129

Kenya

CHAPTER 2.3

KenyaMohamed Ali Bob*, Benjamin Odhiambo**,

Gilbert Kibata** and Jason Ong’aro**

Introduction

Geographical context

Field surveys covering 94 farms wereconducted in 1997-98 following themethodology agreed among projectpartners. The farms represented:Kirinyaga, Thika, Muranga, Kiambu,and Nairobi Districts in CentralProvince; Machakos, Athi River, Meru,Kitui, and Kibwezi Districts in EasternProvince; Kwale, Kilifi, and Taita TavetaDistricts in Coast Province; Kajiadoand Naivasha Districts in Rift ValleyProvince; Kisii, Migori, Homa Bay, andSiaya Districts in Nyanza Province; andVihiga and Busia Districts in WesternProvince (Figure 1). The districts wereselected based on the prevalence of thecultivation of tomato (Lycopersiconesculentum Mill.). Table 1 gives themean annual temperatures andprecipitation for each of the surveyedregions.

In the surveyed regions, tomato isplanted throughout the year with theexceptions noted below. EasternProvince is a region of semi-arid lowermid-altitude land becoming transitionalsemi-humid towards Central Province;here, tomato is not planted in July.

* International Center of Insect Physiology andEcology (ICIPE), Nairobi, Kenya.

** Kenya Agricultural Research Institute (KARI),Nairobi, Kenya.

Towns

Surveyedareas

Eastern

Kenya

NortheasternWestern

Central

Nairobi

Nakuru

Kitale

Kisumu

Figure 1. Tomato-growing areas surveyed forwhitefly incidence in Kenya.

Table 1. Mean annual temperatures andprecipitation for the surveyed regionsof Kenya.

Region/ Mean min.-max. MeanProvince temperature precipitation

(°C) (mm)

Eastern 15-26 728

Rift Valley 14-27 591

Central 13-25 1166

Coast 21-30 872

Western and 16-29 1289Nyanza

Whitefly surveys took place betweenNovember and March of 1997-98 and1998-99. Rift Valley Province consists

130

Whiteflies and Whitefly-borne Viruses in the Tropics

of transitional lower to higher mid-altitude land. In the surveyed areas ofKajiado and Naivasha Districts in theRift Valley, tomato is not planted inJanuary and July. Whitefly surveyswere carried out in September 1998and March 1999. Central Provinceconsists of sub-humid highlands;whitefly surveys were carried out herebetween November and March in1997-98 and in 1998-99. In the CoastProvince, a region of sub-humidcoastal lowlands, tomato is notplanted from November to January.Whitefly surveys were carried out inJuly 1998. In Western and NyanzaProvinces (sub-humid to humid uppermid-altitude land and lower highlands)tomato is not planted in January.Whitefly surveys in this region werecarried out in February and November1998.

El Niño rains devastated tomatoproduction in Kenya from September1997 to June 1998. Whiteflies werescarce during this period.

The emergence of Bemisiatabaci as a pest and virusvector

Tomato is one of the most widelycultivated vegetable crops in EastAfrica, predominantly grown by small-scale producers for fresh consumption(GTZ, 1995), while in Kenya inparticular, vegetable-growing both forlocal consumption and for exportprovides an increasingly importantsource of income to farmers. In 1995,Kenyan national scientistsparticipating in a regional workshopranked the whitefly Bemisia tabaci(Gennadius) as the most importantpest/vector in tomato (Varela andPekke, 1995). This whitefly hasbecome an important productionconstraint to crops of the familiesSolanaceae, Leguminosae, Malvaceae,

Cucurbitaceae and Euphorbiaceae,particularly in the semi-arid EasternProvince of Kenya. Extension workersand scientists have reportedinformally the occurrence of severalwhitefly-transmitted viruses: Tomatoyellow leaf curl virus (TYLCV), Cowpeamild mottle virus (CPMMV),Sweetpotato mild mottle virus (SPMMV)and Watermelon chlorotic stunt virus(WmCSV).

Despite the perceived importanceof these horticultural crops and theirvirus disease problems in Kenya,previously no systematic attempt hasbeen made to document theirdistribution and incidence. Virtuallyall work related to whitefliesconducted in Kenya has focused oncassava mosaic disease (CMD), itsprevalence in the country, itstemporal and spatial distribution andassociated yield loss (Seif, 1982; Bock,1987). Some research also had beenconducted on the citrus woollywhitefly Aleurothrixus floccosus(Maskell), which is reported to havebecome a common pest of citrus(Citrus spp. L.) as well as coffee(Coffea arabica L.) and guava (Psidiumguajava L.) in Kenya (Löhr, 1997). Inaddition, the whitefly parasitoids,Eretmocerus mundus Mercet. andEncarsia sublutea (Silvestri), havebeen recorded in Kenya (Dr. AbdurabiSeif, personal communication, 1998).Kenya’s participation in the TropicalWhitefly Integrated Pest Management(TWF-IPM) Project provided anopportunity to gain an overview of thegeographical range and economicimportance of whiteflies, whitefly-transmitted viruses and whiteflynatural enemies in the country—information that would serve as asound basis for planning furtherresearch and pest managementefforts.

131

Kenya

Increased BiologicalUnderstanding

Characterization ofbegomoviruses and whiteflybiotypes

During the surveys in Kenya,325 collections of whitefly adults andnymphs were made. From these,549 specimens of whitefly nymphswere mounted on 310 slides andidentified to species level. Most of thewhiteflies collected were B. tabaci(57%), while Trialeurodes vaporariorum(Westwood) represented 34%. B. afer(Priesner and Hosney) wereoccasionally found (7%) in fields ofcassava (Manihot esculenta Crantz)adjacent to tomato fields. A number ofother whitefly species (2%) also wereencountered, including Trialeurodesricini (Misra), A. floccosus, Siphoninusphillyreae (Haliday) and Bemisia spp.Table 2 shows the composition ofwhitefly species based on the identifiedspecimens from each Province.

Disease incidence andsymptom severity

Whiteflies in the surveyed farms werefound breeding on host plants thatbelong to the following families:Acanthaceae, Amaranthaceae,Asteraceae, Commelinaceae,Convolvulaceae, Cucurbitaceae,Euphorbiaceae, Lamiaceae,Leguminosae, Malvaceae, Rutaceae,

Solanaceae and Verbenaceae (Table 3).Surprisingly, whiteflies were foundbreeding only occasionally on tomato inKenya. In Muranga, Kiambu and MweaDistricts in Central Province, it wasdifficult to find a single whitefly nymphin a whole tomato field. In these areas,whiteflies, mainly T. vaporariorum, werefound breeding on crops such ascommon bean (Phaseolus vulgaris L.)and cassava and on non-cultivatedplants, most significantly Euphorbiaheterophylla L., in all areas surveyed.However, in Kihara (Nairobi) andKigumo (Muranga) in Central Provinceand in Nguruman in Rift ValleyProvince, whiteflies, mainlyT. vaporariorum, were found breedingreadily on tomato.

Host plant feeding preference andstatus as reproductive hosts(oviposition and offspring survival) ofB. tabaci was studied in field cages atthe International Center of InsectPhysiology and Ecology (ICIPE),Nairobi, in 1997 using adult insectscollected from a tomato field nearby.Among the four host plants studied,common bean was the most preferredhost plant, followed by lablab (Lablabpurpureus [L.] Sweet), cowpea (Vignaunguiculata [L.] Walp.) and okra(Abelmoschus esculentus [L.] Moench).There were significant differencesbetween plants with regard to theparameters tested. These resultsconcur with observations in the fieldduring the survey.

Table 2. Species composition (%) of whitefly samples from surveyed regions of Kenya.

Province Species compositiona

B.t. T.v. B.a. T.r. A.f. S.p.

Eastern 61 29 6 0 0 4

Rift Valley 20 80 0 0 0 0

Central 35 60 4 1 0 0

Coast 79 0 21 0 0 0

Western 90 0 7 3 0 0

Nyanza 79 6 5 0 10 0

a. B.t., Bemisia tabaci; T.v., Trialeurodes vaporariorum; B.a., Bemisia afer; T.r., Trialeurodes ricini;A.f., Aleurothrixus floccosus; S.p., Siphoninus phillyreae.

132

Whiteflies and Whitefly-borne Viruses in the Tropics

Table

3.

Rep

rodu

ctiv

e h

ost

pla

nts

of w

hit

efly

spec

ies

enco

un

tere

d d

uri

ng

surv

eys

in K

enya

.

Wh

itef

ly s

pec

ies

Cro

ps

Non

-cu

ltiv

ated

hos

ts

Fam

ily

Spec

ies

Fam

ily

Spec

ies

Bem

isia

ta

ba

ci (G

enn

adiu

s)C

onvo

lvu

lace

aeIp

omoe

a b

ata

tas

(L.) L

am.

Aca

nth

acea

eA

syst

asi

a s

chim

per

i T.

An

der

s.C

ucu

rbit

acea

eC

itru

llus

lan

atu

s (T

hu

nb.) M

atsu

m &

Nak

aiA

mar

anth

acea

eA

chyra

nth

es a

sper

a L

.C

ucu

rbit

a p

epo

L.

Ach

yra

nth

es s

icu

la L

.M

omor

dic

a c

ha

ran

tia

L.

Ast

erac

eae

Age

ratu

m c

onyzo

ides

L.

Eu

ph

orbia

ceae

Ma

nih

ot e

scu

len

ta C

ran

tzA

spili

a m

ossa

mbic

ensi

s (O

liv.

) W

ild

Leg

um

inos

aeC

rota

lari

a s

p.

Bid

ens

pilo

sa L

.La

bla

b p

urp

ure

us

(L.) S

wee

tE

mili

a d

isci

folia

(O

liv.

) C

. Jef

frey

Vig

na

un

guic

ula

ta (L.) W

alp

.G

alin

soga

pa

rvif

lora

Cav.

Cro

tala

ria s

p.

Gu

ten

ber

gia

cor

dif

olia

Ben

th. ex

Oliv.

Tit

hon

ia d

iver

sifo

lia (H

emsl

.) A

. G

ray

Ph

ase

olu

s vu

lga

ris

L.

Com

mel

inac

eae

Com

mel

ina

ben

gha

len

sis

L.

Lam

iace

ae

Oci

mu

m b

asi

licu

m L

.C

onvo

lvu

lace

aIp

omoe

a tri

colo

r C

av.

Mal

vace

aeA

bel

mos

chu

s es

cule

ntu

s (L

.) M

oen

chJa

cqu

emon

tia

tam

nif

olia

(L). G

rise

b.

Gos

sypiu

m h

irsu

tum

L.

Eu

ph

orbia

ceae

Eu

ph

orbia

het

erop

hylla

L.

Sol

anac

eae

Ca

psi

cum

an

nu

um

L.

Leg

um

inos

aeR

hyn

chos

ia h

irta

(A

ndre

ws)

Mei

kle

& V

erdc.

Lyco

per

sico

n e

scu

len

tum

Mill.

Lam

iace

ae

Leo

not

is n

epet

ifol

ia (L.) R

. B

r.S

ola

nu

m m

elon

gen

a L

.O

cim

um

kili

ma

nd

sha

ricu

m G

uer

ke

Mal

vace

aeS

ida

acu

ta B

urm

f.

Sol

anac

eae

Da

tura

str

am

oniu

m L

.N

ica

nd

ra p

hysa

lod

es (L.) G

aer

tn.

Ver

ben

acea

eLa

nta

na

ca

ma

ra L

.A

ster

acea

eTit

hon

ia d

iver

sifo

lia (H

emsl

.) A

. G

ray

Tri

ale

uro

des

va

por

ari

oru

m (W

estw

ood)

Cu

curb

itac

eae

Cu

curb

ita s

pp.

Eu

ph

orbia

ceae

Eu

ph

orbia

het

erop

hylla

L.

Mom

ord

ica

ch

ara

nti

a L

.Leg

um

inos

aePh

ase

olu

s vu

lga

ris

L.

Mal

vace

aeA

bel

mos

chu

s es

cule

ntu

s (L

.) M

oen

chS

olan

acea

eLyco

per

sico

n e

scu

len

tum

Mill.

Sol

an

um

mel

onge

la L

.

Bem

isia

afe

r (P

ries

ner

an

d H

osn

ey)

Con

volv

ula

ceae

Ipom

oea

ba

tata

s L. Lam

.Leg

um

inos

aePh

ase

olu

s vu

lga

ris

L.

Mal

vace

aeG

ossy

piu

m h

irsu

tum

L.

Ast

erac

eae

Tit

hon

ia d

iver

sifo

lia (H

emsl

.) A

. G

ray

Tri

ale

uro

des

ric

ini (

Mis

ra)

Sol

anac

eae

Lyco

per

sico

n e

scu

len

tum

Mill.

Lam

iace

ae

Leo

not

is n

epet

ifol

ia (L.) R

. B

r.

Ale

uro

thri

xus

floc

cosu

s (M

ask

ell)

Ru

tace

ae

Cit

rus

spp.

Sip

hon

inu

s ph

illyre

ae

(Haliday)

Un

iden

tified

tre

e

133

Kenya

Samples of 10 plants per field, onaverage, collected from all surveyedareas, were squashed on nylonmembranes and sent to the John InnesCentre, UK, for hybridization.Hybridization signals varied fromstrong to very weak, probably becauseof differences in the degree of nucleicacid homology. This could suggestdifferent strains of begomoviruses ordifferent begomoviruses altogether. Thesignal strength also may have beeninfluenced by the sampling method,because the virus titre may have beenlow in the particular tissues sampled orat that stage of infection.

At Kibwezi, begomovirus symptomswere observed on sweet pepper(Capsicum annuum L.) and a lowincidence of WmCSV (about 5%) wasrecorded on watermelon (Citrulluslanatus [Thunb.] Matsum. & Nakai). Avariety of mosaic symptoms that needfurther investigation also were noticedin pepper and watermelon. There weresymptoms that suggest that tomato inthis region should be tested forpotyviruses such as Potato virus Y(PVY). Aphid infestation levels werevery high at Kibwezi and some of thesymptoms observed could be a result ofaphid-borne viruses. High incidenceand severe symptoms of CMD wereobserved in Mombasa and KilifiDistricts (Coast Province) and in KituiDistrict (Eastern Province). Thesymptoms appeared more severe thanthose observed in western Kenya.Whether these symptoms are causedby infections of East African cassavamosaic virus (EACMV) or are associatedwith the occurrence of the Ugandanvariant needs to be ascertained.

Leaf curl symptoms were foundduring the survey on non-cultivatedhost plants. These were Achyranthesaspera L., E. heterophylla and Nicandraphysalodes (L.) Gaertn. These plantsare therefore likely to act as reservoirs

of TYLCV. Severe begomovirussymptoms also were observed on aweed identified as Gutenbergiacordifolia Benth. ex Oliv. E.heterophylla was found to be a majorreproductive host of both B. tabaci andT. vaporariorum at any site wherewhitefly infestation occurs and thisplant species is present. Therefore, therole of E. heterophylla in survival of thedisease and the vector outside thecrop-growing season needs to beinvestigated.

Natural enemy species

Forty whitefly parasitoid samples wereobtained from the surveys andpreliminary identification was carriedout. Encarsia sophia (Girault and Dodd)was found to be the predominantspecies, while Eretmocerus sp. wasencountered in the Central Province.Encarsia formosa Gahan was onlycollected from Mwea in the CentralProvince. At Nguruman, Rift Valley, thelevel of parasitization appeared higherthan at other survey sites in Kenya.

Some of the 24 predators collectedwere found feeding on the whiteflypupae. Most were coccinnellids, while afew were spiders and mites but nonewere identified to species level.Macrolophus caliginosus Wagner, bugsknown to prey on T. vaporariorum inother parts of the world, were collectedat several locations in Nguruman (RiftValley Province), Kibwezi (Easternprovince) and Mwea (Central Province).

Geographical range of whiteflyand whitefly-transmitted virusinfestation

Areas identified as “hot spots” forTYLCV infection include Kibwezi, Kitui,Machakos, Athi River (EasternProvince) and Naivasha (Rift ValleyProvince). Low incidence of TYLCVsymptoms was observed in all othersurveyed provinces with the exception

134

Whiteflies and Whitefly-borne Viruses in the Tropics

of Western and Nyanza Provinces,where no TYLCV symptoms wereobserved.

B. tabaci was found to be the mostcommon whitefly species in the country(Table 3). T. vaporariorum was commonin the highlands and was encounteredfrequently in the Central, Rift Valleyand Eastern Provinces. T. vaporariorumwas not recorded from the Coastal andWestern Provinces. The abundancepattern and economic importance ofB. tabaci/TYLCV differed significantlybetween provinces. It was found toattain particularly high epidemicincidence (>25% of the tomato withTYLCV symptoms) in Eastern Province(Kibwezi 30%-100%, Kitui 30%-100%,Machakos and Athi River 15%-30%)and in Rift Valley Province (Naivasha15%-40%), where whiteflies and TYLCVwere of great concern to the tomatoproducers. This suggests that theimpact of the B. tabaci/TYLCV complexis generally more severe in the dryareas.

In Coast Province, where theweather is hot and humid throughoutthe year, B. tabaci constituted 79% ofthe collected whitefly specimens.However, only a moderate incidence ofTYLCV symptoms (5%-25%) wasobserved. In Western Province,B. tabaci constituted 90% of thecollected whitefly specimens and inNyanza Province, 79% (weather is alsohot and humid throughout the year)but no TYLCV symptoms were observedon tomato, other vegetable crops, orweeds, and collected samples werenegative when hybridized using a probeto the original Israeli isolate of TYLCV.In Western and Nyanza Provinces, mostproducers plant tomato as a secondarycrop after the main staple cereal crops,maize (Zea mays L.) and sorghum(Sorghum bicolor [L.] Moench). TheB. tabaci population at Kibwezi(Eastern Province) is usually very high

from December to early March (hot anddry season) but the whiteflies virtuallydisappear from April to August (longrainy season followed by cold season),despite the presence of the same cropsin the irrigated fields throughout theyear.

High T. vaporariorum populationswere observed in Nguruman (Rift ValleyProvince) where the underside of leavesof tomato, common bean and eggplant(Solanum melongena L.) could be foundcovered with adults and nymphs. Inthis area, the infestation level onlydrops during a brief rainy season,around April-May. A similar level ofinfestation by the same whiteflyspecies was observed in the relativelycooler area of Kihara (CentralProvince). Direct feeding damage byT. vaporariorum was observed in bothCentral Province (Kihara and Mwea)and Rift Valley Province (Nguruman),where it was the predominant whiteflyspecies.

Since completion of the survey,very high whitefly infestations havebeen reported in the cut-flower andornamental plants on export-orientedflower and vegetable farms aroundLake Naivasha (Rift Valley Province).The highest TYLCV incidences wereobserved at altitudes ranging fromabout 500 to 1000 m altitude (Table 4).B. tabaci was the dominant whiteflyspecies up to an altitude of about1700 m, above which T. vaporariorumbecame dominant.

Increased Socio-economicUnderstanding

Farmers’ assessment ofwhitefly-related problems

Smallholders dominate vegetableproduction in Kenya and account for70% of the total production (Mulandi,1998). Producers in the survey areas

135

Kenya

were surveyed using agreed protocols.Most of the interviewed producers(75%) used their own land, while 13%used rented lands and 12% had otherarrangements. The great majority ofproducers (80%) were men.

Tomato is one of the mostimportant vegetable crops in thecountry. Virtually all interviewedproducers (98%) regarded tomato astheir most profitable vegetable crop and58% of the producers interviewed havecultivated tomato for more than5 years. Main varieties of tomato grownin the country were Moneymaker andCal-J, while other varieties includeRoma, Marglobe, Petomech, EarlyBeauty, Mecheast, Bonny Best, Heinz,Floradade, Marmande, Ponderosa andHotset. However, no tomato varietygrown in Kenya was known to beresistant or tolerant to TYLC. Mostproducers (75%) bought their tomatoseed from the local market, while 20%used their own seed and 5% reportedbuying directly from the seedcompanies. Other vegetable cropsinclude kale (Brassica oleracea L.),common bean, sweet pepper, cabbage(Brassica oleracea var. capitata L.),eggplant, sweetpotato (Ipomoea batatas[L.] Lam.), okra, bitter gourd(Momordica charantia L.), melon(Cucumis melo L.) and onion (Allium

cepa L.). Most producers (94%)practised crop rotation.

Fifty-five percent of the interviewedproducers listed whiteflies and/orassociated virus diseases among theirprincipal pests and diseases andranked these problems, on average, astheir third-most-important problem intomato. For tomato producers inKenya, late blight Phytophthorainfestans (Mont.) de Bary is the maincause of concern as a yield-limitingfactor. Other pest and diseaseproblems were early blight (Alternariasolani [Ell. and Mart.] Jones andGrout), red spider mites (Tetranychusurticae Koch.), bollworms (Helicoverpaarmigera [Hubner]), leaf miners(Lyriomyza spp.), bacterial wilt (Erwiniatracheiphila [Smith]), aphids, cutworms(Spodoptera litura [Fabricius]),nematodes (Rhabditida:Steinernematidae & Heterorhabditae),end rot (Godronia cassandrae f.vaccinii), thrips (Thrips palmi Karny)and rats (Rattus sp.).

Most producers (88%) were able torecognize whiteflies, while only 64% ofproducers were able to recognize thesymptoms of TYLCV and only 1% knewthat whiteflies and TYLCV were inter-related. Although only about half of theproducers (55%) listed whiteflies and

Table 4. Maximum incidence (% of sampled plants showing disease symptoms) of Tomato yellow leafcurl virus (TYLCV) encountered and species composition (%) of whitefly samples collected atdifferent altitudes in Kenya.

Altitudea (m) TYLCV Whitefly species compositionb

incidence B.t. T.v. T.r. B.a. S.p. A.f. B. spp.

20-270 8 72 6 0 14 8 0 0

520-770 100 81 6 0 13 0 0 0

770-1020 80 33 65 2 0 0 0 0

1020-1270 30 66 25 1 8 0 0 0

1270-1520 15 81 16 0 1 0 0 2

1520-1770 27 77 9 2 5 0 7 0

1770-2020 39 43 53 0 4 0 0 0

a. No survey was conducted in the 270-520 altitude range.b. B.t., Bemisia tabaci; T.v., Trialeurodes vaporariorum; T.r., Trialeurodes ricini; B.a., Bemisia afer;

S.p., Siphoninus phillyreae; A.f., Aleurothrixus floccosus; B. spp., Bemisia spp.

136

Whiteflies and Whitefly-borne Viruses in the Tropics

associated viruses among theirprincipal pests and diseases, 79% ofinterviewed producers believed thatwhiteflies cause problems in the cropson their farms. Among theseproducers, 50% attributed theproblems to both whiteflies and TYLCV,47% attributed the problem towhiteflies only, while only 3%mentioned TYLCV alone.

Areas where producers were notable to recognize whiteflies were mainlyin Western Province (Funyula, Mokonjaand Buhuma), Nyanza Province (SouthUgenya and Nyabondo) and CoastProvince (Shimba Hills and Kikambala),where whiteflies were scarce. Mostproducers in these localities also couldnot recognize TYLCV symptoms andwere not aware of its association withthe vector. Some thought that TYLCVwas either a soil-borne or seed-borneproblem, while some confused it withblight.

Most producers who recognizedwhiteflies and TYLCV symptoms hadnames for both but in most cases theproducers gave non-specific names forwhiteflies (e.g., “insects”) and forTYLCV (e.g., “disease”). Most of thenames given to the whiteflies literallymean “flies” or “small insects” and localnames include ebichuni (insect),keguneta, kimbulutwa (butterfly), kudni(insect), mbuu (mosquito), ngaturia,okogataa (mites), oulolo (white moulds),rwagi rweru (white mosquito), twihurutatweru (small white butterflies), ume(aphids), umuu (moulds), vipuvutevidogo and obororo (plant fleas). Namesgiven to the disease problem mean “leafcurling”, “folding” or “stunted plantgrowth” and local names include chanamatawi (combing of leaves), gathuri(stunted old man), gikware, gukunjamathangu (folding of leaves), kanyaria(the devastating disease), kugogonyara(curly) and mbaa (blight).

Estimation of diseaseincidence and yield losses

Among the surveyed farms, 55% hadTYLCV symptoms in their tomato crop.However, very low incidences (1%infection) were found in most parts ofCentral, Eastern and CoastalProvinces. Incidence of TYLCVsymptoms of more than 30% wasrecorded in 11 out of 94 farms. Thesewere in Kibwezi (3), Kitui (3),Machakos/Athi River (1) and Naivasha(4). High incidence (80%-100%) wasonly recorded in two farms in EasternProvince, in Kibwezi and Kitui areas.The incidence in Kibwezi resulted in atotal loss of the tomato crop. Severeleaf curling plus subsequent yielddepression of nearly 100% in sweetpepper was also observed in oneexperimental plot in Kibwezi. Despitethe presence of B. tabaci in westernKenya (Western and Nyanza Provinces),no symptoms of TYLCV infection wererecorded in this part of the country.

On average, producers’ estimate ofthe yield loss in tomato production was30%. According to the survey, 1%reported a total yield loss, 5% reportedthree-quarter yield loss, 26% reportedhalf yield loss and 50% of theproducers reported one-quarter yieldloss due to the whitefly/TYLCVcomplex. The problem was moreserious in drier areas where irrigationwas practised and this could bebecause of the presence of TYLCVreservoirs and the virus vector.Eighteen percent of the producersinterviewed, mainly from WesternProvince, thought that their tomatocrops did not suffer any yield lossbecause of the whitefly/TYLCVcomplex. However, 8% of producersinterviewed reported abandoning theirtomato crops in at least 1 year becauseof the problem. These were fromEastern, Rift Valley, Central and CoastProvinces. Half of them reported thatthe loss occurred in 1997.

137

Kenya

Seventy-nine percent of tomatoproducers believe that they encounterthe whitefly/TYLCV complex everyyear. Almost one-third of the producers(30%) believed that the incidence andseverity of whitefly-transmitted viruses,especially in tomato, was exceptionallyhigh in early 1997 before the El Niñophenomenon. Most (89%) also believethat there is a direct relationshipbetween the seasonal weather changesand the whitefly/disease incidence.Seventy-five percent of producersreported that the period fromDecember to March (hot and dryseason) was the time when theyexperience the most serious whitefly/TYLCV problem, while few believed thatthey have the problem all year round.Producers gave estimation of the costsinvolved (see below).

Costs estimated by producersin control of whitefly/TYLCV

complex per hectare of tomato

US$ % producers0-49 1350-99 17

100-199 30200-299 17300-400 23

Pesticide use

Almost half of the producers (43%)received advice and recommendationson the use of insecticides fromtechnical advisors, 37% from otherproducers, neighbours or familymembers, 10% from pesticidessalesmen, while 10% claimed to usechemicals on their own initiative. Mostproducers (85%) made their owndecision on what, when and whichinsecticides to apply, while only 10%relied on the technical advice theyreceived. Half of the producers (48%)applied insecticides as a preventivemeasure, 25% applied insecticideswhen they observed whitefly/TYLCVdamage, while 18% applied insecticides

routinely according to the calendar.The high level of insecticide usagecoupled with prophylactic chemicalspraying could easily lead to thewhiteflies developing resistance, whichwould hamper their control.

Because of the lack of virus-resistant tomato cultivars, producersresort to the use of pesticides tocombat the whitefly/TYLCV complex.The use of pesticides in mostproduction areas in Kenya hasincreased, even though it is hazardousand sometimes ineffective anduneconomical. Eighty-six percent oftomato producers used insecticides,mostly pyrethroids andorganophosphates, to combat thewhitefly problem on their farms.However, 9% of the producers did notpractice any sort of control against theproblem, while 4% practised croprotation as a means of combating it.Some producers used a Bacillusthuringiensis formulation (Dipel) forcontrol of whiteflies. One producer inMwea (Central Province) reported usingash, in addition to commercialchemicals, to manage the whitefly/TYLCV complex on his farm. Anotherproducer in Kihara (Central Province)used a home-prepared blend ofbotanicals and a detergent to managethe whitefly problem but apparentlywith little success. Chemical controlwas in many cases directed againstblight (i.e., using fungicides). The mostcommonly used insecticides for controlof the whitefly/TYLCV complex in thecountry were pyrethroids (lambda-cyhalothrin and cypermethrin) andorganophosphates (dimethoate anddiazinon). Other commonly usedinsecticides include alphacypermethrin (pyrethroid), amitraz(amidine), malathion, fenitrothion, andomethoate (organophosphates) andendosulfan (organochlorines).Acaricides such as dicofol, which has

138

Whiteflies and Whitefly-borne Viruses in the Tropics

little effect on insects (EXTOXNET,1998), were also commonly appliedagainst whiteflies. The use of fungicidessuch as metalaxyl, copper sulphate andMancozeb (ethylenebisdithiocarbamate), in an attempt tocombat TYLCV was very common.Almost half the producers (48%) madebetween one and four insecticideapplications per crop season to controlthe whitefly/TYLCV complex, while only8% applied nine to ten applications.Most producers in Nyanza and WesternProvinces did not apply insecticides orpractice any sort of control measureagainst the whitefly/TYLCV complex.

There was lack of knowledge onproper use of pesticides. Some tomatogrowers overused chemicals, whileothers did not spray until symptoms ofdamage were observed, and subsequentintensive chemical application failed tosave the crop. In many cases, producerscontinued spraying on a crop alreadyseverely affected by the virus, using ahigher dosage in the hope of curing theplants. Producers were ignorant aboutthe safety period between the lastspraying of the crop and date of sale orconsumption. Some producers said theydid not care about the pre-harvestinterval as long as the produce lookedgood for sale. Some of the localagrochemical suppliers bought largevolumes of pesticides and measured outsmall volumes according to the financialcapability of the buyers. The applicationof inappropriate pesticides and doserates was common and led to controlfailure. Chemicals without any labelswere usually bought from local dealers.

For the first time in Kenya,information was generated on whiteflyresistance to insecticides in selectedareas of the country. Assessment of theresistance of B. tabaci and T.vaporariorum to methomyl (acarbamate), cypermethrin andbifenthrin (pyrethroids) using both

glass-vial and leaf-dip bioassays wascarried out at Kibwezi, Kitui (EasternProvince), Nguruman (Rift ValleyProvince) and Mwea and Kihara (CentralProvince). Generally, T. vaporariorumwas found to be more resistant to thethree insecticides than was B. tabaci. Ofthe three insecticides used in the tests,cypermethrin was the most commonlyused in the survey farms and bifenthrinthe least. The resistance levels deducedfrom the analysis showed that both B.tabaci and T. vaporariorum were moreresistant to cypermethrin than to eitherbifenthrin or methomyl. The resultsobtained suggest that repeatedexposure of whiteflies to specificinsecticides could have selected themfor resistance to the chemicals.However, this was not a universalconclusion because there were somedeviations from the general trend. Forinstance, a 6.8-fold resistance tobifenthrin was detected at Kihara wherethe insecticide had reportedly neverbeen used. Whiteflies from Kajiado werethe most resistant to cypermethrin,giving a resistance factor value (RF) of87 in the leaf-dip method, while thosefrom Nairobi were the most resistant tomethomyl.

Strengthened ResearchCapacity

The present work has substantiallyimproved our understanding of thewhitefly problem in the major vegetableproduction areas of Kenya. Quantitativeinformation on whiteflies and naturalenemy species diversity and theoccurrence of the whitefly-transmitteddiseases, mainly TYLCV, as well as thesocio-economic understanding of theproducers’ perception and practicesrelating to management of the whitefly/TYLCV complex, has provided Kenyanresearchers with a solid background forsetting a research agenda for furtherwork, some of which may be carried out

139

Kenya

under Phase 2 of the TWF-IPM Project(see Introduction to this volume). Thefirst, diagnostic, phase of the projectalso contributed to capacity buildingthrough a planning workshop and amid-term meeting for all participatingscientists, training of one M.Sc. thesisstudent in assessment of insecticideresistance and training of one B.Sc.student in host plant preferencestudies.

Conclusions

The following areas in Kenya have beenidentified as hot spots for futureresearch on whiteflies and whitefly-transmitted viruses: Kibwezi, Kitui,Machakos/Athi River and Naivasha.The survey revealed that the prevalentwhitefly species in the country areB. tabaci, T. vaporariorum, B. afer,T. ricini, A. floccosus and S. phillyreae.The fact that 55% of the surveyedfarms had TYLCV symptoms indicatesa potential for whitefly/TYLCVproblems in Kenya. The majorconstraint in producers’ knowledge onthe whitefly/TYLCV complex was acomplete lack of understanding (99% offarmers) of the connection between thedisease and the vector.

To minimize the devastating effectsof whitefly-transmitted viruses intomato, sustainable vector- anddisease-management options need tobe developed, tested and deployed,including the use of resistant and/ortolerant cultivars, as well as culturaland biological control practices. Thedevelopment of such options needs tobe based on, and supported by,appropriate research. In addition, weneed to understand the source ofwhitefly infestations and the reasonsbehind the striking between-fieldvariability in whitefly population levelsthat is often observed among fields inclose proximity to each other. For

example, very heavy whitefly infestationwas observed in two adjacent farms atKihara but not in any of the othernearby farms. The contrast could havebeen due to the insects’ feedingpreference, because the two farms hadeggplant and squash (both preferredhosts) that were not on the other farmsbut this possibility would need to beinvestigated further. We need tounderstand why there is low or noincidence of TLCV in western Kenya,despite the prevalence of the vector.The potential of local isolates andcommercial formulations ofentomopathogenic fungi, as well asparasitoids and predators, for controlof whiteflies could be evaluated. Thereis also a need to look at other whitefly-transmitted viral diseases and at therole of weeds as reservoirs ofbegomoviruses.

There is a need for training ofentomologists and virologists in areasof detection, identification andcharacterization of whitefly-transmittedviruses, whiteflies and their naturalenemies, as well as training inepidemiological researchmethodologies. The continuedinvolvement of scientists from ICIPEand the Asian Vegetable Research andDevelopment Center (AVRDC), whohave provided technical support duringthe first phase of this project, would behelpful. Collaborative work involvingthe network of regional stations of theKenya Agricultural Research Institute(KARI) should be extended.

References

Bock, K. R. 1987. Some aspects of theepidemiology of African cassavamosaic virus in coastal districts ofKenya. In: Procs. InternationalSeminar on African cassava mosaicdisease and its control, 3-4 May1987, Yomoussokro, CI.p. 125-130.

140

Whiteflies and Whitefly-borne Viruses in the Tropics

EXTOXNET (Extension ToxicologyNetwork). 1998. Pesticideinformation profile, 3 Nov 1998.Available in: http://pmep.cce.cornell.edu/profiles/extoxnet/carbaryl-dicrotophos/dicofol-ext.html. (Accessed 2 May 2000)

GTZ (Deutsche Gesellschaft fürTechnische Zusammenarbeit).1995. Major pests of tomatoes ineastern and southern Africa:Compilation of past research workand identification of IPMopportunities. GTZ-Integrated PestManagement (IPM) HorticultureProject Report, Nairobi, KE. 32 p.

Löhr, B. 1997. The citrus wooly whitefly,a new pest in eastern and southernAfrica. Agrofor. Today Oct-Dec:21-22.

Mulandi, M. A. S. 1998. The status of thehorticultural industry, challengesand strategies, now and beyond theyear 2000. Horticultural CropsDevelopment Authority, Nairobi,KE. 34 p.

Seif A. A. 1982. Effect of the cassavamosaic virus on yield of cassava.Plant Dis. 66:661-662.

Varela, A. M.; Pekke, A. 1995.Proceedings of Tomato PlanningWorkshop for Eastern and SouthernAfrica Region, 16-20 October 1995,Harare, ZW. Deutsche Gesellschaftfür Technische Zusammenarbeit(GTZ)- Integrated Pest Management(IPM) Horticulture Project, Nairobi,KE. 32 p.