Aedes aegypti (L.) and Aedes albopictus (Skuse) in Singapore City

Bull. Org. mond. Sante 11971, 44, 617-627Bull. Wld Hlth Org. J Aedes aegypti (L.) and Aedes albopictus (Skuse) in Singapore City 1. Distribution and Density * Y. C. CHAN,' K. L. CHAN 2 & B. C. HO 3 The distribution and density of Ae. aegypti and Ae. albopictus in Singapore were assessed from extensive larval surveys carried out from 1966 to 1968 to evaluate their respective roles in the epidemiology ofdengue haemorrhagicfever and to study their ecology in the urban areas. Ten urban areas where the majority of dengue haemorrhagic fever cases occurred were surveyed. The results showed that both species were common in the city, with Ae. aegypti being the dominant species. The distribution of Ae. aegypti was more uniform and related to the prevailing housing types and conditions. Its premise index was highest in slum houses, intermediate in shop houses, and lowest in multistorey flats. Ae. albopictus, on the other hand, did not seem to be related to the prevailing housing type in its distribution but tended to be more widespread in areas with open spaces. The larval density index (the average number of larvae per housing unit) was higher for Ae. aegypti than for Ae. albopictus, in agreement with the relative densities shown by their premise indices. The larval density index correlated well with the premise index and correlated best with the infested-receptacle index. For practical purposes, the most suitable, convenient, and reliable measure of density of Ae. aegypti population seems to be the infested-receptacle index. An attempt was made to estimate the rate of dispersal of Ae. aegypti from a stable population to an adjacent area of multistorey flats. The rate of dispersal, estimatedfrom the premise index and the larval density index, was approximately 2 % per year of the " donor " population. Outbreaks of dengue haemorrhagic fever have been occurring annually in Singapore since the dis- ease was first recognized in 1960 (Chan, Lim & Ho, 1967). Limited epidemiological observations during the 1960 outbreak incriminated Aedes aegypti as the primary vector of the disease. Adult mos- quitos of this species were found in more than 50% * This study was supported in part by a grant from the Rockefeller Foundation, New York, to the Department of Bacteriology, University of Singapore. Requests for reprints should be addressed to Dr Y. C. Chan, Department of Bacteriology, University of Singapore, Sepoy Lines, Singa- pore 3. 1 Lecturer, Department of Bacteriology, University of Singapore, Singapore. 2 Chief Vector Control Officer, Vector Control & Research Branch, Ministry of Health, Singapore. 3Lecturer, Department of Parasitology, University of Singapore, Singapore. of the houses in the urban areas and their distribu- tion was consistent with that of the cases of dengue haemorrhagic fever (Lim, Rudnick & Chan, 1961). Five strains of dengue type 2 virus were also iso- lated from Ae. aegypti and one strain of dengue type 2 virus from Aedes albopictus, collected during the outbreak (Rudnick & Chan, 1965). Although dengue virus was isolated from Ae. albopictus and the species was also common in the urban areas, it was not incriminated as a vector of dengue haemorrhagic fever because its distribution was not consistent with that of the disease, which was almost entirely urban (Lim, Rudnick & Chan, 1961; Rud- nick, 1966). Apart from the limited observations in 1960, only two other studies on Ae. aegypti and Ae. albopictus in the urban areas of Singapore have been reported. 2675 -617- 618 Y. C. CHAN, K. L. CHAN & B. C. HO Macdonald (1956), in a survey of Ae. aegypti in Singapore City, found that the premise index ranged from 9% to 37 %, depending upon the type of house surveyed. Yin-Coggrave & Pong (1964) reported an abundance of both Ae. aegypti and Ae. albopictus adults in their house collections in the urban areas. In view of the isolation of dengue virus from Ae. albopictus and the common presence of this species in the urban areas, further investigations of this species and of Ae. aegypti were carried out between 1966 and 1969. The objectives of the investigations were to study the ecology of the two mosquitos in the urban areas and to evaluate their respective roles in the epidemiology of dengue hae- morrhagic fever. The results of the study are reported in this and the following papers of the series. This paper describes the distribution and density of the two species. GENERAL DESCRIPTIONS OF SINGAPORE AND STUDY AREAS The Republic of Singapore is situated 137 km (85 miles) north of the equator, at latitude 1°15' N and longitude 1040 E, and lies at the extreme tip of the Malay Peninsula. The main island measures about 42 km (26 miles) from east to west and 22.5 km (14 miles) from north to south. There are a few small islands scattered around the north- eastern and southern shores of the main island and together they comprise the land area of Singapore, which is approximately 581.5 kmi (224.5 mi2). Sin- gapore City is the only city and it is situated at the southern tip of the main island and has an area of 96 km2 (37.2 mi2). The resident population in 1968 was estimated as 2 003 800, of which about 60% lived in the city. The densest population (about 34 250 persons per kM2, or 89 000 per mi2 is found in the 9 km2 (3.5 mi2) of the inner city on either side of the Singapore River. Half of the population live in degenerated slums in the central area, in attap houses1 in the rural area, or in squatter settlements on the urban fringe. A third of the population live in low-cost multistorey flats in housing estates, and the remaining population live mainly in shop houses. About 90% of the multistorey flats are 10-16 storeys high, and each block of flats is built on the basis of about 50 families or 250 persons to an acre of land (125 families or 620 persons per ha). Houses in the urban areas are supplied with 1 Houses with thatched roofs often made with palm leaves. piped water. The water is distributed from 5 service reservoirs situated on the island. The climate of Singapore is characterized by uni- form temperature, high humidity, and abundant rainfall. The average daily maximum temperatures for the three years 1966, 1967, and 1968 were 31.0°(C, 30.5°C, and 32.10C, respectively, while the average daily minimum temperatures were 23.7°C, 23.4°C, and 22.9°C, respectively. The average humidity is around 70%. The island generally has rainfall throughout the year, with the drier months averaging 16.5-17.8 cm (6.5-7.0 in) per month and the wetter months occurring during the monsoon season from November to January. The average monthly rainfall for the year is about 24.1 cm (9.5 in). Study areas Ten areas in the city were surveyed between September 1966 and May 1968 (Fig. 1). These areas included some of the most densely populated parts of the city, and represented the areas where the majority of dengue haemorrhagic fever cases have been reported since 1960. Pi The 10 areas are classified according to the most prevalent type of housing in the area as slum house, shop house, or flat areas. Slum houses and shop houses are very similar to those described by Mac- donald (1956). Slum houses were generally con- structed forj"temporary occupation but many of them have been occupied for more than 10 years. They usually have a galvanized iron or attap roof,' wooden walls, and a cement floor. They do not have individual piped water supplies but share a com- munal standpipe, and containers for water storage are generally plentiful, both inside and outside the houses. Slum-house areas are generally overcrowded, insanitary, and littered with discarded small con- tainers, such as tin cans and bottles. Many families in these areas also rear pigs and chickens. Shop houses usually have two or three storeys and adjoin one another. They are built of brick or cement and have tiled roofs. The entire shop house, or a por- tion of it, may be used for business or residence and each house is usually occupied by more than one family. Many shop houses have what is known as an " air well ", a section of the house not covered by a roof. A piped water supply is available in each house and also on each floor. The sanitary condi- tion of shop houses varies with area. Blocks of flats may have two or more storeys, and they are generally clean. They are used mainly AE. AEGYPTI AND AE. ALBOPICTUS IN SINGAPORE. 1 619 m Survey areas 1. Alexandra 2. Henderson ri ¢:' :x^^__ /3. Tiong Bahru CD,s 8 / 4. Cantonment CITY " 65. Chinatown 6. Rochor SINGAPORE 7. Geylang West 8. Macpherson 9. Geylang 10. Kaltang 8.' Kallong basin >>jS / ~~City area * Singapore R.r \ ~213a Mile 0 1 2 3 Fig. 1. Singapore city with the ten larval survey areas of Ae. aegypti and Ae. albopictus for residence but in some areas the ground floors Outdoor. That part of a house that is not under are occupied by shops. Blocks of flats are usually a roof but is within the confines of a housing separated from each other by an open grass field, unit, e.g., an " air well ", a garden, or a compound which is planted with shrubs and shade trees. outside a house. METHODS Open space. Any piece of land that is not part Definitions of a housing unit, such as vacant land between two For the purposes of this study, the following housing units or between two blocks of flats or a terms are used as defined: public park. Housing unit. The entire area of a house that bears one house number. A block of flats thus has Collection of larvae and pupae as many housing units as there are house numbers. Both larvae and pupae were collected in this study Indoor. Any part of a house that is under a roof. but, for simplicity, they are collectively referred to "Air wells ", although within a house, are not as larvae. Trained surveyors were used to search covered by a roof and are thus not considered as for larval breeding places in all the accessible houses indoor. and open spaces in an area. All the larvae from 620 Y. C. CHAN, K. L. CHAN & B. C. HO Table 1. Distribution of Ae. aegypti and Ae. albopictus in Singapore city Ae. aegypti Ae. albopictus Area Date surveyed Housing type unitssurveyed No uofuunitsurv No. ofhuig Premise No fhosn Premiseunits positive ine() units Positive ine %for larvae dx% for larvae dxv% Henderson Sep.-Nov. 1967 Slum house 460 136 29.6 73 15.9 Geylang West June 1967 Slum house 917 228 24.9 35 3.8 Geylang Sep.-Oct. 1966 Shop house 3 032 563 18.6 205 6.8 Chinatown Apr.-Dec. 1967 Shop house 8 422 1 343 15.9 68 0.8 Rochor Feb.-Mar. 1967 Shop house 6 026 894 14.8 173 2.9 Alexandra Aug.-Oct. 1967 Flat 2 120 201 9.5 62 2.9 Cantonment Jul.-Aug. 1967 Flat 1 326 90 6.8 4 0.3 Macpherson May 1968 Flat 363 23 6.3 47 12.9 Tiong Bahru Jun.-Aug. 1967 Flat 3 093 164 5.3 53 1.7 Kallang Sep.-Oct. 1966 Flat 4 828 109 2.3 50 1.0 Total 30 587 3 751 770 Mean 12.3 2.5 each breeding habitat, except those from large ones RESULTS such as jars, drums, and rubber tires, were collected into bottles with the aid of pipettes and ladles. In Distribution of Ae. aegypti and Ae. albopictus the case of large containers as many larvae as pos- A total of 30 587 housing units in 10 areas in the sible were collected in one or more samples of water. city centre and surrounding areas was surveyed for Each bottle was labelled with the area of collection breeding of Ae. aegypti and Ae. albopictus. Table 1 and the type and location of the habitat. Potential shows the premise indices of the two species. The breeding habitats-that is, habitats with water but premise index of Ae. aegypti ranged from 2.3 % to no larvae-were also recorded. Information was 29.6 %, with a mean index of 12.3% for the 10 areas. also obtained on the type of housing, the number The distribution of this species appears to be related of occupants in each housing unit, and other relevant to the type of housing in the area, with the highest factors. The larvae were taken to the laboratory for premise index in slum houses, an intermediate index identification and counting. A detailed description in shop houses, and the lowest index in flats. of the method of identification of larvae and pupae The mean premise indices for slum houses, shop of Ae. aegypti and Ae. albopictus will be published houses, and flats were 27.2%, 16.4%, and 5.0%, elsewhere. respectively (Table 2). Table 2. Distribution and density of Ae. aegypti in three types of housing Housing type J No. of housing Premise index(%) Larval density index a Range Mean Range Mean Slum house 1 377 24.9-29.6 27.2 2.93-8.11 5.52 Shop house 17 480 14.8-18.6 16.4 1.90-2.59 2.31 Flat 11 730 2.3- 9.5 5.0 0.20-1.86 0.81 a Average number of larvae per housing unit. AE. AEGYPTY AND AE. ALBOPICTUS IN SINGAPORE. 1 621 Table 3. Distribution of Ae. aegypti and Ae. albopictus in Chinatown Ae. aegypti Ae. albopictus Section Date surveyed housing No. of Premise No. of Premise units housing index housing index surveyed units positive (%) units positive (%) for larvae for larvae Tanjong Pagar Jul.-Aug. 1967 1 273 318 25.0 18 1.4 Havelock Apr.-May 1967 897 163 18.2 6 0.7 Anson Jun.-Jul. 1967 1 211 200 16.5 11 0.9 Hong Liml Apr.-May 1967 1 594 243 15.2 6 0.4 Telok Ayer Sep.-Oct. 1967 1 496 220 14.7 26 1.7 Hong Lim II Sep.-Dec. 1967 1 951 199 10.2 1 0.05 Ae. albopictus, in general, was less widely distri- species in an area is expressed as the average number buted than Ae. aegypti, except in the Macpherson of larvae per housing unit, which, for simplicity, is area where the premise index of Ae. albopictus was referred to as the larval density index. Table 4 higher than that of Ae. aegypti. The premise index shows the larval density indices of the two species. of Ae. albopictus ranged from 0.3 % to 15.9%, with Ae. aegypti had a higher density (mean larval a mean index of 2.5% (Table 1). In contrast to density index of 1.87) than Ae. albopictus (mean Ae. aegypti, the premise index of Ae. albopictus does larval density index of 0.26) in the surveyed areas, not seem to bear any relationship to the prevalent except in the Macpherson area where the density of type of housing in the area. The Henderson (slum Ae. albopictus was about 3 times that of Ae. aegypti. house) and Macpherson (flat) areas had the highest Like the premise index, the larval density index of premise indices of 15.9% and 12.9%, respectively, Ae. aegypti shows an apparent relationship with the while the Chinatown (shop house) and Cantonment type of housing prevalent in the area, but no such (flat) areas had the lowest indices of 0.8% and 0.3%, relationship is seen with the larval density index of respectively. Ae. albopictus. The larval density index of Ae. aegypti In any one area with a single type of housing, was highest in slum houses, intermediate in shop wide variations in the premise indices of Ae. aegypti houses, and lowest in flats (Table 2). and Ae. albopictus were also found. In the shop Ae. aegypti had a higher density indoors than house area of Chinatown, for instance, the premise outdoors, as shown by a higher larval density index indices in the different sections varied from 10.2% of the indoor population in 7 of 8 areas investigated to 25.0% for Ae. aegypti and 0.05% to 1.7% for (Table 5). In 4 of these areas, however, more than Ae. albopictus (Table 3). 20% of the population was found outdoors. The The distribution of the two Aedes species in larval density index of Ae. albopictus, on the other Geylang and Kallang is shown in Fig. 2, in which hand, was higher outdoors than indoors in 4 of the location of each breeding habitat is indicated. 8 areas surveyed (Table 5). The indoor population Both species occurred widely in the two areas but of this species was high in all areas, varying from a particularly high concentration of Ae. albopictus 24% to 100%. was found in the part of Geylang adjacent to Table 6 compares the number of larvae of the Kallang Basin. Kallang Basin was originally a tidal aede6pcies te n of arvae ouse* * ~~~~~~~twoAedes speies collected in and around h esswamp, and it became a large breeding ground of pesthis species as a result of the dumping of refuse and in open spaces, in three of the areas surveyed. including containers. A very small population, not exceeding 7% of thetotal population in the area, of Ae. aegypti was Density of Ae. aegypti and Ae. albopictus found in open spaces far removed from the imme- In this study, the population density of each diate vicinity of houses. The open-space population 622 Y. C. CHAN, K. L. CHAN & B. C. HO Ge lan 0 MILE 112 Kallang Fig. 2. Distribution of Ae. aegypti and Ae. albopictus in Geylang and Kallang. Each dot or cross represents one larval breeding habitat. Table 4. Density of Ae. aegypti and Ae. albopictus in houses in Singapore city ______________ - F No. of Ae. aegypti Ae. albopictus Area Housing type housing LarvangadKlngl| dot o repsto units Larval Larvaly Larval densitysuveed con index a count index a Henderson Slum house 460 3 729 8.11 1 819 3.95 Geylang West Slum house 917 2 683 2.93 296 0.32 Geylang Shop house 3 032 5 756 1.90 1 487 0.49 Chinatown Shop house 8 422 20 503 2.43 625 0.07 Rochor Shop house 6 026 15 617 2.59 1 339 0.22 Alexandra Flat 2 120 3 951 1.86 853 0.40 Cantonment Flat 1 326 815 0.61 12 0.01 Macpherson Flat 363 157 0.43 446 1.23 Tiong Bahru Flat 3 093 2 914 0.94 655 0.21 Kallang Flat 4 828 973 0.20 566 0.12 a Average number of larvae per housing unit. AE. AEGYPTI AND AE. ALBOPICTUS IN SINGAPORE. 1 623 Table 5. Density of indoor and outdoor Ae. aegypti and Ae. albopictus Ae. aegypti Ae. albopictus Area Larval count Larval density index a Larval count Larval density index a Indoor Outdoor Indoor Outdoor Indoor Outdoor Indoor Outdoor (%) (%) (%) (%) Henderson 2 264 1 465 4.92 ( 60.7) 3.19 (39.3) 433 1 386 0.94 ( 23.8) 3.01 (76.2) Geylang West 2 100 583 2.29 ( 78.2) 0.64 (21.8) 171 125 0.19 ( 59.4) 0.13 (40.6) Chinatown 15 191 5 312 1.80 ( 74.1) 0.63 (25.9) 272 353 0.03 ( 42.9) 0.04 (57.1) Rochor 13 310 2 307 2.21 ( 85.3) 0.38 (14.7) 688 651 0.11 ( 50.0) 0.11 (50.0) Alexandra 3 743 208 1.77 ( 95.2) 0.09 (4.8) 281 572 0.13 ( 32.5) 0.27 (67.5) Cantonment 815 0 0.61 (100.0) 0.00 ( 0-0) 12 0 0.01 (100.0) 0.00 ( 0.0) Macpherson 67 90 0.18 ( 41.9) 0.25 (58.1) 135 311 0.37 ( 30.1) 0.86 (69.9) Tiong Bahru 2 259 655 0.73 ( 77.7) 0.21 (22.3) 505 150 0.16 ( 76.2) 0.05 (23.8) a Average number of larvae per housing unit. Table 6. Density of Ae. aegypti and Ae. albopictus of Ae. albopictus, however, was much larger; in the in houses and in open spaces a Rochor area, nearly 80% of the total population occurred in open spaces. These large populations No. of Ae. aegypti No. of Ae. albopictus of Ae. albopictus in open spaces were due to the ____larvae _______larvae ___ Area H%in presence of dumps of rubber tires or tin cans. These House Open house House Opaen open types of container are two of the major breedingspace j____ spac space habitats of Ae. albopictus (Chan, Ho & Chan, 1971). Geylang 5 756 54 99.1 1 487 475 24.2 Replicate surveys of Ae. aegypti Rochor 15 617 1 137 93.2 1 339 5 121 79.3 A part of Chinatown was surveyed on two different Kallang 973 21 97.9 566 65 10.3 occasions, the first during March-April and the second during November-December of the same _______ -______- year. The premise index and the larval density index a All breeding habitats in open spaces were examined, except in of Ae. aegypti obtained in the two surveys are shown the case of large scrap dumps where only a random sample of in Table 7.10 containers from each dump was examined. Table 7. Replicate surveys of Ae. aegypti in Chinatown (Hong Lim I) No.of ousngNo. of housing Premise No. of Larval densitySurvey date Notsofsursien units positive index larvae index aunitsrveyed for larvae () a e Mar.-Apr. 1967 737 102 13.8 1 470 1.99 Nov.-Dec. 1967 1 452 196 13.5 2 526 1.74 a Average number of larvae per housing unit. 624 Y. C. CHAN, K. L. CHAN & B. C. HO Table 8. Natural rate of dispersal of Ae. aegypti Area I Population 1 Housing type Age of I Premise Larval densityArea status j_____________|houses (years) | index (%) index Geylang | donor | Shop house 50 18.6 1.90 Kallang recipient Flat 6 2.3 0.20 Estimated rate of dispersal = population in - recipient - area 100population in - donor - area x age of houses in - recipient - area 2.3 100 Rate based upon premise index = x6 -- = 2.06 % per year of ' donor ' population 0.20 100 Rate based upon larval density index = i- x-6 = 1.75 % per year of - donor - population Observation on the rate of dispersal of Ae. aegypti " recipient " area to the population of the "donor" area, expressed as a percentage per year of theThe distribution and density of Ae aegyptiein the population of the " donor" area. The rates ofGeylang and Kallang areas afford a unique oppor- dispersal of Ae. aegypti from the Geylang area to tunity for the estimation of the rate of dispersal of the Kallang area, as estimated from the premise this species. These two areas, although adjoining index and the larval density index, are, respectively, each other, are different historically and topogra- 2.06% and 1.75% per year of the population of the phically. The Geylang area comprises mainly shop Geylang (" donor ") area (Table 8). houses that are about 50 years old, and the Ae. aegypti The Henderson and Alexandra areas are two other population in this area is, presumably, a stable one. areas with a similar type of donor-recipient relation- In contrast, the multistorey flats in the Kallang ship. The Henderson area is a slum that is believed area are relatively new, having been built only to have existed for at least 50 years and the adjacent 6 years ago. The housing estate is built over a Alexandra area is about 14 years old. The Ae. aegypt. reclaimed swamp adjacent to an old airport where population in the flats of the Alexandra area is Aedes control had previously been carried out to believed to have originated from the stable popula- some extent. This area can thus be considered as tion in the Henderson area. When the premise having been Ae. aegypti-free prior to the develop- index and the larval density index of this species in ment of the estate. It is assumed that the present the Alexandra area were calculated on the basis of distribution and density of this species in the Kallang the rates of dispersal obtained from the Geylang- area is primarily the result of spread from the neigh- Kallang example, the estimated indices were not sig- bouring stable population in the Geylang area, since nificantly different from the actual indices (Table 9). the latter is the only heavily built-up area in close proximity to the Kallang area. Measurement of Ae. aegypti populations The rate of dispersal of a mosquito population is The premise index, the receptacle index, and the defined here as the ratio of the population of the infested-receptacle index have been used by various Table 9. Estimated indices of Ae. aegypti in Alexandra area Population - Age of Premise index (%) Larval density indexArea statusi Housing type houses (years) _statusousing tyj~ Actual Estimated a Actual Estimated b Henderson ' donor Slum house 50 29.6 8.11 Alexandra - recipient - Flat 14 9.5 8.7 1.86 1.96 a Based on a natural rate of dispersal of 2.06 % per year of ' donor - population. b Based on a natural rate of dispersal of 1.75 % per year of - donor - population. AE. AEGYPTI AND AE. ALBOPICTUS IN SINGAPORE. 1 625 workers to indicate the density of Ae. aegypti popula- tions (Tinker, 1967). The receptacle index is the per- centage of the inspected receptacles that are infested, and the infested-receptacle index is the average number of infested receptacles per house. These three indices and the larval density index calculated for the 10 areas are compared in Table 10. There Table 10. Comparison of various indices measuring Ae. aegypti population 1eLarval Premise Recep- Infested- Area density index tacle recep-index (% (%) a indexb Henderson 8.11 29.6 15.4 0.76 Geylang West 2.93 24.9 11.7 0.29 Rochor 2.59 14.8 15.6 0.24 Chinatown 2.43 15.9 17.7 0.22 Geylang 1.90 18.6 12.9 0.26 Alexandra 1.86 9.5 61.7 0.15 Tiong Bahru 0.94 5.3 10.3 0.09 Cantonment 0.61 6.8 10.0 0.07 Macpherson 0.43 6.3 5.2 0.07 Kallang 0.20 2.3 2.4 0.02 Correlation coefficient with larval density index 0.866 0.123 0.985 a Percentage of the inspected receptacles that were infested. b Average number of infested receptacles per housing unit. is good correlation between the larval density index and the premise index (correlation coefficient r= 0.866), and a very strong correlation between the larval density index and the infested-receptacle index (r=0.985), but no correlation between the larval density index and the receptacle index. DISCUSSION The Ae. aegypti premise index of slum houses, in the present study, was higher than that reported by Macdonald (1956) for similar types of house in Singapore, but much lower than the index he re- ported for slum houses in many towns in Malaya (Macdonald, 1956). Our mean Ae. aegypti premise index for slum houses was 27.2 %, in contrast to the figure of 9.0% that Macdonald reported for Singapore slum houses. However, our mean pre- mise index for shop houses (16.4%) was close to the 15.0% obtained by Macdonald. The premise index of multistorey flats has never been studied in the past because this type of housing has only recently been developed. The lowest mean premise index (5.0%) was found in flats. Our finding that the Ae. aegypti premise index was higher in slum houses than in shop houses was consistent with the data obtained by Macdonald in Malaya. The reasons advanced by Macdonald to explain his finding are equally applicable to the present study, as the types of houses and the habits of the people in the two countries are very similar. The general living conditions and the supply of piped water in slum areas are poor, and the need to store water has resulted in a larger accumulation of containers in and around houses. We have found that there were twice as many containers per house in slum areas as in shop-house areas.' The use to which houses are put is also an impor- tant factor affecting the local distribution of Ae. aegypti. Commercial houses, in general, have a lower premise index than residential houses even though they are of a similar type of housing. Hong Lim II in Chinatown, which had a premise index of 10.2%, is a highly commercialized area where living space is limited. The houses in this area are generally clean and have fewer unused containers. In contrast, Tanjong Pagar in Chinatown has more residential houses than commercial houses and the premise index of this area was 25.0%. Because of the large number of people living in this area there are more used and unused containers available for mosquito breeding. Our finding of the mean Ae. aegypti premise index for shop houses is of interest. It appears that there had been no significant change in the distribu- tion of this species in the shop-house area in Singa- pore since the survey by Macdonald (1956) 10-11 years ago. The population of Ae. aegypti has prob- ably reached an equilibrium between its natural rate of increase and the effects of human interference. This is possible, as the majority of shop houses are between 50 and 100 years old and human habits, in the absence of any drastic external influence, are not likely to have changed to any great extent. In order to compare Aedes populations between different areas and between different times of survey, we have introduced the larval density index, which 1 Unpublished data. 626 Y. C. CHAN, K. L. CHAN & B. C. HO is the average number of larvae per housing unit in tacle index, and infested-receptacle index. Tinker an area, as a measure of density. Wide differences (1967) has shown that the best index for measure- in Ae. aegypti density were found in the surveyed ment of this species is the premise index. Soper areas but there was general agreement between the (1967), however, pointed out that the premise index larval density index and the premise index. The does not adequately measure the density of Ae. highest density was found in slum houses and the aegypti, which is affected not only by the number lowest density in flats, with shop houses occupying and distribution of infested houses, but also by the an intermediate position. The indoor population of number and size of available breeding habitats. We Ae. aegypti was generally denser than the outdoor have found good correlation between the larval population but in many areas the outdoor popula- density index (which is based upon the absolute tion constituted more than 20% of the total popula- larval count) and the premise index, and the best tion in the area. This species was also found in correlation between the larval density index and the open spaces far removed from houses, although the infested-receptacle index. For practical purpose, population was small. the convenient and reliable infested-receptacle index Replicate surveys of Ae. aegypti in one area seems to be the most suitable measure of the density showed no difference in the premise index or in the of Ae. aegypti populations. larval density index. Subsequent studies, however, Our study of Ae. albopictus was the first systematic showed that the periods selected for the replicate survey of this species in the urban areas of Singapore. surveys coincided with peaks in the annual popula- The premise index and the larval density index were tion cycle of this species (Ho, Chan & Chan, 1971). also used to show the distribution and density, According to the studies carried out by Macdonald respectively, of this species. These indices are, how- in Malaya (Macdonald, 1956) the dispersal of Ae. ever, biased against Ae. albopictus, which was often aegypti is relatively slow. We have attempted to esti- found breeding in dumps of rubber tires and tin mate, from data obtained in this study, the rate of cans on vacant land, open fields, and parks. In some dispersal of this mosquito from a presumably sta- areas, the population of Ae. albopictus in open ble population to an adjacent multistorey flat area. spaces was larger than the population found in and The rate of dispersal, as estimated from the premise around houses. index and the larval density index, was approxi- Our surveys showed that Ae. albopictus was pres- mately 2% per year of the "donors" population. ent in the urban areas, although it was less widely This rate of dispersal may be applicable only to the distributed than Ae. aegypti. Large populations of spread of Ae. aegypti from a stable population into Ae. albopictus in houses were found in areas where multistorey flats. Such flats are characteristically there were open spaces, irrespective of the type of occupied by large numbers of people within a very housing prevalent in the area. In the Macpherson short period of time, and this is especially true in flat area, where the setting was more suburban than Singapore where the flat-dwellers, in most instances, urban, the premise index and the larval density have been transferred from demolished slum areas. index of Ae. albopictus were higher than those of It is not known how much the population increase Ae. aegypti. in the " recipient " area was the result of mechanical In all the areas surveyed, more than 200% of the transport of adult mosquitos or of larval stages, house-associated populations of Ae. albopictus was or both, by the people when they moved into the found indoors. In 4 of 8 areas surveyed, the indoor flats and how much of the population increase was population reached or exceeded 50% of the total due to natural increase of the species. population, showing that this species has become Several indices have been used by workers for mea- well-established in the domestic environment in the suing Ae. aegypti density, e.g., premise index, recep- urban areas of Singapore. ACKNOWLEDGEMENTS The authors thank the Deputy Director of Medical Singapore, for his help with the statistical analysis. They Services, Ministry of Health, Singapore, for permission also acknowledge the valuable assistance ofthe staff of the to publish this paper, and Mr C. Y. Tye of the Depart- Vector Control and Research Branch, especially Miss Tay ment of Social Medicine & Public Health, University of Siew Khuan, Mr Gan Sing Hai, and Mr Teo Boo Teck. AE. AEGYPTI AND AE. ALBOPICTUS IN SINGAPORE. 1 627 RESUMt AEDES AEGYPTI (L.) ET AEDES ALBOPICTUS (SKUSE) DANS LA VILLE DE SINGAPOUR: 1. REPARTITION ET DENSITE On a procede de 1966 a 1969 a une serie d'investiga- tions dans la ville de Singapour afin de determiner la repartition et la densite d'A. aegypti et d'A. albopictus. Ces enqu8tes, basees essentiellement sur la recolte de larves et de nymphes, devaient permettre d'etudier l'eco- logie de ces especes et d'appr6cier leur r6le respectif dans l'epidemiologie du syndrome dengue/fievre hemorragique. On a d6cele la presence de l'une et l'autre especes. A. aegypti etait le plus frequent, avec une repartition plus uniforme et en rapport avec le type predominant d'habitation. L'indice de positivite des locaux pour A. aegypti 6tait le plus elev6 (27,2%) dans les bas quar- tiers, intermediaire (16,4%) dans les magasins et le plus faible (5,0 %) dans les immeubles a plusieurs etages. Quant 'a A. albopictus, sa repartition ne semblait pas dependre du type d'habitation, mais on le trouvait avec une frequence accrue dans les secteurs comportant des espaces depourvus de constructions et specialement dans les dep6ts de vieux pneus et boites de conserve. L'indice de densite larvaire (nombre moyen de larves par unite d'habitation) etait plus eleve pour A. aegypti (indice moyen: 1,87) que pour A. albopictus (indice moyen: 0,26). On notait une correlation satisfaisante entre cet indice et l'indice de positivite des locaux. Il apparait n6anmoins que le meilleur moyen d'eva- luer la densite d'A. aegypti consiste a determiner le nombre moyen de recipients contenant des larves par maison. Des enquetes effectuees a deux epoques de l'annee n'ont montr6 aucune variation significative des indices de positivite des locaux et des indices de den- site larvaire. On s'est efforce par ailleurs d'evaluer le taux de disper- sion d'une population d'A. aegypti depuis longtemps etablie dans un secteur et presumee stable vers une zone de constructions plus recentes. D'apres les indices de positivite des locaux et de densite larvaire, il semble que ce taux soit annuellement de 2% environ de la population d'origine. REFERENCES Chan, K. L., Ho, B. C. & Chan, Y. C. (1971) Bull. Wld Hlth Org., 44, 629-633 Chan, Y. C., Lim, K. A. & Ho, B. C. (1967) Jap. J. med. Sci. Biol., 20, 81-88 Ho, B. C., Chan, K. L. & Chan, Y. C. (1971) Bull. Wld Hlth Org., 44, 635-641 Lim, K. A., Rudnick, A. & Chan, Y. C. (1961) Singapore med. J., 2, 158-161 Macdonald, W. W. (1956) Ann. trop. Med. Parasit., 50 385-398 Rudnick, A. (1966) Bull. Wld Hlth Org., 35, 77-78 Rudnick, A. & Chan, Y. C. (1965) 149, 638-639 Soper, F. L. (1967) Bull. Wld Hlth Org., 36, 536-538 Tinker, M. E. (1967) J. econ. Ent., 60, 634-637 Yin-Coggrave, M. & Pong, W. S. (1964) Ann. trop. Med. Parasit., 58, 359-366