Two continents and two names for a Neotropical colletid bee species (Hymenoptera: Colletidae: Neopasiphaeinae): Hoplocolletes ventralis (Friese, 1924)

Neopasiphaeine bees (Apoidea: Colletidae) are known for their Amphinotic distribution in the Australian and Neotropical regions. Affinities between colletid taxa in Australia and South America have been speculated for decades, and have been confirmed by recent phylogenetic hypotheses that indicate a biogeographic scenario compatible with a trans-Antarctic biotic connection during the Paleogene. No neopasiphaeine species occurs on both sides of the Pacific Ocean, but the Neotropical species Hoplocolletes ventralis (Friese, 1924) was described as an Australian taxon due to an error in the specimen labels. This mistake was recognized by CD Michener 50 years ago. We herein report that the same labeling problem also happened with Dasycolletes chalceus Friese, 1924, which remained as a tentatively placed species in the Australian genus Leioproctus until now. Moreover, Dasycolletes chalceus is interpreted as a synonym of Hoplocolletes ventralis. We also provide a revised diagnosis for Hoplocolletes, describe the male of H. ventralis in detail for the first time, including a comparative study of its genitalia and associated sterna.


INTRODUCTION
Affinities between taxa of Colletidae distributed in Australia and South America have been speculated for decades (Michener, 1965;Michener, 1989), and have been confirmed by recent phylogenetic hypotheses that indicate a biogeographic scenario compatible with a trans-Antarctic biotic connection during the Paleogene (Almeida et al., 2012). Dasycolletes ventralis Friese, 1924 was described as an Australian colletid species based on a single female specimen labeled as having been collected in Sydney (Australia). The species actually is endemic to Brazil, and there is no species occurring in Australia that could be confused with it. The confusion certainly results from an error in the label, as concluded by Michener (1965: p. 41), an interpretation followed by subsequent authors (e.g., Moure, Graf & Urban, 2007;Rasmussen & Ascher, 2008). After the species description, it was moved to the genus Paracolletes by Cockerell (1929), and later placed in Leioproctus (Hoplocolletes), created by Michener (1965) to accommodate it based on clear affinities to other taxa classified as Leioproctus, but also recognizing its uniqueness (see also Michener, 1989;Michener, 2007). Hoplocolletes remains a monotypic taxon in Neopasiphaeinae (Colletidae), having been classified as genus (e.g., Silveira, Melo & Almeida, 2002;Moure, Graf & Urban, 2007;Almeida & Danforth, 2009;Almeida et al., 2012) or as subgenus of Leioproctus (e.g., Michener, 1965;Michener, 1989;Michener, 2007), the former being followed in this paper.
Hoplocolletes ventralis has been recorded in three states in southeastern Brazil: Espírito Santo, Minas Gerais, Rio de Janeiro (Silveira, Melo & Almeida, 2002;Moure, Graf & Urban, 2007). Nevertheless, it remains a poorly known genus, with relatively little distributional information, the male undescribed, host-plant preferences unknown, and the only piece of bionomical information for this species is that it is a soil nesting bee (EAB Almeida, pers. obs., 2001). The phylogenetic affinities of Hoplocolletes and other neopasiphaeine taxa were uncertain until molecular phylogenetic hypotheses placed this taxon in a clade comprising Eulonchopria and Nomiocolletes (Almeida & Danforth, 2009;Almeida et al., 2012). Michener (1989: p. 630) suggested that Hoplocolletes could be part of a "Basal Group," characterized by the fully developed sternal scopa. Based on the phylogenetic hypotheses currently available, it seems that this scopa arose multiple times in the Neopasiphaeine clade, since taxa with this character, Hoplocolletes, Cephalocolletes, Reedapis, and Tetraglossula are otherwise not close relatives (Almeida & Danforth, 2009;Almeida et al., 2012).
The aim of this work is three fold. To resolve a taxonomic problem related to a new synonymy involving Hoplocolletes ventralis and Dasycolletes chalceus, which are here interpreted as synonyms. To report that the above mentioned labeling problem that made the taxonomic history of Hoplocolletes ventralis problematic also happened with Dasycolletes chalceus Friese, 1924, which remained as a tentatively placed species in the Australian genus Leioproctus until now (Michener, 1965;Cardale, 1993;Almeida, 2008;Rasmussen & Ascher, 2008). To increase the knowledge about the morphology and distinctiveness of Hoplocolletes, particularly by providing a novel description of the male genital complex for this species.

MATERIAL & METHODS
Part of the material studied is deposited in the Entomological Collection "Prof. JMF Camargo" [RPSP] in Departamento de Biologia (FFLRP/USP, Ribeirão Preto, Brazil). A male specimen of Hoplocolletes ventralis was obtained on loan from Entomological Collection "Pe. JS Moure" [DZUP], Departamento de Zoologia (UFPR, Curitiba, Brazil), and the female type specimen of Dasycolleletes chalceus Friese, 1924 was studied and photographed at the entomological collection of Museum für Naturkunde [ZMB] (Berlin, Germany). Photographs of the female specimen of Dasycolletes ventralis Friese, 1924, deposited at the American Museum of Natural History (AMNH) collection, were kindly made available for this study.
The general morphological terminology follows Michener (2007). Antennal flagellomeres are indicated as F1, F2, etc.; metasomal terga and sterna, respectively, as T1 to T7, and S1 to S8. The density of punctation and intervals between the punctures are based on relative puncture diameter, pd (e.g., <1 pd: less than 1× the puncture diameter between the punctures). Color images were obtained on a Zeiss Axiocam 206 color camera associated to a Zeiss Discovery. V12 stereomicroscope, or with an AmScope MU1000A Digital Camera adapted onto a Leica MZ6 stereomicroscope; pictures were assembled with the software Helicon Focus 6.2.

RESULTS
The species Dasycolletes chalceus was not studied after its original description. It was described in the same publication and same page as Dasycolletes ventralis (Friese, 1924: p. 218). After 1924, it was only mentioned in catalogues and revisionary works (e.g., Michener, 1965;Cardale, 1993;Almeida, 2008;Rasmussen & Ascher, 2008), but the type specimen was never studied again. The only exemplar of Dasycolletes chalceus located and bearing Friese's original labels is deposited in ZMB (Fig. 1). It clearly has all diagnostic characters for Hoplocolletes as currently circumscribed, and no differences were found in relation to Hoplocolletes ventralis either. Hence, they are herein synonymized. The only known specimen of Dasycolletes ventralis bearing Friese's original labels is in the American Museum of Natural History collection (New York, USA) (Fig. 2) and it is the same female studied by Michener (1965) that lead him to conclude that it was not an Australian taxon, as indicated by the collecting labels, but a specimen probably collected in Brazil. The interpretation of Friese's types is a controversial subject and it is likely that the AMNH specimen is a duplicate, not the primary type (Rasmussen & Ascher, 2008;JS Ascher, pers. comm., 2015). But, so far, it is the only specimen labeled by Friese himself as Dasycolletes ventralis available for study. It is worth noting that both specimens were probably collected together, have locality labels that are identical, "Australia \\ Sydney \\ 14.9/06." The collector's name is lacking from the D. chalceus specimen label but is in the species' description (Friese, 1924: p. 218  frons (<1 pd), on vertex variable (denser medially, sparser (≤1 pd) laterally as well as on gena) integument smooth and shiny between punctures; coarse and dense on mesosoma, sparser toward center of disc of mesoscutum, and inferior on mesepisternum; metapostnotum smooth and shiny, delimited from pronotum by a pit-row; T1 smooth and shiny, with very sparse (2-7 pd) moderately coarse punctation, transversal line of barely aligned punctures delimiting marginal region of T1; on T2 slightly denser than on T1, but punctation leaving broad shiny areas as well; T3 and T4 with basal portion finely and densely punctated, sparser and coarser distad.