Aedes larval bionomics and implications for dengue control in the paradigmatic Jaffna peninsula, northern Sri Lanka

The larval bionomics of Aedes across the Jaffna peninsula in northern Sri Lanka was investigated to obtain information needed for developing more effective larval source reduction measures to control endemic arboviral diseases. The habitats of preimaginal stages of Aedes mosquitoes were surveyed, and ovitrap collections were carried out in densely populated areas of the Jaffna peninsula. Aedes larval productivities were analysed against habitat characteristics, rainfall and dengue incidence. Adults emerging from collected larvae were tested for dengue virus (DENV). Only Aedes aegypti, Ae. albopictus and Ae. vittatus were identified in the field habitat collections and ovitraps. Aedes aegypti was the predominant species in both the field habitat and ovitrap collections, followed by Ae. albopictus and small numbers of Ae. vittatus. Tires and open drains were the preferred field habitats for Ae. aegypti, although larval productivity was higher in discarded plastic containers. The three Aedes species differed in field habitat preferences. Concomitant presence of the three Aedes species was observed in the field habitats and ovitraps. Larval productivities were inversely correlated with the salinity of the field habitat. Rainfall in the preceding month significantly correlated with larval productivity in the field habitats. DENV serotype 2 was detected in Ae. aegypti collected from ovitraps in the city of Jaffna. High Breteau, House and Container indices of 5.1, 5.1 and 7.9%, respectively, were observed in the field habitat surveys and ovitrap indices of up to 92% were found in Jaffna city. Aedes larval indices in populated areas of the peninsula showed a high potential for dengue epidemics. Unacceptable littering practices, failure to implement existing dengue control guidelines, vertical transmission of DENV in vector mosquitoes and preimaginal development in brackish water and open surface drains, as well as in domestic wells that provide potable water, are serious constraints to the current Aedes larval source reduction methods used to control dengue in the Jaffna peninsula. Similar shortcomings in arboviral disease control are likely present in other resource-constrained tropical coastal zones worldwide.


Introduction
Dengue virus (DENV) is estimated to cause 390 (95% confidence interval: 284-528) million infections worldwide every year, with 96 million of these (95% confidence interval: 67-136) accompanied by clinical symptoms [1]. Approximately 70% of dengue infections occur in Asia [1]. Aedes aegypti and Ae. albopictus are the primary and secondary vectors, respectively, of dengue and are also vectors of other important arboviral diseases, such as chikungunya, yellow fever, Rift Valley fever and Zika worldwide [1][2][3]. Dengue has recently spread from the tropics to the temperate zone as a result of range expansion of the two vectors and the development of a cold-resistant, diapausing egg stage in Ae. albopictus [1]. The prevalence of dengue vectors, and hence the rate of dengue transmission, is influenced by many factors, including climate, human population density, the availability of habitats for preimaginal development in the human-environment and the vector control measures in use [1,4,5].
The Jaffna peninsula is located in northern Sri Lanka and is separated from the south Indian state of Tamil Nadu by the 64-to 137-km-wide Palk Strait (Fig. 1). The Jaffna lagoon lies between the peninsula and mainland Sri Lanka. The peninsula has a limestone geology and a maximum altitude of 10.4 m a.s.l., and it contains many small lagoons and seawater inlets. All locations in the peninsula are < 10 km from the sea, and therefore the entire peninsula can be considered to be a coastal zone. The Jaffna administrative district, which includes most of the peninsula and islands in the Palk Strait, has a land area of 1100 km 2 and an average population density of approximately 700 persons/km 2 . Jaffna city is the largest and most populous urban center on the Jaffna peninsula, with an overall population density of 3048 persons/km 2 and is the administrative capital of Jaffna district. The Jaffna peninsula has a tropical climate and receives much of its annual rainfall of 60-190 cm during the northeast monsoon that typically prevails from October through to January, with a variable, lower rainfall during the southwest monsoon from April through to June. The annual average temperature is 31.5 °C, and average relative humidity is 74.4%, with the temperature slightly decreasing and relative humidity increasing more markedly during the monsoon. The peninsula is undergoing rapid salinization of its groundwater due to the incursion of sea water as a result of unsustainable rates of water extraction from freshwater aquifers and rising sea levels [4,6,7]. Dengue has been present in Sri Lanka since the beginning of the 20th century, with a high prevalence in the Jaffna district in recent years; there have been approximately 200 reported cases of dengue per 100,000 persons in the district in 2019 [8,9]. Chikungunya is the other main arboviral disease reported in the Jaffna district [10].
Features of the Jaffna peninsula make it a paradigmatic coastal zone for studying the impacts of population growth, anthropogenic environmental changes, global climate change and rising sea levels on mosquito-borne diseases [4,11]. The preimaginal development of Ae. aegypti and Ae. albopictus in brackish water (BW) and freshwater (FW) habitats has been documented in the peninsula [8,[12][13][14], with FW, BW and saline water being defined as containing < 0.5 ppt (parts per thousand), 0.5-30 ppt and > 30 ppt salt, respectively [12].
The larval bionomics of anopheline mosquitoes in Jaffna city was recently described in the context of controlling a potential resumption of malaria transmission [15]. In contrast, only a very limited number of studies have been performed on the bionomics of Aedes vectors in the Jaffna peninsula [16,17]. We therefore investigated the preimaginal habitats of Aedes species in areas of varying land use within the densely populated city of Jaffna and surrounding areas in 2018 and 2019 in relation to anthropogenic environmental factors, habitat characteristics, rainfall and dengue incidence. a Location of Sri Lanka in relation to India, b location of Jaffna peninsula in northern Sri Lanka, c larval field collection locations in Nallur, Uduvil and Jaffna city (Gurunagar and Navanthurai wards in the city). Ovitrap collections were also conducted at several sites within Jaffna city Ovitrap collections of larvae were also carried out at several sites within the city of Jaffna. We additionally investigated the presence of DENV in adult mosquitoes that emerged from the collected Aedes larvae.

Field collection of larvae and larval indices
Surveys for Aedes larvae and larval collections were performed every 2 weeks from January 2018 to December 2019. Residential premises and a relatively smaller number of public and commercial premises and urban farms (for details, see Additional file 1) were randomly selected and then used throughout the study. Each location was inspected for water collection sites (termed habitats) suitable for the development of Aedes larvae (Fig. 2). Preimaginal stages were collected from habitats using a standard procedure consisting of ten dips with 350-ml dippers for water containers with wide openings (e.g. for drains and water storage tanks), pipetting with 5-ml plastic pipettes for smaller containers and ten dips with string-connected conical drop nets (diameter: 15 cm, depth: 10 cm) for wells, as described previously [15]. Sampling was performed in all selected potential habitats regardless of the presence or absence of preimaginal stages. Collected samples were linked to the habitat type and GPS location. The salinity of the habitat water was measured with a hand-held refractosalinometer (Atago Co. Ltd., Tokyo, Japan). Collected preimaginal stages were brought to the laboratory of the Department of Zoology, University of Jaffna where pupae and dead preimaginal stages were discarded and live larvae were maintained under contained insectary conditions as previously described [16]. Adult mosquitoes emerging from collected larvae were identified to the species level with a standard key [18]. Larval productivity per month of each Aedes species in field habitats was determined as the cumulative number of adults emerging from the larvae collected each month from all the field habitats.
The Container Index (CI), House Index (HI) and Breteau Index (BI) for Aedes larvae were determined as recommended by the World Health Organization [19]. The preferences of Aedes species collectively and individually for the different types of habitats surveyed was assessed by calculating the habitat preference ratio (HPR). The overall HPR for all Aedes species was defined as the ratio of the proportion of each type of Aedes larva-positive habitat in all types of Aedes larva-positive habitats to the proportion of the same type of habitat among all types of surveyed habitats [20]. The HPR for individual Aedes species was calculated as the ratio of the proportion of each type of that Aedes species larvapositive habitat in all types of that Aedes species larvapositive habitats to the proportion of the same type of habitat among all types of surveyed habitats. The HPR indicates the preference for a particular type of habitat when varying numbers of different types of habitats are surveyed [20].
Associations between any two Aedes species in the field habitats were determined by 2 × 2 contingency tables for pairwise comparison [21].

Ovitrap collections
Conventional black plastic ovitraps (capacity: 650 ml, radius: 4.5 cm, height: 10 cm) containing 300 ml of water obtained from the nearest domestic water supply (i.e. well or tap) with a 2 × 15-cm plywood paddle resting against the inside upper rim, as shown in Fig. 3(1a), were utilized, as previously described [22]. The salinity of the habitat was measured with a hand-held refracto-salinometer (Atago Co. Ltd.). Forty outdoor ovitraps were placed at different locations in Jaffna city, with a minimum of distance of 15 m between ovitraps [ Fig. 3(1b-f )], in areas of varying land use. Specifically, 20 ovitraps were located in built-up areas of the city, and the remaining 20 ovitraps were placed in the relatively limited number of areas in the city available for other types of land use, with nine ovitraps placed in selected sites in farmlands, five near water bodies, five on barren land and one on grassland [ Fig. 3(2)].
Weekly collections were carried out from March 2019 to December 2019. After each collection of eggs and larvae at each site, ovitraps with a replacement paddle were filled with an equal amount of water from the same source as the preceding ovitrap for the next round of collection. Eggs and larvae from the ovitraps were brought to the laboratory and reared to adulthood. Emerging adults were identified to the species level [18]. The monthly ovitrap index (OI) for each of the 40 ovitraps was calculated as the percentage of the ratio of the number of times an ovitrap was positive for Aedes in the specific location to number of times the ovitrap was surveyed at the specific location. The average OI over the 9-month collection period was used for geospatial interpolation. Larval productivity per month of each Aedes species in the ovitraps was determined as the cumulative number of larvae produced each month from all the ovitraps.

Statistical analysis of relationships between Aedes larval productivity, habitat characteristics, rainfall, environmental factors and dengue incidence
Monthly rainfall data for the Jaffna district were obtained from the Government Meteorological Department. Monthly dengue incidence reported for the Jaffna district was obtained from the Epidemiology Unit of the Ministry of Health.
The Pearson correlation coefficient analysis was performed to test relationships between rainfall, habitat salinity and dengue incidence with Aedes larval productivity in the different field habitats. This analysis was also performed with the exclusion of salinity for ovitrap collections. To determine the dominant Aedes species collected in field habitats and ovitraps, separate estimations were done using the generalized linear model followed by least-squares means. Statistical analyses were performed using the SAS University Edition (SAS Institute Inc., Cary, NC, USA). The threshold for statistical significance was set at P < 0.05.

DENV nonstructural protein 1 antigen assay and serotyping
Adult Aedes mosquitoes emerging from larvae collected in field habitats and ovitraps, and subsequently identified to the species level, were screened for DENV, as described previously [8], in pools of 30-40 mosquitoes of each species. Aedes mosquitoes identified from field surveys were pooled based on location and type of habitat, and those from ovitrap surveys were pooled based on the different land-use environment. Each pool of mosquitoes was triturated in a 1.5-ml microfuge tube with a plastic disposable pestle for extracting soluble antigens in 300 µl of 0.01 M phosphate buffered saline (PBS) containing 1% v/v Triton X-100. The extract was then centrifuged at 10,000 g for 5 min, and approximately 30 µl of supernatant tested for the DENV nonstructural protein 1 antigen (NS1Ag) using the Biocredit Dengue NS1Ag antigen test (RapiGen Inc., Gyeonggi-do, Republic of Korea).
Of the remaining supernatant from DENV NS1 antigen-positive extracts, approximately 140 µl was used for viral RNA extraction and DENV serotyping, as previously described [23]. Essentially, viral RNA was extracted using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany) and transcribed to cDNA with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's protocol, in 20-µl reaction mixtures containing 1 × RT buffer, 1 × RT random primers, 10 × dNTPMix (100 mM), 50 U of MultiScribe ™ Reverse Transcriptase, 20 U of RNase Inhibitor (Applied Biosystems), 10 µl of RNA and PCR-grade water (Applied Biosystems). The qPCR assay was performed using the TaqMan ® Multiplex Master Mix (Applied Biosystems). The reaction mixture (total volume: 20 μl) contained: 1 × TaqMan multiplex master mix (containing Mustag Purple dye), 900 nM of each primer, 250 nM of each probe, 2 μl of cDNA and PCR-grade water (Applied Biosystems). Following an initial denaturation at 95 °C/20 s, the reaction was carried out at 95 °C/3 s and 60 °C/30 s for 40 cycles. The threshold cycle value (Ct) for each reaction was set manually. All assays were done in triplicate. After the primers and probes were validated, a multiplex method was optimized to quantify the four serotypes in a single reaction.

Aedes larvae in field habitats
A total of 3719 Aedes larvae of three Aedes species were collected from the different types of field habitats illustrated in Fig. 4. During the rainy season 1117 Ae. aegypti, 540 Ae. albopictus and 254 Ae. vittatus were collected, and during the dry season 1164 Ae. aegypti, 623 Ae. albopictus and 21 Ae. vittatus were collected; the monthly productivities are shown in Fig. 5a, b. Larval productivities were significantly different for each Aedes species throughout the study (F (2, 33) = 21.52, R 2 = 96.6, P < 0.001), with Ae. aegypti being the most abundant species with a mean (± standard deviation) monthly larval productivity of 97 ± 25.
Of a total of 2543 potential habitats examined during the field surveys, 202 were found to contain Aedes larvae. In these 202 habitats, the sole presence of either Ae. aegypti, Ae. albopictus or Ae. vittaus was observed in 133, 13 and three habitats respectively. The presence of larvae of more than one species was found in 53 habitats. A total of 2281, 1163 and 275 Ae. aegypti, Ae. albopictus and Ae. vittaus, respectively, were identified in several different types of habitats (Table 1). Discarded waste containers, particularly those made of plastic, were the most productive in terms of the number of Aedes larvae, accounting for 68% of all larval collections during the 2 years of survey. Aedes aegypti was the only species collected in open drains, for which it had the highest HPR of 2.5. Both Ae. aegypti and Ae. albopictus showed high HPRs of 2.2 and 6.3, respectively, for discarded tires. Although relatively fewer Ae. vittatus were collected, it showed the highest HPR of 20.1 for water puddles and was the only species collected in water puddles; both the BI and HI were 5.1% and the CI was 7.9% in the surveyed areas.
Aedes aegypti was collected throughout the year, Ae. albopictus was collected more variably throughout the year and Ae. vittatus was collected mainly during the rainy season (Fig. 5a, b).

Presence of larvae of different Aedes species in the same field habitat
Aedes aegypti was the most abundant species in the field collections, but the three species were sometimes present in the same or nearby habitats in the four areas studied (Additional file 1). Aedes aegypti and Ae. albopictus were found together in wells, cement water tanks, discarded plastic containers, flower pots/bases and tires; Ae. aegypti and Ae. vittatus were found together in discarded plastic containers; Ae. albopictus and Ae. vittatus were found together in flowerpots/bases. The concomitant presence of Ae. aegypti and Ae. albopictus was recorded in 25 and 20 habitats during the dry and rainy seasons, respectively. Larvae of all three species were found together in a single discarded plastic container during the wet season; however, Chi-square analysis for the concomitant presence of different species in the habitats (Additional file 2) suggested that Ae. aegypti and Ae. albopictus ( χ 2 (1,202) = 47.69, df = 1, P < 0.01) as well as Ae. aegypti and Ae. vittatus ( χ 2 (1,202) = 45.39, df = 1, P < 0.01) demonstrated significant development in different habitats. There was no statistical evidence to support the separate development of Ae. albopictus and Ae. vittatus.

Aedes larval productivity in field habitats in relation to habitat salinity
The salinity in the habitats where Aedes larvae developed ranged between 0 ppt salt and 8 ppt salt (Table 1). Larval productivities of all three Aedes species combined (r (24) = − 0.69, P = 0.0002) and Ae. aegypti alone (r (21) = − 0.79, P < 0.0001) were significantly and negatively correlated with increasing salinity in the different habitats (Additional file 3).

Aedes larval productivity in field habitats in relation to rainfall
The monthly larval productivities of Ae. aegypti, Ae. albopictus and Ae. vittatus alone or all three species collectively were not significantly related to rainfall during the same month (r (24) = 0.11-0.43, P = 0.1-0.58; Additional file 3). However, as can be seen in Fig. 5c, there was a delay in the increase in larval productivity following the onset of rains. When rainfall in the preceding 1-month period was analysed in relation to the monthly larval productivity, only Ae. aegypti larval productivity showed a significant positive correlation with the preceding month's rainfall (r (23) = 0.42, P < 0.05).

Aedes larval productivity in field habitats in relation to dengue incidence
As shown in Fig. 5b, there was a delay between the increase in larval productivities of Aedes in field habitats and the increase in dengue incidence. Statistical analysis of the larval productivities with the monthly incidence of dengue in the Jaffna district is shown in Additional file 4. The monthly larval productivity of Ae. aegypti showed a tendency to correlate with monthly dengue incidence (r (24) = 0.38, P = 0.06), while that of Ae. vittaus showed a significant positive association with monthly dengue Fig. 4 Examples of field habitats containing Aedes larvae. a Disposable plastic cups, b open surface drain, c discarded tire, d discarded coconut shell, e flower pot, f well, g discarded metal containers, h discarded ceramic containers, i water puddle incidence (r (24) = 0.48, P = 0.016). Larval productivities of all three Aedes species collectively tended to approach significance for an association with dengue incidence (r (24) = 0.38, P = 0.06). However, there were significant positive associations between the larval productivities of Ae. aegypti (r (20) = 0.52, P = 0.01) and collectively for all three Aedes species (r (20) = 0.59, P = 0.005) when larval productivities for the preceding 2 months were   separately tested against dengue incidence in any given month.

Aedes larval productivities in ovitraps
Larval productivities in the ovitrap collections were significantly different between the three species, with Ae. aegypti being the dominant species (F (2,27) = 51.05, P < 0.001). Different species could be collected from the same or relatively proximal ovitraps (Additional file 5). The OIs for individual ovitraps varied from 5 to 92% in built-up areas, from 0 to 13% in farmlands and from 0 to 33% near water bodies; they were 0% on grasslands and barren land [ Fig. 3(2)].
Of the 40 ovitraps, the salinity of the water in six ovitraps was in the range of 1-3 ppt salt; all six of these ovitraps utilized water drawn from the nearest wells. All three species laid eggs and developed into preimaginal stages in a single ovitrap containing water with a salinity of 1 ppt salt. Aedes aegypti and Ae. albopictus laid eggs and developed into preimaginal stages in five ovitraps containing water with 1-3 ppt salt BW. Rainfall did not significantly influence the larval productivities of Ae aegypti, Ae. albopictus and collectively all larvae of all three species (r (10) = 0.20-0.49, P = 0.14-0.57), with the exception of Ae. vittatus (r (10) = 0.74, P = 0.01), which was only collected from ovitraps during the rainy season (Fig. 6).

DENV NS1Ag test and DENV serotyping
Forty pools representing emergent adults of Ae. aegypti (21 pools), Ae. albopictus (15 pools) and Ae. vittatus (4 pools) collected from field habitats (8, 6 and 3 pools, respectively) and ovitraps (13, 9 and 1 pools, respectively) were separately tested for DENV NS1 antigen. A single pool of ovitrap-derived Ae. aegypti was found to be positive in the NS1Ag test. This pool of Ae. aegypti was collected in the month of October 2019 during the monsoon season, from a built-up area with the highest observed ovitrap index of 92% (Fig. 3(2)). It was shown to contain DENV-serotype 2 by reverse transcription-qPCR.

Discussion
A previous survey of peridomestic mosquitoes during 1986 and 1987 detected only Ae. aegypti, Ae. albopictus and Ae. novalbopictus in Jaffna city [24]. Since 2011, however, Ae. aegypti and Ae. albopictus larvae have been collected from FW and BW field microhabitats of discarded food and beverage containers on beaches at concentrations of up to 15 and 14 ppt salt, respectively [11,13], and in the water of domestic wells at up to 9 ppt salt located in the coastal areas of Jaffna peninsula [12]. Aedes vittatus is present in parts of mainland Sri Lanka [25,26], including its largest city Colombo [27]. Aedes vittatus has been regarded as a sylvatic or peridomestic mosquito in rural environments, developing mainly in rock-holes [21,  28,29]. Taken together, our data from Jaffna city and previous data from Colombo [29] suggest that Ae. vittatus has now adapted to develop in densely populated urban areas, showing a preference for water puddles over discarded waste containers in the limited number of collections carried out to date. Additional investigations to determine whether Ae. vittatus now demonstrates greater anthropophagic and anthropophilic behaviour in the Jaffna peninsula are justified. Aedes vittatus has been reported to be a competent vector of DENV, yellow fever virus and chikungunya virus [29,30]. Our observation of a significant association between Ae. vittatus larval productivity and dengue incidence in the Jaffna peninsula suggests that further investigations are needed on Ae. vittatus as a potential vector of DENV in Sri Lanka.
Our observations show that the Aedes vectors are able to oviposit and undergo preimaginal development in a large variety of habitats in urban and semi-urban areas across the Jaffna peninsula. The habitats were associated with (i) domestic water storage and supply (i.e. open wells and cement water tanks); (ii) drainage (i.e. open surface drains); (iii) man-made litter (i.e. plastic, metal, glass/ ceramic, rubber containers, coconut shells); (iv) rain-fed water puddles; (v) flower pots and their bases. A recent relevant study of urban farms in Miami-Dade County, Florida, USA, reported that storm drains, phytotelmata, discarded tyres and garbage cans contained significant numbers of Ae. aegypti [31]. Phytotelmata were not investigated in our study due to the limited availability of personnel for the study, but they are also likely to serve as habitats for Aedes in the Jaffna peninsula. High larval productivities were recorded in collections from the numerous discarded plastic containers. The data suggest varying preference for different types of habitats among the three Aedes species that warrants more detailed investigation than was possible in our study. Our findings suggest that Aedes vectors in the Jaffna peninsula are evolving to adapt to anthropogenic changes in the environment [5,7]. In the Jaffna peninsula, Ae. aegypti, the urban vector of dengue, has become the dominant Aedes species in Jaffna city as well as in suburban Nallur and semi-rural Uduvil. Similarly, Ae. albopictus, typically regarded to be a peri-urban vector of dengue, has expanded its range to built-up areas of Jaffna city. Urban farms have been reported to provide favourable environments for Aedes mosquitoes in Florida, USA [31,32], and this also appears to be the case in the Jaffna peninsula. A previous GIS-based study showed that the built-up areas of Jaffna city have a high local incidence of dengue [33], which is consistent with the present findings that show high ovitrap indices in the city. These factors have to be taken into consideration in developing appropriate vector control measures in the peninsula.
Our findings show that Ae. aegypti, Ae. albopictus and Ae. vittatus can be collected from field habitats and ovitraps, either the same ovitrap or ovitraps that are in relative proximity to each other, suggesting an overlap in the distribution of blood-fed female Aedes species in the studied areas. Interspecific larval competition between Ae. aegypti and Ae. albopictus and associated competitive population displacement reported in the field [34,35] does not appear to be a significant factor at the locations studied in the Jaffna peninsula. The preimaginal stages of Ae. aegypti and Ae. albopictus are known to co-exist in the same water collection although one species may predominate [36][37][38]. However, our findings suggest that Ae. aegypti and Ae. albopictus show variations in preferences for different field habitats. We observed that Ae. vittatus showed a more marked temporal variation in larval abundance than either Ae. aegypti and Ae. albopictus in being collected mainly during the rainy season in both the field habitats and ovitraps. This temporal variation may be related to a lower density of Ae. vittatus in the peninsula that is particularly dependent on rainfall to expand populations, but more detailed investigations are needed to establish this.
The preimaginal habitats in our present study contained FW or BW of up to 8 ppt salt in field habitats and ovitraps, consistent with our previous finding of salinity-tolerant Ae. aegypti and Ae. albopictus in the Jaffna peninsula [12][13][14]. Anophelines prevalent in Sri Lanka have also been shown to develop in BW habitats in the Jaffna peninsula and along nearby coastal mainland areas, including the well-known malaria vectors Anopheles culicifacies [14,39] and An. stephensi [15]. The entire human population of the peninsula relies on groundwater extracted from the aquifers to meet all water needs (e.g. drinking water, and water for domestic, agricultural and industrial needs), but this water is becoming increasingly brackish [4,6,7,[40][41][42][43]. The water in many domestic wells in the coastal areas nearby Jaffna city are brackish and yet support the preimaginal development of Ae. aegypti and Ae. albopictus [13,14]. Heritable changes contribute to the greater salinity tolerance of Ae. aegypti in BW habitats in the peninsula [44,45]. However, the inverse relationship between Aedes larval productivities and habitat salinity in field habitats suggests that the three Aedes species in the Jaffna peninsula retain a preference to oviposit in habitats of lower salinity. The tendency of BW-derived laboratory colonies of Ae. aegypti in the peninsula to prefer to oviposit in BW rather than FW in the laboratory [44] has therefore not yet spread widely into the wild populations of Ae. aegypti in the peninsula. However, the present findings reinforce our previous recommendations [7,44] that dengue vector control measures in the Jaffna peninsula, which are currently only directed towards FW habitats, need to be extended to BW habitats.
The vertical or transovarial transmission of DENV to eggs, through larvae and then to emergent F1 generation adults in Ae. aegypti [46][47][48] and Ae. albopictus [49] is well documented. We previously observed extensive vertical transmission of DENV from infected bloodfed female Ae. aegypti and Ae. albopictus to larvae and adults of the F1 generation in the laboratory [8]. Vertical transmission may supplement more direct transmission of DENV through blood-feeding alone, but its role in dengue epidemiology needs to be more firmly established. Vertical transmission in dengue vectors that are able to develop in BW, which is not presently targeted in vector control programmes, may provide a reservoir of infected vectors that can maintain a basal level of DENV persistence in coastal areas during the dry season. This has been postulated to facilitate the increased transmission that occurs with the onset of the monsoon due to the rapid proliferation of vectors developing in FW in the Jaffna peninsula, a phenomenon that may be shared with similar locations worldwide [8].
The reduction of dengue vector larval sources in the Jaffna peninsula is an ongoing activity that is enforced with penalties by the Ministry of Health. Our findings from the field surveys showed a BI and HI of > 5% and a CI of 7.9%, with an OI of up to 92% in Jaffna city. These indices show that the current management approach has been ineffective in the Jaffna peninsula and also that there is a very high risk of dengue epidemics [50,51], emphasizing the urgent need for more effective preventive measures. The present findings provide important information that is necessary for improving larval source reduction measures by: (i) identifying areas, habitats and times where and when these measures are likely to be most effective; (ii) pinpointing the need for improving waste management; (iii) highlighting the importance of extending control measures to BW habitats, surface drains and domestic wells; (iv) re-emphasizing the need to replace open surface drains with underground drains [16]; (v) reinforcing the need for using larvivorous fish in both BW and FW domestic wells that have previously been shown to effectively eliminate mosquito larvae in the Jaffna peninsula [15,52,53]; (vi) stressing the need for installing mosquito-proof coverings for domestic wells and water storage tanks; (vii) underscoring the need for improving surveillance and research on both adult and preimaginal stages of Aedes vectors.

Conclusion
The two established arboviral vectors, Ae. aegypti and Ae. albopictus, and a potential vector, Ae. vittatus, develop in many different types of discarded containers near dwelling places, open surface drains and domestic wells, yielding high larval indices in populated areas of the Jaffna peninsula. The vertical (transovarial) transmission of DENV, salinity tolerance, preimaginal development in open surface drains, domestic wells that provide potable water and litter in the humanenvironment are particular challenges that need to be addressed. The findings, which may serve as a paradigm for many other tropical coastal locations, identify the urgent need to improve Aedes larval source reduction practices in order to control dengue and chikungunya in the Jaffna peninsula.