When compared to other regions of the world, southern South America seems to depict low bumble bee diversity (^{Williams, 1998}). Only ten out of the ca. 250 species of Bombus described worldwide have been reported to inhabit Argentina. Two of them, Bombus ruderatus (Fabricius, 1775) and B. terrestris (L., 1758), are invasive species of relatively recent entry into the southwest of the country from Chile, while the remaining eight are native ( ^{Abrahamovich et al., 2007} and ^{Schmid-Hempel et al., 2014}).

According to the last available surveys on their geographic distribution in Argentina (^{Moure and Sakagami, 1962}, ^{Abrahamovich and Díaz, 2001}, Abrahamovich et al., 2004 and Abrahamovich et al., 2007), B. pauloensis Friese, 1913 (=B. atratus Franklin, 1913) , B. bellicosus Smith, 1879, B. morio (Swederus, 1787), and B. opifex Smith, 1879 are known to exhibit wide ranges. Bombus tucumanus Vachal, 1904, B. baeri Vachal, 1904, and B. dahlbomii Guérin, 1835 appear to show more limited ranges, while B. brasiliensis Lepeletier, 1836 may possibly occur only in Misiones province at the northeastern tip of the country. Although B. brasiliensis, an assiduous visitor of bromeliad flowers (Bromeliaceae) like Aechmea spp. ( ^{Kaehler et al., 2005} and ^{Schmid et al., 2011}) appears to be widespread in Brazil ( Abrahamovich et al., 2004 and ^{Santos Júnior et al., 2015}), surveys carried out by our group since January 2008 suggest that its distribution in Argentina may be nowadays considerably reduced. This communication reports the detection of the microsporidium Nosema ceranae infecting B. brasiliensis and argues on a possible effect on the distribution of this bee species on the southeastern part of its range.

After campaigns in search of B. brasiliensis since 2008 done by authors and other team members surveying 32 localities in the provinces of Formosa (Bañado La Estrella, Colonia Perin, El Colorado, Gran Guardia, Ibarreta, Ingeniero Juárez, Laguna Yema, Las Lomitas, Palo Santo, Pirané, Posta Cambio Zalazar, Pozo del Tigre), Chaco (38 km North of Resistencia, Colonia Elisa, Juan José Castelli, Resistencia, Presidencia Roque Sáenz Peña), Corrientes (Colonia Carlos Pellegrini, Corrientes, Estero Santa Lucía, Laguna Iberá, Santo Tomé), and Misiones (Aristóbulo del Valle, Cuña Pirú, El Alcázar, 30 km East of María Magdalena, Montecarlo, Leandro N. Alem, Posadas, Urugua-í, Wanda) (Fig. 1A, Table 1) with no positive results, only six adult workers were collected in February 2015 in the surroundings of Puerto Iguazú, Misiones, northeastern Argentina (Fig. 1A). They were captured while foraging using entomological nets, conserved frozen (-32° C), and identified based on information provided by ^{Moure and Sakagami (1962}), ^{Abrahamovich et al. (2005}), and ^{Santos Júnior et al. (2015}).

Fig. 1 (A) The 32 surveyed localities in northeastern Argentina. Formosa province [Fo]: (1) Ingeniero Juárez; (2) Laguna Yema; (3) Las Lomitas; (4) Bañado La Estrella; (5) Posta Cambio Zalazar; (6) Pozo del Tigre; (7) Colonia Perin; (8) Ibarreta; (9) Palo Santo; (10) Pirané; (11) Gran Guardia; (12) El Colorado. Chaco province [Ch]: (13) J. J. Castelli; (14) Presidencia Roque Saenz Peña; (15) Colonia Elisa; (16) 38 km North of Resistencia; (17) Resistencia. Corrientes province [Co]: (18) Corrientes; (19) Estero Santa Lucía; (20) Colonia Carlos Pellegrini; (21) Laguna Iberá; (22) Santo Tomé. Misiones province [Mi]: (23) Posadas; (24) Leandro N. Alem; (25) Cuña Pirú; (26) Aristóbulo del Valle; (27) El alcázar; (28) Montecarlo; (29) 30 km East of María Magdalena; (30) Wanda; (31) Urugua-í. (★) indicates Puerto Iguazú, the only locality where six Bombus brasiliensis workers were found. (B) Habitat where B. brasiliensis was found. (C) Two distended midgut cells of B. brasiliensis with spores of Nosema ceranae inside. 

Examination of each individual was performed following dissection techniques under stereoscopic microscopy (×10, ×40) (^{Lacey and Solter, 2012}). Briefly, small portions of different tissues and organs were extracted in order to prepare fresh smears with one-quarter-strength Ringer's solution (^{Poinar and Thomas, 1984}) for detection of microsporidia and protists (^{Lange and Lord, 2012} and Solter et al., 2012). Observations were done using phase-contrast microscopy (×400, ×1000). Each infected individual was then homogenized in 2 mL of double distilled water and infection intensity (spore load) was quantified using an Improved Neubauer hemocytometer (^{Undeen and Vávra, 1997}). Spore suspensions were obtained by repeated filtration and centrifugation (15 min; 7500 × g) ( Lange and Henry, 1996). Double distilled water was replaced by absolute ethanol in spore suspensions and stored at -32 °C until genetic analysis were performed. Microsporidium rRNA (16S small subunit) was amplified by real time PCR according to ^{Medici et al. (2012}) with specific primers for Nosema apis Zander, 1907 (321APIS) and N. ceranae (218MITOC). Amplicons were separated on ethidium bromide-stained 1% agarose gel. Genetic material was purified with an ExoSap-IT kit (Amersham, Biosciences) and sequenced in an automatic MegaBACE Sequence Analyzer (Amersham, Biosciences). Sequences were aligned using SMS software ( ^{Stothard, 2000}) and submitted to Genbank.

Microsporidian infections were detected in two individuals of B. brasiliensis collected while foraging on Solanum sp. (Solanaceae) in a farmland 5 km southeast of Puerto Iguazú (25°38'52" S; 54°32'40" W) (Fig. 1B). Diseased individuals did not show obvious external signs of infection. The microsporidium was found infecting cells of the midgut epithelium, many of which were distended due to the heavy presence of spores (Fig. 1C). Infection intensity for each bumble bee was 5.5 × 10⁷ and 4.4 × 10⁷ spores/insect. Samples of spores from both infections were amplified by 218MITOC primers and produced amplicons indicating presence of N. ceranae. One of the sequences (220 bp) was deposited on GeneBank under accession number KX024757.

Nosema ceranae was originally described from the Asian honey bee Apis cerana Fabricius, 1793 ( ^{Fries et al., 1996}). Ten years after its description, it was also detected in the European honey bee A. mellifera L., 1758 in Spain ( ^{Higes et al., 2006}). ^{Teixeira et al. (2013}) evidenced that N. ceranae has been present in Brazil for at least 36 years infecting Africanized honey bees. ^{Plischuk et al. (2009}) reported for the first time the presence of this microsporidium in bumble bees infecting three native South American species of genus Bombus [B. pauloensis (named as B. atratus), B. bellicosus, B. morio].

Transmission of N. ceranae is mainly horizontal ( ^{Higes et al., 2008}). This microsporidium completes its lifecycle inside midgut epithelial cells. Spores leave the host with the feces and may remain viable in the environment until they enter a new individual per os. During the last decade, numerous studies worldwide have demonstrated high virulence of N. ceranae against A. mellifera ( ^{Paxton et al., 2007} and Higes et al., 2007), suggesting a role in the honey bees Colony Collapse Disorder (CCD) (Higes et al., 2008). Effects of N. ceranae in honey bees are relatively well known and include lesions in the midgut epithelium causing metabolic stress and suppressing immune response ( ^{Antúnez et al., 2009} and ^{Mayack and Naug, 2009}). Higes et al. (2007) have reported 100% bee mortality 8 days post inoculation under experimental conditions. On the contrary, pathological effects of N. ceranae in bumble bees are not completely clear. Experimental infections on B. terrestris have shown that this pathogen would be highly virulent, colonizing the midgut epithelium and reducing host survival by 48% within one week after exposure ( ^{Graystock et al., 2013}).

Because these are the first detections of N. ceranae in B. brasiliensis, effects on the host are yet unknown. However based on the observation that infections in both individuals were advanced and heavy, a significant virulence possibly similar to that described in other hosts would not be unexpected.

An important subject is the possible spread of N. ceranae to other species. Since the vectoring of pathogens by the sharing use of flowers seems to be a relatively common process, the transferring of infective spores between bees and other pollinators appears highly likely ( ^{Graystock et al., 2013} and Graystock et al., 2015), and B. brasiliensis [possibly along with A. mellifera ( ^{Teixeira et al., 2013})] could act as source of N. ceranae to other sympatric species. Recent studies have demonstrated that N. ceranae is capable to infect not only Bombus and Apis species but also solitary bees belonging to genus Andrena (Andrenidae), Osmia, and Heriades (Megachilidae) in Belgium ( ^{Ravoet et al., 2014}). Numerous endemic bees inhabit the Neotropical region, as several stingless bees (Meliponini) (^{Freitas et al., 2009}) and orchid bees (Euglossini) (^{Nemésio, 2009}) that could be potential hosts. If these bees are susceptible, the effects of N. ceranae might be enhanced as seen in other, several new host-parasite associations ( ^{Goulson and Hughes, 2015}).

Although data about the presence of B. brasiliensis is fragmented, the extension of its geographic distribution appears to be from the southern portions of the Araguaia-Tocantins basin in Mato Grosso, Brazil to northern Argentina and Uruguay in the South. The western limit would be along the Paraná River basin (Mato Grosso do Sul and southern Goiás), whereas to the East it would reach the Atlantic coast ( ^{Santos Júnior et al., 2015}).

However, recent studies have shown that the eastern range would actually be narrower than previously thought since that area is inhabited by B. bahiensis^{Santos Júnior et al., 2015}, a recently described sister species ( Santos Júnior et al., 2015). The southwestern range into Argentina also seems to have historically diminished from Chaco, Formosa and Misiones provinces ( ^{Holmberg, 1903}, ^{Moure and Sakagami, 1962}, ^{Abrahamovich and Díaz, 2001} and Abrahamovich et al., 2004) to only the latter (Abrahamovich et al., 2007). Absence of detections of B. brasiliensis after several campaigns since 2008 at more than 30 localities ( Fig. 1A, Table 1) within the original range may suggests a more restricted presence nowadays, apparently limited to a small area at the northeastern tip of the country (Fig. 1A). Bombus brasiliensis was also present in Uruguay during the last century ( Moure and Sakagami, 1962) but it has not been detected in that country since at least 2005, suggesting that the southernmost distribution could also have suffered a retraction (Santos et al., 2013). In Uruguay, N. ceranae is highly prevalent in B. pauloensis and B. bellicosus ( ^{Arbulo et al., 2015}).

Although evidencing a cause-effect link between occurrence of N. ceranae and the apparent range retraction of B. brasiliensis constitute an elusive goal, we feel it is important to report on the infections we have found. Nosema ceranae may be at least one of the drivers modulating the populations of this host.