Diversity of yellow fever mosquito vectors in the Atlantic Forest of Rio de Janeiro , Brazil

Introduction: Environmental modifications caused by human activities have led to changes in mosquito vector populations, and sylvatic species have adapted to breeding in urban areas. Methods: Mosquitoes were collected using ovitraps in three sampling sites in the Atlantic Forest in the State of Rio de Janeiro, Brazil. Results: We collected 2,162 Culicidae specimens. Haemagogus janthinomys and Haemagogus leucocelaenus, both sylvatic yellow fever virus vectors, were the most common species found. Conclusion: There is a potential for the transmission of arboviruses in and around these natural reserves. Therefore, it is necessary to maintain entomological surveillance programs in the region.

Understanding the biodiversity of mosquito species in the Atlantic Forest and their response to both human disturbance and forest recovery is important for predicting changes in mosquito populations, especially those commonly associated with sylvatic habitats.Although the mosquito fauna of the Atlantic Forest is diverse and includes potential vectors for yellow fever virus as well as other arboviruses, from an epidemiological point of view, the Haemagogus and Sabethes spp.are the most important in the transmission of yellow fever virus because they are the primary vectors in the forest areas of the Americas (1) .Haemagogus spp., in particular, are sylvatic, active during the warmest hours of the day, and found primarily in the tree canopy of tropical forests.Nonetheless, they will take blood meals at ground level in deforested areas and some of these species also show a tendency toward domiciliation (2) .However, behavioral tendencies may vary across regions and seasons.Therefore, we collected mosquito eggs in order to evaluate the mosquito diversity in environmental preservation areas in the Southeastern Brazilian State of Rio de Janeiro.1).PARNA-Itatiaia, situated 176km from the City of Rio de Janeiro, was the first national reserve in Brazil.It covers an area of 28,155ha and is heavily affected by anthropogenic activities, including housing development and palm cabbage harvesting.The reserve includes two ecologically distinct areas between 400 and 2,791m above sea level: one with rock formations at higher elevations and one lower with numerous waterfalls and small lakes.Rainfall in PARNA-Itatiaia is heavy and occurs mainly in the summer.Annual precipitation averages 2,400mm with the heaviest rainfall in January (27 rainy days and 388mm of rainfall on average).The collection site was located at 22º25'52.1"S and 44º37'16.7"W.
RBioPA is situated in the municipality of Silva Jardim and encompasses an area of 5,000ha.Constituted in 1914, the reserve includes several areas that were previously orchards, houses, or pastures; however, the forest has gradually recovered, and primary forest fragments with original vegetation can be found on the alluvial plains and in the lower areas of the reservation.The climate is hot and wet with most rainfall occurring in the summer (total rainfall = 1,000mm concentrated between October RPPNBR is situated 140km from the City of Rio de Janeiro.It covers an area of approximately 556ha and is almost completely covered by primary Atlantic Forest.The region is heavily influenced by intense solar radiation and Atlantic Ocean humidity producing a tropical wet climate (4) .The collection site was at 22º27'15.3"S and 42º18'02.4"W.

Rio de Janeiro
Mosquito egg sampling was conducted over 5 months (December 2014 to April 2015) using ovitraps consisting of a 1L black bucket installed 2-3m from the soil and containing water, leaf litter, and four wood plates.These plates were collected twice a month and examined in the laboratory.Plates with mosquito eggs were immersed in transparent trays filled with Milli-Q® water and maintained at 28 ± 1°C.Emerged adults were identified (5) by checking original descriptions and redescriptions when necessary.
We calculated the index of species abundance for each species and then standardized this on a scale from zero to one [standardized index of species abundance (SISA)] as described by Roberts & Hsi according to (6) .This index is determined by the number of specimens collected and the distribution pattern across samples.Species dominance categories were defined as eudominant (>10%), dominant (<10% and >5%), subdominant (<5% and >2%), recessive (<2% and >1%), and rare (<1%) (7) .
We also compared the mosquito diversity between sites with the Shannon-Wiener Diversity Index (H' = Σp i lnp i , where p i is the proportional abundance of species i in the collection) using the DivEs Species Diversity program (W.C.Rodrigues; LIZARO Soft).In addition, we calculated the species richness (S) and the Sørensen similarity index (SI).An SI > 0.50 was considered significant.Since collections were not conducted in April 2015 in PARNA-Itatiaia, all comparisons were restricted to the period from December 2014 to March 2015.
Between December 2014 and April 2015, 2,217 specimens from six mosquito species were collected.Since the studied areas were within the distribution of both Haemagogus janthinomys and Haemagogus capricornii spp.and because the females are very difficult to differentiate (2) and the only male was collected in RPPNBR, the females were only identified as H. capricornii/janthinomys. Five specimens could only be identified as Wyeomyia spp.(Table 1).In addition, Haemagogus leucocelaenus (Dyar & Shannon 1924) was the most abundant in all locations, followed by H. capricornii/janthinomys in RPPNBR and RBioPA, and by Limatus durhamii Theobald, 1901, in PARNA-Itatiaia (Table 1).H. leucocelaenus was the most dominant in all areas (Table 2).There was no significant difference in diversity among the localities (t-test, p > 0.05), and all localities had a similar species richness (more than 50% similarity).Two species were observed in RPPNBR, whereas five species were found and species richness was higher in PARNA-Itatiaia and RBioPA (Table 1).In RPPNBR and RBioPA, the population density was highest in December and April, respectively, and lowest in January.In PARNA-Itatiaia the population density was highest in March and lowest in February.
Of the three areas studied, the highest Shannon Diversity Index (H' = 0.37) was found for the RBioPA sample site and the greatest species richness (S = 5) was found for the PARNA-Itatiaia site.In addition, at the PARNA-Itatiaia collection site we found three epidemiologically important species: H. leucocelaenus, H. janthinomys, and A. albopictus (Table 1).The species diversity comparisons confirmed no significant differences between the different sampling areas (RBioPA x RPPNBR t-test = 22.8851; RBioPA x PARNA-Itatiaia t = 7.0586; RPPNBR x PARNA-Itatiaia t = 10.3493;p > 0.05 for all).
We also used the dominance index to analyze the species composition in each of the three study areas.In RBioPA, H. leucocelaenus and H. capricornii/janthinomys were eudominant, A. albopictus and C. iridescens were subdominant, and A. terrens was recessive.In RPPNBR, H. leucocelaenus was eudominant and H. janthinomys was recessive.In PARNA-Itatiaia, H. leucocelaenus was eudominant and L. durhamii was dominant (Table 2).
Nevertheless, ovitraps have some limitations.For example, they cannot be used to determine absolute population densities, the infusions are not standardized preventing comparison between different traps and occasions, and they are labor intensive (8) .However, the only alternative is to sample eggs from natural habitats; therefore, ovitraps should be complemented by human landing catches and larval surveys.In addition, ovitraps do not capture some species, such as flood mosquitoes (e.g., A. scapularis and A. albifasciatus); therefore, it is ideal to utilize several sampling methods (such as light traps), baits, and breeding places.However, ovitraps may provide useful data on seasonal fluctuations as well as height and environmental preferences.For example, H. janthinomys shows a clear preference for foraging at the highest levels of the forest canopy and lays eggs in tree holes situated in very high and unreachable places (5) , indicating preference for egg-laying in higher traps (9) .Except for A. albopictus, which has adapted to breeding in bamboo internodes and bromeliads (among other places), all species are adapted to several phytotelmata and some of them have also been found in artificial containers (10) .For example, Culex (Carrollia) spp.are commonly associated with several different phytotelmata (11) , including bamboo internodes, the fungus Aquascypha hydrophora, palm spathes, Heliconia, Araceae, and artificial containers.However, since immature forms of C. (Carrollia) iridescens (Lutz, 1905) are frequently found in natural habitats in Serra do Mar, São Paulo (12) , but remain absent from human landing catches in the same area (13) , these mosquitoes seem to have low anthropophily and thus may not be medically important.
Although the studied areas seemed to be quite ecologically different, they were not significantly different in terms of mosquito diversity.However, species dominance was different across sites.
Among the species already identified as potential vectors of yellow fever virus, H. janthinomys stands out as the principal vector in the Americas.This species appears to be highly adapted to different biomes and different abiotic conditions (e.g., temperature and humidity).The potential for virus transmission is enhanced by the geographic distribution of this mosquito, which coincides with areas known to be endemic for the disease (2) .
Three mosquito species epidemiologically important to the transmission of arboviruses (H.leucocelaenus, H. janthinomys, and A. albopictus) were collected in the present study; however,