap was baited with both the

worn sock and CO2 compared with CO2 only baitedRESULTS traps. However, the effectiveness of the addition of the

worn sock varied with species. FewerAe. triseriatus( 4

Olfactometer Studies vs 5) and only one more specimen of Cx. salinarius( 21In the olfactometer trials utilizing femaleAe. aegypti, vs 20, index= 1. 05) were collected when the sock was

Limburger cheese resulted in a mean attractive response combined with the CO2 compared to CO2 alone. Theof 6.4% versus 0.0% for the simultaneously run control; index for the other 15 species ranged from 1. 35x morethe worn sock resulted in a 66. 1% mean response mosquitoes collected with the addition of the sock for

TABLE 1. Mean1,2 percent attraction ( SE) of Aedes aegypti to humanhand, human worn sock, and Limburger cheese odors in dual-port olfactometer.

Treatment Mean% Responding( S. E.)

Hand 80. 12 2. 38) A

Sock 66.08 3. 21) B

Cheese 6.44 1. 47) C

Control 0.04 0.04) D

n= 10 for hand, sock and cheese treatments and 30 for control.

Means with the same letter are not significantly different; SAS PROCGLM/means REGWQ multiple range test( SAS Institute, 1985).


TABLE 2. Effect ofDLKWS worn sock on the mean number(± SE) of mosquitoes collected by PJ traps baitedwith CO2 ( 500 ml/ min)( n= 8 nights).

Species No Sock Sock ITI' Pr> 1T12 Index3

All species 615. 88 ( 44.42) 1285. 75( 112.93) 7.93 0.0001 2. 09Aedes atlanticus 224. 50 ( 22. 85) 486.75 ( 51. 43) 6. 86 0.0002 2. 17Ae. canadensis 48. 88 ( 10. 44) 77.63 ( 14. 07) 3. 88 0.0060 1. 59Ae. dupreei 40.25 ( 7. 86) 66.25 ( 7.92) 2.92 0.0224 1. 65Ae. fulvus-pallens 0.50 ( 0.33) 1. 88 ( 1. 13) 1. 55 0. 1643 3. 76Ae. infirmatus 98. 13 ( 9.45) 233. 38 ( 27. 98) 6. 42 0. 0004 2. 37Ae. sticticus 0. 13 ( 0. 13) 0. 75 ( 0.25) 2. 38 0. 0492 5. 77Ae. triseriatus 0.63 ( 0.26) 0. 50 ( 0. 38) - 0.23 0. 8264 0.79Ae. vexans 10.38 ( 2. 05) 14. 00 ( 2. 82) 1. 45 0. 1892 1. 35Anopheles crucians 105. 50 ( 14. 97) 173. 25 ( 27. 60) 3. 39 0.0116 1. 64An. punctipennis 0.25 ( 0. 16) 0. 13 ( 0. 13) - 0. 55 0.5983 0.52An quadrimaculatus 0.50 ( 0.27) 0.75 ( 0. 31) 0.68 0.5165 1. 50Coquillettidia perturbans 10. 50 ( 1. 87) 16.00 ( 2. 81) 1. 54 0. 1663 1. 52Culex( Melanoconion) spp. 4. 88 ( 0.77) 9. 38 ( 2. 22) 2. 17 0.0663 2. 50Cx. nigripalpus 0.63 ( 0.42) 1. 75 ( 0.65) 2. 35 0.0514 2. 78Cx. salinarius 2.50 ( 0.65) 2. 63 ( 0.42) 0.20 0.8436 1. 05Culiseta melanura 4. 88 ( 0.77) 9. 38 ( 2. 22) 2. 17 0.0663 1. 92Psorophora ciliata 0.50 ( 0.38) 1. 63 ( 0.98) 1. 76 0. 1218 3. 26Ps. ferox 65. 13 ( 8. 00) 193. 00 ( 26. 08) 5. 51 0.0009 2.96Ps. howardii 0.63 ( 0.26) 3. 38 ( 0. 82) 2. 92 0.0222 5. 37

T test statistic for paired- difference t- test determined by SAS Institute Proc Univariate.2Pr> ITI values less than 0.05 indicate the average difference is significantly different from zero.3Index= # mosquitoes in sock+ CO2- baited trap

mosquitoes in CO2- only baited trap

Ae. vexans to 5. 37x for Ps. howardii. The collection size mosquito species were caught by traps baited with atof the Psorophora spp. was increased the most by the least one of the treatments( TABLE 4). The trap baitedaddition of the worn sock to CO2. with DLKBS worn sock + CO2 caught 16 species

JDKWS+ CO2 vs CO2 only. No An. punctipennis, followed by 13 species for the CO2 only baited trap, 12Cx. quinquefasciatus, Ps. ciliata, Ps. columbiae, or Ur. species for the sock only baited trap and 7 species for thesapphirina were collected( TABLE 3). Overall 2. 45x unbaited trap. Total mosquitoes collected in the wornmore mosquitoes were collected when the sock was sock only trap was 3. 71x greater than that collected inused in combination with the CO2 compared to CO2 the unbaited trap. Individual species varied in theiralone. Only a few specimens of Ae. sticticus, Ae. relative collection size between sock only baited andtriseriatus, and An. quadrimaculatus were collected unbaited trap from 1. 6x ( Ae. canadensis) to 10.5xand these were in the traps baited with the sock+ CO2. ( Ae. infirmatus) greater for the sock baited trap. TheAlthough all the other species were collected by traps sock only baited trap caught three species consideredwith and without the worn socks, the sock baited traps to be major annoyance pests ( Ae. atlanticus, Ae.always had at least twice the collection size as the CO2 infirmatus and Cq. perturbans) at 0. 18x, 0.25x andonly baited traps. The index for these other species 0. 14x, respectively, their collection level in CO2 onlyranged from 2. 11x for Ae. dupreei to 6.46x for Ae. baited traps. In traps baited with worn socks+ CO2 thesevexans. same three species were caught at 2. 39x, 2. 84x and

2. 48x greater, respectively, than in traps baited withLatin Square Experiments CO2 only.

1: DLKBS and CO2. In this experiment, 17 different 2: Limburger cheese and CO2.

Eighteen mosquito


TABLE 3. Effect of JDKWS worn sock on the mean number(± SE) of mosquitoes collected by PJ traps baitedwith CO2 ( 500 ml/ min)( n= 3 nights).

Species No Sock Sock ITI' Pr> ITI2 Index3

All species 378.67 ( 88. 05) 928.00 ( 129. 33) 11. 66 0.0073 2. 45

Aedes atlanticus 150.67 ( 38. 92) 362. 67 ( 45. 48) 7. 22 0.0187 2.41

Ae. canadensis 49.00 ( 14. 57) 115. 67 ( 21. 88) 6. 64 0.0219 2.36

Ae. dupreei 27.00 ( 4.04) 57. 00 ( 9. 00) 5. 12 0.0361 2. 11

Ae. fulvus-pallens 1. 67 ( 1. 20) 4.67 ( 2.40) 1. 44 0.2863 2. 80

Ae. infirmatus 51. 67 ( 14. 72) 127. 33 ( 16. 02) 5. 51 0.0314 2.46

Ae. sticticus 0.00 ( 0.00) 0.67 ( 0.33) 2.00 0. 1835

Ae. triseriatus 0.00 ( 0.00) 1. 00 ( 0.58) 1. 73 0. 2254

Ae. vexans 0.67 ( 0.67) 4. 33 ( 1. 20) 3. 05 0.0927 6.46

Anopheles crucians 21. 67 ( 7. 84) 56. 00 ( 6.66) 2.71 0. 1136 2.58

An quadrimaculatus 0.00 ( 0.00) 0. 33 ( 0.33) 1. 00 0.4226

Coquillettidia perturbans 3. 33 ( 0.88) 7. 67 ( 1. 33) 4.91 0.0390 2. 30

Culex( Melanoconion) spp. 1. 00 ( 0.58) 5. 33 ( 2.96) 1. 23 0.3442 5. 33

Cx. nigripalpus 0.67 ( 0. 33) 2. 67 ( 1. 45) 1. 15 0.3675 3. 99

Cx. salinarius 2. 33 ( 0. 33) 3. 00 ( 1. 73) 0.46 0.6914 1. 29

Culiseta melanura 2. 67 ( 0. 88) 8. 33 ( 2.91) 2.43 0. 1358 3. 12

Psorophora ferox 66.33 ( 21. 99) 171. 00 ( 44.64) 4. 35 0.0490 2. 58

Ps. howardii 0.00 ( 0.00) 0.33 ( 0. 33) 1. 00 0.4226

T test statistic for paired- difference t-test determined by SAS Institute Proc Univariate.2Pr> ITI values less than 0.05 indicate the average difference is significantly different from zero.3lndex= # mosquitoes in sock+ CO2- baited trap

mosquitoes in CO2- only baited trap

species were collected in traps during this experiment: maculatus and Ae. sticticus were collected only in the17 by CO2 only baited traps, 13 by traps baited with sock+ CO2 baited trap. Overall, mosquito collectionsLimburger cheese+ CO2 and 8 each in the Limburger in this experiment were only slightly better( 1. 33x) forcheese only and unbaited traps( TABLE 5). Only one the sock+ CO2 compared to CO2 only baited traps. Withspecimen of Cx. quinquefasciatus was collected and this the exception of Ae. infirmatus species, which in the

was in the unbaited trap. A total of only 38 mosquitoes previous field experiments had their collection size

were collected in each the Limburger cheese only baited > 1. 5x in traps baited with the combination sock+ CO2and unbaited traps. Overall, traps baited with Lim- compared to CO2 alone, still maintained the same typeburger cheese+ CO2 caught fewer( 0.84x) mosquitoes increase between the two baited trap types. Differentthan traps baited with CO2 alone. Only five species, species dominated during this last experiment.Ae. infirmatus ( 1. 05x), Ae. fulvus-pallens( 2.73x), Cs.

melanura ( 1. 22x), Cx. ( Melanoconion) spp. ( 1. 56x), DISCUSSION

and An. quadrimaculatus( 4.0x) were caught in greater

numbers in traps baited with Limburger cheese + These data show that there are chemical compounds

CO2. produced by the human foot that actively attract3: DLKWS awl CO2. Twenty species ofmosquitoes mosquitoes without the need for CO2. This was most

were collected during this experiment: six by unbaited, obvious in the olfactometer studies where a human sock

seven by sock only, 17 each by CO2 only and sock+ CO2 worn for three days consistently attracted ca. 50% of the

baited traps ( TABLE 6). Two specimens of Ur. available Ae. aegypti mosquitoes. The results of the

sapphirina were collected only in the sock alone baited field studies were also encouraging in that the socktrap; Ae. triseriatus and Ps. columbiae were collected alone baited trap attracted significantly( p< 0. 05) moreonly in the CO2 alone baited trap; and An. quadri- mosquitoes overall than did the unbaited traps in one of


TABLE 4. Mean' catch( standard error)( raw data) ofmosquitoes associated with woodlands in PJ traps baited withCO2 and/ or black human worn socks( DLKBS) near Gainesville, FL( May 26- 30, 1997).

Bait

Species None DLKBS CO22 DLKBS+CO2 Index'

All species 7.75 ( 2. 56) D 28. 75 ( 7. 89) C 223. 00 ( 23. 16) B 537.75 ( 73. 39) A 2.41Aedes atlanticus 5.00 ( 1. 87) C 17. 25 ( 6.26) B 95. 00 ( 17. 87) A 227.50 ( 38. 96) A 2.39Ae. canadensis 1. 25 ( 0.48) B 2.00 ( 0.91) B 31. 25 ( 5. 02) A 58. 75 ( 7. 82) A 1. 88Ae. dupreei 0.00 ( 0.00) C 0.50 ( 0.50) C 8. 75 ( 1. 65) B 21. 25 ( 0.48) A 2.43Ae. fulvus-pallens 0.00 ( 0. 00) B 0.00 ( 0.00) B 1. 75 ( 0.48) B 9. 00 ( 3. 03) A 5. 14Ae. infirmatus 0.50 ( 0. 29) D 5. 25 ( 1. 38) C 20.75 ( 7. 32) B 59. 00 ( 11. 05) A 2. 84Ae. triseriatus 0.00 ( 0.00) A 0.25 ( 0.25) A 0.25 ( 0. 25) A 0.50 ( 0.29) A 2. 00Ae. vexans 0.00 ( 0.00) A 0.00 ( 0. 00) A 0.00 ( 0.00) A 2. 25 ( 1. 11) AAnopheles crucians 0.25 ( 0.25) C 0.25 ( 0. 25) C 23.25 ( 5. 89) B 66. 50 ( 24.97) A 2. 86An. quadrimaculatus 0.00 ( 0.00) A 0.00 ( 0.00) A 0.25 ( 0.25) A 0. 75 ( 0.48) A 3. 00Coquillettidia perturbans 0.25 ( 0.25) B 0.75 ( 0.48) B 5. 25 ( 2. 14) B 12. 75 ( 3. 40) A 2. 48Culex( Melanoconion) spp. 0. 00 ( 0.00) B 0.25 ( 0.25) B 2.50 ( 0. 87) AB 5. 75 ( 2. 14) A 2. 30Cx. nigripalpus 0.00 ( 0.00) A 0.00 ( 0.00) A 0.00 ( 0.00) A 0.25 ( 0.25) ACx. quinquefasciatus 0. 25 ( 0.25) A 0. 25 ( 0.25) A 0.00 ( 0.00) A 0.00 ( 0.00) ACx. salinarius 0.00 ( 0.00) B 0. 75 ( 0.75) B 7. 00 ( 3. 67) AB 8. 75 ( 0.63) A 1. 25Culiseta melanura 0.00 ( 0.00) B 0. 50 ( 0.50) B 6.25 ( 0.48) A 3. 00 ( 1. 50) A 0.88Psorophora ciliata 0.00 ( 0.00) A 0.00 ( 0.00) A 0.00 ( 0.00) A 0.25 ( 0. 25) APs. ferox 0.25 ( 0.25) C 0.75 ( 0.48) C 20.75 ( 4. 61) B 59.00 ( 8. 46) A 2. 84

n= 4 nights; means in the same row followed by the same letter are not significantly different( P> 0.05);

REGWQ multiple range test( SAS Institute 1985) applied to log- transformed data.2CO2 release rate= 500 ml/min.

Index= mean# mosquitoes in DLKBS+ CO2 baited trapmean# mosquitoes in CO2 baited trap

the Latin square sock studies( TABLE 4). In the other Also encouraging is the persistence of the synergisticLatin square sock study( TABLE 5), relative mosquito effect observed with the worn socks over time observedcomposition had changed from those attracted strongly in the first paired difference study. This effect wasto the sock volatiles to those with little or no attractancy relatively consistent over the eight day period that thisto sock volatiles. Overall mosquito density was also study was conducted, which indicates that the attractantmuch reduced. In the first study Ae. atlanticus was the substances were probably bound up in the sock fibersmost abundant species, but in the last study An. crucians and subsequently slowly released. The relative collectionwas ca. 3. 5x more abundant than Ae. atlanticus. Aedes index(# mosquitoes captured in sock+ CO2 baited trap/atlanticus appeared to be more attracted to the sock mosquitoes captured in the CO2 only baited traps)alone baited trap than An. crucians. There also seemed remained relatively consistent ( 2. 13, 2. 16, 2. 28, 1. 73,to be a synergistic response by most mosquito species to 2. 21, 1. 68, 2. 84, and 1. 75 for days 1 through 8,the combination of CO2 + worn sock. respectively).

Another encouraging factor is that different kinds The data reported herein are also in basic agreementof socks and the same kind of socks worn by two with recent results obtained in Europe. Mboera et al.different people showed the same basic response patterns. 1998 reported that in dual- choice olfactometer studiesThe overall response ofmosquitoes toJDKWS appeared female Cx. quinquefasciatus were significantly moreto be slightly greater than to DLKWS. This is consistent attracted to foot skin emanations collected on nylonwith observations made over several years on relative stockings( by wearing them for four to five days betweenmosquito attractive responses to these two individuals 0600 hr and 1800 hr) than clean stockings. In contrast tounpublished data). these data, no significant interaction between CO2 and


TABLE 5. Mean' catch( standard error)( raw data) of mosquitoes associated with woodlands in PJ traps baited with

CO2 and/ or Limburger cheese near Gainesville, Fl( May 30- June 02, 1997).

Bait

Species None Cheese CO22 Cheese+ CO2 Index'

All species 7.50 ( 3. 28) B 8. 75 ( 4. 21) B 313. 00 ( 22. 89) A 305. 75 ( 86. 58) A 0.98

Aedes atlanticus 4.50 ( 2.72) B 5. 25 ( 2. 87) B 115. 50 ( 30.01) A 109. 25 ( 36. 83) A 0. 89

Ae. canad'' nsis 0.00 ( 0.00) B 0.75 ( 0.48) B 37. 25 ( 6. 18) A 34. 25 ( 10. 83) A 0.73

Ae. dupreei 0.00 ( 0.00) B 0.50 ( 0.50) B 14. 50 ( 7. 35) A 9. 25 ( 1. 75) A 0.98

Ae. fulvus-pallens 0. 25 ( 0.25) B 0.00 ( 0.00) B 2. 75 ( 1. 25) A 10. 75 ( 5. 34) A 2.73

Ae. infirmatus 0.00 ( 0.00) B 0.00 ( 0.00) B 24. 75 ( 7. 39) A 35. 50 ( 13. 15) A 1. 05

Ae. triseriatus 0. 00 ( 0.00) A 0.00 ( 0.00) A 0. 50 ( 0. 29) A 0.00 ( 0.00) A

Ae. vexans 0. 00 ( 0.00) A 0.00 ( 0.00) A 1. 00 ( 0. 41) A 0.00 ( 0.00) A

Anopheles crucians 0. 50 ( 0.29) B 0.75 ( 0.48) B 61. 50 ( 32. 12) A 55. 75 ( 15. 02) A 0. 63

An. quadrimaculatus 0.00 ( 0.00) B 0.00 ( 0.00) B 0. 25 ( 0.25) A 1. 00 ( 0.00) A 4.00

Coquillettidia perturbans 0.25 ( 0.25) B 0.25 ( 0.25) B 8. 50 ( 4. 50) A 12. 00 ( 3. 14) A 0. 85

Culex( Melanoconion) spp. 0.25 ( 0.25) B 0.25 ( 0.25) B 2. 25 ( 0.48) A 5. 00 ( 1. 78) A 1. 22

Cx. nigripalpus 1. 00 ( 1. 00) B 0.50 ( 0.50) B 8. 50 ( 3. 97) A 6.50 ( 4.27) A 0.91

Cx. quinquefasciatus 0.25 ( 0.25) A 0. 00 ( 0.00) A 0. 00 ( 0.00) A 0.00 ( 0.00) A

Cx. salinarius 0.50 ( 0.50) B 0. 00 ( 0.00) B 6. 50 ( 1. 19) A 8. 75 ( 3. 09) A 0.96

Culiseta melanura 0.00 ( 0.00) B 0. 50 ( 0.50) B 5. 50 ( 1. 32) A 6. 25 ( 0.95) A 1. 22

Psorophora ferox 0.00 ( 0.00) C 0.00 ( 0.00) C 23. 50 ( 7. 10) B 11. 50 ( 2.63) A 0.57

Ps. howardii 0.00 ( 0.00) A 0.00 ( 0.00) A 0.25 ( 0.25) A 0.00 ( 0.00) A


REGWQ multiple range test( SAS Institute 1985) applied to log- transformed data.

2CO2 release rate= 500 ml/min.

Index= mean# mosquitoes in Limburger Cheese+ CO2 baited trapmean# mosquitoes in CO2 baited trap

foot emanations was observed in these laboratory ( occurred in very low numbers), and Ae. infirmatus

experiments. Also, these investigators have so far failed collections were slightly increased with the combinationto attract wild Cx. quinquefasciatus in Tanzania to of cheese+ CO2.unlighted CDC traps baited with worn stockings Why our Limburger cheese failed to yield a goodMboera, unpublished data). attractive response from mosquitoes either in the

The initial results of the Limburger cheese studies olfactometer or field studies is interesting since Knolswere not as encouraging. Response of Ae. aegypti to ( 1996) reported such good responses for An. gambiae

Limburger cheese in the olfactometer was relatively s. s. Knols( personal communication) suggested that too

low (6. 4%). Similar responses were noticed for most much cheese was used in these studies, which may havemosquito species in the field studies. No significant resulted in repelling some mosquitoes. He suggesteddifference was noted for any species in unbaited versus that a study needs to be conducted on the dose- responseLimburger cheese only baited traps. There was no behavior of Ae. aegypti to Limburger cheese. Further

similar synergistic effect between Limburger cheese+ discussion with Knols concerning Limburger cheeseCO2 as was noted with the worn sock+ CO2 baited traps. revealed that there are many different kinds ofLimburgerIn fact, the addition of Limburger cheese resulted in a cheese. Knols ( 1996) did not indicate the brand of

slight reduction in overall mosquito collection size. The Limburger cheese used for tests in The Netherlands.

combination of Limburger cheese+ CO2 resulted in a Indeed, other Dutch colleagues have since provided

significant decrease( p< 0.05) in Ps. ferox collections several brands of Limburger cheese from The

when compared to CO2 only baited traps. In contrast Netherlands for evaluation. Preliminary studies in theCulex spp., Cq. perturbans, An. quadrimaculatus olfactometer with one of these brands against A e. aegypti


TABLE 6. Mean' catch( standard error)( raw data) of mosquitoes associated with woodlands in PJ traps baited with

CO2 and/ or White human worn socks( DLKWS) near Gainesville, Fl( June 8- 12, 1997).

Bait

Species None DLKWS CO22 DLKWS+ CO Index3

2

All species 1. 50 ( 1. 22) B 4.00 ( 1. 22) B 178. 50 ( 36.97) A 237.00 ( 37. 68) A 1. 33

Aedes atlanticus 0.00 ( 0.00) C 1. 25 ( 0.48) C 15. 75 ( 3. 86) B 34.25 ( 3. 50) A 2. 17

Ae. canadensis 0.25 ( 0.25) B 0.50 ( 0.29) B 18. 75 ( 5. 45) A 23. 19 ( 11. 59) A 1. 89

Ae. dupreei 0.00 ( 0.00) B 0.00 ( 0.00) B 0.75 ( 0.48) B 3. 00 ( 0.71) A 4.00

Ae. fulvus-pallens 0.25 ( 0.25) B 0.00 ( 0.00) B 0. 50 ( 0.29) B 4.25 ( 1. 55) A 8. 50

Ae. infirmatus 0.00 ( 0.00) B 0.00 ( 0.00) B 4. 50 ( 1. 19) A 5. 75 ( 1. 93) A 1. 28

Ae. triseriatus 0.00 ( 0.00) A 0.00 ( 0.00) A 0.25 ( 0.25) A 0.00 ( 0.00) A

Ae. vexans 0.00 ( 0.00) B 0.00 ( 0.00) B 1. 50 ( 0.65) A 1. 75 ( 1. 18) A 1. 17

Anopheles crucians 0.00 ( 0.00) B 1. 00 ( 0.41) B 67. 00 ( 20.24) A 74.00 ( 19. 40) A 1. 10

An. quadrimaculatus 0.00 ( 0.00) A 0.00 ( 0.00) A 0.00 ( 0.00) A 0.25 ( 0.25) A

Coquillettidia perturbans 0.25 ( 0.25) B 0.50 ( 0.50) B 21. 00 ( 5. 70) A 17. 50 ( 9. 54) A 0. 83

Culex( Melanoconion) spp. 0.00 ( 0.00) B 0.50 ( 0.29) B 3. 25 ( 1. 03) A 7.00 ( 1. 22) A 2. 15

Cx. nigripalpus 0.50 ( 0.50) B 0.00 ( 0.00) B 12. 25 ( 5. 71) A 19. 75 ( 9. 68) A 1. 61

Cx. salinarius 0.00 ( 0.00) B 0.25 ( 0.25) B 5. 25 ( 1. 55) A 5. 25 ( 2. 46) A 1. 00

Culiseta melanura 0.25 ( 0.25) B 0.00 ( 0.00) B 20.75 ( 9.08) A 15. 50 ( 6. 17) A 0.75

Psorophora ciliata 0.00 ( 0.00) A 0.00 ( 0.00) A 0.75 ( 0.75) A 0.25 ( 0.25) A 0.33

Ps. ferox 0.00 ( 0.00) C 0.00 ( 0.00) C 4.75 ( 1. 80) B 11. 00 ( 2. 27) A 2.32

Ps. howardii 0.00 ( 0.00) B 0.00 ( 0.00) B 1. 50 ( 0.65) A 2.00 ( 1. 68) A 1. 50


REGWQ multiple range test( SAS Institute 1985) applied to log- transformed data.2

CO2 release rate= 500 ml/min.

3Index= mean# mosquitoes in DLKWS+ CO2 baited trapmean# mosquitoes in CO2 baited trap

using the same amount of cheese as before has resulted studies. I want to thank H. T. McKeithen for his

in a mean attractive response of 24. 3% compared to technical assistance and for performing the statistical6. 4% for our American brand. Thus, the brand ofcheese analyses of these data. I also want to thank K. Posey formay also be acontributing factor to the response behavior his assistance in the conduct of the olfactometer tests.

of the test mosquitoes. Further studies are planned to see

if these trends remain consistent. REFERENCES CITED

Cooperative studies are in progress with analytical

chemists to determine what volatiles are present in the De Jong, R. and B. G. J. Knols. 1995. Selection ofbitingtwo types of Limburger cheese and in worn socks. So sites on man by two malaria mosquito species onfar 24 volatile compounds have been identified from the man. Experientia 51: 80- 84.

worn socks and testing in the dual port olfactometer Hall, D. R., P. S. Beevor, A. Cork, B. F. Nesbitt, and G.

with these compounds has begun. Several of these A. Vale. 1984. 1- Octen- 3- ol: A potent olfactorycompounds have elicited> 50% attractive response by stimulant and attractant for tsetse isolated from

Ae. aegypti alone and in combination. Field tests with cattle odours. Insect Sci. Appl. 5: 335- 339.

these combinations are in the planning stage. Kline, D. L. 1994. Olfactory attractants for mosquitosurveillance and control: 1- octen- 3- ol. J. Am.

Acknowledgments Mosq. Contr. Assoc. 10: 280- 287.Kline, D. L. 1998. Comparison of two American

I want to thank American Biophysics Corporation, Biophysics mosquito traps: the professional and a

East Greenwich, RI, for supplying the traps for these new counterflow geometry trap. J. Am. Mosq.


Contr. Assoc. ( In Press). Culex quinquefasciatus Say( Diptera: Culicidae) inKnols, B. G. J. 1996. Odour-mediated host-seeking a dual- choice olfactometer. J. Vect. Ecol. 23( 2):

behaviour of the Afro- tropical malaria vector 107- 113

Anopheles gambiae Giles. Ph.D. Thesis, University Posey, K. H., D. R. Barnard, and C. E. Schreck. 1998.of Wageningen, The Netherlands, 213 pp. A triple cage olfactometer for evaluating mosquito

Knols, B. G. J. and R. De Jong. 1996. Limburger cheese Diptera: Culicidae) attraction responses. J. Med.as an attractant for the malaria mosquito Anopheles Entomol. 35: 330- 334.gambiae s. s. Parasitol. Today 12: 159- 161. SAS Institute. 1985. SAS Procedure Guide, version 6

Knols, B. G. J. and J. Meijerink. 1997. Odors influence ed. SAS Institute, Cary, NCmosquito behavior. Sci. Med. 4: 56- 63. Schreck, C. E., D. L. Kline, and D. A. Carlson. 1990.

Knols, B. G. J., J. J. A. van Loon, A. Cork, R. D. Mosquito attraction to substances from the skin ofRobinson, W. Adam, J. Meijerink, R. De Jong, and different humans. J. Am. Mosq. Contr. Assoc. 6:W. Takken. 1997. Behavioural and electrophysio- 406- 410.logical responses of the female malaria mosquito Schreck, C. E., N. Smith, and D. A. Carlson. 1981. AAnopheles gambiae ( Diptera: Culicidae) to material isolated from human hands that attractsLimburger cheese volatiles. Bull. Entomol. Res. female mosquitoes. J. Chem. Ecol. 8: 429- 438.87: 151- 159. Takken, W. and D. L. Kline. 1989. Carbon dioxide and

Mboera, L. E. G., B. G. J. Knols, W. Takken, and P. W. 1- octen- 3- ol as mosquito attractants. J. Am. Mosq.T. Huisman. 1998. Olfactory responses of female Contr. Assoc. 5: 311- 316.


Egg Maturation in Neotropical Malaria Vectors:One Blood Meal is Usually Enough

L. P. Lounibos', D. Couto Lima2, R. Lourenco-de- Oliveira2,R. L. Escher', and N. Nishimura'

University of Florida, Florida Medical Entomology Laboratory,200 9th St. SE, Vero Beach FL 32962, USA

2Departamento de Entomologia, Instituto Oswaldo Cruz,Av. Brasil 4365, 21045-900 Rio de Janeiro, RJ-Brazil

Received 30 June 1998; Accepted 28 September 1998

ABSTRACT: The requirement of more than one blood meal to complete a gonotrophic cycle is reportedhere to be infrequent among field-collected Anopheles( Nyssorhynchus), including Anopheles darlingi,South America' s most important malaria vector. The overall frequencies of host- seeking females that didnot develop eggs after one blood feeding were low for Anopheles darlingi (6.6%), Anopheles rondoni5.0%), and Anopheles triannulatus( 2. 2%); the majority of wild-caught females that did not develop eggs

after one blood meal were nulliparous. Laboratory- reared Anopheles albimanus and Anopheles albitarsis,other species of the same subgenus, were grown on enriched and impoverished larval diets to yield a range

of adult sizes and to examine relationships between egg maturation, body size, and blood meal intake.10.7% of nulliparous An. albimanus and 22.9% ofAn. albitarsis failed to mature eggs after sugar and oneblood meal, and shorter wings and smaller blood meals predisposed females to require multiple blood feedsfor egg maturation. Unmated An. albitarsis were significantly more likely than mated females not todevelop eggs. Multiple blood meals within a single gonotrophic cycle appear to be less important in thelife histories of neotropical Nyssorhynchus species compared to afrotropical malaria vectors.

Keyword Index: Anopheles( Nyssorhynchus), egg development, blood feeding, female mass, winglengths.

INTRODUCTION 1993). In West African localities, 42% of An. gambiaes. s. and 63% of An. funestus required more than one

The occurrence of multiple blood feeds within a blood meal to develop eggs( Brengues and Coz 1973),single gonotrophic cycle has received renewed attention and pre- gravids were more common in the dry seasonfrom vector ecologists studying mosquito disease vectors. ( Molineaux and Grammicia 1980). Field studiesFor Africa' s most important malaria transmitters elsewhere in the Old World tropics have established thatAnopheles( Cellia) gambiae s. l. and Anopheles( Cellia) other important malaria vectors commonly imbibe bloodfunestus, this habit is associated with endophily and the more than once during the first gonotrophic cycle( e. g.,easy accessibility of human blood as a nutrient source Reisen and Aslamkhan 1976, Mogi et al. 1995).Beier 1996, Lyimo and Takken 1993). Gillies( 1954, Laboratory investigations have demonstrated that

1955) demonstrated that the majority of field-collected the pre- gravid condition of African anophelines mayfemales of these two species that required blood more result from inadequate larval nutrition ( e. g. El-Akadthan once to mature eggs were recently emerged nullipars, and Humphreys 1990, Takken et al. 1998). The use offor which he coined the term ' pre-gravids' to describe the first blood meal to satisfy energy requirements in-these new emergences that needed two or more blood stead ofvitellogenesisismorecharacteristicofAnophelesmeals to complete the first gonotrophic cycle. than Aedes mosquitoes ( Briegel 1990). Observations

Approximately 20% of host-seeking An. gambiae s. l. that several species of sugar- deprived Anopheles tookand An.funestus were pre- gravid at two sites in Tanzania, repeated blood meals within a single gonotrophic cyclewhere the incidence of this phenomenon varied have led to the conclusion that this behavior" prevails inseasonally ( Gillies 1954, 1955; Lyimo and Takken many anopheline( s)"( Briegel and Hurler 1993).


Blood requirements for egg maturation have been The purpose of the present paper is to demonstrate

relatively poorly studied in neotropical anophelines, that multiple blood meals are not required for eggalthough multiple blood feeding within one gonotrophic maturation by most wild- caught, host- seekingcycle has been documented in a laboratory colony of Nyssorhynchus, including An. darlingi, South America' sAnopheles( Nyssorhynchus) albimanus( Briegel 1990, most important malaria vector ( Zimmerman 1992).

Briegel and Wirier 1993, Klowden and Briegel 1994), Secondly, we describe results of experiments onan important malaria vector in Central America laboratory colonies of two other Nyssorhynchus speciesFrederickson 1993). Davis( 1928) reported from field- to examine whether small body size or small blood

collected specimens in Argentina that Anopheles meals might predispose females to take blood more than

Anopheles) pseudopunctipennis, butnot several species once to mature eggs. Finally, our discussion comparesof Nyssorhynchus, required multiple blood meals to the multiple blood feeding habit in neotropical andmature eggs. The latter subgenus includes the majority afrotropical malaria vectors.

of species, such as Anopheles( Nyssorhynchus) darlingi The older literature on this topic proposed terms

and An. albimanus, regarded as important neotropical such as gonotrophic dissociation( Swellengrebel 1929),

malaria vectors( Deane 1988, Zimmerman 1992). Only gonotrophic discordance( Rao 1947), and gonotrophic

4- 10% of Anopheles ( Nyssorhynchus) nuneztovari, disharmony( Adam et al. 1964) to describe the absencefield-collected in western Venezuela, where this species or incompletion of egg development after a single bloodtransmits Plasmodium vivax to humans, required more meal. We consider these terms, which predate a better

than one blood meal to develop eggs ( Lounibos and understanding of the physiological processes involvedConn 1991). e. g., Washino 1977, Briegel 1990), to be unnecessary

By contrast, most unmated, sugar-deprived females in the context of the present paper.

of An. nuneztovari would not mature eggs after one

blood meal in the laboratory( Lounibos and Conn 1991). MATERIALS AND METHODS

Subsequently, circumstantial evidence indicated thatmating may be necessary for egg development in some Anopheline mosquitoes were collected 1800- 2100

Nyssorhynchus species, such as An. nuneztovari; during hours from human bait, using protocol approved by thecopulation females of such species receive substances University of Florida' s Institutional Review Boardneeded for ovarian maturation from enlarged male acces- ( Projects No. 94.299 and 96. 011), in the vicinity of

sory glands( Lounibos 1994a). A requirement for mating Porto Velho( site CR), State ofRondonia and Peixoto de

to complete egg development had been reported decades Azevedo( site PX), State ofMato Grosso, in Amazonian

earlier forAnopheles( Cellia) subpictus from India( Roy Brazil( TABLE 1). Mosquitoes were transported live in

1940). Among An. nuneztovari, the probability of humidified containers to laboratories and, on the morning

maturing some eggs was significantly increased by a after collection, lightly anesthetized for specieslarge blood meal( Lounibos and Conn 1991). identifications under a dissecting microscope. Females

TABLE 1. Rates of insemination, parity, and need for another blood meal to develop eggs(= To with stationaryovaries) among field-collected Anopheles( Nys.) females.

No. with

Collections Blood No. No. stat. Parous

Species Site Date( s) sources dissect. insemin. ovaries % rateb( n)

An. darlingi CR 21- 28/ VI1195 M,Q 48 11 22.9 39.4( 94)

An. darlingi CR 11/ X-9/ XI/95 Q 140 140 2 1. 4 —

An. darlingi CR 4/ IV-9/ V/96 Q 120 119 8 6. 7 —

An. darlingi PX 20NII/96 G 83 5 6. 0 30. 5( 59)

An. rondoni PX 14- 18/ VII/96 G 40 2 5. 0 49. 0( 37)

An. triannulatus CR 26/ V11/ 95 Q,H 26 26 0 0. 0 7. 7( 13)

An. triannulatus CR 4/ IV-9/ V/96 Q 20 20 1 5. 0 —

all= human, G= guinea pig, M= mouse, Q= quail

bBased on dissections of non- fed females collected from the same site on the same date.


with no signs ofrecent blood were confined individually Nyssorhynchus excrete 23- 40% by weight of ingestedin 9- dram vials containing moist cotton and filter paper fluid( Perez de Valderrama and Scorza 1976). Ovarian

and fitted with screened lids. Within two days of development was allowed to proceed for three days at

collection, females were blood- fed to repletion on a 27± 0.5°C forAn. albimanus or 28± 1° C forAn. albitarsis,

restrained animal or, for species considered low risk for whereafter stages of development were assessed, as

malaria transmission, on one of the investigators. After indicated previously, by dissection of females that didthe blood meal, females were maintained for two to not oviposit in vials. Individual An. albitarsis, but not

three days without access to sugar in ambient- temperature An. albimanus, were examined for insemination. In

insectaries at 26± 2°C. Females rarely oviposited during comparable conditions of a previous study ( Lounibosthe post- blood holding period prior to dissection. 1994a), all females of colony An. albimanus held in

Egg development was assessed from ovaries cages with males were inseminated. Wing lengths weredissected in saline( Hayes 1953), which were classified measured as described in Lounibos ( 1994b). The

as either stationary(< stage 2 of Christophers[ 19111) or relationships between wing lengths, mass after blooddeveloping > stage 3 of Christophers). For the vast feeding, and the proportions of females developing eggsmajority of specimens, the classification decision was were tested by logistic regression with the procedurestraightforward, most developing follicles being in stages CATMOD of SAS( 1985) for fit to the sigmoid function:

4 or 5 of Christophers at the time of dissection. The

detection ofstationary ovaries indicated that a subsequentln[ f/(1- f)]= A+ B( wing length)+ C(blood meal mass)

blood meal would be required to complete eggwhere f is the proportion maturing eggs and A, B, and

maturation. Ovaries classified as stationary were rinsedC are parameters to be estimated ( Trexler and Travis

in distilled water and allowed to dry on microscope1993). Previous studies support the use of similar

slides forparity determinations by the tracheolar skeininglogistic regression models to describe the relationship

method ( Detinova 1962). Among certain collections,

between measures of mosquito body size and suchspermathecae of dissected females were placed in saline

binary variables as egg development or infection withand examined for the presence of sperm under high

power of a compound microscope. Non- blood fedparasites ( e. g., Lounibos 19946, Lyimo and Koella

specimens from four collections were dissected to obtain1992).

parity incidence estimates of the sampled field RESULTSpopulations.

Laboratory colonies of two An.( Nys.) species were Estimates From Field Collections

used to obtain females of known developmental history Four collections of An. darlingi from two sites at

and age. General colony maintenance was as described three times of year yielded a range from 1. 4 to 22.9% of

previously for An. albimanus originating from El females that did not mature eggs after blood( TABLE 1).

Salvador( Ford and Green 1972) and An. albitarsis( s. s.) Only 6.6% of all An. darlingi dissected( n= 391) failed

from Brazil ( Horosko et al. 1997). Larvae of An. to develop eggs after one blood meal. The frequency ofalbimanus were fed a suspension of finely ground liver stationary ovaries after blood feeding was also low forpowder and Brewer' s yeast and those of An. albitarsis An. triannulatus( 2. 2%) and An. rondoni( 5.0%).

a pollen- fish food-dog chow mixture ( Lounibos et al. Among An. darlingi with stationary ovaries( n= 26)1998). Diet concentrations were adjusted to yield adults after blood, three ( 11. 5%) were parous. Insemination

of varied sizes ( Lounibos 1994b). Newly emerged rates of field-collected females were high forAn. darlingi

females were allowed access to 10% sucrose on the first ( 99.6%) and An. triannulatus( 100%)( TABLE 1). Parityday after eclosion and maintained in the presence of an estimates at times of collection ranged from 7. 7% for

excess ofmales for the first two days. On day 3, females An. triannulatus to 49% for An. rondoni.

were allowed to consume blood from a restrained chicken

An. albimanus) or a human arm( An. albitarsis). Within Size Variables and Egg Maturation among Nyssor-two hours after blood feeding but after diuresis was hynchus Colonies

complete, masses of lightly anesthetized specimens Among laboratory- reared, nulliparous females,were determined gravimetrically on an electrobalance 10.7% of An. albimanus and 22. 9% of An. albitarsis

Lounibos and Conn 1991), after which females failed to develop eggs after one blood meal( TABLE 2).recovered in individual vials. The purpose was to estimate The distribution ofwing lengths was significantly skewedrelative blood mealintakeofalargenumberofspecimens, to the right among An. albimanus only, indicating anrather than precisely measure blood meal volume( e. g., overabundance of larger females of this species. WingBriegel 1986). During diuresis, related species of lengths and mass after blood feeding were significantly


correlated in both species. was relatively uncommon among these NyssorhynchusOf the An. albitarsis, 11. 0% ( n= 38) were not species. Most collected females were nulliparous and,

inseminated, and 65. 8%( n= 25) of these did not mature in contrast to pre- gravid afrotropical malaria vectorseggs. A G- test for independence using Williams' caught athosts( Gillies 1954, 1955), inseminated. Thesecorrection( Sokal and Rohlf 1995) indicated that failure observations suggest a different sequence ofreproductiveto develop eggs and lack of insemination were events for the neotropical anophelines after emergencesignificantly associated ( Gadi= 35. 2, P< 0.001) for An. compared to afrotropical malaria vectors. Whereas

albitarsis. teneral An. gambiae usually seek blood before mating,Both mass after blood meal and wing length nearly all An. darlingi and An. triannulatus were

explained significant variation in the logistic regression inseminated before their first blood meal. This trendmodel for the proportion of An. albimanus maturing was also observed for Nyssorhynchus field-captured ineggs after one blood meal( TABLE 3). ForAn. albitarsis, Trinidad, where more than 90% of host- seekingonly mass after blood meal was statistically significant. Anopheles aquasalis and more than 97% ofAn. albitarsisLikelihood ratios showed a significant deviation from ( s. l.) had been inseminated( Senior White 1955). Sincethe logistic regression model for An. albitarsis but not the principal afrotropical vectors are more endophagicfor An. albimanus. and endophilic than their neotropical counterparts( e. g.

Beier 1996 cf. Zimmerman 1992), the availability ofDISCUSSION human blood lends itself to more multiple blood feeding

in the African situation. Less frequent contact withOur preliminary data from field collections of host- human hosts by Nyssorhynchus species may be one

seeking females indicate that the pre- gravid condition cause of low malaria infection rates among neotropical

TABLE 2. Mean wing lengths and masses after blood feeding, plus associated statistics, for laboratory- rearedAnopheles( Nyssorhynchus).

Mean mass Mean

with after wingstat. blood length

Species n ovaries ( mg) S. D. ( mm) S. D. C.V. ra

skewnesstb

An. albimanus 564 10. 7 2. 84 0.67 3. 26 0. 17 5. 22 0. 567** - 0. 37 3. 60***

An. albitarsis 347 22.9 2. 58 0.46 3. 56 0.20 5. 62 0. 434** 0. 04 0. 30ns

a product-moment correlation between mass after blood and wing lengthbt- test for significance of skewness( Sokal and Rohlf 1995)

P< 0.01, ***= P< 0.001, ns= non- significant

TABLE 3. Maximum likelihood analyses of variance for logistic regressions of eggdevelopment( dependent, binary variable) versus wing length and mass afterblood feeding( independent, numeric variables).

Species n source d. f. xz P

An. albimanus 564 mass after blood 1 52.02 < 0.001

wing length 1 18. 65 < 0.001

likelihood ratio 494 124. 33 1. 000

An. albitarsis 347 mass after blood 1 6.64 0.010

wing length 1 3. 17 0.075

likelihood ratio 141 206. 01 < 0.001


vectors( e.g., Arruda et al. 1986, Oliveira-Ferreira et al. relationship between the pre-gravid state and lack of1990, Frederickson 1993). insemination of An. albitarsis confirms other evidence

The size range of An. albimanus derived from for a role ofmating in egg maturation( Lounibos 1994a)enriched and impoverished larval diets mimicked the but does not, by itself, explain the relatively highdistribution of wing lengths and their coefficients of pre-gravid rate in the experimental cohort of thisvariation and skewness in natural populations of related species. The contrasting results between An. albimanusspecies( Lounibos 1994b). The preponderance of large- and An. albitarsis may be attributable to differentexperi-sized females in biting collections ( Lounibos 1994b) mental protocols, species- specific genetic properties,suggests that Nyssorhynchus species may be less subject or both.

in nature to the larval stresses that induce afrotropical The present study calls into question whethermalaria vectors to seek blood more frequently( Takken repeated host- seeking within a gonotrophic cycle iset al. 1998). Approximately 90% of the experimental normal behavior for An. albimanus, An. darlingi, andAn. albimanus matured eggs after one blood feeding, related species in the field. Even though sugar- deprivedconfirming a low incidence ofpre-gravids observed in a females of a strain of the former species were attractedsmaller sample of unmated females of the same species repeatedly to hosts and took multiple blood mealsLounibos 1994a). The average mass of An. albimanus during one egg cycle( Briegel and Hurler 1993, Klowden

after post-blood diuresis was similar to values calculated and Briegel 1994), there is reason to doubt that thisfor An. nuneztovari and An. darlingi ( Perez de behavior is common in natural, largely exophilicValderrama and Scorza 1976), species of similar size populations ( Frederickson 1993). Ovaries of host-based upon wing length comparisons( Lounibos 1994b, seeking An. darlingi captured in Brazil are rarelyLounibos et al. 1995). advanced to stage 3 of Christophers ( DCL, LPL, and

The probability of developing some eggs was RL-O, unpublished), and Senior White( 1955) reportedpositively correlated to both mass after blood feeding only 2% of wild-caught An. aquasalis in stages 3- 5.and wing length of An. albimanus. Although the chi- Although malaria transmission by An. albimanus andsquare value from regression estimates by maximum An. darlingi may occur indoors in some circumstanceslikelihood was much higher for post-blood mass than ( Zimmerman 1992), females of these species typicallywing length, multicollinearity of these two variables leave dwellings during the night or morning of bloodimpedes estimates of their independent impact on egg feeding ( Loyola et al. 1993, Lourenco-de- Oliveiramaturation( Trexler and Travis 1993). Although smaller 1994). Thus, replete females would need to reinitiateAn. albimanus are more likely to be pre- gravid, as the host- location process to refeed which, based uponobserved for An. gambiae( Lyimo and Takken 1993), a the data reported here, is usually unnecessary for eggmuch higher percentage of the small An. gambiae( 69. 2- maturation. The exophilic behavior of these neotropical100%) examined by Takken et al. ( 1998) required at species suggests that they might more readily consumeleast two experimentally delivered blood meals to mature plant sugar for energy, a behavior nearly discarded byeggs. the endophilic afrotropical vectors ( Beier 1996). The

The An. albitarsis females were generally larger scanty field data available indicate that nectar feedingthan An. albimanus, which may be attributable to rearing is common in female Nyssorhynchus ( Senior-Whiteconditions, such as a richer larval diet, or the temperate 1952).

source of this species from southern Brazil( Horosko etal. 1997). The laboratory rearings of An. albitarsis on Acknowledgmentshigh and low diets did not yield a skewed size distribu-tion, which in field samples of related species is We thank personnel ofFundacao Nacional da Saudecommonly negatively skewed ( Lounibos 1994b). ( FNS) for hospitality and assistance in Porto Velho andMoreover, An. albitarsis females took a smaller blood of USAMRU-Rio for the same in Peixoto de Azevedo.meal, on average, than An. albimanus in proportion to S. Horosko and J. Bento graciously provided An.their body size, as estimated from wing lengths. albitarsis s. s. from their colony. We are grateful to S.Perhaps related to the smaller blood meal and/ or its Juliano for statistical advice and to R. Darsie, C. Lord,source from the less nutritious human host ( Shelton W.Takken, R. Zimmerman, and an anonymous reviewer1972), a higher percent( 22.9%) of An. albitarsis were for critical comments on an earlierdraftofthemanuscript.pre- gravid compared to An. albimanus ( 10. 7%). The Research was supported in part by grants from NIH( AI-logistic regression model provided a poorer fit to the 31034) of the USA and CNPq, PAPES/ FIOCRUZ anddata for this species, with only mass after blood FNS of Brazil. This is Florida Agricultural Experimentexplaining significant variation. The significant Station, Journal Series No. R-06536.


REFERENCES C11'ED Gillies, M. T. 1954. The recognition of age- groups

within populations of Anopheles gambiae byAdam, J. P., G. Vattier, and F. X. Pajot. 1964. the pre-gravid rate and the sporozoite rate.

Dysharmonie gonotrophique chez deux Ann. Trop. Med. Parasit. 48: 58- 74.Anopheles cavernicoles du Congo ( Brazza- Gillies, M. T. 1955. The pre- gravid phase of ovarian

ville). Bull. Soc. Pathol. Exot. 57: 397-400. developmentinAnophelesfunestus. Ann. Trop.Arruda, M. de, M. B. Carvalho, R. S. Nussenzweig, M. Med. Parasit. 49: 320- 323.

Maracic, A. W. Ferreira, and A. H. Cochrane. Hayes, R. O. 1953. Determination of a physiological

1986. Potential vectors of malaria and their saline for Aedes aegypti( L.). J. Econ. Entomol.

different susceptibility to Plasmodium 46: 624-627.

falciparum and Plasmodium vivax in northern Horosko, S., J. B. P. Lima, and M. B. Brandolini. 1997.

Brazil identified by immunoassay. Am. J. Trop. Establishment of a free- mating colony ofMed. Hyg. 35: 873- 881. Anopheles albitarsis from Brazil. J. Am. Mosq.

Beier, J. C. 1996. Frequent blood- feeding and restrictive Contr. Assoc. 13: 95- 96.

sugar- feeding behavior enhance the malaria Klowden, M. J. and H. Briegel. 1994. Mosquito

vector potential ofAnophelesgambiaes. l. and gonotrophic cycle and multiple feedingAn. funestus ( Diptera: Culicidae) in Western potential: contrasts between Anopheles and

Kenya. J. Med. Entomol. 33: 613- 618. Aedes. J. Med. Entomol. 31: 618- 622.

Brengues, J. and J. Coz. 1973. Quelques aspects Lounibos, L. P. 1994a. Variable egg development amongfondamentaux de la biologie d' Anopheles Anopheles ( Nyssorhynchus): control bygambiae Giles( sp. A) et d' Anophelesfunestus mating? Physiol. Entomol. 19: 51- 57.Giles, en zone de savane humide d' Afrique de Lounibos, L. P. 1994b. Geographical and developmental

l' Ouest. Cah. ORSTOM Ser. Entomol. Med. components of adult size of neotropical

Parasit. 11: 107- 126. Anopheles ( Nyssorhynchus). Ecol. Entomol.

Briegel, H. 1986. Protein catabolism and nitrogen 19: 138- 146.

partitioning during oogenesis in the mosquito Lounibos, L. P. and J. Conn. 1991. Fecundity, parity andAedes aegypti. J. Ins. Physiol. 32: 455- 462. adult feeding relationships among

Briegel, H. 1990. Fecundity, metabolism and body size Nyssorhynchus malaria vectors from

in Anopheles( Diptera: Culicidae), vectors of Venezuela. Mem. Inst. Osw. Cruz 86: 57- 66.

malaria. J. Med. Entomol. 27: 839- 850. Lounibos, L. P., N. Nishimura, J. Conn, and R. Lourenco-

Briegel, H. and E. Horler. 1993. Multiple blood meals as de- Oliveira. 1995. Life history correlates ofa reproductive strategy in Anopheles( Diptera: adult size in the malaria vector Anopheles

Culicidae). J. Med. Entomol. 30: 975- 985. darlingi. Mem. Inst. Osw. Cruz 90: 769- 774.

Christophers, S. R. 1911. The development of the egg Lounibos, L. P., D. Couto Lima, and R. Lourenco- de-

follicle in Anophelines. Paludism 2: 73- 88. Oliveira. 1998. Prompt mating of releasedDavis, N. C. 1928. Notes on the development of ovarian Anopheles darlingi in western Amazonian

follicles in Argentine Anopheles. Am. J. Hyg. Brazil. J. Am. Mosq. Contr. Assoc. 14: 210-8: 467- 473. 213.

Deane, L. M. 1988. Malaria studies and control in Lourenco- de- Oliveira, R. 1994. Qual a importancia de

Brazil. Am. J. Trop. Med. Hyg. 38: 223- 230. hematofagia extradomiciliar do Anopheles

Detinova, T. S. 1962. Age-grouping methods in Diptera darlingi, na Amazonia? Rev. Patol. Trop. 23:of medical importance. WHO, Geneva, 216 p. 100- 101.

El-Akad, A.S. and J. G. Humphreys. 1990. Effects of Loyola, E. G., L. Gonzalez-Ceron, M. H. Rodriguez, J.

larval and adult diet plus mating/ insemination J. Arredondo-Jimenez, S. Bennett, and D. N.

upon ovarian development of laboratory- reared Bown. 1993. Anopheles albimanus ( Diptera:

Anopheles pharoensis in Egypt. J. Am. Mosq. Culicidae) host selection patterns in three eco-

Contr. Assoc. 6: 96-98. logical areas of the coastal plains of Chiapas,

Ford, H. R. and E. Green. 1972. Laboratory rearing of southern Mexico. J. Med. Entomol. 30: 518-

Anopheles albimanus Wiedemann. Mosq. 523.

News 32: 509- 513. Lyimo, E. O. and J. C. Koella. 1992. RelationshipFrederickson, E. C. 1993. Bionomics and control of between body size of adult Anopheles gambiae

Anopheles albimanus. PAHO Tech. Paper No. s. l. and infection with the malaria parasite34, 76 p.+ vi. Plasmodium falciparum. Parasitol. 104: 233-


237. Senior White, R. A. 1952. Studies on the bionomics ofLyimo, E. O. and W. Takken. 1993. Effects of adult Anopheles aquasalis Curry, 1932 ( concld.).

body size on fecundity and the pre-gravid rate Ind. J. Malar. 6: 29- 72.of Anopheles gambiae females in Tanzania. Senior White, R. A. 1955. On the evening biting activityMed. Vet. Entomol. 7: 328- 332. of three neotropical Anopheles in Trinidad,

Mogi, M., I. Miyagi, T. Toma, M. Hasan, K. Abadi and British West Indies. Bull. Entomol. Res. 44:Syafruddin. 1995. Age structure ofAnopheles 451- 460.

subpictus ( Diptera: Culicidae) collected by a Shelton, R. M. 1972. The effects of blood source andlight trap in Halmahela, Indonesia. S. E. Asia J. quantity on production of eggs by CulexTrop. Med. Publ. Hlth. 26: 760- 766. salinarius Coquillett ( Diptera: Culicidae).

Molineaux, L. and G. Gramiccia. 1980. The Garki Mosq. News 32: 31- 37.Project. WHO, Geneva, 311 p. Sokal, R. R. and F. J. Rohlf. 1995. Biometry. 3rd edit.

Oliveira-Ferreira, J. de, R. Lourenco-de- Oliveira, A. W.H. Freeman, NY, 887 p. + xix.

Teva, L. M. Deane, and C. T. Daniel- Ribeiro. Swellengrebel, N. H. 1929. Ladissociation des fonctions1990. Natural malaria infections in anophelines sexuelles et nutritives ( dissociation gontro-in Rondonia State, Brazilian Amazon. Am. J. phique) d' Anopheles maculipennis comme

Trop. Med. Hyg. 43: 6- 10. cause du paludisme dans les pays- bas et sesPerez de Valderrama, M. and J. V. Scorza. 1976. rapports avec l' infection domiciliare. Ann. Inst.

Experimentos e indagaciones para explicar el Past., Paris 43: 1370- 1389.

comportamiento evasivo de Anopheles nunez- Takken, W., M. J. Klowden, and G. M. Chambers.tovari Gabaldon, 1940 de Venezuela. Bol. Div. 1998. Effect of body size on host seeking andMalar. San. Amb. 16: 212- 220. blood meal utilization in Anopheles gambiae

Rao, V. V. 1947. On gonotrophic discordance among sensu stricto ( Diptera: Culicidae): the

certain Indian Anopheles. Ind. J. Malar. 1: 43- disadvantage of being small. J. Med. Entomol.50. 35: 639- 645.

Reisen, W. K. and M. Aslamkhan. 1976. Observations Trexler, J. C. and J. Travis. 1993. Nontraditionalon the swarming and mating behaviour of regression analyses. Ecology 74: 1629- 1637.Anopheles culicifacies Giles in nature. Bull. Washino, R. K. 1977. The physiological ecology ofWld. Hlth. Org. 54: 155- 158. gonotrophic dissociation and related

Roy, D. N. 1940. Influence of spermathecal stimulation phenomena in mosquitoes. J. Med. Entomol.on the physiological activities of Anopheles 13: 381- 388.

subpictus. Nature 145: 747- 748. Zimmerman, R. H. 1992. Ecology of malaria vectors inSAS Institute Inc. 1985. SAS User' s Guide. Statistics. the Americas and future direction. Mem. Inst.

Version 5 edit. Cary, NC 956 p+ xvi. Osw. Cruz 87( suppl. III): 371- 383.

ADVERTISING

Commercial advertising space is available in the Journal; fullpage (black and white) at $ 150.00 per issue, half page (black

and white) at $ 90.00 per issue. Inquiries may be addressed toH. B. Munns, Advertising Manager, SOVE, P. O. Box 338,Pioneer, California 95666, telephone ( 209) 295- 3540; Fax

209) 295- 3563; or e- mail; munzy @volcano.net

The publisher reserves the right to approve or refuse anyadvertisement. The publisher is not responsible for anyclaims, litigations, or expenses resulting from the advertiser' sunauthorized use of any name, photograph, sketch, or wordsprotected by registered trademark or copyright.

volume 23, issue 2, december 1998

Documents