Ants ( Hymenoptera : Formicidae ) in different green areas in the metropolitan region of Salvador , Bahia state , Brazil

This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. One of the main causes for biodiversity loss is urbanization, mostly due to city growth in highly diverse areas and priority conservation areas, such as the Atlantic Forest (Brasil, 2002; Melo and Delabie, 2017; Conservation International, 2019). Some of the major motivations for the conservation of urban biodiversity, include the preservation of local diversity, the protection of important populations and rare species, as well as the provision of ecosystem services (Dearborn and Kark, 2010). To achieve this, it is essential to understand species diversity and distribution in urban areas. In urban environments, green areas are important habitats for conservation as they house a large number of species (Nielsen et al., 2014; Melo and Delabie, 2017). With more than 16 thousand species/morph species of described ants (Bolton, 2019), data on ants in urban environments has increased worldwide over the last 20 years, with the majority of studies being focused on species distribution (Santos, 2016). In Brazil, more than 490 ant species have been recorded in urban green areas (Melo and Delabie, 2017), with a large number of species being registered in the city of Salvador, Bahia (Melo et al., 2014; Melo and Delabie, 2017). Thus, with the aim of increasing the available knowledge on the ant species from Salvador, recorded by Melo et al. (2014), we present an ant checklist with additional data for their occurrence in green areas. Ants were collected in Salvador, Bahia, Brazil (12°58’S 38°30’W), between April and June 2019 at 62 sample points (SP). At each sample point we removed an area of leaf litter of size 50 × 50cm and installed a bait line as described below. Four green area categories were sampled: fragment (25 samples), road median strip (18), squares (13) and vacant lots (6), totaling 62 samples (Table 1). The SPs were at least 100 meters apart. At each SP we sampled the leaf litter and plants to ensure the detection of ant diversity. To sample the leaf litter fauna, we used a Winkler extractor in units of 50 × 50 cm, where we left the collected material exposed for 24h for fauna extraction. We collected the vegetation fauna in arboreal strata with the adapted bait line technique (Leponce et al., 2019). This technique consists of putting a rope over the top of a tree, using a slingshot. We allocated baits every two meters along the rope from one meter above soil level to the highest tree point, and left them for three hours. We conducted surveys with authorization licenses no 62268-1 from MMA/SISBIO and no 2018-003254/TEC/PESQ-0006 from INEMA/DIRUC. The identification of morph species was acheived following Melo et al. (2014) and species nomenclature followed Bolton (2019). Ants were deposited in the collection of the Laboratório de Mirmecologia (CPDC, curator: J. Delabie), at the Comissão Executiva do Plano da Lavoura Cacaueira (CEPLAC, Itabuna, Bahia, Brazil), voucher #5846. We evaluated the variation in ant species richness according to the type of sampling method and the different types of green areas sampled. Additionally, we measured the similarity between green area types (fragment, road median strip, public square and vacant lots), and we used a Permutational Analysis of Variance (PERMANOVA), with Jaccard’s similarity index as the association measure to evaluate differences in ant species composition. Rarefaction curves and Non-Metric Multidimensional Scaling (NMDS, Jaccard’s distance) were produced in order to compare the richness between the environments. The statistical analyzes were performed using PAST 4.02 software (Hammer et al., 2001). We collected 93 ant species/morph species of 39 genera and six subfamilies (Table 2). Myrmicinae was the richest (S = 57 species), followed by Ponerinae (S = 12), Formicinae (S = 9), Dolichoderinae (S = 7), Pseudomyrmecinae (S = 4), Ectatomminae (S = 3), and Amblyoponinae (S = 1). In Brazil, Melo and Delabie (2017) found a high number of species in urban environments from Atlantic Forest. Five cities in the metropolitan region of Salvador (Bahia, S = 198 species) (Melo et al., 2014), ten cities in an inland city in Santa Catarina (S = 140) (Lutinski et al., 2013), three cities at Alto Tietê (São Paulo, S = 86) (Munhae et al., 2009), and the city of São Paulo (São Paulo, S = 79) (Morini et al., 2007), presented the same number of species found in this study, even Ants (Hymenoptera: Formicidae) in different green areas in the metropolitan region of Salvador, Bahia state, Brazil

One of the main causes for biodiversity loss is urbanization, mostly due to city growth in highly diverse areas and priority conservation areas, such as the Atlantic Forest (Brasil, 2002;Melo and Delabie, 2017;Conservation International, 2019). Some of the major motivations for the conservation of urban biodiversity, include the preservation of local diversity, the protection of important populations and rare species, as well as the provision of ecosystem services (Dearborn and Kark, 2010). To achieve this, it is essential to understand species diversity and distribution in urban areas.
In urban environments, green areas are important habitats for conservation as they house a large number of species (Nielsen et al., 2014;Melo and Delabie, 2017). With more than 16 thousand species/morph species of described ants (Bolton, 2019), data on ants in urban environments has increased worldwide over the last 20 years, with the majority of studies being focused on species distribution (Santos, 2016). In Brazil, more than 490 ant species have been recorded in urban green areas (Melo and Delabie, 2017), with a large number of species being registered in the city of Salvador, Bahia (Melo et al., 2014;Melo and Delabie, 2017). Thus, with the aim of increasing the available knowledge on the ant species from Salvador, recorded by Melo et al. (2014), we present an ant checklist with additional data for their occurrence in green areas.
Ants were collected in Salvador, Bahia, Brazil (12°58'S 38°30'W), between April and June 2019 at 62 sample points (SP). At each sample point we removed an area of leaf litter of size 50 × 50cm and installed a bait line as described below. Four green area categories were sampled: fragment (25 samples), road median strip (18), squares (13) and vacant lots (6), totaling 62 samples (Table 1). The SPs were at least 100 meters apart. At each SP we sampled the leaf litter and plants to ensure the detection of ant diversity. To sample the leaf litter fauna, we used a Winkler extractor in units of 50 × 50 cm, where we left the collected material exposed for 24h for fauna extraction. We collected the vegetation fauna in arboreal strata with the adapted bait line technique (Leponce et al., 2019). This technique consists of putting a rope over the top of a tree, using a slingshot. We allocated baits every two meters along the rope from one meter above soil level to the highest tree point, and left them for three hours. We conducted surveys with authorization licenses nº 62268-1 from MMA/SISBIO and nº 2018-003254/TEC/PESQ-0006 from INEMA/DIRUC. The identification of morph species was acheived following Melo et al. (2014) and species nomenclature followed Bolton (2019). Ants were deposited in the collection of the Laboratório de Mirmecologia (CPDC, curator: J. Delabie), at the Comissão Executiva do Plano da Lavoura Cacaueira (CEPLAC, Itabuna, Bahia, Brazil), voucher #5846. We evaluated the variation in ant species richness according to the type of sampling method and the different types of green areas sampled. Additionally, we measured the similarity between green area types (fragment, road median strip, public square and vacant lots), and we used a Permutational Analysis of Variance (PERMANOVA), with Jaccard's similarity index as the association measure to evaluate differences in ant species composition. Rarefaction curves and Non-Metric Multidimensional Scaling (NMDS, Jaccard's distance) were produced in order to compare the richness between the environments. The statistical analyzes were performed using PAST 4.02 software (Hammer et al., 2001).
A higher number of species was found in the leaf litter (Winkler extractor, S = 77; 46 exclusive species) than in the vegetation (bait line, S = 47; 16 exclusive species). Ant assemblage composition shows important differences according to strata (mostly in soil when compared to vegetation) (Wilson and Hölldobler, 2005). Thus, a lower richness of arboreal species is expected as evolutionary history shows that ground ants specialize in resource selection, nesting places, and dispersal mechanisms and Ants in green areas in Salvador, Brazil composition according to the type of green areas (F 3,58 = 2.07; p < 0.001). Ant assemblages from road median strips and public squares were 42% more similar and 35% of those were similar to ants from vacant lots (Figure 3). Ants from forest fragments were 30% similar to all other green urban areas here studied. We recorded 10 species common to all green areas:  (Table 2). A high richness of native ants has been reported in cities with different levels of anthropogenic disturbance (Santos, 2016;Melo and Delabie, 2017). Conserved environments, such as native fragments, have higher ant species richness therefore, have an advantage when compared to vegetation assemblages (generalists) (Wilson and Hölldobler, 2005). The sampling technique may be related to the lower arboreal richness observed, since baits can attract only a proportion of ant assemblies. Despite the recorded arboreal ant richness, nine new species were sampled: Azteca prox. alfari, A. severini Emery, 1896, Cephalotes atratus (Linnaeus, 1758), Dorymyrmex pyramicus (Roger, 1863), Myrmelachista sp.1, Pseudomyrmex sp.5 (group pallidus), P. curacaensis (Forel, 1912), Rogeria subarmata (Kempf, 1961) and Xenomyrmex sp.1.
Among the different types of green urban areas, forest fragments showed higher richness (S = 74 species; 32 exclusive species), followed by road median strips (S = 44; three exclusives), squares (S = 37; five exclusives), and vacant lots (S = 26; one exclusive) (Figure 1 and 2). We detected significant differences in ant species  compared to impacted ones such as road median strips, public squares, and vacant lots (Melo and Delabie, 2017). Therefore, we highlight the importance of maintaining more conserved green areas in order to preserve ant species. Although additional research and methods would allow for the detection of more ant species, this study has deepened the knowledge available on ants from Salvador.