Andrena (Truncandrena) Warncke 1968

Abstract

SubgenusTruncandrenaWarncke, 1968 Genetics. The Subgenus Truncandrena is taxonomically challenging, with many “ subspecific ” names available due to Warncke’s broad species concepts and a high level of “ variation ” comprising different colour patterns (in both the integument and pubescence). This has led to the elevation of many of these subspecific concepts to full species status (e. g. Gusenleitner and Schwarz 2002; Pisanty et al. 2022 b; Wood 2023 b, 2024 a). In this context, many new DNA barcodes were available for the East Mediterranean fauna. For many taxa, they offered good resolution; although females are challenging to separate morphologically, Andrena caneae Strand, 1915 and Andrena serraticornis Warncke, 1965 were strongly separated genetically (Fig. 54). Three major issues were detected in this subgenus. Firstly, Andrena tscheki Morawitz, 1872 sensu lato formed three distinct subclades, with sequences from 1) Greece, 2) Germany, and 3) Israel and Lebanon. Greek specimens were separated from German specimens by an average genetic distance of 9.63 % (range 9.57–9.73 %), and from Levantine sequences by an average of 11.21 % (range 11.09–11.25 %). German and Levantine sequences were separated by an average of 7.38 % (range 7.29–7.45 %). This issue is discussed below, along with morphological recognition of the taxa. Secondly, Andrena schmiedeknechti Magretti, 1883 (described from Italy) formed two lineages, with sequences from Italy (Sicily) strongly separated from sequences from southern Greece by an average genetic distance of 6.28 % (range 5.87–7.12 %). Although smaller in absolute terms than the distances seen in A. tscheki sensu lato, this led to a topology where Italian A. schmiedeknechti were placed closer to members of the rufomaculata - group, and Greek A. schmiedeknechti were placed closer to Andrena urfanella Scheuchl & Hazir, 2012. The name Andrena schmiedeknechti ssp. flavopilis Warncke, 1965 is available for Greek populations, but additional study is required. The topology of the tree separating Italian and Greek sequences is unstable, with bootstrap support grouping Italian A. schmiedeknechti + rufomaculata - group of only 37. The addition of more sequences from this group (including A. schmiedeknechti ssp. nigrorum Warncke, 1967 from North Africa) could result in a monophyletic clade, as in A. tscheki sensu lato. There also do not appear to be structural differences between the genital capsules of A. schmiedeknechti sensu lato seen in Italy and Greece, and the only difference is in the colouration of the hair (Warncke 1965 a). As such, no action is taken here, and the broad species concept of Warncke is maintained. The third issue surrounds the rufomaculata - group which contains the taxa Andrena rufomaculata Friese, 1921, Andrena paramythensis Mavromoustakis, 1957, and Andrena optata Warncke, 1975. Warncke (1974 a, 1975) treated this as a broad taxon A. rufomaculata, with A. rufomaculata s. str. found in southern Turkey and the Levant (locus typicus: Turkey, Airan [an Armenian village, modern name not currently traceable] in the Taurus Mountains), A. rufomaculata ssp. paramythensis found in Cyprus (locus typicus: Cyprus, Paramytha), and A. rufomaculata ssp. optata found from central Anatolia across the Mediterranean coast to Greece, Albania, the Dalmatian coast, southern Hungary, and Crimea (locus typicus: Greece, Lamia). Although Warncke and Gusenleitner and Schwarz (2002) used this broad concept, subsequent authors (e. g. Hazir et al. 2014) have treated the three as distinct, though this has led to some confusion between A. paramythensis and A. optata, with the latter being reported as the former in some studies (e. g. Hazir et al. 2014; Pisanty et al. 2018). Although Warncke considered these “ subspecies ” not to have overlapping ranges, A. optata and A. rufomaculata actually appear to have overlapping ranges in the Levant, with A. optata descending to Lebanon and northern Israel (Mount Hermon) (Pisanty et al. 2018, as A. paramythensis misidentification; Wood et al. 2020; Fig. 55). Whilst the morphological characters appear solid, available genetic sequences indicate that the differences within this group are very small. Three sequences of A. paramythensis showed very low intraspecific variation of 0.04 % (range 0.00–0.06 %). These were separated from two sequences of A. optata (Greece: Peloponnese) plus one sequence of A. rufomaculata (Lebanon) by an average genetic distance of 3.61 % (range 3.33–3.81 %). The sequences of A. optata were separated from A. rufomaculata by only 0.84 % (range 0.76–0.91 %). In absolute terms, this is a small genetic difference. Inclusion of additional sequences from Cyprus and the Levant actually show more complexity, with no consistent topological differences and sequences of all three taxa intermixed (G. Pisanty, pers. comm.). However, the ecological affinities of the two taxa suggest genuine species-level differences. In the Levant, A. optata is found at high altitude on Mount Hermon and the Mount Lebanon chain between 1200–2200 m a. s. l. whereas A. rufomaculata is much more thermophilic and is found at low to moderate altitudes. Specimens morphologically conforming to A. paramythensis are found only on the island of Cyprus. Consequently, as a temporary solution, A. optata is retained here as a good species despite the only weak to non-existent genetic separation from A. rufomaculata. This remains an area for future investigation using more powerful genetic tools to quantify realised gene flow in this group. Overall, the genetic results from within Subgenus Truncandrena paint a challenging picture in which it is difficult to consistently translate high genetic distances into species-specific differences within existing morphological species concepts, and some existing morphological species concepts receive low to non-existent genetic support. Whilst the conclusions presented here are conservative, they maintain nomenclatural stability and allow morphological recognition of taxa which is beneficial for ecological study and monitoring of bee species pending further study.

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