Species of Drosophila ( Diptera : Drosophilidae ) attracted to dung and carrion baited pitfall traps in the Uruguayan Eastern Serranías

This study investigates the species richness and abundance of Drosophila Fallén, 1823 attracted to dung and carrion baited pitfall traps in natural areas with heterogeneous habitats at the Sierra de Minas, Eastern Serranías, southeastern Uruguay. Collecting was carried out on a monthly basis (May 2002 through April 2003). Drosophilids accounted for 0.84% (n = 131) and 3.61% (n = 158) of the Diptera collected from dung (n = 15,630) and carrion (n = 4,382) pitfall traps, respectively. A total of 12 species were identified, 11 of which belong to the subgenus Drosophila (the richest) and one to the subgenus Sophophora Sturtevant, 1939. Over 90% of the Drosophila specimens collected belong to five species of the subgenus Drosophila, namely D. gaucha Jaeger & Salzano, 1953, D. immigrans Sturtevant, 1921, D. mediovittata Frota-Pessoa, 1954, D. aff. nappae Vilela, Valente & Basso-da-Silva, 2004, and D. ornatifrons Duda, 1927. Drosophila cardini Sturtevant, 1916 is recorded for the first time from Uruguay. Drosophila abundance and species richness in the four habitats sampled in the Uruguayan Eastern Serranías, namely woodlands sierra, riparian forest, pine plantation and grazing grassland, were considered to be a function of habitat conservation. Diversity indices were low in all habitats. Different habitats supported particular coprophilous and necrophilous Drosophila species. The woodland sierra represents the most preserved habitat, and contributed with the highest species richness observed. Drosophila ornatifrons was the dominant species, with a restricted habitat distribution. On the other hand, grazed grassland, an environment modified by livestock management, had the lowest species richness: only a few specimens of D. repleta Wollaston, 1858. Regarding species composition, significant differences were found in some pairwise comparisons of groups of Drosophila species that included D. ornatifrons. Fly attraction to dung can be exploited as an alternative and/ or complementary collecting method in ecological studies of Drosophila assemblages in natural areas.

indicated that Drosophilidae were identified among necrophilous insects sampled from the Mexican regions of Jojutla (DELOYA et al. 1987) and Tamaulipas (SÁNCHEZ-RAMOS et al. 1993).In these studies, squid or marine fish meat were used as baits in pitfall traps.Those were replaced biweekly or monthly and 95 parts of ethanol 70° and 5 parts of glacial acetic acid (MORÓN & TERRÓN 1984) or commercial antifreeze liquid (SÁNCHEZ-RAMOS et al. 1993) were used as preservative solutions.As stated by MORÓN & TERRÓN (1984), the preservative solutions they used do not mask the fetid volatile odors from bait, but instead enhance the attractiveness of the bait to other saprophagous insects (as Drosophilidae).
At southern latitudes, REMEDIOS et al. (2012) reported flies of Drosophilidae among Diptera sampled in pitfall traps in natural areas at the Sierras de Minas, Department of Lavalleja, Uruguay.
Since the samples in the research papers cited above have not been identified to species, the efficacy of pitfall traps as an alternative collecting method for Drosophilidae diversity and ecological studies remains to be assayed.
Mid-latitude South American lowlands gain particular importance when the ecological patterns for faunal groups are examined.Besides some ecological and distributional considerations on Drosophila species, mostly from collecting at south and southeast Uruguay, reported by GOÑI et al. (1997GOÑI et al. ( , 1998)), drosophilid diversity in the natural areas locally known as sierras or cuchillas is almost unknown.The most extensive sierras in Uruguay, known as the Serranías del Este (Eastern Serranías), cover a half moon -shaped area at southeastern Uruguay (Fig. 1, forward-slashed area at right).They have a rugged topography, with altitudes ranging from 200 to 500 m, where the key element is the unevenness between peaks and foothills, and the steep slopes.The geological basement of the Eastern Serranías System is varied and complex, especially in the southern area.Different geological formations (interspersed granitoid and last postectonic granitoids, basaltic lavas, limestone, and Pliocenic sandstones and Pleistocenic loess) are found within very short distances.This geological diversity is reflected into a variety of soil types (from the dominant shallow soils to moderately and very deep ones) associated with rocky and stony outcrops.The average annual temperature is 17°C, with an annual rainfall between 1,000 mm and 1,200 mm (EVIA & GUDYNAS 2000).The typical sierra landscape consists of fault ridges with a regular range of hills and streams interspersed between them, and covered predominantly by patches of woody and shrubby vegetation and rocky outcrops.The Sierra de las Animas is located at the southernmost point of the Eastern Serranías, and the Sierra de Minas to the west, whereas nearby the Serranías de Sauce, Cabral and Caracoles are formed.These sierras present a major south-north axis, separated by valleys of several streams.This region is ecologically very significant, because of the wide extent of natural areas that harbor a rich fauna and flora, and has suffered little human impact.More-over, soils of the sierras are in general suitable for forestry-related activities, in particular pine and eucalyptus plantations, which have been implemented in the last 20 years in areas traditionally occupied by grasslands (EVIA & GUDYNAS 2000).Considering the mosaic distribution of the typical diverse natural ecosystems of the sierras, such as woodlands and shrub lands, grasslands, rocky and riparian forests, and the rich fauna and flora, the sierras are of great interest to researchers and conservationists (EVIA & GUDYNAS 2000, BRAZEIRO et al. 2008).
The purpose of the present study was to investigate the abundance and species composition of Drosophila attracted to dung and carrion baited pitfall traps placed in natural areas of Sierra de Minas, with different grades of anthropogenic disturbance, and compare the spatial distribution of the species and their association with habitat disturbance, as part of a research project on ecological studies on Uruguayan dung and carrion beetles (GONZÁLEZ-VAINER & MORELLI 2008).

MATERIAL AND METHODS
Collecting was carried out monthly within an area of 10 km 2 at the Sierra de Minas, Department of Lavalleja, Uruguay, from May 2002 through April 2003.The area is adjacent to Route 81, km 105-112 at the south of the Eastern Serranías system, between 34°30'59"S, 55°20'07"W and 34°30'54"S, 55°19'53"W (Fig. 1).This area includes Cerro Mirador at 370 m altitude and other unnamed lower elevations, whose eastern side gently slopes down to the Mataojo creek.It is characterized by silt loam uliginous soil, with altitude varying from 110 to 230 m (GONZÁLEZ-VAINER & MORELLI 2008).Four areas, representing different habitats, were selected for sampling (Fig. 2 Coprophilous and necrophilous insects were sampled using pitfall traps baited with cow dung (from livestock fed local grasses) or cow liver, buried with the rim at ground level and filled with an aqueous solution of formaldehyde (10%) and a drop of detergent (Figs 3-5).The bait (150 g) was placed in the middle of the upper plastic jar, which was perforated in its bottom, allowing insects to fall into a second jar containing the preserving fluid (Fig. 3).In each habitat, three pitfall traps were placed at 20 m intervals, resulting in a total of 288 traps.
Baited traps remained in place for a week; the captured insects were sorted in the laboratory.All dipteran insects were preserved in ethanol 70% until identification.Drosophila samples were counted and identified to species using characters of the external morphology and, in some cases, the male and/or female terminalia, according to the method detailed by BÄCHLI et al. (2004).et al. (2000), andVILELA et al. (2004) were used.Male and female specimens used for species identification were doublemounted (glued to cardboard tips soon after drying, according to BÄCHLI et al. 2004: 3), labeled and deposited as vouchers at the Museu de Zoologia, Universidade de São Paulo (Brazil), and at the collection of Sección Entomología, Departamento de Biología Animal of the Facultad de Ciencias, Universidad de la República, Montevideo (Uruguay).
Total sample estimates of species richness and abundance in each habitat were used to compute the following ecological indexes: Shannon (H' = -⌺ pi ln pi), Berger-Parker (d = Nmax/N) and Pielou evenness (E = H'/Hmax) (MAGURRAN 1988).Pairedsample Student's t tests (ZAR 1994) were used to examine differences in species diversity between dung and carrion baited pitfall traps.A matrix was constructed using absolute abundance of samples in each habitat and each species at dung and carrion baited traps.Sample-based rarefaction curves were constructed to assess species richness for each habitat according to baited trap.The resultant curve is a plot of the number of species as a function of the number of samples, which also minimizes the effect of the absence of samples by loss of traps.Rank abun-

RESULTS
In this study over 20,000 Diptera were collected at the Sierra de Minas in a twelve-month period (Tab.I).Most of them were coprophagous (78.10%, n = 15,630) with a lower percentage of necrophagous (21.90%, n = 4,382).Temporal and spatial variations in abundance were observed among dung and carrion traps samples.Diptera were most abundant from October to December, then again in March and April, most probably due to the total precipitation peaks recorded in those (or immediately previous) months (Fig. 6).Considering the accidental loss of some monthly samples, the spatial and temporal data on Diptera composition in each habitat should be viewed as tentative.
A total of 289 drosophilids (131 in dung traps and 158 in carrion traps), belonging to 12 species of Drosophila, were identified: eleven species in the subgenus Drosophila and one spe-cies (D. willistoni Sturtevant, 1916) in the subgenus Sophophora (Tab.II).With the exception of Drosophila cardini Sturtevant, 1916 and two undescribed species of the tripunctata group, the remaining nine species had been previously recorded from Uruguay (GOÑI et al. 1997(GOÑI et al. , 1998)).
The rarefaction curves of Drosophila species attracted to dung and carrion traps have shown that more (unrecorded) species may occur in the study area (Figs 7-8).Drosophila richness at the woodlands sierra (WS) habitat is likely to yield nearly 14 species, whereas the steep slope of species richness observed in the other three habitats indicates that additional species are expected to occur.
The diversity indices were low in all habitats (Tab.II).Nonsignificant differences were found at the 5% level (Student's t test) regarding species diversity in relation to the bait, dung vs carrion, in three out of the four analyzed habitats, WS (t = 1.17, v = 229.1,p = 0.24), RP (t = -1.47,v = 2.8, p = 0.24), and PP (t = -0.02,v = 8.9, p = 0.99).The abundance of species attracted to dung and carrion baited traps in each habitat are shown in Figu-      Different habitats supported particular coprophilous and necrophilous Drosophila species as shown by the ANOSIM analyses (Global R = 0.09, p = 0.008 for the dung baited trap group, and R = 0.04, p = 0.04 for the carrion-baited trap group, respectively).As expected, significant differences in species composition were found within some pairwise comparisons of Drosophila species (Tab.III).The SIMPER analysis pointed to D. ornatifrons as the taxon primarily responsible for the observed differences between some groups of samples.In dung baited traps, this species contributed with 35.01%and 33.74% of the observed differences between WS and RF, and between WS and PP habitats, respectively (the overall dissimilarity was 65.07% and 64.36%, respectively).In carrion baited traps, D. ornatifrons contributed with 30.11% between WS and RF habitats (showing an overall dissimilarity of 54%).
Cluster analysis using habitats association (Figs 12 and 13) showed that WS, RF and PP were associated in dung and carrion traps, with low and medium similarity values, respectively.In carrion traps (Fig. 13), GG habitat was unassociated, probably due to the low abundance and species richness, with no samples in dung traps (Tab.II).Different associations of species were obtained in dung and carrion traps (Figs 14 and 15).In dung traps (Fig. 14), two clusters of species showed the highest level of association: one formed by six species (D. cardini, D. hydei, D. ornatifrons, D. meridionalis Wasserman, 1962, D. aff. paraguayensis Duda, 1927, andD. willistoni) and the other by two members (D. immigrans and D. aff.nappae).They were associated to other Drosophila species at varying levels of similarity (values above 0.4).In carrion traps (Fig. 15), seven out of eight species were associated to each other at varying levels of similarity.In both types of traps, D. repleta was unassociated and divergent from the other species.

DISCUSSION
Drosophila ornatifrons had been previously registered in a sample collection at Cerro del Toro (34°51'15"S; 55°15'23"W), an isolated topographic accident of Sierra de las Animas, located at the southernmost region in the Eastern Serranías System, using conventional banana-baited traps (GOÑI et al. 1998).Drosophila ornatifrons (also cited under its junior synonym, D. guarani Dobzhansky & Pavan, 1943) was formerly believed to be endemic to the Atlantic Forest biome of Brazil (SENE et al. 1980).However, later on, it was collected in gallery forests, xerophitic enclaves, highland rocky fields and other transitional areas between the Atlantic Forest and adjacent biomes (ARAÚJO & VALENTE 1981, TIDON-SKLORZ et al. 1994, VILELA & MORI 1997, MATEUS et al. 2006, TIDON 2006).It was also recorded from Colombia, Bolivia and Ecuador (RAFAEL & VELA 2000).Temporal and spatial studies of drosophilids in the cerrado biome indicate that D. ornatifrons has been more frequently found in gallery forests than in the savanna-like vegetation (or cerrado), and has low but significant positive correlation with monthly humidity and light in   Two species of the tripunctata species group, D. mediovittata and D. aff.nappae, were not very abundant among the Drosophila samples collected both in dung and carrion traps.Drosophila mediovittata belongs to the subgroup IV (FROTA-PESSOA 1954) and seems to be endemic to Brazil and Uruguay.In Uruguay, this species breeds on a limited number of substrates such as flowers of Erythrina crista-galli L. (Fabaceae) and on rotting cladodes of Opuntia arechavaletai Speg.(Cactaceae) collected from the Atlantic wetlands in the Department of Rocha (GOÑI et al. 1998).
Drosophila immigrans is a cosmopolitan species that had been previously collected in several localities in Uruguay (GOÑI et al. 1997(GOÑI et al. , 1998)).
Drosophila gaucha is the only member of the mesophragmatica species group that seems to be widely distributed outside the Andean region.It is found in Ecuador, Bolivia, Argentina, Brazil, and Uruguay (JAEGER & SALZANO 1953, BRNCIC & SANTIBAÑEZ 1957, BRNCIC et al. 1971, GOÑI et al. 1997, RAFAEL & VELA 2000).According to BRNCIC & SANTIBAÑEZ (1957) there are no morphological differences between D. gaucha and its sibling species, D. pavani Brncic, 1957.Since the latter species has never been collected in the southernmost Brazilian state of Rio Grande do Sul, which includes the type locality of the former and extends contiguously to Uruguay, the specimens collected were tentatively identified as belonging to D. gaucha.Three species of the repleta species group, D. hydei Sturtevant, 1921, D. mercatorum Patterson &Wheeler, 1942, andD. repleta were identified in our samples.Drosophila hydei and D. repleta are cosmopolitan species, whereas D. mercatorum is semicosmopolitan.In previous collections by GOÑI et al. (1997GOÑI et al. ( , 1998)), D. hydei was frequently found associated with pristine as well as man-modified environments.In the present study, only one female specimen of D. hydei was sampled at the WS habitat (Tab.II).Drosophila repleta is a domestic species, often found in restaurants, urinals and inside houses, especially kitchens (STURTEVANT 1918, COE 1943), as well as in poultry manure, from which BICHO et al. (2004) identified 9,229 adults of this species among 28,720 Diptera collected, representing 32% of the flies.As just one female imago of D. repleta was collected in dung baited traps placed at the RF habitat, it seems that this species is not as attracted to cow dung baits, containing only feces, as it is to the semisolid mixture of feces and of uric acid of the poultry manure.Whether the uric acid is responsible for the strong attraction that poultry manure and restaurant's urinals exert on flies (COE 1943) remains an open question.Although D. repleta was absent from necrophagous insect communities sampled during the decomposition of medium-sized domestic pig carcasses at a central European urban habitat (GRASSBERGER & FRANK 2004), in the present survey one male and one female of this species were collected in carrion baited pitfall traps in the GG habitat (Tab.II).The occurrence of D. repleta in the latter traps may indicate the presence of human settlements near the sampled areas (Fig. 2).Just one female of D. mercatorum, probably belonging to D. mercatorum pararepleta, was collected in the present study, in a carrion baited trap set at the WS habitat (Tab.II).
Finally, a few specimens of D. cardini were collected in dung traps at the most preserved habitat, the WS.This species has been found in Florida (USA), Mexico, Central and South America, and the West Indies, and later recorded from the Hawaiian Islands (HERFORTH et al. 1984), where it became one member of the established immigrant drosophilid species (LEBLANC et al. 2009).In Brazil, this species has been collected in semi-arid areas such as caatingas, but can also be found in cerrados (savannas), the coast and the restingas (sand dune vegetation) (VILELA et al. 2002).This is the first inventory of Drosophila species in Sierra de Minas using pitfall traps placed across heterogeneous natural areas within a relatively small geographic area (altitude of sampled sites varied between 110-230 m, and the longest distance between these sites was about 5 km).Endemic and cosmopolitan species were collected in almost all habitats.Among the endemic species, one previously unrecorded species from Uruguay, D. cardini, and two undescribed ones of the tripunctata group were registered.However, some widespread cosmopolitan generalists, such as D. simulans Sturtevant, 1919, which is frequently collected in natural areas of Uruguay (GOÑI et al. 1998), was absent from our traps.This observation indicates that the use of pitfall traps to collect Drosophila might result in an underestimation of the true species richness.This hypothesis is consistent with the ZOOLOGIA 29 (4): 308-317, August, 2012 (higher) expected richness estimated in this study for the sampled area.The abundance and richness of endemic species (D. cardini, D. ornatifrons, D. meridionalis, D. aff. nappae, D. aff. paraguayensis and D. willistoni) are associated with habitat differences.We infer this based on the fact that statistical differences, as well as low habitat association (estimated in dung and carrion baited traps), were observed among the Drosophila assemblages of the sampled areas.At the most preserved area, the WS habitat, most of the species are at least partially (or totally) dependent on the forest environment, where the greatest number of species were sampled by both dung and carrion baited traps.Furthermore, in this habitat, D. ornatifrons was dominant with a restricted distribution, suggesting its association with local resource(s) availability and/or particular forest environmental conditions.On the other hand, in the natural modified habitat where anthropogenic disturbance was present, the abundance and richness of Drosophila shifted to low indices, as observed in the GG habitat, with only one species, D. repleta, recorded in carrion baited traps.
The results presented here indicate that carrion and dung baited pitfall traps, though not efficient in terms of expected richness, it can be exploited as an additional and reliable sampling method.However, it does not replace the more conventional traps (based only on banana or bakers' yeast-fermented banana bait) used to study Drosophila biodiversity and ecology.Comparative studies involving insect collections using different baits (fruits, carrion, and dung), conducted simultaneously, can be employed to test the usefulness of alternative baits for understanding more fully the dynamics of frequency and abundance of drosophilid species in natural environments.Our findings support the view that the Drosophila assemblage attracted to carrion and dung baited pitfall traps is quite sensitive to natural forest or natural grasslands, and can be used to monitor environmental disturbance in conservation programs along with other necrophilous and coprophilous insects.

Figures
Figures 1-2.(1) Location of the study area in the Uruguayan Eastern Serranías (arrowhead).Northern and Eastern Serranías unit (forward slashes), modified from EVIA & GUDYNAS (2000).(2) Satellite photograph (from Google Earth free virtual globe, map and geographical information program) showing part of the sampled area where coprophilous and necrophilous insects were collected.Woodlands sierra (WS), riparian forest (RF), grazing grassland (GG), and pine plantation (PP).Arrows indicate rural houses.
Whittaker plots) were built to display the relative species abundance attracted to both dung and carrion baited trap in each habitat.The rank abundance curve provides a means for visually representing species richness and species evenness.Analysis of similarities (ANOSIM, CLARKE 1993) was performed to statistically test whether the species attracted to dung or carrion baited traps varied significantly between habitats.Analysis of Similarity Percentage (SIMPER, CLARKE 1993) was used to assess the contribution of each species to the average Bray-Curtis dissimilarity between habitats.The relationships between habitats or species composition in dung or carrion baited traps were analyzed using Morisita similarity index (KREBS 1999), followed by cluster analysis using the UPGMA method(SNEATH & SOKAL 1973).For statistical analysis (natural logarithmic transformation in number of individuals) ln (n + 1) was used.All data analyses were carried out using PAST version 2.09(HAMMER et al. 2001) and Biodiversity Pro version 2.0 (Biodiversity Pro: Free Statistics Software for Ecology).

Figures 3
Figures 3-5.Schematic representation of the dung (or carrion) baited pitfall trap (3) used in this study containing the bait (above) and the fixative (below).The pitfall trap (4) before being buried with the rim at ground level (5).

Figure 6 .
Figure 6.Temperature and rainfall monthly averages during the study period.Source: Dirección Nacional de Meteorología, Uruguay.
and carrion baited traps, respectively (Tab.II).By contrast, there is a sharp fall in the species abundance curves for the RP and PP habitats.These habitats, surrounded by open fields, contributed and almost 87% of the individuals attracted to dung and carrion baited traps, respectively.This endemic species was restricted to the WS habitat.Other species, D. gaucha Jaeger

Table I .
Abundance of Diptera collected in four habitats at the Sierra de Minas, Department of Lavalleja, Uruguay, from May 2002 to April 2003.

Table II .
Species richness, abundance of Drosophila spp.and ecological indexes calculated for each habitat surveyed at the Sierra de Minas, Department of Lavalleja, Uruguay, from May 2002 to April 2003.Dung (D) and carrion (C) baited pitfall traps.
Male Female Male Female Male Female Male Female Male Female Male Female Male Female cardini

Table III .
ANOSIM.Pairwise comparison of Drosophila species abundance and composition between habitats for dung (D) and carrion (C) baited pitfall traps.