Aedes species in treeholes and fruit husks between dry and wet seasons in southeastern Senegal

Authors


ABSTRACT

During the dry season in February, 2010 and the wet season in September, 2011 we sampled mosquito larvae and eggs from treeholes of seven native hardwood species and the husks of Saba senegalensis in 18 sites in the PK-10 forest in southeastern Senegal. Larvae were reared to adults for species identification. In the dry season, we recovered 408 Aedes mosquitoes belonging to seven species. Aedes aegypti s.l. comprised 42.4% of the collection, followed by Ae. unilineatus (39%). In contrast to reports from East Africa, both Ae. aegypti aegypti and Ae. aegypti formosus were recovered, suggesting that both subspecies survive the dry season in natural larval habitats in West Africa. In the wet season, 455 mosquitoes were collected but 310 (68.1%) were the facultatively predaceous mosquito Eretmapodites chrysogaster. The remaining 145 mosquitoes consisted of ten Aedes species. Aedes aegypti s.l. comprised 55.1% of these, followed by Ae. apicoargenteus (15.2%) and Ae. cozi (11.7%). Similar to East Africa, most (90%) of Ae. aegypti s.l. in the wet season were subspecies formosus.

INTRODUCTION

The mosquito Aedes aegypti (Linneus) is the major vector of yellow fever (YF) and dengue fever (DENV1–4) flaviviruses (Gubler 2012), and chikungunya alphavirus (CHIK) (Dupont-Rouzeyrol et al. 2012) throughout most tropical and subtropical regions of the world. The species is taxonomically subdivided on the basis of scaling patterns on the abdominal tergites into three subspecies, two of which occur in Africa. The initial need for designation of Ae. aegypti s.l. subspecies in Africa arose from observations made in East Africa in the late 1950s that the frequency of pale “forms” of Ae. aegypti was higher in populations in and around human dwellings than in those of the nearby forest (McClelland 1960). The implied correlation between color and behavior prompted a revision of the biology and taxonomy of Ae. aegypti (Mattingly 1957). He described formosus (Walker) as a subspecies of Ae. aegypti that was restricted to sub-Saharan Africa and in West Africa “is the only form known to occur except in coastal districts and in one or two areas of limited island penetration” (Mattingly 1957, p. 395). He also suggested that it most frequently breeds in natural containers such as tree holes and feeds on wild animals. Mattingly also stated that in addition to the dark-scaled parts of the body being generally blacker, “ssp. formosus never has any scales on the first abdominal tergite (p. 395).” The type form of Ae. aegypti aegypti was alternatively defined as “either distinctly paler and browner (at least in the female) than ssp. formosus or with pale scaling on the first abdominal tergite or both (p. 395).” These two forms are hereafter referred to as Aaa and Aaf.

Mattingly (1957) suggested that Aaa breeds in artificial containers provided by humans, will breed indoors, and has a preference for feeding on human blood. A comprehensive study was subsequently made of differences in scale patterns on the abdominal dorsum in 74 Ae. aegypti s.l. collections from 69 different worldwide locations (McClelland 1974). A 15-point scale running from F (formosus-like = no scales on the first abdominal tergite) to Q (queenslandensis-like = entire abdominal dorsum covered with scales) was used to categorize Ae. aegypti s.l. from these many collections. McClelland (1974) concluded that many of Mattingly's distinctions between subspecies were not always clear-cut. For example, collections from North America (east coast of Mexico, United States, and various Caribbean Islands) contained Ae. aegypti lacking any scales on the first abdominal tergite (McClelland 1974). Both Aaa and Aaf were collected throughout Africa but forms with minimal scaling tended to occur in coastal regions.

In a recent study carried out in Senegal, Sylla et al. 2009 demonstrated that there was a northwest–southeast cline in the abundance of Aaf, with Aaa occurring exclusively in the northwest and Aaf exclusively in the southeast. The southern margins of the arid Acacia-Savannah habitat run from east to west through central Senegal. Along this margin, mixed Aaa and Aaf collections were observed. Paupy et al. 2010 examined one such mixed site in and around Niakhar, a town 115 km southeast of Dakar, Senegal. They examined the abundance of subspecies in adults raised from larvae collected in domestic larval habitats during the dry season and from domestic, peridomestic, and natural containers during the wet season. They noted a significant seasonal shift towards female mosquitoes with higher McClelland scores (greater scaling) during the wet season. Paupy et al. 2010 also performed an analysis of genetic differentiation at eight microsatellite loci. Allele frequencies were then compared between sexes, among the two subspecies, and among collections in the dry vs rainy seasons. In agreement with earlier studies (Huber et al. 2008, Sylla et al. 2009), they found no differences between sexes nor among the two subspecies. However, a significant difference was detected between mosquitoes collected in the dry vs rainy seasons.

There are two deficiencies in the existing literature on seasonal shifts in subspecies abundance in West Africa. Sylla et al. 2009 only made collections during the wet season while Paupy et al. 2010 did not examine subspecies abundance in natural containers during the dry season. To partially address these deficiencies, we have tested for seasonal shifts in subspecies abundance in treeholes and fruit husks in a natural forested area. During the dry season in February, 2010 and in the wet season in September, 2011, we sampled mosquito larvae and eggs from treeholes of seven native hardwood species and the husks of Saba senegalensis in 18 collection sites in the PK-10 forest near the town of Kedougou in southeastern Senegal, 610 km southeast of Dakar. We also found many other Aedes species in abundance in treeholes and fruit husks in PK-10 and these are also reported here.

MATERIALS AND METHODS

From February 15–19, in 2010, treeholes were excavated and flooded at 14 locations in the PK-10 forest gallery. The locations of all sites were recorded with a GPSMAP® 62s (Garmin Inc. Wichita KS) and coordinates are listed in Table 1 and mapped in Figure 1. All native trees were identified to species using Berhaut (1967).

Table 1. Location of the 18 collection sites in PK10 and the tree species from which mosquito larvae were collected. Approximately 50 husks were collected per site at PK1003, 11, 12, and 18. The locations of collecting towers are indicated for reference with earlier literature.
SiteTree speciesLatitudeLongitude
PK10–01Adansonia digitata12°36’45.11”N12°14’51.24”W
PK10–02Diospyros mespiliformis12°36’45.11”N12°14’51.24”W
PK10–03Saba senegalensis husks12°36’45.11”N12°14’51.24”W
PK10–04Anogeissus leiocarpus12°36’41.66”N12°14’49.35”W
PK10–05Cola nitida12°36’41.24”N12°14’49.46”W
PK10–06Combretum glutinosum12°36’41.10”N12°14’50.15”W
PK10–07A. leiocarpus12°36’41.10”N12°14’50.15”W
PK10–08A. leiocarpus12°36’41.10”N12°14’50.15”W
PK10–09A. leiocarpus12°36’40.59”N12°14’49.84”W
PK10–10A. leiocarpus12°36’40.80”N12°14’49.54”W
PK10–11S. senegalensis husks12°36’40.80”N12°14’49.54”W
PK10–12S. senegalensis husks12°36’39.99”N12°14’49.57”W
PK10–13D. mespiliformis12°36’38.63”N12°14’50.44”W
PK10–14A. leiocarpus12°36’38.63”N12°14’50.44”W
PK10–15Parkia biglobosa12°36’36.09”N12°14’47.49”W
PK10–16C. nitida12°36’31.68”N12°14’45.17”W
PK10–17A. leiocarpus12°36’37.88”N12°14’45.16”W
PK10–18S. senegalensis husks12°36’41.65”N12°14’46.16”W
Maginot Tower 12°36’39.73”N12°14’50.39”W
Station Tower 12°36’43.41”N12°14’47.27”W
Figure 1.

Map of locations of treeholes and tree species in the PK-10 forest sampled in the present study. Also shown are the locations of Saba senegalensis. In each of these four locations, only fruit husks were sampled. The locations of all sites were recorded with a GPSMAP® 62s (Garmin) and coordinates are listed in Table 1. Marigot and Station Towers are shown to provide reference points relative to earlier studies that included PK-10 (Raymond et al. 1976, Monlun et al. 1993, Cornet et al. 1975).

All treeholes were completely dry during the February “dry season” collection. In treehole cavities that were large enough to reach into, we scooped all loose detritus into a plastic bag. A cold chisel was then used to scrape the inside of the cavity and this loosened material was collected into the same plastic bag. The tree hole was immediately filled to the brim with water and then, using a siphon, the majority of this water was recovered from the treehole into a collecting container. This was immediately followed by a second wash and siphoning. Treeholes that were too small to reach into were washed and siphoned twice. We intentionally made no collections during August-September, 2011 to allow PK10 treeholes and S. senegalensis husks to go through one wet and dry season following our February, 2010 collections.

The same procedures were followed and the same treeholes were sampled from September 7–30, 2011. All treeholes contained free standing water or moist detritus and mud during the September “wet season” collection. All free standing water was first removed into a plastic pail. The hole was subsequently immediately filled to the brim with water brought from the laboratory and stirred. All liberated material was scooped into the pail. Any remaining water was removed to the pail using a siphon. This was immediately followed by a second wash, stirring, and siphoning. Approximately 50 Saba senegalensis husks that had been split open (presumably by foraging monkeys) were collected into a plastic pail at sites 3, 11, 12, and 18 (Figure 1).

All 14 plastic bags and pails containing treehole water and four pails containing S. senegalensis husks were returned to our local laboratory where husks and dry contents of treeholes were separately flooded. All containers were checked daily in the laboratory, larvae were collected and transferred into cups, supplemented with Brewer's yeast, and reared to adults in Bug-dorm® DP1000 cages (Bug-dorm Store, Taichung, Taiwan). Larval mortalities were not recorded. Adult mosquitoes which had eclosed within the previous 12–16 h period were aspirated and transferred into one-pint cartons that had been covered with mesh. These were then knocked down (in most cases killed) with FlyNap (Triethylamine) (Carolina Biological Supply Company, Burlington, NC). Adults were then individually removed with forceps to the stage of a dissecting microscope (Olympus) where they were identified to species. We progressively developed a regional key of adult Aedes for southeastern Senegal. Initially we used Huang and Ward 1981, followed by Huang (2004) for the subgenus Stegomyia and Huang (1990) for the africanus group. When these failed (e.g., Ae. capensis and Ae. simulans), we used Edwards (1941). Aedes aegypti s.l. were also scored on a scale from 1 (“F”) to 15 (“Q”) by separating the wings and examining the amount of scaling on the abdominal tergites and scoring these based upon the diagrams in McClelland (1974) to distinguish between Aaa and Aaf. All mosquitoes were then individually stored in Purell® Advanced Hand Sanitizer as voucher specimens and for eventual extraction of DNA.

Proportions of species were compared among larval habitats, collection methods, and seasons by calculating Bayesian 95% Highest Density Intervals (95% HDI) using WinBUGS (Lunn et al. 2000) and the analysis of contingency tables script (Box 6.13 in McCarthy 2012). Species diversity was summarized with Shannon's diversity index H (Shannon 1948) where H = −∑ pi ln(pi) and pi is the proportion of species i in a collection. The 95% HDI around H was estimated to compare H among collections using WinBUGS and the script in Box 3.15 of McCarthy (2012). Proportions and diversity indices with non-overlapping 95% HDI were considered credible.

RESULTS

Seasonal shifts in abundance of Aedes aegypti s.l. subspecies

There was a credible lower abundance of Aaf during the dry season (60.7%) as compared with the wet season (90.0%), mainly due to an excess of McClelland G mosquitoes in the dry season (Table 2a). During the dry season, percentages of Aaf were the same regardless of whether larvae were collected in treeholes by excavation, flooding, or by submerging S. senegalensis husks (Table 2b). This was also true during the wet season (Table 2c).

Table 2. Abundance of each Aedes aegypti subspecies and McClelland's forms during the dry and wet seasons in the PK-10 forest of southeastern Senegal. Values defining the 95% HDI appear in parentheses. The first number in parentheses is the 2.5% HDI, the second underlined value is the mean percentage, while the third value is the 97.5% HDI. Rows in which mean percentages had overlapping 95% HDI are highlighted in grey. Values are not highlighted when credible differences exist between mean percentages.
Image

Dry season treehole collections captured 355 mosquitoes of which 147 (41.4%) were Ae. aegypti s.l. (Table 3). A similar percentage was obtained from the same treeholes sampled during the wet season (43.1%, Table 4). Flooded treehole collections contained the same percentage of Ae. aegypti s.l. in the dry (34.4%, Table 3) and wet seasons (43.1%, Table 4). However, excavated contents of treeholes in the dry season produced a credibly higher percentage of Ae. aegypti s.l. (64.8%, Table 3) than flooded treehole collections. Half (49.1%, Table 3) of the mosquitoes from flooded S. senegalensis husks in the dry season were A. aegypti s.l. compared to 6.6% during the wet season (Table 4).

Table 3. Dry season abundance of Aedes species in different sampling habitats in the PK10 forest of southeastern Senegal. Values defining the 95% HDI appear in parentheses. The first number in parentheses is the 2.5% HDI, the second underlined value is the mean percentage, while the third value is the 97.5% HDI. Rows in which mean percentages had overlapping 95% HDI are highlighted in grey. Values are not highlighted when credible differences exist across mean percentages.
SpeciesExcavated treeholes(95% HDI)Flooded treeholes(95% HDI)Saba senegalensis husks(95% HDI)
Ae. aegypti s.l.53(54.0%, 64.6%, 74.8%)94(28.9%, 34.4%, 40.2%)26(35.6%, 49.1%, 62.4%)
Ae. apicoargenteus0(0.0%, 0.0%, 1.6%)7(1.0%, 2.6%, 4.6%)1(0.0%, 1.9%, 7.0%)
Ae. luteocephalus7(3.2%, 8.5%, 15.6%)25(6.0%, 9.2%, 12.9%)7(5.4%, 13.2%, 23.3%)
Ae. metallicus3(0.4%, 3.7%, 8.5%)5(0.6%, 1.8%, 3.7%)5(2.5%, 9.4%, 18.5%)
Ae. opok0(0.0%, 0.0%, 1.4%)1(0.0%, 0.4%, 1.2%)0(0.0%, 0.0%, 2.9%)
Ae. stokes4(0.9%, 4.9%, 10.4%)11(2.0%, 4.0%, 6.6%)0(0.0%, 0.0%, 3.4%)
Ae. unilineatus15(10.7%, 18.3%, 27.5%)130(41.7%, 47.6%, 53.6%)14(15.5%, 26.4%, 39.3%)
TOTAL82 273 53 
Shannon Diversity Index (95% HDI) (0.92, 1.11, 1.29) (1.19, 1.29, 1.39) (1.12, 1.30, 1.45)
Table 4. Wet season abundance of each Aedes species in different sampling habitats in the PK10 forest of southeastern Senegal. Values defining the 95% HDI appear in parentheses. The first number in parentheses is the 2.5% HDI, the second underlined value is the mean percentage, while the third value is the 97.5% HDI. Mean percentages with overlapping 95% HDI are highlighted in grey. Values are not highlighted when credible differences exist between mean percentages.
SpeciesFlooded treeholes(95% HDI)Saba senegalensis husks(95% HDI)
Ae. aegypti s.l.59(34.8%, 43.1%, 51.6%)21(4.1%, 6.6%, 9.7%)
Ae. apicoargenteus22(10.5%, 16.1%, 22.5%)0(0.0%, 0.0%, 0.5%)
Ae. capensis4(0.9%, 2.9%, 6.1%)0(0.0%, 0.0%, 0.3%)
Ae. cozi17(7.4%, 12.4%, 18.3%)0(0.0%, 0.0%, 0.5%)
Ae. furcifer-taylori4(0.9%, 2.9%, 6.1%)0(0.0%, 0.0%, 0.3%)
Ae. luteocephalus0(0.0%, 0.1%, 1.0%)1(0.0%, 0.3%, 1.4%)
Ae. metallicus7(0.1%, 5.1%, 4.7%)0(0.0%, 0.0%, 0.0%)
Ae. simulans1(0.1%, 0.7%, 2.4%)0(0.0%, 0.0%, 0.2%)
Ae. unilineatus7(0.1%, 5.1%, 4.7%)0(0.0%, 0.0%, 0.0%)
Ae. vittatus1(0.1%, 0.7%, 2.7%)1(0.0%, 0.3%, 1.1%)
E. chrysogaster15(6.0%, 10.9%, 16.8%)295(89.6%, 92.9%, 95.3%)
TOTAL137 318 
Shannon Diversity Index (1.639, 1.79, 1.927) (0.234, 0.32, 0.414)
E. chrysogaster excluded from diversity analysis (1.46, 1.63, 1.779) (0.211, 0.50, 0.810)

Seasonal shifts in species abundance

Seven species were collected during the dry season and their abundances were similar in excavated treeholes, flooded treeholes, and S. senegalensis husks with two exceptions. Credibly more Ae. aegypti s.l. were obtained from excavated treeholes, while more Ae. unilineatus were obtained from flooded treeholes (Table 3). Shannon diversity indices were uniform among species collected in excavated and flooded treeholes, and S. senegalensis husks (Table 3).

Thirteen species were collected during the wet season and the Shannon diversity index (1.79, Table 4) was credibly greater than in the dry season (1.11–1.30, Table 3). However, in S. senegalensis husks an opposite trend was seen, wherein the diversity index was credibly greater in the dry season (1.30, Table 3) than in the wet season (0.32, Table 4). Only three species were collected during the wet season and two of these were represented by a single individual and the remainder was all E. chrysogaster s.l.

DISCUSSION

There are five principal findings of this study. First, we detected a lower abundance of Aaf during the dry season as compared with the wet season. This contrasts with Paupy et al. 2010 who instead reported a greater abundance of Aaf in the dry season. However, they only sampled domestic containers in the dry season. Nevertheless, they reported that ∼45–50% of female mosquitoes had high McClelland scores of 8–10 (letter scores of L-M) during the rainy season as compared to ∼10% during the wet season. Paupy et al. 2010 observed a decrease in the relative abundance Aaf during the wet season, while we noted an increase. However, we only noted this trend in treeholes not in S. senegalensis husks. They also found a difference in McClelland scores between males and females while McClelland (1974) and the present study detected no such differences (analyses not shown). Furthermore, gender differences were not large in natural containers in the wet season (Paupy et al. 2010). The present study and Paupy et al. 2010 document shifts in the relative abundance Aaf between the dry and wet seasons.

Secondly, while dry and wet season flooded treeholes yielded similar percentages of Ae. aegypti s.l., its abundance differed greatly between S. senegalensis collections in the dry and wet seasons. It is most likely that the lower percentage in wet season husks was due to the abundance of facultatively predaceous E. chrysogaster s.l. (92.8%) larvae as compared to treeholes (10.9%, Table 4). Eretmapodites (Theobald) is a small genus containing 24 species. Haddow (1946) first commented on the predatory habits of Eretmapodites larvae in Africa noting that the mouthparts “have all been found to possess a group of thickened, comb-like hairs on the medio-ventral aspects of the mouth brushes (p. 58).” He further observed that “an Eretmapodites larva, after seizing its victim, holds it between the half-flexed head and ventral surface of the thorax. The prey is consumed rapidly – a large larva may be devoured in about 10 minutes – and larvae are attacked even in the presence of abundant other food material (p. 59).” We made similar observations. Haddow (1946) also made the first major taxonomic review of Eretmapodites species in Africa and identified four groups: E. chrysogaster s.l. (Graham) group (five species), E. inornatus (Newst.) group (two species), E quinquevittatus (Theo.) group (two species) and E. oedipodius (Edw.) (two species). Near Mombasa, Kenya, Lounibos (1978) showed that E. subsimplicipes (E. chrysogaster s.l. group) has a strong preference for fruit husks including the congeneric species Saba florida, while E. quinquevittatus has a preference for tap water. Lounibos (1981) reported a peak abundance of E. subsimplicipes in August and September from bamboo traps in forested collecting sites near Mombasa in Kenya. Raymond et al. 1976 sampled treeholes and fruit husks near Kedougou and also reported a large number of E. chrysogaster s.l. in husks collected from August-September, 1974. As in the present study, E. chrysogaster s.l. constituted 90–94% of mosquitoes collected in fruit husks and adults were easily identified based upon their large size, lack of ornamentation on the scutum, and the presence of large plumes on the hind tarsi of males. Eretmapodites males lacking these plumes or females with scutum ornamentation were never recovered in the present study.

Third, we detected large seasonal shifts in species abundance and diversity. Fewer species were collected during the dry season. This result is not surprising given that species richness is expected to be greater in wet vs dry treeholes. Furthermore, ephemeral fruit husks would not be expected to support the number of species as are found in stable treeholes. Large seasonal shifts in larval species composition have been previously documented by Haddow in East Africa (Haddow 1945), Teesdale (Teesdale 1959) and later Lounibos along the Kenya coast (Lounibos 1981), and Corbert in Uganda (Corbet 1964). The same was noted in West Africa in the Northern Guinea Savannah of Nigeria (Service 1965), in southern Nigeria (Dunn 1927, Kerr 1933, Mattingly 1949a,1949b), in Liberia (Rozeboom and Burgess 1962), in Ghana (Addy et al. 1996), and in Senegal (Raymond et al. 1976).

Fourth, sampling technique and location affected species composition during the dry season (Table 3). Specifically far more Ae. unilineatus were collected by flooding while the relative abundances of Ae. aegypti were greater in excavated material. The effect of location on species composition was even greater during the wet season in which the relative abundances of the 11 species were significantly different between flooded treeholes (ten species) and S. senegalensis husks (four species), and most (92.8%) of the mosquitoes collected from husks were E. chrysogaster s.l. Lounibos (1981) described different hatching patterns among treehole Aedes and demonstrated that Stegomyia spp. hatched first, followed by other subgenera. This may explain why wet season samples contained a higher proportion of non-Stegomyia aedines compared to flooded, dry-season tree holes. Also Sota and Mogi 1992 showed that eggs of forest-dwelling Stegomyia are less desiccation-resistant than non-forest counterparts, which may explain the differences between Aaf and Aaa in treeholes in the dry vs wet seasons. At present, we are uncertain of the reasons that treehole excavation and flooding would yield different numbers of species. More careful excavation of different regions of the treehole might provide answers. In addition, more careful sampling might have detected differences in species abundance among different sizes of treeholes and possibly differences among tree species.

Fifth, this study suggests that both Aaa and Aaf may survive the tropical dry season in natural larval habitats such as treeholes and husks in West Africa. Many reports from coastal Kenya (e.g., Trpis & Hausermann 1975, 1978) concluded that Aaf and Aaa may hybridize peridomestically in the rainy season. Sylla et al. 2009 reported Aaf domestic indoor larval habitats in Senegal. This is consistent with earlier reports from Nigeria and Gabon (Dunn 1927, Kerr 1933, Mattingly 1949a,1957, Service 1963, 1965). Furthermore, in our collections from treeholes in PK-10 we recovered an abundance of Aaa. Both trends contrast with previous reports from East Africa (Mattingly 1957, Trpis and Hausermann 1975, 1978, 1986) of 1) strict endophily in Aaa and exophily in Aaf, and 2) household containers as the exclusive larval habitats for Aaa and treeholes as the predominant larval habitats for Aaf.

Acknowledgments

This work was supported by NIH R01AI0833680. We thank Pete Graham (petit pete) for his assistance in mosquito collection and processing. We thank Dr. Phil Lounibos and two anonymous reviewers for invaluable corrections and historical insights.

Ancillary