Andrena (Euandrena) Hedicke 1933

Abstract

SubgenusEuandrenaHedicke, 1933 Genetics. The Subgenus Euandrena is very young by Andrena standards (Pisanty et al. 2022 a), and contains a swarm of morphologically variable and challenging to characterise taxa. The situation is particularly acute in the East Mediterranean, where multiple faunas meet. There are also many mountainous areas; these landscape features create many areas with phenologically distinct microclimates, and seem to support high Euandrena species richness when considered at a global scale (Osytshnjuk et al. 2008; Praz et al. 2019; Wood et al. 2021; Wood 2021, 2023 b, 2023 c, 2024 a). Since the work of Praz et al. (2019) which resolved taxon delineation issues in Central European Euandrena, a large number of sequences have now becomes available from the East Mediterranean. This is pertinent, as Praz et al. left a number of questions open; whilst elevating Andrena pileata Warncke, 1975 and treating it distinct from Andrena allosa Warncke, 1975 (Spain, France, Switzerland). Andrena pileata was described from Chelmos mountain in the Peloponnese, and Praz et al. considered the described paratype male not to be conspecific with the holotype male. They also identified three uncertain taxa supported by DNA barcodes which could not at the time be named; sp. 1, sp. 2, and sp. 3. Analysis of DNA barcodes, including the most pertinent barcodes of Praz et al., provides resolution to these open questions (Fig. 7). To begin with, sp. 2 sensu Praz et al. (HYMAA 038-18) comes from the island of Lesbos, and almost perfectly matches sequences of Andrena robusta Warncke, 1975 from Lesbos and Samos (separated by 0.26 %); the species was originally described from the island of Rhodes. This question at least is straightforward to answer. More challenging is sp. 1 sensu Praz et al.; sequences from Chelmos and Mount Olympus (HYMAA 037-18; HYMAA 043-18) placed adjacent to a long series of sequences of A. pileata from Mount Maenalon (Peloponnese), Mount Parnassus (south Central Greece), and the Pirin Mountains (south-western Bulgaria) (Fig. 8), separated by an average genetic distance of 2.02 % (range 1.67–2.44 %). As in the analysis of Praz et al., sp. 1 was placed close to A. allosa, but the addition of new Balkan sequences has resulted in a topology where clades of A. pileata and A. allosa are reciprocally monophyletic, although bootstrap support is relatively low, potentially indicating an unstable topology. Taking A. pileata (inclusive of sp. 1), it was separated from A. allosa by an average genetic distance of 1.46 % (range 0.95–2.28 %); excluding sp. 1, A. pileata was separated by an average genetic distance of 1.35 % (range 0.95–1.67 %). sp. 1 alone was separated from A. allosa by an average genetic distance of 1.89 % (range 1.55–2.28 %). Adjacent to A. allosa are two sequences, NATBE 120-25 which is a female Euandrena from Petrich (Bulgaria) on low altitude grassland at the Rupite mineral baths (20 May 2024) and WPATW 1645-25 which is a male Euandrena from Armenia. They were separated from A. allosa by an average genetic distance of 2.22 % (range 1.27–3.19 %). The male specimen is badly abraded and essentially uncharacterisable given the challenges associated with identification of isolated male Euandrena specimens. The female specimen shows the densely plumose scutal hairs typical of allosa + pileata, but the clypeus is comparatively slightly shorter and the terga are shinier than in A. pileata. Moreover, it was collected at an altitude of 90 m a. s. l., when the one available sequenced specimen of A. pileata from Bulgaria was collected at an altitude of approximately 2,000 m a. s. l. (3.5 km NNE of Popina Luka). Given what is known about Euandrena and their ecological and altitudinal niches, and the likely break in gene flow between low and high altitude in this group, it is difficult to imagine these being conspecific. In this context, it is best to currently consider A. allosa and A. pileata (inclusive of sp. 1) as distinct following Praz et al. (2019), and to consider the Bulgarian + Armenian sequences of uncertain identity at the present time until more data become available. Praz et al. (2019) also considered that Warncke’s paratype male of A. allosa ssp. pileata was not conspecific with the holotype (even though they were collected in sympatry, just two days apart), but based on the examination of a barcoded male of A. pileata (NATBE 056-25; Fig. 9), it seems to match Warncke’s specimen in all aspects including the shape of the genital capsule, and so Warncke’s existing male paratype is considered to have been correctly associated. Before moving to sp. 3 sensu Praz et al., it is also necessary to report Andrena amieti Praz, Müller & Genoud, 2019 as new for the southern Balkans. Four sequences were available from northern Greece collected at an altitude of 1800–1900 m a. s. l. which were separated from A. amieti group 1 (Switzerland, including the holotype; bootstrap support of 86) sequences by an average genetic distance of 0.96 % (range 0.91–1.07 %) and from A. amieti group 2 (Switzerland and Italy; bootstrap support of 95) by an average genetic distance of 2.45 % (range 2.31–2.58 %). Compared to the analysis of Praz et al. (2019: 16), all Andrena amieti sequences formed a monophyletic clade with bootstrap support of 86, rather than fall into two distinct clades which were interrupted by the clade of A. allosa. The Greek clade of A. amieti received bootstrap support of 89, and overall seems to have have helped resolve the topology, grouping all three clades together with bootstrap support of 86. Although there is slight population structuring, the morphology of A. amieti remains recognisable over this range. Examination of 111 A. amieti specimens from Greece and North Macedonia (see Suppl. material 2) shows activity between 12 May and 25 August, supporting the bivoltine behaviour reported in Switzerland by Praz et al. (2019). In this context, A. amieti can be considered a species found in high mountains in the Alps, southern Apennines, and southern Balkans (Fig. 8). Collectively, covering all three clades, a low average intraspecific genetic distance of 1.04 % (range 0.00–2.74 %) was observed. The distribution is not unlike that of Andrena montana Warncke, 1973; the one available sequence from northern Greece (Mount Kajmakčalan) clustered with sequences from Switzerland, showing minimal average genetic differentiation of 0.25 % (range 0.15–0.46 %). For sp. 3 sensu Praz et al. (HYMAA 036-18), the sequence clustered with a long series of sequences from Bulgaria, Greece, Lebanon, and Jordan. This group of sequences requires discussion, and an analysis of what Warncke referred to as Andrena rufitibialis Friese, 1899 in a broad sense. Friese (1899: 342) described A. rufitibialis from two female specimens from 13 Apr. 1899 collected at Jericho by Francis Morice. Gusenleitner and Schwarz (2002: 647) give ZMHB or OUMNH as the possible type repositories. Examination of these collections produced three female specimens (Figs 10, 11), but only the specimen in the ZMHB was actually labelled by Friese with one of his red type labels. Friese’s type material is often difficult to interpret (see Rasmussen and Ascher 2008); although the OUMNH material may well be syntypic, it is best to treat as only possibly syntypic, and the female in the ZMHB collection illustrated here is designated as the lectotype. It was later redetermined as Andrena biguttata Friese, 1923 by Stöckhert for inexplicable reasons, given that this name is younger, regardless of specimen morphology. Warncke (1969 a) listed A. rufitibialis, and then described Andrena rufitibialis ssp. limosa Warncke, 1969 from Turkey (locus typicus), Syria, and Israel which is darker, with the metasoma entirely dark, but still with partial reddish colouration of the hind tibiae (Fig. 12). This reddish colouration of the hind tibiae is absent on many “ dark ” specimens that can be collected in the southern Balkans and the Levant. Rasmont et al. (2017: 18) elevated Andrena rufitibialis ssp. limosa to species status based on personal communication from Erwin Scheuchl and reported it as new for Greece based on specimens from the East Aegean Islands. No species concept was discussed, or global range indicated, for example whether Turkish material is conspecific with dark material from the southern Balkans and Levant. However, this elevation is justified due to their differing morphologies combined with the ecological affinities of the taxa; A. rufitibialis s. str. is a very rarely collected species of arid habitats (see Material examined below), and is completely absent from the Mediterranean zone in the Levant whereas darker specimens (lacking extensive red tergal markings, A. limosa sensu lato) are common throughout the Mediterranean zone (G. Pisanty, pers. comm.). In the current analysis, it has not been possible to generate a barcode from either i) a specimen with orange hind tibiae and abundant red markings on the terga (A. rufitibialis s. str.) or ii) a generally dark specimen still displaying partially orange hind tibiae (A. limosa), but many sequences are available from entirely dark specimens which form a broad clade from southern Bulgaria to Lebanon and north-western Jordan (Fig. 7). It is necessary to highlight the sequence NATBE 129-25 which was generated from a specimen that was part of a long series of dark specimens collected from Campanula (Campanulaceae) flowers near Gorno Spanchevo in south-western Bulgaria. It shows no morphological differentiation from other specimens from the same collecting event (e. g. NATBE 128-25, NATBE 131-25), but shows a level of genetic differentiation from all remaining dark specimen sequences of 3.38 % (range 2.93–3.67 %). It is unclear if this is variation or if the sequence has problems, for example, if it was generated from nuclear mitochondrial DNA (a NUMT), sequences that are notoriously difficult to differentiate fr

Authors 1

1. Thomas James Wood corresponding Aachen Naturalis Biodiversity Center

   Naturalis Biodiversity Center

   Affiliation as printed

   Naturalis Biodiversity Center, Darwinweg 2, 2333 CR, Leiden, Netherlands