An update and revision of the Andrena fauna of Morocco (Hymenoptera, Apoidea, Andrenidae) with the description of eleven new North African species

Abstract Morocco has a diverse bee fauna, but one that has also been relatively understudied in recent years. Here a revision of the species-rich genus Andrena is presented that reveals eleven new species for science and substantially improves our understanding of North African Andrena. From Morocco, Andrena (Aciandrena) semiadesus Wood, sp. nov., Andrena (Aciandrena) triangulivalvis Wood sp. nov., Andrena (Campylogaster) sparsipunctata Wood sp. nov., Andrena (Carandrena) hebescens Wood sp. nov., Andrena (Cnemidandrena) niveofacies Wood sp. nov., Andrena (incertae sedis) tenebricorpus Wood sp. nov., Andrena (Notandrena) acutidentis Wood sp. nov., Andrena (Poliandrena) breviceps Wood sp. nov., and Andrena (Poliandrena) farinosoides Wood sp. nov. are described and their ecology is discussed. Andrena (Aciandrena) astrella Warncke, 1975 is synonymised with Andrena (Aciandrena) fulica Warncke, 1974 syn. nov. The unknown female of Andrena (Nobandrena) ounifa Warncke, 1974, and the unknown male of Andrena (Poliandrena) guichardi Warncke, 1980 are described. Andrena (incertae sedis) gafsensis Wood sp. nov. from Tunisia is described due to its similarity to Andrena tenebricorpus. Andrena (Poecilandrena) nigriclypeus Wood sp. nov. from Algeria is also described as it was collected within 10 km of the Moroccan border. A further 18 species are recorded in Morocco for the first time. Andrena (Melandrena) nitida (Müller, 1776) and Andrena (Notandrena) nitidiuscula Schenck, 1853 are removed from the Moroccan list due to historic problems in the application of these names to Mediterranean taxa.


Introduction
Andrena are the second most speciose genus of bees worldwide after Lasioglossum (Ascher and Pickering 2020). The genus has a primarily Holarctic distribution and with the greatest diversity found in Mediterranean and xeric regions (Gusenleitner and Schwarz 2002). The Mediterranean basin of the Western Palearctic is therefore one of the hotspots of Andrena diversity, and new Andrena species continue to be described from North Africa and the Levant at regular intervals (e.g., Scheuchl 2009;Scheuchl and Gusenleitner 2009;Scheuchl et al. 2011;Schwenninger 2015;Pisanty et al. 2016;Pisanty et al. 2018). Compared to the Levant, the Andrena fauna of north-western Africa is historically better known and described thanks to a long history of taxonomic activity (e.g., Pérez 1895;Schmiedeknecht 1900;Saunders 1908;Alfken 1914;Schulthess 1924;Benoist 1961). The Andrena fauna of this region was later revised by Warncke (1967Warncke ( , 1974Warncke ( , 1980Warncke ( , 1983 who described many new species. Since then there has been relatively little work conducted on North African Andrena until recently (Scheuchl 2009;Scheuchl et al. 2011;Benarfa et al. 2013;Cherair et al. 2013;Schwenninger 2015;Djouama et al. 2016) as interest in the conservation, ecology, and taxonomy of wild bees has increased. However, nothing on the taxonomy of Andrena in Morocco has been published since 1980 (Warncke 1980), and a critical reappraisal of this rich fauna is therefore warranted.
Inspection of 5,685 female and 2,275 male Andrena specimens from contemporary and ongoing collections in Morocco and undetermined material from the Oberösterreich Landesmusum, Linz, the Naturalis Biodiversity Center, Leiden, and the personal collection of Maximillian Schwarz, Ansfelden, comprised 155 valid species, with many new records for Morocco and undescribed taxa. These are described herein, and the identity of several problematic taxa are reviewed.

Materials and methods
Identification was enabled through a comprehensive review of the literature in combination with visits to the Warncke collection in Linz as part of the lead author's ongoing review of West Palearctic Andrena species (Wood 2020;Wood et al. 2020a, b).
The subgeneric classification of Andrena continues to pose problems. Because of the very large number of species, taxonomic workers have largely been restricted to only a part of the global fauna (broadly, West Palearctic, Central and Eastern Palearctic, and Nearctic falling into three groups), and therefore a co-ordinated global system of subgeneric classification has not yet been possible. In both a Moroccan and West Palearctic context, the system of Warncke (1968a) is the most relevant, and so we broadly follow his assigned subgenera. However, though no globally consistent system currently exists, major steps towards such a reclassification have recently been made (Pisanty et al. 2020). Molecular analyses show that many existing Andrena subgenera are strongly polyphyletic (particularly groups such as Poecilandrena and Poliandrena), and are in need of deep revision. It is currently beyond the scope of this study to revise the subgeneric status of Moroccan Andrena taxa, but we take the opportunity to discuss problematic taxa, highlight outstanding areas of confusion, and to place newly described taxa with an appropriate degree of confidence in light of the findings of Pisanty et al. (2020).
Body length was measured from the vertical plane of the front of the head to the tip of the metasoma. Morphological terminology follows Michener (2007). Photographs were taken using an Olympus E-M1 Mark II with a 60 mm macro lens and were stacked using Zerene Stacker 1.04 (Zerene Systems, USA) and plates were prepared in GNU Image Manipulation Program (GIMP) 2.10.
In order to investigate a possible synonymy (see section on Andrena alchata Warncke, 1974), DNA was extracted and amplified from two specimens. DNA was extracted from two legs of each individual using a NucleoSpin Tissue (Macherey-Nagel, Düren Germany) extraction kit and following manufacturer´s instructions. The region selected for amplification was the LEP fragment within the mitochondrial gene cytochrome oxidase I (cox1), a region widely used in Hymenoptera taxonomy (Sheffield et al. 2009). However, as retrieving genetic information from pinned specimens can be challenging (Wandeler et al. 2007), a section within the fragment was targeted for amplification using primers LEP-F1 and LEP-R2 (Hebert et al. 2004). PCR profile consisted of an initial denaturation step at 94 °C for 3 min followed by 36 cycles of denaturation at 94 °C for 1 min, annealing temperature at 50 °C for 1 min and elongation at 72 °C for 1 min followed by a final extension at 72 °C for 10 min. Results of the PCR reaction were checked in a 1.5% agarose gel. Amplicons were sequenced using forward and reverse primers (Eurofins, Germany).
Diagnosis. The finely shagreened propodeal triangle, the narrow facial foveae, the mately 90% of surface (Fig. 6). Yellow marking does not extend to base of clypeus. Laterad of clypeal centre, yellow marking reduced, two black triangular markings extending in towards the centre, giving marking a broad, inverted T-shape. In basal half, clypeus dull and shagreened, becoming shiny in apical half. Clypeal punctures shallow and irregular, separated by 2-5 puncture diameters. Process of labrum rectangular, twice as broad and long, shiny. Face, gena, and vertex with long white hairs, longest on underside of head, equalling length of the scape. Scape and pedicel black, following antennal segments dark brown to black, A3 longer than A4, shorter than A4+5 combined. Gena and ocelloccipital distance as in female. Mesosoma: Scutum and scutellum dull, strongly shagreened, sparsely and shallowly punctured, punctures separated by 1-4 puncture diameters. Propodeal triangle weakly marked, scarcely differentiated from propodeum, shagreened, slightly shining laterally. Mesosoma with white hairs, sparse on scutum, denser and longer on episternum, exceeding length of scape. Legs dark, only final tarsal segment slightly lightened to dark brown, pubescence white. Wings hyaline, venation brown, stigma brown, nervulus interstitial. Metasoma: Tergites as in female, but margins more strongly lightened, transparent whitish apically (Fig. 7). Tergal margins laterally with loose white hair bands of irregular length, very widely interrupted. Genitalia simple (Fig. 8), gonocoxites forming rounded dorsal lobes, gonostyli long, tapering to rounded end.
Distribution. Central and eastern Morocco (Fig. 145a). Floral preferences. None recorded. Etymology. The name semi (partly or partially) + adesus (eaten, worn, eroded) was chosen to illustrate the clypeus of the male, where the central yellow marking is laterally diminished by intruding black marks.
Diagnosis. The male of A. triangulivalvis resembles other Aciandrena with a yellow polished clypeus such as A. pratincola. However, it can be instantly separated from all other Aciandrena with a yellow clypeus by the structure of the genitalia. Normal Aciandrena genitalia are simple, with a relatively narrow, unmodified penis valve (Figs 8, 100, 102, 104, 126, 128, 130;Warncke 1968bWarncke , 1972Warncke , 1974Warncke , 1975aPisanty et al. 2016). However, in A. triangulivalvis the penis valve is inflated and forms a triangle that sites between the broad gonostyli that have a slightly raised internal margin (Fig. 12).
Diagnosis. Andrena sparsipunctata can be easily placed into the Campylogaster because of the large, dense, and clear punctures on the episternum combined with its large body size. However, recent evidence shows that Campylogaster is strongly polyphyletic, and the species in northwestern Africa do not fall close to A. (Campylogaster) erberi Morawitz, 1871, the type species of Campylogaster that differs by its tormentose pilosity. A new subgenus is probably needed for the species around A. sparsipunctata (Pisanty et al. 2020). Against this context, it is close to the two most widespread Campylogaster species in North Africa, A. pruinosa Erichson, 1835 ssp. succinea Dours, 1872 andA. caroli Pérez, 1895, both of which are also stained red over the majority of the metasoma. The female of A. sparsipunctata is instantly recognisable because the punctures of the first tergite are sparse, separated by 2-4 puncture diameters (Fig. 16) whereas in the other two species the punctures are dense and separated by 1-2 puncture diameters (Figs 18,20). Moreover, the foveae of A. sparsipunctata females are narrow and depart from the inner eye margins so that at the top of the compound eye they are separated from the compound eye by a distance subequal to the width of a fovea itself (Fig. 15), whereas in the other two species the foveae are never clearly separated from the top of the compound eye (Figs 17,19). The scutal hairs are also normal, not squamous (Fig. 15), strongly contrasting with the squamous hairs of the other two species (Figs 17,19).
In males, A. sparsipunctata can be recognised by the shape of the labrum which is wide and rectangular (Fig. 23), whereas in A. pruinosa it is trapezoidal and markedly emarginate (Fig. 25). In A. caroli it is less strongly trapezoidal and emarginate (Fig. 27 (Fig. 28), whereas A. sparsipunctata has only a few scattered hairs on the tergal margins, never forming well-marked bands, as in A. pruinosa (Fig. 26).
Description. Female: Body length 12 mm (Fig. 13). Head: Black, a little wider than long. Clypeus domed, evenly and shallowly punctured, punctures separated by 0.5-1 puncture diameters. Clypeus surface weakly shining, fore margin slightly upturned. Process of labrum trapezoidal, with weak lateral striations, fore margin weakly and narrowly emarginate. Gena 1.2 times wider than width of compound eye, densely and deeply punctate, punctures extend to the vertex where they become much shallower and encircle two impunctate areas immediately adjacent to lateral ocelli, these areas equivalent to inter-ocellar area in size (Fig. 15). Gena and vertex with a mane of bright ginger hairs, achieving a maximum length of 2/3 rds of the length of the scape. Face, scape, and clypeus with shorter ginger hairs, these achieving a maximum length of ½ the length of the scape. Antennae dark, A3 apically and A4-12 lightened orange below, A3 longer than A4+5, shorter than A4+5+6. Foveae narrow, slightly narrower than the width of an antenna, close to the internal margin of compound eye below level of antennal insertions but diverging above so that at top of compound eye each fovea is separated by a distance subequal to its own width, area between a fovea and the internal margin of compound eye shiny and sparsely punctured. Ocelloccipital distance broad, two times the width of lateral ocellus. Mesosoma: Scutal punctures dense, punctures almost touching, underlying surface shiny. Scutum and scutellum with non-squamous hairs, these achieving a maximum length of ½ the length of the scape (Fig. 14). Episternum and propodeum with exception of propodeal triangle densely and evenly punctate, punctures separated by 0.5 puncture diameters, underlying surface finely and subtly shagreened, weakly shining. Propodeal triangle weakly rugose. Episternum and propodeum with normal ginger hairs, these achieving the length of the scape. Femora dark, tibia and tarsi orange. Femoral and tibial scopa orange. Wings slightly infuscate, venation dark, stigma slightly translucent. Nervulus interstitial. Metasoma: Tergites amber, becoming amber to dark brown on T1. T1 strongly contrasting with following tergites, T1 sparsely punctate, punctures separated by 2-4 puncture diameters, surface of disc clear of hairs, shining (Fig. 16). T2-4 densely punctured, punctures separated by 0.5-1 puncture diameters, covered in short ginger hair across entire surface. Tergal margins slightly depressed, covered in short ginger hairs that form weak apical fringes, poorly differentiated from hairs on discs of T2-4. T5-6 with long ginger-golden hairs flanking pygidial plate, pygidial plate simple, triangular, without raised margins.
Male. Body length 11-12 mm (Fig. 21). Head: Similar to female, wider than long, pubescence whitish throughout but probably faded. All punctation weaker and finer with exception of clypeus where it is equally strong as in the female. Labrum broadly rectangular, wider than long, fore margin straight, without emargination (Fig. 23). Mandibles long, crossing, very slightly bidentate with tiny internal tooth. Scape darkorange, A2-13 evenly orange coloured, A3 equalling A4. Ocelloccipital distance broad, 1.5 times width of lateral ocellus. Mesosoma: Similar to female, punctation of the scutum and scutellum weaker, punctation of the episternum and propodeum absent, replaced by weak rugosity. Venation of wings amber, nervulus interstitial to antefurcal. Metasoma: Tergites amber, extent variable, becoming darker on centre of tergal discs, in some cases forming extensive dark bands across entire surface (Fig. 24). Punctation shallow and even, separated by 1-2 puncture diameters, not noticeably sparser on T1 compared to other tergites. Sternite 8 long, rectangular, three times longer than broad, covered in hairs, apex weakly emarginate. Genitalia simple (Fig. 22), gonocoxites slightly pointed, weakly separated from each other apically.
Distribution. Southern and eastern Morocco in desert environments (Fig. 145c). Floral preferences. None recorded.
Remarks. There is uncertainty over the exact status of A. pruinosa succinea, as the North African animals differ in colouration of the metasoma and scutal hairs (red) from A. pruinosa sensu stricto (black-brown) from Spain. However, structural differences are minor (e.g., the shape of the foveae is the same), and molecular investigation is warranted. It is however clear that the name succinea cannot apply to the bees described here as the first tergite of females is much less densely punctured than A. pruinosa s.l. from either Spain or North Africa, and Dours (1872) makes no mention of contrasting tergal punctation in his description 'Abdomen… très-finement ponctué et hérissé de poils…'. The sampling locations of A. sparsipunctata, both found at the extreme edges of Morocco, may provide an explanation as to why this species has gone undetected to date.
Diagnosis. The female of A. hebescens can be placed in the Carandrena because the dorsolateral angle of the pronotum has a transverse ridge, the propodeal triangle is shagreened and weakly rugose at the base, and there is almost no punctation on the metasoma, and the head has a typical Carandrena shape, broader than long, with the inner eye margins slightly converging below (Fig. 30). It is most similar to A. euzona Pérez, 1895 andA. microthorax Pérez, 1895 given its dark, non-metallic appearance (A. aerinifrons Dours, 1873, A. bellidis Pérez, 1895, A. daphanea Warncke, 1974, A. deserta Warncke, 1974, A. nigroviridiula Dours, 1873, and A. reperta, Warncke, 1974 with metallic green or blue integument, A. binominata Smith, 1853, A. eremobia Guiglia, 1933, and A. leucophaea Lepeletier, 1841 with partially red metasoma), dull domed clypeus, and white hair bands on the tergites (A. eddaensis Gusenleitner, 1998 andA. decaocta Warncke, 1967 with felt-like hair on the tergal discs). However, the scutum is shagreened and only weakly shining (Fig. 31) whereas in A. euzona it is extensively smooth and shiny (Fig. 33), and the hair bands are narrower (Fig. 32) whereas in A. euzona they are wider (Fig. 34). The clypeus is also evenly punctured and shagreened, weakly shining with only a subtle impunctate mid-line whereas in A. microthorax it is shiny and strongly punctured with a conspicuous impunctate mid-line, particularly at the fore margin of the clypeus where it forms a broad impunctate triangle.
The male of A. hebescens has a yellow marking on the clypeus, but this is reduced in size and does not cover the entire clypeal surface, with two black markings that extend in from the clypeal margins giving the marking an inverted 'T' shape (Fig. 36).
The male is therefore superficially similar to A. semiadesus, but can be separated by the typical Carandrena head which is clearly wider (more than 1.5 times) than the width of a compound eye in A. hebescens (Figs 37,38) but only slightly wider in A. semiadesus and lacking an angulate hind corner. The male of A. microthorax can have a yellow clypeal marking, but the scutum and tergites of this species are shiny, whereas in A. hebescens they are shagreened and at most, weakly shining (Fig. 39).
Male. Body length 8 mm (Fig. 35). Head: Similar to female, clypeus slightly arched, evenly and shallowly punctured, punctures separated by one puncture diameter, no impunctate central line. Clypeus centrally with yellow mark, this not reaching the lateral or basal margin of the clypeus, laterally invaded by two black marks therefore forming a broad inverted T-shape (Fig. 36). Underlying surface weakly shagreened, moderately shiny. Process of labrum trapezoidal, fore margin inflated, slightly bulbous, very weakly emarginate. Antennae dark, A4-13 slightly lightened dark brown, A3 exceeding A4, shorter than A4+5. Gena enlarged, most 1.2 times wider than compound eye, noncarinate, with weakly angulate hind corner (Fig. 37). Gena, vertex, and face below the level of the antennal insertions with long white hair equalling length of the scape. Scape, frons, and inner margin of compound eyes with mixture of black and white hairs. Ocelloccipital distance short, 2/3 width of lateral ocellus. Mesosoma: Similar to female but scutum with stronger shagreenation, dull except for central shining line (Fig. 38). Scutellum centrally weakly shining, contrasting with the scutum. Episternum, propodeum, and mesosomal pubescence as in the female. Legs dark, tarsi lightened brown, with white pubescence. Wings hyaline, venation dark brown, stigma centrally light brown. Nervulus interstitial to slightly antefurcal. Metasoma: Similar to the female. Tergites more clearly punctured, punctures visible against the microreticulation with slightly raised margins giving the overall surface an uneven impression (Fig. 39). T1-5 consistently punctured, punctures separated by 3-4 puncture diameters. T2-4 laterally with weak fringes of white hair, T5-6 with complete fringes of whitish to golden hairs. Sternites forming loose white hair bands apically. Genitalia simple, of a typical Carandrena form, gonocoxites apically forming weak points, rounded, diverging apically (Fig. 40).
Distribution. South-western Morocco in the Souss valley (Fig. 145a). Floral preferences. None recorded. Etymology. The name hebescens was chosen because this member of the Carandrena, though morphologically similar to several species with metallic green colouration, is completely dark, therefore heb-(dull or blunt) + escens (becoming). Diagnosis. Andrena niveofacies can easily be recognised as part of the Cnemidandrena because of the transverse ridge on the dorsolateral angle of the pronotum, the triangular hind tibiae in the female, the upturned process of the labrum in the male (Fig. 48), the thick and distinct hair bands on the metasoma (Figs 45,46), and the late summer activity period.
Within the Cnemidandrena, A. niveofacies females have completely brown hair on the scutum (intermixed with black in A. denticulata (Kirby, 1802)), the galea is dull (shiny in A. fuscipes (Kirby, 1802)), T5-6 are black haired (light haired in A. tridentata (Kirby, 1802)), and the face is pale haired (black at least in part in A. nigriceps (Kirby, 1802) and A. freygessneri Alfken, 1904). The female A. niveofacies is most similar to A. simillima Smith, 1851 but is can be easily separated as it has strikingly bright white hairs on the face and gena (Fig. 43), these being buff-brown in A. simillima.
The male has a non-carinate gena (Fig. 49, carinate in A. denticulata and A. tridentata), the galea is dull (shining in A. fuscipes), S8 is short and relatively densely haired (long and sparsely haired in A. freygessneri, see illustrations in Ember 2001), and the face has bright white hairs, with no trace of black hairs on the discs of the tergites (face with brown hairs in A. nigriceps, T4-5 with black hairs basally). However, the male material also differs from A. simillima as A3+4 are equal in length, whereas in A. simillima A3 is a little longer than A4.
Description. Female: Body length 10 mm (Fig. 41). Head: Black, a little wider than long. Clypeus slightly arched, evenly and clearly punctured, punctures dense, separated by 0.5-1 puncture diameters with the exception of subtle central impunctate line than widens into a triangle immediately before fore margin of clypeus. Underlying surface weakly shagreened, slightly shining. Process of labrum trapezoidal, fore margin clearly emarginate. Gena slightly wider than width of compound eye. Gena, face, and scape with bright white hairs, longest achieving length of the scape (Fig. 43). Vertex with contrasting brown and black hairs of equal length. Foveae generally wide, occupying three quarters of distance between top of compound eye and lateral ocellus. Antennae black, ventral surface of A4-12 lightened brown, A3 slightly shorter than A4+5. Ocelloccipital distance equalling width of lateral ocellus. Mesosoma: Scutum densely punctate, punctures separated by 0.5 puncture diameters over majority of disc but becoming sparser towards rear in centre and on scutellum, here separated by one puncture diameter. Underlying surface microreticulate, but this weakens to become fine shagreenation, shining, centrally and laterally on scutum and particularly on scutellum. Episternum with microreticulation, dull. Propodeum laterally shagreened, weakly shining, posterolaterally microreticulate and dull with large shallow punctures, these absent from the propodeal triangle which is therefore well defined. Scutum, scutellum, and propodeum with buff-brown hairs of moderate length, contrasting with white hairs on episternum. Legs dark, tarsal segments lightened brown. Wings hyaline, venation black, stigma centrally dark brown. Nervulus clearly antefurcal. Metasoma: Tergites black, microreticulate, weakly shining, moderately punctured, punctures separated by 1-2 puncture diameters (Fig. 45). T1 with long buff-brown hairs over whole disc and margin, T2-4 with clear buff-brown hair bands occupying marginal area, surface of disc with sparser and shorter hairs of same colour. T5-6 with black-brown hairs. S2-4 with sparse white hairs forming weak bands.
Male. Body length 10 mm (Fig. 47). Head: As in the female, but clypeus without central impunctate line, process of labrum narrow and centrally upturned (Fig. 48). Galea dull (Fig. 49). Antennae dark, A4-13 lightened to dark brown below, A3 equalling A4. Gena wide, 1.5 times wider than width of compound eye, non-carinate, forming weak posterior angle. Face, gena, scape, and vertex all with long white hairs achieving length of the scape. Ocelloccipital distance broad, 1.5 times width of lateral ocellus. Mesosoma: Scutum and scutellum as in female, microreticulation weaker so the whole surface appears shinier. Propodeal punctures weaker and shallower, slightly obscured by weak reticulation, contrast with the propodeal triangle as in the female. Metasoma: As in female (Fig. 51), T2-5 with pale hair bands, T6 with brown apical fringe. S8 short and densely hairy (Fig. 53).
Distribution. Probably restricted to the High Atlas Mountains of Morocco. This material represents the first record of the subgenus Cnemidandrena from Morocco and more broadly the whole of North Africa (Fig. 145b).
Remarks. There has historically been uncertainty over the species status of taxa assigned to A. simillima. The former subspecies A. freygessneri was returned to species status by Ebmer (2001), but the status of A. s. bremensis Alfken, 1900 andA. s. sischkai Warncke, 1988 remain unclear. This latter subspecies is found in Bulgaria and Greece (Figs 42, 44, 46, 50, 52, 54;Warncke 1988) and the facial hairs are buff coloured as in the nominate form ( Fig. 44), but the tergites are much less densely hairy and the hairs are shorter, giving the overall impression that the bee is darker (Figs 46,52). It is likely that both taxa are distinct from A. simillima, but more work is required. Against this context, we describe A. niveofacies as a good species, not only because of the morphological differences but also because of the degree of geographical separation, with the locus typicus in the High Atlas approximately 1,600 km from the nearest A. simillima records in southern France. The collection labels themselves are a small mystery, as they have the collector name Warncke struck through by hand (Warncke). They may have been used by a collector accompanying Warncke who used his spare preprinted labels, but if this is the case then their identity is not clear.
Etymology. The subspecific epithet niveofacies from niveus (snow) + facies (face) was chosen to illustrate the bright white facial hair. Diagnosis. Placement of this species into a subgenus is not immediately obvious. Andrena gafsensis is a small to medium sized Andrena with a strongly flattened clypeus that is extensively shiny in both the female (Fig. 56) and the male where it is coloured yellow, with this colouration extending onto the lower paraocular areas (Fig. 60). This pattern of yellow colouration is typically found in males from the subgenera Holandrena Pérez, 1980, Ulandrena Warncke, 1968, and Nobandrena Warncke, 1968. Specimens can quickly be separated from the first two because of the sculpturing of the episternum and propodeum (simple, not honeycomb-areolate) and from the second by the shape of the female hind tibial spur (straight not medially broadened) and the genitalia of the male (penis valve not inflated).
Nobandrena normally contains large species in the range of 13-14 mm in length, most of which have males with domed clypei which are extensively yellow-marked, with the yellow extending onto the lower paraocular areas. However, there are two smaller species A. iliaca Warncke, 1969 (7-8 mm, Turkey, Israel, and Jordan) and A. ounifa Warncke, 1974 (7-7.5 mm, Algeria, see below for new description of female) placed into the Nobandrena by Warncke where the male clypeus is shinier and flatter, whilst retaining the same colour pattern. This flattened clypeus can be seen in the female of A. ounifa which has a strongly flattened and shining clypeus in both the female (Fig. 94) and the male (Fig. 98). However, these two taxa clearly do not belong in the Nobandrena, with molecular analysis placing A. iliaca in the Fuscandrena (Pisanty et al. 2020). The position of A. ounifa is unclear without molecular work, and so confident placement of A. gafsensis is not possible at this time.
Given these problems, Andrena gafsensis is therefore best recognised in the female by the distinctive clypeal structure (Fig. 56), but then separated from A. ounifa by the short propodeum (Fig. 57) and by the colouration of the tergites which are almost completely red except for the basal parts of T1 and two lateral black spots on T2 (Fig. 58). The male can be recognised by the clypeal structure and colouration, but can also be separated by the short propodeum as in the female and by the structure of the genitalia (Fig. 62) which are most similar to A. iliaca but the gonocoxites form lateral points and the gonostyli lack an emargination in the outer margin.
Description. Female: Body length 9 mm (Fig. 55). Head: Black, as wide as long. Clypeus very slightly arched laterally, flattened over the majority of its area (Fig. 56). Laterally with punctures separated by one puncture diameter but only immediately adjacent to clypeal margin, remaining parts more sparsely punctured, punctures separated by 2-3 puncture diameters, underlying surface smooth and shiny. Process of labrum broad, trapezoidal, approximately twice as wide basally as long. Gena approximately width of compound eye. Gena, vertex, face, and scape with white hairs, the longest not exceeding length of the scape. Foveae of average width, occupying just over half the distance between top of compound eye and lateral ocellus. Antennae black, A4-12 extensively lightened orange below, A3 equalling A4+5+6. Ocelloccipital distance short, less than 1/3 width of lateral ocellus. Mesosoma: Scutum and scutellum densely punctured, punctures separated by 1-1.5 puncture diameters, underlying surface of the scutum shagreened anteriorly and laterally, shiny centrally, posteriorly and on scutellum. Episternum microreticulate, dull. Lateral faces of propodeum (propodeal corbicula) shagreened, weakly shining. Hind face of propodeum microreticulate and weakly rugose, rugosity outlining propodeal triangle which is rugose and therefore defined by change in texture (Fig. 57). Scutum and scutellum with short brown hair, at most achieving the length of A3. Episternum and propodeum with longer white hair, never exceeding length of the scape. Legs dark, tarsi lightened brown, pubescence brown to brownish white. Femoral scopa white, tibial scopa white to light brown. Wings hyaline, venation brown, stigma light brown. Nervulus slightly antefurcal. Metasoma: Tergites light red with exception of black basal and lateral parts of T1 and two black lateral spots on T2, and apex of T5 which is slightly lightened brown (Fig. 58). Sternites light red with exception of S1 and apical margins of S4-6. Tergites densely, finely, and evenly punctured, punctures separated by one puncture diameter. T2-4 with thin white hair bands of fine hairs, on T2-3 widely interrupted, on T4 complete. T5-6 with golden hairs flanking pygidial plate, laterally white. Pygidial plate with slightly raised margins, centrally with longitudinal slightly raised area.
Male. Body length 9-10 mm (Fig. 59). Head: Black, slightly wider than long. Clypeus very shallowly arched, essentially flat, ivory-coloured with exception of two faint dark spots centrally, yellow colouration extending onto bottom parts of paraocular areas (Fig. 60). Underlying clypeal surface shiny, moderately and shallowly punctured, punctures separated by 1-2 puncture diameters. Process of labrum trapezoidal, fore margin broadly and shallowly emarginate, shiny. Pubescence and antennal colouration as in female, A3 equalling A4+5. Ocelloccipital distance slightly shorter than width of lateral ocellus. Mesosoma: Scutum and scutellum dark, shagreened, dull laterally and anteriorly, weakly shining only centrally, moderately punctured, punctures separated by one puncture diameter. Sculpturing of episternum and propodeum as in female. Pubescence as in female. Wings hyaline, venation brown, stigma light brown. Nervulus clearly antefurcal. Metasoma: Tergites predominantly dark, tergal margins slightly depressed, lightened yellow to orange (Fig. 61). Tergites microreticulate, weakly shining, shallowly punctured, punctures separated by 1-2 puncture diameters. T2-5 with loose white hair fringes, that on T2 weakly interrupted, the following complete. T6 with fringe of longer white-golden hairs. Genitalia simple, gonocoxites diverging apically and forming points (Fig. 62) Diagnosis. Andrena tenebricorpus is very similar to A. gafsensis and faces the same problems of subgeneric classification, and is therefore also not currently placed in one until molecular data are available. Both species have a flattened and shiny clypeus (Fig. 64) and a short propodeum (Fig. 65), but differ most obviously in colour, with A. tenebricorpus having dark brown terga (Fig. 66) in strong contrast to the red terga of A. gafsensis. In terms of structural difference, in A. tenebricorpus the foveae are slightly wider, the hairs on the scutum and scutellum are relatively much shorter (compare Figs 57, 65) at the same time that the hair bands on the apical margins of the tergites are relatively much thicker (compare Figs 58,66), suggesting that this is not as a result of abrasion, and the scopa are a dirty brown (Fig. 63) rather than white (Fig. 55).
In terms of overall colouration and appearance A. tenebricorpus is extremely similar to A. guichardi Warncke, 1980 (also found only in south-western Morocco) with the same dark to dark brownish tergal colouration with contrasting white hair bands (see Fig. 104). However, A. tenebricorpus can be clearly separated by the flattened clypeus (domed in A. guichardi), more densely punctate scutum (punctures separated by 1-2 puncture diameters, by 2-4 puncture diameters in A. guichardi, see Fig. 103), and by the propodeal triangle which is weakly differentiated only by sculpturing (clearly marked by a small raised carina in A. guichardi).
Description. Female: Body length 10 mm (Fig. 63). Head: Dark, as wide as long. Clypeus slightly arched laterally, broadly flattened on the disc, underlying surface laterally with shagreenation, the majority of disc shiny (Fig. 64). Moderately punctured, punctures separated by 1-2 puncture diameters, with generally impunctate line in centre of clypeus. Process of labrum trapezoidal with faint lateral striations, fore margin weakly emarginate, slightly bulging. Gena equal to width of compound eye. Gena, face, and scape with white hairs, longest not exceeding length of the scape. Vertex with mixture of whitish and yellowish hairs of same length as on rest of face. Foveae of average width, occupying slightly more than half of distance between top of compound eye and lateral ocellus. Antennae dark, A4 apically and A5-12 lightened orange below, A3 equalling A4+5. Ocelloccipital distance equalling width of lateral ocellus. Mesosoma: Scutum and scutellum densely punctured, punctures separated by 1-1.5 puncture diameters, underlying surface of the scutum shagreened anteriorly, elsewhere and on the scutellum shiny. Episternum microreticulate, dull. Lateral faces of the propodeum shagreened, weakly shining. Hind face of propodeum microreticulate and weakly rugose, rugosity outlining the propodeal triangle which is itself rugose and therefore defined by the change in texture between rugosities (Fig. 65). Scutum and scutellum with very short and fine hairs, scarcely longer than width of lateral ocellus. Episternum and propodeum with longer white hair, never exceeding length of the scape. Legs dark, tarsi brown, pubescence brown. Femoral scopa white, tibial scopa dirty brown. Wings hyaline, venation brown, stigma dark brown. Nervulus slightly antefurcal. Metasoma: Tergites dark brown throughout with exception of two lateral black spots on T2 (Fig. 66). Tergites densely, finely, and evenly punctured, punctures separated by one puncture diameter. T2-4 with dense white hair bands that obscure underlying surface, on T2 interrupted medially, on T3-4 complete. T5-6 with thick brown hairs flanking pygidial plate, with a few white hairs laterally. Pygidial plate with slightly raised margin, centrally flat.
Remarks. Whilst A. tenebricorpus is structurally very similar to A. gafsensis, the subtle morphological differences suggest that A. tenebricorpus is distinct. Moreover, the Souss valley to the Guelmim region is bio-climatically quite distinct to the deserts of southern Tunisia and hosts many unique bee species within Morocco itself, further suggesting a distinct specific identity.
Etymology. The name tenebricorpus from tenebris (dark) + corpus (body) was chosen because of the structural similarities between A. tenebricorpus and A. gafsensis, but without the same reddish colouration.  Biougra-Tafraout, 13.ii.1987, 1♂, leg H. Teunissen, NMNL. Paratypes are deposited at the OÖLM and NMNL, with a male and female retained in the personal collection of TJW.
Diagnosis. A small Notandrena recognised in the subgenus by the dorsolateral angle of the pronotum with a transverse ridge, the clearly punctured metasoma, and the weakly rugose (not shagreened) propodeal triangle. Because of its small size it can be placed into the nitidiuscula group, and it is most similar to A. fulvicornis Schenck, 1853 (see also species newly recorded for Morocco below). It differs by the clypeus which has a central, shining, impunctate line (Fig. 68, evenly punctured and shagreened in A. fulvicornis) and the shiny scutum and scutellum (Fig. 69, shagreened and dull in A. fulvicornis).
The male can easily be recognised as a Notandrena because of the greatly enlarged and carinate gena (Fig. 73) and the broadened apex of the gonostyli in combination with a punctured metasoma (Fig. 75). It can be further recognised within the Notandrena by the shape of the genitalia which are short and compact (Fig. 76), placing it close to species like A. chrysosceles, A. fulvicornis, A. nitidiuscula, and A. pallitarsis (see illustrations in Schmid-Egger and Scheuchl 1997). However, the apices of the gonostyli are produced into points whereas they are rounded in the other species, the clypeus is yellow (Fig. 72, black in species like A. fulvicornis and A. nitidiuscula), and the apex of sternite 8 is also emarginate. The combination of these characters is unique.
Distribution. The Souss valley in south-western Morocco (Fig. 145d). Floral preferences. None recorded. Other members of the Notandrena are associated with Apiaceae (Wood et al. 2020a).
Etymology. The name acuti (sharp) + dentis (teeth) was chosen because of the male genitalia where the apices of the gonostyli are produced into points in contrast to other members of this group where they are rounded. Diagnosis. The subgenus Poecilandrena is poorly defined and has been treated as a wastebasket for species with few apomorphies (Pisanty et al. 2018), and so unsurprisingly it has been found to be strongly polyphyletic, containing at least five clades (Pisanty et al. 2020), and in future revisions, only the labiata and viridescens groups are likely to remain in the Poecilandrena. However, until this point it is desirable to keep superficially similar species together, even if the subgenus itself is clearly polyphyletic, so that they may be dealt with together in future revisions. Based on the criteria outlined by Pisanty et al. (2018), A. nigriclypeus can be placed into the Poecilandrena sensu lato by the combination of small body size (< 10 mm), red marked abdomen, non-carinate pronotum, mesepisternum weakly areolate with large, spaced shallow punctures, propodeal triangle moderately rugose, genal area not broadened, and sternite eight columnar.
Currently only one species of Poecilandrena sensu lato is known from North Africa, the similarly red-marked A. maximiliani Scheuchl, 2009 which was described from Tunisia (see Scheuchl 2009 for discussion of previous unconfirmed records of Poecilandrena from North Africa). Andrena nigriclypeus is similar to A. maximiliani as both have male genitalia with pronounced but apically truncate (square-ended) gonocoxal teeth. However, it can be easily separated because the clypeus is black, not yellow, and the gonostyli are completely different, small and narrow and with a shallow emargination in the outer edge before the apex whereas in A. maximiliani they are long, broad, and flattened. The only other similar species is A. paradisaea Warncke, 1975 from Turkey which has a red abdomen and a black clypeus in the male. However, this species has clearly hooked hind tibial spurs and the genitalia are completely different with a broad-based penis valve, spatula-shaped gonostyli with an evenly rounded outer margin, and pointed gonocoxal teeth (see photograph in Pisanty et al. 2018: fig. 122).
Distribution. North-western Algeria close to the Moroccan border (Fig. 145d). Floral preferences. None recorded. Etymology. The name nigri (black) + clypeus (clypeus) was chosen because of the entirely black male clypeus, which is unusual within small red-marked Andrena species. Diagnosis. The subgenus Poliandrena Warncke, 1968 is currently unsatisfactorily defined (females keying out in three places in Warncke's 1968a key) and will probably be broken up in the future as it has been shown to be strongly polyphyletic (Pisanty et al. 2020). It predominantly contains Mediterranean species that have heads that are short and broad (Fig. 84), short facial foveae, strongly punctate metasomas (Fig. 86), and often show a propodeal triangle that is defined by an external carina and internal rugosity (but not honeycomb-areolate, Warncke 1968a). As such, members of the Poliandrena are often recognised by their similarity to each other rather than by a single character per se, in other words a wastebasket taxon.
Andrena breviceps is small relative to the rest of the Poliandrena, comparable in size to A. marsae Schmiedeknecht, 1900 andA. laurivora Warncke, 1974, two of the smallest Poliandrena, but A. breviceps can easily be separated from them by the colour of the tergites which are dark brown (red in A. marsae, dark metallic greenblue in A. laurivora) and the sculpturing of the scutum which is shiny and (relatively within the Poliandrena) sparsely punctate, punctures separated by 2-3 puncture diameters (Fig. 75, separated by less than ½ a puncture diameter in A. marsae and by one puncture diameter in A. laurivora, underlying integument shagreened, with a metallic glint). It is also similar to the larger A. relata Warncke, 1967 which is newly recorded for Morocco (see below) because of their similar dark tergites and general appearance. However, it can also be separated using the same scutal punctation character (in A. relata punctures dense laterally, centrally separated by at most two puncture diameters) and also by the clypeus where A. breviceps has a central longitudinal impunctate line that is not present in A. relata. Overall, the shiny and relatively sparsely punctate scutum (Fig. 75) in combination with its small size should allow separation from other Poliandrena from North Africa.
Description. Female: Body length 8.5-9 mm (Fig. 83). Head: Black, clearly wider than long (Fig. 84). Clypeus broad, arched, with large punctures, punctures separated by one puncture diameter except for a central slightly raised impunctate line, underlying surface uneven, slightly but irregularly raised between punctures, shining. Process of labrum broadly trapezoidal, fore margin weakly emarginate. Gena as wide as width of compound eye. Gena, vertex, face, and scape with moderately dense white hairs, longest of these not exceeding length of the scape. Foveae moderately broad, occupying 2/3 of distance between the of compound eye and lateral ocellus, of normal length, not extending below the level of the antennal insertions. Antennae bright, scape and pedicel dark, A3 apically marked with orange, A4-12 predominantly orange, A3 exceeding A4+5, shorter than A4+5+6. Ocelloccipital distance short, ½ width of lateral ocellus. Mesosoma: Scutum and scutellum moderately punctured, punctures separated by 1-2 puncture diameters, underlying surface shiny, very weakly shagreened laterally and anteriorly (Fig. 85). Margins of scutum and scutellum with short, dense whitish hairs, these extending only very sparsely onto the disc. Episternum and propodeum microreticulate, dull, propodeum with weak rugosity, propodeal triangle defined with small but clear slightly raised carina, propodeal triangle with sparse and weak rugosity centrally. Episternum and propodeum with white hairs, longest not exceeding length of the scape. Legs dark, tarsi lightened brown, pubescence whitish to brownish. Femoral and tibial scopa white. Wings hyaline, venation and stigma brown. Nervulus antefurcal. Metasoma: Tergites brownish, margins slightly depressed, lightened white to yellow, apically translucent (Fig. 86). Tergites densely, evenly, and finely punctate, punctures separated by one puncture diameter, underlying surface very weakly shagreened, shining. T1 with two small lateral hair patches of white hair on apical margin, T2-4 apically with complete white hair bands obscuring underlying surface. T5-6 with golden hairs flanking pygidial plate. Pygidial plate rounded triangular, flat, without raised margin.
Floral preferences. None recorded. Etymology. The name brevi (short) + ceps (head) was chosen because of the particularly short and wide head of this species, even within the Poliandrena. Diagnosis. Andrena farinosoides can also be placed in the Poliandrena because of its short and broad head (Fig. 88), short facial foveae, strongly punctate metasoma (Fig. 91), and propodeal triangle marked by a carina and internal rugosity (but not honeycomb-areolate). It can be recognised within the Poliandrena as very similar to A. farinosa Pérez, 1895 from Spain and France (Gusenleitner and Schwarz 2002) but the two species differ in the pubescence of the abdomen and the scutal punctation. In addition to thick hair bands on the tergal margins (Figs 90, 92), A. farinosa also has short hairs on the tergal discs forming a sparse velvety pubescence when viewed laterally (Fig. 94, not as dense as in A. eddaensis or A. decaocta). In contrast, A. farinosoides has only a few short hairs on the tergal discs, not forming a velvety pubescence (Fig. 93). In addition, the scutum of A. farinosoides is clearly less densely punctate, with punctures separated by one puncture diameter, punctures denser and almost confluent in A. farinosa.
Diagnosis of the male is more difficult, but both A. farinosa and A. farinosoides are in the group with dark, densely punctate tergites and black (not yellow) clypei with an upturned fore margin (Fig. 96). The male of A. farinosoides is then less densely and more finely punctate on the scutum (punctures separated by one puncture diameter) whereas A. farinosa has larger punctures that are separated by less than a puncture diameter (Fig. 97).
Description. Female: Body length 8.5-9.5 mm (Fig. 87). Head: Black, clearly wider than long (Fig. 88). Clypeus broad, slightly arched, densely and evenly punctured, punctures separated by ½ a puncture diameter. Process of labrum broad, trapezoidal, twice as wide as long, fore margin compressed and slightly upturned, surface with weak transverse striations. Gena slightly narrower than width of compound eye. Gena, face, and scape with dense white hairs, longest not achieving ¾ of length of the scape, hairs on the vertex of similar length but becoming yellowish. Foveae normal, occupying half of distance between top of compound eye and lateral ocellus. Antennae dark, A5-12 lightened orange below, A3 long, slightly shorter than A4+5+6. Ocelloccipital distance short, less than 1/3 width of lateral ocellus. Mesosoma: Scutum and scutellum densely punctured, punctures separated by one puncture diameter, underlying surface smooth and shiny (Fig. 89). Margins of scutum and scutellum with short whitish-brownish hairs, densest on the margins and becoming sparser as hairs move into centre of the disc. Episternum and propodeum with weak rugosity, dull, propodeal triangle clearly defined by raised carina, propodeal triangle itself rugose with weak longitudinal carinae. Episternum and propodeum with white hairs, not exceeding ¾ length of scape. Legs dark, tarsi lightened brown, pubescence light brown to whitish. Femoral and tibial scopa white. Wings hyaline, venation and stigma light brown. Nervulus interstitial. Metasoma: Tergites dark, margins depressed and slightly lightened to brown (Fig. 91). Tergites densely and evenly punctured, punctures separated by ½ puncture diameter. T1-4 with thick white hair bands that exceed length of margins, completely obscuring underlying surface, all bands complete in fresh individuals. T5 and T6 with golden hairs flanking pygidial plate, white haired laterally. Pygidial plate rounded triangular, without raised margin, very subtly domed centrally.
Distribution and remarks. Andrena fulica (Morocco, Algeria, and Tunisia) and A. astrella (Spain and Portugal) were both described by Warncke, with A. fulica described as A. astrella ssp. fulica (Warncke, 1974) despite being formally published before A. astrella (Warncke 1975b), this problem arising due to differences in publishing speed between the two journals, with the paper on Iberian Andrena originally submitted in 1971. The most recent global treatment listed them as distinct (Gusenleitner and Schwarz 2002).
The two species can be rapidly recognised within the Aciandrena because they have punctured tergites with the punctures extending onto the tergal margins, a character that is unique within this subgenus. According to Warncke (1974), Andrena fulica is characterised as having finer tergal punctation and, importantly, a black clypeus in the male sex (Fig. 99), whereas in A. astrella the male clypeus is yellow (Figs 101, 103). The description of A. fulica by Warncke (1974) is extremely short, just two sentences for the female and two short sentences for the male which read: '♂ wie beim ♀ feiner punktiert. Clypeus dunkel gefärbt!'.
Inspection of Aciandrena males from Ifrane in Morocco (9.v.1997) revealed the two colour forms in sympatry, though unfortunately only two specimens in total (one A. fulica, one A. astrella) were collected during this sampling making an assessment of variation impossible. Additionally, specimens from Benidorm in Spain (20.iv.1982) also showed the two colour forms in sympatry. Across all Iberian material studied, 46/47 (97.8%) males showed a consistently yellow clypeus. Inspection of the genitalia showed no obvious differences between the two Ifrane specimens (Figs 100, 102), and neither specimen shows obvious differences from the genitalia of Spanish specimens (Fig. 104).
Aciandrena do not always show consistent clypeal colouration. In the recently described A. (Aciandrena) abruptifovea Wood, 2020, a series of six males from the same type locality site collected on the same day vary considerably in colouration from a clypeus with 80% yellow coverage, to 40%, to entirely black (Wood et al. 2020b). Given the extremely similar genitalia and the fact that both colour forms can be found together in both Iberia and North Africa, we synonymise A. astrella with A. fulica, with A. fulica taking priority as senior synonym as it was published first. Warncke, 1974 Figs 105-112 Material examined. Morocco: Drâa-Tafilalet, Errachidia, 11.iv.1995, 2♀, leg. Ma. Halada, one female deposited in the OÖLM, with one female retained in the personal collection of TJW.

Andrena (Nobandrena) ounifa
Distribution and remarks. Previously known only from the type locality in the western Algerian part of the Sahara Desert. The specimens from Errachidia are approximately 350 kilometres to the west. They agree with the male in size, the flattened and shiny clypeus (Figs 106, 108), the length of the propodeum (Figs 107, 111), and general structural characters. The female differs most strongly from classical Nobandrena (see Warncke 1968a) in the shape of the foveae which are long and very narrow, narrower than the width of an antenna (Fig. 106) and on the tergites it lacks the impressed centre line on T(2)3-4. In common with A. iliaca which was also placed in the  may be moved there in future, but without molecular data this move is premature, though it clearly does not belong in the Nobandrena. The holotype male is illustrated here for comparison (Figs 109-112). A single pollen load contained pure Raphanustype pollen (Brassicaceae).
Description. Female: Body length 7.5 mm (Fig. 105). Head: Black, clearly longer than wide (Fig. 106). Clypeus strongly flattened, strongly shining, weakly shagreened only on margins, moderately punctured, punctures separated by 2-3 puncture diameters. Process of labrum weakly trapezoidal, almost triangular, broad, twice as wide as long. Gena slightly narrower than width of compound eye with sparse, short white hair extending to vertex. Foveae long, narrow, narrower than width of scape, separated from inner margin of compound eye by less than their own width. Antenna dark, A4-12 extensively lightened orange below, A7-12 almost entirely orange, A3 slightly exceeding A4+5 in length. Ocelloccipital distance extremely short, linear, almost nonexistent. Mesosoma: Scutum strongly shagreened, dull to weakly shining centrally, strongly contrasting with shiny scutellum. Scutum and scutellum moderately punctured, punctures separated by 2-3 puncture diameters, with short whitish-brown pubescence. Episternum and propodeum strongly shagreened, dull. Dorsal area of propodeum longer than scutellum, propodeal triangle indicated by an increase in granular size of shagreenation (Fig. 107). Episternum and propodeum with white hair, longest attaining ¾ of length of the scape. Legs dark, tarsi lightened brown, pubescence whitish-brownish. Femoral and tibial scopa white. Wings hyaline, venation brown, stigma light brown. Nervulus interstitial. Metasoma: Tergites dark, margins clearly lightened yellow to dark orange, apically partly translucent (Fig. 108). Tergal discs microreticulate, dull. Tergal discs and margins with sparse white hair, T5 apically and T6 with short golden hairs flanking the pygidial plate. T5 basally with long white hairs, these overlaying but not obscuring the apical golden hairs. Pygidial plate narrow, centrally with a longitudinal slightly raised area.

Andrena (Poliandrena) guichardi Warncke, 1980
Distribution and remarks. Known only from south-western Morocco (Warncke 1980). Gusenleitner and Schwarz (2002) provided comments on the appearance of the male but this was not a formal description and so one is provided here. Andrena guichardi can be easily recognised within the Poliandrena by the extremely shiny and very sparsely punctate scutum (Fig. 115).
Other Distribution. Andrena alchata is known from Morocco and Algeria and was only described from the male sex (Warncke 1974); the subspecies A. doursana agadira was known only from south-western Morocco and was described only from the female sex (Warncke 1980), though the presence of putative males was noted.
Remarks. Sampling at both Laanoucer and Oueled Sghir in northern Morocco resulted in the capture of males of A. alchata and females corresponding to A. doursana agadira on the wing at the same time. Other Truncandrena species were present, with females of Andrena ferrugineicrus Dours, 1872 andA. schmiedekneckti Magretti, 1883 captured at Laanoucer and males and females of Andrena varia Pérez, 1895 at Oueled Sghir. However, no males or females of A. doursana citreola Warncke, 1975, the form occurring in this part of Morocco, were recorded. This situation raised the possibility that A. alchata and A. doursana agadira are actually synonymous.
Captured females are darker than A. d. citreola (compare Figs 117,118) and can be separated by the colouration of the hairs of T5-6 flanking the pygidial plate which are dark brown (Fig. 119) not light brown as in A. d. citreola (Fig. 120). The contrast of these hairs against the abdomen is greater in A. d. citreola as the underlying integument is lighter and weakly metallic green (see also Fig. 118) with overlying pale hair bands, whereas in captured females the integument is dark and the hair bands are darker (Fig. 119). The colouration of the hairs on the frons and along the inner margin of the compound eye are also important; these are a mixture of brown and black in captured females (Fig. 121), and white to light brown in A. d. citreola s.s. (Fig. 122). In these characters, captured females match A. d. agadira perfectly (Warncke 1980). Warncke did not describe male of A. d. agadira, commenting that they were indistinguishable from other A. doursana subspecies. In contrast, males of A. alchata are easy to distinguish as the white markings on the face of A. alchata are much more extensive, covering the clypeus and the lower paraocular areas (Fig. 123) whereas in A. d. citreola the single white marking is small and is restricted to the very tip of the clypeus which forms a clear raised protrusion (Fig. 124). There are also genitalia differences, in A. alchata the outer margin of the gonostyli is curved inwards (evenly rounded in A. d. citreola), and the penis valve is comparatively wider.
A female corresponding to A. d. agadira and a male A. alchata were selected from the same site (Laanoucer, 2.v.2018) for molecular investigation. A cox1 fragment of 263 base pairs was obtained after sequencing. A complete homology between the sequences of the female (GENBANK SUB7440720) and the A. alchata male (SUB7440720) was found, confirming their conspecificity. However, this result raises a difficult issue, as Warncke believed the males of A. d. agadira to be identical to other A. doursana subspecies. This therefore means that either A. alchata and A. d. agadira are not synonymous, and simply very similar morphologically in the female sex, or the undescribed A. d. agadira males are incorrectly associated with the type series females. Given this uncertainty, it is not appropriate to propose synonymy between A. d. agadira and A. alchata until genetic sequences can be obtained from the locus typicus to clarify sex associations in this region. What is clear is that captured females represent A. alchata, but given the large degree of variation in the colour of Truncandrena pubescence, it is possible that they simply resemble females of A. d. agadira. As such, we do not describe females here until this situation can be clarified further.

Other species newly recorded for the Moroccan fauna
Andrena (Aciandrena) pratincola Warncke, 1974 Material examined. Morocco: Guelmim-Oued Noun, 10 km E Guelmim, 15-16. iv.1995, 5♂, leg. Ma. Halada, OÖLM;Drâa-Tafilalet, 20 km W Boudnib, 9.iv.1995, 1♂, leg. Ma. Halada, OÖLM. Distribution and remarks. Previously recorded from Egypt and Tunisia (Gusenleitner and Schwarz 2002), this species has red tergites in the female sex but the male is fairly generic within the Aciandrena; it has an almost completely yellow and polished clypeus with loose white hair (Fig. 127), the tergites are dull, the nervulus is pronouncedly antefurcal, and the genitalia are simple (Fig. 128, also illustrated in Warncke 1974). Material from Morocco was entirely male and so comparison with the female type was not possible, but all other characters in the male sex were a good match (Figs 115, 126). These records considerably extend the range of A. pratincola to the west. The species may be found in the Algerian part of the Sahara but more survey effort is required.

Andrena (Aciandrena) varicornis Pérez, 1895
Material examined. Morocco: Oriental, 70 km S Oujda, 8.iv.1995, 1♂, leg. Ma. Halada, OÖLM. Distribution and remarks. Known from Algeria, Tunisia, Egypt, and Israel (Gusenleitner and Schwarz 2002;Pisanty et al. 2018), so the presence of this species in eastern Morocco is not unexpected. The male is unique within the Aciandrena in that the yellow markings on the clypeus extend on to the lower paraocular areas (Fig. 129), and additionally the tips of the gonostyli are produced into points (Fig. 130).

Andrena (Andrena) synadelpha Perkins, 1914
Material examined. Morocco: Fès-Meknès, Ifrane environs, 1700m, 10.v.1997South of Azrou, 25.iv.2017, 1♂, M. Snižek, both OÖLM. Distribution and remarks. This is a predominantly central and northern European bee, but there are isolated southern populations in upland areas of Spain, Portugal, and Turkey (Gusenleitner and Schwarz 2002;Wood et al. 2020a). Andrena synadelpha is polylectic, but it prefers pollen from trees and shrubs of broadleaf woodland such as Acer, Rubus, Crataegus, Quercus, Rhamnus, Frangula and Ilex (Wood and Roberts 2017). Its presence in the Middle Atlas is therefore unexpected, but it has precedence. Andrena haemorrhoa (Fabricius, 1781), another predominantly central and northern European bee which is able to live as far north as the Arctic circle, was also recently found in the Ouled Nail mountains of Algeria at an elevation of 761 meters (Cherair et al. 2013). It remains to be seen if there are more typically 'European' Andrena species that persist in the mountains of North Africa, presumably having been isolated at the end of the Ice Age as the lowland areas of North Africa returned to a more arid climate.

Distribution and remarks.
A West Mediterranean species, found in Algeria, Italy, Spain, and Tunisia (Gusenleitner and Schwarz 2002), and recently Portugal (Wood et al. 2020a). In Portugal, the bee has been recorded between January and March, and so it may have been missed by previous collectors in lowland areas of Morocco.

Andrena (Distandrena) merimna Saunders, 1908
Material examined. Morocco: Drâa-Tafilalet, Er-Rich, 1253meters, 28.ii.2019Fès-Meknès, Laanoucer, 1387meters, 11.v.2018 Distribution and remarks. Previously known only from Algeria and Tunisia (Gusenleitner and Schwarz 2002). At 9 mm in length, this one of the larger Distandrena species, almost as large as Andrena fria Warncke, 1975 which is restricted to Spain. Female Moroccan material is very slightly more striate centrally on the clypeus than comparative material from Algeria and Tunisia, but the male genitalia are a clear match.
Whilst overall the Moroccan material has more extensive reddish tergal margins and darker pubescence, the lack of any major structural differences mean that we consider this material to be conspecific with that from Israel, despite the large degree of geographic separation. This record extends the range of A. menahemella some 3,600 km to the west, giving a disjunct distribution of Morocco and Israel with no records from Algeria, Tunisia, Libya, or Egypt. However, this situation is not unprecedented amongst Andrena, with Andrena aegyptiaca Friese, 1899 showing a disjunct distribution being absent from much of the central part of North Africa (Gusenleitner and Schwarz 2002).
Floral preferences. As the two known specimens were caught in pan traps, no information on floral preferences is available. This reflects the situation in Israel, where females are known only from pan traps (Pisanty et al. 2016). The pollen preferences of Margandrena are incompletely known (Standfuss and Standfuss 2010;Gogala 2011). For species that fly in the winter through to the spring, they seem to be associated with monocotyledons such as Bellevalia (Asparagaceae), Colchium (Colchicaceae), and Crocus (Iridaceae) (Mavromoustakis 1952(Mavromoustakis , 1958Scheuchl and Gusenleitner 2009 (Warncke 1974) and is also known from southern Spain (Dardon et al. 2014). Warncke originally compared A. icterina to the central European Andrena strohmella Stöckhert, 1928 and suggested an association between the two species. In a revision of Micrandrena species from Iberia, Dardon et al. (2014) redescribed the type of A. icterina noting that the distinctive carina of A. strohmella found on the sides of tergite 1 is absent. However, in our material the carina is present, but weak. On examination of the type material we found that the carina is present, but equally weak when compared to A. strohmella.

Andrena (Pallandrena) byrsicola
The bee is similar to Andrena braunsiana Friese, 1887 which is found in central Europe eastwards to Greece, Turkey and the Caucasus (Gusenleitner and Schwarz 2002). However, both the scutum (Fig. 134) and the tergites (Fig. 135) are much less strongly punctured. In Schmiedeknecht's original description he wrote 'Abdomen nitidum, sparsim et subtilier punctlatum, depressionibus latis apicalibus fere laevibus' which corresponds very well to this material. In German, Schmiedeknecht draws parallels with members of the Chlorandrena in general impression, but notes the differences including the wide reddish margins of the tergites.
Confusion exists over this taxon because the location of the type of Schmiedeknecht collected from Tunisia is unclear and it may be lost (Gusenleitner and Schwarz 2002), and furthermore Warncke (1967) described Andrena oblita Warncke, 1967 from southern Italy (females, including holotype) and Tunisia (males). Grünwaldt (unpublished manuscript) examined the type series of A. oblita and compared it to ma- terial including a male of A. byrsicola collected by Schmiedeknecht from Tunis (nontype material). He found that Tunisian A. oblita males were identical to A. byrsicola, but that Italian females were identical to A. braunsiana females. At this moment, it is not possible to confirm these observations and to propose a formal synonymy, but both male and female Moroccan material is consistent with Schmiedeknecht's description of A. byrsicola, and this taxon is likely restricted to Morocco, Algeria, and Tunisia (Kuhlmann et al. 2020).
Floral preferences. There are no flower records associated with these specimens, but the unusual modified tibial scopa of the Pallandrena suggests some kind of floral specialisation. None of the specimens had full pollen loads, but fragments could be removed from four females (Fig. 136). Some samples were contaminated with Cistustype pollen, but the dominant pollen was Geraniaceae, probably Erodium (84.5%). Visual inspection of the scopa of specimens of A. (Pallandrena)  Distribution and remarks. This species was previously known only from a small region of Tunisia and eastern Algeria (Scheuchl et al. 2011). The discovery of this species in northern Morocco extends the range of this species approximately 1,500 km to the west. It is surprising that it has not previously been detected in Morocco because these records come from four quite geographically different provinces.
Other material examined.  Warncke, 1967 but has a slimmer body shape than the former and has hairbands on the metasoma in contrast to the latter, as well as more subtle morphological differences in the structure of the integument. Warncke, 1967 Material examined. Morocco: Oriental, 40 km south of Guercif, 15-17.v.1995, 100♀, leg. Ma. Halada, OÖLM;Fès-Meknès, Ifkern, 25 km E Boulemane, 24.v.1995, 2♀, leg. Ma. Halada, OÖLM Distribution and remarks. Previously known only from Spain (Gusenleitner and Schwarz 2002), a very long series of this species was collected in eastern Morocco. Though not previously recorded from North Africa, this material matches the holotype perfectly. In Iberia, A. relata is known from eastern and southern Spain including from the provinces of Almería and Murcia which are the driest and most desert-like habitat in Europe, so its presence in the similarly desertic regions south of Guercif and east of Boulemane is ecologically consistent.

Andrena (Ulandrena) speciosa Friese, 1899
Material examined. Morocco: Fès-Meknès, 20 km north of Missour, 14.v.1995, 1♀, leg. Ma. Halada, OÖLM. Distribution and remarks. Previously known from Syria, Jordan, and Israel through North Africa to Algeria (Gusenleitner and Schwarz 2002). There are only two Ulandrena in north-western Africa, and A. speciosa is instantly separable in the female sex from A. (Ulandrena) tadorna Warncke, 1974 as the former has a yellow clypeus and lower paraocular areas, whereas those of the latter are completely black.

Species removed from the Moroccan fauna
Andrena (Melandrena) nitida (Müller, 1776) Discussion. Andrena nitida was recorded from Morocco by Warncke (1974) under the name Andrena nitida mixtura Warncke, 1967. This subspecies was described from Portugal and Spain under the name Andrena limata Smith, 1853 mixtura (Warncke, 1967). However, Warncke transferred this subspecies into combination with A. nitida (Warncke 1974(Warncke , 1976 without any apparent justification. Examination of the type series of A. n. mixtura shows that its original placement was justified based on its bivoltine behaviour and the darker colouration of its tergal and scopal hairs (Wood et al. 2020a), and that therefore records of A. n. mixtura from North Africa (none of which are part of the type series) refer to A. limata and that A. nitida should be removed from the Moroccan list.
The distribution of sampling locations for species newly described for science ( Fig. 145) shows that most were collected in desert areas, mainly in the eastern provinces of Oriental and Drâa-Tafilalet, and in the southern provinces of Guelmim-Oued Noun and Souss-Massa (the Drâa and Souss valleys). Considering the history of bee recording in Morocco, these southern and eastern areas are the most difficult to reach, and would not have been visited by historical collectors and taxonomists in the 19 th and early 20 th century. The papers of Pérez (1895), Schmiedeknecht (1900), Saunders (1908), and Alfken (1914) focused primarily on the more coastal and northern areas of North Africa that naturally were easiest to reach and therefore be visited by European collectors. Though Warncke (1974Warncke ( , 1980 described species collected from the Drâa and Souss valleys (e.g., A. guichardi), these collections were clearly not exhaustive. Other recent collecting efforts show that these desert areas are highly likely to contain further undescribed species from other bee groups; for example, a new species of Osmia endemic to the Souss valley was just described, (Müller 2020) and other taxonomic work has revealed a major range expansion for the poorly known genus Borgatomelissa (Panurginiae) that was recently found in Morocco for the first time (Ortiz-Sánchez and Patiny 2019), being previously known from the Arabian Peninsula and East Africa to Mauritania across the Sahelo-Sudanian belt.
There are many outstanding problems in North African Andrena arising from the large number of subspecies described or erected by Warncke such as in A. pandosa Warncke, 1968 andA. medeninensis Pérez, 1895 that require focused taxonomic attention to resolve satisfactorily, and which are beyond the scope of this paper. Dealing with these species complexes should form the base for future taxonomic research on Andrena in this region.