Mencn. 1985 J. Ar"r.Mosq. CoNrhol Assoc. 93 Beck, S. D. 1980. Insect photopefiodism, 2nd ed. ural history of RVF. In this study we exatnined AcademicPress, New York. 387 pp. 739 blood-fed mosquitoes trapped during and Gallaway,W. J. and R. A. Brust. 1982.The occur- following a period of particularly heavy rainfall rence of Aedes hend,ersoniCockerell and Aedes (October-December1982), which did not gen- triseriatus(Say) in Manitoba. MosQ. Syst. 14:262- eratean epizooticof RVF. However,during this 264. Holzapfel,C. M. and W. E. Bradshaw.lg8l. Geogra- period the virus was isolated from mosquitoes phy of larval dormdnci ln the tree-hole mosquito, at the trapping sites,and from orle dead calf on Aed,estriseriatus (Say). Can. J. Zool.59:l0l,l-1021. a nearby farm; there were also 4 seroconver. Sims,S. R. 1982.Larval diapauseiir the easterntree- sionsin a group of 80 yearling cattle testedai hole mosquito, Aed.estriseriatui; Latitudinal varia- one of the trapping sites(Davies, unpublished tion in induction and intensity.Ann. Entorhol.Soc. data). The emergenceof mosquito speciesap- Am. 75:195-200. peared to be similar to that occurring during Wood, D. M., P. T. Dang and R. A. Ellis. 1979.The the early stagesof RVF epizootics(Linthicum et insectsand arachnids Part of Canada. 6. The mos- al. 1983,1984a). quitoesof Canada(Diptera: Culicidae).Agric. Can. Publ. 1686.390 pp. Mosquitoeswere trapped with Solid State Zavortink, T. J. 1972. Mosquito studies (Diptera: Army Miniature light traps (John W. Hock, Co., Culicidae)XXUII. The New World speciesfor- Gainesville,FL) at known RVF epizootic sitesin merly placedin Aedrs(Finlaya). Contrib. Am. En- ecologicalzones II (1"14'30'S, 36'50'E; 1700 tomol. Inst. (Ann Arbor) 8(3):l-206. m) and III (l' l2'S, 37"E; 1500m) in the vicin- ity of Nairobi, Kenya. Studiesconcerning the natural history of RVF during epizootics (Davies 1975, Daviesand Highton 1980), the population biology of immature mosquitoesin A BLOOD MEAL ANALYSIS OF dambos(Linthictrm ENGORGED MOSQUITOES FOUND IN et al. 1983, 1984a)and the feeding habits RIFT VALLEY FEVEREPIZOOTICS AREAS of mosquitoesat human and calf (Linthicum TN KENYAI bait et al, 1984b) have been con- ducted at these sites. The domestic animals K. J. LTNTHICUM,H. F. A. KABURIA,'F. G. presentin theseareas were predominantly cattle, DAVIESaNo K. J. LINDQVIST, with a few goats, horses and sheep. The com- U.S.Army MedicalResearch Unit (WRAIR)Kenya, mon wild animals present were: bushbuck Box 401,APO New York 09675,and Veterinary (Tragelaphus scriptru), eland (Taurotragus oryx), ResearchLiboratoty, P.O. Kabete, Kenya Grant's gazelle(Garclla granti) and Thomson's gazelle (Gazella thomsonii), giraffe (Gira,ffa Epizootic Rift Valley fever (RVF) occurs in camzlopardali;),jackal (Canisrnesamzlzs), ostrich Kenya in grasslandsadjacenr to, within or de- (Struthio carnehu),reedbuck (Reduncaredunca) rived from natural forest. xnd in the bushed and zebra (Equus burchelli). Light traps were and wooded grasslandswith predominantly suspended5G-80 cm above the ground and Combretumor Acacin tree cover (Davies 1975). baited with CO" (2 kg dry ice). The traps were These are in ecological zones II-IV described generally. placed adjacent to emergence sites by Pratt et al. (1966), in their classificationof (flooded dambo formations). Specimenswere EastAfrican habitat.Epizootics occur after pe- obtained in 380 rap night collectionsmade riods of prolonged and heavyrainfall (Scottet during the period October 15, 1982 to Febru- al. 1956, Davies 1975, Davies and Highton ary 15, 1983.One to eight rraps were set 4-7 1980, Davies et al., unpublished data). The days a week at each study site. The catcheswere virus is transmlttedby mosquitoesduring RVF taken to the laboratory live, frozen at - 70oCfor epizootics(Daubney and Hudson 1933, McIn- 30 min and then examined. All blood-fed tosh 1972,Davies and Highton 1980).A knowl- specimenswere removed and their abdomens edge of the mosquitoesfound in such areas smearedonto filter paper) labelledand stored during epizootic and interepizooricperiods is in a desiccatorjar. The smeared area of the relevantto a greater understandingof the nat- filter paper waslater cut out and placedin 0.5 ml of a phosphatebuffered saline which con- tinued 0.1% M sodium preservative. t azideas a This wcirk was supported by Research Grant No. Blood meal identificationwas carried out by DAMDIT-83-c-9517 from the U.S. Army Medical an enzyme immunoassaysystem developed by Research and Development Command, Ft. Detrick, Lindqvist,Gathuma and Kaburia (1982)for use MD 21701 and by a Research Grant for Project No. 3792 by the Overseas Development Administration, in EastAfrica, and many conjugateshad been United Kingdom. prepared againstthe domesticand wild rumi- 2 Department of Public Health, Pharmacology and nants found in this region. The antispecies- Toxicology, Faculty of Veterinary Medicine, Univer- antisera were produced in rabbits, goats or sity of Nairobi, P. O. Box 29053 Nairobi, Kenya. sheep by weekly inoculationsof the different 94 J. Ar. Mosq. CoNrRoLAssoc. Vou l, No. I species globulin with Freud's adjuvant, these 92.3%, respectively, positive ro the specific continued for severalmonths. The sera were anti-cattle conjugate (Table l). Comparatively subjected to repeated absorption cycles to few Culex specimens were obtained, but of these achievespecies specificity. This wasconsidered 93.3Vo of the haemoglobin positive Culex antm- completewhen serum samplesfrom other spe- natus (Becker) had taken cattle blood meals. ciesgave reactionssimilar to the buffer controls One specimen of Ae. lineanpmnis had taken a in double immunodiffusion tests.Enzyme con- human and one a rabbit blood meal. A single jugates were prepared from either IgG or im- Ae. dmntus had fed upon a giraffe and one A?. munoglobulins with horse-radish peroxidase. cumminsii upon a dog or jackal and 2 on a rabbit. Testswere carried out directly with speciesspe- The anti-rabbit reactions were weak, suggesting cific conjugates, or indirectly using antisera that they were not due to the closely related produced in rabbits and a conjugated anti- Lepus or Pronolagus species. Pedetes capensis rabbit globulin prepared in a goat. This method (Rodentia, Pedetidae), the jumping hare, was regularly detected dilutions of test blood be- commonly encountered at the sites and may yond 10-6.The test thus givesa high degreeof have been the host causing the weak reactions. sensitivity and specificity, and is particularly Four specimens of Aa. circurnluteohu (Theobald) suitable for this type of investigation. The and one Ae. qtas'iuniaittatus (Theobald) were eluants were first examined for the presence strongly positive with the anti-horse or -zebra of haemoglobin by oxidizing reduced phe- conjugate. Of the 104 specimens which con- nophthalein in the presenceof FIrO, in alka- tained haemoglobin but were not identified by line solution. Those positive for haemoglo- the l3 specific conjugates, 8 reacted strongly bin were then tested against the following en- when tested later with the broad spectrum zyme conjugates: goat anti-cattle, -human, anti-bovidae conjugate. These 8 may have fed -sheep, -rabbit (Oryctogalus),-bushbuck, on a less abundant wild ruminant at these -reedbuck, -Grant's and Thomson's gazelle trapping sites, most likely impala (Aepyceros IgG; sheep anti-goat, -dog (ackal) lgG; and melampus) or kongoni (Alcelaphus buselaphus rabbit anti-ostrich, -chicken, -horse (donkey, cokii). There was a residue of 96 (l3Vo) blood- zebra), -mouse (rat), -giraffe whole serum; and fed mosquitoes that have not been identified. an anti-bovidae conjugate which detects all Over 50% of the mosquitoes containing un- membersof this family. identified haemoglobin were Ae. lineatopennis. The results of the haemoglobin screening Aefus lineatopennis does feed on birds (Chandler and blood meal identification testsare shown in et al. 1976). The anti-chicken conjugate may Table l. Of 800 specimensexamined, 6l did have failed to react with the many diverse bird not give a reaction for haemoglobin. The most species encountered at the site. It remains pos- common mosquito species collected over the sible that some of the unidentified specimens trapping period was Aedeslineatopennis (Lud- were avian feeds. low), which comprised more than half the The most abundant mosquitoes obtained in blood-fed specimensexamined. Of the 389 this study after heavy rains in RVF epizootic specimensshown to contain haemoglobin 333 areas were Aedes spp. Cattle were the blood (85.6Vo)had taken cattle blood meals. Three meal source for most of the Aedes spp. which other common Aedesspecies, Ae. dentatus were collected. They were also the most com- (Theobald),Ae. cummircii(Theobald), and Ae. mon large animal species in the study areas. su.danens'is(Theobald) were 90.7%, 79.2Voand Cattle are considered to be the major amplify- Table l. Results of haemoglobin screening and blood meal identification tests of engorged mosquitoes collected from October 15" 1982 to February 15, 1983 in Rift Valley fever epizootic areas in Kenya. Uniden- Unknown tified Rabbit Horse/ Species Hb+ Bovidae Cattle Human Canine (Oryctalagus)Giraffe Zebra Total Ae. li,neatoPennis 50 + 333 (85.6Vo)* I 389 Ae, circumluteolus 8 I 27 (67.5Vo) 40 Ae. cumminsii 25 9 lr4 (79.2Vo) 9 t44 Ae. sudanmsis J 36 (92.3Vo) 39 Ae. dzntatru I 88 (90.7Va') 97 Ae. quasiunivittatus 2 6 I Cx, antennatus I l4 l5 Cx. spp. 6 6 Total 96 624 739 * Percentage of total number of blood fed specimens. MencH. 1985 J. Av. Mosq. CoNrnol Assoc, 95 ing and diseasehosts for RVF in many parts of Mclntosh, B. M, 1972. Rift Valley fever I. Vector Africa (Davies 1975), although local farming studiesin the field. J. S. Afr. Vet. Assoc.43:391- practices often dictate whether sheep assume 395. G. I. dos Santos and this role in some areas (Davies, unpublished Mclntosh, B. M., P. Jupp, B.J. H. Barnard, 1980.Vector studies on Rift Valley data). These results suggest that Aedes mos- fever virus in South Africa.