The knowledge of the butterfly fauna in Brazil is unevenly distributed through the territory, with a clear concentration in the Southeastern Region along the Atlantic Coast, in the Atlantic Forest domain (Brown & Freitas 1999, Santos et al. 2008). Not coincidentally, this is also the region in Brazil with higher human densities and with a longer history and intensity of anthropogenic disturbance (Brown & Brown 1992, Coimbra-Filho & Câmara 1996, Dean 1996). In the five centuries since the beginning of European colonization, more than 80% of the Atlantic Forest has been converted into pastures, fields and urban areas, and the remaining forest persists as highly fragmented landscapes (Ribeiro et al. 2009). As a consequence, several species of animals and plants in this region are threatened; for example, 51 out of the 55 species of Brazilian threatened butterflies are endemics of the Atlantic Forest, some of them facing real risks of extinction (Machado et al. 2008, Freitas et al. 2011, 2014, Freitas & Marini-Filho 2011).

In spite of the above scenario, there are still large forest remnants in SE Brazil, especially in the complex and heterogeneous landscapes in the montane regions of Serra do Mar and Serra da Mantiqueira, where butterfly diversity peaks in the Atlantic Forest biome (Brown & Freitas 2000). In these mountain ranges, several new butterfly taxa have been described in the last years (Freitas et al. 2012 and references therein), and new taxa are being discovered continuously with further field trips (Freitas 2010).

During an outing in a montane region in the state of Minas Gerais in February, 2011, the first author collected a female of an unknown Phyciodina (Fig. 2). A subsequent expedition to the same site on May 2014 resulted in three additional males collected. Based on facies, this proved to be an undescribed subspecies of Anthanassa dnisilia (C. Felder & R. Felder, 1861), and also the first record of this species of Anthanassa Scudder, 1875 for Southeastern Brazil. The present paper describes this new subspecies and discusses the biogeographical and conservation implications of this discovery.

Study Sites and Methods

Adults were collected and studied in the field in the region of Pocinhos do Rio Verde, Caldas municipality, Minas Gerais, Southeastern Brazil, a montane region with altitudes varying from 1100 to 1700 m, which is part of the Poços de Caldas Plateau. The climate in the region is Cwb of Koppen's classification (Rezende et al. 2013), with average temperature of 18.2°C and mean annual rainfall of 1500 mm, with a dry winter and a wet summer. The region is mostly covered by a mosaic of Seasonal Semidecidual Montane Forest, Dense Ombrophilous High-Montane Forest and Mixed Ombrophilous Forest, including some altitudinal grasslands and rocky fields in the highest peaks (Rezende et al. 2013). The region is now highly fragmented, intermixed with pastures, small plantations including com, coffee and grapevines and Pinus and Eucalyptus plantations. The whole region is considered of extreme biological importance and a conservation priority for several biological groups in Minas Gerais (Drummond et al. 2005). The specific collecting site was “Sítio Mar de Minas” (21°56′41.53″S, 46°24′41.09″W 1170–1190 m a.s.l.), an area of ca. 12 ha of secondary Seasonal Semidecidual Montane Forest neighbor to a large river (the “Rio Verde”) (Fig. 1A).

Fig. 1.

Habitat and live adults of Anthanassa drusilla rioverde in Pocinhos do Rio Verde, Caldas, MG, SE Brazil. Above — general view of the habitat in the type locality in Sitio Mar de Minas; Below — Two views of a living male landing in the vegetation near the forest edge. Photos by TML.

Taxonomic nomenclature followed Lamas (2004) and Wahlberg & Freitas (2007). Species of Anthanassa were studied based on the material available in the Zoology Museum of Unicamp, and compared with the Lamas collection of neotropical butterfly type specimen photographs at the MUSM (also available online in Warren et al. 2013), representing most currently relevant names and recognized species of Anthanassa (Lamas, 2004), and also with the revision of Higgins (1981). The acronyms for the collections are: DZUP, Departamento de Zoologia, Universidade Federal do Paraná, Curitiba, Paraná, Brazil; ZUEC, Museu de Zoologia da Universidade Estadual de Campinas, Unicamp, Campinas, São Paulo, Brazil. The following abbreviations were used: FW: forewing; HW: hind wing; D: dorsal; V: ventral.

Fig. 2.

Holotype male (left) and paratype female (right) of Anthanassa drusilla rioverde.

Genomic DNA was extracted from two legs of freshly caught butterflies, using Invisorb® Spin Tissue Mini Kit (STRATEG Molecular, Germany) in the laboratory at Campinas, and using the QIAgen DNeasy Tissue Kit (Qiagen, Hilden, Germany) in the laboratory at Turku. Four gene regions were sequenced according to published protocols (Wahlberg & Wheat 2008), these were cytochrome c oxidase I (COI, 1471 bp), elongation factor 1 alpha (EF1α, 1240 bp), ribosomal protein S5 (RpS5, 617 bp) and wingless (403 bp). Sequences were aligned with those of other Anthanassa obtained either in our laboratory or from GenBank (Table 2), comprising 6 out the 19 described species within Anthanassa.

The final matrix comprised 34 individuals of Anthanassa and two outgroups, (Table 1). Bayesian analyses (Huelsenbeck et al. 2001, 2002) were carried out for the combined data set under the model GTR+Γ+I, using the program MrBayes 3.2 (Ronquist & Huelsenbeck 2003). Four simultaneous chains were run for 10×10® generations for two runs, sampling trees every 1,000 cycles. The first 2,500 trees were discarded as “burn in” based on when the runs had converged and reached equilibrium.

Discussion

The importance of the present paper goes far beyond the description of a new taxon of Phyciodina in Southeastern Brazil, which by itself is surprising, since this is one of the most thoroughly collected regions for butterflies in Brazil (Santos et al. 2008). However, even more surprisingly, this is the first record of Anthanassa drusilla for Southern South America. Anthanassa drusilla is known to occur from Mexico through Central America and in the Andean region, in Venezuela, Colombia, Ecuador, Peru, Bolivia and Northwestern Brazil (Higgins 1981, Neild 2008, AVLF and K. S. Brown Jr., unpublished data), with one isolated subspecies in the region of Cerro Neblina (Venezuela) (Neild 2008). Based on the available data, the nearest subspecies to A. drusilla rioverde known so far are Anthanassa drusilla verena (Hewitson, 1864), from Cochabamba, Bolivia (ca. 2,000 km Northwest) and Acre, Brazil (ca. 3,000 km Northwest), and A. drusilla higginsi, from Cerro Neblina, Venezuela (ca. 3,000 km North) (Gareca et al. 2006, Neild 2008, AVLF and K. S. Brown Jr., unpublished data). This makes A. drusilla rioverde the most isolated subspecies of A. drusilla, and the only one from Southeastern South America. It is also worth noting that A. drusilla rioverde is much more similar to A. drusilla higginsi than to the geographically nearer A. drusilla verena. The reasons for the isolation of A. drusilla rioverde are unknown, and the similarity in wing pattern with A. drusilla higginsi cannot be assigned to kinship with the available information. A broader study, including samples of most species and subspecies in multiple localities is required to answer this and other questions about the distribution of the species in this genus, and to explain the isolation of A. drusilla rioverde in SE Brazil.

Fig. 3.

Relationships within the genus Anthanassa analysed using Bayesian inference. Posterior probabilities of nodes are given near the branches.

This subspecies is only known from the region of the Poços de Caldas plateau, in southwestern Minas Gerais (searches in three Brazilian museums resulted in no additional individuals found, except the listed paratypes from the H. Ebert collection, now deposited in the Universidade Federal do Paraná), a region considered of extreme biological importance, including several endemisms of plants and animals (Drummond et al. 2005, Machado et al. 2008, Martinelli & Moraes 2013), and a conservation priority in the Atlantic Forest (MMA 2000). The Poços de Caldas Plateau has been intensively surveyed because of its outstanding levels of natural radiation associated with a complex mixture of ores (Amaral et al. 1992). Consequently, the entire region of Poços de Caldas Plateau, has been highly impacted by mining of bauxite, uranium, and several other mineral deposits, with several socio-environmental impacts, and great injury to the local fauna and flora (Fernandes et al. 1995, Barros et al. 2012).

The type locality of A. drusilla rioverde lies within in the “Santuário Ecológico da Pedra Branca” (Pedra Branca Ecological Sanctuary), a protected area in the municipality of Caldas. However, despite its protected status, this region suffers continuously from heavy impacts arising from mining of syenite (an igneous rock used for several purposes in the region) (Conforti et al. 2007).

Table 1.

Species of Anthanassa and outgroups with code, sampling sites data, and GenBank accession numbers for sequenced genes.

Continued.

The entire distribution and conservation status of this subspecies is unknown. However, it has never been recorded in the many localities surveyed in the Mantiqueira range, which extends east and northward into the states of Rio de Janeiro and Minas Gerais, nor in the more extended Serra do Mar or the inland Serra do Espinhaço. Therefore at present we have to presume that this new subspecies is indeed endemic to the Poços de Caldas plateau and possible to adjoining ridges. Hence it is highly recommended that suggestions and guidelines of the ecological zoning of the area are followed (see Conforti et al. 2007). This will not only enhance protection of this notable new endemic butterfly, but also other threatened plants and animals of the entire region, several of which are of high conservation interest.