Entomological aspects of filariasis control in Sri Lanka

Bull. Org. niond. Sante' 1974, 51, 133-143 Bull. Wid Hith Org.J Entomological aspects of filariasis control in Sri Lanka F. L. LAMBRECHT 1 Historical events and suitable environmental conditions in the southwestern coastal areas of Sri Lanka have led to the establishment of a zone of endemic filariasis caused by Wuchereria bancrofti and transmitted by Culex pipiens fatigans. The previous Brugia malayi foci, scattered over widely dispersed areas of the island, were apparently completely eliminated as a result of control of the Mansonia vectors by the destruction of the larval host plants in their swamp habitats. Control measures by the Anti-Filariasis Campaign against W. bancrofti and C. p. fatigans have greatly reduced the human infection rates in the endemic coastal belt and have kept the rate in the dense population to less than I % over the last several years. This paper assesses the entomological aspects of the control programme during the years 1970-72. Elephantiasis has been known in Sri Lanka for many centuries. Surveys during the late 1930s showed widespread, scattered foci of Brugia malayi in many parts of the island (Dassanayake, unpub- lished report, 1939). The transmission of this parasite was successfully controlled by the elimi- nation of the Mansonia mosquito vectors during the 1947-52 campaign for the destruction of larval host plants in their swamp habitats. During those same years, however, the incidence of filariasis caused by Wuchereria bancrofti rose rapidly in the southwestern coastal areas where now the disease has become endemic and is transmitted by Culex pipiens fatigans mosquitos, which multiply under extremely favourable conditions. Control measures, based upon case finding and treatment, and mosquito control by the use of larvicides (e.g. fenthion) are carried out from several stations maintained by the government-financed Anti-Filariasis Campaign. An analysis of current and earlier entomological data indicates that the present mosquito control measures are inadequate to prevent the circulation of the parasite. This is confirmed by the persistence of infections in the vector which, even if low, casts doubt on the finding (by the thick smear technique) of human infection rates of less than 1 %. I Acting Project Leader, Arbovirus Vector Research Unit, P.O. Box 104, Enugu, Nigeria. GEOGRAPHY Sri Lanka lies in the Indian Ocean between5055' and 9°50' north latitude, and between 79042' and 81°52' east longitude. The island is located within the monsoon belt, where the predominant wind directions are from the southwest during May to September, and from the northeast during December to February, and where weather conditions are variable during the intermonsoon periods. Climate and topography divide the island into a " wet zone " in the southwest, and an " intermediate zone " and a " dry zone " in the rest of the island. This accounts for the distribution of different types of vegetation, ranging from wet tropical evergreen forests in the southwest to tropical dry savannah forests in the north, as well as forests in the central mountain range and various types of intermediate evergreen or deciduous vegetation in the rest of the island (Fig. 1). The high rainfall in the southwest, in conjunction with other factors, has encouraged the settlement of people who now average over 386 per km2 in the coastal areas. The climatic conditions and the dense human population, which is distributed in an almost unbroken string of coastal towns with often very poor sanitation, have created ideal conditions for the breeding of C. p. fatigans and the transmission of W. bancrofti all the year round. 3247 -133- 134 F. L. LAMBRECHT BIOCLIMATIC MAP OF SRI LANKA *9 JCtonbaffna * Hot Arid. Lowland IMl Hot Dry Lowland 2E Hot Moist Lowland U Hot Wet Lowland u Cool Wet Highland ig T it z ~~~Endemic Filariasis Belt Trincomnalee Bl Puttalam | -1Batticaloa Negombo Coiombos 1 1 Ka_ nltatragama Galut F.L.L72 Fig. 1. Map showing major bio-geographic regions in Sri Lanka. Ecologically, both the hot wet lowland and parts of the moist lowland would seem suitable for trans- mission of W. bancrofti throughout the year. Because of its high degree of urbanization, Jaffna, in the north, would also seem a potentially suitable area for the transmission of filariasis. ENDEMIC AREAS AND SUSCEPTIBLE AREAS Endemic filariasis is localized in the southwestern coastal zone, a 5-km-wide strip from Negombo to Matara, a distance of about 200 km, with a population at risk of roughly 2 million. Scattered cases of filariasis are found in adjacent areas and in a number of large towns inland, owing mainly to the movements of people. As indicated earlier, the present human infection rates have been brought down to less than 1% and have hovered around the 1% mark for the last several years.a The total infection rate in C. p. fatigans has remained largely stationary between 2.0% and 1.2 , with no change in the infectivity rate of around 0.6% a The Millipore technique would perhaps demonstrate a much higher infection rate by detecting cases of low parasitaemia missed by the thick smear technique. This would explain the consistent finding of infections in the vector. It is thus clear that an equilibrium has been reached between the different factors concerned with filarial transmission, i.e., the efficiency of mos- quito control and parasite control, selection of vector and parasite strains, and movements and increase of population. No marked change in the infection rates are expected to occur as long as this equilibrium remains unbroken. The fact that the introduction of W. bancrofti parasites by human carriers into towns outside the endemic zones could start satellite foci was demon- strated by Wijetunge (7) in a survey of the student population at Peradeniya University, near Kandy. An examination of night blood from 7 467 students in 1957-62 revealed W. bancrofti parasites in 154, giving an infection rate of 2%. Dissection of 160 C. p. fatigans mosquitos from the campus dormito- ries showed that 6 carried W. bancrofti larvae, giving a mosquito infection rate of 3.7%. The failure of transmission of W. bancrofti in the rest of the island in the presence of C. p. fatigans cannot easily be explained, unless it is assumed that the disease in the endemic belt was brought down in time to a level low enough to prevent its further spread inland. However, the movements and resettlement of people could not have failed to introduce W. bancrofti infections into many areas. The circumstances that led to the start of a small focus in Peradeniya, described above, could be easily repeated in other densely populated cities. In Fig. 2 are given the numbers of bus passengers travelling between the infected coastal towns and important inland towns. Since about 1% of the passengers from the endemic areas could be pre- sumed to be infected, the frequent introduction of infected cases to any inland town could lead to the establishment of an independent focus of disease in that town. During the investigations in 1972, 3 098 C. p. fa- tigans mosquitos from 28 different locations outside the endemic filariasis belt were dissected. Infected mosquitos were found in two towns only: Puttalam (2 infected mosquitos in 908), 100 km north of Negombo (considered to be the northern limit of the endemic zone), and Polgahawela (1 infected mosquito in 133), about 25 km east of the eastern limit. The spread of W. bancrofti from the coastal areas into other parts of the wet lowlands must therefore be considered to be possible when human population densities and the number of introduced cases have reached a critical level. What, then, are the chances for the spread of ENTOMOLOGICAL ASPECTS OF FILARIASIS CONTROL SRI LANKA EUCOP. Farm hsfmlisogM _m , Akmamssa ~~~~~~~~~~~4, 0 ! k. if... r f . mstm'4i)~ ho. *p>sF d * Eisidsi FIdmubl b Fig. 2. Map of the endemic filariasis belt (shaded areas) showing the 15 control stations along the coast and their infection rates in host and vector. The figures along the lines connecting coastal areas with inland towns are the average numbers of bus passengers travelling between these points every month. bancroftian filariasis into other climatic zones of Sri Lanka? Areas such as Jaffna peninsula in the north are extremely densely populated and the absence of filariasis in the presence of C. p. fatigans mosquitos is difficult to explain, unless the trans- mission of W. bancrofti is inhibited by certain factors in those areas. One such factor could be a shorter life-span of C. p. fatigans, perhaps resulting from unfavourable climatic conditions. The normal development cycle of W. bancrofti in C. p. fatigans in Sri Lanka takes from 10 to 12 days (4,6). Shorter periods, however, were demonstrated by Abdulcader et al. (4), who found that W. bancrofti microfilariae did not develop beyond the first stage at Nuwara Eliya, a hill station at about 1 900 m above sea level, although the vector itself was commonly found there. Development was normal, however, when C. p. fatigans mosquitos from Nuwara Eliya were used in experiments carried out at the laboratory in Colombo. It would seem that, all the other factors being equal, the climate at the hill station (average annual temperature: maximum 20°C, minimum 12°C; compare with Colombo, average temperature: maximum 29°C, minimum 23°C) was responsible for the failure of W. bancrofti to develop at Nuwara Eliya. The possibility of a shorter life-span of C. p. fatigans mosquitos, owing to climatic conditions, in the dry zones of Sri Lanka should not be excluded, and this might explain the absence of transmission of bancroftial filariasis in those areas. Age-grading, determined from ovariole dilatations by Detinova's method (5), of C. p. fatigans mosquitos from various areas of the endemic belt and from localities in the dry zones, during 1971, indicated a generally shorter life-span for mosquitos from the latter (Lambrecht & Fernando, unpublished report to WHO, 1972). Thus, the average proportions of C. p. fatigans mosquitos in the P3 and P4 groups, i.e., old enough to harbour mature W. bancrofti larvae, was 8.3% (range 3.3-13.6 Y.) from the endemic belt but only 2% (range 1.0-3.0%) in the dry zones. Among the factors influencing longevity, climatic conditions would seem of major importance, especially the relative humidity. Meteorological data from the wet and dry zones, however, show only slight differences between the averages for relative humidity, in spite of a marked difference in rainfall between the 2 zones. Thus, although some gross meteorological characteristics do not seem to be relevant, it is possible that the micro- climatic characteristics of the resting places of C. p. fatigans mosquitos may help to explain the differences in life-span. 135 F. L. LAMBRECHT BIONOMICS The following 25 mosquito species have been identified from collections of adults and larvae in the filarial areas of Sri Lanka: Nine Culex: bitaeniorhynchus, fatigans, fuscanus, fuscacephalus, gelidus, minutissinus, nigropunctatus, sitiens, and tritoeniorhynchus. Six Anopheles: barbirostris, hyrcanus, jamesi, subpictus, tesselatus, and vagus. Six Aedes: aegypti, albopictus, lineatopennis, pallidostritatus, piperselatus, and vexans. Three Mansonia: annulifera, crassipes, and uni- formis. One Armigeres: subalbatus. It is to be noted that Anopheles culicifacies, the sole known malaria vector in Sri Lanka, is found mainly in the dry and intermediate zones, where it breeds almost entirely in riverbed pools formed during periods of minimal rainfall. It is rarely found in the lowland streams of the coastal filarial areas where the rainfall is regular and the rivers are swift. A major problem in the filariasis control pro- gramme is that there is no natural barrier, which might help to contain the infection, on the land side of the usually described endemic belt. Filarial infections are carried and maintained in the populated coastal towns, between which are situated coconut plantations and small hamlets. Owing to budgetary limitations, mosquito control is carried out, in practice, only in the densely populated areas which, of course, harbour the highest numbers of filaria carriers. This piecemeal method is detrimental to the efficiency of mosquito control as a whole and makes evaluation difficult. Besides, although both controlled and uncontrolled zones belong to the same natural environment, the congested urban areas create a biotope no longer comparable with conditions in the rural parts. However, there are no ecological barriers that could prevent reinfestation of the mosquito- controlled areas from the uncontrolled parts. The distribution of species in collections of adult mosquitos and larvae differs only slightly between the treated and untreated zones, except for a higher proportion of C. p. fatigans in the collections of larvae from the controlled zones (Tables 1-3). Table 1. Numbers of mosquito larvae (all species) in breeding sites in controlled and uncontrolled areas a Controlled areas Uncontrolled areas Both areas Type of breeding site no. of no. of no. of larvae larvae larvae husk pits 6 429 9.8 969 9.6 7 398 9.8 discarded receptacles 27 017 41.3 3 723 36.7 30 740 40.5 trenches 3 472 5.3 850 8.3 4 322 5.7 tanks 1 777 2.7 144 1.4 1 921 2.5 catch pits 7 384 11.2 403 4.0 7 787 10.2 drains 3 312 5.0 225 2.2 3 537 4.6 boats 228 .3 0 0 228 .3 arecanut pots 7 401 11.3 1 859 18.3 9 260 12.2 wells in use 339 .5 41 .4 380 .5 unused wells 108 .2 1 .01 109 .2 tree holes b' 1 199 1.8 145 1.3 1 344 1.7 spent nuts b 4151 6.3 1 160 11.5 5311 7.0 swamps b 2 454 3.7 203 2.0 2 657 3.5 crab holes b 74 .1 0 0 74 .1 others b 413 .6 443 4.3 856 1.2 total 65 758 10166 75 924 a Breeding in water plants is not represented in the list, because these plants were not examined. b These were not made by man. 136 ENTOMOLOGICAL ASPECTS OF FILARIASIS CONTROL W N CWawII. ui C- 1: *i ' *.. * C6 C" * ,W I_,I. r- r- o. Co0N N r- Co 'm lot co 0 Co W co 0 co 0 Co4 ,Co N C')q- m 0 Co N Co4 Co to co N N4 N4 N4 ao _ ~ _N N CY (0 CF c :9 Co 0 co W" W N N CC 0 CD Co Co 0 N 0 0 N o CC) CoCo Co Cw Co co W u) ) co N Nm lt CD CIO co * li 0Co o Co 0 L 0CC v- 0 m o CoCC) Co co .' 0 co N- Co CC) co Co N co 9 Co co co ODN r- co CC) C') 0 W- 0 Co Nrr- 0 oo N N Co Co co -e CON4 Co CC) 0C CD cs Co 0C Co COR Co N 04 Co co N COO _C C') m co LO co C. v- 04 co N C T- Coi co W 0 CC) co - CC CO N 0 Co 0 N Co CC) Co co CoO 0CC Co 04 Co N CIO ND N U0 N Co0 Co co Co co- r. CD a Ct 0 w 0 OC ._ = oC 137 co U) N CD Co4 N U) Co Co Co4 _ 0 N N00C OD0NO 0N (D 0 Co ,-.C C o 0a _0a 0 ~ 0 C,. q o Co 0CD .C D - Co C) 04 Co 1-4 CD co . Co co co D r- CC, N co Co 0 N co 04 N ow 0 a) N co CN 0 0 Co C') 0N l) o N Co Co co N r- a) CD Co 0 am N N Co Co) C') D Co N Co4 co 0. It 0l 06 *0 CN CC, co oi Cs Co t- Co N Co 19t cli co 0 Co ci r. - 0 Co 0 t 0 0 0 coOD a, C14 ._ c CT CO) Co CU E ._ CU *0CD 0 C. 0 ._ 0 %6-Q o 0) C cn aD a, ._ Q 0 *0 a, 0 Co Co -o 0 Co 0 cv ad 0 U) co 0 U) F. L. LAMBRECHT o N C I o w N oo) (V U) 0 a) o N U) s w N U) u) _ 1-0 . .co ) _ _ m) > oo U CD co cli N 0 0 cn C') 0 T- U)0 0 U) co N 0 U) uo o LO 0 U ,O ) 0 CD 0 W) 0- C') UV) N~N1 U)4 C') C) CY) U) U) o U)00 U,) 0co 0 o _- 0 I- uo N O N LO 0 0 0 0 Nl U) LO 0 0 N C) le o UO cnV 0 N o U ) 0 NOq T- o) - N coCD, cN co 0 0 0 00 U) o0 0 0 b _ X o ONr CD '- U) C') T-N CD CD CY) 0L CL . o U co 0~~~~~ L) C3 (C -, 0 coC') U) O N C'f) Uz) 0 0 0 N C' CY) LO 0 LU) N, A 0) (A co C C 06 0) cl oC') 0 s) co a) Q 0 0 0) 138 cr0m U) U) N0) NvCO N U) omNo n _ CO NO U) U- co %- oo N U) CU) U) ND _ 0 -e co 0 le ,it 0 O N U) 0 C'_ at 0 I- RZ. 6 ae C ID cO5 o a s; 6 0.o q0 '0 6M 0 o 0, . -- co CD CD 0 -o ., a) ._ CL an cO 0 0 co CL an 0 Ca cn .) Q -o E - C~) an a) 0 0._ E a) C cl a) a) CO C_ 0 0* a) 0 E z .0 U)co 0T- oo U) r- o _ eq N c 0 Nc6 co co .4 ,% - I- C') w O 0 0 o80, U) U) () I-) LO 0 ND LO ,i_ 0 to 0 0 C'j 0 C') ENTOMOLOGICAL ASPECTS OF FILARIASIS CONTROL Heavy breeding is found in both areas in discarded receptacles, e.g., spent coconuts and other types of small water-containers. Adult stages Of the 10 216 dwellings examined in the filariasis belt in 1971, 61.5% harboured mosquitos; 22758 mosquitos were collected, giving an overall average of 2.2 mosquitos per house and an average of 3.6 per mosquito-positive house (figures based on a 5-mi- nutes search per house). A total of 824 hours were spent in collecting resting mosquitos, yielding an average of 27.8 mosquitos per man per hour. The catching rates were 3 times higher in the urban areas than in the rural parts in spite of the control measures in the former. The dwellings examined in 29 settlements and towns outside the filariasis belt numbered 3 899, of which 45% were found to harbour mosquitos; 7 585 mosquitos were collected, giving averages of 2 mosquitos per house and 4.3 per mosquito- positive house. A total of 340 hours of collecting yielded an average of 22.4 mosquitos per man per hour. Densities, however, varied in these " out- stations " from an occupation rate of 9% to 88% of the dwellings, from 0.1 to 5.3 mosquitos per house, and 1.1 to 8.7 per mosquito-positive house; the catch rate per man per hour varied from 1.3 to 59.1. As might be expected, all the high figures related to the larger towns, such as Kandy, Kata- ragama, Puttalam, Jaffna, Horana, Polonnaruwa, Kegalle, and Polgahawela, and to communities living close to these places. From past records on the prevalence of the most common species of mosquitos collected from houses in the filariasis belt (Table 4), it is seen that the filarial vector, C. p. fatigans, largely dominates the mosquito fauna in the houses; about 7 of every 10 mosquitos were of that species despite a slight decrease since the 1949-62 period. Ae. ae- gypti, on the other hand, made a relatively sub- stantial gain in both the controlled and uncontrolled areas. The average total infection rate for all stages of W. bancrofti larvae has oscillated around the 1.5% mark-from 2.0% in 1967 to 1.2% in 1970 and 1.6% in 1972. The average infectivity rate since 1967 remained stationary at 0.6%, but went down in 1972 to 0.3 %. As indicated earlier, it would seem that an equilibrium has been reached between transmission potentials and control meas- ures, so that the infection and infectivity rates should stay the same for as long as this equilibrium is maintained. Larval stages Data concerning the collections of larvae are summarized in Tables 1-3. Tables 2 and 3 list the species most commonly found and their proportions in 15 of the most common types of breeding sites. The predominant species in both controlled and uncontrolled zones was Armigeres subalbatus, Table 4. Data from past records on the prevalence of different species of mosquitos in houses in the filariasis belt during the periods 1949-62 and 1969-70 a Period: 1949-62 Period: 1969-70 controlled uncontrolled controlled uncontrolled areas areas areas areas C. p. fatigans 77.5 86.1 68.5 66.5 C. gelidus 3.3 2.1 3.6 4.5 C. tritaeniorhynchus 3.6 1.9 3.0 3.3 Ae. aegypti 2.3 .05 5.3 3.2 M. uniformis 8.6 6.4 6.8 8.2 M. annulifera 1.0 1.3 .6 .4 Ar. subalbatus .9 .3 5.4 6.1 A. subpictus .8 .6 .7 .9 other species 2.0 1.25 6.1 6.9 a Figures are expressed as percentages. 3 139 F. L. LAMBRECHT followed by C. p. fatigans and Ae. albopictus (the last-named is a vicious biter during the 2 hours after sunset and would not normally be sampled in our routine adult collections, which were carried out mainly in the early mornings). It is seen from Table 1 that only a small proportion of breeding sites were not man-made, i.e., 12.5% in the controlled zones and 19.1 % in the uncontrolled zones. From data on the larvae collected in the uncontrolled zones, it is possible to calculate the extent of suppression of larval development if the present control method (based on weekly applications of larvicides to " permanent breeding pools ") were to be extended to the following breeding sites: A. Breeding sites that could be treated: trenches (with 8.3% of all larvae collected), catch pits (4.0 Y.), drains (2.2 %), and others (4.3 %). Thus, a total of 18.8% of all larvae could be suppressed. B. Breeding sites that cannot be treated: husk pits (with 9.6% of all larvae; cannot be treated because of danger of poisoning to labourers), discarded receptacles (with 36.7% of all larvae; impracticable to spray), tanks (with 1.4% of all larvae; danger of poisoning), arecanut pots (with 18.3% of all larvae; danger of poisoning), wells in use (with 0.4% of all larvae; danger of poisoning), tree holes and plants (with 1.3% of all larvae; impracticable to spray), spent nuts (with 11.5% of all larvae; impracticable to spray), and swamps (with 2.0% of all larvae; only partially treatable). Thus, a total of 81.2% of all larvae cannot be suppressed. The above figures indicate that the control methods at present applied by the Anti-Filariasis Campaign could only be partially effective since they leave a very large proportion of breeding sites undisturbed. In all fairness, it should be pointed out that the above calculations are based on the numbers of positive breeding pools without consideration of the fact that the larvae produced in one catch pit or trench, for instance, might be higher than the numbers of larvae found in one discarded tin. Table 5 shows how mosquito breeding would be affected and how control measures would be improved by specific or combined control methods. The action of destroying only discarded household receptacles in the uncontrolled areas would result in the inhibition of almost twice the number of breeding pools than could be achieved by the present spraying programme: 36.7% (not including spent coconuts) against 20.2%. If, in addition to the removal of the discarded receptacles, breeding in spent coconuts and in arecanut pots could be prevented, the breeding of all species could be Table 5. Expected percentage reduction of mosquito species as a result of treatment or elimination of specific groups of breeding sites in uncontrolled zones a c c01i 3a t B -e: 3 R ".0a larvicidal spraying in catch pits, drains, tanks, boats, trenches, etc. 27.8 31.5 67.0 21.2 15.4 7.8 20.2 elimination of discarded household receptacles and 41.5 15.9 18.3 74.4 66.0 55.1 48.2 spent nuts 69.3 47.4 85.3 95.6 81.4 62.9 68.4 avoidance of breeding in 21.5 2.0 6.0 0 2.5 30.1 18.3 arecanut pots 90.8 49.4 91.3 95.6 83.9 93.0 86.7 prevention of breeding in 4.7 41.6 8.2 0 11.6 5.7 9.5 husk pits 95.5 91.0 99.5 95.6 95.5 98.7 96.2 a The figures in the bottom rows are cumulative percentages. 140 ENTOMOLOGICAL ASPECTS OF FILARIASIS CONTROL brought down by more than half: 48.2 + 20.2 + 18.3, or 86.7%; this would reduce breeding of the filarial vector, C. p. fatigans, by 90.8%. The elimination of these breeding sites in the controlled areas would increase the efficiency of control against C. p. fatigans by 34.2 %. In regard to other mosquito species, especially Ae. aegypti and Ae. albopictus, which are potential virus carriers, the elimination of discarded receptacles alone would reduce the breeding potential of these species by respectively 76.5% and 80.2% in the controlled zones, and 74.4% and 66.0% in the untreated areas. The breeding of mosquitos in husk pits in certain parts of Sri Lanka presents a particular problem (Lambrecht & Kulasingam, unpublished report to WHO, 1973). Husk pits are excavations of variable size, usually not more than 1 m deep, filled with water from a nearby stream, lagoon, or tidal pool, and are used during the making of "coir ", a fibre obtained from the husk of the coconut. After the removal of the nut, the husks are left in the pits to soak for a variable length of time, on an average from 2 to 4 months. After "maturation ", the husks are pounded, often by hand, and the fibres are removed and left to dry. The coir is used for rope-making, as a filling for mattresses or under upholstery, for doormats, and so on. During the period of soakage, the pits become active breeding sites for a number of mosquito species, especially C. gelidus and C. p. fatigans. Although larvicides, such as fenthion, may be used in routine mosquito abatement pro- grammes in the filarial areas of Sri Lanka, these compounds cannot be used in the husk pits, because people have to enter and stay in the water of the pits for prolonged periods of time when stocking or removing the husks. A study of husk pits in the Ambalangoda- Balapitiya area, about 88 km south of Colombo, for 3 months showed that the culicine fauna in these pits was quite different from that in other breeding places. C. gelidus was predominant and accounted for 76.2% of the 3 246 larvae examined; next was C. p. fatigans with only 7.2%. Larvae collected in a variety of pools other than husk pits, but in an environment similar to that of husk pits, were of the following species: C. p. fatigans (37.2%), Ar. subalbatus (25.4%), Ae. albopictus (11.8%), C. gelidus (10.6%), and C. tritaeniorhynchus (7.2 %). In a densely built-up area south of Colombo, the following 5 species: Ar. subalbatus (42.2%), Ae. albopictus (18.8 %.), Ae. aegypti (12.8 Y,), C. p. fati- gans (10.3 %), and C. gelidus (7.0 .), were identified in a variety of breeding sites. Clearly, mosquitos proliferate in the husk pits where the alkalinity of the water seems particularly to favour the larvae of C. gelidus; the latter are found in their highest numbers above pH 7, reaching a peak at pH 10. In other breeding sites, however, the highest densities of C. gelidus are found at pH 6. It would thus seem that the correlation between the hydrogen-ion concentration and the development of C. gelidus is only secondary, because the favour- able conditions in the husk pits for this species are probably due to certain organic or inorganic compounds formed during maturation of the husks or as a result of bacterial growth. These compounds are perhaps also detrimental to the normal development of other mosquito species. Larvae of Ar. subalbatus predominate in a variety of pools, especially in arecanut pots, but are com- pletely absent from the husk pits. In the arecanut pots, the only other species commonly found in association with Ar. subalbatus is C. p. fatigans. Some observations suggest that in these small breeding sites the 2 species may be in competition, leading to the elimination of C. p. fatigans at a certain stage, perhaps as a result of an excess of organic contents released during the fermentation of the arecanuts. Although the proportion of C. p. fatigans in the husk pits is relatively small, the presence of large numbers of these pits in certain areas of the filariasis endemic belt makes them an important breeding site for the vectors of W. bancrofti. In addition, the spread of C. gelidus adds to the mosquito annoyance and there is a risk that it may be a possible vector of viruses. Ae. aegypti larvae were collected from urban breeding sites and were especially numerous in collections from the harbour areas of Colombo. The predilection of Ae. aegypti for harbour areas and their sometimes restricted distribution in those areas were also noted in the Seychelles and may be related to their transport and introduction by sea-craft (Lambrecht, unpublished report to WHO, 1969). In a survey carried out in the Colombo harbour area by the Municipal Public Health Service during 1966-67, the following data were obtained; (1) Resting adult mosquitos: of 3 818 dwellings examined, 439 (11.5%) were found to harbour mosquitos; 1 380 mosquitos were collected (average 141 F. L. LAMBRECHT 0.4 mosquito per house or 3.1 mosquitos per mosquito-positive house). Species distribution: Ae. aegypti (31.5 %), Ae. albopictus (1.0 %), C. p. fatigans (66.1 %), and others (1.4%). (2) Larvae collected from the same areas: of 3 289 breeding places examined, 5.1 % proved to be positive and yielded 1 683 larvae. Species distri- bution: Ae. aegypti (32.4 %), Ae. albopictus (12.6%), C. p. fatigans (53.5 %), and others (1.5 %). (3) Larvae from floating harbour service craft: 273 were examined; 10.1 % of them yielded a total of 320 larvae. Species distribution: Ae. aegypti (88.7%) and C. p. fatigans (11.3 %). (4) Larvae from floating local craft: 537 were examined; 4.4% of them yielded a total of 264 larvae. Species distribution: Ae. aegypti (47.3 %), Ae. albopictus (10.3 %), C. p. fatigans (29.9 Y.), and others (12.5%). The pH values of 680 breeding places from where larvae were collected were estimated by means of wide-range and narrow-range BDH indicator papers. The average pH range of the breeding pools tested was 6.2-8.6 and the mean overall average for all breeding pools was 7.1. MOSQUITO CONTROL PROGRAMME At the moment of writing, mosquito control is carried out from 15 stations in the endemic belt. In addition, the city of Colombo with a population of around 650000 has its own mosquito control unit within the city health services. Mosquito control work is also carried out by local bodies in 5 other stations, financed by the Anti-Filariasis Campaign. Mosquito control is based essentially on the treatment of permanent breeding places with fenthion (BAYTEX EC), 28.5 ml diluted in 4.5 litres of water, at a target dose of 1 mg/litre, which is applied with the aid of pressurized HUDSON/X-PERT sprayers, equipped with a pressure gauge. The number of breeding sites sprayed weekly totals about 85 000 in the 15 control areas. Drains and trenches account for about 55 % of these sites, followed by burrow pits (15%) and catch pits (9%) Husk pits (6.5%) were only occasionally sprayed when they were not in use. When calculated in relation to the total number of breeding sites, the treated sites account for only 20%. Indeed, the control programme would be far more effective if it were possible to prevent breeding in discarded receptacles, arecanut soaking pots, and spent nuts in conjunction with the present spraying programme UME ASPECTS ENTOMOLOGIQUES DE LA LUTTE CONTRE LA FILARIOSE EN SRI LANKA L'e16phantiasis est connu en Sri Lanka depuis plu- sieurs si6cles. Actuellement, la filariose A Wuchereria ban- crofti est end6mique dans la r6gion c6tiere du sud-ouest de l'ile oiu vivent quelque 2 millions de personnes. Le parasite est transmis par Culex pipiens fatigans. Un programme de lutte contre la filariose a ete lance en 1952. Le taux d'infection humaine, qui atteignait A cette epoque 15% dans certaines regions, a ete ramene en 1972 A moins de 1% grace au depistage et au traite- ment des cas et A la lutte antivectorielle par larvicides. En 1970-1972, les taux d'infection et d'infectivite chez C. p. fatigans etaient respectivement de 1,6 et 0,6%. Au cours des cinq dernieres annees, les taux d'infection chez l'homme et chez le vecteur sont restes pratiquement identiques, indiquant un 6quilibre entre l'efficacit6 des mesures de lutte et le potentiel de transmission du para- site et du vecteur. II ne faut pas s'attendre A une modifi- cation de ces taux aussi longtemps que les facteurs de l'equilibre resteront inchang6s. La presence de C. p. fatigans dans la plupart des r6gions de l'ile et les analogies d'environnement rendent possible l'extension de la filariose A d'autres zones que la zone actuelle d'end6micite. On en a signal6 un certain nombre de cas parmi les etudiants de l'Universite de Peradeniya, a environ 100 km de la zone d'endemicite. L'absence d'infections dans certaines r6gions tres peu- plees et apparemment propices, comme Jaffna dans le nord, laisse supposer que des diff6rences de longevit6 du vecteur suivant les regions peuvent jouer un role. De l'analyse des donn6es entomologiques, il ressort que: a) la lutte antivectorielle menee dans des secteurs separes faisant partie d'un territoire integralement infeste ne donne que des resultats incomplets et doit etre pour- suivie sans relache par suite de la reintroduction conti- nuelle du vecteur; b) 1'efficacite de la lutte antilarvaire est entrav6e par le risque pour la sante qu'implique parfois l'usage du fenthion et par l'impossibilit6 d'utiliser les larvicides dans certains gites larvaires; c) un grand nombre de petits gites, aux alentours des habitations, ne peuvent etre 6limines efficacement que par des mesures d'assainissement, ce qui suppose la cooperation active de la population. 142 ENTOMOLOGICAL ASPECTS OF FILARIASIS CONTROL 143 REFERENCES 1. ABDULCADER, M. H. M. & PADLEY, R. Filariasis records in Ceylon. Indian J. Malar., 14: 521-543 (1960). 2. ABDULCADER, M. H. M. Mosquito fauna of the W. bancrofti endemic belt in Ceylon and their role in the transmission of filariasis. Bull. Indian Soc. Malar., 2: 201-212 (1965). 3. ABDULCADER, M. H. M. The significance of the Culex pipiens fatigans Wiedmann problem in Ceylon. Bull. Wld Hlth Org., 37: 245-249 (1967). 4. ABDULCADER, M. H. M. ET AL. Vectorial capacity of Culex pipiens fatigans in Ceylon. J. trop. Med. Hyg., 68: 254-256 (1965). 5. DETINOVA, T. S. Age-grouping methods in Diptera of medical importance. Geneva, World Health Organization, 1962 (Monograph Series, No. 47). 6. SAMARAWICKREMA, W. A. A study of the age-compo- sition of natural populations of Culex pipiensfatigans in relation to the transmission of filariasis due to Wuchereria bancrofti in Ceylon. Bull. Wld Hlth Org., 37: 117-137 (1967). 7. WIJETUNGE, IH. P. A. Filariasis in a university campus in Ceylon. J. trop. Med. Hyg., 70: 25-28 (1967).