JUNE 1990 Cur,tctNnMoseurroEs rN WESTERNKnNve 207

RELATIVE ABUNDANCEAND BLOODFEEDING BEHAVIOR OF NOCTURNALLYACTIVE CULICINE MOSQUITOES IN WESTERN

JOHN C. BEIER,I WALTER O. ODAGO,FRED K. ONYANGO, CHARLES M. ASIAGO,DAVY K. KOECH aNo CLIFFORD R. ROBERTS

Kenya Medical ResearchInstitute and U.S. Army Medical ResearchUnit, Nairobi, Kenya

ABSTRACT. At 2 sites in western Kenya targeted for future malaria vaccine trials, adult culicine mosquitoeswere sampledover one year by 5 collection techniquesto assesshuman exposureto potential vectorsofpathogens other than malaria.Collections included 20,910 females representing 19 speciesin Kisian and 4,312females of 11 speciesin Saradidi.Common species in Kisian includedCul.erquinque- fasciatus(71.4%), uniforrnis (15.8%),Ma. afrirana (6.2Vo),Aed.es mcintoshi (2.0Vo),Coquill.et- tidia fuscopennata(7.9%) and Ae. ochraceus(1.8%). Common speciesin Saradidi included Cx. quinque' fasciatus(92.7%), Cx. nebulosus(4.5%) and Ma. uniformis (1.0%).Human-bait collectionsidentified 16 man-biting culicine speciesin Kisian and 9 in Saradidi.Man-biting rates at Kisian for the 5 most common specieswere 1.8, 14.6 and 13.5 times higher than at Saradidi for indoor, outdoor and tent collections,respectively. Exposure indoors was estimatedtobe 7,277bites/man/year at Kisian and720 at Satadidi. Blood meal identification for 1,083mosquitoes confirmed that the common culicine species feed primarily on humans and cows.

INTRODUCTION Kenya. Study sites, collection techniquesand meteorologicalsampling were describedin detail Culicine populations on the Kano for concurrent studies on anophelines (Beier et Plain in western Kenya have been studied ex- al. 1990).Briefly, Kisian is located on the shores tensively to determine the effects of irrigation of Lake Victoria where extensive papyrus on mosquitoproduction and abundance(Surtees swamp provides year-round larval development 1970,Surtees et al. 1970,Khamala 1971,Chan- sites. During the rainy seasons,larval develop- dler and Highton 1975,Chandler et al. 1975b, ment also occurs in temporary ground water 1976a,1976b) and arbovirus epidemiology(Bow- pools and containers. At Saradidi, a site with den et al. 1973,Johnson et al. 1977a,t977b, typical savanna-tlpe vegetation, larval devel- 1981).The vectorpotential of mosquitoesin this opment occurs in temporary and permanent area was evaluated further by blood meal iden- ground water pools along streams, pit Iatrines (Boreman tification et al. 1975,Chandler et al. and containers. 1975a). Weekly all-night human-bait collections at This study comparedthe speciescomposition, eachsite were conducted with pairs ofcollectors, relative abundanceand blood feeding patterns working in 0.5-hperiods from sunsetto sunrise, of culicine mosquitoesat 2 sites,west of Kisumu, inside 6 houses,2 outdoor stations and in one in western Kenya. Proposed malaria vaccine army generalpurpose medium tent (Beier et al. trials involving nonindigenousvolunteers at 1990). Resting mosquitoes inside houses were these 2 sites necessitatethe evaluation of the collectedby aspiration from over 50 housesper potentialfor transmissionof diseasesother than week at each site; additionally, pyrethrum spray malaria.This study providesbaseline informa- catches(World Health Organization1975) were populations tion on culicine mosquito and eval- done in 12 housesper week in Kisian. Six CDC uates the potential for human feeding among light traps at each site were operated outdoors the common speciescollected by 5 sampling each week on the night of human-bait collec- methods. tions. Twenty outdoor aspiration collections (0700-0930h), using a mechanicalaspirator, were done in Kisian from Septemberto Decem- MATERIALS AND METHODS ber 1987.Mosquitoes attracted to a tethered cow (2 Adult culicine mosquitoeswere sampled over in a an outdoor bednet ftap x 1 x 2 mm; a year, beginningin February 1987,in the vil- elevated15-20 cm) were aspiratedat 2-h inter- lagesof Kisian (10 km west of Kisumu) and vals during weekly all-night collectionsin Ki- Saradidi (55 km west of Kisumu), western sian, and from Septemberto December 1987 in Saradidi. Culicine mosquitoesfrom eachcollection were -20"C I Present address: Department of Immunology and killed, placed in vials, frozen at and Infectious Diseases, The Johns Hopkins University, transported to Nairobi for identification, using School of Hygiene and Public Health, 615 N. Wolfe the taxonomickeys ofEdwards (1941)and other Street, Baltimore, MD 21205. descriptions (van Someren1949, Huang 1985). 208 JouRNel oF THEAveRIcllt Mosqulto Contnol AssocInTIon VoL.6,No.2

The number of mosquitoes for each species vegetation yielded 7 speciesat Kisian, but this was relatedto samplingeffort for the 5 collecting method was not used at Saradidi. techniques.For human-bait collections,abun- Average man-biting rates for 5 species,rep- dancewas expressedas the number of females/ resenting over 957oof the total mosquitoesob- man/night. Man-biting ratesinside houses, out- tainedin human-baitcollections, were compared doors and in tents were transformedto logro(n for indoor, outdoor and tent collections (Table + 1) to normalizevariance, and were examined 2). Inside houses,highest biting rates were ob- by analysisof variance(ANOVA) and the Fisher servedfor Cx.quinquefosciatus.In Kisian, biting PLSD test to evaluatedifferences in biting rates rates for Cx. quinquefasciotusinside houseswere among types of sampling sites. Overall mean significantly higher than outdoors or in tents. man-biting rates are presented as Williams' Biting rates for Ae. mcintoshi, Cq.fuscopennata, mean values (M-) (Williams 1937),calculated Ma. africana and Ma. uniformis were highest as back-transformedvalues of the monthly outdoors.In Saradidi, differencesin man-biting mean, rates among collection techniqueswere detected Blood-fed mosouitoesfrom each collection only for Cx. quinquefasciatus.Overall man-bit- were later air-dried at room temperature for up ing ratesin Saradidifor these5 specieswere 1.8, to 5 days, and placed individually in labeled 14.6 and 13.5 times lower than in Kisian for vials. Each mosquitowas ground in 50-pl PBS, indoor, outdoor and tent collections, respec- with 950-pl PBS added afber grinding; these tively. were stored frozenat -70'C. Blood mealswere Monthly average man-biting rates were ex- identified by a direct enzyme-linkedimmuno- amined in relation to rainfall for Cx. quinque- sorbent assay (ELISA) using anti-host (IgG) fasciatus, Ma. africana and Ma. uniformis at conjugates (Kirkegaard and Perry, Gaithers- Kisian (Fig. 1) and for Cx. quinquefasciatusat burg, MD) against human, cow, donkey, goat, Saradidi(Fig. 2). dog, cat, rat, rabbit and chicken (Beier et al. Blood meals were identified by ELISA for 1988).Blood mealswere screenedfirst for hu- 88.3%of 1,226specimens tested (Table 3). Culex man and cow, then nonreacting samples were quinquefasciatLrs,representing 97.9% (1,054/ tested for the other 7 hosts. 1,077)ofthe specimenstested from indoor rest- ing collections, fed primarily on humans (87.6%); mealswere RESULTS outdoors,most of the blood identified as cow (61.3%)and human (22.7%). Sampling during 12 months yielded 20,910 Blood meals from both Mansonia specieswere adult femalesrepresenting 19 culicine speciesin primarily human and cow. Aedesmcintoshi and Kisian, and 4,312females consisting of 11 spe- Cq. fuscopennata fed mainly on cows. Two cies in Saradidi (Table 1). Common speciesin chicken blood meals were detectedin Cx. nebu- Kisian, representing99.1% of the total collected, losus; avian feeding was also observed for Cq. included Culex quinquefasciatus, M ansonia uni- fuscopennata,Cx. quinquefasciatusand Ma. af- formis, Mo africana,Aedes mcintoshi, Coquillet- ricana. Mixed blood mealsfor Cx. quinquefascia- tidia fuscopennataand Ae. ochraceus.In Sara- tus included25 human and cow, 5 chicken and didi, Cr. quinquefasciatus,Cx.nebulosus and Ma. cat, and outdoors, one goat and rabbit meal. uniformis comprised98.2% of the total. Addi- Mixed human and cow mealswere also detected tionaily, 7,680males of 14 specieswere collected for one Ae. mcintoshl and 3 Ma. uniformis. One in Kisian and 2,033 males of 4 specieswere cat-rabbit mixed blood meal was identified for collectedin Saradidi;males of speciescollected Cq.fuscopennata. at eachsite are noted in Table 1. Ovet 90% of the maleswere Cr. quinquefasciafus.Males were DISCUSSION collectedprimarily in CDC light traps, pyreth- rum spray catchesand by aspiration in vegeta- The speciescomposition of culicine mosqui- tion. toes from the 2 sites west of Kisumu was similar There were16 speciescollected in human-bait to previousreports from the Kano Plain, eastof collectionsin Kisian comparedwith 9 in Sar- Kisumu (Surtees19?0, Chandler et al. 1975b, adidi (Table 1). Fewerspecies were detectedby 1976b).Culicine speciesdiversity and relative the other collectiontechniques. In indoor resting abundanceat Kisian were considerablyhigher collections, Culex quinquefasciatus comprised than at Saradidi.Kisian has extensiveperma- about 98% of 6 speciescollected at Kisian and nent Iarval developmentsites, including papyrus 8 speciesat Saradidi.Cow-baited traps yielded swamp,due to its location on the shore of Lake 14 speciesat Kisian and 4 speciesat Saradidi. Victoria. The savanna habitat of Saradidi ap- Light traps collected 10 speciesat Kisian and 6 pears to support fewer species as a result of species at Saradidi. Aspiration collections in reducedlarval habitat diversity. The 21 species JUNE 1990 Cur,rcrNnMoseurrons rN WsstnRN KrNve 209

tl tl tl tl tl tl tl .il ^ll co ll qil@tl i io, -tl NA htl Etl 6tl 5tl o@ tr- dll i<) '- ll a €llotl F- tl 6ll * 6)119 ooH<61 roN >il.: rod iN hil E 1il I iil o ; -o ll -

N NCO dct @ dc6 NS o >:ll X i a

vil x il\ Eil O !il o @ io roc{@ N ca ll a od E €tl X all

i @c6

6 tl @ r tl o 6 (g tl 6 ul ll

d NiiroromOJi@ N c9c{ ll dd r CDi +r i rv r (!tl NCI dd N 6t atl tl sil .l '-tldtl F o tl i iooiooic)ir ihrooo o\QQ o ll a ddo dc;d dododa + qiddooo -jo N cll o c! vv v v 6) m .dll;: (g c ll tr a r lle ' dll G oll ! .til; i i iiod r@io6)o<. s i i i : NcOi ll dll : AAA-i^i--;.idc;-icid .joood@dd o ril v VVVV VV TVVVV iV = a j xllse w atl N oI oil olt o o. tl . ll o tl I '0trtl u 1 L? 5tl o e ll x.3 -E.F o tl d tl Etl a b Ei EHia*a *tl.o tl o Etl o *tua*€=EgE:ra.Eft€ii ll gd

n oo otl a OO oo .g tl srF+i€ir€ sg{iiE $r €EEE _tlF< tl tl a@ tl ,oQ tl tl tl $€$iE r€ i i{ r€t$trEiE tl sr tl SSS{{S{{{S*s!!i.ssfi E E E 3 tl {{{{{{+ti{sdEdE.i.iASSS€ 210 JounNer, oF THE AueRrceN Moseurro CoNrnol AssocrnTror VoL.6, No. 2

Table 2. Man-biting rates of 5 common mosquito species collected during all-night biting collections inside houses, outdoors and in tents at Kisian and Saradidi sites, western Kenya (February 198? to January 1988).

Abundance (My)'

Kisian Saradidi lndoors Outdoors Tent lndoors Outdoors Tent Mosquito species (258)' (82) (6e) (204) (70) (62) Aedes mcintoshi 0.05a 1.54b 0.24a 0.00a 0.05a 0.00a Co quille t tid ia f usc o pe nnat a 0.06a 0.48b 0.23c 0.04a 0.00a 0.00a Cul.ex quinqu.efasciatus 2.25a 1.14b O.25c 1.86a 0.92a/b 0.33b Mansonia africana 0.31a 2.03b 1.64b 0.03a 0.01a 0.02a Mansonia uniformis 0.83a 9.12b 3.7Lc 0.04a 0.00a 0.10a I Williams' mean number of mosquitoesper man-night. For each speciesin each site, meansfollowed by the sameletter are not significantly different (P > 0.05) (Fisher PLSD). 'Number of man-nights

could be detectedby further adult sampling and larval surveys. This study identified 16 man-biting culicine speciesat Kisian and 9 at Saradididuring 745 man-nightscollection effort over oneyear. Only 5 speciesin Kisian, Ae. mcintoshi, Cq.fuscopen- nata, Cx. quinquefasciatus,Ma. africana and.Ma. uniformis, and one speciesin Saradidi, Cx. quin- c quefasciatus,were collected commonly at night - :l feeding on humans. Resident exposure inside c I houses,based on monthly average man-biting rates, was estimated to be ca. !,277 E bites/man/ t year at Kisian and 720 in Saradidi. There were o significant monthly differences in biting rates -o I .o--o--t------. E which were associatedwith seasonal rainfall, l especially at Saradidi where populations z ;l were more dependent upon temporary larval devel- opmentsites during the rainy seasons. Blood meal identification confirmed that hu- mans and cows were the primary hosts for the commonculicine species. A higherproportion of human blood meals were detected for mosqui- toescollected inside houses compared with those o collectedoutdoors. Chandler et al. (1975a),in a ^ 200 more extensive study on the Kano Plain, found gE similar feeding patterns. This study provides baseline information on o-- E culicine mosquitoesin 2 sites targeted for future malaria vaccinetrials. These sites were initially tO F[,1AI\4JJA NDJ selectedbecause there was no evidencefor filar- Month iasis transmission, and the prevalenceof arbo- viruses was suspectedto be low (Johnson et al. Fig. 1. Monthly average indoor and outdoor man- 1977a,l977bl. In this study, biting rates for Cr. biting rates for (A) Culex quinquefasciatus, (B) Man- quinquefasciatr.c,a vector of filariasis, were sonia africana and (C) Ma. uniformis, in relation to about 15 times lower than in a filariasis endemic total monthly rainfall, at the Kisian site, February urban area on the Kenyan coast (Wijers and 1987 to January 1988. Kiilu 1977).During initial malaria vaccinetrials involving nonindigenousvolunteers, exposure to natural malaria transmission is expectedto be collectedas adults in these 2 sites representonly Iess than one month. Based on average daily the common nocturnally-active speciesbiting biting rates for common culicine species,vol- man and domestic , and resting in and unteers exposed inside houses, outdoors or in around houses.Undoubtedly, additional species tents would be expectedto receive more than 5 Cur,rcrNnMoseultons rN Wnsrnnr,tKnNve

l5l I -l ol EI SI 3 s-.1 FI 2l-l I 6J

250-l I I 7 2oo1 - r50l I EI : tool =t -'t I or FMAN,IJJASONDJ Monlh

Fig. 2. Monthly average indoor and outdoor man-biting rates for Culex quinquefasciatus, in relation to total monthly rainfall, at the Saradidi site, March 1987 to January 1988. t-t" t. "".r rt""a *

% of identified blood meals Total % Goat/ Mosquito species Habitatr tested identified Human Cow Donkey sheep Dog Cat Chicken Mixed Aedesmcintoshi o 11 90.9 90.0 10.0 CoquiLbttidiafusco- I 1 100.0 100.0 wnruna o 20 45.0 55.t) 11.1 11.1 11.1 11.1 Culexquinquefascia- I 1,054 89.2 87.6 r.2 r.4 0.5 1.6 0.8 3.? 3.2 tus o 87 86.2 22.7 61.3 2.7 4.0 8.0 1.3 Culex nebulosus o 3 66.7 100.0 Mansonia afrirarn I 7 100.0 57.r 14.3 14.3 14.3 o 8 87.5 100.0 Mansonia uniformis I 15 86.7 84.6 7.7 7.7 o 20 95.0 5.3 84.2 10.5 1Collection techniquesinside housesincluded aspiration collectionsand pyrethrum spray catches;collections outdoors included CDC light traps and aspiration collections in vegetation. bites per 30-day period from only 5 culicine Kenya) for assistance in mosquito identifica- speciesin Kisian and only one speciesin Sar- tion. Excellent technical assistancewas pro- adidi. vided by the following: C. Oyaro, G. Oyare, S. Odour, A. Masinya, S. Towett, I. Ouma, J. Mak- asa,J. Kamanza and other Walter Reed Project L. Peyton ACKNOWLEDGMENTS Staff. We thank R. E. Harbach, E. and C. J. Peters for helpful commentson the We are grateful to M. Mugambi, former Di- manuscript. Informed consent was obtained rector of the Kenya Medical ResearchInstitute from volunteers participating in all-night mos- (KEMRI), for facilitating this research.We also quito collections. The views of the authors do thank E. L. Peyton (Walter ReedArmy Institute not purport to reflect the position of the U.S. of Research)and Yiau-Min Huang (Smithso- Departmentof the Army, the U.S. Department nian Institution) and T. P. Gargan(USAMRU- of Defenseor the governmentof Kenya. 2r2 JouRlrlr, oF THE AvrRrclN Mosourro CoNtRor, Assocrlrrorl VoL.6,No.2

REFERENCES CITED seum(Nat. Hist.). London. Huang, Y.-M. 1985.A new African speciesof Aedes Beier, J. C., P. V. Perkins, F. K. Onyango,T. P. (Diptera:Culicidae). Mosq. Syst.17:108-120. Gargan,C. N. Oster,R. E. Whitmire, D. K. Koech Johnson, B. K., A. C. Chanas,P. Shockley,E. J. and C. R. Roberts.1990. Characterization of malaria Squires,P. Gardner,C. WaIIace,D. I. H. Simpson, transmissionby Anopheles(Diptera: Culicidae)in E. T. W. Bowen, G. S. Platt, H. Way, J. Parsons western Kenya in preparation for malaria vaccine and W. E. Grainger. 1977a. Arbovirus isolations trials. J. Med. Entomol. 27:(Inpress). from, and serologicalstudies on, wild and domestic Beier,J. C., P. V. Perkins,R. A. Wirtz, J. Koros, D. vertebratesfrom Kano Plain, Kenya. Trans. R. Soc. Diggs,T.P. Gargan and D.K. Koech. 1988.Blood Trop. Med. Hyg. 7l:512-517. meal identificationby direct enzyme-linkedimmu- Johnson,B. K., G. Gichogo,G. Gitau, N. Patel, G. norsorbent assay (ELISA), tested on Anopheles Ademba,R. Kirui, R. B. Highton and D. H. Smith. (Diptera: Culicidae)in Kenya. J. Med. Entomol. 1981.Recovery of o'nyong-nyongvirus hom Anoph- 25:9-16. eLesfunestus in westernKenya. Trans. R. Soc.Trop. Boreman,P. F. L., J. A. Chandlerand R. B. Highton. Med. Hyg. 75:239-247. 1975.Studies on the feedingpatterns of mosquitoes Johnson, B. K., P. Shockley,A. C. Chanas,E. J. of the generaFicalbia, Mimomyia and Uranotaenia Squires,P. Gardner,C. Wallace,D. I. H. Simpson, in the Kisumu area of Kenya. Bull. Entomol. Res. E. T. W. Bowen,G. S. Platt, H. Way, J. A. Chandler, 65:69 74. R. B. Highton and M. N. Hill. 1977b.Arbovirus Bowden,E. T. W., D. I. H. Simpson,G. S. Platt, H. isolations from mosquitoes:Kano Plain, Kenya. Way, W. F. Bright, J. Day, S. Achapaand J. M. D. Trans. R. Soc.Trop. Med. Hyg. 7L:518-521. Roberts.1973. Large scaleirrigation and arbovirus Khamala,C. P. M. 1971.The biting of the Kano epidemiology,Kano Plain, Kenya. II. Preliminary Plains,Kenya: Part II. Larval habitatsof common serologicalsurvey. Trans. R. Soc.Trop. Med. Hyg. mosquitospecies (Dipt., Culicidae).Bull. Entomol. 67:702-709. Res.61:299-307. Chandler,J. A. and R. B. Highton. 1975.The succes- Surtees,G. 1970.Large irrigation and arbovirusepi- sion of mosquitospecies (Diptera, Culicidae) in rice demiology,Kano Plain, Kenya.I. Descriptionof the fields in the Kisumu areas of Kenya, and their areaand preliminary studieson the mosquitoes.J. possiblecontrol. Bull. Entomol.Res. 65:295-302. Med. Entomol. 7:509-517. Chandler,J. A., P. F. L. Boreham,R. B. Highton and Surtees,G., D. I. H. Simpson,E. T. W. Bowen and M. N. Hill. 1975a.A study of the host selection W. E. Grainger. 1970. Ricefield development and patterns of the mosquitoesof the Kisumu area of arbovirus epidemiology,Kano Plain, Kenya. Trans. Kenya. Trans. R. Soc.Trop. Med. Hyg. 69:415-425. R. Soc.Trop. Med. Hyg. 64:511-518. Chandler,J. A., R. B. Highton and M. N. Hill. 1975b. van Someren,E. C. C. 1949.Ethiopian Culicidae- Mosquitoesof the Kano Plain, Kenya. L Resultsof Eretmapodites Theobald: description of four new indoor collections in irrigated and nonirrigated speciesof the chrysogastergroup with notes on the areasusing human bait and light traps. J. Med. five known speciesofthis group.Proc. R. Entomol. Entomol.12:501-510. Soc.Lond. (B),18:119 129. Chandler,J. A., R. B. Highton and P. F. L. Boreham. Wijers, D. J. B. and G. Kiilu. 1977.Bancroftian filar- 1976a.Studies on some ornithophilic mosquitoes iasisin Kenya. III. Entomologicalinvestigations in (Diptera,Culicidae) of the Kano Plain, Kenya.Bull. Mambrui,a smallcoastal town, andJaribuni, a rural Entomol.Res. 66:133-143. areamore inland (CoastProvince). Ann. Trop. Med. Chandler,J. A., R. B. Highton and M. N. Hill. 1976b. Parasitol.7 I:347 -359. Mosquitoesof the Kano Plain, Kenya. II. Results Williams, C. B. 1937.The use of logarithms in the of outdoor collections in irrigated and nonirrigated interpretation of certain entomological problems. areasusing human and bait and light traps. Ann. Appl. Btol. 24:404-414. J. Med. Entomol. 13:202-207. World Health Organization.1975. Manual on practi- Edwards,F. W. 1941.Mosquitoes of the Ethiopian cal entomologyin malaria. Part II. Methods and region.III. Culicineadults and pupae.British Mu- techniques.WHO Offset Pub. 13,Geneva.