- Email: [email protected]
Highly efficient dry season transmission of malaria in Thailand
22 Tm~s~mom
OP THE ROYAL Socmn DF TROPICALMED~NE AND HYGIENE(1990) 84, 22-28
Highly efficient
dry season transmission
of malaria in Thailand
Ronald Rosenberg, Richard G. Andre and Lek Somchit Departments of Entomology, Walter Reed Atmy Institute of Research, Washington, DC, USA and Armed ForcesResearchInstitute of Medical Sciences,Bangkok, Thailand Abstract Man-biting collections were made for 7 consecutive nights per month for 24 months at 2 sites in a Thai village regularly treated with DDT and fenitrothion yet hyperendemic for Plasmodiumfalcipanrm and P. vivax. Only Anopheles dirus was incriminated as a vector: 1.6% were infective and 2.4% were infected (median numbers of oocysts=3-5). Transmission occurred within the village, which was located in groves of rubber and fruit trees, during the dry months of November to May only, when rates of parity (64%) and biting (2/mannight) were higher than during the monsoon (38% and 0.8%/mannight). Vectorial capacity and inoculation rates surged and then fell during 30 d at the end of the monsoon, quickly reinitiating transmission. Sporozoite species were identified using indirect fluorescent antibody tests or enzyme-linked immunosorbent assays: 76% were P. fulcipatwm, compared to 78% of gametocytes; one mosquito was infected with both species. Vector survival and inoculation rates differed between similar sites 800 m apart. Dry season breeding occurred at the bottom of a deep, concrete-lined well. Much of the natural forest habitat of An. dirus in south-eastern Thailand that was once destroyed for farming is now being replaced with orchards; this ecological change may reintroduce malaria to a wide area. Introduction Malaria in south-east Asia is rural and has always been most stable and dangerous in areascovered with forest or scrub. The notable reduction of malaria risk in some countries during this century may have been as much an effect of extensive deforestation as of malaria control programnres. As recently as 1960, forest blanketed most of the hills of south-eastern Thailand from the bight of Bangkok to the Cambodian border. Endemic malaria in this region was only partly alleviated by indoor application of DDT, which nearly eliminated one vector, Anophelesminimus, but had no effect on another, An. dir-us. Over the next 20 years cultivation replaced nearly all the forest-mostly with cassavafor export-and malaria prevalence greatly declined. By 1980 it became clear that cassava, which had become the subject of European Economic Community trade restrictions, would no longer be so profitable and the Thai government began actively to promote the planting of rubber and fruit trees in its place. Spurred by high market prices for rubber, rambutan, durian, and mangosteen, the area in the south-east devoted to tree crops doubled between 1974 and 1984 (Ministry of Agriculture, Thailand, Correspondence: AFRIMS, APO SanFrancisco,CA 96346, USA.
unpublished report) and is expected to continue increasing at the expense of cassava (Fig. 1). The roles played by An. dirus, An. balabaceuis, and other members of the An. leucosphyrusgroup in transmission of human malaria have been appreciated during the last 40 years only (IMACARTHUR, 1951; COLLESS, 1957; SCANLON & SANDINAND, 1965), a period of unprecedented exploitation of Asian forests. Intrinsically rather rare forest mosquitoes that apparently prefer to feed on arboreal primates, they readily adapt to the edge of man-made clearings and to tree plantations such as tea and rubber. The danger from An. dirw is not only that it is very resistant to control within its habitat but that it is an extraordinarily efficient vector, so long-lived and anthropophilit that only a small population is necessary to maintain high malaria endemicity. For 24 months we studied malaria transmission in a stable, prosperous, tree-farming community in southeastern Thailand. Despite its outwardly healthy appearance, the village of Ban Phluang was highly endemic for both Plasmodiumfalciparum and P. vivax (ROSENBERG et al.,
1990a).
In this report
we shall
describe how, during the same period, malaria was transmitted there. Methods The first year’s researchwas from June 1983to May 1984; the second, from June 1984 to May 1985. Site Details of the location, inhabitants, dwellings, and
Fig. 1. Extent of existing and planned rubber plantations in south-eastern Thailand (after a map prepared by Ministry of Agriculture, Bangkok).
23
economy of the study village, in the south-eastern sector of Thailand, as well as the census methods used, were described in a.previous paper (ROSENBERG et al., 199Oa).Ban Phluang lies in a valley that is extensively planted with rubber (Henea brusiliensis) and several types of fruit trees, particularly rambutan (Nepkdium lappaceum), durian (Durio zibetkinus), and mangosteen (Garcinia mungosrunu);the valley is well drained by a network of perennial streams whose sources are on the surrounding steep hillsides. The relationship of houses to shade and water and the relative positions of the 2 collection sites are shown in Fig. 2. The collection site designated A was a cluster of houses surrounded by mature rubber trees about 25 m tall; site B was surrounded by rambutan trees about 3 m tall and by rice fields. Shade and undergrowth were appreciably less under rambutan than under rubber. Rainfall and continuous temperature and relative humidity were recorded daily. Houses were mostly elevated wooden platforms, about 2 m high, covered by synthetic roofing; ventilation was enhanced by the common practice of constructing no more than 3 complete walls. Interiors and under&oring were sprayed by government malaria control workers each March with DDT at 2 g/m’ and each October with fenitrothion at 1 g/m2
during 1984 and 1985; in 1983 fenitrothion was used for both applications. An. &as from the area have routinely been shown to be susceptible to both insecticides (PATIPONGSE,1986). All farming was mechanized and the only domestic animals in the valley were chickens, dogs, and cats. During the dry season, November-May, about 55% of those older than 12 years reported spending at least one night per week tapping rubber, an activity usually done between 0100 and 0600 h. All but 3 of 48 houses in the study area reported having mosquito nets and 82.8% of residents claimed to have used them during May 1985; those nets examined were generally in good condition. Mosquito collections Man-biting Anopkeles were collected simultaneous-
ly at each of two sites (Fig. 2) for 7 consecutive nights each month during the period of the full moon (ROSENBERG & MAHESWARY,1982). At each site captures were made for 50 min periods beginning on the hour from 1900 to 0450 h bv a two-man team sitting at ground level l-2 m outside an inhabited dwelling; during rain, collections continued under the nearest houses. No adjustments were made for the slight seasonal variation in daylight length. Teams
Fig. 2. Plan of the study site, Ban Phluang, Thailand, showing relationships of houses and mosquito collection sites to vegetation. Vegetation was classified using aerial photographs (19 000) with ground control. In general, rice fields and cleared land have no shade; scrub and immature orchard are partly shaded; plantations and forest are heavily shaded all day. A and B indicate collection sites.
24 were replaced between 2350 and 2410 h; consequently the sixth collection period was only 40 min long. Approximately midway between the regularly scheduled October and November 1983 collections an extraordinary 6 d collection was made using normal procedures; except where explicitly stated, data from this collection were excluded from analyses. Several days each month were devoted to unsystematic surveys of breeding sites within a 5 km radius of the biting collection sites. Whenever possible, larvae were reared to adults and preserved as matched sets with their 4th instar and pupal exuviae for identi6cation. Adult mosquitoexamination Each morning after a collection species were identified, ovaries examined for parity, and the midguts and salivary glands of parous females examined for parasites. During the one month each autumn when populations peaked only An. dirus, An. minimus, and An. muculatus were checked for parity and parasites. During the first year glands with sporozoites were dried on microscope slides and tested with the immunofluorescent antibody (IFA) technique using sera from rabbits immtmized with either P. tivax or P. falciparum sporozoites collected from the salivary glands of laboratory-reared An. dirus fed on gametocyte carriers. Rabbits each received an initial intravenous inoculation of lo6 sporozoites followed by 3 weekly boostersof lo5 before blood was drawn; specificity was checked against 6 sporozoite isolates. In the second year, sporozoites were idenTable 1. AnopMw species caught biting man on 672 man-nights in Ban Pbluang, Thailand
Soecies An. An. An. An. An. An. An. An. An. An. An. An. An.
aconitus annulati dints jam&i
kanvarii kochi
letifer maculatus minimus peditaeniatttf
philippknsisb teselahu umbrosus
No. infected
No. cauaht
oocvsts
Swrozoites
156
0
0
18
0
0
1744
42 0
29 0
821
5
0
1
0
0
2 567
0 4 25 2
0 0 0 0
1 0
0 0
0
0
6
1563 2179 132 194 5
‘Includes An. peditaeniatw, An. barbirostris, An. campestris, An. argvrops, An. sinens&An. nigenimus,An. nit&k, and An. hyrcanur. %&da An. philippinensis and An. niwipes.
tified using an enzyme-linked immunosorbent assay (ELISA) (WIRTZ et al., 1987); all specimens were tested in triplicate. Calculations To facilitate analysing seasonaldifferences in vectorial capacity, several estimates and assumptions were made. Daily survival rates were computed after the method of DAVIDSON (1954) according to the relationship p= Ti,,, where T is the proportion parous and x is the length of the blood f&g interval, here assumed to be 3 d. Vectoral capacity (GARRET-JONES & SHIDRAWI, 1969)was calculated as C=(ma)apWnp, in which mu is the daily man-biting rate, a is the daily rate of blood feeding on man, p is the daily rate of survival, and n is the length of the sporogonic cycle. The rate ma is calculated from the biting collections and p from the proportion parous, as above. The value of a was chosen as O-33, which assumesa blood meal every third day on man; captured An. dirus usually oviposited the second night after blood feeding; how soon after that they refed is unknown. For convenience the length of sporogonic cycle for both parasite species was taken to be 12 d, although P. vivax would usually mature at about 11 d and P. fakiparum at 13 d at 27°C. In this paper inoculation rate (h’) means the number of bites received per person per day by mosquitoes with sporozoite-infected glands; the rate is estimated from dissection results. A mosquito with infected glands is described as ‘infective’, whereasone carrying only oocysts on the gut is termed ‘infected’. Results
In 168 nights of regular biting collection (672 man-nights), 13 speciesor speciesgroups of Atwpkeles were caught (Table 1). Only A. dirus was found to be infective (Table 2): 1.55% of the salivary glands and 2.41% of the midguts were parasitized. An. minimus was found to be infected relatively often during the dry months with oocysts which appeared healthy (1*60%), occasionally differentiated, and may have been a secondary vector. The median numbers of oocysts for both specieswere similarly small, 3-5 for An. dirts and 3.0 for An. minimus; oocyst ranges and frequency distributions are compared in Fig. 3. Daily inoculation rate for the annual 7 month transmission season(Fig. 4) was 0.102 in the first year and O-036for the second; this implies that each village resident received nearly 22 inoculations of sporozoites
Table 2. Seasonal changes in vectorial components for Anoplieles dhs Year 1 2 Transition Season Wet Wet Transition (Jun.-Sep.) (Oct.) (Nov.)” (No?%ay) Months (Jun.-Oct.) (Nov.) (De?%ay) Man-biting rate (tn~)~ 1.22 12.46 24.21 1.93 0.50 11.61 2.33 Parous 044 044 0.51 0.70 0.21 0.59 0.58 Daily survival @) 0.76 0.76 0.80 0.89 0.59 0.84 O-83 Vectoral capacity (C) 0% 0.56 2.42 1.28 0.001 2.60 0.48 Inoculation rate (h’) 0 0 0.29 0.10 0 0.07 0.03 ‘Extraordinary 6 d collection midway between regular October and November collections. bAbbreviations: ma=number of vector biting/man/day; p=parous rate1’3;C=(mu)up”l-lnp, where a--0.33, n= 12; h’=number of infective mosquitoes biting/man/day.
JJASONDJFMAMJJASONDJFMAM
140120-
8
loo-
0
80-
2
IO-
60-
; > B
5-
03
0
0
8
c
n
,ilnfl
rln
nn
-
I-L
4020-
1983
1984
1985
Fig. 4. Monthly relationships at Ban Phhmg, Thailand between (A) rainfall >I00 mdmonth (upper rectangle) and percentages of Amphela dina with (B) sporozoites, (C) oocysts, (D) parous, and (E) number of An. dims biting/man-day. Data for (BHE) collected for 7 nights/month. Parity was not calculated for July-September 1984 (marked with asterisks) because of small sample sizes.
Dirus
Minimus
Fig. 3. Frequency distribution of malaria oocysts for Anopheladims and An. minimusdissected at Ban Phluang, Thailand from June 1983 to May, 1985.
Table
3. Malaria
infection
in Anopbeles dhs: A
No. caught Parity (%) Inoculation rate Percentage of catchb with Oocysts only Oocysts+sporozoites Sporozoites only Oocysts (total) Sporozoites (total)
548 65.7 4.2
comparison
Year 1 B 316 73.9 7.7
in the first seasonand nearly 8 in the second. Of the 20 infective mosquitoes collected in the lirst year, 14 reacted by IFA with P. falciparum only, 3 with P. vivux only, 1 with both, and 2 with neither, apparently becausesporozoites had washed off the slide. In the second year 4 mosquitoes reacted with P. falciparum,
by year and site (A and B) of collection Total 864 69.3 6.0
A 571 54.8 1.2
Year 2 B 307 63.8 3.0
Total 878 58.3 2.1
0.9 (5)
1.0 (3)
0.9 (8)
2.1 (12)
2.3 (7)
2.2 (19)
0.9 (5) 0.4 (2)
:‘;. ii;
1.4 (8) (12) 0.9
0.2 (1) (1) 0.2
0.7 1.0 (2) (3) 2.9 (9) 1.6 (5)
;:; 1;;
3.2 (10) 4.1 (13)
“Calculated for dry months only: November-May bNumhers of infected mosquitoes in parentheses.
2.3 (20) 2.3 (20) in year 1, December-May
2.3 (13) 0.4 (2) in year 2.
26 catch during that hour. During the wet seasonAn. dirus larvae were found in and around the study area
19 20 21 22 23 24 01 02 03 64 d5 Time [ Hourof Day ) Fig. 5. Cumulative, outdoor night biting pattern of 1744Anopheles dim caught during 672 man-nights. The 24-01 h collection period was 40 mill long; all others were 50 min.
2 with P. vivax, and 1 was negative by ELISA. The 2-year proportion of 76.0% P. falciparum sporozoites compares to a prevalence rate of 789% for P. fakiparum gametocytes (ROSENBERG et al., 1990a). Abundance, parity and infectivity of An. dirus occurred in annual cycles that appeared to be linked to rainfall (Fia. 4). Durina the height of the summer rains both ih; size of the biting population (Fig. 4E) and the percentage parous (Fig. 4D) fell to relatively low levels, but as the monsoon ended in October and November the number caught increased explosively and parity began to rise. The population remained large for less than 60 d, but parity continued high through the dry season; virtually all infected and infective An. dirus were caught in the latter half of this brief population surge or during the dry months that followed, when bitine averaaed 2.l/man/ninht (Fie. 4B,C). In Table 2 these sea&al events are-a&y&d by assuming that changesin parity indicated changes in daily survival rates; survivorship was then used to calculate vectorial capacity for each of 3 annual periods; wet, transitional, and dry. Vectorial capacity and inoculation were highest during the latter part of the transitional period, owing in particular to the high man-biting rates. The second year inoculation rates were 65% lower than the first; calculated survival rates (Table 2) did not rise in the post-transition period as they had in the first year. The possibility that a decreasein mosquito longevity was responsible for lowering both parity and infectivitv was sunported by the observation that the number -of gut-ii&ted Ah. dirts also positive for swrozoites fell from 60% the first vear to 14% in the skcond (Table 3); this implies that fewer infected females were living long enough for oocysts to mature. Although 44% fewer An. dirts were caught biting at site B. a resident was twice as likelv to be bitten bv an infective mosquito there as at site A, about Sk m away (Table 3). The same proportion of females at both sites was infected with oocysts alone, but at B parity was consistently about 85% higher than at A, and the proportion carrying sporozoites was 3-2 to 4.0 times greater. This suggests that vector survival was higher at B and that the females caught tended to spend their lives in a small area. Peak densities of An. dirus bit outdoors between 2200 and 0100 h (Fig. 6). There was no discernible deviation from this biting pattern for sub-populations of parous or infected females, or between the 2 collection sites. The replacement at 2400 h of the collection team may have caused the relatively low
in wheel ruts under mature rambutan trees, in abandoned charcoal pits under mature rubber, in a discarded 55 gallon [250 litre] metal drum under forest, and in a pit dug next to a shaded stream. During the dry months larvae were occasionally found at the bottom of a concrete lined well. about 1.5 m diameter and 5 m deep, at collection site B; sampling along stream edges in the village and nearby forest was negative. Discussion Ban Phluang was hyperendemic for malaria because it was a suitable habitat for An. dirus. The contribution of An. minimus to transmission was inconsequential and the other species sometimes found to be infected with a few small oocysts of indeterminate parasites were unlikely to have been vectors. The annual An. dirus infectivity rate of 1.6% compares closely to other long-term rates for this and related species (ISMAIL et al., 1975; KHIN-IMAuNG-KYI & & WINN, 1976; WILKINSONet al., 1978; ROSENBERG MAHESWARY,1982). Since sporozoites were found during only 7 months per year, inoculation rates were nearly twice as intense as the annual rate indicates. The species ratio of sporozoites was remarkably similar to that of gametocytes in humans, implying that both P. falcipanrm and P. vivax were being transmitted equally efficiently, even though there is evidence that glands naturally infected with P. falciparum may contain more sporozoites (BURKOTet al., 1987). The small median number of oocystsfound (3.5) was commensurate with the low gametocyte densities prevalent (MUIRHEAD-THOMSON, 1957; ROSENBERG et al., 1990a) and suggests that most gland infections, and possible most sporozoite (ROSENBERG et al., 199Oc), were small. The most striking characteristic of transmission at Ban Phluang was the brief transitional period at the end of the monsoon. A. dirus breeds in small, fresh, temporary pools and its numbers and proportion infective have usually been observed to peak sometime during the wet season(COLLESS,1952; SCANLON & SANDINAND, 1965; WILKINSON et al., 1978; ROSENBERG & MAHESWARY,1982). At Ban Phluang, however, transmission began as the rains ended, when, in less than 60 d, the man-biting rate rose steeply, survival lengthened, and the inoculation rate climbed to nearly its highest level of the year. The rapid start of transmission caused an equally rapid increase in P. falciparum gametocyte prevalence (ROSENBERG et al., 1990b), a phenomenon that may have compensated for the subsequent decline in vectorial capacity. This pattern contrasts with a gradual increase in seasonal transmission seen in Africa (MOLINEAUX & GRAMICCIA,1980). The large differences in inoculation rates observed between seasons,years, and collection sites probably resulted from differences in vector survival rates, which can have a hugely disproportionate effect on the probability of a mosquito becoming or remaining infective (ONORI & GRAB, 1980). Parity rates can be rough measures of daily adult survival when emergence is stable (GARRETT-JONES & GRAB, 1964), and these moved in concert with changes in observed transmission rates. The ratio of gland-infected to
27
gut-infected An. dims also implies that survival was longer in the first year than in the second, and longer at site B than at A. (It has been suggested by DYE (1986) that changes in vector population size may predict infectivity as accurately asestimatesof longevity; clearly this was not true at Ban Phluang.) The methods we used were not sensitive enough to reveal why survival rates differed; there was no clue in rainfall pattern and the possible influences of microclimate, predators, and pathogens were not explored. It may not be coincidental, for instance, that the reduced transmission rate in the second year was preceded by a very low density of adults during the monsoon. Almost twice as many An. dim-s were caught at site A, where vegetation was denser and less cultivated, than at site B, 800 m away, but a much smaller proportion of those at A lived long enough to become infective. If female An. dims dispersed randomly from points of blood feeding and oviposition, then there should have been no sharp discontinuities in survival rates within the small study area. The rice paddies between sites A and B may have limited random dispersion, but An. dims is capable of flights over uncovered ground of at least half a mile [0*8 km] (COLLESS,1953). We infer that most females spent their adult lives within a short radius of the same hosts. Such behaviour would increase the probability of a given mosquito feeding repeatedly in the same house and could lead to clustering of cases if transmission were low. It also suggests that many females may have deposited ova & places such as moist denressions (WILKINSON et al.. 1978: ROSENBERG, lY82) or streams which were ‘convenient for them but unfavourable for their offspring. The difficulty of finding dry season breeding places supports this hypothesis. Mosquito nets hung in most houses and were usually in good condition; when used with a sleeping mat, as was customary, they should have excluded mosquitoes. Nevertheless, everyone was infected with malaria. The population of An. dims may have been suppressed by heavy rainfall but other man-biting species thrived during the monsoon; if the nets were used mostly to protect against the nuisance of being bitten rather than of getting malaria, then their use was probably greatest during the season when the danger was least. Furthermore, nets were usually made to accommodate several people and so the probability of a sleeper being bitten through the net seemsgreater. Malaria incidence among the hundreds of pilgrims, mostly non-immune, who spent one or more nights at the village monastery each February is unknown. Natural forest is not essential habitat for An. dims, which has commonly been found in hilly, shaded terrain in which the dominant vegetation was scrub, tea, or rubber. Once it was indigenous to large areas of south-eastern Thailand but it disappeared, along with much malaria, when trees were felled in the 1960s to make room for cassavacultivation. More recently, the government has successfully promoted the planting of rubber and eucalyptus (ANONYMOUS, 1987) on this same land in response to depressed cassavaprices. If An. dim recolor&es this substantially restored habitat the.n malaria endemicity may also return.
Acknowledgements
Expert technicalhelp wasgiven by Ruan Thaopha,Sanit Nakngen, Vi&it Phunkitchar, Inkam Inlao, Sompom Chanaimongol,ChunmongNoigamol,PrajimBoonyakanist, SuwattanaVongpradit,andPhillip E. Fleischer.Kol Mongkol entereddatainto a computerand RampaRattanarithikuI con6rmedsomespeciesidentifications. GeorgeWard and Robert Wirtz of the WalterReedArmy Institute of Research aidedin the adaptationof IFA and ELISA; J. Gingrich and JetsumonSattabongkotperformedthe ELISAs. The Director and staff of the Malaria Division, Ministry of Public Health, Thailand provided invaluable cooperation.
References Anonymous (1987). Shell seekssupport for eucalyptus plan. The Nation (Bangkok), 13, (3807), 25. Burkot, T. R., Graves, P. M., Cattani,J. A., Wirtz, R. A. & Gibson, F. D. (1987). The efficiency of sporozoite transmission in the human malarias. Plasmodium falciparum and P. vivax. Bulletin of, the World Health Organization,
65, 375-380.
Colless,D. H. (1952).Observationson the periodicity of natural infections in the anopheline mosquitoes of Borneo. Medical Journal of Malaya, 6, 234-240. Colless. D. H. (1953). Observations on the fliaht range ofA. balabacensis~ Baisas (= A. leucosphyrus, &ct.). Medical Journal of Malaya,
7, 179-184.
Colless, D. H. (1957). The Anopheles leucosphyrus group. Proceedings of the Royal Entomological Society of London,
108. 37-116. Davidson, G. (1954). Estimation of the survival rate of anopheline mosquitoes in nature. Nancre, 174, 792-793. Dye, C. (1986). Vectorial capacity: must we measure all its components? Parasitology Today, 2, 2-8. Garrett-Jones, C. & Grab, B. (1964). The assessment of insecticidal impact on the malaria mosquito’s vectorial capacity, from data on the proportion of parous females. Bulletin
of the Worki Health Grganization,
31, 71-86.
Garrett-Jones, C. & Shidrawi, G. R. (1969). Malaria vectorial capacity of a population of Anopheles gambiae, an exercise in epidemiological entomology. Bulletin of the World Health Grganixation, 40, 531-545. Ismail, I. A. H., Notananda, V. & Schepens, J. (1975). Studies on malaria and responses of Anopheles balabacensis and Anopheles minimus to DDT residualspraying in Thailand. Pan I. Acta Tropica, 31, 129-164. Khin-Maung-Kyi & Winn, S. M. (1976). Studies on malaria and Anopheles balabacensisbalabacensis and Anopheles minimus in Burma. Geneva: World Health Organization,
Mimeographeddocumentno. WHO/MAL/76.875.
MacArthur, J. (1951). The importance of Anopheles leucosphyrus. Transactions of the Royal Society of Tropical Medicine and Hygiene, 44, 683-694. Molineaux, L. & Gramiccia, G. (1980). The Garki Project.
Geneva: World Health Organization. Muirhead-Thomson, R. C. (1957). The malaria infectivity of an African village population to mosquitoes (Anopheles gambiae). American Journal of Tropical Medicine and Hvaiene. 6. 971-979.
Onori, E. & Grab, B. (1980). Indicators for the forcasting of malaria epidemics. Bulletin of the World Health Grganization, 58, 91-98. Patiuonase, S. (1986). Susceotibilitv of Thai malaria vectors to insecticides. In: Se&f Conference on Malaria Research. Bangkok: Ministry of Public Health, p. 170. Rosenberg, R. (1982). Forest malaria in Bangladesh. III. Breeding habits of Anopheles dirus. American 3oumal of Tropical Medicine and Hygiene, 31, 192-201. Rosenberg, R. & Maheswary, N. P. (1982). Forest malaria in Bangladesh, II. Transmission by Anopheles dirus. f8y.r; Journal of Tropical Medicine and Hygiene, 31, Rosenberg, R., Andre, R. G., Ngampatom, S., Hatx, C. & Burge, R. (1990a). A stable, oligosymptomatic malaria focus in Thailand. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84, 14-21.
28 Rosenberg, R., Andre, R. G. & Ketrangsee, S. (1990b). Seasonalfluctuation of Plrrmrodiumfalcipan gametocytaemia. Transactions of the Royal So&g of Tropical Medicine and Hygiene, 84, 29-33. Rosenberg, R., Wirtz, R. A., Schneider, I. & Burge, t(. (l!BOc). An estimation of the number of malana sporoz&tes ejected by a feeding mosquito. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84, in
press. Scanlon, J. E. & Sandinand, U. (1%5). The distribution and biology of Atwpheks balabacensis in Thailand. Journal of Medical Entomology, 2, 61-69.
I
Corrections
Wilkinson, R., Gould, D. J., Boonyakanist, P. & Segd,, H. E. (1978). Observations on Anopheks balubacenm in Thailand. Journal of Medical Entomology, 14, 666-671. Wirtz, R. A., Burkot, T. R., Graves, P. M. & Andre, R. A. (1987). Field evaluation of enzyme-linked immunosorbent assaysfor Plusm&ium fakiparum and Pkwtwdium &ax sporozoites in mosquitoes (Diptera: Culicidae) from Papua New Guinea. Journal of Medical Entomology, 24, 433-437.
Received 15 December 1988; accepted fm publication 30 March
1989
I
7’ransactims, vol. 83, part 4, p. 510. The code number of the isolate of Leishmznia tropica was cited incorrectly by the authors; it should have been MHOMIMAI871LPN 39. Transactions, vol. 83, part 6, pp. 788 At 789. Figures 1 and 4 should be transposed, as should Figures 2 and 3; the lengeds as printed are correct. We apologise for this error.
OP THE ROYAL Socmn DF TROPICALMED~NE AND HYGIENE(1990) 84, 22-28
Highly efficient
dry season transmission
of malaria in Thailand
Ronald Rosenberg, Richard G. Andre and Lek Somchit Departments of Entomology, Walter Reed Atmy Institute of Research, Washington, DC, USA and Armed ForcesResearchInstitute of Medical Sciences,Bangkok, Thailand Abstract Man-biting collections were made for 7 consecutive nights per month for 24 months at 2 sites in a Thai village regularly treated with DDT and fenitrothion yet hyperendemic for Plasmodiumfalcipanrm and P. vivax. Only Anopheles dirus was incriminated as a vector: 1.6% were infective and 2.4% were infected (median numbers of oocysts=3-5). Transmission occurred within the village, which was located in groves of rubber and fruit trees, during the dry months of November to May only, when rates of parity (64%) and biting (2/mannight) were higher than during the monsoon (38% and 0.8%/mannight). Vectorial capacity and inoculation rates surged and then fell during 30 d at the end of the monsoon, quickly reinitiating transmission. Sporozoite species were identified using indirect fluorescent antibody tests or enzyme-linked immunosorbent assays: 76% were P. fulcipatwm, compared to 78% of gametocytes; one mosquito was infected with both species. Vector survival and inoculation rates differed between similar sites 800 m apart. Dry season breeding occurred at the bottom of a deep, concrete-lined well. Much of the natural forest habitat of An. dirus in south-eastern Thailand that was once destroyed for farming is now being replaced with orchards; this ecological change may reintroduce malaria to a wide area. Introduction Malaria in south-east Asia is rural and has always been most stable and dangerous in areascovered with forest or scrub. The notable reduction of malaria risk in some countries during this century may have been as much an effect of extensive deforestation as of malaria control programnres. As recently as 1960, forest blanketed most of the hills of south-eastern Thailand from the bight of Bangkok to the Cambodian border. Endemic malaria in this region was only partly alleviated by indoor application of DDT, which nearly eliminated one vector, Anophelesminimus, but had no effect on another, An. dir-us. Over the next 20 years cultivation replaced nearly all the forest-mostly with cassavafor export-and malaria prevalence greatly declined. By 1980 it became clear that cassava, which had become the subject of European Economic Community trade restrictions, would no longer be so profitable and the Thai government began actively to promote the planting of rubber and fruit trees in its place. Spurred by high market prices for rubber, rambutan, durian, and mangosteen, the area in the south-east devoted to tree crops doubled between 1974 and 1984 (Ministry of Agriculture, Thailand, Correspondence: AFRIMS, APO SanFrancisco,CA 96346, USA.
unpublished report) and is expected to continue increasing at the expense of cassava (Fig. 1). The roles played by An. dirus, An. balabaceuis, and other members of the An. leucosphyrusgroup in transmission of human malaria have been appreciated during the last 40 years only (IMACARTHUR, 1951; COLLESS, 1957; SCANLON & SANDINAND, 1965), a period of unprecedented exploitation of Asian forests. Intrinsically rather rare forest mosquitoes that apparently prefer to feed on arboreal primates, they readily adapt to the edge of man-made clearings and to tree plantations such as tea and rubber. The danger from An. dirw is not only that it is very resistant to control within its habitat but that it is an extraordinarily efficient vector, so long-lived and anthropophilit that only a small population is necessary to maintain high malaria endemicity. For 24 months we studied malaria transmission in a stable, prosperous, tree-farming community in southeastern Thailand. Despite its outwardly healthy appearance, the village of Ban Phluang was highly endemic for both Plasmodiumfalciparum and P. vivax (ROSENBERG et al.,
1990a).
In this report
we shall
describe how, during the same period, malaria was transmitted there. Methods The first year’s researchwas from June 1983to May 1984; the second, from June 1984 to May 1985. Site Details of the location, inhabitants, dwellings, and
Fig. 1. Extent of existing and planned rubber plantations in south-eastern Thailand (after a map prepared by Ministry of Agriculture, Bangkok).
23
economy of the study village, in the south-eastern sector of Thailand, as well as the census methods used, were described in a.previous paper (ROSENBERG et al., 199Oa).Ban Phluang lies in a valley that is extensively planted with rubber (Henea brusiliensis) and several types of fruit trees, particularly rambutan (Nepkdium lappaceum), durian (Durio zibetkinus), and mangosteen (Garcinia mungosrunu);the valley is well drained by a network of perennial streams whose sources are on the surrounding steep hillsides. The relationship of houses to shade and water and the relative positions of the 2 collection sites are shown in Fig. 2. The collection site designated A was a cluster of houses surrounded by mature rubber trees about 25 m tall; site B was surrounded by rambutan trees about 3 m tall and by rice fields. Shade and undergrowth were appreciably less under rambutan than under rubber. Rainfall and continuous temperature and relative humidity were recorded daily. Houses were mostly elevated wooden platforms, about 2 m high, covered by synthetic roofing; ventilation was enhanced by the common practice of constructing no more than 3 complete walls. Interiors and under&oring were sprayed by government malaria control workers each March with DDT at 2 g/m’ and each October with fenitrothion at 1 g/m2
during 1984 and 1985; in 1983 fenitrothion was used for both applications. An. &as from the area have routinely been shown to be susceptible to both insecticides (PATIPONGSE,1986). All farming was mechanized and the only domestic animals in the valley were chickens, dogs, and cats. During the dry season, November-May, about 55% of those older than 12 years reported spending at least one night per week tapping rubber, an activity usually done between 0100 and 0600 h. All but 3 of 48 houses in the study area reported having mosquito nets and 82.8% of residents claimed to have used them during May 1985; those nets examined were generally in good condition. Mosquito collections Man-biting Anopkeles were collected simultaneous-
ly at each of two sites (Fig. 2) for 7 consecutive nights each month during the period of the full moon (ROSENBERG & MAHESWARY,1982). At each site captures were made for 50 min periods beginning on the hour from 1900 to 0450 h bv a two-man team sitting at ground level l-2 m outside an inhabited dwelling; during rain, collections continued under the nearest houses. No adjustments were made for the slight seasonal variation in daylight length. Teams
Fig. 2. Plan of the study site, Ban Phluang, Thailand, showing relationships of houses and mosquito collection sites to vegetation. Vegetation was classified using aerial photographs (19 000) with ground control. In general, rice fields and cleared land have no shade; scrub and immature orchard are partly shaded; plantations and forest are heavily shaded all day. A and B indicate collection sites.
24 were replaced between 2350 and 2410 h; consequently the sixth collection period was only 40 min long. Approximately midway between the regularly scheduled October and November 1983 collections an extraordinary 6 d collection was made using normal procedures; except where explicitly stated, data from this collection were excluded from analyses. Several days each month were devoted to unsystematic surveys of breeding sites within a 5 km radius of the biting collection sites. Whenever possible, larvae were reared to adults and preserved as matched sets with their 4th instar and pupal exuviae for identi6cation. Adult mosquitoexamination Each morning after a collection species were identified, ovaries examined for parity, and the midguts and salivary glands of parous females examined for parasites. During the one month each autumn when populations peaked only An. dirus, An. minimus, and An. muculatus were checked for parity and parasites. During the first year glands with sporozoites were dried on microscope slides and tested with the immunofluorescent antibody (IFA) technique using sera from rabbits immtmized with either P. tivax or P. falciparum sporozoites collected from the salivary glands of laboratory-reared An. dirus fed on gametocyte carriers. Rabbits each received an initial intravenous inoculation of lo6 sporozoites followed by 3 weekly boostersof lo5 before blood was drawn; specificity was checked against 6 sporozoite isolates. In the second year, sporozoites were idenTable 1. AnopMw species caught biting man on 672 man-nights in Ban Pbluang, Thailand
Soecies An. An. An. An. An. An. An. An. An. An. An. An. An.
aconitus annulati dints jam&i
kanvarii kochi
letifer maculatus minimus peditaeniatttf
philippknsisb teselahu umbrosus
No. infected
No. cauaht
oocvsts
Swrozoites
156
0
0
18
0
0
1744
42 0
29 0
821
5
0
1
0
0
2 567
0 4 25 2
0 0 0 0
1 0
0 0
0
0
6
1563 2179 132 194 5
‘Includes An. peditaeniatw, An. barbirostris, An. campestris, An. argvrops, An. sinens&An. nigenimus,An. nit&k, and An. hyrcanur. %&da An. philippinensis and An. niwipes.
tified using an enzyme-linked immunosorbent assay (ELISA) (WIRTZ et al., 1987); all specimens were tested in triplicate. Calculations To facilitate analysing seasonaldifferences in vectorial capacity, several estimates and assumptions were made. Daily survival rates were computed after the method of DAVIDSON (1954) according to the relationship p= Ti,,, where T is the proportion parous and x is the length of the blood f&g interval, here assumed to be 3 d. Vectoral capacity (GARRET-JONES & SHIDRAWI, 1969)was calculated as C=(ma)apWnp, in which mu is the daily man-biting rate, a is the daily rate of blood feeding on man, p is the daily rate of survival, and n is the length of the sporogonic cycle. The rate ma is calculated from the biting collections and p from the proportion parous, as above. The value of a was chosen as O-33, which assumesa blood meal every third day on man; captured An. dirus usually oviposited the second night after blood feeding; how soon after that they refed is unknown. For convenience the length of sporogonic cycle for both parasite species was taken to be 12 d, although P. vivax would usually mature at about 11 d and P. fakiparum at 13 d at 27°C. In this paper inoculation rate (h’) means the number of bites received per person per day by mosquitoes with sporozoite-infected glands; the rate is estimated from dissection results. A mosquito with infected glands is described as ‘infective’, whereasone carrying only oocysts on the gut is termed ‘infected’. Results
In 168 nights of regular biting collection (672 man-nights), 13 speciesor speciesgroups of Atwpkeles were caught (Table 1). Only A. dirus was found to be infective (Table 2): 1.55% of the salivary glands and 2.41% of the midguts were parasitized. An. minimus was found to be infected relatively often during the dry months with oocysts which appeared healthy (1*60%), occasionally differentiated, and may have been a secondary vector. The median numbers of oocysts for both specieswere similarly small, 3-5 for An. dirts and 3.0 for An. minimus; oocyst ranges and frequency distributions are compared in Fig. 3. Daily inoculation rate for the annual 7 month transmission season(Fig. 4) was 0.102 in the first year and O-036for the second; this implies that each village resident received nearly 22 inoculations of sporozoites
Table 2. Seasonal changes in vectorial components for Anoplieles dhs Year 1 2 Transition Season Wet Wet Transition (Jun.-Sep.) (Oct.) (Nov.)” (No?%ay) Months (Jun.-Oct.) (Nov.) (De?%ay) Man-biting rate (tn~)~ 1.22 12.46 24.21 1.93 0.50 11.61 2.33 Parous 044 044 0.51 0.70 0.21 0.59 0.58 Daily survival @) 0.76 0.76 0.80 0.89 0.59 0.84 O-83 Vectoral capacity (C) 0% 0.56 2.42 1.28 0.001 2.60 0.48 Inoculation rate (h’) 0 0 0.29 0.10 0 0.07 0.03 ‘Extraordinary 6 d collection midway between regular October and November collections. bAbbreviations: ma=number of vector biting/man/day; p=parous rate1’3;C=(mu)up”l-lnp, where a--0.33, n= 12; h’=number of infective mosquitoes biting/man/day.
JJASONDJFMAMJJASONDJFMAM
140120-
8
loo-
0
80-
2
IO-
60-
; > B
5-
03
0
0
8
c
n
,ilnfl
rln
nn
-
I-L
4020-
1983
1984
1985
Fig. 4. Monthly relationships at Ban Phhmg, Thailand between (A) rainfall >I00 mdmonth (upper rectangle) and percentages of Amphela dina with (B) sporozoites, (C) oocysts, (D) parous, and (E) number of An. dims biting/man-day. Data for (BHE) collected for 7 nights/month. Parity was not calculated for July-September 1984 (marked with asterisks) because of small sample sizes.
Dirus
Minimus
Fig. 3. Frequency distribution of malaria oocysts for Anopheladims and An. minimusdissected at Ban Phluang, Thailand from June 1983 to May, 1985.
Table
3. Malaria
infection
in Anopbeles dhs: A
No. caught Parity (%) Inoculation rate Percentage of catchb with Oocysts only Oocysts+sporozoites Sporozoites only Oocysts (total) Sporozoites (total)
548 65.7 4.2
comparison
Year 1 B 316 73.9 7.7
in the first seasonand nearly 8 in the second. Of the 20 infective mosquitoes collected in the lirst year, 14 reacted by IFA with P. falciparum only, 3 with P. vivux only, 1 with both, and 2 with neither, apparently becausesporozoites had washed off the slide. In the second year 4 mosquitoes reacted with P. falciparum,
by year and site (A and B) of collection Total 864 69.3 6.0
A 571 54.8 1.2
Year 2 B 307 63.8 3.0
Total 878 58.3 2.1
0.9 (5)
1.0 (3)
0.9 (8)
2.1 (12)
2.3 (7)
2.2 (19)
0.9 (5) 0.4 (2)
:‘;. ii;
1.4 (8) (12) 0.9
0.2 (1) (1) 0.2
0.7 1.0 (2) (3) 2.9 (9) 1.6 (5)
;:; 1;;
3.2 (10) 4.1 (13)
“Calculated for dry months only: November-May bNumhers of infected mosquitoes in parentheses.
2.3 (20) 2.3 (20) in year 1, December-May
2.3 (13) 0.4 (2) in year 2.
26 catch during that hour. During the wet seasonAn. dirus larvae were found in and around the study area
19 20 21 22 23 24 01 02 03 64 d5 Time [ Hourof Day ) Fig. 5. Cumulative, outdoor night biting pattern of 1744Anopheles dim caught during 672 man-nights. The 24-01 h collection period was 40 mill long; all others were 50 min.
2 with P. vivax, and 1 was negative by ELISA. The 2-year proportion of 76.0% P. falciparum sporozoites compares to a prevalence rate of 789% for P. fakiparum gametocytes (ROSENBERG et al., 1990a). Abundance, parity and infectivity of An. dirus occurred in annual cycles that appeared to be linked to rainfall (Fia. 4). Durina the height of the summer rains both ih; size of the biting population (Fig. 4E) and the percentage parous (Fig. 4D) fell to relatively low levels, but as the monsoon ended in October and November the number caught increased explosively and parity began to rise. The population remained large for less than 60 d, but parity continued high through the dry season; virtually all infected and infective An. dirus were caught in the latter half of this brief population surge or during the dry months that followed, when bitine averaaed 2.l/man/ninht (Fie. 4B,C). In Table 2 these sea&al events are-a&y&d by assuming that changesin parity indicated changes in daily survival rates; survivorship was then used to calculate vectorial capacity for each of 3 annual periods; wet, transitional, and dry. Vectorial capacity and inoculation were highest during the latter part of the transitional period, owing in particular to the high man-biting rates. The second year inoculation rates were 65% lower than the first; calculated survival rates (Table 2) did not rise in the post-transition period as they had in the first year. The possibility that a decreasein mosquito longevity was responsible for lowering both parity and infectivitv was sunported by the observation that the number -of gut-ii&ted Ah. dirts also positive for swrozoites fell from 60% the first vear to 14% in the skcond (Table 3); this implies that fewer infected females were living long enough for oocysts to mature. Although 44% fewer An. dirts were caught biting at site B. a resident was twice as likelv to be bitten bv an infective mosquito there as at site A, about Sk m away (Table 3). The same proportion of females at both sites was infected with oocysts alone, but at B parity was consistently about 85% higher than at A, and the proportion carrying sporozoites was 3-2 to 4.0 times greater. This suggests that vector survival was higher at B and that the females caught tended to spend their lives in a small area. Peak densities of An. dirus bit outdoors between 2200 and 0100 h (Fig. 6). There was no discernible deviation from this biting pattern for sub-populations of parous or infected females, or between the 2 collection sites. The replacement at 2400 h of the collection team may have caused the relatively low
in wheel ruts under mature rambutan trees, in abandoned charcoal pits under mature rubber, in a discarded 55 gallon [250 litre] metal drum under forest, and in a pit dug next to a shaded stream. During the dry months larvae were occasionally found at the bottom of a concrete lined well. about 1.5 m diameter and 5 m deep, at collection site B; sampling along stream edges in the village and nearby forest was negative. Discussion Ban Phluang was hyperendemic for malaria because it was a suitable habitat for An. dirus. The contribution of An. minimus to transmission was inconsequential and the other species sometimes found to be infected with a few small oocysts of indeterminate parasites were unlikely to have been vectors. The annual An. dirus infectivity rate of 1.6% compares closely to other long-term rates for this and related species (ISMAIL et al., 1975; KHIN-IMAuNG-KYI & & WINN, 1976; WILKINSONet al., 1978; ROSENBERG MAHESWARY,1982). Since sporozoites were found during only 7 months per year, inoculation rates were nearly twice as intense as the annual rate indicates. The species ratio of sporozoites was remarkably similar to that of gametocytes in humans, implying that both P. falcipanrm and P. vivax were being transmitted equally efficiently, even though there is evidence that glands naturally infected with P. falciparum may contain more sporozoites (BURKOTet al., 1987). The small median number of oocystsfound (3.5) was commensurate with the low gametocyte densities prevalent (MUIRHEAD-THOMSON, 1957; ROSENBERG et al., 1990a) and suggests that most gland infections, and possible most sporozoite (ROSENBERG et al., 199Oc), were small. The most striking characteristic of transmission at Ban Phluang was the brief transitional period at the end of the monsoon. A. dirus breeds in small, fresh, temporary pools and its numbers and proportion infective have usually been observed to peak sometime during the wet season(COLLESS,1952; SCANLON & SANDINAND, 1965; WILKINSON et al., 1978; ROSENBERG & MAHESWARY,1982). At Ban Phluang, however, transmission began as the rains ended, when, in less than 60 d, the man-biting rate rose steeply, survival lengthened, and the inoculation rate climbed to nearly its highest level of the year. The rapid start of transmission caused an equally rapid increase in P. falciparum gametocyte prevalence (ROSENBERG et al., 1990b), a phenomenon that may have compensated for the subsequent decline in vectorial capacity. This pattern contrasts with a gradual increase in seasonal transmission seen in Africa (MOLINEAUX & GRAMICCIA,1980). The large differences in inoculation rates observed between seasons,years, and collection sites probably resulted from differences in vector survival rates, which can have a hugely disproportionate effect on the probability of a mosquito becoming or remaining infective (ONORI & GRAB, 1980). Parity rates can be rough measures of daily adult survival when emergence is stable (GARRETT-JONES & GRAB, 1964), and these moved in concert with changes in observed transmission rates. The ratio of gland-infected to
27
gut-infected An. dims also implies that survival was longer in the first year than in the second, and longer at site B than at A. (It has been suggested by DYE (1986) that changes in vector population size may predict infectivity as accurately asestimatesof longevity; clearly this was not true at Ban Phluang.) The methods we used were not sensitive enough to reveal why survival rates differed; there was no clue in rainfall pattern and the possible influences of microclimate, predators, and pathogens were not explored. It may not be coincidental, for instance, that the reduced transmission rate in the second year was preceded by a very low density of adults during the monsoon. Almost twice as many An. dim-s were caught at site A, where vegetation was denser and less cultivated, than at site B, 800 m away, but a much smaller proportion of those at A lived long enough to become infective. If female An. dims dispersed randomly from points of blood feeding and oviposition, then there should have been no sharp discontinuities in survival rates within the small study area. The rice paddies between sites A and B may have limited random dispersion, but An. dims is capable of flights over uncovered ground of at least half a mile [0*8 km] (COLLESS,1953). We infer that most females spent their adult lives within a short radius of the same hosts. Such behaviour would increase the probability of a given mosquito feeding repeatedly in the same house and could lead to clustering of cases if transmission were low. It also suggests that many females may have deposited ova & places such as moist denressions (WILKINSON et al.. 1978: ROSENBERG, lY82) or streams which were ‘convenient for them but unfavourable for their offspring. The difficulty of finding dry season breeding places supports this hypothesis. Mosquito nets hung in most houses and were usually in good condition; when used with a sleeping mat, as was customary, they should have excluded mosquitoes. Nevertheless, everyone was infected with malaria. The population of An. dims may have been suppressed by heavy rainfall but other man-biting species thrived during the monsoon; if the nets were used mostly to protect against the nuisance of being bitten rather than of getting malaria, then their use was probably greatest during the season when the danger was least. Furthermore, nets were usually made to accommodate several people and so the probability of a sleeper being bitten through the net seemsgreater. Malaria incidence among the hundreds of pilgrims, mostly non-immune, who spent one or more nights at the village monastery each February is unknown. Natural forest is not essential habitat for An. dims, which has commonly been found in hilly, shaded terrain in which the dominant vegetation was scrub, tea, or rubber. Once it was indigenous to large areas of south-eastern Thailand but it disappeared, along with much malaria, when trees were felled in the 1960s to make room for cassavacultivation. More recently, the government has successfully promoted the planting of rubber and eucalyptus (ANONYMOUS, 1987) on this same land in response to depressed cassavaprices. If An. dim recolor&es this substantially restored habitat the.n malaria endemicity may also return.
Acknowledgements
Expert technicalhelp wasgiven by Ruan Thaopha,Sanit Nakngen, Vi&it Phunkitchar, Inkam Inlao, Sompom Chanaimongol,ChunmongNoigamol,PrajimBoonyakanist, SuwattanaVongpradit,andPhillip E. Fleischer.Kol Mongkol entereddatainto a computerand RampaRattanarithikuI con6rmedsomespeciesidentifications. GeorgeWard and Robert Wirtz of the WalterReedArmy Institute of Research aidedin the adaptationof IFA and ELISA; J. Gingrich and JetsumonSattabongkotperformedthe ELISAs. The Director and staff of the Malaria Division, Ministry of Public Health, Thailand provided invaluable cooperation.
References Anonymous (1987). Shell seekssupport for eucalyptus plan. The Nation (Bangkok), 13, (3807), 25. Burkot, T. R., Graves, P. M., Cattani,J. A., Wirtz, R. A. & Gibson, F. D. (1987). The efficiency of sporozoite transmission in the human malarias. Plasmodium falciparum and P. vivax. Bulletin of, the World Health Organization,
65, 375-380.
Colless,D. H. (1952).Observationson the periodicity of natural infections in the anopheline mosquitoes of Borneo. Medical Journal of Malaya, 6, 234-240. Colless. D. H. (1953). Observations on the fliaht range ofA. balabacensis~ Baisas (= A. leucosphyrus, &ct.). Medical Journal of Malaya,
7, 179-184.
Colless, D. H. (1957). The Anopheles leucosphyrus group. Proceedings of the Royal Entomological Society of London,
108. 37-116. Davidson, G. (1954). Estimation of the survival rate of anopheline mosquitoes in nature. Nancre, 174, 792-793. Dye, C. (1986). Vectorial capacity: must we measure all its components? Parasitology Today, 2, 2-8. Garrett-Jones, C. & Grab, B. (1964). The assessment of insecticidal impact on the malaria mosquito’s vectorial capacity, from data on the proportion of parous females. Bulletin
of the Worki Health Grganization,
31, 71-86.
Garrett-Jones, C. & Shidrawi, G. R. (1969). Malaria vectorial capacity of a population of Anopheles gambiae, an exercise in epidemiological entomology. Bulletin of the World Health Grganixation, 40, 531-545. Ismail, I. A. H., Notananda, V. & Schepens, J. (1975). Studies on malaria and responses of Anopheles balabacensis and Anopheles minimus to DDT residualspraying in Thailand. Pan I. Acta Tropica, 31, 129-164. Khin-Maung-Kyi & Winn, S. M. (1976). Studies on malaria and Anopheles balabacensisbalabacensis and Anopheles minimus in Burma. Geneva: World Health Organization,
Mimeographeddocumentno. WHO/MAL/76.875.
MacArthur, J. (1951). The importance of Anopheles leucosphyrus. Transactions of the Royal Society of Tropical Medicine and Hygiene, 44, 683-694. Molineaux, L. & Gramiccia, G. (1980). The Garki Project.
Geneva: World Health Organization. Muirhead-Thomson, R. C. (1957). The malaria infectivity of an African village population to mosquitoes (Anopheles gambiae). American Journal of Tropical Medicine and Hvaiene. 6. 971-979.
Onori, E. & Grab, B. (1980). Indicators for the forcasting of malaria epidemics. Bulletin of the World Health Grganization, 58, 91-98. Patiuonase, S. (1986). Susceotibilitv of Thai malaria vectors to insecticides. In: Se&f Conference on Malaria Research. Bangkok: Ministry of Public Health, p. 170. Rosenberg, R. (1982). Forest malaria in Bangladesh. III. Breeding habits of Anopheles dirus. American 3oumal of Tropical Medicine and Hygiene, 31, 192-201. Rosenberg, R. & Maheswary, N. P. (1982). Forest malaria in Bangladesh, II. Transmission by Anopheles dirus. f8y.r; Journal of Tropical Medicine and Hygiene, 31, Rosenberg, R., Andre, R. G., Ngampatom, S., Hatx, C. & Burge, R. (1990a). A stable, oligosymptomatic malaria focus in Thailand. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84, 14-21.
28 Rosenberg, R., Andre, R. G. & Ketrangsee, S. (1990b). Seasonalfluctuation of Plrrmrodiumfalcipan gametocytaemia. Transactions of the Royal So&g of Tropical Medicine and Hygiene, 84, 29-33. Rosenberg, R., Wirtz, R. A., Schneider, I. & Burge, t(. (l!BOc). An estimation of the number of malana sporoz&tes ejected by a feeding mosquito. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84, in
press. Scanlon, J. E. & Sandinand, U. (1%5). The distribution and biology of Atwpheks balabacensis in Thailand. Journal of Medical Entomology, 2, 61-69.
I
Corrections
Wilkinson, R., Gould, D. J., Boonyakanist, P. & Segd,, H. E. (1978). Observations on Anopheks balubacenm in Thailand. Journal of Medical Entomology, 14, 666-671. Wirtz, R. A., Burkot, T. R., Graves, P. M. & Andre, R. A. (1987). Field evaluation of enzyme-linked immunosorbent assaysfor Plusm&ium fakiparum and Pkwtwdium &ax sporozoites in mosquitoes (Diptera: Culicidae) from Papua New Guinea. Journal of Medical Entomology, 24, 433-437.
Received 15 December 1988; accepted fm publication 30 March
1989
I
7’ransactims, vol. 83, part 4, p. 510. The code number of the isolate of Leishmznia tropica was cited incorrectly by the authors; it should have been MHOMIMAI871LPN 39. Transactions, vol. 83, part 6, pp. 788 At 789. Figures 1 and 4 should be transposed, as should Figures 2 and 3; the lengeds as printed are correct. We apologise for this error.