Impact of livestock on a mosquito community ( Diptera : Culicidae ) in a Brazilian tropical dry forest

Introduction: This study evaluated the effects of cattle removal on the Culicidae mosquito community structure in a tropical dry forest in Brazil. Methods: Culicidae were collected during dry and wet seasons in cattle presence and absence between August 2008 and October 2010 and assessed using multivariate statistical models. Results: Cattle removal did not signifi cantly alter Culicidae species richness and abundance. However, alterations were noted in Culicidae community composition. Conclusions: This is the fi rst study to evaluate the impact of cattle removal on Culicidae community structure in Brazil and demonstrates the importance of assessing ecological parameters such as community species composition.

Livestock are a main driver of biodiversity loss in ecosystems, mainly through alterations of local habitat through deforestation, destruction of riparian areas, and wetland drainage (1) .The impact of livestock on biodiversity may also be related to the use of production animals as a resource by other organisms.Hematophagous insects can readily exploit domestic animals as hosts.Cattle herds serve as important blood sources and are among the main hosts for mosquitoes (Diptera: Culicidae).
Three non-mutually exclusive epidemiological scenarios of disease transmission by insect vectors can result from the introduction of livestock to an area: 1) reduction in number of feedings on the human population leading to a reduction in disease transmission; 2) an increase in the vector population as a result of the extra blood source and/or an increase in habitat availability to larvae, leading to an increase in the number of feedings per person; and 3) an increase in the risk of disease transmission because the introduced domestic animals serve as reservoirs for human diseases (2) .To circumvent negative impacts of vectors on human hosts coexisting with animals, zooprophylaxis, is a commonly used method.This practices is defi ned by World Health Organization as the use of wild or domestic animal that are not hosts of a particular disease as reservoirs to divert mosquito vectors from human hosts.
One disadvantage of adopting zooprophylactic practices is that raising several domestic animals may change environmental conditions and lead to the production of suitable breeding sites for mosquito larvae, increasing the mosquito populations in these regions and thus increasing the risk of mosquito-borne disease transmission and co-infection.In Brazil, a knowledge gap exists regarding the impact of livestock on Culicidae fauna.Most studies that have addressed the infl uence of cattle on Culicidae species were performed primarily to determine the feeding preferences of mosquito species.Thus, the aim of our study was to evaluate the effect of cattle removal on Culicidae community composition in a tropical dry forest in the northern region of Minas Gerais, Brazil.The following hypotheses were tested: I) cattle presence increases Culicidae species richness and abundance and II) cattle removal alters Culicidae community composition.
The study was performed in the Mata Seca State Park (Parque Estadual da Mata Seca -PEMS) (14°48′36′′S-43°55′12′′), located in the Manga municipality; Northern Minas Gerais, Brazil (Figure 1).The region's climate is classifi ed as tropical dry with a mean annual temperature of 24.4°C and rainfall of 871mm (3) , characterized by a marked dry season in the winter, with seven months with rainfall levels lower than 60mm and forests areas dominated by deciduous vegetation, with 90-95% leaf loss.The park was created through the expropriation of four farms with extensive cattle farming and plantations of common beans and corn in two pivots of 80 hectares each (3) .In 2008, this area still housed approximately 800 cattle that were removed in the second half of 2009, according to information obtained from the offi ce of the Supervisory Agency of Farming in Minas Gerais [Instituto Mineiro de Agropecuária (IMA)].
Culicidae sampling was performed in three tropical dry forests fragments and one plot located in the gallery forest of a large pond (Lagoa da Prata) in different periods: August (rainy and dry seasons without cattle).Culicid collections were performed at dusk using one Shannon trap exposed for two hours per site for 8 nights (16 hours total) during each collection period.The mosquitoes were taxonomically identifi ed using dichotomous keys (4) and incorporated into the entomological collection of the Laboratory of Ecology and Biological Control of Insects at Montes Claros State University.
To identify the effect of the presence of cattle on Culicidae species richness and abundance, the data were fi t to generalized linear models (GLMs).The signifi cance of each variable was tested by an analysis of deviance using the chi-squared test.All of the GLMs underwent residual analyses to assess the adequacy of the error distributions.All analyses were performed with R 2.15 software (5) .
To evaluate the effect of cattle removal on Culicidae community composition, non-metric multidimensional scaling (NMDS) was performed using the Bray-Curtis dissimilarity index.To test for differences in community composition between sampling periods (presence vs. absence of cattle), analysis of similarity (ANOSIM) was performed.Moreover, a similarity percentage analysis (SIMPER) was used to determine which species contributed the most to the formation of distinct groups (when cattle were present or absent).The analyses were performed with the PAST software ver.2.17c (6) .A total of 5,170 mosquitoes were collected in the presence (n = 2,045 mosquitoes) or absence of cattle (n = 3,125 mosquitoes), and 21 and 27 species were represented, respectively (Table 1).The presence of cattle had no effect on of Culicidae richness (p = 0.531) or abundance (p = 0.521).
The Mansonini tribe represented 81.2% of mosquitoes collected (1,720 individuals during the period with cattle and 2,480 individuals after cattle removal).Mansonia titillans was the most abundant species sampled representing (29.1%) of the total number of collected mosquitoes and 1,509 (60.8%) of those collected after cattle removal.The species with the highest abundance during the period with cattle present was Coquillettidia nigricans, comprising 510 individuals (9.8% of the total mosquitoes collected).
The NMDS analysis indicated a change in the culicid community: two distinct groups were observed, suggesting that cattle removal affects community composition (Figure 2).The analysis of similarity (ANOSIM) showed that Culicidae community composition in PEMS differs signifi cantly between sampling periods (R = 0.4989; p < 0.01).
The SIMPER analysis showed that the species contributing most to the formation of clusters with or without cattle include Mansonia titillans, Mansonia humeralis, Mansonia pseudotitillans, Coquillettidia nigricans, and Coquillettidia venezuelensis, together comprising 71.6% of the two distinct groups reported by NMDS (Figure 2).These results suggest that the presence of cattle results in changes in Culicidae community structure, without necessarily altering species richness and total abundance.Multivariate analytical approaches are rarely used in studies on Culicidae, despite the broad applications of understanding species composition and ecological community changes.The NMDS analysis has been used, however, to predict annual cycles of the composition of mosquitoes infl uenced by rainfall in Darwin, Australia (7) .The high abundance of mosquitoes of the Mansonini tribe may be explained by the presence of dense aquatic vegetation in Lagoa da Prata, which serves as a breeding site for these mosquitoes.Immature mosquitoes can attach to this aquatic vegetation through modifi ed siphons for respiration and become dependent on these macrophytes for survival (4) .
Mansonini mosquitoes have been considered by other Brazilian authors, preferring to feed on avian species (8) .Cattle are large animals with few defensive behaviors and may have opportunistically shifted these insects to a hematophagic diet in cattle.Culex tarsalis can change their diet seasonally, feeding on mammals when their preferred hosts, herons and other aquatic birds, migrate to other locations (9) .Moreover, the increase in bird diversity and abundance after cattle removal from PEMS (unpublished data) may have contributed to the return of Ma. titillans to an ornithophilic diet.This hypothesis would need to be verifi ed by longitudinal studies and the analysis of bloodmeals.
The choice of hosts for hematophagy may be driven by several factors, including the availability and defensive behavior of hosts, blood nutritional value, and the energy cost of digestion (10) .Mosquitoes within a community depend primarily on host availability, and contact with specifi c hosts is infl uenced more by host presence/absence than by innate selection by the mosquitoes.Several studies have investigated the feeding preference of Culicidae, indicating that several mosquito species are more selective in host preference and feed on one or a few closely related species.Indeed, some hosts, including hoofstock (cattle and horses), other domestic animals, humans, and birds are preferred by certain mosquito species.Other mosquito species are not as discriminatory and select their hosts from among mammals, birds, and reptiles (11) .
Coquillettidia nigricans and Coquillettidia albicosta were more abundant when cattle were present, corroborating reports of a preference for large mammals by mosquitoes of this genus (12) .Among the Aedini mosquitoes, the most abundant species was Aedes scapularis, which was sampled in both collection periods but found to be especially prevalent in the absence of cattle.Ae. scapularis has received attention for its adaptation to environments modifi ed by humans.The domiciliation of Ae. scapularis in Brazil was documented in 1961, suggesting that this mosquito readily adapts to human-mediated environments (4) .
The results presented here suggest an association between species of the genus Anopheles and the presence of cattle.Anopheles albitarsis l.s.typically co-occurs with cattle, reinforcing previous results that this anopheline has a strong tendency to perform hematophagy on large animals, such as cattle and horses (13) (14) .
The abundance of Anopheles darlingi co-occurring with cattle is nine times higher than its abundance without the host.This species is considered the most widely distributed anopheline in South America and is the main vector of Plasmodium spp.(the protozoan parasite that causes malaria in humans) in Brazil.Bloodmeal analysis of An. darlingi collected in resting habitats at three rural villages located in State of Amapá found that 57% of non-mixed bloodmeals are from cattle, while only 12% are from human, demonstrating that this species may be more eclectic in its food habits (15) .Most of the literature regarding the effects of livestock on mosquito populations consists of paired studies comparing different areas with or without cattle near human dwellings, measuring the risk of acquiring a disease, in most cases malaria.The present study, although limited by the small number of samples, is the fi rst in Brazil to characterize the effects of mosquito composition in the same area with and without cattle, and to demonstrate a change in Culicidae community structure in the absence of this important blood source.Furthermore, this study improves our understanding of the ecology of mosquito fauna in this semi-arid region with an important economic base on livestock and points out the risks and the need for surveillance to prevent possible disease outbreaks.