Print
Hidden diversity in European bees: Andrena amieti sp. n., a new Alpine bee species related to Andrena bicolor (Fabricius, 1775) (Hymenoptera, Apoidea, Andrenidae)
expand article infoChristophe Praz§, Andreas Müller|, David Genoud
‡ University of Neuchâtel, Neuchâtel, Switzerland
§ InfoFauna – Swiss Zoological Records Center, Neuchâtel, Switzerland
| ETH Zurich, Institute of Agricultural Sciences, Zurich, Switzerland
¶ Unaffiliated, Arzens, France
Open Access

Abstract

We revise the Alpine bee taxa related to Andrena bicolor (Fabricius, 1775), including A. montana Warncke, 1973 and A. allosa Warncke, 1975, the status of which has remained contentious. Phylogenetic analyses of one mitochondrial gene and one nuclear gene, as well as morphological examination reveal the presence of four Alpine species in this complex, one of which is new to science, A. amieti sp. n. This new species is widely distributed in the Alps from southern France throughout Switzerland, northern Italy and southern Germany to Austria; a single record is known from the Apennines. The type locality is located within the Unesco World Heritage site “Swiss Alps Jungfrau-Aletsch”. Two widely divergent mitochondrial lineages are found in sympatry in A. amieti sp. n.; the status of these lineages, which together form a paraphyletic unit from which A. allosa arose, is briefly discussed. We show that A. allosa, A. amieti sp. n. and A. montana are polylectic but that each species exhibits a distinct spectrum of pollen hosts: the univoltine A. allosa shows affinities for pollen of the early-blooming Alpine plant genus Crocus. A. amieti sp. n. is bivoltine and, as in A. bicolor, the summer generation exhibits a distinct preference for Campanulaceae, while the spring generation is widely polylectic. A. montana has a single generation in the summer and forages on a diversity of flowers such as Campanulaceae, Cistaceae and Caryophyllaceae. An identification key is presented for central European members of the subgenus Euandrena Hedicke, 1932. Lastly, the new Alpine species appears to represent the tip of the iceberg of substantial cryptic diversity in southern European Andrena (Euandrena): A. croatica Friese, 1887 is resurrected from synonymy with A. bicolor and treated as a valid species (stat. rev.), A. pileata Warncke, 1875, described as a subspecies of A. allosa, is elevated to species rank (stat. n.), and three additional unclear taxa are briefly described.

Key Words

Species delimitation, species paraphyly, host-plant preferences, polyphenism, DNA-barcoding

Introduction

The conservation of species depends on accurate data regarding their distributions and a sound understanding of their ecological requirements. While the distributions and ecology are well-known for a large part of the bee fauna in northern and central Europe (Peeters et al. 2012, Scheuchl and Willner 2016, Else and Edwards 2018, Westrich 2018), some Alpine and numerous southern European bee taxa remain poorly investigated. In part, this situation is due to open taxonomic questions and ambiguous morphology-based identifications, both of which strongly limit studies on species’ distribution and biology.

Within central European bees, one complex of species that remains insufficiently investigated is the group of taxa related to Andrena bicolor (Fabricius, 1775), referred to here as the “bicolor-group”. Andrena bicolor is one of the most widely distributed species of wild bees in Europe (Gusenleitner and Schwarz 2002, Scheuchl and Willner 2016). In Switzerland it is found in all biogeographic regions in numerous habitats from low elevations to the tree line (Amiet et al. 2010). The taxonomic status of two alpine taxa related to A. bicolor remains debated, A. allosa Warncke, 1975 and A. montana Warncke, 1973. These two taxa were recognized as specifically distinct from A. bicolor by Schmid-Egger and Scheuchl (1997), although both Westrich (2006) and Schmid-Egger (2012) indicated that the taxonomy of the bicolor-group was not fully elucidated. In agreement with this finding, Amiet et al. (2010) report that the morphological delimitation of A. allosa and A. montana in the Swiss Alps was difficult because of intermediate forms that bridge the morphological gap among taxa. In particular, Amiet et al. (2010) point to an unclear, Alpine “form” of A. bicolor in which the vestiture of both female and male is extensively white instead of yellow-brown and thus transitional between A. bicolor and A. montana.

In a recent, comprehensive DNA barcoding study of central European bees, Schmidt et al. (2015) presented DNA barcodes for three taxa of the bicolor-group and indicated that these taxa were clearly divergent from one another, supporting the hypothesis that there are three well separated species. While this result appears to solve the controversy on the status of these species, sequencing more specimens, especially those intermediate forms mentioned by Amiet et al. (2010), is still critical to our understanding of how to delimitate these species using morphological criteria. Such morphological delimitation is a necessary step for the revision of museum specimens and for obtaining details on species’ life-histories and distributions.

The biology of A. allosa and A. montana remains virtually unknown. The nominal subspecies of A. allosa is so far known from less than 20 individuals from France and Switzerland. Two subspecies have been described from Greece and Turkey, respectively, A. allosa pileata Warncke, 1975 and A. allosa canigica Warncke, 1975. Gusenleitner and Schwarz (2002: 100) raised doubts that these two infraspecific taxa were conspecific with A. allosa and stated that “more taxa are certainly present in the close relatedness of A. allosa“. Andrena montana is slightly better known and Amiet et al. (2010) indicated that this species flies from late June to mid-August, suggesting a single generation per year. Observations on the floral associations of these two taxa, however, are completely lacking. In contrast, the biology of A. bicolor is well-known. This species is bivoltine (e.g., Westrich 2006, 2018), at least at low elevations (Amiet et al. 2010), and broadly polylectic (Stöckhert in Schmiedeknecht 1930, Peeters et al. 2012, Scheuchl and Willner 2016). The cuckoo bee Nomada fabriciana (Linnaeus, 1767), similarly bivoltine, is a cleptoparasite of A. bicolor (e.g., Westrich 2006).

In the present study, we investigate the systematics of the Alpine taxa of the bicolor-group using DNA sequences of a mitochondrial and a nuclear gene. The nuclear gene was used in addition to the mitochondrial gene given the limitations of relying on a single DNA marker for species delimitation, for example because of incomplete lineage sorting, mitochondrial introgression or deep mitochondrial genetic divergence not accompanied by nuclear differentiation (e.g. Nicholls et al. 2012, Andriollo et al. 2015, Klopfstein et al. 2016). Based on combined genetic and morphological results, we demonstrate the presence of four well-separated species in the Alps, one of which is new to science. We also examine the status of some southern European taxa of the bicolor-group; this taxonomic treatment does not seek to resolve the taxonomy of this group outside the Alps but simply to ensure that the new species that we describe is not conspecific with a named taxon in southern Europe. Lastly, we revise the available material to determine the distribution of the Alpine taxa and to examine their phenology, study their host plant preferences based on analyses of pollen loads and report on field observations on habitat selection and floral preferences.

Methods

Specimen selection for molecular analyses

The selection of specimens for molecular analyses was performed iteratively. During an initial phase of this project, 14 specimens representing the observed morphological variation in the bicolor-group were sequenced, including 11 Alpine specimens as well as three specimens from the Apennines. Based on genetics and subsequent morphological examination of these specimens, we were able to recognize four well-separated species in the Alps. For simplicity, these species are named hereafter in a way that anticipates the species delimitation and the description presented below: A. allosa, A. amieti sp. n., A. bicolor and A. montana. Subsequently, 35 additional specimens originating from the Alps were identified using morphology prior to genetic analyses and added to the molecular dataset. Whenever possible, for all Alpine localities where A. amieti sp. n. and A. bicolor were found in sympatry we sequenced at least one specimen of each species. One specimen of A. montana from Greece was sampled, and we included one unclear specimen from the northern Pyrenees that showed morphological affinities with A. allosa.

Morphological examinations of further specimens of the bicolor-group revealed several additional taxa in southeastern Europe, some of which were very similar to A. amieti sp. n. Since at least one species-group name is available for this group of taxa (A. allosa pileata), we expanded our sampling to include six specimens of the bicolor-group from southeastern Europe. These six specimens include three unclear taxa of the bicolor-group, referred to as Andrena sp1, sp2 and sp3. Lastly, we sequenced three specimens of A. rufula Schmiedeknecht, 1883 to generate reference material for this species since females have so far been misidentified in Switzerland (Artmann-Graf 2017).

Details on locality information and BOLD and Genbank accession numbers, as well as GenSeq categories following Chakrabarty et al. (2013) are given in Table 1. All available sequences for the bicolor-group (Schmidt et al. 2015) were downloaded from the BOLD platform (www.boldsystems.org), as well as a selection of European species from the subgenus Euandrena Hedicke, 1933 and representatives of the subgenera Larandrena LaBerge, 1964, Andrena s. str. Fabricius, 1775, Leucandrena Hedicke, 1933, Micrandrena Ashmead, 1899, Melandrena Pérez, 1890 and Ptilandrena Robertson, 1902.

Locality data (including coordinates), BOLD (for the mitochondrial gene cytochrome oxidase, or COI) or Genbank (LW rhodopsin, or opsin) accession numbers and GenSeq category (after Chakrabarty et al. 2013; 1 = holotype; 2 = paratype; 3 = topotype; 4 = vouchered specimens) for sequences produced in this study.

Voucher Name Sex Country Locality Genseq category Coord N Coord E COI Opsin
833 A. chrysopus CH Leuk (S. Gerber) 4 46.303, 7.678 NA MK157241
483 A. fulvata CH Orvin (C. Praz) 4 47.163, 7.213 NA MK157240
873 A. rufula FR St-Guilhem-Le-Désert (34) (J. Litman/C. Praz) 4 43.738, 3.542 HYMAA032-18 MK157242
874 A. rufula FR Blandas (30) (J. Litman/C. Praz) 4 43.912, 3.513 HYMAA033-18 NA
1146 A. rufula CH Laupersdorf (SO) (G. Artmann-Graf) 4 47.332, 7.645 HYMAA021-18 NA
558 A. montana CH Zermatt, Gornergrat (VS) (J. Litman/C. Praz) 4 45.980, 7.790 HYMAA049-18 NA
559 A. montana CH Zermatt, Gornergrat (VS) (J. Litman/C. Praz) 4 45.980, 7.790 HYMAA050-18 MK157255
800 A. montana IT Maiella (Abruzzo) (J. Litman/C. Praz) 4 42.144, 14.112 HYMAA053-18 MK157256
801 A. montana CH Fully, Sorniot (VS) (D. Bénon) 4 46.177, 7.089 HYMAA054-18 MK157257
962 A. montana GR Mt Olympus (Piera) (K. Minachilis) 4 40.105, 22.390 HYMAA044-18 MK157258
1201 A. montana CH Visperterminen (CH) (D. Bénon) 4 46.257, 7.944 HYMAA025-18 NA
1204 A. montana CH Glarus Süd, Linthal (GL) (R. Neumeyer) 4 46.841, 8.932 HYMAA004-18 NA
961 A. sp1 GR Mt Olympus (Piera) (K. Minachilis) 4 40.095, 22.369 HYMAA043-18 MK157259
1252 A. sp1 GR Chelmos mountains (Archaea) (A.W Ebmer) 4 38.002, 22.195 HYMAA037-18 NA
1254 A. sp2 GR Lesbos, Agiassos (North Aegean) (A.W Ebmer) 4 39.054, 26.401 HYMAA038-18 NA
928 A. sp3 GR Kosmas (Arcadia) (J. Litman/C. Praz) 4 37.107, 22.728 HYMAA036-18 NA
842 A. allosa FR Moulinet, Mille Fourches (06) (E. Dufrêne) 4 43.992, 7.435 HYMAA059-18 NA
848 A. allosa CH Gündlischwand (BE) (F. Amiet) 4 46.649, 7.928 HYMAA060-18 NA
1265 A. allosa CH St-Martin, Lovégno (VS) (C. Praz/J. Litman) 4 46.175, 7.470 HYMAA020-18 NA
1271 A. allosa CH Leuk (VS) (S. Giriens) 4 46.335, 7.657 HYMAA022-18 MK157244
1285 A. allosa FR Allos, La Foux d’Allos (04) (C. Praz) 3 44.300, 6.566 HYMAA039-18 MK157243
1293 A. aff allosa FR Larrau (64) (D. Genoud) 4 43.035 -1.023 HYMAA045-18 NA
798 A. amieti group 1 IT Monte Pollino (Basilicata) (J. Litman/C. Praz) 2 39.904, 16.181 HYMAA051-18 NA
799 A. amieti group 1 IT Monte Pollino (Basilicata) (J. Litman/C. Praz) 2 39.904, 16.181 HYMAA052-18 MK157260
1202 A. amieti group 1 CH Kandersteg, Oeschinensee (BE) (C. Praz) 2 46.503, 7.707 HYMAA006-18 NA
802 A. amieti group 2 CH Fully, Sorniot (VS) (D. Bénon) 2 46.167, 7.108 HYMAA055-18 MK157261
804 A. amieti group 2 CH Orsières (VS) (D. Bénon) 2 46.060, 7.091 HYMAA057-18 MK157262
805 A. amieti group 2 CH Fully, Sorniot (VS) (D. Bénon) 2 46.167, 7.108 HYMAA058-18 NA
875 A. amieti group 2 CH Bagnes (VS) (D. Bénon) 2 46.008, 7.346 HYMAA023-18 MK157263
1028 A. amieti group 2 CH Domleschg (GR) (H. Martz) 2 46.785, 9.453 HYMAA031-18 NA
1029 A. amieti group 2 CH Grindelwald (BE) (M. Haider) 2 46.649, 8.077 HYMAA029-18 NA
1031 A. amieti group 2 CH Grindelwald (BE) (M. Haider) 2 46.649, 8.077 HYMAA028-18 NA
1159 A. amieti group 2 FR Le Bourg d’Oisan, Lauvitel (38) (M. Aubert) 2 44.954, 6.060 HYMAA046-18 NA
1193 A. amieti group 2 CH Ennenda (GL) (A. Müller) 2 47.028, 9.105 HYMAA010-18 NA
1197 A. amieti group 2 CH St-Martin, Lovégno (VS) (C. Praz) 2 46.169, 7.472 HYMAA007-18 NA
1203 A. amieti group 2 CH Glarus Süd, Linthal (GL) (R. Neumeyer) 2 46.839, 8.922 HYMAA005-18 NA
1207 A. amieti group 2 CH Kandersteg, Oeschinensee (BE) (J. Litman) 2 46.501, 7.712 HYMAA002-18 NA
1268 A. amieti group 2 CH St-Martin, Lovégno (VS) (C. Praz/J. Litman) 2 46.175, 7.470 HYMAA024-18 MK157264
1284 A. amieti group 2 CH Kandersteg, Oeschinensee (BE) (J. Litman) (HT) 1 46.501, 7.712 HYMAA001-18 NA
1286 A. amieti group 2 FR Allos, La Foux d’Allos (04) (C. Praz) 2 44.300, 6.566 HYMAA040-18 MK157265
556 A. bicolor clade 1 CH Emdt, Schalb (VS) (J. Litman/C. Praz) 4 46.221, 7.816 HYMAA047-18 MK157245
803 A. bicolor clade 1 CH Fully, Sorniot (VS) (D. Bénon) 4 46.172, 7.123 HYMAA056-18 MK157246
871 A. bicolor clade 1 GR Kryoneri (East Attica) (J. Litman/C. Praz) 4 36.955, 22.363 HYMAA034-18 NA
900 A. bicolor clade 1 CH Leuk (VS) (S. Gerber) 4 46.303, 7.678 HYMAA011-18 NA
903 A. bicolor clade 1 CH Vex (VS) (S. Gerber) 4 46.208, 7.411 HYMAA013-18 NA
904 A. bicolor clade 1 CH Vex (VS) (S. Gerber) 4 46.208, 7.405 HYMAA014-18 NA
925 A. bicolor clade 1 CH Leuk (VS) (S. Gerber) 4 46.303, 7.678 HYMAA016-18 MK157247
927 A. bicolor clade 1 CH Gampel-Bratsch (VS) (C. Praz) 4 46.329, 7.720 HYMAA018-18 MK157248
1199 A. bicolor clade 1 CH St-Martin, Lovégno (VS) (C. Praz) 4 46.169, 7.472 HYMAA008-18 NA
557 A. bicolor clade 2 CH St-Niklaus (VS) (J. Litman/C. Praz) 4 46.179, 7.800 HYMAA048-18 MK157249
876 A. bicolor clade 2 GR Kosmas (Arcadia) (J. Litman/C. Praz) 4 37.107, 22.728 HYMAA035-18 NA
902 A. bicolor clade 2 CH Naters (VS) (N. Evéquoz) 4 46.321, 7.967 HYMAA012-18 MK157250
924 A. bicolor clade 2 CH Yvonnand (VD) (M. Khadraoui) 4 46.783, 6.716 HYMAA015-18 NA
926 A. bicolor clade 2 CH Nods (BE) (J. Litman/C. Praz) 4 47.135, 7.063 HYMAA017-18 MK157251
1030 A. bicolor clade 2 CH Vals (GR) (A. Müller) 4 46.621, 9.199 HYMAA026-18 NA
1032 A. bicolor clade 2 CH Bonaduz (GR) (R. Neumeyer) 4 46.809, 9.412 HYMAA019-18 NA
1033 A. bicolor clade 2 CH Domleschg (GR) (H. Martz) 4 46.782, 9.453 HYMAA030-18 NA
1194 A. bicolor clade 2 CH Ennenda (GL) (A. Müller) 4 47.026, 9.104 HYMAA009-18 NA
1195 A. bicolor clade 2 CH Vitznau (LU) (A. Müller) 4 47.024, 8.502 HYMAA027-18 NA
1205 A. bicolor clade 2 CH Kandersteg, Oeschinensee (BE) (J. Litman) 4 46.507, 7.722 HYMAA003-18 NA
1287 A. bicolor clade 2 FR Allos, La Foux d’Allos (04) (C. Praz) 4 44.293, 6.596 HYMAA041-18 MK157253
1288 A. bicolor clade 2 FR Allos, La Foux d’Allos (04) (C. Praz) 4 44.300, 6.566 HYMAA042-18 MK157254

Lab protocols

We first amplified the 658-bp “DNA barcode” fragment of the mitochondrial gene cytochrome oxidase I (hereafter COI) using the LepF and LepR primers (Hebert et al. 2004) with the PCR conditions given therein; if no amplification was obtained, we used the alternate forward primer UAE3 (TAT AGC ATT CCC ACG AAT AAA TAA; Lunt et al. 1996) together with LepR to amplify a 409-bp fragment of the same gene applying the same PCR conditions. If the obtained chromatograms were not clean, possibly because of nuclear pseudogenes, contamination or co-amplification of Wolbachia sequences, the PCR was repeated with the following specific primers designed for the bicolor-group: bicolor_COX1_F: RGC AGG AAT RRT TGG WGC AT and bicolor_COX_R: CTC CVC CYA TWG GRT CAA A, with the following PCR conditions: 4’ 94°; 35 × 45’’ 94°, 45’’ 55°, 45’’ 72°; 7’ 72°. These primers amplified a 563-bp fragment nested within the 658-bp barcoding fragment with a high success rate.

For 22 specimens of the bicolor-group as well as for A. chrysopus Pérez, 1903, A. fulvata Stöckhert, 1930 and A. rufula, we additionally sequenced an approximately 400-bp fragment of the single-copy nuclear gene LW-rhodopsin (hereafter opsin; Danforth et al. 2004) using the following primers specific to the genus Andrena: OpsFor3_Andrena: TTC GAC AGA TAC AAC GTR ATY GTM AAR GG and OpsRev3_Andrena: GCY ARC TTK GCC TCY GCG CT, with the following PCR conditions: 4’ 94°; 35× 45’’ 94°, 45’’ 56°, 45’’ 72°; 7’ 72°.

Lab protocols follow Trunz et al. (2016). DNA was extracted from a single leg using DNA extraction kits (Nucleospin tissue, Macherey-Nagel). PCR-reactions were performed with Hotstart GoTaq polymerase (Promega) with a blank sample as a negative control. PCR products were examined visually using agarose gel electrophoresis and purified enzymatically with a mix of exonuclease and FastAP thermosensitive alkaline phosphatase (Fermentas). Purified PCR products were sequenced bidirectionally with the primers used in the PCR using BigDye Terminator v3.1 technology (Applied Biosystems).

Phylogenetic analyses

Chromatograms were trimmed and assembled in Geneious 6.0.6 (Kearse et al. 2012) and exported consensus were aligned using MAFFT (Katoh and Standley 2013). The resulting matrices were examined and edited in Mesquite (Maddison and Maddison 2018) and converted to amino acid sequences to verify that no stop codon was present. Maximum likelihood analyses were performed in RAxML 8.2.10 (Stamatakis 2014) on the CIPRES server (Miller et al. 2010), performing 1000 bootstrap replicates and applying a GTR + G model to each partition. COI was partitioned into three partitions corresponding to the three nucleotide positions, while opsin was partitioned into four partitions corresponding to the three nucleotide positions and the intron. Trees were rooted at the base of the subgenus Euandrena. Uncorrected p-distances were computed in a test version of Paup 4.0 (Swofford 2002) kindly provided by D. Swofford.

Morphology

Terminology follows Michener (2007); when numbered, metasomal terga and sterna are abbreviated by T and S (e.g., T2 for metasomal tergum 2). When describing vestiture colour in the descriptions and the identification key, “brown” means yellowish-brown, orange-brown or grey-brown and is opposed to “dark brown”, which means dark brown to black. The length of the mouthparts, especially of the maxillary and labial palps, is an important criterion for separating some species of Euandrena. The palps are not cylindrical; instead, several segments are rather flattened and triangular. For estimating the length-to-width ratio of the segments, the width was measured on the flattened side, at the widest part, near the apex, perpendicular to the longitudinal axis of the segment. Measurements were made with ImageJ 1.48 (Abràmoff, Magalhães and Ram 2004) on pictures taken with a Keyence digital microscope VHX 1000 with the stacking option turned off. All pictures of specimens presented here were taken with the same digital microscope with the stacking option turned on.

Databasing

The material of the bicolor-group deposited in the main entomological collections in Switzerland (Table 2) has been revised. In addition, the material deposited in the Warncke collection (OLML) as well as in some additional collections in Europe (Table 2) has been examined. For A. amieti sp. n., every type specimen deposited in a public collection in Switzerland has been provided with a label bearing a unique identifier. Swiss data have been submitted to the Swiss Zoological Records Center (InfoFauna; www.cscf.ch) and data from other countries to GBIF (www.gbif.org) if the specimens are deposited in a public collection in Switzerland. To compare the phenology and altitudinal distribution of the examined species, records from the Alps were filtered to keep only one record per day and locality (regardless of the number of individuals collected). The distribution map was produced with ArcGIS using the “Digital Elevation Model over Europe, EU-DEM” data for relief (available at https://www.eea.europa.eu/data-and-maps/data/eu-dem).

Collections examined with acronyms.

AEC Private collection of P. Andreas Ebmer, Puchenau, Austria
AMC Private collection of Andreas Müller, Wädenswil, Switzerland
BNM Bündner Naturmuseum, Chur, Switzerland
CPC Private collection of Christophe Praz, Neuchâtel, Switzerland
CSEC Private collection of Christian Schmid-Egger, Berlin, Germany
DGC Private collection of David Genoud, Arzens, France
ESC Private collection of Erwin Scheuchl, Ergolding, Germany
ETHZ Eidgenössische Technische Hochschule, Zürich, Switzerland
FAC Private collection of Felix Amiet, Solothurn, Switzerland
GAC Private collection of Georg Artmann-Graf, Olten, Switzerland
HSC Private collection of Hans Schwenninger, Stuttgart, Germany
JVC Private collection of Johannes Voith, Augsburg, Germany
KHC Private collection of Karl Hirt, Menziken, Switzerland
MAC Private collection of Matthieu Aubert, Saint-Jean-De-Buèges, France
MBC Private collection of Markus Bur, Rechthalten, Switzerland
MHC Private collection of Mike Herrmann, Konstanz, Germany
MHNN Muséum d’histoire naturelle de la ville de Neuchâtel, Switzerland
MNHN Muséum Nationale d’Histoire Naturelle, Paris, France
MZL Musée cantonal de zoologie, Lausanne, Switzerland
NMBE Naturhistorisches Museum der Burgergemeinde Bern, Switzerland
OLML Oberösterreichisches Landesmuseum, Linz, Austria
POL-AEGIS University of the Aegean Pollinator collection, Mytilene, Lesbos, Greece
PRUN Research collection of Christophe Praz, University of Neuchâtel, Switzerland
RNC Private collection of Rainer Neumeyer, Zürich, Switzerland
RNF Collection of the Natural Reserves of France
SRLC Collection of the Swiss Bee Red List project, Neuchâtel, Switzerland
ZMHB Museum für Naturkunde, Berlin, Germany

Pollen host preferences

The pollen host preferences of Andrena allosa, A. amieti sp. n. and A. montana were assessed by microscopic analysis of the scopal pollen contents of 24, 50 and 16 female specimens, respectively. These females all originated from the Swiss, French and Italian Alps and were collected from 1921 to 2018. For the bivoltine A. amieti sp. n., the results of the pollen analysis were split between females of the first (or spring) generation (n = 22) and females of the second (or summer) generation (n = 28). The methodology for pollen removal, pollen identification and data evaluation follows Müller (2018).

Results

Genetic analyses

Phylogenetic analyses of COI revealed the presence of six taxonomic units (either monophyletic clades or non-monophyletic grades) for Alpine members of the bicolor-group (Fig. 1): 1. a clade comprising all specimens of Andrena montana. 2. a poorly resolved paraphyletic grade comprising two southern Italian specimens and one Alpine specimen of A. amieti sp. n.; this grade is referred to as “A. amieti group 1”. 3. a clade comprising most Alpine specimens of A. amieti sp. n; this clade is referred to as “A. amieti group 2”. 4. a poorly resolved grade comprising all specimens of A. allosa, rendered paraphyletic by an unclear taxon from Greece, Andrena sp1. 5 and 6. two monophyletic clades in A. bicolor, referred to as A. bicolor clades 1 and 2.

Figure 1. 

Phylogenetic tree based on maximum likelihood analyses of sequence data of the mitochondrial gene COI; numbers above branches indicate statistical support based on 1000 bootstrap replicates (values below 50 are omitted).

The two clades recovered within A. bicolor are those reported in Schmidt et al. (2015); specimens reported therein as A. montana in fact belong to A. amieti sp. n group 2; no specimens of A. montana were sequenced by Schmidt et al. (2015). For Alpine members of A. allosa, A. amieti sp. n., A. bicolor and A. montana, all COI analyses were in agreement with our morphological identifications.

Andrena montana was only distantly related to A. bicolor, A. allosa and A. amieti sp. n. (Fig. 1); genetic distances were on average 6.88% to A. bicolor and 9.48% to A. amieti sp. n., and maximal genetic distance within A. montana was 0.35%. Both groups of A. amieti sp. n. were separated by an average distance of 2.41% (range 1.85 - 3.01%), and maximal within-group distances were 0.93 and 0.26% for groups 1 and 2, respectively. These two groups together formed a paraphyletic unit with respect to a clade formed by A. allosa and the unclear taxon referred to as Andrena sp1 from Greece (bootstrap support 67%; Fig. 1). Maximal genetic distance within Alpine populations of A. allosa was 0.48%; the minimal distance between A. allosa and the divergent specimen tentatively attributed to A. allosa from the Pyrenees (number 1293 on Fig. 1) was 0.73%. Distances between A. allosa and A. amieti sp. n. ranged from 2.77 to 4.71% (average 3.33%). Maximal distance within A. bicolor clade 1 was 0.89%, that within A. bicolor clade 2 (with the exception of the divergent specimen 876 from Greece) 1.44%; the distances between the two clades of A. bicolor (except specimen 876) ranged from 3.13 to 4.40% (average 3.68%). Specimen 876 was separated by distances ranging from 1.04 to 1.46% from the rest of clade 2.

Three unclear taxa are additionally revealed in our analyses: Andrena sp1 represented by two females (specimens 961 from Mount Olympus, Greece, and 1252 from the Chelmos or Aroania Mountains, Greece) was separated by a minimum distance of 1.82% from A. allosa and 4.27% from A. amieti sp. n.; Andrena sp2, represented by a single female from Lesbos (specimen 1254), was separated by a minimum distance of 5.11% from A. amieti sp. n. group 1; and Andrena sp3 from the Peloponnese (specimen 928) separated by a distance of 5.71% from A. bicolor clade 1. Details on these unclear taxa are provided below.

In the ML analysis of COI, A. (Ptilandrena) fulvata and A. (Ptilandrena) angustior (Kirby, 1802) were nested within the subgenus Euandrena, with a clade composed of A. ruficrus Nylander, 1848 and A. fulvida Schenck, 1853 sister to a clade composed of these two species of Ptilandrena and the rest of Euandrena (Fig. 1); A. amieti sp. n., A. allosa, Andrena sp1 and Andrena sp2 formed a well-supported monophyletic group (bootstrap support 96%) that was sister to another well-supported clade (bootstrap support 86%) composed of both clades of A. bicolor and Andrena sp3 (Fig. 1).

Analyses of opsin (Fig. 2) indicated that all specimens of A. montana formed a well-supported clade (bootstrap support 95%). Specimens of A. allosa, of A. amieti sp. n. as well as Andrena sp1 formed another well-supported clade (bootstrap support 98%) separated from A. bicolor; Andrena sp1 was sister to an unresolved clade that contained all specimens of A. allosa and A. amieti sp. n., both of which were not recovered as reciprocally monophyletic clades. The two groups found within A. amieti sp. n. in analyses of COI were not recovered in analyses of opsin. All specimens of A. bicolor formed an unresolved assemblage; all four specimens representing A. bicolor clade 2 (specimens 557, 926, 1287 and 1288) formed a poorly-supported clade separated from clade 1 specimens. Andrena amieti sp. n. was separated from A. bicolor by distances ranging from 1.32% to 2.12%. Lastly, A. (Ptilandrena) fulvata was nested in the subgenus Euandrena as in COI analyses, with at least A. ruficrus (no opsin sequence was available for A. fulvida) being sister to a clade including the other members of Euandrena and A. (Ptilandrena) fulvata.

Figure 2. 

Phylogenetic tree based on maximum likelihood analyses of sequence data of the nuclear gene LW rhodopsin; numbers above branches indicate statistical support based on 1000 bootstrap replicates (values below 50 are omitted).

Species delimitation

Both groups revealed in analyses of COI (Fig. 1) within Andrena amieti sp. n. included specimens of the spring and of the summer generations. The lone Alpine specimen of group 1 (specimen 1202) originates from the type locality of A. amieti sp. n., thus was found in sympatry with numerous specimens of group 2 and was morphologically not visibly divergent from group 2 specimens. The two southern Italian specimens of A. amieti sp. n., both of which were included in group 1, were morphologically slightly distinct from the Alpine specimens (including the lone specimens of group 1 and group 2 specimens). Differences were found in the colour of the vestiture but not in the sculpture. We thus tentatively consider these two groups to represent two mitochondrial lineages within the same biological species. Group 1 lineage appears to be present at low frequency in the Alps since it was represented by only one specimen in the 18 sequenced specimens of A. amieti sp. n.

We also consider the Alpine populations of A. allosa to constitute a distinct and well-separated species despite lack of monophyly (Fig. 1), lack of divergence with A. amieti sp. n. in the nuclear marker (Fig. 2), and the fact that A. allosa (and Andrena sp1) rendered A. amieti sp. n. paraphyletic in analyses of COI. Alpine populations of A. allosa are found in sympatry with A. amieti sp. n. but these two taxa are morphologically, phenologically and biologically well-differentiated from one another. In addition, in spite of non-monophyly in COI-based trees, COI sequences of A. allosa were unique and diagnostic for this species.

Alpine taxa of the bicolor-group

Andrena bicolor is not treated here since this species is widespread and well-known.

Andrena allosa Warncke, 1975

Figs 14, 18, 23, 25, 28, 32, 35, 43, 45, 49, 54.

Andrena allosa Warncke 1975a: 311, ♀, “Allos, Basses-Alpes” [France]. Holotype ♀, OLML, paratypes ♀♀.

Material examined

Type material: holotype female of A. allosa (Fig. 18); additional material: 57 females, 2 males from various localities in France and Switzerland (Suppl. material 1).

Distribution

Western Alps from “Alpes Maritimes” in Southeastern France to Western Switzerland (Fig. 3); possibly Northern Spain and Pyrenees (see note below). A mention from northeastern Italy on the map presented in Gusenleitner and Schwarz (2002: 967) was based on a misidentified specimen of A. amieti sp. n. from “Monte Baldo” (OLML).

Figure 3. 

Distribution of Andrena allosa (red triangles), A. amieti sp. n. (green circles) and A. montana (blue squares) in the Alps.

Notes

We examined one specimen from Spain labelled as follows: “P. Leon Las Senales, 4,875 10.v-12.vi.1967 I & E Yarrow BM 1967-352” (OLML), which agrees with A. allosa in the sculpture of the terga and in its vestiture, but which has a shorter malar space, the clypeus that is not flattened apically as well as shorter mouthparts. This specimen appears to be superficially similar to the sequenced female collected at Larrau, in the “Pyrenées Atlantiques” department in France (number 1293; Fig. 1), which was 0.73% divergent from Alpine specimens of A. allosa. For now, we consider these two specimens to belong to A. allosa in spite of the lack of the most conspicuous diagnostic characters of this species, namely the long malar space and the flattened clypeus.

Phenology

Andrena allosa has only one generation per year from the end of March (one isolated record) to the end of June (Fig. 4).

Figures 4–6. 

Phenology (left diagram) and altitudinal distribution (right diagram) of the three Alpine taxa of the bicolor-group; only data originating from the Alps were used to produce these graphs. 4, Andrena allosa. 5, A. amieti sp. n. 6, A. montana.

Habitat

We found Andrena allosa in various subalpine and alpine grasslands from 1370 m to slightly above the tree line at around 2100 m (Fig. 4), often in close proximity to stands of Crocus albiflorus (Fig. 9). Nesting sites are unknown.

Pollen host preferences

Andrena allosa collected the pollen on eight plant families (Table 3) but had a pronounced affinity for the pollen of Crocus (Iridaceae), which contributed 62.4% to the total pollen grain volume and was recorded in 20 out of 24 scopal loads. This finding is supported by field observations at five different localities in the Swiss and French Alps, where several females of A. allosa simultaneously harvested pollen on Crocus (Fig. 14). Numerous females were observed to collect the pollen from Crocus flowers late in the season, at a time when blooming Crocus were restricted to small patches where snow had remained particularly long. Because of the particularly short blooming period of Crocus, this observation suggests that A. allosa females may locate short-lived stands of Crocus throughout their flying season.

Pollen host preferences of the three Alpine taxa of the bicolor-group. n = total number of pollen loads, N = number of pollen loads from different localities. Countries: CH = Switzerland, F = France, IT = Italy. Plant families: ACE = Aceraceae, AST = Asteraceae, BRA = Brassicaceae, CAM = Campanulaceae, CAR = Caryophyllaceae, CIS = Cistaceae, CLU = Clusiaceae, COL = Colchicaceae, CRA = Crassulaceae, ERI = Ericaceae, FAB = Fabaceae, GER = Geraniaceae, IRI = Iridaceae, LAM = Lamiaceae, LIL = Liliaceae, ORO = Orobanchaceae, PLA = Plantaginaceae, RAN = Ranunculaceae, ROS = Rosaceae, SAL = Salicaceae. Definitions of bee host ranges after Müller and Kuhlmann (2008).

Bee species n N Origin (and number) of pollen loads % pollen grain volume (number of loads) Preferred host % pollen grain volume of preferred host % pure loads of preferred host % loads with preferred host Host range
Andrena allosa Warncke 1975 24 9 CH (20), F (4) IRI (Crocus) 62.4% (20), AST (Cichorioideae) 7.2% (11), AST (Asteroideae) 2.7% (1), AST (Carduoideae) 1.1% (1), RAN 6.8% (7), BRA 6.2% (4), ROS (cf. Potentilla) 5.0% (5), ROS (Geum) 0.9% (2), CIS (Helianthemum) 4.9% (3), LIL (Gagea) 1.7% (2), FAB 0.1% (1), unknown 1.0% (1) Crocus 62.4% 16.7% 83.3% polylectic (8 plant families) with affinity for Crocus (Iridaceae)
Andrena amieti sp. n. (first generation) 22 9 CH (22) AST (Cichorioideae) 28.9% (14), AST (Asteroideae) 5.3% (2), ROS (cf. Potentilla) 9.2% (6), ROS (Geum) 2.5% (2), ROS (other) 9.4% (4), SAL (Salix) 15.6% (7), BRA 7.5% (4), RAN 6.0% (3), GER (Geranium) 3.1% (1), ORO (cf. Euphrasia) 2.6% (1), IRI (Crocus) 2.6% (1), CAR 2.2% (3), ACE (Acer) 2.1% (1), CAM 1.8% (2), ERI 0.7% (1), unknown 0.5% (1) - - - - polylectic (12 plant families)
Andrena amieti sp. n. (second generation) 28 26 CH (28) CAM 79.9% (24), GER (Geranium) 7.1% (6), CAR 4.0% (2), BRA 3.7% (3), COL (Colchicum) 1.8% (1), CIS (Helianthemum) 1.5% (3), AST (Cichorioideae) 1.4% (6), RAN 0.3% (1), PLA (Plantago) 0.2% (1), unknown 0.1% (1) Campanulaceae 79.9% 39.3% 85.7% polylectic (9 plant families) with strong preference for Campanulaceae
Andrena montana, Warncke 1973 16 15 CH (14), F (1), IT (1) CAM 27.1% (6), CIS (Helianthemum) 24.3% (6), CAR 22.0% (7), ORO (cf. Euphrasia) 7.8% (2), AST (Cichorioideae) 5.5% (6), AST (Asteroideae) 0.3% (1), RAN 4.0% (3), PLA (Veronica) 2.7% (1), GER (Geranium) 2.1% (2), CRA 1.5% (2), CLU (Hypericum) 1.1% (1), LAM (Nepetoideae) 1.0% (1), ROS 0.6% (2) - - - - polylectic (12 plant families)

Andrena amieti Praz, Müller & Genoud, sp. n.

Figs 11, 12, 15–17, 20, 22, 24, 26, 27, 29–31, 33, 37, 39, 40, 42, 44, 46–48, 50, 52, Suppl. material 2: S1.

Type locality

Switzerland, Canton of Bern, Municipality of Kandersteg, Northern shore of Lake Oeschinen [Oeschinensee] 46.502N 7.723E, 1590m. This locality is located within the Unesco World Heritage site “Swiss Alps Jungfrau-Aletsch” (Fig. 7).

Figures 7–10. 

Habitats of Andrena amieti sp. n., A. allosa and A. montana. 7, the lake of Oeschinen in the Bernese Alps in Switzerland, UNESCO world heritage site and type locality of A. amieti sp. n. (Picture J. Litman). 8, nesting site of A. amieti sp. n. near Disentis in the Grisons, Switzerland (Picture A. Müller). 9, habitat of A. allosa near Chandolin in the Valais, Switzerland, at a time where the females were already actively collecting pollen (Picture S. Giriens, www.swisswildbees.ch). 10, habitat of A. montana near Zermatt in the Valais, Switzerland (Picture S. Giriens, www.swisswildbees.ch).

Figures 11–14. 

Habitus of Andrena amieti sp. n., A. montana and A. allosa. 11, A. amieti sp. n. male, summer generation, on Caryophyllaceae spec. (Picture S. Giriens, www.swisswildbees.ch). 12, A. amieti sp. n. female, summer generation, with Campanulaceae pollen load (Picture S. Giriens, www.swisswildbees.ch). 13, A. montana female on Caryophyllaceae spec. (Picture D. Bénon, www.swisswildbees.ch). 14, A. allosa female on Crocus albiflorus (Picture D. Genoud).

Holotype

Female of second generation, pinned (Figs 15–17, 26). Original labels: 1. “CH Oeschinensee, 621750/150000, 1640m, 6.vii.2014, leg. J. Litman 029472“. 2. “ACU 29472, GBIFCH00103472” [unique identifier with a graphical barcode]. 3. “Voucher specimen for DNA extraction, Sample 1284, Christophe Praz, University of Neuchâtel” [yellow label, printed]. 4. Holotype female Andrena amieti sp. n. Praz, Müller & Genoud [red label, printed]. Deposited in MHNN.

Figures 15–25. 

Structure of female of Andrena amieti sp. n., A. allosa and A. montana. 15, holotype of A. amieti sp. n. in lateral view. 16, holotype of A. amieti sp. n. in frontal view. 17, clypeus of holotype of A. amieti sp. n. 18, clypeus of holotype of A. allosa. 19, clypeus of A. montana. 20, labrum of A. amieti sp. n. 21, labrum of A. montana. 22, malar space of A. amieti sp. n. 23, malar space of A. allosa. 24, section of mouthparts of A. amieti sp. n. 25, section of mouthparts of A. allosa.

Paratypes

250 males and 237 females from various localities in France, Switzerland, Italy, Germany and Austria (Suppl. material 1).

Note

The description and diagnosis are based on Alpine populations of this species. Southern Italian specimens (Suppl. material 2: Fig. S1) are slightly divergent in the colour of the vestiture; this variation is presented at the end of the description.

Diagnosis

In the female sex, Andrena amieti sp. n. is highly similar to A. bicolor and A. allosa. All three species can easily be separated from A. montana by the dark prepygidial and pygidial fimbria (Fig. 37) (orange-brown in A. montana; Fig. 38) and by the presence of at least some dark hairs laterally on the mesosoma (Fig. 15) (hairs on lateral parts of mesosoma entirely brownish-grey in A. montana). The female differs from A. bicolor in the presence of comparatively long dark hairs on the mesonotum (Fig. 26), usually also on the scutellum; these dark hairs are intermixed with longer, orange brown or grey brown hairs, and many of them are longer than half the length of the light brown hairs. In A. bicolor, there are either no dark hairs on the mesonotum and scutellum, or only very short dark hairs on the mesonotum, their length being visibly smaller than half the length of the light brown hairs. In addition, there are in most cases numerous dark hairs on the propodeum in A. amieti sp. n. (all hairs usually light in A. bicolor), the light hairs on T1–T4 are snow white (Fig. 33) (always yellowish-white in fresh specimens of A. bicolor; Fig. 34), and some hairs between the antenna are always grey-brown (Fig. 16) (hairs between the antennae commonly entirely dark in A. bicolor, although sometimes also grey-brown).

The female of A. amieti sp. n. differs from that of A. allosa by the shorter clypeus with convex preapical area (Fig. 17) (clypeus produced apically, preapical zone flattened medially in A. allosa; Fig. 18) and the short malar space (Fig. 22) (malar space as long as the basal width of antennal segment 3 in A. allosa; Fig. 23); in addition, A. allosa has shagreened, more sparsely punctate tergal discs (especially medially on T2), the margin of T2 hardly impressed medially (Fig. 35), and comparatively longer mouthparts (compare Figs 24, 25).

The male of A. amieti sp. n. (Figs 39, 42) is similar to that of A. bicolor, A. allosa and A. montana, and superficially also to that of A. ruficrus. Separation from the male of A. bicolor may be difficult in worn specimens. In A. amieti sp. n., the light vestiture on mesosoma and metasoma is greyish-white without yellowish hue (Figs 39, 42, 50) (hairs on dorsal side of mesosoma and first terga with yellowish hue in fresh specimens of A. bicolor; Fig. 51). The vestiture is comparatively long on the terga (Fig. 50), in particular medially on T4, the apical fringe of hairs is longer than the tergal margin (in A. bicolor fringe of hairs as long as or shorter than tergal margin; Fig. 51). The disc of T4 has at most a few isolated, dark hairs (disc of T4 with numerous dark hairs in A. bicolor). Lastly, the facial vestiture (Figs 39, 42) is on average lighter than in A. bicolor, at least some grey hairs are found between and around the antennal sockets (in A. bicolor frequently entirely dark).

The male of A. amieti sp. n. is highly similar to that of A. allosa, especially in the first generation (A. allosa has only one generation). Differences are summarized in the key: the maxillary palps are comparatively slightly shorter in A. amieti sp. n. (Fig. 44) than in A. allosa (Fig. 45), and in A. amieti sp. n. the head is comparatively short and broad with little protruding clypeus (Fig. 42) (head comparatively long with protruding clypeus in A. allosa; Fig. 43). In addition, the first recurrent vein usually enters the second submarginal cell in or near its middle and the second submarginal cell is subquadrate or longer than broad (Fig. 48) (in A. allosa, the first recurrent vein enters the second submarginal cell in its basal half, second submarginal cell broader than long; Fig. 49).

In its vestiture, the male of A. amieti sp. n. is similar to that of A. montana; the best diagnostic characters are the shape of the labral appendix, which is wider than long in A. amieti sp. n. (Fig. 40) and as long as apically wide in A. montana (Fig. 41), and the width of the penis valves and of the gonostylus: the gonostylus is more slender and the penis valves broader in A. montana (Fig. 61) than in A. amieti sp. n. (Fig. 52), although for these genitalic characters direct comparison with reference material is necessary. In addition, the mesonotum is nearly entirely shagreened in males of A. montana, while it is at least partly shiny in males of the second generation of A. amieti sp. n. (Fig. 47). Lastly, males of A. amieti sp. n. are also superficially similar to those of A. ruficrus, which usually have the apex of the hind tibia partly orange and the clypeal vestiture entirely show-white; in addition, the penis valves are broader basally in A. ruficrus (Fig. 60) than in A. amieti sp. n (Fig. 52).

Description

Female: Body size and proportions: Very similar to A. bicolor. Body length approximately 9mm, slightly smaller on average than A. allosa (body length approximately 9.5mm-10mm). Head slightly broader than long (Fig. 16), clypeus broader than long. Malar space short (Fig. 22), as in A. bicolor, shorter than in A. allosa (Fig. 23), length without impressed area at most equal to half the base of third antennal segment. Gena slightly broader than compound eye in lateral view. Interocellar distance approximately 2 times diameter of lateral ocellus. Ocelloccipital distance approximately 0.9 times diameter of lateral ocellus. Third antennal segment longer than fourth and fifth together, the latter two broader than long, segments 6–11 subquadrate, 12 longer than broad (Fig. 16). Labral process trapezoidal, its apical width larger than its length (Fig. 20), comparatively broader than in A. montana (Fig. 21). Mouthparts (Fig. 24) as in A. bicolor, shorter than in A. allosa (Fig. 25), in particular segment 4 of maxillary palpus only three times as long as apically broad (in A. allosa at least four times as long as apically broad), and segment 2 of labial palpus hardly longer than broad (in A. allosa at least twice as long than broad).

Wing venation: As in A. bicolor; the first recurrent vein enters the second submarginal cell in its middle or nearly so, and second submarginal cell subquadrate or longer than broad (Fig. 27) (in A. allosa, the first recurrent vein usually enters the second submarginal cell in its basal half; second submarginal cell broader than long; Fig. 28).

Figures 26–32. 

Structure of female of Andrena amieti sp. n. and A. allosa. 26, vestiture of dorsal side of mesosoma of holotype of A. amieti sp. n. 27, section of right forewing of A. amieti sp. n. 28, section of right forewing of A. allosa. 29, dorsal view of mesosoma of spring generation of A. amieti sp. n. 30, dorsal view of mesosoma of summer generation of A. amieti sp. n. 31, propodeum of A. amieti sp. n. 32, propodeum of A. allosa.

Integument colour: As in A. bicolor, integument black or dark brown, including flagellum and tegulae, apical margin of T1–T4 slightly lighter, tarsal segments 2–4 dark orange-brown, tarsal claws weakly ferruginous, hind tibial spurs light brown. Wing venation (including stigma; Fig. 27) predominantly brown as in A. bicolor.

Vestiture: Entire body vestiture made of simple to weakly branched hairs, more strongly branched hairs are present between antennal sockets, plumose hairs in propodeal corbicula, floculus, and prepygidial and pygidial fimbria. Hairs on head predominantly dark brown (Fig. 16), always with grey-brown to greyish-white hairs between antennal sockets (sometimes also around antennal sockets) and medially on vertex (in A. bicolor hairs on head often entirely dark); extent of grey vestiture on face variable but vestiture on average lighter than in A. bicolor. Hairs on mesosoma predominantly grey-brown, including on lateral and ventral sides (Fig. 15); hairs never bright orange-brown as in fresh specimens of A. bicolor; mesonotum and usually also scutellum with intermixed long, grey-brown hairs and short, dark brown hairs (Fig. 26) (in A. bicolor either no dark hairs, or dark hairs shorter; see above); tegulae covered with short dark hairs (Fig. 26) (in A. bicolor light brown hairs); propodeum with intermixed grey and dark brown hairs (in A. bicolor usually without dark brown hairs); on lateral side of mesosoma, hairs on average lighter than in A. bicolor, sometimes entirely grey-brown (Fig. 15), dark hairs mostly with grey tips except sometimes under tegula, where hairs can be entirely dark; ventral part of mesosoma with grey-brown hairs becoming lighter apically. Leg vestiture nearly as in A. bicolor, flocculus grey-brown to whitish grey, on average lighter than in A. bicolor; scopa yellowish brown, hairs on basitarsi 2 and 3 usually yellowish brown (in A. bicolor usually dark brown). T1 mostly covered with snow-white hairs (in A. bicolor yellowish-white), only a few isolated, dark hairs on anterior part of disc and numerous dark hairs on vertical anterior part of tergum; T2 and T3 covered with long, snow white hairs (Fig. 33) (in A. bicolor hairs yellowish-white, Fig. 34 and usually slightly shorter medially) forming very loose apical fringes, not hiding cuticula, discs with short, erect dark hairs, more numerous on T3 than T2; T4 covered with long, dark hairs forming loose apical fringe and with short, erect dark hairs on disc; prepygidial and pygidial fimbriae dark brown (Fig. 37). Sterna covered with dark, erect hairs, apical margin with apical fringes of dark, plumose hairs. In first generation, vestiture on average longer than in second, especially on clypeus, mesosoma and basal terga.

Sculpture: The variation observed in the sculpture of A. amieti sp. n. represents the middle range of the much wider scuptural variation observed in A. bicolor. Head: Facial fovea narrow (Fig. 16), as in A. bicolor; clypeus (Fig. 17) densely punctured, more sparsely punctured laterally and basally than apically, there with shiny interspaces that are on average one puncture diameter wide, often with longitudinal irregularities (more so than in A. bicolor); apical part of clypeus without flat area as observed in A. allosa; frons unpunctured with numerous longitudinal ridges, as in A. bicolor. Mesosoma: mesonotum (Figs 29, 30) with well-visible punctures, interspaces nearly entirely dull and on average 3–4 puncture diameters in first generation (Fig. 29), shiny or silk-shiny and 2–3 puncture diameters in second generation (Fig. 30). Propodeum (Fig. 31) as in A. bicolor, propodeal triangle without punctures, dull, finely sculptured, anteriorly with a few longitudinal wrinkles, wrinkles less visible than in A. allosa (Fig. 32). Metasoma: terga sparsely punctured (Fig. 33), within variation observed in A. bicolor (Fig. 34), on average slightly more shagreened, especially apical tergal margin of T2 (compare specimens of same generation); first generation: disc of T1 very sparsely punctured, that of T2 sparsely punctured with interspaces equal to 4–5 puncture diameters; surface of tergal discs mostly shagreened to silk shiny, apical margins shagreened, mostly unpunctured, that of T2 weakly impressed medially, as in A. bicolor but more so than in A. allosa; second generation (Fig. 33): punctation on average denser (interspaces equal to 3–4 puncture diameters on disc of T1, and 2–3 puncture diameters on disc of T2; sculpture of surface on average shinier, although usually still weakly shagreened; apical margins as in first generation, usually at least slightly shagreened.

Figures 33–38. 

Structure of female of Andrena amieti sp. n., A. bicolor, A. allosa and A. montana. 33, T1–T4 of A. amieti sp. n. 34, T1–T4 of A. bicolor. 35, T1–T4 of A. allosa. 36, T1–T4 of A. montana. 37, prepygidial and pygidial fimbria of A. amieti sp. n. 38, prepygidial and pygidial fimbria of A. montana.

Male: Body size and proportions: Body length approximately 7–8mm, similar to A. bicolor; the two males of A. allosa examined were 8-8.5mm long. Head slightly broader than long, clypeus little protruding apically (Fig. 42) (in A. allosa, clypeus slightly more strongly protruding; Fig. 43); gena slightly broader than compound eye in lateral view; interocellar distance approximately 2.5–3 times diameter of lateral ocellus; ocelloccipital distance approximately equal to diameter of lateral ocellus; length of third antennal segment approximately 1.3 times length of fourth antennal segment (in A. bicolor third antennal segment often shorter, 1.1 times length of fourth segment), fourth slightly shorter than broad or subquadrate, segments 5–9 slightly longer than broad, 10–12 subquadrate, 13 longer than broad. Labral process trapezoidal, its apical margin slightly emarginate (Fig. 40) (in A. montana, process longer; Fig. 41). Mouthparts (Fig. 44) as in A. bicolor, slightly shorter than in A. allosa (Fig. 45) although differences not as striking as in female; segment 4 of maxillary palpus only three times as long as apically broad (in A. allosa approximately four times as long as apically broad).

Figures 39–51. 

Structure of male of Andrena amieti sp. n., A. montana, A. allosa and A. bicolor. 39, male of A. amieti sp. n. in lateral view. 40, labrum of A. amieti sp. n. 41, labrum of A. montana. 42, male of A. amieti sp. n. in frontal view. 43, male of A. allosa in frontal view. 44, section of mouthparts of A. amieti sp. n. 45, section of mouthparts of A. allosa. 46, dorsal view of mesosoma of spring generation of A. amieti sp. n. 47, dorsal view of mesosoma of summer generation of A. amieti sp. n. 48, section of right forewing of A. amieti sp. n. 49, section of right forewing of A. allosa. 50, T1–T4 of A. amieti sp. n. 51, T1–T4 of A. bicolor.

Wing venation: As in female, the first recurrent vein usually enters the second submarginal cell in or near its middle; second submarginal cell subquadrate or longer than broad (Fig. 48) (in two examined specimens of A. allosa, the first recurrent vein enters the second submarginal cell in its basal half, second submarginal cell broader than long; Fig. 49).

Integument colour: As in female.

Vestiture: Entire body vestiture grey-white without yellowish hue (Figs 39, 42, 50) (light hairs always with yellowish hue in fresh specimens of A. bicolor; this yellowish hue is not apparent in worn specimens), except as follows: facial vestiture varying from predominantly dark (as on Fig. 42), with only a few grey hairs between and around antennal sockets, on scape and on vertex medially, to nearly predominantly grey-white, with dark hairs restricted to apical and lateral parts of clypeus (facial vestiture often entirely dark in A. bicolor; clypeus entirely covered by snow-white vestiture in A. ruficrus, and by light brownish-grey hairs and dark hairs laterally in A. montana), area along inner margin of compound eye, and lateral parts of head. Mesonotum with a few isolated dark hairs among longer, grey hairs (Figs 46, 47); lateral side of mesosoma with a few isolated dark hairs below tegula. Propodeum with intermixed dark and light grey hairs. Legs mostly covered with grey-white hairs, hairs on external surface of all tibiae yellowish and on ventral surface of all tarsal segments yellowish-white. Surface of T1 with a few isolated, dark hairs on anterior, vertical part. Vestiture of T2–T4 snow white (Fig. 50) without dark hairs or at most with a few isolated dark hairs basally on disc of T4 (in A. bicolor, vestiture of T1–T4 yellowish white in fresh specimens, Fig. 51; disc of T4 predominantly covered with dark hairs, and with numerous dark hairs basally on disc of T3; dark hairs may appear light grey in worn specimens).

Sculpture: As for female, the sculpture of the male of A. amieti sp. n. is similar to that of A. bicolor, which is particularly variable. Head: clypeus densely punctured, interspace shiny or weakly shagreened, mostly narrower than one puncture diameter (Fig. 42); frons nearly unpunctured with numerous longitudinal ridges. Mesosoma: mesonotum in first generation (Fig. 46) entirely shagreened with sparse and little visible punctures, interspaces equal to 4–5 puncture diameters (similar to A. allosa or to first generation of A. bicolor), in second generation (Fig. 47) nearly always with shiny area medially, punctation well-visible, on average slightly sparser than in second generation of A. bicolor, interspaces commonly over 3–4 puncture diameters (in A. bicolor rarely over 3 puncture diameters); propodeum as in female. Mesosoma: terga (Fig. 50) similar to A. bicolor, on average slightly more sparsely punctate and more shagreened, but within variation observed in A. bicolor; punctures fine (slightly coarser on T1), interspaces often more than 5 times puncture diameters. Tergal margins usually partly shagreened (Fig. 50) (in second generation of A. bicolor often entirely shiny). Structure of S8 not visibly different from A. bicolor.

Genitalia: As on Fig. 52, very similar to Andrena bicolor (Fig. 53), dorsal lobe of gonocoxite weakly developed, gonostylus simple, regularly spatulate, external margin regularly rounded, penis valves narrow, hardly broadened basally.

Figures 52–61. 

Male genitalia of species of Euandrena. 52, Andrena amieti sp. n. 53, A. bicolor. 54, A. allosa. 55, A. chrysopus. 56. A. symphyti. 57, A. fulvida. 58, A. vulpecula. 59, A. rufula. 60, A. ruficrus. 61, A. montana.

Geographic variation

In females from Southern Italy (Suppl. material 2: Fig. S1), the vestiture is nearly entirely grey-white, including on all parts of the mesosoma (isolated dark hairs on mesonotum and scutellum excepted), with no brownish hue, in contrast to Alpine specimens where the mesosonal vestiture is predominantly brown. No difference is found in vestiture colour in male specimens.

Etymology

This species is named in honor of Felix Amiet, who has greatly contributed to our understanding of Central European bees, including the four species presented here.

Additional comments

There are numerous species-group names currently treated as junior synonyms of A. bicolor (Gusenleitner and Schwarz 2002). We carefully examined the description of each of these names if their type locality was located within the known range of A. amieti sp. n. The description of none of these species-group names points to A. amieti sp. n. (see also comments on A. croatica Friese, 1887 stat. rev., below). A. gwynana var. testacea Dalla Torre, 1877, described from Seefeld (Tyrol) and whose type material is presumably lost, originates from a region where A. amieti sp. n. might occur. The very brief original description merely mentions the “testaceous” hind tibia, which corresponds neither to A. bicolor nor to A. amieti sp. n., and which does not point to a differential character between these two species. For these reasons, we keep this taxon as a junior synonym of A. bicolor. Lastly, several taxa of Euandrena have been described from Ukraine, Central Asia or Russia (Gusenleitner and Schwarz 2002). Given that A. amieti sp. n. appears to be restricted to the Alps, the Apennines and the Pyrenees, we consider it unlikely that any of these taxa will turn out to be conspecific with A. amieti sp. n.

Distribution

Alps from Southern France to East Tyrol (Austria), including most of the Swiss Alps and the Allgäu Alps in Southern Germany, as well as the Italian Alps (Fig. 3); Monte Pollino in Southern Italy; one isolated record from the Pyrenees.

Phenology

According to our data and field observations, Andrena amieti sp. n. has two generations per year (Fig. 5). The earliest males of the first generation were collected at the end of March, and some worn females collected during the first half of June probably belonged to the first generation (thus some females collected in June and seemingly belonging to the second “peak” on Fig. 5 were in fact most likely of the first generation). The second generation is active from mid-June to the first half of September. While the first generation was underrepresented in the examined material, intensive surveys in three Swiss localities where the species was found to be abundant consistently revealed the presence of two generations, with several females observed from May to early June, then numerous fresh males and fresh females observed from the end of June to mid-August.

Habitat

Andrena amieti sp. n. has been found from an altitude of approximately 1000m up to slightly above the tree line at around 2300m in the Valais; one isolated record is from an elevation of 2500m (Fig. 5). Foraging females were observed in various habitats, often in clear forests but also in meadows, subalpine grasslands and scree slopes. Patrolling males, probably indicating the location of nesting aggregations, were regularly found on disturbed terrains with little vegetation, such as dry river beds, scree slopes or avalanche corridors (Fig. 8). In these presumed nesting sites, the soil was not particularly sandy, but rather consisting of gravel mixed with sand or clay.

Pollen host preferences

Andrena amieti sp. n. collected the pollen from flowers belonging to 15 plant families (Table 3). Females of the spring generation exploited a wide spectrum of pollen hosts, among which Asteraceae, Rosaceae and Salix (Salicaceae) predominated; pollen of these three plant taxa represented 70.9% of the total pollen grain volume. In striking contrast, females of the summer generation exhibited a strong preference for the pollen of Campanulaceae, which contributed 79.9% to the total pollen grain volume and was recorded in 24 out of 28 scopal loads (Fig. 12). Field observations revealed that flowers of both Campanula and Phyteuma serve as pollen hosts among the Campanulaceae.

Andrena montana Warncke 1973

Figs 13, 19, 21, 36, 38, 41, 61.

Andrena montana Warncke 1973: 33, ♀♂, “Bozen/Italien [Bolzano, Italy]” Holotype ♀, OLML, paratypes ♀♀♂♂.

Material examined

Holotype female (OLML); additional material: 31 males, 29 females originating from France, Switzerland, Germany, Italy, Greece and Macedonia (Suppl. material 1).

Distribution

Alps from “Alpes Maritimes” in Southeastern France to Switzerland, Northern Italy and southern Germany (Fig. 3); Apennines; Balkans (Greece, Macedonia). The species is expected to also occur in the western Austrian alps. One mention from Albania (Warncke 1973) in fact refers to A. ruficrus.

Phenology

Andrena montana has only one generation per year from mid-June to August (Fig. 6).

Habitat

Andrena montana has been found from 1500 m to 2750m (Fig. 6) and is more often observed above the tree line than any other Alpine species of the bicolor-group. Pollen-collecting females were observed in flower-rich alpine grasslands (Fig. 10). A few males were observed to patrol on areas with bare soil sparsely covered by vegetation; the soil was not particularly sandy but rather made of gravel mixed with sand and clay.

Pollen host preferences

Andrena montana collected the pollen on 12 plant families (Table 3). Flowers of the Campanulaceae, Helianthemum (Cistaceae) and Caryophyllaceae were the most important pollen hosts (Fig. 13); their pollen contributed 73.4% to the total pollen volume, whereas the pollen of all other taxa was represented by less than 10% each.

Identification key to Central European species of Andrena subgenus Euandrena

Unique characters in bold; non-unique characters in regular font; characters given in order of importance. This key will work in most of Continental Europe except in the Pyrenees, in Spain and in the Balkans.

Females

Euandrena : body-length 9–11mm. Fovea comparatively narrow, its broadest width less than 1.5 times broadest width of scape (slightly broader in A. fulvida). Third antennal segment longer than segments 4 and 5 together. Metasomal integument predominantly dark-brown. Mesonotal vestiture consisting of long, simple or shortly branched hairs. Flocculus neither particularly long nor bent apically. Propodeal triangle at most weakly wrinkled, coarsely wrinkled anteriorly in A. fulvida and in A. allosa (Fig. 32); propodeum laterally with little-developed corbicula. Hind femur without row of thorn-like projections. Dorsolateral angle of pronotum without elevated transverse carina.

1 Head clearly longer than broad. T2-T4 with dense apical fringes of snow-white hairs, fringes continuous on T3 and T4 and hiding cuticula. Head vestiture entirely light grey. Hind basitarsus ferruginous. Surface of metasomal terga smooth, not shagreened. Prepygidial and pygidial fimbria orange to brown-orange. Andrena chrysopus
Head at most as long as broad. If present on T2-T4, apical fringes of hairs not hiding cuticula or interrupted medially. Combination of other characters different. 2
2 Hind basitarsus and hind tibia ferruginous. Terga regularly finely shagreened, silk-shiny, with comparatively fine, little visible punctures. Fovea comparatively short and broad, little narrowed inferiorly, 1.5 times broader than broadest width of flagellum. Facial vestiture brown medially, dark brown laterally. Andrena ruficrus
Hind basitarsus and tibia dark brown. Terga shagreened or shiny with often distinct punctures. Fovea usually longer and often narrower (not clearly so in A. fulvida). Vestiture variable. 3
3 Clypeus shiny, with particularly coarse punctures and slightly elevated, impunctate longitudinal line (Fig. S2). Entire body vestiture brown, except dark on T6, sterna, tibia and tarsi, as well as a few isolated dark hairs on face along compound eye. Mesonotum and terga shagreened, the latter with isolated, little visible punctures. Andrena symphyti
Clypeus shiny or shagreened, without particularly coarse punctures, with or without impunctate longitudinal line, but longitudinal line never elevated. Mesosomal vestiture variable. Sculpture of mesonotum and terga variable. 4
4 Facial vestiture nearly entirely brown, with at most a few isolated dark hairs along compound eye and vestiture on mesonotum brown intermixed with numerous, slightly shorter dark hairs. Anterior area of propodeum comparatively coarsely wrinkled (as on Fig. 32). Surface of T1 shiny or only weakly shagreened, medially between disc and depression with a few longitudinal irregularities [as in the distantly related A. (Melandrena) vaga]. Vestiture on lateral side of mesosoma entirely brown, without black hairs. T2 and T3 fringed laterally with weak, brownish-white apical fasciae. Mesonotum and scutellum mat, the former comparatively densely punctate. Terga regularly, finely punctate, shiny or only slightly shagreened. Andrena fulvida
Vestiture of head and mesonotum different, never with the combination of predominantly brown facial hairs and intermixed brown and dark hairs on mesonotum. Propodeum more finely wrinkled (except in A. allosa). Surface of T1 shagreened or shiny, without longitudinal irregularities. Vestiture and sculpture variable. 5
5 Integument of apical margin of T1–T4 ivory-colored (Figs S3, S4), basal part of tergal margins weakly translucid. Entire body vestiture brown, except facial vestiture, which is predominantly dark with more or less brown hairs medially, a few isolated, dark hairs on mesonotum and some dark hairs on T5 and T6 6
Integument of apical margin of T1–T4 not ivory-colored, at most narrowly light brown apically; basal part of tergal margins not translucid; vestiture variable. 7
6 Tergal discs on average more coarsely and densely punctate, interspaces on disc of T2 equal to 1.5-2 puncture diameters (Fig. S3). Surface of tergal discs strongly convex, apical margin strongly impressed. Fovea comparatively long, starting below middle of antennal sockets. Andrena granulosa
Tergal discs on average more finely and sparsely punctate, interspaces on disc of T2 equal to 3-4 puncture diameters or more (Fig. S4). Surface of tergal discs nearly flat, apical margins less strongly impressed. Fovea comparatively short, starting well-above middle of antennal sockets. Andrena vulpecula
7 Vestiture of mesosoma entirely brown-orange (quickly fading to grey brown), including on lateral and ventral sides, mesonotum without intermixed long or short dark hairs. Facial vestiture predominantly brown orange, only a few dark hairs along compound eyes and sometimes on clypeus (Fig. S5). Scopa orange. Prepygidial and pygidial fimbria dark brown. Facial fovea comparatively little narrowed ventrally (Fig. S5). Inferior part of clypeus shiny, sparsely and coarsely punctate (Fig. S5). Mesonotum strongly shagreened, with shallow punctures. Tergal discs superficially shagreened, silk-shiny, sparsely punctate, interspaces 4-5 puncture diameters medially on disc of T3. Apical margins of T2 and T3 slightly depressed, shagreened, with few, shallow punctures, often light-brown apically. [Females of A. rufula are often confused with those of A. (Ptilandrena) fulvata, which can be recognized by the less shiny clypeus with a weak, longitudinal impression medially (sometimes not clearly visible!), by the strongly shagreened and very sparsely punctate terga, and by the medially not depressed apical margins of T2 and T3]. Andrena rufula
Lateral and ventral parts of mesosoma either with at least some dark hairs, or entirely grey-brown, but then mesonotum with some intermixed dark hairs; other characters variables. 8
8 Prepygidial and pygidial fimbria yellowish-orange (Fig. 38), at most with a few brown hairs next to pygidial plate. Hairs on face nearly entirely grey-white, only a few dark hairs laterally and on vertex (Fig. 19). Hairs on lateral and ventral parts of mesosoma light grey, at most with isolated dark hairs. Labral appendix long, approximately as long as wide apically (Fig. 21). Hairs on T1-T4 entirely grey-white, without erect, dark hairs (Fig. 36). Fovea comparatively broad, broadest width wider than broadest width of flagellum and little narrowed ventrally. Tergal discs entirely finely shagreened (Fig. 36), silk-shiny, regularly sparsely punctate, punctures fine, interspaces 3-4 puncture diameters. Margins of T2-T4 hardly depressed, shagreened Andrena montana
Prepygidial and pygidial fimbria dark brown (Fig. 37), never orange (but whitish brown in old specimens). Hairs on face predominantly dark, light brown or grey only between and around antennae. Lateral and ventral sides of mesosoma with extended areas covered with dark hairs. Labral appendix wider than long (Fig. 20). Sculpture variable. 9
9 Smallest length of malar space (without impressed margin close to mandible) comparatively long, approximately as long as basal width of antennal segment 3 (Fig. 23). Clypeus laterally convex, medially nearly flat, densely and coarsely punctate, interspaces conspicuously shiny at least apically (Fig. 18). Terga entirely shagreened, disc of T2-T4 very sparsely punctate, nearly impunctate medially (Fig. 35); T2 and T3 with apical margin weakly impressed medially (Fig. 35) [sculpture nearly as in A. (Ptilandrena) fulvata]. Anterior part of propodeum with comparatively strong wrinkles (Fig. 32). First recurrent vein enters second submarginal cell in basal half, second submarginal cell usually longer than broad (Fig. 28). Vestiture dark on head (except for some brown hairs on vertex), lateral and ventral sides of mesosoma, and T4-T6. Mesonotum with brown hairs intermixed with some dark hairs that are 2/3 as long as the brown hairs. Andrena allosa
Malar space shorter (Fig. 22) (in some rare females with particularly long head, malar space nearly as long as in A. allosa). Clypeus regularly convex medially (Fig. 17), shiny or shagreened. Sculpture of terga variable, but disc of T2-T4 comparatively more densely punctate (Figs 33, 34); apical margin of T2 and T3 visibly impressed medially (Figs 33, 34). Anterior part of propodeum at most with only fine wrinkles (Fig. 31). First recurrent vein usually enters second submarginal cell in its middle, second submarginal cell usually quadrate (Fig. 27). Mesonotal vestiture variable. 10
10 Brown vestiture of mesonotum and usually also of scutellum intermixed with numerous, comparatively long dark brown hairs; dark hairs on average at least half as long as light hairs (Fig. 26). Light hairs on T1-T4 snow white without yellowish hue (Fig. 33). Hairs on mesonotum greyish brown (Figs 15, 26), never bright orange brown. Hairs laterally and ventrally on mesosoma dark grey to dark brown (Figs 15, 26), but dark hairs often with greyish tips. Propodeum with numerous dark hairs. Hairs on face predominantly dark but always grey between or around antennal sockets (Fig. 16). Sculpture variable, mesonotum sparsely punctate, medially with interspaces up to 3 puncture diameters, punctures well visible (Figs 29, 30), interspaces often silk-shiny medially, never entirely smooth with no sculpture. Tergal discs shagreened and comparatively sparsely punctate (Fig. 33). Andrena amieti sp. n.
Mesonotum either without intermixed dark brown hairs, or with very short dark hairs, then dark hairs on average less than half the length of light hairs (very rarely longer). Scutellum without intermixed dark hairs. Light hairs on T1-T4 yellowish-brown (Fig. 34) in fresh specimens (but grey-white in worn specimens). Hairs on mesonotum orange-brown in fresh specimens. Hairs laterally and ventrally on mesosoma mostly dark, dark hairs without grey tips, except in second generation, especially in the southern and western Alps; propodeum usually without dark hairs. Hairs on face usually entirely dark, although in the southern and western Alps often with numerous grey-brown hairs in second generation. Sculpture of mesonotum and of terga variable and encompassing variation seen in previous species. A. bicolor

Males

Euandrena : The delineation of the subgenus Euandrena is challenging in males and it is thus advised to use available keys to the genus Andrena (e.g., Schmid-Egger and Scheuchl 1997; Amiet et al. 2010) in addition to this diagnosis. Body-length 7–11mm. Integument of clypeus and of terga dark brown. Propodeal triangle at most weakly wrinkled (coarsely wrinkled anteriorly in A. fulvida). Third antennal segment usually clearly longer than fourth (except in A. rufula). Labral appendix not projecting anteriorly. Apical margin of clypeus regularly rounded, not curved anteriorly. Male genitalia usually simple (Figs 52–61), dorsal gonocoxite lobes short and rounded, penis valves usually narrow basally, their maximal width subequal to apical width of gonostylus (except in A. fulvida, Fig. 57 and A. chrysopus, Fig. 55), gonostylus not enlarged apically (except in A. fulvida, Fig. 57).

1 Antennal segment 3 as long as segments 4 and 5 together, segment 4 clearly wider than long, segment 5 wider than long to approximately as long as wide. Entire vestiture brown or grey, especially facial vestiture nearly entirely grey, at most with a few isolated dark hairs laterally. 2
Antennal segment 3 shorter than segments 4 and 5 together, segment 4 as long as wide or longer, segment 5 longer than wide. Vestiture variable 3
2 Terga shiny and densely punctate, interspaces approximately 1-2 puncture diameters on terga 2-4. Hind basitarsus and usually also hind tibia ferruginous. Penis valves basally broadened with lateral, angular extension (Fig. 55) Andrena chrysopus
Terga dull with little visible and sparse punctation. Hind basitarsus dark. Penis valves basally without lateral extension (Fig. 56), although broader than in A. bicolor. Andrena symphyti
3 Penis valves basally broad, broader than gonostylus apically (Fig. 57). Gonostylus apically broad, spatulate, inner angle nearly quadrate (Fig. 57). Propodeum anteriorly coarsely wrinkled. Entire body vestiture brown, except a few isolated dark hairs laterally on face. Andrena fulvida
Penis valves basally narrow, subequal to apical width of gonostylus (Figs 52–54, 58–61). Gonostylus apically not broadened (Figs 52–54, 58–61). Propodeum comparatively finely wrinkled anteriorly. Vestiture variable. 4
4 Apical margin of terga 1-5 ivory coloured (Figs S6, S7). Terga impressed basally and apically, thus tergal discs comparatively strongly convex. Penis valves basally slightly broader than in A. bicolor, approximately as broad as apical width of gonostylus (Fig. 58). 5
Apical margin of terga 1-5 not ivory-coloured, at most T3-5 with a narrow, yellowish brown margin. Terga little impressed basally, disc comparatively flat. Width of penis valves variable 6
5 Tergal discs coarsely and densely punctate, on T2 interspaces equal to 2 puncture diameters (Fig. S6). Surface of tergal discs strongly convex, apical margin strongly impressed, separated from disc by an abrupt step even medially (Fig. S6). Andrena granulosa
Tergal discs finely and sparsely punctate, on T2 interspaces equal to 3-4 puncture diameters (Fig. S7). Surface of tergal discs not strongly convex (similar to A. bicolor), apical margin weakly impressed, separated from disc medially by a gradual step (Fig. S7). Andrena vulpecula
6 Apex of gonostylus triangular, external angle acute (Fig. 59). Third antennal segment subequal to or only slightly longer than fourth. Vestiture entirely brown, except a few dark hairs laterally on face. Penis valves slightly broadened basally (Fig. 59). Mesonotum shagreened with little visible punctures. Andrena rufula
Apex of gonostylus rounded, without acute angle (Figs 52–54, 60, 61). Third antennal segment visibly longer than fourth. Vestiture, basal width of penis valves and sculpture of mesonotum and terga variable. 7
7 Hind tibia usually more or less ferruginous near apex, hind basitarsus sometimes ferruginous (hind leg rarely entirely dark). Clypeal vestiture predominantly snow white and propodeum laterally with numerous dark hairs. Clypeus densely and finely punctate, interspaces very narrow, mat. Sterna with dense apical fringes of snow-white hairs. Vestiture laterally on mesosoma and on terga predominantly snow white, on mesonotum grey-brown. Terga silk-shiny, punctation very fine, little visible. Penis valves slightly broadened basally (Fig. 60). Andrena ruficrus
Hind tibia and basitarsus dark. Clypeal vestiture dark or light, but if light grey-brown (fading to grey-white), then propodeum without dark hairs. Clypeus more coarsely punctate, interspaces wider and at least partly shiny (Figs 40–42). Vestiture and basal width of penis valves variable. 8
8 Facial hairs predominantly grey-white, especially on clypeus, including hairs along apical clypeal margin; dark hairs restricted to area along compound eyes and on frons. Labral appendix longer than apically wide (Fig. 41). Mesonotum with grey-brown hairs with a yellowish hue in fresh specimens and with isolated, comparatively long intermixed dark hairs. Propodeum and lateral parts of mesosoma without dark hairs. Penis valves basally slightly enlarged (difference with the following species only visible in direct comparison!), its maximal width approximately equal to the width of the gonostylus; apex of gonostylus comparatively narrow, parallel-sided, exterior margin straight or very slightly concave (Fig. 61). Discs of T2-T5 finely shagreened, silk-shiny, never entirely smooth, finely and regularly punctate, punctures comparatively small. Andrena montana
Facial hairs usually predominantly dark, grey hairs, if present, restricted to area around antennal sockets or to base of clypeus (Figs 39, 42), hairs along apical clypeal margin predominantly dark. Labral appendix wider than long (Fig. 40). Vestiture on mesosoma variable, at least some dark hairs present laterally. Penis valves narrow basally, not broadened and clearly narrower than apical width of gonostylus; gonostylus apex regularly rounded, external margin convex (Figs 52–54). Sculpture of tergal discs variable, often with smooth and shiny area, or punctation coarser. 9
9 Light hairs on dorsal parts of mesosoma, terga (Fig. 51) and sterna yellowish white in fresh specimens (although snow white in worn specimens). T2-T4 with comparatively short apical fringes of hairs, long only laterally, fringes medially not longer than apical tergal margin. Erect hairs on disc of T4 usually predominantly dark (lighter in worn specimens). Facial vestiture often entirely dark (but frequently light medially in second generation, especially in southern and western parts of the Alps) Andrena bicolor
Light hairs on dorsal parts of mesosoma, terga and sterna snow white, without yellowish hue even in very fresh specimens (Figs 39, 50). T2-T4 with comparatively long, snow white apical fringes of hairs, fringes medially clearly longer than apical tergal margin (Fig. 50). Disc of T4 at most with a few isolated, erect dark hairs. Face always with some grey-white hairs medially, at least between and around antennal sockets (Figs 39, 42). 10
10 Maxillary palps comparatively longer, segments 4 and 5 at least four times as long as maximal apical width (Fig. 45). Body length 8.5mm. First recurrent vein entering second submarginal cell comparatively closer to first submarginal crossvein (Fig. 49), second submarginal cell broader than long. Head comparatively long with protruding clypeus (Fig. 43). Mesonotum matt, without extended silk-shiny area medially (as on Fig. 46). Andrena allosa
Maxillary palps comparatively shorter, segments 4 and 5 at most three times as long as maximal apical width (Fig. 44). Slightly smaller, body length on average 7-8mm. First recurrent vein entering second submarginal cell comparatively closer to middle of cell, second submarginal cell often subquadrate (Fig. 48). Head comparatively short with little protruding clypeus (Fig. 42). Sculpture of mesonotum variable, in first generation matt (Fig. 46), in second generation with extended silk-shiny area medially (Fig. 47). Andrena amieti sp. n.
(Only two males of A. allosa were examined and the characters mentioned to distinguish A. allosa from A. amieti sp. n. are tentative.)

Additional taxonomic notes on selected Andrena (Euandrena) species

Andrena croatica Friese, 1887, stat. rev.

Suppl. material 3: Figs S8, S9 Andrena croatica Friese 1887: 85, ♀, ♂, “Prope Orechovicam non procul ab urbe St. Veithii, mense Martio” [Near Orehovica (nowadays a district of Rijeka), not far from the city of St. Veith (Rijeka, Croatia) month of March]. Syntypes ♀♀♂♂ (NMBE, OLML, ZMHB).

Notes

We examined two possible syntype females of Andrena croatica, which perfectly agree with the original description. No specimen was available for genetic study. These two females are markedly divergent from A. bicolor, and thus A. croatica, which is currently considered as a synonym of A. bicolor (Gusenleitner and Schwarz 2002), probably represents a valid species (stat. rev.). Striking divergent features in the female sex include the white scopa, the fine, regular punctation of the terga (Suppl. material 3: Fig. S9), the mesonotal vestiture comprising intermixed grey-brown hairs and numerous, comparatively long, dark hairs, and the numerous dark hairs on the propodeum. This species also has long mouthparts approaching the condition observed in A. allosa (Suppl. material 3: Fig. S8). The male is unknown to us.

Examined material

2 possible syntype females from Fiume [Croatia, Rijeka] and an additional female from Croatia (Suppl. material 1).

Andrena pileata Warncke, 1975, stat. n.

Suppl. material 3: Figs S10–S12

Andrena allosa pileata Warncke 1975b: 85, ♀ nec ♂, “Chelmos, Peloponnes” [Chelmos mountains, Greece]. Holotype ♀, OLML.

Notes

Warncke (1975b) described Andrena allosa pileata from the Chelmos mountains in Greece. The type series of A. allosa pileata is mixed and includes two distinct species. The first species, represented by the holotype, is briefly redescribed here. No specimen was available for genetic study. This first species appears clearly distinct from A. allosa and A. amieti sp. n., and we treat it as a valid species, A. pileata Warncke stat. n.

Description

Female: Body length 9mm. Clypeus shorter than in A. allosa, as in A. bicolor, without flattened apical area, regularly convex, densely punctate, interspaces weakly shagreened even apically (Suppl. material 3: Fig. S10). Malar space as in A. bicolor, shorter than in A. allosa. Fovea as in A. bicolor or A. allosa, markedly narrower inferiorly than superiorly. Mesonotum nearly entirely matt, silk-shiny only medially, comparatively densely punctate, except medially (there interspaces up to 3–4 puncture diameters). Terga entirely shagreened, finely, regularly punctate even medially on discs, interspaces 3 puncture diameters (Suppl. material 3: Fig. S12). Apical tergal margins weakly impressed apically, more so than in A. allosa. The sculpture of the terga is similar to that of A. ruficrus and unlike A. amieti sp. n. or A. allosa. Vestiture predominantly brown (both specimens are not fresh), medially on face grey-brown, laterally dark; vestiture on mesonotum orange-brown, made of comparatively short, strongly plumose hairs, with only few, very short dark hairs (Suppl. material 3: Fig. S11). Hairs on lateral and ventral sides of mesosoma grey-brown. Scopa brown-orange. Hairs on metasoma comparatively short, metasomal terga without loose apical fringes of white hairs, at most with isolated, short hairs (Suppl. material 3: Fig. S12).

Male: unknown; the male paratype is likely not conspecific and probably belong to Andrena sp1.

Examined material

Holotype female of A. pileata; one additional female from the Chelmos mountains, Greece (Suppl. material 1).

Andrena sp1

Suppl. material 3: Figs S13–S15

Andrena allosa pileata Warncke, 1975b, partim

Notes

The second species included in the type series of Andrena allosa pileata is referred to as Andrena sp1. This species shares similarities with both A. amieti sp. n. and A. allosa. It is superficially similar to specimens of A. amieti sp. n. from Southern Italy (Suppl. material 2: Fig. S1), which have entirely grey vestiture even in the female sex. Two specimens were available for genetic study (specimens 961 and 1252; Figs 1, 2). These two specimens were closely related to A. allosa and separated from that species by a genetic distance of 1.82%. We do not consider this species to be conspecific with A. amieti sp. n. for three reasons: first, genetic distances to A. amieti sp. n. (4.27%) were higher than to A. allosa; second, based on the available material, this species appears to have only one generation per year unlike A. amieti sp. n.; third, there are slight, but consistent morphological differences compared to A. amieti sp. n. Andrena sp1 is characterized as follows:

Description

Female: Body size, 7–8mm. Clypeus shorter than in A. allosa, as in A. bicolor, without flattened apical area, in most specimens, apical area with irregular, longitudinal concavity (not clearly visible on Suppl. material 3: Fig. S15). Clypeus less densely punctate than in A. pileata stat. n., interspaces shinier, very weakly shagreened or completely shiny medially (Suppl. material 3: Fig. S15). Malar space as in A. bicolor, shorter than in A. allosa. Fovea as in A. bicolor or A. allosa, markedly narrower inferiorly than superiorly. Mesonotum nearly entirely matt, finely sculptured, silk-shiny medially, comparatively sparsely punctate, interspaces medially up to 5 puncture diameters. Terga weakly shagreened, entirely silk-shiny, shinier than in A. allosa, irregularly, very sparsely punctate to nearly impunctate, apical margin weakly impressed apically (Suppl. material 3: Fig. S14) although more so than in A. allosa. Vestiture predominantly white-grey, without yellowish or orange hue (Suppl. material 3: Figs S13, S14). Face darker than in A. pileata stat. n., grey only around antennal sockets. Vestiture on mesonotum predominantly grey, made of comparatively long, weakly plumose grey hairs and numerous, comparatively long dark hairs (Suppl. material 3: Fig. S13) (as in A. amieti sp. n. but unlike A. pileata stat. n.). Hairs on lateral side of mesosoma dark, on ventral side grey. Scopa white. Hairs on metasoma comparatively long, nearly entirely whitish grey, except dark on disc of T4, and on T5 and T6. The most distinctive features are the white scopa, the grey mesonotal vestiture with long, dark hairs, and the nearly impunctate terga.

Male: No difference was found between the three males examined and the first generation of A. amieti sp. n.

Examined material

3 males and 10 females from the Chelmos Mountains, the Mount Olympus and the Mount Tymfristos, Greece (Suppl. material 1).

Andrena sp2

Suppl. material 4: Figs S16, S17

Notes

One female specimen from Lesbos included in our genetic study (specimen 1254) is morphologically distinct from both A. pileata stat. n. and Andrena sp1 and is referred to as Andrena sp2. In the female sex, it is similar to Andrena sp1, except that the clypeus lacks the longitudinal impression (Suppl. material 4: Fig. S17) observed in Andrena sp1, the mesonotum is more densely punctate, the terga are more shiny (disc of T1 and T2 nearly completely shiny; Suppl. material 4: Fig. S16) and more clearly punctate. The mesonotum has grey-brown vestiture, the hairs are weakly plumose, there are a few long dark hairs on the mesonotum and numerous dark hairs on the propodeum. The scopa is yellowish white, in contrast to Andrena sp1. Male unknown.

Examined material

2 females from two localities on the Island of Lesbos, Greece (Suppl. material 1).

Andrena sp3

Suppl. material 4: Figs S18, S19

Notes

Specimen 928 in our genetic dataset, a female collected in Greece, appears to be morphologically distinct from A. bicolor and from any taxa mentioned above. In addition, we have examined a series of females from the Island of Lesbos, as well as one female from Northern Italy, all of which putatively belong to Andrena sp3. This species is presumably widely distributed in southeastern Europe. In the female sex, Andrena sp3 is similar to some forms of A. bicolor. The vestiture is nearly entirely brown-orange, including on all sides of the mesosoma, with only a few dark hairs laterally on face (Suppl. material 4: Fig. S19), on the disc of T4, as well as numerous dark hairs on T5 and T6. The mesonotum is conspicuously shiny with coarse and dense punctation (Suppl. material 4: Fig. S18). In its vestiture this species is highly similar to A. rufula, from which it can easily be separated by the shiny sculpture of the mesonotum. Male unknown to us.

Examined material

5 females from the Island of Lesbos, Greece; 1 female from the Peloponnese, Greece; one female from northwestern Italy (Suppl. material 1).

Discussion

Species delimitation in the bicolor-group

The taxonomy of the bicolor-group in the Alps has long remained controversial. In light of our results, there appears to be little doubt that four well-separated biological species co-occur in the Alps, for the following reasons. First, DNA barcodes are diagnostic for each of these species in spite of the observed paraphyly in Andrena amieti sp. n. in mitochondrial trees (Fig. 1); these distinct barcodes were concordant with morphology in all cases, suggesting no mitochondrial introgression. Second, all four species occur in sympatry and maintain distinct morphology. Andrena montana differs from the other three species in numerous sculptural characters in addition to conspicuous characters in vestiture colour; similarly, this species was only distantly related to the other three species (Figs 1, 2). Andrena allosa differs from A. bicolor and A. amieti sp. n. in subtle but constant sculptural characters in the female, while A. amieti sp. n. and A. bicolor can be distinguished by constant differences in vestiture colour. Third, A. allosa and A. montana have a single generation per year while A. bicolor and A. amieti sp. n. have two generations per year. Fourth, A. allosa differs from A. bicolor and A. amieti sp. n. in its pollen host choice: the second generation of the latter two species have a pronounced affinity for Campanulaceae, a pollen type that is completely absent from the pollen diet of A. allosa.

These clear differences in morphology and life-history traits are in contradiction with the results of our analyses of the nuclear gene, where one specimen of A. allosa (1285; Fig. 2) was not separated from some specimens of A. amieti sp. n. A. amieti sp. n. and A. allosa have probably diverged recently, and the observed pattern in the nuclear gene sequences is likely due to either lack of variation in the selected marker, incomplete lineage sorting or rare instances of horizontal gene flow, all of which are expected to occur in complexes of closely related species.

While our study solves a long controversy in the systematics of central European bees, it also raises additional questions that will require future work. In both A. bicolor and A. amieti sp. n., two distinct clusters of sequences were found in phylogenetic analyses of COI. In both cases these two clusters were found in sympatry and they conflict with current species concepts. The same two clades occurring in sympatry within A. bicolor were previously reported (Schmidt et al. 2015). Based on available sequences on BOLD these two clades appear to be widespread in Europe. In contrast to the situation in A. bicolor where both clades are sister groups, the two clusters found in A. amieti sp. n. form a paraphyletic unit with respect to a distinct biological species occurring in sympatry, A. allosa.

Two hypotheses can be formulated to explain these discrepancies between mitochondrial gene trees and species trees (Mutanen et al. 2016). First, these clusters may point to additional cryptic diversity. For example, species paraphyly in the pierid butterfly Leptidea reali Reissinger, 1989 has recently been demonstrated to be due to the presence of an additional cryptic species, L. juvernica Williams, 1946 (Dincă et al. 2011). Second, these clusters may merely represent deep intraspecific divergences within one biological species, without reproductive isolation between mitochondrial lineages. Such a situation has been suggested for two divergent clades within the Common Redstart Phoenicurus phoenicurus (Linnaeus, 1758) (Hogner et al. 2012) and for the paper wasp Polistes dominula (Christ, 1791) (Neumeyer et al. 2014, Schmid-Egger et al. 2017). To discriminate between these two alternative hypotheses for the bicolor-group, future research is needed, ideally using genomic-scale, nuclear markers.

Our study reveals that there are probably five additional species in the bicolor-group in southeastern Europe: A. croatica stat. rev., A. pileata stat. n., Andrena sp1, Andrena sp2 and Andrena sp3. How these species relate to A. allosa, A. amieti sp. n., or to additional species from Eastern Europe (A. asperula Osytshnjuk, 1977 from Ukraine) or from the East Palearctic (e.g., A. capilosella Osytshnjuk, 1986, A. khosrovi Osytshnjuk, 1993), remains unclear and requires future investigation. We strongly recommend the use of genetic analyses to clarify the taxonomy of these challenging taxa.

Delimitation of Euandrena and Ptilandrena

Our study suggests that Andrena fulvata and the closely related A. angustior, currently placed in the subgenus Ptilandrena, may in fact belong to the subgenus Euandrena. These two species differ from other species of Euandrena in the long mandibles of the male and the presence of an elevated carina on the pronotum in both sexes. They differ from other members of Ptilandrena such as A. vetula Lepeletier, 1841, in the narrow facial fovea of the female, a character shared with species of Euandrena. Future studies should examine the phylogenetic placement of other members of the morphologically heterogeneous subgenus Ptilandrena.

Pollen host preferences

The high proportion of Crocus pollen in the larval diet of Andrena allosa was unexpected. Crocus albiflorus, which is the only native Crocus species in the Central Alps, starts to bloom immediately after the snowmelt (Fig. 9). The early flight period of A. allosa in combination with its restricted host plant choice might explain why this bee species has so rarely been found in the past. We hypothesize that A. allosa will turn out to be more common than hitherto assumed when flowering Crocus stands will be systematically checked for the presence of this species.

Surprisingly, the spring and summer generations of the bivoltine A. amieti sp. n. distinctly differ in their pollen host selection in that the females of the latter exhibit a strong preference for the pollen of Campanulaceae. This preference is also suggested by our observations that males of the summer generation patrolled Campanula flowers in their search for females and that summer generation males and females visited Geranium sylvaticum (Geraniaceae). The latter observation is in agreement with the finding that Geranium was the second most important pollen host of the summer generation (Tab. 3). Geranium flowers often serve as a source of nectar or a “rendezvous” point for oligolectic bees specialized on Campanula, such as Andrena curvungula Thomson, 1870, A. pandellei Pérez, 1895 or Chelostoma rapunculi (Lepeletier, 1841), which might possibly be due to similar colour or olfactory cues exerted by the flowers of the two unrelated taxa. Attraction to similar cues might also explain why Geranium is regularly visited by A. amieti sp. n. Interestingly, a preference for Campanulaceae was also reported for the summer generation of the bivoltine A. bicolor, which is closely related to A. amieti sp. n. (Stöckhert in Schmiedeknecht 1930, Müller et al. 1997, Peeters et al. 2012, Scheuchl and Willner 2016). Differing pollen host choice by spring and summer generations of the same species as in A. amieti sp. n. and A. bicolor suggest that the two generations have evolved their own phenotypes as a result of different selection regimes in spring and summer (Simpson et al. 2011). In a recent study, however, Milet-Pinheiro et al. (2016) did not detect seasonal polyphenism in A. bicolor and found instead that females of the first and the second generation have the same innate flower search image in spite of the different spectrum of pollen hosts they exploit.

Conclusion

Genetic analyses, in combination with morphological analysis, were decisive in solving an important controversy in the taxonomy of European wild bees. However, our study also raised new questions, in particular the intriguing case of paraphyly observed in mitochondrial gene trees in Andrena amieti sp. n. This new species appears to be the tip of the iceberg of cryptic diversity in southern Europe. Time and funds should be dedicated to taxonomic research on wild bees in species-rich Mediterranean ecosystems.

Acknowledgements

We thank the institutions and people who generously contributed specimens for this study, in particular: H. Baur (Naturhistorisches Museum Bern), R. Eastwood (ETH Zürich), A. Freitag (Musée de Zoologie Lausanne), S. Liersch (Bündner Naturmuseum Chur), J. Litman (Muséum d’Histoire Naturelle Neuchâtel), S. Schmidt (Zoologische Staatssammlung München), F. Gusenleitner and M. Schwarz (Biologiezentrum Linz), Claire Villemant (Muséum Nationale d’Histoire Naturelle Paris), Matthieu Aubert, Eric Dufrêne, Karl Hirt, Johannes Voith, Hans Schwenninger, Markus Bur, Georg Artmann, Felix Amiet, Claudio Sedivy, Dimitri Bénon, Sophie Giriens, Mike Herrmann, Jessica Litman, Erwin Steinmann, Hansueli Tinner, Rainer Neumeyer, Jelle Devallez, Theodora Petanidou and Christian Schmid-Egger. Particular thanks to P. Andreas W. Ebmer who generously provided specimens from mountainous regions of Greece, Fritz Gusenleitner, Martin Schwarz and Maximilian Schwarz for facilitating our stays in Linz, Erwin Scheuchl for numerous discussions on the taxonomy of Andrena bees, Claudio Sedivy and Sophie Giriens for contributing field observations on A. allosa, Jessica Litman for numerous joint excursions with the first author in the Alps. Collecting permits were delivered by the canton of Bern to Jessica Litman, by all Swiss cantons for people participating in the actualization of the red list of the Swiss bees, and by the Mercantour National Parc (permits to David Genoud, Christophe Praz, Matthieu Aubert). Databasing was greatly facilitated by the staff of InfoFauna (Neuchâtel), in particular François Claude, Sofia Wyler and Pascal Tschudin. The map presented on Fig. 3 was kindly prepared by Fabien Fivaz (InfoFauna, Neuchâtel). Dimitri Bénon produced the diagrams presented on Figs 4–6. Pictures were taken with the Keyence digital microscope jointly owned by the Museum of Natural History of Neuchâtel and InfoFauna. The molecular work was jointly funded by the University of Neuchâtel and InfoFauna; molecular work and databasing of French specimens were performed within the “inventaire du patrimoine naturel (inpn.mnhn.fr)” project with funding from UMS PatriNat (AFB, CNRS, MNHN); and microscopic analysis of female pollen loads was funded by the Tierschutzverein Glarus. We thank Jessica Litman, Erwin Scheuchl, Rainer Neumeyer and Gidi Pisanty for numerous constructive comments on an earlier version of this manuscript.

References

  • Abràmoff MD, Magalhães PJ, Ram SJ (2004) Image Processing with ImageJ. Biophotonics International 11: 36–42.
  • Amiet F, Hermann M, Müller A, Neumeyer R (2010) Apidae 6, Andrena, Melliturga, Panurginus, Panurgus. Fauna Helvetica 26, CSCF & SEG, Neuchâtel, 316 pp.
  • Artmann-Graf G (2017) Zwei Erstfunde von Andrena rufula Schmiedeknecht, 1883 in der Schweiz (Hymenoptera, Apoidea). Ampulex 9: 40–41.
  • Chakrabarty P, Warren M, Page L, Baldwin C (2013) GenSeq: An updated nomenclature and ranking for genetic sequences from type and non-type sources. ZooKeys 346: 29–41. https://doi.org/10.3897/zookeys.346.5753
  • Dincă V, Lukhtanov VA, Talavera G, Vila R (2011) Unexpected layers of cryptic diversity in wood white Leptidea butterflies. Nature Communications 2: 324. https://doi.org/10.1038/ncomms1329
  • Else GR, Edwards M (2018) Handbook of the Bees of the British Isles. Ray Society Monographs, The Ray Society, c/o The Natural History Museum, London, 775 pp.
  • Friese H (1887) Species aliquot novae generis Andrena FABR. Természetrajzi Füzetek 11: 21–26.
  • Gusenleitner F, Schwarz M (2002) Weltweite Checkliste der Bienengattung Andrena mit Bemerkungen und Ergänzungen zu paläarktischen Arten (Hymenoptera, Apidae, Andreninae, Andrena). Entomofauna, Supplement 10, 1280 pp.
  • 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. Proceedings of the National Academy of Sciences 101: 14812–14817. https://doi.org/10.1073/pnas.0406166101
  • Hogner S, Laskemoen T, Lifjeld JT, Porkert J, Kleven O, Albayrak T, Kabasakal B, Johnsen A (2012) Deep sympatric mitochondrial divergence without reproductive isolation in the common redstart Phoenicurus phoenicurus. Ecology and Evolution 2: 2974–2988. https://doi.org/10.1002/ece3.398
  • Katoh K, Standley DM (2013) MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution 30: 772–780. https://doi.org/10.1093/molbev/mst010
  • Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S, Cooper A, Markowitz S, Duran C, Thierer T, Ashton B, Meintjes P, Drummond A (2012) Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28: 1647–1649. https://doi.org/10.1093/bioinformatics/bts199
  • Klopfstein S, Kropf C, Baur H (2016) Wolbachia endosymbionts distort DNA barcoding in the parasitoid wasp genus Diplazon (Hymenoptera: Ichneumonidae). Zoological Journal of the Linnean Society 177: 541–557. https://doi.org/10.1111/zoj.12380
  • Michener CD (2007) The Bees of the World, second edition. Johns Hopkins University Press, Baltimore, Maryland, 953 pp.
  • Milet-Pinheiro P, Herz K, Dötterl S, Ayasse M (2016) Host choice in a bivoltine bee: how sensory constraints shape innate foraging behaviors. BMC Ecology 16: 20. https://doi.org/10.1186/s12898-016-0074-z
  • Miller MA, Pfeiffer W, Schwartz T (2010) Creating the CIPRES science gateway for inference of large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop (GCE), New Orleans LA, USA.
  • Müller A (2018) Pollen host selection by predominantly alpine bee species of the genera Andrena, Panurginus, Dufourea, Megachile, Hoplitis and Osmia (Hymenoptera, Apoidea). Alpine Entomology 2: 101–113. https://doi.org/10.3897/alpento.2.29250
  • Müller A, Krebs A, Amiet F (1997) Bienen - Mitteleuropäische Gattungen, Lebensweise, Beobachtung. Naturbuch Verlag, Augsburg, 384 pp.
  • Müller A, Kuhlmann M (2008) Pollen hosts of western palaearctic bees of the genus Colletes (Colletidae) – the Asteraceae paradox. Biological Journal of the Linnean Society: 95, 719–733. https://doi.org/10.1111/j.1095-8312.2008.01113.x
  • Mutanen M, Kivelä SM, Vos RA, Doorenweerd C, Ratnasingham S, Hausmann A, Huemer P, Dincă V, van Nieukerken EJ, Lopez-Vaamonde C, Vila R, Aarvik L, Decaëns T, Efetov KA, Hebert PD, Johnsen A, Karsholt O, Pentinsaari M, Rougerie R, Segerer A, Tarmann G, Zahiri R, Godfray HC (2016) Species-Level Para- and Polyphyly in DNA Barcode Gene Trees: Strong Operational Bias in European Lepidoptera. Systematic Biology 65: 1024–1040. https://doi.org/10.1093/sysbio/syw044
  • Neumeyer R, Baur H, Guex GD, Praz C (2014) A new species of the paper wasp genus Polistes (Hymenoptera, Vespidae, Polistinae) in Europe revealed by morphometrics and molecular analyses. ZooKeys (400): 67–118. https://doi.org/10.3897/zookeys.400.6611
  • Peeters TMJ, Nieuwenhuijsen H, Smit J, van der Meer F, Raemakers IP, Heitmans WRB van Achterberg C, Kwak M, Loonstra AJ, de Rond J, Roos M, Reemer M (2012) De Nederlandse Bijen (Hymenoptera: Apidae s.l.). Natuur van Nederland 11, Naturalis Biodiversity Center and European Invertebrate Survey, Leiden, 544 pp.
  • Scheuchl E, Willner W (2016) Taschenlexikon der Wildbienen Mitteleuropas. Quelle & Meyer, Wiebelsheim, 920 pp.
  • Schmid-Egger C (2012) Die taxonomischen Situation der Gattung Andrena (Apidae) in Mitteleuropa – Kommentar zu: Amiet et al. (2010): Fauna Helvetica, Apidae 6. Ampulex 4: 55–60.
  • Schmid-Egger C, Scheuchl E (1997) Illustrierte Bestimmungstabellen der Wildbienen Deutschlands und Österreichs unter Berücksichtigung der Arten der Schweiz, Band III, Andrenidae. Eigenverlag, Velden/Vils, 180 pp.
  • Schmid-Egger C, van Achterberg K, Neumeyer R, Morinière J, Schmidt S (2017) Revision of the West Palaearctic Polistes Latreille, with the descriptions of two species - an integrative approach using morphology and DNA barcodes (Hymenoptera, Vespidae). ZooKeys (713): 53–112. https://doi.org/10.3897/zookeys.713.11335
  • Schmidt S, Schmid-Egger C, Morinière J, Haszprunar G, Hebert PD (2015) DNA barcoding largely supports 250 years of classical taxonomy: identifications for Central European bees (Hymenoptera, Apoidea partim). Molecular Ecology Resources 15: 985–1000. https://doi.org/10.1111/1755-0998.12363
  • Schmiedeknecht O (1930) Die Hymenopteren Nord- und Mitteleuropas. Gustav Fischer, Jena, 1062 pp.
  • Swofford DL (2002) PAUP 4.0 Phylogenetic Analysis using Parsimony (and other Methods). Version 4. Sinauer Associates, Sunderland.
  • Trunz V, Packer L, Vieu J, Arrigo N, Praz CJ (2016) Comprehensive phylogeny, biogeography and new classification of the diverse bee tribe Megachilini: Can we use DNA barcodes in phylogenies of large genera? Molecular Phylogenetics and Evolution 103: 245–259. https://doi.org/10.1016/j.ympev.2016.07.004
  • Warncke K (1973) Beitrag zur Bienenfauna Mazedoniens (Colletidae, Andrenidae und Melittidae/Apoidea). Mitteilungen aus dem Zoologischen Museum in Berlin 49: 13–36.
  • Warncke K (1975a) Die Bienengattung Andrena F., in Iberien (Hym. Apidae). Teil A. Eos 49: 293–314.
  • Warncke K (1975b) Die Sandbienen der Türkei (Hymenoptera. Apoidea, Andrena). Mitteilungen der Münchner Entomologischen Gesellschaft 65: 29–102
  • Westrich P (2006) Ein weiterer Beleg für den Bivoltinismus und das Wirt-Parasit-Verhältnis von Andrena bicolor (FABRICIUS 1804) und Nomada fabriciana (KIRBY 1802) (Hym. Apidae). Linzer biologische Beiträge 38: 919–923.
  • Westrich P (2018) Die Wildbienen Deutschlands. Ulmer, Stuttgart, 824 pp.