Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Hitting an Unintended Target: Phylogeography of Bombus brasiliensis Lepeletier, 1836 and the First New Brazilian Bumblebee Species in a Century (Hymenoptera: Apidae)


This work tested whether or not populations of Bombus brasiliensis isolated on mountain tops of southeastern Brazil belonged to the same species as populations widespread in lowland areas in the Atlantic coast and westward along the Paraná-river valley. Phylogeographic and population genetic analyses showed that those populations were all conspecific. However, they revealed a previously unrecognized, apparently rare, and potentially endangered species in one of the most threatened biodiversity hotspots of the World, the Brazilian Atlantic Forest. This species is described here as Bombus bahiensis sp. n., and included in a revised key for the identification of the bumblebee species known to occur in Brazil. Phylogenetic analyses based on two mtDNA markers suggest this new species to be sister to B. brasiliensis, from which its workers and queens can be easily distinguished by the lack of a yellow hair-band on the first metasomal tergum. The results presented here are consistent with the hypothesis that B. bahiensis sp. n. may have originated from an ancestral population isolated in an evergreen-forest refuge (the so-called Bahia refuge) during cold, dry periods of the Pleistocene. This refuge is also known as an important area of endemism for several animal taxa, including other bees. Secondary contact between B. bahiensis and B. brasiliensis may be presently prevented by a strip of semi-deciduous forest in a climate zone characterized by relatively long dry seasons. Considering the relatively limited range of this new species and the current anthropic pressure on its environment, attention should be given to its conservation status.


Despite the fact that, taxonomically, bees are among the best known insects in the world (for example [1]), there are taxonomic problems to be solved in the group, including sex association in dimorphic species and distinction of cryptic species, even in relatively well-known taxa. In these cases, molecular data, including the so called DNA barcode, alone or integrated to other types of characters, have proven to be much useful (for example [2, 3]).

The bumblebees (genus Bombus Latreille, 1802) comprise approximately 250 species widely distributed in the world, but occurring mainly in the cool subtropical and temperate areas of the Nearctic and, especially, Palearctic regions [4]. Defining species limits in Bombus is frequently difficult, because of the lack of useful structural characters and the great variability in hair-color patterns. This has led to an array of recent studies employing DNA markers to elucidate bumblebee-species boundaries (for example [512]). Molecular tools have also been used in the genus to investigate intraspecific genetic structure, phylogeography and phylogenetic relationships (for example [1322]).

In South America, most of the species of Bombus are distributed along the Andes and in temperate regions, with only a few species recorded in the warm lowlands—the later, actually, the only bumblebees to occur in such environments in the world [1]. Only six species of the genus are generally referred to occur in Brazil [23], all belonging to the same subgenus—Fervidobombus Skorikov, 1922 or Thoracobombus Dalla Torre, 1880, depending on which classification one adopts (for example [1, 24] or [25]). Five of these species (B. bellicosus Smith, 1879; B. brasiliensis Lepeletier, 1836; B. brevivillus Franklin, 1913; B. pauloensis Friese, 1913; and B. transversalis (Olivier, 1789)) seem to be very closely related, while the sixth one, B. morio (Swederus, 1787) belongs to a distinctive clade in the same subgenus [19, 21]. Bombus pauloensis has been widely treated as B. atratus Franklin, 1913 in the literature (for an explanation for the adoption of B. pauloensis as a valid name, see [24]). The occurrence of a seventh species in western Brazil, B. pullatus Franklin, 1913, was reported by Milliron [26], but considered with suspicion by Abrahamovich & Díaz [27], and ignored by Moure & Melo [24], and needs confirmation. An eighth species, B. rubriventris Lepeletier, 1836, was recorded with doubts by Milliron [26] as possibly occurring in the state of Goiás. According to him, this species, known only from its female holotype, would be very rare or extinct. One cannot rule out the possibility that the assignment of B. rubriventris to Brazil resulted from mere labeling mistake or from wrong interpretation of the type locality (see [26] for details).

In their thorough revision of the Brazilian bumblebees, Moure & Sakagami [23]) noted that B. brasiliensis Lepeletier, 1836 was the only species to be more common on mountaintops (mainly in the ranges along the southern and eastern Brazilian coasts) than in lowlands, and that it was the only one recorded in elevations above 1,800 m. Later, Silveira & Cure [28] noted that B. brasiliensis is absent in the lowlands surrounding the mountain ranges where it is relatively common further inland in southeastern Brazil. The occurrence of these isolated mountaintop populations of B. brasiliensis raised the suspicion that they might in fact belong to a species different from the one present in the lowlands along the coast and in the Paraná River valley. Here, this hypothesis is tested, mainly by means of phylogenetic and population-genetic analyses of two mtDNA genes. Since the question of how to delimit species boundaries is still in debate (see, for instance, the revision by Wiley & Lieberman [29]), different character sources (molecular and morphological) and analytical methods (tree and non-tree based) were complementarily employed to decide whether or not a set of populations should be considered as a new taxon.

Materials and Methods

Morphological and Molecular Procedures

The description of the new species presented below was based on six specimens—three workers and a queen from the municipality of Ilhéus, in the Brazilian state of Bahia, and two workers collected in the municipality of Conceição da Barra, state of Espírito Santo. These specimens are deposited at the Taxonomic Collections of the ‘Universidade Federal de Minas Gerais’—UFMG—and in the ‘Padre Jesus Santiago Moure’ entomological collection, of the ‘Universidade Federal do Paraná’—DZUP—, as detailed in the section “Taxonomic treatment”, below. Two collecting expeditions were conducted in southern Bahia and northern Espírito Santo states (coordinates of sampling sites are, in Bahia: Eunápolis—16°25'8"S, 39°34'55"W, 162m; Itamaraju—16°58'39"S, 39°33'16"W, 81m; Porto Seguro—16°27'3"S, 39°17'16"W, 114m; and, in Espírito Santo: Conceição da Barra—18°20'54"S,39°51'06"W, 47m; São Mateus—18°45'13"S, 39°51'39"W, 39 m; Sooretama—19°03'01"S, 40°08'02"W, 94m), under the collecting permit number 23784 (granted by “Instituto Chico Mendes de Conservação da Biodiversidade” to JESJ), in an attempt to increase molecular sample size and geographic representation. No additional specimens were found, however, during these expeditions.

The specimens were examined under a dissection microscope (Leica M125) and compared to the descriptions in Moure & Sakagami [23]. Comparisons were also made with photographs of the lectotype of B. brasiliensis Lepeletier, 1836, reproduced here in Fig 1A and 1B. Direct observation of this type was not done because, after examining hundreds of specimens of B. brasiliensis from all its geographic range, an easily-observable diagnostic character was found to distinguish the two concerned species, which could be readily checked in the photographs obtained from the type. Morphological terminology employed here is mostly that of Moure & Sakagami [23]. Flagellomeres are designated as F1, F2, F3 etc. Accordingly, metasomal terga and sterna are designated as T1, T2, T3 etc and S1, S2, S3 etc. The apical width of the malar area is the shortest width of the malar area, measured along the mandible base.

Fig 1. Types of Bombus brasiliensis Lepeletier, 1836 and Bombus bahiensis sp. n.

Lectotype of B. brasiliensis—A: lateral view of head, mesosoma and metasoma; B: dorsal view of mesosoma and metasoma. Holotype of B. bahiensis sp. n.—C: lateral view of head, mesosoma and metasoma; D: dorsal view of mesosoma and metasoma.

Two mitochondrial markers, Cytochrome oxidase I (COI) and Cytochrome B (CytB) were employed in addition to the morphological characters. DNA was extracted from one of the hind legs of each of 168 Bombus specimens, 142 of them belonging to B. brasiliensis, three to the species described as new below, and the rest belonging to the other species of the genus known to occur in Brazil (S1 Table). This was done through the phenol–chloroform method [30]. The extracted DNA was re-suspended in 40μL of TE buffer. The 5’region of the COI gene was amplified using the primers LepF1 and LepR1 [31], mtd_6 and mtd_9 [32], and that of the CytB was amplified using the primers mtd_26 [32] and CytB_R1, 5'TTCAATTATTTGACTTCCTAATCAAG3' (designed for this analysis). Amplification of both genes was done in a 30μL polymerase chain reaction (PCR) mix, including 0.5 units of Taq DNA polymerase, 2 mM MgCl2, in 1x PCR buffer, 0.5 μM of each primer, 2.5 mMdNTPs and about 20 ng of genomic DNA. Amplification was carried out in a thermocycler using a program consisting of 5 min of denaturation at 94°C, followed by 35 30-second cycles at 94°C, 40 s at 48–51°C, 1 min at 72°C, and a final extension for 10 min at 72°C. PCR products were visualized in a 0.8% agarose gel. All PCRs that generated a single product were purified using polyethylene glycol 20% (PEG) precipitation (modified of Sambrook & Russel [30]—S2 Table). This purified PCR product was sequenced in both directions using a MegaBACE 1000 sequencer (GE Healthcare) with the same amplification primers following the manufacturer's recommendations. The raw sequences were obtained using the software Phred v. 0.20425 [33]. The final sequences were assembled with the software Phrap v. 0.990319 (http// Consed 19.0 [34] was used to view and edit the high quality consensus sequences. Alignments of the consensus sequences for all individuals were generated using MUSCLE [35] implemented in the program MEGA 5.01 [36].

Molecular extraction and sequencing were done in the Biodiversity and Molecular Evolution Lab (LBEM), at the Universidade Federal de Minas Gerais, Belo Horizonte, Brazil. In addition to the 306 sequences generated here (168 COI, and 138 CytB sequences) for five species, 49 COI sequences available in the GenBank were employed for 16 species belonging to the subgenus Thoracobombus (sensu Williams et al. [25]) (Table 1). Additionally, sequences from three others specimens of B. bellicosus, stored in the Barcode of Life Database, and kindly made available by Dr. P. Tubaro, from the Museo Argentino de Ciencias Naturales, were also employed (Table 1).

Table 1. Specimens belonging to GenBank and BOLD Systems databases, which were used in analyzes with the COI gene.

The computer programs Arlequin version [37], DnaSP version 5.10.01 [38] and MEGA 5.01 were used to estimate the following intra and interpopulational parameters: 1) haplotype diversity (H); 2) average number of nucleotide differences (k); 3) mean number of pairwise differences (π); 4) number of polymorphic sites (S); and 5) Tajima’s D and Fu’s FS test of selective neutrality. DnaSP was also used to verify non-synonymous and synonymous substitutions. Haplotype networks constructed using the median-joining algorithm (MJ) [39], available in the NETWORK 4.5 software, were used for inferences about phylogenetic relationships among haplotypes and their possible geographical correlation. Population analyses were performed separately for B. brasiliensis (124 specimens) and B. bahiensis sp. n. (3 specimens) using a CytB and COI concatenated matrix (127 specimens).

Average intra and interspecific genetic distances were obtained with Mega 5.01, using the parameters of the Kimura 2 model—K2P [40]. Two analyses were performed with different data sets, one using only the COI gene and the other employing a concatenated data set of the COI and CytB genes. This was done because: 1) There were more COI than CytB sequences available in GenBank and BOLD for species of Thoracobombus (sensu Williams et al.[25]); and 2) There were no CytB sequences available for many of the specimens from which DNA was extracted for this work.

Since primers used here were not the same as those employed for obtaining the sequences in GenBank and BOLD, only 402 bp were present in all sequences used in the COI-only analyses. Intraspecific distances were estimated only for species represented by three or more specimens, with at least two different haplotypes.

Phylogenetic analyses were performed using a concatenated matrix with CytB and COI-gene data, including a total of 138 specimens, with B. morio as the outgroup. The best fit substitution model estimated with Modeltest 3.7 [41] for this analysis was the GTR+G model.

Phylogenies were generated through Bayesian Inference (BI) in MrBayes 3.1 [42], and through Maximum Parsimony (MP) and Maximum Likelihood (ML) algorithms using the program PAUP* 4.0b10 [43]. Inferences of trees with MP and ML methods were performed using heuristic search, with the following parameters: stepwise addition (random) starting from a single initial tree and branch swapping (tree bisection and reconnection—TBR). MP analysis was set to retain up to 1000 most parsimonious trees. The bootstrap method was used as a measure of branch support for the recovered phylogenies, using a total of 10,000 and 100 replications, respectively, for MP and ML. Phylogenetic analyses using BI used two sets of Markov chains, each containing three hot chains and one cold, with 20 million generations with a 25% burn-in, to seek for convergence to the same subset of best trees.

Nomenclatural Acts

The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "". The LSID for this publication is: The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.


Concatenated population analyses

The phylogenetic relationships among haplotypes and their geographical correlation, using the concatenated CytB (683 bp) and COI (471 bp) sequences (totaling 1154 bp), showed that populations from almost all the geographic range of B. brasiliensis share haplotypes, as shown by the haplotype network, which suggests lack of population structure (Fig 2). In contradiction to the initial hypothesis, thus, this result suggests that populations of B. brasiliensis isolated in mountain tops away from the coast belong in the same species as those living in coastal forests from the state of Rio de Janeiro southward to the state of Santa Catarina and westward to Paraguay in the Paraná-river valley.

Fig 2. Network of haplotypes for Bombus brasiliensis Lepeletier, 1836 and Bombus bahiensis sp. n.

Median-joining network of haplotypes for B. brasiliensis and B.bahiensis sp. n. The B. brasiliensis on isolated moutaintops in southeastern Brazil are coloured of different shades of green and the B. brasiliensis continuously distributed are coloured of different shades of blue. B.bahiensis sp. n. is coloured yellow. All lines joining haplotypes are one mutation step long, except for three of them, marked with “a”, “b” and “c”, which are three, two and 38 steps long, respectively.

On the other hand, these results indicate that the three specimens collected in the coastal forest of southern Bahia state (municipality of Ilhéus, all with the same haplotype) diverged from B. brasiliensis by 38 mutation steps (Fig 2), suggesting that this population belongs to a different species (described below as B. bahiensis sp. n.)

The population analysis of B. brasiliensis included a total of 124 specimens, from 30 localities (Fig 3). It resulted in 27 haplotypes, 17 polymorphic sites (11 synonymous and six nonsynonymous mutations), 10 of which were parsimoniously informative and seven were autapomorphies (singleton sites). The total haplotype diversity (Hd), nucleotide diversity (π), and average number of nucleotide differences (K) were 0.886, 0.003 and 3.057, respectively. The statistics for the neutrality tests were D = -0.083 (p = 0.5) for Tajima’s, and FS = -12.28 (p = 0.001) for Fu’s. Although the Tajima’s D was not statistically significant, the significant value of Fu’s FS, associated with the large number of unique haplotypes and the star-shaped haplotype network, suggests recent population expansion.

Fig 3. Distribution of Bombus brasiliensis Lepeletier, 1836 and Bombus bahiensis sp. n.

Records for B. bahiensis are represented by triangles (the black triangle, the gray triangle and the gray triangle with black point represents a population used in mtDNA analyzed, a population used in morphologic analyzed and a population inferred to belong to the species based on descriptions in Moure &Sakagami [23], respectively); records for B. brasiliensis are represented by circles (the black circle denote sites represented by mtDNA). Records were compiled from the literature [23, 60, 73] and from specimens deposited at UFMG and/or listed in S1 Table. Acronyms represent Brazilian states, as follows: BA = Bahia; ES = Espírito Santo; MG = Minas Gerais; RJ = Rio de Janeiro; SP = São Paulo; PR = Paraná; SC = Santa Catarina; RS = Rio Grande do Sul; MT = Mato Grosso; GO = Goiás; MS = Mato Grosso do Sul; DF = Distrito Federal. Localities mentioned in the text are indicated. The stippled line indicated by black arrow represents the Doce river.

The genetic distance between B. bahiensis sp. n. and B. brasiliensis was 3.7 ± 0.6%, while the average intraspecific distance within B. brasiliensis was 0.26 ± 0.09%. The population parameters above were not calculated for B. bahiensis sp. n., since all three specimens had the same haplotype.

COI-only analyses

Of the 402 bp used in the COI-only analyses, 241 were constant and, among the variable characters, 28 were parsimony-uninformative, and 133 parsimony-informative. The mean genetic distance between B. bahiensis sp. n. and B. brasiliensis estimated with the COI gene was 2.89 ± 0.8%. Among the comparisons made for the 21 Thoracobombus species employed in this analysis, distances smaller than that between B. brasiliensis and B. bahiensis were obtained for seven other species pairs (S3 Table). Moreover, while the largest distance between any two specimens of B. brasiliensis was of 1 ± 0.5%, the smallest distance between a specimen of B. brasiliensis and a specimen of B. bahiensis was of 2.5± 0.8%.

The intraspecific distances among the Thoracobombus species analyzed here, estimated with the COI sequences, were all equal or smaller than 0.5% (Table 2). One exception was the mean intraspecific distance for B. brevivillus, which was very high when all specimens originally attributed to this species where considered together (Table 2). Moreover, the average distance between the two “B. brevivillus” clades obtained in the phylogenetic results described below is 3.92 ± 0.93%. This suggests that the samples considered here as belonging to B. brevivillus include specimens of two distinct species (B. brevivillus (1) and (2) in S3 Table). This finding is being examined in detail and results will be published elsewhere.

Table 2. Genetic distances for COI sequences (%) within species of bumblebees.

Phylogenetic analyses

Of the 1157 bp employed in the phylogenetic analyses (686 bp for CytB and 471 for COI), 916 were constant, three were gaps; eight variable characters were parsimony-uninformative, and 233 were parsimony-informative. Those three gaps are due to one CytB codon appearing exclusively in B. morio (included as the outgroup) and were the cause for the difference in pair-base numbers between the data employed in the phylogenetic and the population analyses.

The phylogenetic analyses yielded the same single topology for Parsimony, Maximum Likelihood and Bayesian analysis (Fig 4). This tree was yielded after 1,102,282 rearrangements in the MP analyses and was 302-steps long. The average standard deviation of split frequencies after 20 million generations in the Bayesian analysis was of 0.002869, indicating convergence to a single subset of trees.

Fig 4. Phylogenetic relationships among Brazilian bumblebees obtained from concatenated phylogenetic analyses performed with CytB and COI sequences, using Bayesian inference and all specimens available for each species.

Values shown under each branch are posterior probabilities. The values shown above each branch are bootstrap values represented here by Maximum Parsimony (MP)/ Maximum Likelihood (ML), because an identical topology was obtained with MP and ML. An analysis employing only unique haplotypes yields the same topology. “*” Represents the same values for MP and ML. Bombus morio (Swederus, 1787) was used as outgroup. The collapsed branch includes all Bombus brasiliensis haplotypes found on the haplotype network from Fig 2 and S1 Fig.

The phylogenetic analyses are congruent with the population analyses in suggesting that populations of B. brasiliensis isolated on the top of mountain ranges in southeastern Brazil belong to the same species as those in the Atlantic coast and in the lowlands of the Paraná-river valley (S1 Fig). They also show that specimens originally attributed to B. brasiliensis in the coastal forest of southern Bahia constitute a clade (B. bahiensis sp. n.), sister to the rest of the populations of B. brasiliensis.

Morphology and taxonomic decisions

Examination of the external morphology of specimens attributed by the molecular analyses to B. brasiliensis and to B. bahiensis sp. n. revealed a character also allowing for their distinction: The pilosity of T1, which is entirely yellow in B. brasiliensis and mostly black in B. bahiensis (see key below). Furthermore, this character led to the recognition that two additional specimens from Conceição da Barra (northern Espírito Santo state) also belong in B. Bahiensis sp.n. (these latter specimens were too old and DNA extraction from them was not attempted). Considering the molecular and morphological evidence, the specimens from southern Bahia attributed to B. brasiliensis in our samples are considered here to be a species distinct from that containing the remaining populations of B. brasiliensis.

Two pieces of evidence indicate that the name B. brasiliensis should be applied to the species occurring along the coast of Rio de Janeiro and southwards: a) the holotype of B. brasiliensis has its first metasomal tergum covered with yellow pilosity (Fig 1B), as opposed to black, as in specimens from southern Bahia (Fig 1C and 1D); and b) the fact that this type specimen was probably collected in the vicinities of the city of Rio de Janeiro [23].

Moreover, two synonyms are recognized for B. brasiliensis: B. venustus Smith, 1861 and B. brasiliensis var. palliventris Friese, 1931 [23, 24], the former described from Teresópolis (on the top of the Serra do Mar mountain range, in the state of Rio de Janeiro—Fig 3) and the latter from Paraguay. The fact that populations sampled in Paraguay and on the mountaintops in southeastern Brazil showed to be conspecific with specimens from the southern Atlantic coast in the phylogenetic and population-genetic analyses, indicate that those names should continue to be considered as synonyms of B. brasiliensis. Taking all above in account, the populations occurring to the north of the Doce river, in the coastal forests of northern Espírito Santo and southern Bahia, are described below as a new species.

Taxonomic treatment

Bombus bahiensis Santos Júnior & Silveira sp. n.

Fig 1C and 1D

Diagnosis. No single morphological character is known to be exclusive of Bombus bahiensis sp. n., which can be distinguished from most other Brazilian bumblebee species, except B. brasiliensis, for the mesosoma dorsally yellow, traversed by a black intertegular band, which extends laterally over the lower half of the mesepisternum (and generally reaching its ventral area), and the predominantly yellow metasoma traversed by a black band in T3; from B. brasiliensis it can be distinguished by the mostly black-haired T1, with yellow and partially-yellow hairs restricted to the mid-apical margin of the tergum. Only two species are potentially sympatric with B. bahiensis sp. n., B. pauloensis (melanic form) and B. brevivillus, which can be readily distinguished from B. bahiensis for their entirely black pilosity.

Description (worker holotype). Measurements (mm): approximate body length—16.2; head (length:width)– 4.5: 4.0; labrum (length: width)– 0.4:1.7; malar area (length × width)– 0.9:1.3; clypeocellar distance—1.9; interantennal distance—0.7; antennocular distance—0.6; ocellocular distance—0.6; interocellar distance—0.7; ocellar diameter—0.2; ocelloccipital distance—0.8; scape (length:diameter)– 2.1:0.3; length of F1 and F2–0.5, 0.3; forewing length—14.3; length of 2ndand 3rd submarginal cells (measured on their posterior margins)– 1.7, 1.3.

Structure: labrum biconvex; disc of clypeus depressed; malar area slightly shorter than wide; interocellar distance larger than ocellorbital.

Body color: black, except reddish-brown on posterior tibia; wings dark brown, with light cupreous hue.

Body surface: clypeus coarsely punctate, except on disc, irregularly punctate with shiny interspaces; area between lateral ocelli and eye impunctate, smooth and shiny, except for a micro-punctate region on upper paraocular area, near upper inner-margin of compound eyes, which is narrower than mid-ocellar diameter, generally occupying one third of ocellorbital distance; glabrous area on disc of mesoscutum ill-defined, mostly restricted to region posterior to median mesoscutal line, lightly micro-reticulate but shiny; glabrous shiny area on inner surface of hind femur relatively narrow (less than half the width of the surface) occupying the apical three-fourths of the femoral length.

Pubescence: on head, black bristles intermixed with greyish plumose hairs, most abundant between antennal sockets; on mesosoma, long, fine hairs with long branches, light yellow, except for black intertegular band; on inner surface of hind tibia, fine, spatulate, the flat apical portion triangular, restricted to the very tip of each seta; on T1- T5 long and fine, with long branches, on T3 light yellow, on other terga, black; on T6 black and short; on sterna and legs, black and short.

Variation: the integument varies from black to reddish brown, especially on legs and metasoma; black hairs with light-yellow tips may occur on T1.

Queen. Similar to worker, except for the vertex much elevated above level of ocelli, as normally occurs in Bombus queens.

Male. Unknown.

Holotype. “Ilhéus BA; Brasil 18/11/2009; A. Nemésio”, “Euglossina da Hiléia Baiana, Campus UESC; 16724–47065”, “Bombus bahiensis; Santos Júnior & Silveira, sp. n.; HOLOTYPUS”. Collected in flight. Deposited at UFMG (accession number 16724–47065). The holotype lacks the right hind-leg, removed for DNA extraction.

Paratypes. “Euglossina da Hiléia Baiana, Campus UESC; 18339–52893”, “Ilhéus BA; Brasil 20/02/2010; A. Nemésio”, “Bombus bahiensis; Santos Júnior & Silveira sp. n.; PARATYPUS”; “Euglossina da Hiléia Baiana, Campus UESC; 18339–52894”, “Ilhéus BA; Brasil 20/02/2010; A. Nemésio”, “Bombus bahiensis; Santos Júnior & Silveira, sp. n.; PARATYPUS” (both workers, deposited at UFMG). “Conceição da Barra ES; Brasil 31/01/1969; C. Elias & T. Elias”, “Bombus bahiensis; Santos Júnior & Silveira, sp. n.; PARATYPUS” (two workers, deposited at DZUP).“Ilhéus BA; Brasil 2003; M. A. Costa”, “Bombus bahiensis; Santos Júnior & Silveira, sp. n.; PARATYPUS” (queen, deposited at DZUP).

Etymology. The name bahiensis refers both to the Brazilian state of Bahia, where the holotype and some of the paratypes were collected, and to its habitat, the Bahia Forest (see explanation below, under “distribution”).

Distribution. The species seems to be restricted to the so-called Bahia Forest, which is the especially luxuriant Atlantic Forest formation that covers southern Bahia and northern Espírito Santo states, in Brazil (Fig 3).

Key to females of Brazilian species of Bombus Latreille, 1802.

(Modified from [23])

  1. - Pilosity entirely black, except for apically-pale hairs on ventral parts of body ........ 2
    - Pilosity at least partially yellow on mesosoma, metasoma or both, forming bands or entirely covering one or more terga .......................................................... 4
  2. - Malar area longer than its apical width; posterior glabrous area on disc of mesoscutum well delimited, micro-reticulate and dull ....B. morio (Swederus, 1787)
    - Malar area shorter than its apical width; posterior glabrous area on disc of mesoscutum poorly or well delimited but never micro-reticulate and shiny ........... 3
  3. - Micro-punctate region on upper paraocular area, near upper inner-margin of compound eyes wide, as wide as or wider than mid-ocellar diameter, occupying half of ocellorbital distance; pilosity relatively dense, velvety, especially dorsally on mesoscutum of queens ....... . . .. . . .. . .. . ....... B. brevivillus Franklin, 1913*
    - Micro-punctate region on upper paraocular area, near upper inner-margin of compound eyes narrow, narrower than mid-ocellar diameter, generally occupying one third of ocellorbital distance; pilosity looser, not velvety ................................... ......................................................... pauloensis Friese, 1913 (black form)
  4. - Pilosity of pronotum, mesoscutum and scutellum light-yellow to yellowish-brown, without black intertegular hair-band; pilosity on T4-T6 ferruginous ................................... ......................................................... B. bellicosus Smith, 1879
    - Yellow pilosity on pronotum, mesoscutum and scutellum relatively well developed, but always with an intertegular black band; T3 generally with yellow hair band; T4-T6 black-haired ................................................................................. 5
  5. - Yellow pilosity on mesosoma never totally covering meso and metepisterna, which are black haired on lower half, including venter ........................................... 6
    - Yellow pilosity entirely covering meso and metepisterna, frequently reaching venter of mesosoma ........................................ . ................................ . ............ 7
  6. - Posterior glabrous area on disc of mesoscutum ill-defined, with sparse punctures reaching its middle; pilosity on pronotum, scutellum and T3 deep yellow, black intertegular band narrower than pronotal band ........ B. transversalis (Olivier, 1789)
    - Posterior glabrous area on disc of mesoscutum well defined, with no punctures on its middle; yellow pilosity entirely light; intertegular hair band wider than pronotal band ........................................... B. pauloensis Friese, 1913 (yellow-banded form)
  7. - T1 covered with yellow pilosity .........................B. brasiliensis Lepeletier, 1836
    - T1 covered with black pilosity, frequently intermingled with yellow or partially yellow hairs mid-apically ................ B. bahiensis Santos Júnior & Silveira sp. n.

* Probably a compound species (see text).


Variation discontinuity, Phylogeny and Species Status

The genetic distance between B. bahiensis sp. n. and B. brasiliensis was relatively small, and this may raise the suspicion that they might in fact be a single species. However, comparable small divergences have been commonly estimated for several species-pairs believed to have diverged recently, in other bee taxa. For example, divergences found for COI-sequences by Dick et al. [44] and Nemésio et al. [45] between species pairs of orchid-bee (Euglossina) in South America were all below 2%, while Gibbs [2] found only a 3.06% average COI-sequence divergence among five species of Lasioglossum previously misinterpreted as a single species (with the lowest recorded value of 1.7%). In Bombus, such small genetic distances were found, also, among cryptic species in northern Europe ([5, 10], see also Table 5 in [12]). Moreover, although the average genetic distance between B. brasiliensis and B. bahiensis sp. n. was the smallest recorded between pairs of Brazilian species of Bombus in this study, it was higher, for example, than those obtained for seven other pairs of undisputed species in Asia, Europe, and South America (S3 Table). The fact that the smallest distance between a specimen of B. brasiliensis and one of B. bahiensis is 2.5 times larger than the largest distance between any two specimens of B. brasiliensis is an additional support for a genetic gap between B. bahiensis and B. brasiliensis.

Traditionally, discontinuity in variation among populations has been considered, both in morphological and molecular-based studies, as evidence that independently-evolving lineages are involved (for example [29]). In the barcoding context, average intra-population divergences much smaller than the inter-population divergences (the so called barcode gap—[46]) is considered evidence that they are cohesive sets of populations, isolated from each other. Much has been discussed in the literature about the reality and convenient thresholds of such a gap (for example [46]). However, there is some agreement that one is safe in recognizing different species when the barcode gap between them reaches the order of one magnitude, as in the case of B. brasiliensis and B. bahiensis sp. n. A problem here is that large intra-specific divergence could not be expected for B. bahiensis sp. n., considering that it is represented by a small sample from a single site. However, the situation could be looked at from the other way around: B. brasiliensis from sites more than 1500 km apart along the latitudinal gradient or from isolated populations in areas differing more than 1500 m in elevation did not show divergences as large as that found between populations of B. bahiensis sp. n. and those of B. brasiliensis about 240 Km apart in lowland evergreen forests.

The fact that B. brasiliensis and B. bahiensis sp. n. are reciprocally monophyletic also supports their recognition as separate species (for example [47]), as it suggests that each of them is a mutually-independent evolutionary lineage. It could be argued that the phylogenetic hypothesis produced here, being based on two linked genes (maternally-inherited mtDNA), may merely indicate a phylogeny for those genes, which may be different from the correct phylogeny of the species involved (for example [48]). Nevertheless, as pointed out by Williams et al. [49], this problem has not been detected for the use of COI in bumblebees and, moreover, mtDNA may actually be more suitable for species delimitation than other molecular and morphological data [50].

Biogeographic considerations

Before the arrival of the Portuguese settlers, the presumed geographic range of B. bahiensis sp. n. was covered by an especially luxurious evergreen forest (from now on, the Bahia forest), coinciding with an important area of endemism for many plant and animal taxa, including insects [51], and including several recently-described endemic species of orchid bees (Apidae: Euglossina) (for example [5256]). Intra- and interspecific disjunctions of populations or sister-species involving this area and other areas in the Amazonian Forest or to the south in the Atlantic Forest also have been recorded for several taxa (for example [5759]).

A closer look to the distributions of B. brasiliensis and B. bahiensis sp. n. shows a 240-km gap between the northernmost known record of the species in the coastal lowlands of Rio de Janeiro state (in the municipality of São Francisco de Itabapoana, see Fig 3 and S1 Table), and its southernmost record in the lowlands of Espírito Santo state (municipality of Linhares; [23]—see comments on this population below) (Fig 3). Bombus brasiliensis does occur in southeastern Brazil, in latitudes between those of Linhares and São Francisco de Itabapoana [23, 28, 60], but all these records (and additional localities represented in the UFMG collection) refer to cloud forests (or their immediate vicinities) on the top of mountain chains, further inland. These mountaintop records represent populations which are isolated from those near the coast (Fig 3) by semi-deciduous forests in areas under climates with relatively long dry seasons (4–6 months)—the same kind of environment found in the gap between the humid northern area occupied by B. bahiensis sp. n. and the humid southern area occupied by B. brasiliensis (compare, for example, the map in Fig 3 with that of the Atlantic Forest vegetation types presented by Carnaval et al. [61] in their supplementary material’ Figure S1).

Close relationship between elements of the Bahia forest and the Amazonian Forest biotas has been pointed out for organisms such as birds (for example [62]) and also for orchid bees, with several pairs of presumed sister species of the latter occurring disjunctly in the two phytogeographic domains (for example [54, 55]). This may suggest that B. bahiensis sp. n. could be sister to the only species in the genus restricted to Amazonia, B. transversalis [23]. However, this hypothesis is not supported by any of the results obtained here.

The presumed range of B. bahiensis sp. n. also falls within a forest refuge that existed in the Atlantic Forest domain during cold, dry periods of the Pleistocene, the so-called Bahia refuge, which was predicted by climatic models and validated by paleopalynological data (for example [61, 63,64]). In the Bahia refuge, limited in the south by the Doce river, in Espírito Santo, and extending northward into southern Bahia, evergreen forests persisted even in the driest periods of the Pleistocene. Carnaval et al. [64] predicted that surveys in this area would still reveal undescribed species and cryptic lineages. If Hines [65] is right in her estimate that the closest common ancestor of B. brasiliensis, B. transversalis and B. pauloensis existed at about 2 Mya, then the most recent common ancestor of B. brasiliensis and B. bahiensis sp. n. lived in the Pleistocene, and this would be consistent with the hypothesis that the isolation of part of its populations in the Bahia refuge during that period could be the vicariance event responsible for its genetic divergence. The close association of B. brasiliensis and B. bahiensis sp. n. with evergreen forests in southeastern Brazil suggests that interbreeding of their populations may have been prevented by their inability to settle in lowland semi-deciduous forests, which occur under climates with relatively long dry seasons and that reaches the coast between southern Espírito Santo and northern Rio de Janeiro states. It should be noted, however, that this hypothesis is weakened by the occurrence of B. brasiliensis in areas dominated by climates with relatively long dry seasons and covered with semideciduous forests along the Paraná River valley, westward into Paraguay and central Brazil (Fig 3). A closer look at the habitats used by this species in that region is necessary for a better understanding of this question.

Morphological distinction and geographic range

Bombus is a relatively monotonous genus, as far as morphological characters are concerned (for example [1]), and this makes them frequently difficult to distinguish, especially among closely-related species (for example [511, 49]. Thus, it is not surprising that not a single exclusive autoapomorphy of B. bahiensis sp. n. could be found, and that the only morphological character distinguishing it from its presumed sister species, the black-haired T1, is a plesiomorphic trait or, at most, a homoplastic apomorphy, also found in all other Brazilian species but B. brasiliensis (for example [23]).

Moure & Sakagami [23] noticed that specimens identified by them as B. brasiliensis from Linhares, by the northern margin of the Doce river, in Espírito Santo state (shown by a grey triangle with a central black dot in Fig 3) had their first metasomal tergum completely covered by black pilosity. They interpreted this population as a melanic variety of B. brasiliensis. These specimens quite probably belong to B. bahiensis sp. n. and may represent the southernmost population of the species, since the Doce river is the southern limit for the evergreen forests of the Bahia refuge and the Bahia forest (although some authors, as Silva & Casteleti [51] and Ribeiro et al. [66] consider the Bahia forest to extend further south in Espírito Santo).

Conservation Status

Local or regional decline or extirpation of bumblebee populations has been reported in the northern hemisphere (for example [49, 6770]) and in Brazil [71]. It seems that the small available sample of B. bahiensis sp. n. reflects a low abundance in nature and not a small sampling effort in its habitat. The bee fauna of northern Espírito Santo was intensively sampled by C. Elias, a collector for the DZOL collection in Curitiba in the 1960’s and early 1970’s, and the only specimens of the species collected among thousands of bee specimens, apparently are those few listed in the “Taxonomic treatment”. Moreover, two expeditions were set to southern Bahia and northern Espírito Santo to search for the species, during the development of this project (January and June/July, 2014), but resulted unsuccessful. With this in mind, two facts should be considered concerning the conservation of this new species: 1) its presumed geographic range is the smallest among all Brazilian Bombus species; and 2) its natural habitat is under heavy anthropic impact—the rain forest that originally covered about 86% of the region, now covers only between 12% and 17% of it [51, 66]—and continues to be fragmented. Moreover, bees in general and bumblebees specifically may be more susceptible to fragmentation than other organisms, due to issues related to effective population size, social behavior, nesting, sex determination mechanisms etc. (see discussions in [72, 68]). Thus, B. bahiensis sp. n. may actually be an endangered species and efforts should be made to map the remaining populations of the species and their abundances, better defining its current geographic range and conservation status.

Supporting Information

S1 Fig. Phylogenetic relationships among Bombus brasiliensis obtained from concatenated phylogenetic analyses performed with CytB and COI sequences, using Bayesian inference.

Values shown under each branch are posterior probabilities.


S1 Table. Specimens sequenced for the genetic analyses with their geographic origins.

Universidade Federal de Minas Gerais—UFMG IHY; Universidade Federal de Santa Catarina—UFSC; Universidade Federal do Norte Fluminense—UENF; Universidade Federal de Ouro Preto—UFOP; Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto—Universidade de São Paulo—USP: FFCLRP; Universidade de São Paulo—USP/SP.


S2 Table. Protocol of the purification of Polyethylene Glycol 20% (PEG 20%) for elimination of bands <300–400 bp.


S3 Table. Genetic distances for COI sequences (%) between bumblebee-species.

The model used was the Kimura 2-parameter. Bees currently considered as B. brevivillus belong in two distinct species, identified below as B. brevivillus (1) and B. brevivillus (2). AD = Average interspecific distance ± standard deviation; MiD = minimum interespecific distance ± standard deviation.



We thank Dr. J. E. Hogan, curator of the Hope Entomological Collection (Oxford University Museum of Natural History), who kindly provided the photos of the lectotype of Bombus brasiliensis (reproduced here in Fig 1A and 1B); Dr. A. J. Santos (Universidade Federal de Minas Gerais), for allowing us to photograph the holotype of Bombus bahiensis with the equipment under his care (provided by the project INCT/HYMPAR/SUDESTE). Thanks are also due to B. Blochtein, G. A. R. Melo, L. A. O. Campos, L. Packer, E. Francoso, M. C. Arias, M. C. Gaglianone, S. C. Krug, S. Matheus, J. Steiner and Y. Antonini for loan of specimens; L. Packer (PCYU) and P. Tubaro (Museo Argentino de Ciencias Naturales), for providing DNA sequences of specimens of Bombus; and L. Packer, A. J. Santos, P. Williams and S. Brady for suggestions on the manuscript. We also acknowledge the Instituto Chico Mendes for the collecting permit granted to JESJ.

Author Contributions

Conceived and designed the experiments: JESJ FAS. Performed the experiments: JESJ. Analyzed the data: JESJ FRS FAS. Contributed reagents/materials/analysis tools: FRS FAS. Wrote the paper: JESJ FAS.


  1. 1. Michener CD (2007) The Bees of the World. Second Edition. Johns Hopkins University, Baltimore, 1016 pp.2.
  2. 2. Gibbs J (2009) Integrative taxonomy identifies new (and old) species in the Lasioglossum (Dialictus) tegulare (Robertson) species group (Hymenoptera, Halictidae). Zootaxa 2032: 1–38.
  3. 3. Sheffield CS, Hebert PDN, Kevan PG, Packer L (2009) DNA barcoding a regional bee (Hymenoptera: Apoidea) fauna and its potential for ecological studies. Mol Ecol Resources 9: 196–207.
  4. 4. Willams PH (1998) An annotated checklist of bumble bees with an analysis of patterns of description (Hymenoptera: Apidae, Bombini). Bulletin of The Natural History Museum (Entomology) 67: 79–152.
  5. 5. Bertsch A, Schweer H, Titze A, Tanaka H (2005) Male labial gland secretions and mitochondrial DNA markers support species status of Bombus cryptarum and B. magnus (Hymenoptera, Apidae). Insectes Sociaux 52: 45–54.
  6. 6. Ellis J S, Knight ME, Goulson D (2005) Delineating species for conservation using mitochondrial sequence data: the taxonomic status of two problematic Bombus species (Hymenoptera: Apidae). J Insect Cons 9: 75–83.
  7. 7. Ellis JS, Carvell C, Goulson D (2006) Cryptic species identification: a simple diagnostic tool for discriminating between two problematic bumblebee species. Mol Ecol 6: 540–542.
  8. 8. Wolf S, Rohde M, Moritz FA (2010) The reliability of morphological traits in the differentiation of Bombus terrestris and B. lucorum (Hymenoptera: Apidae). Apidologie 41 45–53.
  9. 9. Lecocq T, Lhomme P, Michez D, Dellicour S, Valterová I, Rasmont P (2011) Molecular and chemical characters to evaluate species status of two cuckoo bumblebees: Bombus barbutellus and Bombus maxillosus (Hymenoptera, Apidae, Bombini). Systematic Entomol 36: 453–469.
  10. 10. Carolan JC, Murray TE, Fitzpatrick U, Crossley J, Schmidt H, Cederberg B et al. (2012) Colour patterns do not diagnose species: Quantitative evaluation of a DNA barcoded cryptic bumblebee complex. PLoS ONE 7(1): e29251. pmid:22238595
  11. 11. Gjershaug J, Staverløkk A, Kleven O, Ødegaard F (2013) Species status of Bombus monticula Smith (Hymenoptera: Apidae) supported by DNA barcoding. Zootaxa 3716: 431–440.
  12. 12. Williams PH, Brown MJF, Carolan JC, An J, Goulson D, Aytekin AM, Best LR et al. (2012) Unveiling cryptic species of the bumblebee subgenus Bombus s. str. worldwide with COI barcodes (Hymenoptera: Apidae). Systematics and Biodiversity 10(1): 21–56.
  13. 13. Estoup A, Solignac M, Cornuet JM, Goudet JS (1996) Genetic differentiation of continental and island populations of Bombus terrestris (Hymenoptera: Apidae) in Europe. Mol Ecol 5: 19–31. pmid:9147693
  14. 14. Pirounakis K, Koulianos S, Schmid-Hempel P (1998) Genetic variation among European populations of Bombus pascuorum (Hymenoptera, Apidae) using mitochondrial DNA sequence data. European J Entomol 95: 27–33.
  15. 15. Widmer A, Schmid-Hempel P, Estoup A, Scholl A (1998) Population genetic structure and colonization history of Bombus terrestris s.l. (Hymenoptera: Apidae) from the Canary Islands and Madeira. Heredity 81: 563–572.
  16. 16. Widmer A, Schmid-Hempel P (1999) The population genetic structure of a large temperate pollinator species, Bombus pascuorum (Scopoli) (Hymenoptera: Apidae). Mol Ecol 8: 387–398. pmid:10199006
  17. 17. Koulianos S (1999) Phylogenetic relationships of the bumblebee subgenus Pyrobombus (Hymenoptera: Apidae) inferred from mitochondrial cytochrome B and cytochrome oxidase I sequences. Entomol Soc Amer 92: 355–358.
  18. 18. Koulianos S, Schmid-Hempel P (2000) Phylogenetic relationships among bumble bees (Bombus, Latreille) inferred from mitochondrial cytochrome b and cytochrome oxidase I sequences. Mol Phylogenetics Evol 14: 335–341. pmid:10712839
  19. 19. Cameron SA, Williams PH (2003) Phylogeny of bumble bees in the New World subgenus Fervidobombus (Hymenoptera: Apidae): congruence of molecular and morphological data. Mol Phylogenetics Evol 28: 552–563. pmid:12927138
  20. 20. Shao ZY, Mao HX, Fu WJ, Ono M, Wang DS, Bonizzoni M (2004) Genetic structure of Asian populations of Bombus ignitus (Hymenoptera: Apidae). J Heredity 95: 46–52.
  21. 21. Cameron SA, Hines HM, Williams PH (2007) A comprehensive phylogeny of the bumble bees (Bombus). Biol J Linn Soc 91: 161–188.
  22. 22. Kim MJ, Yoon HJ, Im HH, Jeong UH (2009) Mitochondrial DNA sequence variation of the bumblebee, Bombus ardens (Hymenoptera: Apidae). J Asia-Pacific Entomol 12: 133–139.
  23. 23. Moure JS, Sakagami SF (1962) As mamangabas sociais do Brasil (Bombus Latreille) (Hymenoptera, Apoidea). Stud. Entomol 5: 65–194.
  24. 24. Moure JS, Melo GAR (2012) Bombini Latreille, 1802. In Moure JS, Urban D & Melo GAR, editors. Catalogue of Bees (Hymenoptera, Apoidea) in the Neotropical Region—online version. Available at Accessed Jun/24/2014
  25. 25. Williams PH, Cameron SA, Hines HM, Cederberg B, Rasmont P (2008) A simplified subgeneric classification of the bumblebees (genus Bombus). Apidologie 39: 46–74.
  26. 26. Milliron HE (1973) A monograph of the Western Hemisphere bumblebees (Hymenoptera: Apidae; Bombinae. II. The genus Megabombus subgenus Megabombus. Mem Entomol Soc Canada 89: 81–237.
  27. 27. Abrahamovich AH, Díaz B (2002) Bumble bees of the Neotropical region (Hymenoptera; Apidae). Biota Colombiana 3: 199–214.
  28. 28. Silveira FA, Cure JR (1993) High-altitude bee fauna of Southeastern Brazil: Implications for biogeographic patterns (Hymenoptera: Apoidea). Studies Neot Fauna Environ 28: 47–55.
  29. 29. Wiley EO, Lieberman BS (2011) Phylogenetics—Theory and Practice of Phylogenetic Systematics. 2 ed. Hoboken, Wiley-Blackwell. Xvi + 406 p.
  30. 30. Sambrook J, Russel DW (2001) Molecular Cloning: A Laboratory Manual. CSH Laboratory Press, Cold Spring Harbor, NY.
  31. 31. Hebert PDN, Penton EH, Burns JM, Janzen DH, Hallwachs W (2004) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proc Natl Acad Sci USA 101:14812–14817. pmid:15465915
  32. 32. Simon C, Frati F, Beckenbach A, Crespi B, Liu H, Flook P (1994) Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Entomol Soc Amer 87: 651–701.
  33. 33. Ewing B, Hillier L, Wendi M, Green P (1998) Basecalling of automated sequencer traces using Phred I: Accuracy assessment. Genome Research 8: 175–185. pmid:9521921
  34. 34. Gordon D, Abajian C, Green P (1998) Consed: a graphical tool for sequence finishing. Genome Research 8: 195–202. pmid:9521923
  35. 35. Edgar R (2004) MUSCLE: multiple sequence alignment with high accuracy and high output. Nucleic Acids Res 32: 1792–1797. pmid:15034147
  36. 36. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Mol Biol Evol
  37. 37. Excoffier L, Lischer HEL (2010). Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10: 564–567. pmid:21565059
  38. 38. Librado P, Rozas J (2009). DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25: 1451–1452. pmid:19346325
  39. 39. Bandelt HJ, Forster P, Röhl A (1999). Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 48: 16–37.
  40. 40. Nei M, Kumar S (2005) Molecular Evolution and Phylogenetics. New York: Oxford University Press.
  41. 41. Posada D, Crandall KA (1998) Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817–818. pmid:9918953
  42. 42. Huelsenbeck J, Ronquist F (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: 754–755. pmid:11524383
  43. 43. Swofford D (1998) PAUP: Phylogenetic Analysis Using Parsimony (and Other Methods). Version 4. Sunderland, Massachusetts, Sinauer Associates Inc.
  44. 44. Dick CW, Roubik DW, Gruber KF, Bermingham E (2004) Long-distance gene flow and cross-Andean dispersal of lowland rainforest bees (Apidae: Euglossini) revealed by comparative mitochondrial DNA phylogeography. Mol Ecol 13: 3775−3785. pmid:15548290
  45. 45. Nemésio A, Santos JE Júnior, Santos FR (2013) Eufriesea zhangi sp. n. (Hymenoptera: Apidae: Euglossina), a new orchid bee from Brazil revealed by molecular and morphological characters. Zootaxa 3609: 568–582. pmid:24699618
  46. 46. Meyer CP, Paulay G (2005) DNA Barcoding: Error rates based on comprehensive sampling. PLoS Biology 3:2229–2238.
  47. 47. Goldstein PZ, DeSalle R (2011) Integrating DNA barcode data and taxonomic practice: Determination, discovery, and description. Bioessays 33: 135–147. pmid:21184470
  48. 48. Edwards SV, Beerli P (2000) Perspective: gene divergence, population divergence, and the variance in coalescence time in phylogeographic studies. Evolution 54: 1839–1854. pmid:11209764
  49. 49. Williams PH, Byvaltsev A, Sheffield C, Rasmont P (2013) Bombus cullumanus—an extinct European bumblebee species? Apidologie 44: 121–132. pmid:24330735
  50. 50. Wiens JJ, Penkrot TA (2002) Delimiting species using DNA and morphological variation and discordant species limits in spiny lizards (Sceloporus). Syst Biol 5: 69–91.
  51. 51. Silva JMC, Casteleti CHM (2003) Status of the biodiversity of the Atlantic Forest of Brazil. In: Galindo-Leal C & Câmara IG editors. The Atlantic Forest of South America: Biodiversity Status, Threats, and Outlook. Washington: Island Press. pp 43–59.
  52. 52. Moure JS (1995) Notas sobre algumas espécies da Bahia, Brasil (Hymenoptera, Apoidea). Revta bras Zool 12: 467–470.
  53. 53. Nemésio A (2011) Exaerete salsai sp. n. (Hymenoptera: Apidae): a new orchid bee from eastern Brazil. Zootaxa 2967: 12–20.
  54. 54. Faria LRR, Melo GAR (2012) Species of Euglossa of the analis group in the Atlantic forest (Hymenoptera, Apidae). Zoologia 29: 349–374.
  55. 55. Nemésio A, Engel MS (2012) Three new cryptic species of Euglossa from Brazil (Hymenoptera, Apidae). ZooKeys 222: 47–68. pmid:23129986
  56. 56. Nemésio A, Cerântola NCM, Vasconcelos HL, Nabout JC, Silveira FA, Del Lama MA (2012) Searching for Euglossa cyanochlora Moure, 1996 (Hymenoptera: Apidae), one of the rarest bees in the world. J Insect Cons 16: 745–755.
  57. 57. Costa LP (2003) The historical bridge between the Amazon and the Atlantic Forest of Brazil: A study of molecular phylogeography with small mammals. J Biogeography 30: 71−86.
  58. 58. Tchaicka L, Eizirik E, Oliveira TG, Cândido JF Jr, Freitas TRO (2007) Phylogeography and Population History of the Crab-eating Fox (Cerdocyonthous). Mol Ecol 16: 819–838. pmid:17284214
  59. 59. Resende HC, Yotoko KSC, Delabie JHC, Costa MA, Campiolo S, Tavares MG, et al. (2010) Pliocene and Pleistocene events shaping the genetic diversity within the central corridor of the Brazilian Atlantic Forest. Biol J Linn Soc 101: 949–960.
  60. 60. Azevedo AA, Silveira FA, Aguiar CM, Pereira VS (2008) Diversidade de abelhas (Hymenoptera, Apoidea) nos campos rupestres da Cadeia do Espinhaço. Megadiversidade 4: 67–94.
  61. 61. Carnaval AC, Hickerson MJ, Haddad CF, Rodrigues MT, Moritz C (2009) Stability predicts genetic diversity in the Brazilian Atlantic Forest Hotspot. Science 323: 785–789. pmid:19197066
  62. 62. Batalha-Filho H, Fjeldså J, Fabre P-H, Miyaki CY (2012) Connections between the Atlantic and the Amazonian Forest avifaunas represent distinct historical events. J Ornithol 154: 41–50.
  63. 63. Carnaval AC, Moritz C (2008) Historical climate modelling predicts patterns of current biodiversity in the Brazilian Atlantic forest. J Biogeography 35: 1187–1201.
  64. 64. Carnaval AC, Waltari E, Rodrigues MT, Rosauer D, VanDerWal J, Damasceno R, et al. (2014) Prediction of phylogeographic endemism in an environmentally complex biome. Proceedings of The Royal Society 281: 1–9.
  65. 65. Hines HM (2008) Historical biogeography, divergence times, and diversification patterns of bumble bees (Hymenoptera: Apidae: Bombus). Syst Biol 57: 58–75. pmid:18275002
  66. 66. Ribeiro M C, Metzger JP, Martensen AC, Ponzoni FJ, Hirota MM (2009) The Brazilian Atlantic Forest: How much is left, and how is the remaining forest distributed? Implications for conservation. Biol Cons 142: 1141–1153.
  67. 67. Colla SR, Packer L (2008) Evidence for decline in eastern North American bumblebees (Hymenoptera: Apidae), with special focus on Bombus affinis Cresson. Biodivers Conserv 17:1379–1391.
  68. 68. Goulson D, Lye GC, Darvill B (2008) Decline and conservation of bumble Bees. Ann Rev Entomol 53: 191–208. pmid:17803456
  69. 69. Grixti JC, Wong LT, Cameron SA, Favret C (2009) Decline of bumble bees (Bombus) in the North American Midwest. Biol Cons 142:75–84.
  70. 70. Williams PH, Osborne JL (2009) Bumblebee vulnerability and conservation world-wide. Apidologie 40: 367–387.
  71. 71. Martins AC, Melo GAR (2009) Has the bumblebee Bombus bellicosus gone extinct in the northern portion of its distribution range in Brazil? J Insect Conserv
  72. 72. Packer L, Owen R (2000) Population genetic aspects of pollinator decline. Conservation Ecology 5: 4. [online] URL:
  73. 73. Abrahamovich AH, Díaz NB, Morrone JJ (2004) Distributional patterns of the Neotropical and Andean species of the genus Bombus (Hymenoptera: Apidae). Acta Zoológica Mexicana (new series) 20: 99–117.