Andrena fuliginata Pérez, 1895, Andrena gregaria Warncke, 1974, Andrena poupillieri Dours, 1872, and Andrena prazi sp. nov.

Praz et al. (2022) identified two putative lineages that could refer to the taxon A. poupillieri, for which the type series has been lost. Wood (2023) designated a neotype (see Fig. 4) for the lineage called poupillieri #2 by Praz et al. (2022). This species now has an unambiguous status. The identity of poupillieri #1 appears to be A. gregaria based on type examination and the generation of additional sequences from Morocco (Fig. 4). This species resembles A. poupillieri morphologically, but the terga are much more densely punctate, including punctures on the declivity of T1 (cf. A. wilkella (Kirby, 1802)). However, two specimens conforming to this concept had identical barcodes to the barcode of A. fuliginata used in the analysis of Praz et al. (2022). This A. fuliginata specimen (HYMAA289-22) conforms to the type concept (Fig. 5A–D) in that it has short dark hairs on the scutum underlying the longer light brown scutal hairs (Fig. 5B), the hind tibiae are orange, the terga are almost impunctate, including the declivity of T1 (Fig. 5C), and the terminal fringe is dark (Fig. 5D). There appears therefore to be some barcode sharing in this group of species; no taxonomic action is taken, and all sequences are grouped together in a broad gregaria / fuliginata aggregate for the time being. This aggregate shows a very low average ‘intraspecific’ variation of 0.37% (range 0.00–0.68%) and has a bootstrap support of 96.

In contrast, a sister clade was formed next to this aggregate consisting of specimens from Morocco originally identified as A. russula Lepeletier, 1841 with dark terga and one specimen from Tunisia with red-marked terga. This clade was separated from the gregaria / fuliginata aggregate by an average of 1.92% (range 1.61–2.74%), and has low average intraspecific variation of 0.44% (range 0.00–1.00%); this clade consequently has a bootstrap support of 94. Morphologically, it cannot be confused with either A. gregaria or A. fuliginata, and so it is described as A. prazi sp. nov. (see diagnosis section for this species).

Material examined

ALGERIA • ♀, lectotype of Andrena fuliginata; Téniet [Théniet El Had]; MNHN (Fig. 5A–D).

MOROCCO • ♀, holotype of Andrena gregaria; Marraquesh [Marrakech]; 1–31 Mar. 1907; Escalera leg.; OÖLM.