The Drosophila communiTy in xerophyTic vegeTaTions of The upper parana-paraguay river Basin maTeus

The genus Drosophila is the most abundant in the Drosophilidae family. some species are endemic to certain regions and others are cosmopolitans. in Brazil, there are several ecosystems to explore regarding the composition and ecological aspects of Drosophila. xerophytic areas are an example. They can be found in the south and southeast of Brazil as islands, a result of paleoclimatic cycle changes. The aim of the present work is to provide information about the composition of the Drosophila community in eight xerophytic areas (where the cactus Cereus hildmaniannus occurs) in the south and southeast of Brazil. This work is an important step in the study of quantitative ecological aspects of the Drosophila community in xerophytic areas that form continental islands in the neotropical region. The composition of the Drosophila community which was found is compatible with previous work in several aspects. The ecological indexes showed a possible positive association between diversity and the degree of preservation of the studied areas. são paulo state communities presented the highest similarity among the sites tested, although no statistical significant correlation between the Jaccard index and geographical distance was found using the mantel test.


IntroductIon
although in the 18 th century scopoli, Linné and fabricius were the first to describe the species which are today allocated to the family Drosophilidae, Duda (1924aDuda ( , 1924bDuda ( , 1925) could be considered the pioneer in the systematics of this group.moreover, sturtevant, well known for his classic work (1921,1939,1942), is one of the most important researchers within this family.The first data on the Brazilian Drosophila community were published by Duda (1925). in the southeast region of Brazil, Dobzhansky & pavan (1943) were pioneers and described several new species.after 1940, much work about Brazilian Drosophila fauna was published (pavan & cunha, 1947;Dobzhansky & pavan, 1950;pavan, 1950;1959) including the first Brazilian Drosophila species list (mourão et al., 1965).Later, other authors analyzed the Drosophila fauna in the morphoclimatic domains of Brazil in an effort to understand the geographical distribution of the most common species better (sene et al., 1980; vilela et al., 1983; val et al., 1981).
The Drosophilidae family is composed by 65 genus and more than 3,500 described species that occur in a number of ecosystems all over the world (Bächli, 1998).most genera are found in tropical regions.The Drosophila genus is the most abundant and comprises around 53% of the total species.some of them are endemic to certain regions and others are cosmopolitan, dispersed mostly in association with human activity.grimaldi (1990) reviewed the classification of the Drosophilidae family based on a cladistic analysis using maximum parsimony.This author studied 217 morphological characters of adults and 120 species representative of the family and concluded that they constitute a monophyletic group.The Drosophila genus probably originated in the tropics (stalker, 1976) and different members of it expanded their distribution to the temperate zones, where the ability to use resources under extreme environmental conditions was selected.This process might have been repeated several times during the evolution of the genus.
Today, most research using Drosophila species is being conducted in the fields of molecular Biology and genetics, but it was only in the early twentieth century that Drosophila melanogaster was recognized as useful for genetic analyses (castle, 1906).Despite the large number of studies using Drosophila as a model organism in these fields, Brazilian species still require further studies regarding many aspects of ecology, systematics, genetics and evolution.some ecological studies with Drosophila have been carried out, especially regarding fly activity during the day (Klaczko et al., 1983;medeiros, 2000), the appropriate period of the day for sampling (pavan et al., 1950;Belo & oliveira-filho, 1978), and the spatial distribution of species (horizontal: Dobzhanzky & pavan, 1950;Burla et al., 1950;pavan, 1959;medeiros, 2000;and vertical: sene et al., 1981;Kratz et al., 1982e Tidon-sklorz & sene, 1992).however, comparison among areas has received little attention (medeiros & Klaczko, 2004, for example) and still remains an open area of study for Brazilian Drosophila species.
Taking this into account, there are several ecosystems to explore in Brazil regarding composition and ecological aspects of Drosophila.xerophytic areas, determined by the presence of drytype vegetation with the occurrence of different types of cacti, are such an example.currently, as well as the center areas of cacti occurrence, such as the Brazilian caatinga and argentinean chaco, these plants can be found in the south, southeast, and central regions of Brazil on islands, a result of the cycled paleoclimatic changes.These cycles alternated between cold/dry and warm/humid. in the first case, xerophytic areas expanded from the center areas over the continent and in the second, they retracted forming refuges (islands) between them (Bigarella et al., 1975;ab'saber, 1977;vanzolini, 1981).Therefore, refuges are important for biodiversity, ecological and evolutionary studies of several groups, including Drosophila species that breed specifically in rotting cactus tissue, as do flies of the D. repleta group.
The aim of the present work is to provide information about the composition of the Drosophila community in eight xerophytic areas (where cactus Cereus hildmaniannus occurs) in the south and southeast regions of Brazil.This will provide valuable information for further studies regarding ecology, genetics and the evolution of geographically isolated Drosophila populations.

mAterIAl And methods
The places and date of collections are summarized in Table 1.The samples were collected in woodlands composed of xerophytic vegetation in mesophytic forests in the valleys of the upper portion of the parana-paraguay river basin.These areas are at different stages of perturbation, ranging from use for livestock raising next to a road (itatiba -sp and sengés -pr) to a state park preservation area (canyon guartelá -pr).all collections were made at the end of the wet period, which corresponds to february to may in two different years (1999 and 2000).

Collections
for three days, the adult flies were attracted to fermented banana and orange in open traps, dispersed no closer than 15 m from each other, hung on branches at ± 1.5 m height from the ground.after this period, the flies were captured with entomological nets, put in glass jars with culture medium and taken to the laboratory alive, where they were identified.
it is difficult to identify members of the Drosophila repleta and D. willistoni groups to a certain degree.The females of the first group are shown as unidentified if the diagnostic characteristic is the male aedeagus.in terms of the Drosophila willistoni group, identification using genetic traits is more reliable but was not done.Therefore, the cryptic species were grouped as "repleta group" and "D.willistoni group", respectively.The unidentified "repleta group like" species were not considered in the analyses.

Data analyses
in order to verify the species occurrence qualitatively, the occurrence constancy method (Dajoz, 1983) was used.The constancy value (c) was obtained by dividing the number of collections in which one species occurred by the total number of collections, and then multiplying that result by 100.species with index c ≥ 50 were considered constants.accessory species were those with 25 ≤ c < 50.accidental species had c < 25. species that occurred in only one area were considered exclusive.
Drosophila communities were analyzed using several ecological indexes: Berger-parker (1/d), shannon-Wienner (h'), simpson (D), margalef (D mg ), menhinick (D mn ) and Jaccard (J). a nonparametric anova, Kruskal-Wallis, was used to determine if there were differences in the median abundance between communities.afterwards, a multiple comparison among samples (similar to the Tukey test) was made to determine whether significant differences occurred between the samples (Zar, 1999).
in order to verify the relationships among communities, two cluster analyses were performed: 1) using presence and absence of species; and 2) using the number of specimens collected for each species in each area.The phenograms were built in minitab® for Windows release 10.1 software selecting euclidean distance and single Linkage methods in both approaches.
To analyze if there is an association of community similarities and geographical distances, a mantel test was done using the Tfpga software (miller, 1997) comparing two matrixes: Jaccard indexes and geographical distances between a pair wise of communities.results And dIscussIon in this study we were able to identify a total of 10,464 flies belonging to 4 subgenera, 10 groups and at least 30 different species of Drosophila (Table 2).constant species (c ≥ 50) represented approximately 63% of the total collected species (19 out of 30).seven species were considered accessory (~ 23%) and 4 accidental species were found (~ 13%).all subgenera showed constant species, except for the Dorsilopha subgenus that showed only one accidental species (Drosophila busckii).
Drosophila polymorpha was collected from guatemala to Brazil and in different types of environments: forests, cerrado, restinga, and also associated with human presence.it is a relatively abundant species in different morphoclimatic domains, except in the caatinga (sene et al., 1980).
Drosophila antonietae was expected to be collected in all areas, as it is a cactophilic species that occurs in the south and southeast regions of Brazil, and north of the eastern boundary of the argentinean chaco.This species always occurs associated with the columnar cactus Cereus hildmaniannus in the xerophytic vegetation of gallery and mesophytic forests in the valleys of the paraná-uruguai river basin (Tidon-sklorz & sene, 2001).This pattern of distribution that isolates Drosophila antonietae from other D. buzzatii cluster species coincides with defined patterns of other dipterans and also with regions of endemism, suggesting vicariant events such as zoogeographic barriers established in earlier geological eras or climatic events that prevented gene flow enabling the accumulation of differences between populations (amorin & pires, 1996).This pattern of distribution is also observed regarding plant distribution restricted to the xerophytic caatinga and chaco regions (Duvernell & eanes, 2000).
Drosophila mercatorum is quite common in natural environments in south america, especially in open areas (sene et al., 1981; vilela et al., 1983).
Drosophila sturtevanti is a widely distributed species in central and south america occurring in different morphoclimatic domains (Tidon-sklorz & sene, 1999).
exclusive species appeared in salto santa rosa -pr (Drosophila busckii) and canyon guartelá -pr (Drosophila guaraja, D. carolinae and D. nigricruria).Drosophila busckii was not expected to be collected in salto santa rosa -pr as it is an introduced species in Brazil and is frequently associated with environments modified by man (Tidon-sklorz & sene, 1999).however, it has been collected in natural environments mainly in dry open vegetations (Bizzo & sene, 1982;Tidon-sklorz & sene, 1992).among the exclusive species collected in canyon guartelá -pr, Drosophila guaraja was registered in Brazil only in Boracéia -sp (val & Kaneshiro, 1988).This is, therefore, the first record of it outside this locality in Brazil.Drosophila carolinae has been collected only in rain forests in Brazil (vilela, 1983) and for this reason was not expected to be found in areas of xerophytic vegetation.however, canyon guartelá -pr is one of the most preserved areas surveyed and the cacti are distributed under a woodland region inside the state park, a fact that could explain its presence.Drosophila nigricruria is the only exclusive species that has wide geographical distribution.nevertheless, it was collected only in this area in this study.medeiros & Klaczko (2004) also collected this species in Barreiro rico farm (between 48° 05' 11'' W -22° 41' 15'' s and 48° 04' 52'' W -22° 41' 26'' s), an interior plateau which has a semi deciduous forest and one of the three remaining forests in são paulo state where collections were made by these authors.
The Drosophila subgenus showed the highest richness (7 groups and 22 species), with 7 species belonging to the Drosophila repleta group.The second highest richness was found in the Sophophora subgenus (3 groups and 6 species).This richness distribution among subgenera has been observed in other studies (sene et al., 1981;Tidon-sklorz & sene, 1995;Tidon-sklorz & sene, 1999;medeiros & Klaczko, 2004). 2 Cryptic species of the group that could not be identified.The introduced drosophilid species, Zaprionus indianus (vilela, 1999), was found in all collections (1,459 specimens in serrana; 317 in itatiba; 409 in santa maria da serra; 302 in itirapina; 42 in salto santa rosa; 117 in guartelá; 598 in sengés; and 8 in cianorte).The events associated to biological invasions are divided into three categories: arrival, establishment, and integration (vermeij, 1996).according to moraes et al. (2000), Zaprionus indianus has gone through all three and is probably competing unfavorably with other similar species in the southeast of Brazil, as a fall in relative frequency has been observed.The introduction of an exotic Drosophila species in Brazil is not an uncommon event. in 1976, val & sene (1980) collected a large number of Drosophila malerkotliana in northeastern Brazil, a species which was not recorded in extensive earlier collections.currently in our surveys, Drosophila malerkotliana is collected almost throughout Brazil, though always at a low relative frequency.
Twenty-three different species were collected in santa maria da serra -sp, 20 species in serrana -sp and in canyon guartelá -pr, 18 in itirapina -sp, 17 in cianorte -pr, 16 in itatiba -sp and salto santa rosa -pr and 9 in sengés -pr (Table 3).medeiros & Klaczko (2004) described the last inventory of Drosophila fauna in Brazil, surveyed in three forests remaining in são paulo state.These authors obtained 57, 76 and 90 different species in three studied areas.These numbers of species are greater than those observed here, perhaps a result of differences in the collection method or simply because the surveyed areas are distinct in their vegetation constitution (forest and xerophytic vegetation).The highest total abundance was detected in serrana -sp (4676 specimens) and the lowest in cianorte -pr (123 specimens) (Table 3).
according to ecological indexes (Table 3), cianorte -pr presented higher values in almost all indexes (1/d = 4.2409; h' = 2.3116; D mg = 3.3249; D mn = 1.5328), except for simpson (D) which showed canyon guartelá -pr to be the most diverse area (D = 0.5196).canyon guartelá -pr also showed one of the highest indexes in two more cases (D mg = 2.8891; D mn = 0.7464).sengéspr, itatiba -sp and serrana -sp showed the lowest values for most indexes.santa maria da serra -sp, itirapina -sp and salto santa rosapr showed intermediate values for most indexes.These observations are compatible with the superficial evaluation of degradation in these areas.cianorte -pr is located next to a road, but on a high slope board of the Ligeiro river, with difficult access for human degradation.canyon guartelá -pr is a woodland area inside the canyon guartelá state park of paraná state, as mentioned earlier.Therefore, it is a highly preserved area.sengés -pr is the most degraded area, next to a road, home to livestock raising and under high influence from human activity.serrana -sp is a preserved area on the top of a hill which is difficult to access.The lower diversity indexes found for this area must be due to the higher number of traps used in this collection (Table 1).The other areas were expected The pairwise comparison of communities using the Jaccard index (Table 4) showed that são paulo state communities (serrana -sp, itatiba -sp, santa maria da serra -sp and itirapina -sp) had the highest similarity among all comparisons (J > 0.62).The same pattern was not observed among paraná state communities and also between both states, with the exception of the index obtained between santa maria da serra -sp and cianorte -pr (J = 0.667), and between salto santa rosa -pr and canyon guartelá -pr (J = 0.636).The cluster analyses using presence and absence of species data (fig.1),    and sengés -pr communities were the most distinct when compared individually to the others (4 significant different comparisons for each one).itirapina -sp and canyon guartelá -pr were the most similar communities when compared to each community sampled (only one significant difference for each one).all these comparisons (Jaccard index, cluster analyses and multiple pairwise comparisons) showed that a higher similarity occurred among são paulo state communities, although no positive significant correlation was found in the mantel test comparing Jaccard index and geographical distances (r = -0.3235;p = 0.0839).
The community composition survey is an important step for several studies in a wide range of fields, such as genetics, ecology and evolution.This work adds new knowledge regarding composition and diversity of the Drosophila community in xerophytic areas which form continental islands in the neotropical region.This approach is important for further studies in any field considering Drosophila as a model.

Fig. 1 -
Fig.1cluster analysis using presence and absence of species in 8 xerophytic areas in Brazil obtained selecting euclidean distance and the single Linkage method.

Fig. 2 -
Fig.2cluster analysis using number of specimens collected for each species in 8 xerophytic areas in Brazil obtained selecting euclidean distance and the single Linkage method.