Culicidae ( Diptera : Culicomorpha ) from the central Brazilian Amazon : Nhamundá and Abacaxis Rivers

Mosquito fauna (Culicidae) from remote areas along the geographical limits of the State of Amazonas were assessed by employing CDC, Shannon, Malaise and Suspended traps, together with net sweeping and immature collections. Two hundred and six collections were performed in seven localities along the Nhamundá and Abacaxis Rivers, State of Amazonas, Brazil, during May and June 2008. The northernmost locality was 120 km from Nhamundá, whereas the southernmost locality was 150 km from the mouth of the Abacaxis River. The 5,290 mosquitoes collected are distributed in 16 genera, representing 109 different species, of which eight are new distributional records for the State of Amazonas. Furthermore, there are nine morphospecies which may represent undescribed new taxa, five of which are also new records for the State of Amazonas. Culex presented the highest number of species and the largest number of individuals. Anopheles, which represents 3% of the total sample, had the second highest number of species, followed by Wyeomyia. Psorophora and Aedes, represent the third and fourth largest number of individuals. The most abundant species was Cx. (Mel.) vaxus Dyar, 1920 followed by Cx. (Mel.) eknomios Forattini & Sallum, 1992, Cx. (Cux.) mollis Dyar & Knab, 1906, Cx. (Mel.) theobaldi Lutz, 1904, and Cx. (Cux.) declarator Dyar & Knab, 1906. The epidemiological and ecological implications of mosquito species found are discussed and are compared with other mosquito inventories from the Amazon region. The results presented represent the largest standardized inventory of mosquitoes of the Nhamundá and Abacaxis rivers, with the identification of 118 species level taxa distributed in seven localities, within four municipalities (Nhamundá, Maués, Borba, Nova Olinda do Norte), of which we have only few or no records in the published literature.

in more recent publications. Most of these new records of distribution are found in publications resulting from inventories (BARBOSA et al. 2008, HUTCHINGS et al. 2002, 2005, 2010, 2011, SUÁREZ-MUTIS et al. 2009) and as a result of the description of new species (FORATTINI & SALLUM 1992, SALLUM & HUTCHINGS 2003, SALLUM et al. 1997. It should be noted that many of the published records are not results of collections made with the purpose of studying the entire mosquito community, but mainly had epidemiological objectives (CERQUEIRA 1961, DEANE 1947. Therefore, any list of species prepared for a given location which is based on published records may be incomplete and/or biased. For example, after collecting 119 species in the Jau National Park, 25% (30 species) were new records for the State of Amazonas (HUTCHINGS et al. 2005) and of 145 species collected north of Manaus (HUTCHINGS et al. 2011), 16% (23 species) are also new records for the State of Amazonas, including seven new records for Brazil. Outside of being biased, the geographical distribution of published records is still unrepresentative given the low coverage of the municipalities. Although the coverage includes 61% of the municipalities within the State of Amazonas, the sampled area of each municipality is still very small.
It is important to consider that an increase in the knowledge of the mosquito fauna of the Amazon region will permit us to obtain basic information of the faunal diversity, distribution and variety of ecosystems where mosquitoes occur, thus providing basic knowledge for studies on the control of diseases which affect humans and animals, whose infectious agents are transmitted by mosquitoes. In this work, we present the first results of mosquito collections from remote regions lo-cated near the political boundaries of the State of Amazonas, as part of the project "Amazonas: Diversidade de insetos ao longo de suas fronteiras" of the Programa de Apoio a Núcleos de Excelência (FAPEAM-CNPq).
Therefore, with the objective of serving as a base inventory for future surveys of Culicidae from the Amazon, the mosquito species collected inside the riparian and terra firme forests along the basins of the Nhamundá and Abacaxis rivers, Amazonas, Brazil, are reported herein.

MATERIAL AND METHODS
A mosquito inventory, conducted during a river expedition in areas near the eastern border of the State of Amazonas, Brazil (Fig. 1), includes collections of mosquitoes from seven different localities: two localities along the Nhamundá River, Municipality of Nhamundá (between 01°35'S, 057°37'W and 01°53'S, 057°03'W); and five localities along the Abacaxis River, including the Municipalities of Maués, Borba and Nova Olinda do Norte (between 05°15'S, 058°41'W and 04°28'S, 058°33'W). These localities are characterized by having most of their area covered by dense upland (terra firme) ombrophilous forests with low plateaus, together with riparian rain forests having dense alluvial and lowland vegetation (Floresta Ombrófila Densa Aluvial e de Terras Baixas) along the rivers, intermixed with areas of transition including Amazonian white sand (campinarana) and floodplain (varzea) forests. The tropical rainforest climate is warm and wet, characterized by being constantly humid, with temperature and precipitation with little annual variation. Based on climatic data from Parintins and Maues (RADAMBRASIL 1976), the  Table I).
region has an annual relative humidity of 86% and a mean annual temperature of 26°C. A shorter dry season occurs from July to November with the lowest monthly precipitation being less than 50 mm along the Nhamundá River and over 150 mm along the Abacaxis River. The rainy season occurs between December and June, with the maximum precipitation in April. For the Nhamundá basin the mean annual precipitation is 1,750 mm, while the Abacaxis basin is greater than 2,750 mm.
Located in the far eastern region of the State of Amazonas, there are many difficulties in accessing the collection locations because of the long distance from urban centers. The most remote locality surveyed along the Nhamundá River is 240 km from Parintins and 630 km from Manaus, while the localities along the Abacaxis River are 360 km from Maués and 530 km from Manaus. The region, with a very low demographic density, only has small settlements which occupy marginal areas along the rivers. The main means of transport is by boat.
Mosquitoes specimens were mostly collected inside the riparian forest along existing and/or newly created trails, perpendicular to the river banks, and within continuous upland terra firme forest using a variety of capture methods including: CDC traps with different types of lighting (incandescent lamp (CDC) or ultraviolet fluorescent tube (UV CDC)); flight intercept traps (Malaise, Shannon, Suspended); and sweeping with nets. The CDC traps were installed every 50 m along trails, placed at 1, 10, 15 or 20 m above the ground, and were activated at dusk for a period of 12 hours, between 18:00 and 06:00 h. The Shannon traps, placed within small open understory areas, using an internal light source for attraction and a portable battery powered aspirator for capturing specimens, were used between 18:00 and 21:00 h. The 6 m long Malaise flight intercept traps, also placed within small open understory areas, were used for periods of up to three days and the Suspended flight intercept traps were hung one meter above the water level, along river margins, or at tree canopy level, also for periods of up to three days. Each sweeping collection, using entomological nets, was performed during a minimum of two hours at each location. Immature mosquitoes were collected from breeding sites found along trails, in the same areas where adults were captured. The immature mosquitoes were reared for the purpose of obtaining adult males and females, associated with larval and pupal exuviae. These reared specimens were mostly used to obtain a more accurate identification of adult female specimens captured using other methods.
Adult and immature mosquitoes were captured, preserved and mounted following techniques detailed by BELKIN et al. (1967). Specimens were identified in the laboratories at INPA in Manaus, and were confirmed at the Laboratório de Sistemática e Ecologia de Culicidae (LASEC), of the Faculdade de Saúde Pública (FSP/USP), in São Paulo and in the Laboratório de Transmissores de Hematozoários of the Instituto Oswaldo Cruz (IOC), in Rio de Janeiro, using the identification keys in LANE (1953a, b), FORATTINI (1965a, b, 2002, ZAVORTINK (1972ZAVORTINK ( , 1979, ARNELL (1973), VALENCIA (1973), BERLIN & BELKIN (1980), SALLUM & FORATTINI (1996), as well as the PECOR et al. (1992) catalog for Culex (Melanoconion). Whenever possible, anatomical characteristics of the male genitalia were examined to confirm the identifications of both females and males. The collected material will be deposited in the Coleção de Invertebrados of the Instituto Nacional de Pesquisas da Amazônia (INPA-Manaus), in the Coleção Entomológica de Referência of the Faculdade de Saúde Pública, Universidade de São Paulo (FSP/ USP) and in the Coleção de Culicídeos of the Instituto Oswaldo Cruz (FIOCRUZ). The collection and specimen data was digitized, stored, archived and organized using the relational database structure provided by the Biota software version 2.04 (COLWELL 2012).

RESULTS
A total of 206 collections distributed in seven locations along the Nhamundá and Abacaxis Rivers, in the State of Amazonas, from May 14 to June 6, 2008, resulted in the capture of over 5,000 mosquitoes (Table I). Each collection corresponds to the capture yield of a trap (i.e. CDC, CDC UV, Shannon, Malaise, and Suspended) or method (i.e. sweeping,  (3); in a tree hole; and in a Bertholletia pixidium. More specific information on the sampling effort for each locality is presented in Table II.
Of the 5,290 specimens captured, 5,231 were identified and are distributed in 16 genera, representing 118 different taxa (among species and morphospecies) (Appendix 1). The morphospecies (identified as near F#) are similar to a known species, but it is believed that some may represent undescribed new taxa. Some identification could not be exact because of the absence of males, whose genitalia usually possess anatomical features that allow the specific diagnosis. These individuals were identified as morphotypes, indicating the species to which they are most similar.
Unfortunately, among the mounted, sorted and examined material, it was not possible to identify 1,815 specimens (Ϸ 34%) to the species level for several reasons: either there are no known characters to separate female individuals of different species or the characters used to separate these species were damaged. For some of these individuals it was only possible to identify to ge-nus level because the characters which are used for identification are damaged and/or lost, and the rest of the collected material was recognized to subgeneric or informal taxonomic group (sections or groups) level (shown with the prefix "gr.", "sG." or "sec." or the suffix "sp." in Appendix 1). Most of the individuals that could not be identified to species level are females (1790 H" 98%) and belong to the genus Culex (91%) (Appendix 1). It is interesting to note that only 13% of the specimens collected in this inventory were males.
The most abundant species was Culex (Mel.) vaxus Dyar, 1920 (587 individuals collected, representing 17% of the material identified to species level) followed by Cx. Lutz, 1904, and Cx. (Cux.) declarator Dyar & Knab, 1906 (with 481, 456, 415 and 255 individuals, respectively). The five most abundant species (<5% of the recorded species) represent 66% of specimens identified to the species level.  Bruijning, 1959. There are also 203 specimens of at least nine morphospecies (marked as near F # in Appendix 1), of which five also represent new geographical records for the State of Amazonas. These morphospecies, which probably represent species not yet described, belong to three different genera. Aedes Together, the nocturnal collecting methods (CDC, CDC-UV and Shannon Traps) were responsible for 90% of the captured mosquitoes, of which the CDC traps (with a total combined sampling effort of 162 trap-nights) were responsible for more than 83%. Net sweeping accounted for 7%, followed by the Suspended and Malaise flight intercept traps, with 1.3% and 1% of the specimens, respectively (Table II). Both types of CDC traps together were responsible for collecting 62% (73) of the species level taxa, of which the CDC-UV trap alone collected 53% of the species level taxa. The adult specimens of Aedeomyia, Orthopodomyia and Uranotaenia were only collected at night (CDC, CDC-UV and Shannon), while Haemagogus was only collected during the day and Onirion was only registered by rearing larvae. The methods of capture for each taxon can be seen in the final columns of Appendix 1. Net sweeping, and the CDC type traps combined, were responsible for the highest number of species which were only and exclusively collected with a specific method of capture, although every method did collect exclusive species (see details in Table II). The diurnal mosquitoes are not equally represented in this inventory, compared to the nocturnal mosquitoes because the sampling effort was greater for the nocturnal collecting methods.
Of the 118 different species level taxa identified during this inventory, 48 (41%) were collected in both river basins, while 29 species (24%) were found only along the Nhamundá River and 41 species (35%) were only found along the Abacaxis River (Appendix 1). The results of the mosquito inventory for each separate river basin are presented below.

Nhamundá River
The inventory along the Nhamundá river basin, sampled from May 16 to 19, includes specimens from 98 collections in two localities (Table I), resulting in 933 mosquitoes from 15 genera, representing 77 different taxa identified to species level (between species and morphospecies) (Appendix 1). It was not possible to identify 256 specimens (H" 27%) to the species level for the reasons previously discussed. For this basin, the sampling effort included 528 CDC trap-hours, 408 CDC (UV) traphours, 17 Shannon trap-hours, 9 Malaise trap-days, 3 Suspended trap-days, 18 net sweeping hours and two immature collections (in a Bertholletia pixidium and a Bromeliaceae leaf axil) (Table II).
In the Nhamundá River basin, the genus Culex presented the highest number of species (26 H" 38%) and the largest number of individuals (712 H" 77%). The genus Wyeomyia, which represents 4% of the total sample (34 specimens), had the second highest number of species (11 H" 16%), followed by Psorophora with only seven species represents the second largest number of specimens (71 H" 8%). Anopheles had 6 species (H" 9%), and less than 4% of the individuals, while Aedes with five species had the third largest number of individuals (45 < 5%). The most abundant taxon was the morphospecies Culex (Mel.) near vaxus F3 (142 individuals collected, representing 23% of the specimens identified to species level) followed by Cx.

Abacaxis River
The inventory along the Abacaxis river basin, sampled from May 26 to June 4, includes specimens from 108 collections in five localities (Table I), resulting in 4,357 mosquitoes, from 15 genera, representing 89 different taxa identified to species level (including species and morphospecies) (Appendix 1). It was not possible to identify 1558 specimens (H" 36%) to the species level for the reasons previously discussed. For this basin, the sampling effort included 672 CDC trap-hours, 336 CDC (UV) trap-hours, 11 Shannon trap-hours, six Malaise trapdays, three Suspended trap-days, 22 net sweeping hours and six immature collections (in lakes and streams (three samples), in two Bromeliaceae leaf axils, and in a tree hole (Table II).
In the Abacaxis River basin, Culex presented the highest number of species (37 H" 44%) and the largest number of individuals (3,941 H" 92%). Anopheles, which represents 3% of the total sample (136 specimens), had the second highest number of species (13 H" 15%), followed by Psorophora and Aedes with six species (H" 7%) each, representing the third largest (107 H" 3%) and the fourth largest number of individuals (45 = 1% Among the 84 species, collected along the Abacaxis River, there are three new species distribution records for the State of Amazonas. There are also 43 specimens of at least five morphospecies (H"6%), of which two represent new geographical records for the State of Amazonas. These morphospecies, which probably represent species not yet described, belong to two different genera: Ae.

DISCUSSION
Among the 118 species level taxa collected in this inventory, there are 13 (11%) new geographical distribution records for the State of Amazonas. Other mosquito surveys from upland terra firme sites have similar results: of the 145 species collected, north of Manaus, 16% (23 species) were new records for the State of Amazonas (HUTCHINGS et al. 2011); of the 119 species collected in the Jau National Park, 25% (30 species) were new records (HUTCHINGS et al. 2005); and of the 44 species recorded in Querari, 27% (12 species) were also new records for the state (HUTCHINGS et al. 2002). We found no previously published mosquito distributional records for the municipalities of Nhamundá and Nova Olinda do Norte. Therefore, the results of this inventory represent the first published report of mosquito taxa for these municipalities.
Epidemiologically, the presence of Anophelines may be important because this genus includes Plasmodium vector species, which cause malaria in humans. Within the Anopheles, it is worth noting the presence of Anopheles (Nys.) konderi s.l., An. (Nys.) oswaldoi s.l. and Anopheles (Nys.) triannulatus are consid-ered secondary vectors, but they can take the role of local or regional primary vectors (FORATTINI 2002). Considering that Anopheles (Nys.) konderi s.l. and An. (Nys.) oswaldoi s.l. were demonstrated to be species complexes (MOTOKI et al. 2009, the vector status of each species needs to be determined in further studies conducted in areas of malaria transmission where species of these complexes are present. Additionally, the absence of An. (Nys.) darlingi and also of species of the An. (Nys.) albitarsis complex may be indicative of an undisturbed natural environment. In several studies conducted inside pristine areas of the State of Amazonas, no specimens of An. (Nys.) albitarsis s.l. and only a few specimens of An. (Nys.) darlingi were found. For example, only seven An. (Nys.) darlingi specimens (2% and 4% of the Anophelines, respectively) were collected in both the Jau National Park (HUTCHINGS et al. 2005) and in the Juami-Japura Ecological Station (HUTCHINGS et al. 2010), while only one An. (Nys.) darlingi specimen (<0.4%) was found in areas north of Manaus (HUTCHINGS et al. 2011). In contrast, An. (Nys.) darlingi can be the most prevalent species inside deforested areas of the Amazon region (CASTRO et al. 2006), whereas species of An. (Nys.) albitarsis complex can become more frequent depending on the land use (CONN et al. 2002).
Furthermore, there are Culex species which are potential vectors of arboviruses. For example, Cx. gnomatos, the second most common Culex species in these samples, is highly susceptible to infection by enzootic (ID and IE) and epizootic strains (IAB and IC) of the Venezuelan Equine Encephalitis Virus (VEEV) (TURELL et al. 2000). It is worth mentioning that Cx. pedroi, also a common species in these collections, is considered a potential enzootic vector of the Eastern Equine Encephalitis Virus (EEEV), in Brazil, as well as of the VEEV and other arboviruses (GALINDO & SRIHONGSE 1967, GALINDO et al. 1966, SRIHONGSE & GALINDO 1967. Moreover, it is interesting to note that AITKEN (1972) observed that Cx. portesi may be involved in the of epizootic and enzootic transmission cycles of the Mucambo virus. Cx. spissipes is a potential vector of the Bimiti, Caraparu, Oriboca and Itaqui viruses, of the Bunyaviridae family and of the VEEV III-B subtype (SHOPE et al. 1988, WALTON & GRAYSON 1988. Considering the number of specimens and/or species resulting from the different methods of capture and sampling efforts of this inventory (Table II), future mosquito surveys should give priority to the use of CDC type traps and net sweeping in order to maximize collecting results, when time and field resources are limited.
This mosquito inventory is part of a larger entomological inventory of different locations within remote and sparsely populated areas near the border regions of the State of Amazonas which also resulted in the collection of a large number of other insects, including Lepidoptera (CASAGRANDE et al. 2012). The information presented here represents the largest standardized mosquito inventory ever executed, within the Nhamundá and Abacaxis river basins, with the identification of 118 taxa distributed in seven different locations within four different counties (Nhamundá, Maués, Borba, Nova Olinda do Norte), of which few or no geographical records have been previously published.

ACKNOWLEDGMENTS
We wish to thank Monique de A. Motta (FIOCRUZ-RJ) for help with the identification of the Wyeomyia and the technicians, Luis Aquino, Jose M. da S. Vilhena, and Isis S. Menezes, for their help collecting and processing specimens. All specimens were collected using the "Autorização para Atividades com  a Indicates the total number of specimens for each taxon.
b Indicates the sex and/or condition of the specimens collected: M = male; F = female; dam = damaged specimen; gen = identified using genitalia dissection. c X = specimens of the taxon were only captured using this method.
d It was not possible to identify these specimens to species level because either there are no known characters which can be used to separate the females of these two species, or the characters used to separate these species are damaged.