Research Article |
Corresponding author: Christophe Praz ( christophe.praz@unine.ch ) Academic editor: Jack Neff
© 2022 Christophe Praz, David Genoud, Killian Vaucher, Dimitri Bénon, Joseph Monks, Thomas J. Wood.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Praz C, Genoud D, Vaucher K, Bénon D, Monks J, Wood TJ (2022) Unexpected levels of cryptic diversity in European bees of the genus Andrena subgenus Taeniandrena (Hymenoptera, Andrenidae): implications for conservation. Journal of Hymenoptera Research 91: 375-428. https://doi.org/10.3897/jhr.91.82761
|
Using a combination of DNA barcodes and morphology, we examine species boundaries in bees of the genus Andrena subgenus Taeniandrena in Europe. First, we solve the long controversy surrounding the status of Andrena ovatula (Kirby, 1802) and A. albofasciata Thomson, 1870, proposed to represent distinct species nearly 100 years ago, but mostly treated as conspecific in recent studies. Our results unambiguously support the presence of two taxa that are often found in sympatry: the first taxon, referred to as A. ovatula, is present in Northern Europe but also in Southern Europe along the Mediterranean coast; the second taxon is referred to as A. afzeliella (Kirby, 1802), stat. rev., with A. albofasciata considered to be a junior synonym (syn. nov.), and is widely distributed in Europe. Second, we show that another widely distributed species has hitherto been overlooked in Europe: A. ovata Schenck, 1853, stat. rev. Third, we demonstrate that two taxa currently treated as subspecies should be given specific rank due to significant morphological and genetic differences: A. croceiventris Morawitz, 1871, stat. rev., so far treated as a subspecies of A. similis Smith, 1849, and A. vocifera Warncke, 1975, stat. nov., so far treated as a subspecies of A. gelriae van der Vecht, 1927. Both A. croceiventris and A. vocifera have particularly restricted ranges in Europe, being known only from central to southern Italy and Sicily, and continental France, respectively. Fourth, we describe a new species from Sardinia and Corsica, A. antonellae sp. nov. Lastly, the following new synonymies are proposed: A. similis, A. ocreata cyprisina Warncke, 1975 and A. similis caraimica Osytshnjuk, 1994 are placed in synonymy with A. russula Lepeletier, 1841 (syn. nov.); A. fuscata (Kirby, 1802), A. canescens Schenck, 1853 and A. pseudovatula Alfken, 1926 are placed in synonymy with A. afzeliella (syn. nov.). Lectotypes are designated for A. afzeliella, A. fuscata (Kirby, 1802), A. ovata and A. wilkella (Kirby, 1802). Our results suggest a particularly fast diversification in this group of bees, leading to the presence of numerous species exhibiting particularly restricted geographic ranges. We discuss the implications for conservation of this astonishing cryptic diversity in European bees.
Cryptic species, DNA barcoding, speciation, species delimitation
Obtaining accurate taxonomic and distribution information on bees is crucial for the conservation and monitoring of native pollinators. However, new species are regularly uncovered even in regions with a long tradition in taxonomy such as northern and central Europe, following changes in taxonomy (e. g.,
The genus Andrena Fabricius, 1775 is exemplary in this context, as it is highly diverse and includes numerous species showing a restricted geographic range. In addition, morphological identifications are challenging, often relying entirely on subtle differences in the colour of the vestiture or in the sculpture. Accordingly, new species of Andrena are being uncovered at an accelerated pace (e.g.,
In the present study, we further examine species boundaries in the subgenus Taeniandrena using a combination of DNA barcodes and morphology. Our investigations lead to the recognition of three taxa hitherto considered to be subspecies or synonyms as valid species, to the description of a new species, and to a resolution of a longstanding controversy in central European bee taxonomy: the status of Andrena ovatula (Kirby, 1802) and A. albofasciata Thomson, 1870, alternately considered as two valid species (e.g.,
We used mitochondrial genetic data (cytochrome oxidase I, hereafter COI;
We aimed at obtaining at least one full-length mitochondrial barcode (658 base pairs, hereafter bp) for as many species as possible in Europe. We amplified two overlapping fragments of COI using the two primer pairs LepF/LepR (658 bp) and UAE3/LepR (409 bp) (see
For specimens older than 15 years, we used the specific primers COX-Taeniandr-F, COX-Taeniandr-R1 and COX-Taeniandr-R2 (see primer sequences in
Two barcodes from specimens collected in Israel were kindly provided by Gideon Pisanty. A selection of DNA barcodes available on BOLD, mostly from
Chromatograms were trimmed and assembled in Geneious 6.0.6 (
For final phylogenetic analyses, we included a subset of all specimens, favouring long sequences and maximising both geographic representation and haplotype diversity within each species. Sequences originating near type localities were included whenever possible. Bayesian phylogenetic analyses were performed in BEAST ver. 1.10.4 (Suchard et al. 2018). The aligned matrix was divided into three partitions corresponding to the three codon positions. A HKY + G model was applied to each partition (more complex models resulted in poor estimation of some parameters); all parameters were unlinked across partitions; the clock model was set to “strict clock” and the tree prior was set to “speciation: Yule process”. One sequence of Andrena lathyri Alfken, 1899 was included in the analysis, and a clade including all other terminals was constrained to be monophyletic to ensure proper rooting of the trees. The analysis was run for 10 million generations, sampling trees and parameters every 1000 generations. The software Tracer ver. 1.7.1 (
Morphological terminology follows
Morphological identifications were performed alongside genetic analyses; specimens identified using DNA sequences were used as reference for morphological examination, and initial genetic results were integrated to revise our morphological identifications. When not specified, the type material was examined by TJW. We do not provide a full list of synonyms for each species, but only list those discussed in the text or new synonyms based on our examination of the type material. The list of examined specimens (Suppl. material
CSEC Private collection of Christian Schmid-Egger, Berlin, Germany;
DGC Private collection of David Genoud, Ambazac, France;
KHC Private collection of Karl Hirt, Menziken, Switzerland;
MHC Private collection of Mike Herrmann, Konstanz, Germany;
MIB Collection of the ZooPlantLab of the University of Milano-Bicocca, Milano, Italy;
OUMNH University Museum of Natural History, Oxford, UK;
PRUN Research collection of Christophe Praz, University of Neuchatel, Switzerland;
SIZK Scientific Collections of the Schmalhausen Institute of Zoology, Kiev, Ukraine;
SMFD Forschungsinstitut und Naturmuseum Senckenberg, Frankfurt am Main, Germany;
STC Private collection of Stefan Tischendorf, Darmstadt, Germany;
TJWC Private collection of Thomas J. Wood, Mons, Belgium;
ZMHB Museum für Naturkunde, Berlin, Germany;
ZML Zoological Museum, Lund, Sweden.
Of the 142 barcoded specimens, partial or full DNA barcodes were obtained for 121 (Suppl. material
Phylogenetic analyses (Fig.
Maximum clade credibility tree found in Bayesian analyses of sequence data of the mitochondrial gene COI, showing the phylogenetic relationships among the different species of Andrena subgenus Taeniandrena investigated in the present study. Numbers above branches represent posterior probabilities (values < 0.5 are omitted). Taxa mentioned in the text are indicated by arrows or by vertical bars.
Specimens of Andrena similis Smith, 1849 were included in a well-supported clade (PP 1.0), which comprised specimens from the Iberian Peninsula, Northern Europe, France, Italy, Cyprus, Morocco, as well as one specimen from Algeria (1533 in Fig.
Andrena croceiventris Morawitz, 1871, treated so far as a subspecies of A. russula (
Our molecular analyses also highlight an additional hitherto overlooked species of Taeniandrena, Andrena ovata Schenck, 1853 (stat. rev.). Specimens of this taxon formed a well-supported clade including one specimen from Greece (2265), four specimens from France and one specimen from southern Switzerland. In addition, one full-length sequence from Spain, four short sequences from Italy (numbers 2371, 2376, 2375, 2377; Suppl. material
One moderately well-supported clade (PP 0.95) included species associated with Andrena gelriae van der Vecht, 1927 and A. intermedia Thomson, 1870 (including A. producta Warncke, 1973, A. gredana Warncke, 1975, A. levante Wood & Praz, 2021 and A. contracta Wood, 2022). This clade is hereafter referred to as the gelriae-clade. Genetic distances within this clade were considerably lower than among the other species of Taeniandrena: on average 1.59% between A. gelriae and A. intermedia (excluding two divergent specimens of A. intermedia from Spain; see below), 3.54% between A. gelriae and A. gredana, 1.98% between A. gelriae and A. levante, 2.35% between A. gelriae and A. producta. Andrena vocifera Warncke, 1975, so far treated as a subspecies of A. gelriae, was separated from the latter species by average distances of 1.21% and is treated here as a distinct species (stat. nov.). Within A. gelriae, one male specimen from southern Italy (1427) was sister to all other specimens of A. gelriae, but support for this relationship was low (PP less than 0.5); this specimen was separated from other specimens of A. gelriae by average distances of 1.12%. Lastly, specimens of A. intermedia did not form a monophyletic clade: central European specimens and one specimen from southern Italy formed a well-supported clade (PP 1.0, maximal genetic distances within this clade 1.12%); two male specimens from Spain morphologically highly similar to central European specimens of A. intermedia (referred to as A. cf. intermedia in Fig.
All specimens initially identified as A. gelriae present in Swiss entomological collections were examined and most females were analysed genetically. The vast majority of occurrences (see map in
Melitta afzeliella
Kirby, 1802: 169, ♀, “Barhamiae” [Barham, Suffolk, England]. Lectotype ♀ (
Melitta fuscata
Kirby, 1802: 167, ♀, “Barhamiae” [Barham, Suffolk, England], Syn. nov. Lectotype ♀ (
Andrena canescens
Schenck, 1853: 140, ♂, “Schierstein” [Schierstein, Wiesbaden, Germany], Syn. nov. Holotype ♂ (
Andrena albofasciata
Thomson, 1870: 154, ♀ ♂, “Sällsynt på Skånes sandmarker” [Rare on the sandy grounds of Scania; Sweden], Syn. nov. Lectotype ♀, by designation of
Andrena afzeliella var. heliopolis Friese, 1914: 227, ♀ [Egypt]. Holotype ♀ (ZMHB)
Andrena pseudovatula Alfken, 1926: 107, ♂ [Egypt]. Syn. nov. Syntype ♂ (SMFD)
Type material: Lectotype ♀ of Melitta afzeliella Kirby, 1802 (Figs
Lectotype
♀ of Melitta fuscata Kirby, 1802, by present designation, a female in good condition (
Holotype
♂ of Andrena canescens Schenck, 1853, a male in good condition (
Holotype of Andrena afzeliella var. heliopolis Friese, 1914. Egypt: ♀, no locality information (ZMHB).
Syntype of Andrena pseudovatula Alfken, 1926. Egypt: ♂, no locality information, A. Andres leg. (SMFD).
Other material: Specimens from the UK, France (including Corsica), Switzerland, Italy (including Sardinia), the Netherlands, Germany, and Israel have been barcoded. Additional sequences are available on the BOLD website for specimens from Austria, Croatia and Georgia. Full details of examined material can be found in Suppl. material
Widespread and abundant in Europe (Fig.
Analysis of 37 pollen loads from 19 localities suggests polylecty with a strong preference for Fabaceae, with 91.7% of pollen collected from this family. Rosaceae, Plantaginaceae, Boraginaceae, Cornaceae, Ranunculaceae, Brassicaceae, Dipsacaceae, and Asteraceae were detected in decreasing order of abundance (TJW, unpublished data). Use of Fabaceae pollen is more pronounced in the summer generation.
Andrena afzeliella is clearly bivoltine, with the peak of the first generation in May and early June, and that of the second generation in July and August (Fig.
The status of this species, in particular its distinctiveness from Andrena ovatula has been a controversial topic, and consequently the identity of A. afzeliella, A. albofasciata and A. ovatula in the literature is confusing (
The distinctiveness of A. ovatula and A. albofasciata was then advocated by J. van der Smissen, who wrote that she could unambiguously separate these two taxa in northern Germany based on Niemelä’s key (van der
Our initial DNA barcodes of Swiss and French A. ovatula sensu lato were similar to those obtained by
With respect to the names that can be applied to these two species, five names proposed by
Three males are preserved under A. ovatula in the Kirby collection (
Preserved syntypes of A. afzeliella in the Kirby collection include three females and one male, all of which probably belong to “A. albofasciata” (one female and the male are stylopised, rendering identifications challenging); the best-preserved female (Figs
We tried to locate the type material of the Taeniandrena taxa described by A. Schenck, namely A. albofimbriata Schenck, 1853, A. canescens, A. distincta Schenck, 1861, A. gibba Schenck, 1853, A. octostrigata Schenck, 1853, A. ovata and A. plantaris Schenck, 1853. In the Schenck collection (SMFD), several specimens are preserved under the names wilkella, xanthura, afzeliella and convexiuscula. These specimens belong to several species including A. lathyri (see
However, Schenck described numerous taxa based on material collected by C. Kirschbaum, and some of the type material for these taxa is preserved in the Kirschbaum collection (
Several additional names are currently treated as synonyms of A. ovatula sensu lato (
Andrena afzeliella and A. ovatula can be separated from the other central European species of Taeniandrena by a combination of characteristics, which are summarised in the identification key given below. The criteria allowing separation of A. afzeliella and A. ovatula are presented in Table
Comparison of Andrena afzeliella stat. rev. and A. ovatula (modified from van der
Females | ||
Characters (order of importance) | Andrena afzeliella | Andrena ovatula |
Apical fringe of T5 | Golden (Figs |
Always dark brown to black medially, often lighter laterally (Fig. |
Pygidial fimbria | Mostly whitish laterally, light brown medially, sometimes entirely golden (Fig. |
Dark brown medially (sometimes a few light brown hairs laterally), without reddish hue (Fig. |
Apical fringe of hairs on femur, covering the base of basitibial plate | Orange to brown orange, never dark brown (Fig. |
Dark brown (Fig. |
Colour of scopal hairs | Entirely white or yellowish (Fig. |
White or yellowish, but nearly always with a few dark hairs dorsally near basitibial plate (Fig. |
Scutal vestiture | Long hairs yellowish white in fresh specimens; without short, dark hairs beneath long hairs (short hairs, if present, yellowish white). | Long hairs brownish yellow in fresh specimens; medially with dark, minute hairs beneath long hairs; this character is clearly visible in northern European populations, less so on the Iberian Peninsula. |
Clypeal punctation and sculpture | On average comparatively fine and sparse (Fig. |
On average comparatively coarse, sculpture shiny (Fig. |
Size | On average slightly smaller, body length mostly 8–9 mm, although some specimens are up to 10 mm. | On average slightly larger, body length 9–10 mm |
Males | ||
Character (order of importance) | Andrena afzeliella | Andrena ovatula |
Antennal ratio | A3 = 0.9–1.0 × A4 (Fig. |
A3 = 0.8–0.9 × A4 (Fig. |
Sculpture of disc of T3 and T4 | Shiny, punctation coarse and comparatively sparse (Fig. |
Shagreened, punctation dense and shallow, punctures little visible (Fig. |
Gonostylus and valve | Gonostylus broader, external margin often straight; valve often slightly broader (Fig. |
Gonostylus more slender, external margin always concave, valve often more slender (Fig. |
Gonocoxae | Internal margins usually diverging apically (Fig. |
Internal margins usually parallel apically (Fig. |
Vestiture | Body vestiture often entirely greyish white even in fresh specimens (Fig. |
Body vestiture yellowish brown in fresh specimens (Fig. |
Width of tergal fringes | comparatively broad (Fig. |
comparatively narrow (Fig. |
Andrena afzeliella is widely distributed in the Western Palearctic. Two barcoded specimens from Corsica and from Sardinia, respectively, clustered with continental populations of A. afzeliella in our phylogenetic tree (Fig.
Comparison of the females of Andrena antonellae sp. nov., and those of the populations of A. afzeliella and A. wilkella from Corsica and Sardinia. Andrena wilkella has so far only been reported from Corsica.
Characters (order of importance) | Andrena afzeliella | Andrena antonellae sp. nov. | Andrena wilkella |
---|---|---|---|
Apical fringe of T5 and pygidial fimbria | Brown-orange (bronze) | Brown-orange (bronze) | Greyish brown |
Sculpture of T1 | Fine but distinct on disc and on margin; underlying surface weakly shagreened, shiny | Mostly impunctate, sometimes with a few, scarcely visible punctures on disc and margin; underlying surface shagreened, weakly shiny (Fig. |
Fine but distinct on disc and margin, denser than in A. afzeliella but less so than in central European specimens of A. wilkella; underlying surface weakly shagreened, shiny |
Sculpture of T2 | Very densely punctate, punctures separated by less than 0.5 puncture diameters; underlying surface weakly shagreened, shiny; margin impunctate | Mostly impunctate, sometimes few barely visible punctures on disc (Fig. |
Moderately densely punctate, punctures separated by 1–2 puncture diameters (punctation less dense and less visible than in central European populations); underlying surface weakly shagreened, shiny; margin impunctate |
Sculpture of T4 | Very finely and very densely punctate, interspaces less than a puncture diameter, surface appearing shagreened | Very finely and very densely punctate, interspaces less than a puncture diameter, surface appearing shagreened (Fig. |
Finely and densely punctate, interspaces above one puncture diameter, surface appearing shiny |
Colour of integument | Margin of T1–4 dark, sometimes slightly brownish apically. Femora 3 predominantly orange. | Margin of T1–4 predominantly yellowish brown, dark brown only basally (Fig. |
Margin of T1–4 entirely dark. Femora 3 dark. |
Body size | 8–9 mm | 8–9 mm | 10–11 mm |
Holotype
♀ (Figs
Paratypes
(Suppl. material
Italy • ♀; Sardinia, Buggerru, Cala Domestica; 39°22'36"N, 8°22'57"E [39.3767°N, 8.3825°E]; 17.iv.2017; leg. J. Litman; C. Praz (PRUN) • ♂; Sardinia, Monte Crescia; 39.295°N, 9.392°E; 7.5.2018; leg. D. Bénon; unique identifier: GBIFCH00117711 (PRUN) [DNA extraction number 1437].
So far known only from the Islands of Sardinia and Corsica. In Corsica, the species is known from several localities in the south near Bonifacio, one locality along the east coast, as well as one locality in the north of the Island. In Sardinia, the species is known from two localities in the south of the Island. Andrena antonellae sp. nov. is probably widely distributed on both islands.
Unknown; two females have been captured on an unidentified yellow Fabaceae shrub (CJP, unpublished data).
Presumably univoltine with one generation from March until the end of June depending on elevation.
Female. Females of Andrena antonellae sp. nov. are characterised by the small body length (8–9 mm), the brown-orange vestiture of the mesosoma and the head (Fig.
Andrena antonellae sp. nov. is morphologically most similar to A. russula (=A. similis), A. fuliginata and A. croceiventris; according to current knowledge, these three taxa are absent from Sardinia and Corsica [the map for Andrena russula (=A. similis) presented in
The Corso-Sardinian populations of A. wilkella (Kirby, 1802) and A. afzeliella, the only two other species of Taeniandrena known so far from Sardinia and Corsica, also have an orange vestiture on the mesosoma, as A. antonellae sp. nov. Since all three species are found in sympatry, the criteria allowing for the identification of the female are summarised in Table
Male. The males of Andrena antonellae sp. nov. are sculpturally nearly identical to those of A. croceiventris and A. fuliginata. All three species have a comparatively elongate, oval-shaped genitalia with a narrow penis valve (Figs
Female. Measurements. Body length 8–9 mm.
Head. Head 1.3 times as wide as long. Clypeus dark, flattened over most of its area, densely and uniformly punctate with exception of a narrow central impunctate line, punctures separated by 0.5 puncture diameters, underlying surface weakly shagreened, usually shiny, especially apically (Fig.
Mesosoma. Scutum densely punctate, punctures separated by < 0.5 puncture diameters over majority of surface except becoming slightly sparser centrally and posteriorly, underlying surface strongly shagreened (Fig.
Metasoma. Terga dark, strongly shagreened, weakly shiny (Fig.
Male. Measurements. Body length 8–9 mm.
Head. Head 1.3 times as wide as long (Fig.
Mesosoma. Scutum, scutellum, episternum, and propodeum structurally as in female, punctation overall sparser. Scutum and scutellum with fine yellowish grey hairs that equal length of scape. Front legs dark, mid legs dark except tarsi, orange, hind legs dark, tibiae and tarsi orange (Fig.
Metasoma. Terga dark, finely shagreened and weakly shiny, apical part of marginal areas lightened, semi-translucent brown (Fig.
Female specimens from one high elevation locality in Corsica (Evisa, Col de Vergio 1477 m; one specimen is included in Fig.
This species is named in honour of Antonella Soro for her contributions to the field of conservation genetics of bees.
Andrena antonellae sp. nov. is sculpturally highly similar to A. croceiventris and A. fuliginata, both in the female and male sexes. Based on current evidence, these three taxa do not occur in sympatry. The highly similar morphology in these three taxa conflicts with our DNA barcoding results, which suggest that the three taxa are only distantly related. This discrepancy between molecules and morphology is reminiscent of the strong genetic divergences between A. afzeliella and A. ovatula; these two cases are puzzling and are further discussed below. While A. antonellae sp. nov., A. fuliginata and A. croceiventris are morphologically close, there are still subtle differences among them; these differences include the colour of the terminal fringe, the colour of the integument, the presence or absence of short dark intermixed hairs on the scutum, the sculpture of T1 and T2, and the sculpture of the clypeus. In Taeniandrena, such differences, although subtle, generally correspond to between-species differences. For this reason and based on the strongly divergent DNA barcodes, we treat these three taxa as distinct.
Andrena croceiventris Morawitz, 1871: 219, ♀, “Calabria” [Italy].
Andrena stefanii
Pérez, 1895: 41, ♀ ‘Sicile’ [Italy]. Synonymy in
Type material: Lectotype ♀ of Andrena stefanii Pérez, 1895. Italy • Sicile [Sicily], no further information (
Other material (Suppl. material
Only known from Italy between Tuscany and Calabria, as well as from Sicily.
Female. The most distinctive feature of Andrena croceiventris is the partly reddish colour of the integument of the metasoma, especially T1–3 (Fig.
Male. As mentioned above, the male of A. croceiventris is sculpturally nearly identical to those of A. antonellae sp. nov. and A. fuliginata. The lone specimen of A. croceiventris examined has part of the metasomal integument dark orange (tergal margins, pregradular areas and lateral, declivous parts), a unique feature in European species of Taeniandrena. Differences compared to A. wilkella and A. russula are mentioned under A. antonellae sp. nov. or in the identification key.
This species has so far been treated as a subspecies of the widespread species A. russula (=A. similis; see below) (
Andrena gelriae
van der Vecht, 1927: 87, ♀ ♂, „Putten“ [Holland]. Syntypes (
Andrena gelriae karelica Niemelä, 1949: 114, ♀ ♂, [Finland].
(Suppl. material
France • ♂; Avignon; [43.908°N, 4.878°E]; 11.5.2012; leg. N.-J. Vereecken; unique identifier: scwe057 (DGC) • ♀; Rouffach; [47.959°N, 7.298°E]; 18.6.1992; leg. F. Amiet; unique identifier: GBIFCH00117688 (
Italy • ♂; Italy, Basilicata, Monte Pollino; 39.904°N, 16.181°E; 4.7.2011; leg. Trunz, Litman, Praz; unique identifier: GBIFCH00117715 (PRUN) [DNA extraction number 1427].
Northern and Central Europe, including Scandinavia, the Netherlands, Belgium, Germany, Switzerland, France as far south as Valensole and Avignon (Fig.
Probably oligolectic on Fabaceae (
Univoltine, in central Europe from early June until mid-July, slightly after A. wilkella (although this species has a long flight period; see below) and A. intermedia, as in northern Europe (
This species is particularly challenging to identify in the female sex. Nearly all recent mentions from Switzerland (
See
Andrena intermedia
Thomson, 1870: 154, ♀ ♂, “Norrland” [Sweden]. Lectotype ♀, by designation of
(Suppl. material
Spain • ♂; Sierra Nevada, El Dornajo, 1700 m; 37.132335°N, -3.439249°E; 6.6.2021; leg. T.J. Wood; unique identifier: TJW0398 (TJWC) [DNA extraction number 2264] • ♂; Cuenca, Cueva de los Morceguillos; 40.181797°N, -2.01305°E; 21.6.2021; leg. T.J. Wood; unique identifier: TJW0460 (TJWC) [DNA extraction number 2270].
Northern and Central Europe, restricted to mountainous areas in Southern Europe.
Probably oligolectic on Fabaceae (
Univoltine, from early May until July depending on the elevation. According to
Andrena intermedia did not form a monophyletic group in our phylogenetic tree (Fig.
See
Andrena ovata
Schenck, 1853: 133, ♀ ♂, “[Nassau, Germany]. Lectotype ♂ (
Type material: Lectotype ♂ of Andrena ovata Schenck, 1853, by present designation, a male in good condition (Fig.
Other material (Suppl. material
Germany • ♀; Rheinland-Pfalz, Mechtersheim NSG Tongruben RLP MV5555 Standort 3; 28.5.1995; leg. Niehuis (CSE) [DNA extraction number 2373] • 2 ♀; Hessen-Höchst a.M. SO 46 Gutachten Industriepark Wiese, Lotus; 50.0907, 8.5494; 22.5.2020; leg. S. Tischendorf (STC and PRUN) • ♀; Hessen, Babenhausen Kiesgrube nördl. Bab, Düne 1; 7.6.1996; leg. S. Tischendorf (STC) • ♀; Hessen-Riedstadt MTB 6116 R3455-H5526 UTM MA52 sw Geinsheim Kiesbaggerei Kiebert; 24.5.2004; leg. S. Tischendorf (STC) • ♀; Hessen-Wattenheim; 49.689595, 8.40162; 1.6.2021; leg. S. Tischendorf (STC) • ♀; Hessen-Kelkheim/T MTB 5816 R3461-H5555 UTM MA65 Streuobstwiese; 14.6.2004; leg. S. Tischendorf (STC) • ♀; Hessen-Höchst a.M. SO 14 Gutachten Industriepark GS 24 trockene Wiese; 50.08654, 8.545; 22–31.5.2020 (STC) • ♀; Hessen, Babenhausen Kiesgrube nördl. Bab, Düne 1; 17.6.1996; leg. S. Tischendorf (PRUN) • ♀; Hessen-Darmstadt TK 6117-7522 Eberstadt, Prinzenberg; 28.5.1998 (STC) • ♂; Hessen-Bensheim TK 6317 3475/5505; 15.5.1994; leg. S. Tischendorf (STC) • ♂; Hessen-Höchst a.M. SO 14, trockene Wiese Gutachten Höchst AG GS 17; 8–17.5.2020; leg. S. Tischendorf (STC) • ♂; Hessen-Biebesheim FO19 Rheinufer Ufer Gras Proj. Blaues Band; 49.768567, 8.45083; 30.5.2017; leg. S. Tischendorf (STC) • ♂; Rhl.-Pfalz Ingelheim Rheindamm 6; 24.5.1992; leg. M. Hauser (STC).
Greece • ♀; Arkadia, 2 km NW Kosmas; 37°06'24"N, 22°43'42"E [37.1067°N, 22.7283°E]; 2.vi.2014; leg. J. Litman & C. Praz; unique identifier: GBIFCH00117721 (PRUN) [DNA extraction number 2265].
Italy • 2♂ 2♀; Lombardia, Magenta; 45.43553, 8.831499; 10.05.2019; leg. P. Biella (MIB) [unique identifiers MIB:ZPL:08811; MIB:ZPL:08812; MIB:ZPL:08813; MIB:ZPL:08814] • ♀; Gargano, San Giovanni; 41.717778, 15.724722; 24.v.2011; leg. S. Gerber, I. Mercerat; unique identifier GBIFCH00132003 (PRUN) [DNA extraction number 2375] • ♀; Gargano, San Giovanni; 41.673056, 15.726111; 24.v.2011; leg. S. Gerber, I. Mercerat; unique identifier GBIFCH00132004 (PRUN) [DNA extraction number 2377] • ♀; Levanto; 31.v–5.vi.1999; leg. M. Herrmann (MHC) [DNA extraction number 2371] • ♂; Levanto; 31.v–5.vi.1999; leg. M. Herrmann (MHC) [DNA extraction number 2376].
Spain: ♂; Segovia, Madrona, 500 m NE, Arroyo del Hocino; 40.9006°N, -4.1559°E; 15.5.2021; leg. T.J. Wood (TJWC).
Switzerland • ♀ Mendrisio TI [Ticino], Meride, 603 m; 717163/83448 [Swiss coordinates; 45.8927°N, 8.9481°E]; 8.v.2020; leg. L. Giollo; unique identifier: GBIFCH00124927 (
Peloponnese (Greece), southern and northern Italy, southern Switzerland (Tessin), western Germany (Hessen and Rheinland-Pfalz), southern and central France, northeast to Dijon and northwest to Orléans, and Segovia Province, Spain (Fig.
Unknown.
Presumably univoltine with one generation from April until mid-June.
Female. Females of Andrena ovata are characterised by the pale, grey vestiture on scutum and head (Figs
Females of Andrena ovata are superficially highly similar to those of A. poupillieri incana, restricted to the Balearic Islands. Both species have light vestiture, snow white tergal hair bands and a comparatively dark terminal fringe. Andrena poupillieri incana has thicker tergal hair bands, slightly coarser punctation on the scutum and denser clypeal punctation; in addition, the scopa is partly dark dorsally near the basitibial plate, while it is entirely yellowish white in A. ovata.
Male. The males of A. ovata are most similar to those of A. wilkella, with which they share the short third antennal segment (A3 = 0.6–0.7× A4; Figs
We initially treated this taxon as an undescribed species, until its presence in Hessen and Rheinland-Pfalz in Germany was brought to our attention. We thus examined the type material of the taxa described by Schenck and found one male syntype (Fig.
Andrena ovata belongs to a group of species characterised in the female by nearly entirely greyish white vestiture and weakly punctate terga. This group includes A. poupillieri incana, known only from the Balearic Islands, as well as A. poupillieri from North Africa (see note regarding this taxon under A. ovatula). No genetic data was available from A. poupillieri incana. The status of this subspecies remains unclear, but the male is sculpturally different from that of A. ovata: the genitalia of A. poupillieri incana are similar to those of A. ovatula, and A3 is subequal to A4, unlike the condition in A. ovata.
Lastly, it is probable that the females of “A. poupillieri” mentioned by
Melitta ovatula Kirby, 1802: 149, ♂ [indicated as female], “Barhamiae” [Barham, Suffolk, UK]. See note below for information on the type material.
Type material: Only three males are preserved in the Kirby collection (
Other material: Barcoded material includes specimens from Belgium, France, Germany, Italy, Portugal, Spain, the United Kingdom; in addition, sequences from Ireland are available on BOLD. Examined material additionally includes specimens from Andorra; see full list of examined specimens in Suppl. material
Widespread in north-western Europe (France, England, Belgium, the Netherlands, Germany; Fig.
Analysis of 30 pollen loads from 20 localities strongly suggest oligolecty on Fabaceae, with 99.6% of pollen collected from this family (TJW, unpublished data). This taxon is particularly associated with members of the tribe Genisteae in Atlantic-influenced environments across Western Europe, such as Cytisus, Genista, and Ulex. This association with Fabaceae shrubs may explain the particular distribution of A. ovatula, which appears to be more frequent in coastal areas than in the central parts of Europe.
Bivoltine, first generation in Northern Europe from the end of March until the end of May, second generation from early June until early September, approximately a month earlier than A. afzeliella (Fig.
The identity of Andrena poupillieri, a species that has been treated as a subspecies of A. ovatula, remains unclear because the Dours collection, presumably including all syntypes of this taxon, has been destroyed.
See under Andrena afzeliella (Table
Andrena russula
Lepeletier, 1841: 251, ♀, “Oran” [Algeria]. Holotype ♀ (
Andrena similis Smith, 1849: lx, ♂, “Bristol” [UK], syn. nov. Syntype or holotype ♂ (OUMNH).
Andrena ocreata cyprisina
Warncke, 1975c: 78, ♀,♂, “Limassol” [Cyprus], syn. nov. Holotype ♀ (
Andrena similis caraimica Osytshnjuk, 1994: 33, ♀,♂ [Crimea], syn. nov. Holotype (SIZK).
Type material: Holotype of A. russula (
Other material (Suppl. material
Crimea • 23♂ 28♀; Karadagh [Kara Dag], Vodianja balka; 44°56'22"N, 35°12'44"E; 21.4.2003; leg. Y. Budaschkin (
Iran • ♀; Lorestan province, Dorud Lanjaban env, 960 m; 33.419°N, 48.986°E; 10.5.2016; leg. M. Kafka (TJWC) [DNA extraction number TJW038].
Italy • ♀; Marche, M. Sibillini, M. Vettore; 42.8011°N, 13.2711°E; 30.6.2011; leg. Trunz, Litman, Praz; unique identifier: GBIFCH00117708 (PRUN) [DNA extraction number 1425] • ♀; Umbria, Castel Viscardo, 10 km NW Orvieto; 42.7574°N, 12.0023°E; 28.5.1991; leg. J. Gusenleitner (
Libya • ♀; Tripolitaine, Djebel Ghariane; 4.1899; leg. Alluaud (
Tunisia • ♀; Tunis; 189? (
Widely distributed throughout Europe, including the Iberian Peninsula, France, Central Italy, Bulgaria, Cyprus, Crimea, northern Africa from Morocco to Libya, Turkey, the Caucasus, the Levant (Israel, Syria, Lebanon, Jordan) and Iran.
Oligolectic on Fabaceae (
Univoltine, in Switzerland from the end of April to early June at low elevations, slightly later at high elevations.
This widespread species has so far mostly been referred to as A. similis. A first change to this view was advocated by
This species is widely distributed in the Western Palearctic; geographic variation in structural morphology is minimal, and variation that does exist such as in the strength of tergal punctation follows no clear pattern or gradient. In north-western Africa, in populations referred to as “A. ocreata ssp. russula” by Warncke (see
See
Andrena gelriae vocifera
Warncke, 1975a: 136, ♂ ♀, “Bordeaux[sic]/Drôme” [likely Bourdeaux, a municipality in the Drôme Departement, France]. Holotype ♂ (
Type material: Holotype ♂ (
Other material (Suppl. material
So far only known from southern France between the Drôme and the Landes departments (Fig.
Unknown.
Presumably univoltine with one generation from the end of May until the end of June.
The map presented for Andrena gelriae s. l. in
Female. The female of A. vocifera is unique among all European Taniandrena for the dense and coarse punctation of the terga, and for the shiny, hardly shagreened underlying sculpture (Figs
Male. This species can be identified by the genital structure, with a comparatively broad valve (width = 1.5 × diameter of lateral ocellus) with long, parallel-sided base (Fig.
Melitta wilkella
Kirby, 1802: 145, ♀, “prope Londinium” [near London, England]. Lectotype ♀ (
Type material: Lectotype ♀ of Melitta wilkella Kirby, 1802, by present designation, a female in good condition (
Other material: numerous specimens, only some of which are listed in Suppl. material
From northern Iberia (absent from Morocco, see
Considered to be univoltine (
In our genetic analysis, A. wilkella forms three clades separated by average distances comprised between 1.49 and 3.04% (Fig.
See
Unique characters in bold, non-unique characters in regular font; characters given in order of importance.
Females
1 |
Surface of tergal discs shiny with hardly any shagreenation (Fig. |
Andrena vocifera |
– | Surface of tergal discs shagreened. Punctation of tergal discs weaker and less dense, tergal margins clearly punctate only in A. wilkella (Fig. |
2 |
2 |
Terga comparatively strongly and densely punctate, punctures comparatively clearly visible in shagreenation (Fig. |
Andrena wilkella |
– | Terga less strongly and densely punctate, punctures less visible and often nearly imperceptible in shagreenation, especially on anterior, declivous part of T1 (interspaces commonly 2–3 puncture diameters) and on apical margin of T2. Width of apical tergal hairbands variable, usually larger (except in A. ovata). Colour of terminal fringe variable | 3 |
3 |
Vestiture on scutum and scutellum bright orange, particularly dense, especially on lateral parts of scutellum. Terga with shallow, nearly indiscernible punctures (Fig. |
Andrena russula |
– | Vestiture on scutum yellowish brown, brown-orange or grey, less dense. Tergal punctation, colour of integument of hind tibia and tarsi, colour of terminal fringe and tergal fringes variable | 4 |
4 |
Integument of T1–T4 partly red (Figs |
Andrena croceiventris |
– | Integument of T1–T4 dark brown. Punctation of terga variable, but if disc of T1 nearly impunctate and disc of T2 very sparsely punctate, then body size > 10 mm. Width of tergal fringes variable. Scutellum usually partly shiny. Body length variable | 5 |
5 |
Terga strongly shagreened, punctures little visible, disc of T1 and margins of T2–T4 appearing impunctate (Fig. |
Andrena ovata |
– | Terga more shiny and distinctly punctate. Scutal punctures deep and clearly visible. Sculpture of clypeus variable. Colour of tergal fasciae variable, often brownish in fresh specimens, wider, interrupted or not on T3. Scutal vestiture usually darker in fresh specimens. Colour of terminal fringe variable. Body length variable | 6 |
6 | Apical hairbands of T3 not interrupted medially in fresh specimens (Figs |
7 |
– | Apical hairbands of T3 interrupted medially even in fresh specimens (Figs |
8 |
7 | Terminal fringe dark brown (Fig. |
Andrena ovatula |
– | Colour of terminal fringe variable (Figs |
Andrena afzeliella |
The following two species are particularly difficult to separate in the female sex.
8 | Vestiture on scutum slightly longer, nearly twice as long as width of antennae; punctation on terga on average less dense and less visible (Fig. |
Andrena intermedia |
– | Vestiture on scutum shorter (as in A. wilkella). Punctation on terga on average denser, punctures more visible, approaching condition observed in A. wilkella, from which A. gelriae can be separated by the broader and more conspicuous apical tergal fringes (Fig. |
Andrena gelriae |
Males
1 | Tergal discs coarsely and sparsely punctate, underlying surface nearly completely shiny with almost no shagreenation (Fig. |
Andrena vocifera |
– | Tergal discs more finely punctate, underlying surface usually shagreened, especially when tergal punctation is coarse (e.g., in A. wilkella and A. gelriae; Figs |
2 |
2 | Valve basally broad, its maximal width approximately equal to 2 diameters of lateral ocellus (Figs |
3 |
– | Valve basally less broad, its width at most 1.5 × diameters of lateral ocellus (Figs |
4 |
3 |
Valve basally approximately as wide as maximal width of gonostylus (Fig. |
Andrena intermedia |
– | Valve less wide than maximal width of gonostylus (Fig. |
Andrena gelriae |
4 |
Integument of terga partly orange brown, especially margins, pregradular area and lateral, declivous parts. Tergal discs very faintly punctate, punctures little visible in shagreenation. Terga with comparatively narrow apical hairbands laterally, hairbands broadly interrupted on all terga. A3 0.7–0–8× as long as A4 (Fig. |
Andrena croceiventris |
– | Integument of terga dark, at most with apical parts of margins light brown. Punctation of terga, apical hairbands and length of A3 variable | 5 |
5 |
A3 short, approximately 0.6–0.7 × A4, the following segments approximately 1.7 times longer than wide (Figs |
6 |
– | A3 longer, 0.8–1.2 × A4, the following segments 1.5 times longer than wide (Figs |
7 |
6 |
Clypeus with dense, snow white vestiture, vestiture hiding underlying sculpture on apical parts of clypeus (Fig. |
Andrena ovata |
– | Clypeus without particularly dense vestiture. Body vestiture yellowish grey in fresh specimens. Terga weakly shiny and distinctly punctate (Fig. |
Andrena wilkella |
7 |
Broad, parallel-sided basal part of penis valve long and narrow (Fig. |
Andrena russula |
– | Parallel-sided part of penis valve shorter and broader (Figs |
8 |
The following two species are difficult to separate in the male sex.
8 | Terga more densely punctate, surface matt (Fig. |
Andrena ovatula |
– | Terga less densely punctate, surface weakly shiny (Fig. |
Andrena afzeliella |
Our results lead to new hypotheses on species delimitation in the subgenus Taeniandrena. In particular, the long controversy on the status of Andrena albofasciata (referred to here as A. afzeliella) and A. ovatula, appears definitively solved: in spite of sympatry in northern Europe as well as on the Iberian Peninsula and along the Mediterranean coast (Fig.
Two taxa hitherto treated as subspecies are raised to valid species, Andrena croceiventris, so far treated as a subspecies of A. russula (=A. similis), and A. vocifera, hitherto treated as a subspecies of A. gelriae. Both species pairs show distinct, but slightly overlapping geographic ranges: we document the presence of A. gelriae sensu stricto in southern France, and of A. russula in mountainous areas of central Italy, suggesting that these species pairs maintain morphological and genetic integrity in spite of range sympatry. This result is of more than academic importance, for the following reasons. First, both A. croceiventris and A. vocifera have a very narrow geographic range, with the former restricted to continental Italy and Sicily, and the latter to a small area in southern France between the Pyrenees and the Alps. A narrow geographic range is the basis for Criterion B in red list assessments, the most widely used criterion for insects. Second, both species appear to be particularly rare: we are aware of only nine specimens of A. vocifera collected in France in the last 40 years, despite intensive faunistic inventories; and the last known record of A. croceiventris in Italy was made in 2004. Both species are rather conspicuous, likely forage on “classical” bee host plants (Fabaceae), and are thus unlikely to be overlooked in surveys, raising questions on their conservation status. The very low number of records from these two taxa further stresses the poor knowledge of the southern European bee fauna and call for additional faunistic inventories in this important hotspot of bee diversity.
Our results highlight the presence of two newly recognized species in Europe. Andrena ovata appears to be widely distributed in Europe as demonstrated by the isolated records in Greece, Italy, Switzerland, Germany, France and Spain. This species has likely been overlooked due to challenging morphological identifications in Taeniandrena. DNA barcoding was pivotal in the detection of this species and in the association of sexes, as specimens were previously typically misidentified as A. ovatula sensu lato (females) or as A. wilkella (males).
The second species described here, Andrena antonellae sp. nov., shows a small geographic range restricted to the islands of Corsica and Sardinia. While several endemic insect species are known from these two islands, including the emblematic Corsican Swallowtail Papilio hospiton Gené, 1839, only few endemic bee species are so far known, including Andrena corssubalpina Theunert, 2007 and Nomada legoffi Dufrêne, 2021. Species delimitation in A. antonellae sp. nov., A. croceiventris and A. fuliginata was highly challenging, these three taxa being sculpturally nearly identical but forming distant clades in the phylogenetic tree, with genetic distances among these taxa comprised between 4.69–5.73% (Fig.
More generally, two opposing trends can be highlighted in European Taeniandrena when comparing morphological results with those based on DNA barcodes: on the one hand, groups of morphologically highly similar taxa were found to be genetically only distantly related, as in the case of Andrena croceiventris, A. fuliginata and A. antonellae sp. nov.; against our expectations, A. ovatula and A. afzeliella were not closely related (genetic distances 9.5–10.6%); an unclear taxon on the Iberian Peninsula, Andrena aff. russula (TJW017), is morphologically nearly identical to A. russula, but only distantly related to that species (6.49–7.87%); and superficially similar specimens corresponding to the description of A. poupillieri were only distantly related in the tree (Fig.
A second, contrasting pattern is observed in the gelriae-clade: in this clade, genetic differences were low, comprised between 1.12 and 3.54%, and even as low as 0.30 and 0.37% between A. contracta, A. levante and the Spanish specimens of A. intermedia. These comparatively low genetic distances strongly contrast with the marked morphological differences in male genitalia, which allow for a clear separation of these taxa (Figs
A first hypothesis to account for these two contrasting patterns is that morphological similarity does not necessarily imply phylogenetic relatedness, especially given the morphological homogeneity in the subgenus. A second hypothesis is that the diversification rate is particularly high in the gelriae-clade, given the high number of species and the low genetic divergences among morphologically clearly delimitated species. In agreement with this fast diversification, some species pairs or triplets in the gelriae-clade exhibit narrow geographic range and parapatric or near parapatric distribution, as observed in A. gelriae and A. producta in Central Europe, and A. levante, A. vocifera and A. gelriae in western Europe. A more extreme example of parapatry can be found around the Sierra Nevada in southern Spain. Andrena levante is present in dry valleys at lower altitudes up to ~1200 metres, A. intermedia is found in subalpine habitats between 1200–2000 metres, and A. contracta is found in the alpine zone above 2000 metres (
The use of DNA barcodes was pivotal in highlighting cryptic diversity in this group of bees, but also in obtaining crucial data on the distribution of several restricted bee species. Using short, but diagnostic barcodes (180 bp) we were able to generate genetic data for important records for which no recent material was available, and for confirming the morphological separation of Andrena ovatula and A. afzeliella performed more than 20 years ago in the absence of genetic data (van der
This study is dedicated to Jane van der Smissen, whose long commitment in the careful study of a local bee fauna has presciently led to the recognition of two closely related species long before genetic data could be used. Felix Amiet kindly donated specimens identified by Jane van der Smissen for genetic analysis. Gideon Pisanty and Sophie Cardinal made two barcodes of A. afzeliella available for this study. We thank Mike Herrmann, Paolo Rosa, Christian Schmid-Egger and Jessica Litman for discussions on earlier versions of this manuscript. Matthieu Aubert, Paolo Biella, Ian Cross, Eric Dufrêne, Lorenzo Giollo, Mike Herrmann, Kobe Janssen, Alireza Monfared, F. Javier Ortiz-Sánchez, Marino Quaranta, Lise Ropars, Bertrand Schatz, Christian Schmid-Egger, Maximillian Schwarz, Stefan Tischendorf, Peter Vanormelingen, and many others (see Suppl. material
Table S1. Specimens used in genetic analyses. The unique identifiers are identical to the Sample-ID on BOLD
Data type: excel file
Explanation note: Table S1. Specimens used in genetic analyses. The unique identifiers are identical to the Sample-ID on BOLD. The primers used in PCR and the size of the fragment sequenced are indicated (primer names are abbreviated; see text). GenSeq category after Chakrabarty et al. 2013, Zookeys 346: 29–41 (1 = holotype; 2 = paratype; 3 = topotype; 4 = vouchered specimens).
Table S2. List of all examined specimens.
Data type: excel file
Explanation note: Table S2. List of all examined specimens. Swiss occurrences are freely available on GBIF (