Mosquitoes of the Caatinga: 2. Species from periodic sampling of bromeliads and tree holes in a dry Brazilian forest

Accepted Manuscript Title: MOSQUITOES OF THE CAATINGA: 2. SPECIES FROM PERIODIC SAMPLING OF BROMELIADS AND TREE HOLES IN A DRY BRAZILIAN FOREST Authors: Let´ıcia Silva Marteis, Delsio Natal, Maria Anice Mureb Sallum, Antˆonio Ralph Medeiros-Sousa, Roseli La Corte PII: DOI: Reference:

S0001-706X(16)31045-2 http://dx.doi.org/doi:10.1016/j.actatropica.2017.03.031 ACTROP 4257

To appear in:

Acta Tropica

Received date: Revised date: Accepted date:

1-12-2016 18-3-2017 19-3-2017

Please cite this article as: Marteis, Let´ıcia Silva, Natal, Delsio, Sallum, Maria Anice Mureb, Medeiros-Sousa, Antˆonio Ralph, Corte, Roseli La, MOSQUITOES OF THE CAATINGA: 2.SPECIES FROM PERIODIC SAMPLING OF BROMELIADS AND TREE HOLES IN A DRY BRAZILIAN FOREST.Acta Tropica http://dx.doi.org/10.1016/j.actatropica.2017.03.031 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

MOSQUITOES OF THE CAATINGA: 2. SPECIES FROM PERIODIC SAMPLING OF BROMELIADS AND TREE HOLES IN A DRY BRAZILIAN FOREST Letícia Silva Marteis1, Delsio Natal2, Maria Anice Mureb Sallum2, Antônio Ralph MedeirosSousa2 and Roseli La Corte3

1

Colegiado de Medicina. Universidade Federal do Vale do São Francisco. Av. José de Sá

Maniçoba, S/N, Centro, 56304-917, Petrolina, PE, Brasil 2

Departamento de Epidemiologia, Faculdade de Saúde Pública, Universidade de São Paulo,

Avenida Dr. Arnaldo, 715, 01246-904, São Paulo, SP, Brasil 3

Departamento de Morfologia, Universidade Federal de Sergipe, Avenida Marechal Rondon,

S/N, 49100-000, São Cristóvão, SE, Brasil

Corresponding author: Letícia Silva Marteis. Colegiado de Medicina. Universidade Federal do Vale do São Francisco. Av. José de Sá Maniçoba, S/N, Centro. Petrolina, PE, 56304-917. Email: [email protected]

2

Abstract

The Caatinga is a dry tropical forest, located in the Brazilian semiarid region and rich in phytotelmata. This study investigated the culicid fauna of phytotelmata of the caatinga by sampling for 19 consecutive months aquatic immatures from tree holes and bromeliads. A total of 127L of water was taken from the plants, containing 6,764 immature culicids of 16 species, of which 11 (69%) are undescribed and respond to 90% of the total abundance of the specimens collected. Epiphytic bromeliads harbor a large number of immature Culicidae, although terrestrial bromeliads are the most abundant and widely distributed in the region. The richness of culicid species was similar between terrestrial and epiphytic bromeliads and lower in habitats represented by tree hole phytotelmata. There was no similarity in the composition of culicid species that developed in bromeliads or tree holes. Temperature and humidity were the environmental parameters most strongly associated with the proportion of positive plants. The Caatinga has a great number of endemic species that remain unknown to science and many additional culicid species may await discovery from there.

Key words. Tropical dry forest, Culicidae, new species, phytotelmata, biodiversity, ecology.

1. Introduction

Mosquitoes are insects widely distributed around the world. There are approximately 490 species in Brazil (WRBU, 2015), mainly found in tropical rainforest biomes such as the Atlantic Forest and the Amazon Rainforest. Mosquitoes have the ability to develop in a variety of habitats, from wild environments to polluted water from urban areas and affinity for ovipositing in a wide range of sites, including those represented by phytotelmata. Phytotelmata are aquatic habitats naturally formed by water retention in plants or parts of terrestrial plants, living or dead, detached or not, e.g., tree holes, bromeliad ponds, bamboo internodes, open fruit and fallen leaves (Kitching, 2000). In these environments, which constitute complex micro-ecosystems despite the space and energy limitations, a high diversity of organisms, especially arthropods, coexists. These organisms belonging to different taxonomic groups interact ecologically and contribute to the diversity of patterns observed in nature (Fish, 1983; Kitching, 2000, 2001).

3

Bromeliad axils play an important role in the development of aquatic insects, including culicids. They store water between the overlaps of their lateral leaves and central well, enabling the establishment of various species in an environment rich in organic matter. In bromeliad phytotelmata, debris not only supports the community of insects but also serves as a source of nutrients for the plant (Ngai & Srivastava, 2006). Like bromeliads, tree holes constitute important habitats for the development of rich aquatic fauna and were once considered the oldest natural larval habitat of culicids (Jenkins & Carpenter, 1946). The main attraction of these phytotelmata stems is their water storage capacity, maintained mainly by the stemflow that brings nutrients into the holes (Walker et al. 1991) along with the deposition of organic matter in sufficient quantities to form habitats that ensure the development of culicids, even after the decline of rainfall (Jenkins & Carpenter, 1946; Gomes et al., 1992). Tree holes, formed from hard plant parts, are potentially more durable than other phytotelmata, lasting ten or more years (Mogi, 2004). The Caatinga is a tropical dry forest located in the Brazilian semiarid region, in which the culicid fauna associated with phytotelmata is virtually unknown. Specific information of culicid fauna in the Caatinga is scarce and restricted to transition areas with other biomes (Rebêlo et al., 2007; Santos et al., 2015). There is a recent record of immature fauna in an exclusive conservation area of the Caatinga biome, but associated to ponds (Marteis et al., 2015). Furthermore, the biological diversity of the Caatinga has been underestimated for decades, and the fauna and flora of the biome had been alleged to be poor, resulting in few areas of the Caatinga being preserved as fully protected units (Silva & Dinnout, 1999; Zanella & Martins, 2003). This ecosystem assessment was probably due to the Caatinga´s peculiar environmental conditions, with a seasonality marked by long periods of drought, scarce and erratic rainfall, high temperatures and low relative air humidity (Andrade-Lima, 1981; Trovão et al., 2007). Although its richness and uniqueness are recognized nowadays, the Caatinga is still one of the least studied biomes and, consequently, one of the most neglected in the literature, along with other tropical dry forests of the world (Tabarelli & Vicente, 2004; Santos et al., 2011; Sunderland et al., 2015). The presence of phytotelmata in this biome suggests that these habitats play a role in the survival dynamics of culicid species adapted to this semi-arid environment. However, there are no systematic studies investigating the diversity and dynamics of culicid that colonize phytotelmata in the Caatinga, nor are there any ecological studies to help understand the behavioral and evolutionary adaptations made by the species based on environmental selective pressures. As such, this study was conducted to investigate both the biodiversity and

4

the population dynamics of the culicid fauna that colonize tank bromeliads and tree holes in the Caatinga biome.

2. Material and methods

2.1 Study area The study was conducted at the Ecological Station (ESEC) Raso da Catarina (38º44'00" W to 39º29'20" W and 9º33'13" S to 9º54'30" S), Conservation Unit (CU) of the Caatinga biome located in Bahia, northeastern Brazil (Fig. 1). The CU is classified as a strictly protected area (IUCN, Category 1a), with approximately 99,772ha and a 135,279m perimeter (Paes & Dias, 2008). Raso da Catarina is the oldest CU in Bahia, the state with the largest area - 30,092,536ha - of the Caatinga biome in Brazil (Hauff, 2008). The ESEC Raso da Catarina is located at low tropical latitudes and has a hot semi-arid climate, characterized by the predominance of high temperatures combined with strong insolation and low cloudiness. The temperature range undergoes pronounced daily fluctuations (around 12 °C) in the warmer months of the year. Rainfall is sparse, irregular, heavy and poorly distributed in space and time throughout the year. There is a long dry season punctuated by short rainy periods, reaching annual rainfall averages ranging between 300 mm and 500 mm (Almeida & Figueroa, 1983). The water found on the surface is scarce and certain parts of streams and creeks are intermittent, resulting in a high water deficit. The landscape is marked by vegetation on sandy soil, predominantly shrubby, very dense and less thorny than that observed in the Caatinga overlying crystalline soils (Velloso et al., 2002). There are two areas with different vegetation types within the CU. The first type covers almost all the CU, with branched and thorny shrub-tree vegetation consisting mainly of Euphorbiaceae, Bromeliaceae and Cactaceae. This area is marked by deciduity in the dry season, with tree species that reach an average of 5m in height, and the predominance of shrubs of less than 3m in height. The second environment is a forest area called Mata da Pororoca, represented by a natural area of isolated forest that occupies 29 ha of the Station, or 0.03% of its total area and has a width between 200 and 400 m. The forest area is not associated with any drainage or watercourse, has trees that sometimes reach as much as 15m in height, has a closed understorey, dry shrubs and the ground is supposedly more fertile than that of its surroundings. Epiphytic species of the Bromeliaceae and Orchidaceae families are found in the forest area. There is also the presence

5

of litter and a large number of lianas. This environment differs both in physiognomy and flora from other areas of the ESEC (Paes & Dias, 2008).

2.2 Collection, maintenance and identification of culicids Immature samples were collected over 19 consecutive months, from March 2013 to September 2014. In each collection, seven tree holes and 40 bromeliads – 30 terrestrial and 10 epiphytes – were investigated. The bromeliads surveyed were close to established trails and roads within the CU and were divided into three areas located in different transects: terrestrial bromeliads present along a trail, in an area with shrub-tree vegetation and sandy soil, from which 20 bromeliads were haphazardly sampled; a cluster of epiphytic bromeliads, also in the same shrub-tree vegetation area and at a distance of approximately 800 meters from the previous cluster, from which 10 bromeliads were haphazardly sampled; forest area, at a distance of about 25 km from the previous locations, from which 10 epiphytic bromeliads were haphazardly sampled. Epiphytic bromeliads sampled ranged from 50 cm to 6 m height from the soil. Of the seven tree holes investigated, six were in the forest area, in a shrub-tree Caatinga area where larger trees are scarce. Other holes were monitored but since they were permanently dry, they were excluded from the study. In each survey, 40 sampled bromeliads were randomly selected, with a total of 760 samples at the end of the study. The seven sampled tree holes were the same in all larval surveys, totaling 133 investigations. All examined bromeliads, both terrestrial and epiphytic, belonged to Aechmea aquilega, abundant throughout the CU region. Trees that had phytotelmata habitats belonged to the species commonly known as Angico-manjola (Parapiptadenia zehntneri), Guava-bravo (Psidium myrsinites) and Jaqueira-brava (Richeria grandis). The water in the central well and the additional cavities formed by the outer axial leaves of each bromeliad and in tree holes was drawn using a manual suction pump (700 mL). The aspirated content was measured and carefully transferred into plastic basins to facilitate the capture of immature culicids. The larvae and pupae found were quantified and transferred into lidded plastic containers, along with water from the larval habitat, before being stored in a Styrofoam box for transportation to Laboratório de Entomologia e Parasitologia Tropical (LEPaT) of the Universidade Federal de Sergipe, where the immature forms were maintained. The fluid contents and imamatures from each sampled plant were maintained separately. The aspirated and unused water was immediately returned to the plants and rainwater was added to replace the volume removed.

6

Specimens collected as pupae were transferred into disposable cups and placed in cages until the emergence of adults. Larvae were kept in the same container with water from their larval habitat and fed with fish food (TetraMarine Saltwater Granules®; Tetra GmbH Herrenteich, 78 – 49324 Melle – Germany). Upon reaching the fourth stage of larval development, specimens were individualized for the emergence of adults and the collection of larval and pupal exuvia. The adults were killed with water at 60°C and preserved in 70% ethanol until mounted on a slide with coverslip, and were then identified at a species level or as morphologically similar species. To ensure reliable taxonomic identification of culicids, the associated sets of larval and pupal exuvia and male and female genitalia were examined. Samples of the species collected were sent to the Laboratório de Entomologia em Saúde Pública – LESP of the Faculdade de Saúde Pública da Universidade de São Paulo, to confirm their species identification and to provide samples of the species to the Entomology Reference Collection.

2.3 Data analysis The species were characterized by means of dominance and constancy indices. Dominance was defined as the total number of culicids collected in all samples using the formula D% = (i/t)*100 (i = number of individuals of a particular species and t = number of individuals captured). According to the value of D, five classifications were obtained: eudominant (above 10%), dominant (5 to 10%), subdominant (2 to 5%), eventual (1 to 2%) and rare (less than 1%). The constancy index was calculated with the formula C = (p/N)*100 (p = number of samples of a given species and N = number of samples tested). According to the value of C, species were classified into three categories: constant (C> 50%), accessory (25
7

was used to verify the change in species composition between samples (β- diversity) (Magurran, 2004) from bromeliads and tree holes and between epiphytic and terrestrial bromeliads, using the formula Ss = 2ab/(a+b) (a is the number of species found in one type of larval habitat; b is the number of species in the other and ab is the number of species shared by the two habitats). Species and individual accumulation curves were made using EstimateS 9.1 Software (Colwell, 2013). For this analysis, each plant was considered a sample and, for each of them, the fluid volume aspirated was verified. In addition, different phytotelmata were compared according to the variation in larval density per liter of water in positive samples. If the larval data density showed no evidence of normal distribution in the Shapiro-Wilk test (p <0.05), data was compared using the nonparametric Mann-Whitney and the probability of the test value obtained was verified based on the Monte Carlo randomness test. The power of the association between the presence and abundance of immature culicids and meteorological variables was investigated using generalized linear models (GLM's). Explanatory variables were: accumulated precipitation of 1-10, 11-20, 21-30 and 1 to 30 days pre-collection, monthly average temperature and relative humidity. Meteorological variable values were obtained from the meteorological database at the Instituto Nacional de Meteorologia (INMET), from the station in Paulo Afonso, Bahia (9º 36' S and 38º 21' W) which was the closest (approximately 40 Km) to the CU. A logistic regression model was used to evaluate the relationship between the proportion of positive larval habitats and meteorological variables. Response variables considered for this analysis were: 1) total number of bromeliads, 2) epiphytic bromeliads, 3) terrestrial bromeliads and 4) tree holes. If the data showed over-dispersion (variance greater than expected in the statistical model), quasi-binomial models were used for the correction of standard errors of the estimates. The relationship between metereological variables and the abundance of immature forms was analyzed through linear models with errors of the negative binomial type, which showed the best fit for the abundance data. The response variable was the number of immature per collection in: 1) total number of larval habitats, 2) epiphytic bromeliads, 3) terrestrial bromeliads and 4) tree holes. The effect of each explanatory variable on the dependent variable was tested, both alone in simple linear models or combined in multiple linear models. An information theoretic approach based on the Akaike Information Criteria corrected for small samples (AICc) was adopted for select the more plausible models. Models eith value of evidence ∆AICc ≤ 2 was considered those with more empirical support

8

(Burnham & Anderson, 2002). The explanatory power of each model was compared with McFadden's Pseudo R-squared (McFadden, 1974). All these analyses were prepared using the computing environment R (R Core Team, 2015).

3. Results

During the study period, 893 samples were collected in phytotelmata habitats, of which 760 were in bromeliads (570 terrestrial and 190 epiphytic) and 133 in tree holes. A total of 127 L of water was aspirated, from which 6,764 immatures were removed, of which 5,521 (82%) were identified (Table 1). The mosquito fauna was composed of 16 species in six genera: Aedes, Culex, Haemagogus, Runchomyia, Toxorhynchites and Wyeomyia (Table 2). 11 (69%) of the species found were unknown to science and correspond to 90% of the total abundance of specimens collected. Considering only the culicid fauna associated with bromeliad phytotelmata, the percentage of unknown species was 90%, corresponding to 99% of the total collected immature forms. The taxonomic units Cx. (Mcx.) Gr. Imitator, Wy. (Pho.) n. sp.1 / Wy. (Pho.) n. sp.2 and Wy. (Pho.) n. sp.1 / Wy. (Pho.) n. sp.2 / Wy. (Pho.) n. sp.3 were designated since they represent groups of species indistinguishable at the development stage available for identification. The Cx. (Mcx.) Gr. Imitator specimens were represented by 78 larvae indistinguishable to Cx. (Mcx.) Gr. Imitator n. sp.1 and Cx. (Mcx.) imitator Theobald, 1903, but distinguishable by morphological characteristics of adult and male genitalia in the associated specimens. Of the 16 species identified in this study, five (31%) – Cx. (Mcx.) Gr. Imitator n. sp.1, Cx. (Mcx.) xenophobus n. sp.1, Cx. (And.) n. sp.1, Wy. (Pho.) n. sp.1 and Wy. (Pho.) n. sp.2 – were considered dominant and constant. Therefore, each one represented 5 to 10% of the collected immatures and was present in over half of the samples. Of the species using tree holes as larval habitat, only Hg.(Con.) leucocelaenus was considered dominant, but accessory (Table 2). The monthly abundance distribution of these species suggests that despite their classification as dominant, considering the total immatures listed in the study, not all were dominant in all samples (Fig. 2). Epiphytic bromeliads had a higher larval density and lower proportion of dried plants compared to terrestrial bromeliads and tree holes. In general, bromeliads were most frequently

9

found harboring larvae although tree holes showed higher average volume of water per plant (Table 1). Immature abundance of culicid fauna in bromeliad and tree hole phytotelmata showed different distribution patterns (Spearman-rho test = 0.213, p= 0.381). In bromeliads, the months with a greater abundance of immatures tended to be those with the greatest richness, with records of all bromeliad species in the same collection (August 2014) occurring when the highest abundance of immatures was also recorded (Fig. 3). In addition, the presence of mosquitoes in bromeliads was maintained during the whole study period, except for the first three months of collection and in March 2014, one year after the research began. In contrast, in tree holes, months with the most abundant culicid fauna had a lower richness and none of the collections had concomitant records of all culicid species that use tree holes as larval habitat. Furthermore, the tree hole phytotelmata had peaks of immature abundance alternating with negative samples (Fig. 3). Of the phytotelmata habitats, a greater total volume of water (110L) and, consequently, a greater number of individuals (n = 6,161) was aspirated from those established in bromeliads compared to the total amount of aspirated water 17L and collected immature (n = 603) in tree holes. However, by eliminating differences in sampling and analyzing the water volume and number of individuals (Fig. 4), it was possible to observe that if the same volume of water was removed from both types of phytotelmata, bromeliads would have a higher richness than tree holes (Fig. 4A). The abundance of immature was similar in both environments (Fig. 4B). Moreover, if we consider the same number of individuals, the highest species richness is still found in bromeliads (Fig. 4C). Results from a comparative analysis of epiphytic and terrestrial bromeliads (Fig. 5) indicate that the former contrasts out compared to the latter since it has a higher species richness (Fig. 5A) and a greater abundance of immatures (Fig. 5B), when considering the same water volume in both environments. In terrestrial bromeliads, the richness is higher when a lower abundance of culicids is considered, but with the increasing number of individuals, the number of species becomes larger in epiphytic bromeliads (Fig. 5C). Data from the larval density per liter of extracted water shows no significant difference between densities recorded in bromeliad habitats and tree holes (U-value = 8125, n1=230, n2=39, Monte Carlo p=0.3049). However, the distribution of larval density in bromeliads, for some samples, had higher values when compared to the density distribution in tree holes. In contrast, when considering the data from the larval density in epiphytic and terrestrial

10

bromeliads, larval habitats in epiphytic bromeliads had a higher larval density compared to those established in terrestrial bromeliads (U-value = 14890, n1=145, n2=317, Monte Carlo p=0.0001). There was no sharing of species between bromeliad phytotelmata habitats and tree holes (Ss = 0). However, the similarity in the composition of culicid species was high (Ss = 0.92) between epiphytic bromeliads habitat and terrestrial bromeliads. Only one species – Toxorhynchites n. sp.1 – did not occur in both phytotelmata, but was restricted to epiphytic bromeliad habitats. Although the Wy. (Pho.) n. sp.1 and Wy. (Pho.) n. sp.2 species were present in both types of bromeliad, the first was predominantly (98%) in epiphytes, while 94% of the second specimens was collected from terrestrial bromeliads. Amongst the meteorological variables, accumulated rainfall was the most irregular parameter throughout the year. In contrast, the monthly average temperature was the most stable variable, with a mean amplitude of 5.6 °C throughout the study (although daily temperatures reached a minimum and maximum of 16.2 °C and 38.2 °C, respectively). The mean monthly relative humidity was the parameter with the most regular distribution pattern, with the highest values of over 70% humidity recorded in May-August 2013 and 2014. After the first three months with no culicid observations, the monthly abundance of immatures rose for six consecutive months, after which it began to decline to a minimum abundance again in March. In the following year, the data of immature abundance had irregular monthly distribution (Fig. 6). The average monthly temperature was the meteorological variable most related to the presence of mosquitoes in phytotelmata larval habitat (Table 3). An analysis of the relationship between meteorological variables and the proportion of positive larval habitats pointed more empirical support to an association between bromeliads (both epiphytic and terrestrial) and average monthly temperatures, with lower AICc values and and good explanatory power for the models while observing the Pseudo-R² values. The other explanatory variables showed no statistically significant association. Tree holes models had significant association and more empirical support for average monthly temperatures, monthly relative humidity and accumulated rainfall between 1-10 days prior to collection also presented statistically significant association. All models showed relatively less explanatory power than that observed in bromeliads and monthly average temperatures models (Table 3). Abundance data show little or no association with meteorological variables, except between temperature and abundance of immature terrestrial bromeliads that, nevertheless, holds low explanatory power.

11

The analysis of multiple variables showed that the Minimum Adequate Model for the proportion of positive larval habitats in total bromeliads, while epiphytic bromeliads and tree holes only took monthly average temperature into consideration. In the case of epiphytic bromeliads, the combined temperature and humidity variables gave better explanatory power to the model. Furthermore, analyses showed that the Minimal Adequate Model is the null model (model with only the overall average) for explaining the abundance of immature found in epiphytic bromeliads and tree holes. This result showed that none of these explanatory variables, when taken from the full model simulations, caused statistically significant worsening in the model. For immatures abundance in terrestrial bromeliads, the temperature and humidity variables were kept in the model; however, the low value of Pseudo-R² indicated low explanatory power. Among all the phytotelmata, only the terrestrial bromeliads were associated, albeit with slight adjustment, with the abundance of immature data and the volume of water stored in the plant. As for the meteorological variables, the volume of water was not associated with precipitation values, but was negatively associated with temperature and positively associated with humidity.

4. Discussion

In this study, 17.5% of the immature collected were Cx. (Mcx.) and 66.5% were from the Sabethini tribe. All were collected from bromeliads (Table 2). This species group is common in surveys of culicid fauna that inhabit bromeliads in other biomes, especially the Atlantic Forest (Silva et al., 2004; Müller & Marcondes, 2006; Marques & Forattini, 2008; Mocellin et al., 2009; Marques et al., 2012; Cardoso et al., 2015; Ceretti-Junior et al., 2016) and the Amazon forest (Torreias et al., 2010). As for the culicid fauna found inhabiting tree holes, previous entomological surveys show a similar species diversity, especially with records of species from Aedini and Culicini tribes, and fewer predatory larvae from the Toxorhynchitini tribe (Fincke et al., 1997; Silva & Lozovei, 1999; Alencar et al., 2010; Medeiros-Sousa et al., 2015). In this study the distinction between closely related species was possible due to the combination of immature and adult stages that helped highlight differences, especially regarding morphological characteristics of male genitalia. In ESEC Raso da Catarina, dominant and constant culicid species were present in all samples and were solely responsible for more than half the collected immature (Table 2). This observation demonstrates the evolutionary success of these species that are highly adapted to

12

surviving in an environment with climatic characteristics that limit the development of culicid fauna. Water storage in the Caatinga phytotelmata is a challenge, especially with regards to the scarce and irregular rainfall, both temporally and spatially, and the associated high temperatures and low humidity rates resulting in prolonged dry periods. Consequently, those species whose development depends on stored water are likely to have survival strategies, selected during the evolutionary process, enabling them to overcome this adverse environmental ecosystem. In adverse environments such as the Caatinga, it is necessary to optimize the available resources, especially water stored in phytotelmata for the development of culicids. However, no species of culicids were observed to show any opportunistic behavior in colonizing both bromeliad habitats and tree holes as a survival strategy (Table 2). The occupation of welldefined and distinct ecological niches decreases competition for available resources in the habitat. Thus, avoiding competition for scarce resources seems to be advantageous for culicid communities in bromeliads and tree holes. A similar situation was observed in temperate rainforest areas of New Zealand. In these regions, tree holes, even irregularly distributed and often dry, have been characterized as ideal habitats for maintaining specialised aquatic invertebrate fauna which evolved to occupy these unique micro-environments and differ from the fauna found in artificial containers, analogous to the tree holes, installed in the natural environment (Blakely et al., 2012). With regards to the two dominant species in the Caatinga phytotelmata habitats – Wy. (Pho.) n. sp.1 and Wy. (Pho.) n. sp.2 – distinct patterns of abundance were observed. Although both species have a similar total abundance, Wy. (Pho.) n. sp.1 seems to be common most of the time, while Wy. (Pho.) n. sp.2 tends to be less abundant, with short peaks of high abundance (Fig. 2). Although both species are found in both types of bromeliad, Wy. (Pho.) n. sp.1 predominated (98%) in the epiphytic, while 94% of the Wy. (Pho.) n. sp.2 were collected in terrestrial bromeliads. Epiphytic bromeliads were restricted to the forest area. In this location, the effects of climate aridity are less evident and the presence of larger trees, compared with the shrub-tree vegetation in the surroundings, establishes a canopy that allows the formation of shaded areas. Therefore, it is likely that in this environment the water in bromeliads is stored for longer than in terrestrial bromeliads, enabling the development of immature culicids and maintaining them at a relatively constant number most of the time. In terrestrial bromeliads, distributed in areas of shrub-tree vegetation, the effects of drought tend to be more intense, with less constant water storage and specific times at which conditions become more favorable. In

13

favorable conditions, punctual peaks in the abundance of Wy. (Pho.) n. sp.2 occur, generally corresponding to twice the number of individuals recorded in the peak abundance of Wy. (Pho.) n. sp.1 (Fig. 2). The characteristics of the environment where the two types of bromeliad are found not only explain the distribution patterns of the dominant species, but also clarify that epiphytic bromeliads provide the most productive larval habitats in comparison with other types of phytotelmata investigated (Table 1). Studies have shown that environmental and metereological conditions where the bromeliad species are established may affect the abundance and composition of macrofauna communities associated with it (Kitching, 2000; Lopez & Iglesias Rios, 2001; Montero et al., 2010). Furthermore, phenological characteristics of bromeliad species (Guimarães-Souza et al., 2006), as well as biological and physiological aspects of the plants, can interfere with the metabolism of the water in their wells, reflecting differences in the production of immature culicids in the Caatinga bioma (Fig. 3). Perhaps the bromeliads´ ability to store water during prolonged dry periods results not only from fluctuations in rainfall, but also from some intrinsic mechanism that allows them, for example, to collect water from moisture in small amounts but enough to keep the wells filled. It is plausible that, if rainwater were the only source of water, the wells would dry quickly, similarly to artificial larval habitats or even tree holes, which tend to be more ephemeral. In this study, the epiphytic bromeliads, despite having less water (Table 1), kept their water volume for longer than the tree holes, thus providing higher levels of moisture in phytotelmata. Despite differences in sampling, it was possible to compare habitats through an analysis of the number of individuals (Fig. 4 and fig. 5). Differences between sample sizes followed the pattern of phytotelmata abundance observed in the CU, where there were fewer epiphytic bromeliads and tree holes, thus hampering the extent of sampling possible, while the terrestrial bromeliads were abundant and widely dispersed throughout the study area. Observation of the general average values analyzed by number of individuals showed that epiphytic bromeliads exhibit more favorable conditions for the formation and maintenance of larval habitats for culicids. Compared to other phytotelmata, epiphytic bromeliads had a lower percentage of dried plants and plants with water, but without any immature forms, in addition to presenting a proportionally greater number of larval habitats and a higher overall average of immature per larval habitat (Table 1). The combined analysis of water volume and abundance of immature, measured as the larval density per liter of stored water, showed that compared to the terrestrial, the epiphytic bromeliads held a higher density

14

of immature (Fig. 5). In addition, if the water volume and number of immatures were the same, bromeliads would have held higher species richness compared with tree holes (Fig. 4). However, while the epiphytic bromeliads appear to be more efficient for the maintenance of the immature, there are few of them and they are restricted to the forest area that occupies only 0.03% of the total area of the CU. Thus, the terrestrial bromeliads, abundant and widely distributed in the region, are suggested as the main larval habitat for culicids at ESEC Raso da Catarina, given that the lower productivity of these bromeliads can be compensated for by their high density in the region. Epiphytic bromeliads and tree holes were distributed in the same forest environment, except for a tree hole located in the shrub-tree vegetation area. Therefore, in general, tree holes provide less arid climatic conditions. Despite large stored volumes of water per plant (maximum registration of 2,400 mL aspirate in a sample), tree holes had lower richness and larval density per liter of water stored, with 56% of the samples dry. The bromeliads´ maximum volume per plant was 890 mL. Due to similar results in other studies tree holes have been considered larval habitats of reduced production of culicids when compared with other types of larval habitats of the same entomofauna (Bonnet & Chapman, 1956). In our study, however, despite lower richness and larval density in comparison with bromeliad, tree holes were used as larval habitats for Hg. (Con.) leucocelaenus, the dominant species in the region and vector of the jungle yellow fever. This corroborates the epidemiological importance of these natural larval habitats in the maintenance of zoonotic arbovirus cycles. The sharing of culicid species between established habitats in terrestrial and epiphytic bromeliads was not complete due to a single species – Toxorhynchites n. sp.1 – which was restricted to epiphytic plants. However, there was no similarity in fauna composition between bromeliad habitats and tree holes. These results suggest that the culicid communities associated with bromeliads and tree holes are particular. This characteristic can probably be explained by a process involving evolutionary patterns of adaptation and the selection of species with different biologies and ecologies that depend on biotic and abiotic factors specific to each phytotelmata group. Meteorological variables, in general, were not associated with the abundance of immature in the phytotelmata, but with the proportion of positive larval habitats, especially for temperature and humidity data. The rainfall data was the least adjusted to the variable responses investigated (Table 3). However, it is possible that the data used in the analysis does not accurately reflect the conditions of the study area. Although data on metereological

15

variables is representative of the region in which the ESEC Raso da Catarina is established, especially temperature and humidity data, it may not correspond to variations that occur within the CU, especially for precipitation. The nearest metereological station area is located outside the unit areas and the Brazilian semiarid region is marked by an accentuated variability in spatiotemporal rainfall (Correia et al., 2012). This suspicion is reinforced by the lack of association of rainfall data to the total volume of water stored in phytotelmata, although temperature and humidity were associated. Data suggests that, regardless of the amount of precipitation, lower temperatures and higher humidity levels ensure the maintenance of the water volume in phytotelmata by promoting the formation of larval habitats and increased abundance of immature. In addition, the volume of stored water was the only parameter associated with the abundance of immature in terrestrial bromeliads. The proportion of larval habitats for culicids established in terrestrial bromeliads also presented the best adjustment to meteorological variables data. Since they are numerous and widely distributed throughout the length of the CU, terrestrial bromeliads are the phytotelmata that can best respond to general abiotic conditions in the region. In this study, the surprising number of species unknown to science and probably typical of the Caatinga, recorded in unusually high amounts even for entomological surveys in general, indicates a high degree of endemism in the biome. Among the 11 new species for science cited in this study are the seven unknown species already found in adult stage survey in the study area (Marteis et al., 2017). Considering the fauna exclusively associated to bromeliad phytotelmata, the number of unknown species represented 90% of the total richness and 99% of the total abundance of individuals, with only one species - Cx. (Mcx.) imitator – already described. Four other named species – Hg. (Con.) leucocelaenus, Ae. (Pro.) terrens, Ae. (How.) fulvithorax e Cx. (And.) conservator – were recorded in tree holes and are commonly found in natural larval habitats of other Brazilian biomes, where environmental conditions are very different from those observed in the Caatinga. Amongst the unknown species probably endemic to the Caatinga, two of them – Toxorhynchites n. sp.1, which occurs exclusively in the forest area, and Wy. (Pho.) n. sp.1 which is a dominant species in the forest area – are characteristic of specific areas within the biome in which metereological conditions are milder, restricting their distribution area. Thus the Caatinga degradation process can affect these species more intensely, leading to their extinction and that of many others that may disappear before they are known to science. This

16

reinforces the importance of studies on the diversity of culicids in the Caatinga, given the lack of culicid fauna surveys in the Brazilian semiarid region. This study exposes how the culicid fauna from the Caatinga is ignored and presents an unprecedented contribution to the knowledge of the culicid fauna diversity whilst highlighting the urgency of prioritizing the conservation of dry forest biodiversity through the delimitation and efficient management of strictly protected areas.

Acknowledgements

Authors thank Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Fundação de Apoio à Pesquisa e Inovação Tecnológica do Estado de Sergipe (FAPITEC), Edital 47/2010 SISBIOTA, for their financial support; Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio/SISBIO) for the permit and logistic support during the scientific activities in ESEC Raso da Catarina; Herbário ASE from Universidade Federal de Sergipe for the identification of bromeliad specimens; Aristides Fernandes for the culicids identification; Lydia Banfield for comments and suggestions on the language, the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for granting the doctoral fellowship to Letícia Silva Marteis and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for granting the doctoral fellowship to Antônio Ralph Medeiros de Sousa.

References

Alencar, J., Gil-Santana, H.R., Oliveira, R.F.N., Dégallier, N. & Guimarães, A.E. (2010) Natural Breeding Sites for Haemagogus Mosquitoes (Diptera, Culicidae) in Brazil. Entomological News, 121, 393-396. Almeida, M.C.B. & Figueroa, L.A. (1983) Estação Ecológica Raso da Catarina – Bahia. Subprojeto: estudos geomorfológicos. Relatório de Pesquisa do Convênio Salvador: Sema/Minter/UFBA, Bahia. Andrade-Lima, D. (1981) The Caatinga dominium. Revista Brasileira de Botânica, 4, 149153. Blakely, T.J., Harding, J.S. & Didham, R.K. (2012) Distinctive aquatic assemblages in water‐filled tree holes: a novel component of freshwater biodiversity in New Zealand temperate rainforests. Insect Conservation and Diversity, 5, 202-212.

17

Bonnet, D.D. & Chapman, H. (1956) The importance of mosquito breeding in tree holes with special reference to the problem in Tahiti. Mosquito News, 16, 301-305. Burnham, K. P. & Anderson, D. R. (2002) Model selection and multimodel inference: a practical information-theoretic approach. Springer Science & Business Media. (2002). Cardoso, C.A.A., Lourenço-de-Oliveira, R., Codeço, C.T. & Motta, M.A. (2015) Mosquitoes in Bromeliads at Ground Level of the Brazilian Atlantic Forest: the Relationship Between Mosquito Fauna, Water Volume, and Plant Type. Annals of the Entomological Society of America, sav040. Ceretti-Junior, W., Christe, R.O., Rizzo, M., Strobel, R.C., Matos Junior, M.O., Mello, M.H.S.H., Fernandes, A., Medeiros-Sousa, A.R., Carvalho, G.C. & Marrelli, M.T. (2016) Species Composition and Ecological Aspects of Immature Mosquitoes (Diptera: Culicidae) in Bromeliads in Urban Parks in the City of São Paulo, Brazil. Journal of Arthropod-Borne Diseases, 10, 102–112. Colwell, R.K. (2013) EstimateS: Statistical Estimation of Species Richness and Shared Species from Samples. Version 9 and earlier 2011.User’s guide and application. . 19 out 2015. Correia, R.C., Kiill, L.H.P., Moura, M.S.B., Cunha, T.J.F., Jesus Júnior, L A.& Araújo, J.L.P. (2012) A região Semiárida brasileira. TV Voltolini, Produção de Caprinos e Ovinos no Semiárido. Fincke, O.M., Yanoviak, S.P. & Hanschu, R.D. (1997) Predation by odonates depresses mosquito abundance in water-filled tree holes in Panama. Oecologia, 112, 244-253. Fish, D. (1983) Phytotelmata: flora and fauna. Phytotelmata: terrestrial plants as hosts for aquatic insect communities. Plexus, New Jersey. Gomes, A.C., Forattini, O.P. & Kakitani; I. (1992) Microhabitats de Aedes albopictus (Skuse) na região do Vale do Paraíba, estado de São Paulo, Brasil. Revista Saúde Pública, 26: 108-118. Guimarães-Souza, B.A., Mendes, G.B., Bento, L., Marotta, H., Santoro, A.L., Esteves, F.A., Pinho, L.; Farjalla, V.F & Enrich-Prast, A. (2006) Limnological parameters in the water accumulated in tropical bromeliads. Acta Limnologica Brasiliensia, 18, 47-53. Hauff, S.N. (2008) Unidades de conservação e terras indígenas do bioma Caatinga. The Nature

Conservance

e

Ministério

do

Meio

Ambiente.

. 4 dez 2015. Jenkins, D.W. & Carpenter, S.J. (1946) Ecology of the tree hole breeding mosquitoes of Neartic North America. Ecological Monographs, 16, 31–47.

18

Kitching, R.L. (2000) Food webs and container habitats: the natural history and ecology of phytotelmata. Cambridge University, UK. Kitching R.L. (2001) Food webs in phytotelmata: “bottom-up” and “top-down” explanations for community structure. Annual Review of Entomology, 46, 729-760. Lopez, L.C.S. & Iglesias Rios, R. (2001) Phytotelmata community distribution in tanks of shaded and sun-exposed terrestrial bromeliads from restinga vegetation. Selbyana, 22, 219–224. Magurran, A.E. (2004) Measuring biological diversity. Oxford, UK. Marques, G.R.A.M. & Forattini, O.P. (2008) Culicídeos em bromélias: diversidade de fauna segundo influência antrópica, litoral de São Paulo. Revista Saúde Pública, 42, 979-985. Marques, T.C., Bourke, B.P., Laporta, G.Z. & Sallum, M.A.M. (2012) Mosquito (Diptera: Culicidae) assemblages associated with Nidularium and Vriesea bromeliads in Serra do Mar, Atlantic Forest, Brazil. Parasites & vectors, 5, 1-9. Marteis, L.S., Sallum, M.A.M., Natal, D., Oliveira, T.M.P., Gama, R.A., Dolabella, S.S. & Santos, R.L.C. (2015) First Record of Anopheles oryzalimnetes, Anopheles argyritarsis, and Anopheles sawyeri (Diptera: Culicidae) in the Caatinga Biome, Semiarid Scrubland of Sergipe State, Brazil. Journal of Medical Entomology, 52, 858–865. Marteis, L.S., Natal, D., Sallum, M.A.M., Medeiros-Sousa, A.R., Oliveira, T.M.P. & La Corte, R. (2017) Mosquitoes of the Caatinga: 1. Adults stage survey and the emerge of seven news species endemic of a dry tropical forest in Brazil. Acta Tropica, 166, 193201. McFadden, D. (1974) Conditional logit analysis of qualitative choice behavior. P. Zarembka (Ed.). Frontiers in econometrics. Academic Press. New York, NY, 105-142. Medeiros-Sousa, A.R., Ceretti-Júnior, W., Carvalho, G.C., Nardi, M.S., Araujo, A.B., Vendrami, D.P. & Marrelli, M.T. (2015) Diversity and abundance of mosquitoes (Diptera:Culicidae) in an urban park: Larval habitats and temporal variation. Acta Tropica, 150, 200–209. Mocellin, M.G., Simões, T.C., Nascimento, T.F.S., Teixeira, M.L.F., Lounibos, L.P. & Oliveira, R.L.D. (2009) Bromeliad-inhabiting mosquitoes in an urban botanical garden of dengue endemic Rio de Janeiro-Are bromeliads productive habitats for the invasive vectors Aedes aegypti and Aedes albopictus?. Memórias do Instituto Oswaldo Cruz, 104, 1171-1176. Mogi,

M.

(2004)

Phytotelmata:

hidden

freshwater

habitats

faunas. Freshwater invertebrates of the Malaysian region, 13-22.

supporting

unique

19

Montero, G., Feruglio, C. & Barberis, I.M. (2010) The phytotelmata and foliage macrofauna assemblages of a bromeliad species in different habitats and seasons. Insect Conservation and Diversity, 3, 92-102. Müller, G.A. & Marcondes, C.B. (2006) Bromeliad-associated mosquitoes from Atlantic forest in Santa Catarina Island, southern Brazil (Diptera, Culicidae), with new records for the State of Santa Catarina. Iheringia, 96, 315-319. Ngai, J.T. & Srivastava, D. (2006). Predators accelerate nutrient cycling in a bromeliad ecosystem. Science, 314, 963. Paes, M.L.N. & Dias, I.F.O. (2008) Plano de manejo: Estação Ecológica Raso da Catarina. Brasília: Ibama. R Core Team. (2015) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 20 nov 2015. Rebêlo, J.M.M., Moraes, J.L.P., Alves, G.A., Leonardo, F.S., Rocha, R.V., Mendes, W.A., Costa, E., Câmara, L.E.M.B., Silva, M.J.A., Pereira, Y.N.O. & Mendonça, J.A.C. (2007) Distribuição das espécies do gênero Anopheles (Diptera, Culicidae) no Estado do Maranhão, Brasil. Cadernos de Saúde Pública, 23, 2959-2971. Santos, C.F., Silva, A.C., Rodrigues, R.A., Jesus, J.S. & Borges, M.A.Z. (2015) Inventory of mosquitoes (Diptera: Culicidae) in conservation units in brazilian tropical dry forests. Revista do Instituto de Medicina Tropical de São Paulo, 57, 227-232. Santos, J.C., Leal, I.R., Almeida-Cortez, J.S., Fernandes, G.W. & Tabarelli, M. (2011) Caatinga: the scientific negligence experienced by a dry tropical florest. Tropical Conservation Science, 4, 276-86. Silva, A.D., Nunes, V. & Lopes, J. (2004) Culicídeos associados a entrenós de bambu e bromélias, com ênfase em Aedes (Stegomyia) albopictus (Diptera, Culicidae) na Mata Atlântica, Paraná, Brasil. Iheringia, 94, 63-66. Silva, J.M.C. & Dinnout, A. (1999) Análise da representatividade das unidades de conservação federais de uso indireto na Floresta Atlântica e Campos Sulinos. In: Pinto, L.P. (coord.) Padrões de biodiversidade da Mata Atlântica do Sul e Sudeste. Campinas, São Paulo. Silva, M.D. & Lozovei, A.L. (1999) Ocorrência de Haemagogus (Conopostegus) leucocelaenus (Dyar & Shannon) e Toxorhynchites (Lynchiella) theobaldi (Dyar & Knab) em ocos de árvore em capão de mata, Curitiba, Paraná, Brasil. Revista Brasileira de Zoologia, 16, 257-267.

20

Silveira Neto, S., Nakano, O., Barbin, D. & Villa Nova, N.A. (1976) Manual de ecologia dos insetos. Ceres, Piracicaba. Sunderland, T., Apgaua, D., Baldauf, C., Blackie, R., Colfer, C., Cunningham, A.B., Dexter, K., Djoudi, H., Gautier, D., Gumbo, D., Ickowitz, A., Kassa, H., Parthasarathy N., Pennington, R.T., Paumgarten, F., Pulla, S., Sola, P., Tng, D., Waeber P. &Wilmé, L. (2015) Global dryforests: a prologue. International Forestry Review, 17, 1-9. Tabarelli, M. & Vicente, A. (2004) Conhecimento sobre plantas lenhosas da Caatinga: lacunas geográficas e ecológicas. In: Silva, J.M.C., Tabarelli, M., Fonseca, M.T. & Lins, L.V. (orgs.). Biodiversidade da Caatinga: áreas e ações prioritárias para a conservação. Torreias, S.R.S., Ferreira-Keppler, R.L., Godoy, B.S. & Hamada, N. (2010) Mosquitoes (Diptera,

Culicidae)

inhabiting

foliar

tanks

of

Guzmania

brasiliensis

Ule

(Bromeliaceae) in central Amazonia, Brazil. Revista Brasileira de Entomologia, 54, 618-623. Trovão, D.M.B.M., Fernandez, P.D., Andrade, L,A, & Neto, J.D. (2007) Variações sazonais de aspectos fisiológicos de espécies da Caatinga. Revista Brasileira de Engenharia Agrícola e Ambiental, 11, 307-311. Velloso, A.L., Sampaio, E.V.S.B & Pareyn, F.G.S. (2002) Ecorregiões: propostas para o bioma Caatinga. Instituto de Conservação Ambiental, Recife. Walker, E.D., Lawson, D.L., Merritt, R.W., Morgan, W.T. & Klug, M.J. (1991) Nutrient dynamics, bacterial populations, and mosquito productivity in tree hole ecosystems and microcosms. Ecology 72, 1529-1546. WRBU (2015) Walter Reed Biosystematics Unit. Systematic catalog of Culicidae. Washington, USA. . 4 out 2015. Zanella, F.C.V. & Martins, C.F. (2003) Abelhas da caatinga: biogeografia, ecologia e conservação. In: Leal, I.R., Tabarelli, M., Silva, J.M.C. Ecologia e Conservação da Caatinga. Ed. Universitária da EFPE, Recife.

21

Table 1. Characterization of the types of phytotelmata habitats associated with the mosquito fauna, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

Table 2. Accumulated species of immature mosquitoes collected in bromeliads and tree holes, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

Table 3. Logistic regression data for the relationship between meteorological variables and proportion of positive larval habitat mosquitoes, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

Fig. 1. Location of ESEC Raso da Catarina, Paulo Afonso, Bahia, Brazil.

Fig. 2. Monthly distribution of abundance of the dominant species (5
Fig. 3. Monthly distribution of abundance and richness species of collected mosquitoes in bromeliads (A) and tree holes (B), ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

Fig. 4. Species richness (A) and individuals number (B) by water accumulated volume and species richness by individuals number (C) in bromeliads and tree holes, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

Fig. 5. Species richness (A) and individuals number (B) by water accumulated volume and species richness by individuals number (C) in epiphytic and terrestrial bromeliads, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

Fig. 6. Monthly abundance of mosquitoes and data on recorded weather variables during 30 days pre-collection, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014.

22

Table 1. Characterization of the types of phytotelmata habitats associated with the mosquito fauna, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014. Terrestrial bromeliads 570 136 (24%) 434 (76%) 117 (27%) 317 (73%) 81 24 (30%) 57 (70%) 187 (6-1,000) 181 (153) 3,278 2,431 (74%) 173 (170) 10 (13) 40 57 9

Epiphytic bromeliads Tree holes 190 133 32 (17%) 74 (56%) 158 (83%) 59 (44%) 13 (8%) 20 (34%) 145 (92%) 39 (66%) 29 17 3 (10%) 4 (23%) 26 (90%) 13 (77%) 181 (10-890) 303 (5-2,400) 178 (144) 365 (675) 2,883 603 2,553 (88%) 537 (89%) 152 (112) 32 (41) 20 (20) 15 (18) 101 36 111 46 10 6

Investigated plants (n) Dry plants (n) Plants with water (n) Plants with water and without immature (n) Positive plants (n) Total water aspirated volume (L) Total water aspirated volume per plant without immature (L) Total water aspirated volume per positive plant (L) Average volume water per plant with water (mL) (min-max) Average volume water per positive plant (mL)* Total abundance collected immature (n) Total abundance identified immature (n) Monthly average abundance collected immature* Average abundance immature per positive plants (n)* Larval density / total liter water Larval density / liter water positive plant Species richness (S)** * In brackets the standard deviation; ** In brackets estimated species based on the samples with 95%IC.

Total 893 242 (27%) 651 (73%) 150 (23%) 501 (77%) 127 31 (24%) 96 (76%) 196 (5-2,400) 193 (235) 6,764 5,521 (82%) 356 (246) 14 (17) 53 70 16

Table 2. Accumulated species of immature mosquitoes collected in bromeliads and tree holes, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014. Taxonomic units

n (%)

Dominance

Constancy

Larval habitat Bromediad Bromeliad

Tree

23

terrestrial Aedini (n=516; 9,3%; S=4) Haemagogus (Con.) leucocelaenus Dyar & Shannon, 1924 Aedes (Pro.) terrens Walker, 1856 Haemagogus (Hag.) spegazzinii n. sp.1 Aedes (How.) fulvithorax Lutz, 1904 Culicini (n=1.279; 23,2%; S=5) Culex (Mcx.) Gr. Imitator n. sp.1 Culex (Mcx.) xenophobus n. sp.1 Culex (And.) n. sp.1 Culex (Mcx.) imitator Theobald, 1903 Culex (And.) conservator Dyar & Knab, 1906 Culex (Mcx.) Gr. Imitator Sabethini (n=3.672; 66,5%; S=5) Wyeomyia (Pho.) n. sp.1 Wyeomyia (Pho.) n. sp.2 Runchomyia n. sp.1 Wyeomyia n. sp.1 Wyeomyia (Pho.) n. sp.3 Wy. (Pho.) n. sp.1/Wy. (Pho.) n. sp.2 Wy. (Pho.) n. sp.1/Wy. (Pho.) n. sp.2/Wy. (Pho.) n. sp.3 Toxorhynchitini (n=54; 1,0%; S=2) Toxorhynchites n. sp.1 Toxorhynchites n. sp.2 Total

epiphytic

297 (5.4) 175 (3.2) 38 (0.7) 6 (0.1)

dominant subdominant rare rare

accessory accessory accessory accidental

476 (8.6) 379 (6.9) 308 (5.6) 35 (0.6) 3 (0.1) 78 (1.4)

dominant dominant dominant rare rare eventual

constant constant constant accessory accidental constant

x x x x x

x

429 (7.8) 378 (6.8) 18 (0.3) 12 (0.2) 11 (0.2) 2,163 (39.2) 661 (12.0)

dominant dominant rare rare rare eudominant eudominant

constant constant accidental accidental accidental constant constant

x x x x x x x

x x x x x x x

36 (0.7) 18 (0.3) 5,521 (100.0)

rare rare

accessory accessory

holes

x x x x

x x x x x

x x

Table 3. Logistic regression data for the relationship between meteorological variables and proportion of positive larval habitat mosquitoes, ESEC Raso da Catarina, Paulo Afonso, Bahia, from March 2013 to September 2014. Positi ve larval habita t phytot elmat a

Response variables

Explanatory variable average temperature relative humidity

Intercept 11.835 (±3.867) -3.581 ( ±2.334)

Slope -0.459 (±0.148) 0.052 (±0.035)

p AICc ∆AICc Pseudo-R² 0 0.295 0.006 313.0 0.084 0.154 405.6 92.6

Positive tree holes

Positive terrestrial bromeliads

Positive epiphytic bromeliads

24

accumulated precipitation 1 - 30 accumulated precipitation 1 - 10 accumulated precipitation 11 - 20 accumulated precipitation 21 - 30 null model average temperature relative humidity null model accumulated precipitation 21 - 30 accumulated precipitation 1 - 30 accumulated precipitation 11 - 20 accumulated precipitation 1 - 10 average temperature relative humidity accumulated precipitation 1 - 30 accumulated precipitation 1 - 10 accumulated precipitation 11 - 20 null model accumulated precipitation 21 - 30 average temperature accumulated precipitation 1 - 10 relative humidity accumulated precipitation 1 - 30 null model accumulated precipitation 11 - 20 accumulated precipitation 21 - 30

-0.457 (±0.428) -0.292 (±0.339) -0.275 (± 0.365) -0.194 (±0.324) -0.123 (±0.067) 19.908 (±9.023) -3.038 (±4.200) 1.003 (±0.164) 0.879 (±0.543) 0.823 (±0.711) 0.972 (±0.618) 0.979 (±0.576) 12.351 (±4.482) -3.672 (±2.611) -0.636 (±0.481) -0.446 (±0.378) -0.465 (±0.405) -0.275(±0.085) -0.338 (±0.359) 10.620 (± 4.417) -1.611 (±0.357) -7.094 (±2.536) -1.814 (0.487) -1.069(±0.198) -1.306 (±0.396) -1.171 (±0.353)

* In brackets the standard error of estimates and in bold the values p <0.05.

0.011 (±0.011) 0.023 (±0.027) 0.011 (±0.018) 0.008 (±0.019) -0.712 (±0.334) 0.062 (±0.065) 0.015 (±0.038) 0.006 (±0.002) 0.002 (±0.032) 0.003 (±0.047) -0.486 (±0.172) 0.051 (±0.039) 0.012 (±0.013) 0.023 (± 0.030) 0.014 (±0.020) 0.007 (±0.021) -0.454 (0.173) 0.064 (0.025) 0.089 (±0.037) 0.023 (±0.011) 0.017 (±0.018) 0.011 (±0.019)

0.322 0.402 0.533 0.673 0.048 0.353 0.686 0.749 0.941 0.944 0.012 0.206 0.338 0.445 0.482 0.734 0.018 0.018 0.026 0.055 0.354 0.568

424.7 429.6 435.3 439.1 439.8 135.2 170.1 175.8 176.7 177.4 178.3 178.3 255.3 322.6 332.2 336.9 338.1 342.4 343.3 65.8 68.1 68.3 71.700 77.5 78.1 79.2

111.7 116.6 122.3 126.1 126.8 0 34.9 40.6 41.5 42.2 43.1 43.1 0 67.3 76.9 81.6 82.8 87.1 88 0 2.3 2.5 5.900 11.7 12.2 13.4

0.04 0.028 0.016 0.007 0.248 0.047 0.009 0.005 <0.001 <0.001 0.263 0.065 0.037 0.023 0.020 0.004 0.188 0.158 0.155 0.110 0.025 0.009

25

26

27

28

29

30