﻿Nomenclature and typification in Verbascum (Scrophulariaceae) from North Africa

﻿Abstract The progress of taxonomic work on native Verbascum L. taxa found in Morocco led to a search for reference specimens in various herbaria. This process was extended to the taxa found in the other four countries of North Africa (Algeria, Tunisia, Libya, and Egypt), which make up the southern shore of the Mediterranean basin. Numerous names were identified as needing typification or requiring corrections of their earlier lectotypifications in order to stabilize their nomenclature and provide a better definition of each taxon. As a result, lectotypes are now designated for 35 names, a neotype is proposed for V.ballii (Batt.) Hub.-Mor., and second-step lectotypes are proposed for V.faureisubsp.acanthifolium (Pau) Benedí & J.M.Monts. and V.pinnatisectum (Batt.) Benedí. Comments have been added for each typified name. Known isolectotypes are also mentioned whenever possible. Furthermore, some new combinations are proposed in this paper, namely V.longirostrevar.antiatlantica (Emb.) Khamar, comb. nov., V.longirostrevar.atlantica (Maire) Khamar, comb. nov., and V.longirostrevar.hoggarica (Maire) Khamar, comb. nov.

In continental North Africa (Fig. 1), the genus is represented by 36 taxa belonging to 31 species (Dobignard and Chatelain 2013;Khamar et al. 2017Khamar et al. , 2022. Endemism rate is estimated to be 48% of all taxa. Verbascum species of North Africa grow in various habitats, i.e. steppes, forests, scrublands, lowland and high mountain pastures, rocky places, dry stony ravines and wadis, and can be found from the seashore up to high mountains (Jahandiez and Maire 1934;Quézel and Santa 1963;Pottier-Alapetite 1981;Qaiser 1982;Boulos 2002;Ibn Tattou 2007).
The present paper constitutes a preliminary step towards a taxonomic and nomenclatural revision of the taxa growing in North Africa. The purpose is to (1) contribute to nomenclatural stability of the species by clarifying the type citations, the designation of lectotypes, holotypes or neotypes when necessary, or the indication of previous typifications, and (2) provide an updated list of synonyms for each taxon.

Materials and methods
All the protologues of each validly published name for Verbascum taxa occurring in continental North Africa, as well as their synonyms, were consulted and critically reviewed. The biogeographic area was considered to extend from the southern shore of the Mediterranean Sea to the southern phytogeographic limit of the genus Verbascum, which is marked in southwestern Morocco by the presence of V. sinuatum L., V. tetrandrum Barr and Murb, V. maroccanum (Ball.) Huber-Morath and V. longirostre (Murb.) Huber-Morath; by V. dentifolium Delile and V. longirostre (Murb.) Huber-Morath in the Algerian Sahara (Hoggar in southern Algeria), by V. tibesticum (Quézel) Hub.-Mor. in the northern part of the Tibesti mountain range in Chad near the Chadian-Libyan border (Fig. 1).
Three key elements were taken into account to verify the original type plants (Turland et al. 2018): (1) the specimen characteristics that matched those in the original description, (2) the date and locality of collection mentioned in the protologue, and (3) all handwritten annotations on the labels.
Under the guidance of the International Code of Nomenclature for algae, fungi, and plants ( ICN;Turland et al. 2018) and statements recently suggested by McNeill (2014), the lectotype was selected among others according to its quality and its accordance with the description and data provided in the protologue. When duplicates were traced in other herbaria, these were designated as isolectotypes. When no appropriate material for use as a lectotype was located, a neotype was designated and its duplicates -if available -were designated as isoneotypes.

Lectotypifications
The taxa covered in this study are sorted in alphabetical order, according to their current accepted name (in bold), followed by the author citation, the bibliographic reference of the protologue or the nomenclatural recombination, the transcription of the original label of the specimen designated as type (lectotype, isolectotype, or neotype), and a barcode number following the herbarium acronym whenever available. Moreover, the homotypic and/or heterotypic synonyms for each name are quoted in chronological order. Comments have been added for each typified name.  (1889) based on material collected south of Oran, Algeria. In the protologue, the authors did not designate a holotype, nor did they provide data about the herbaria housing the original material. According to Stafleu andCowan (1976-1988), the plants collected by Battandier and Trabut are kept at Herb. MPU, but a number of duplicates can be found at Herbs P and RAR. We traced three specimens stored at Herb. MPU (MPU007787, MPU007788) and Herb. P (P00083087). The morphology of all three specimens agreed with the original description, and their locality was also in agreement with the locality data given in the protologue. It should also be noted that the specimens housed at Herb. MPU were mounted as two preparations for the same collection -sheets MPU007788 and MPU007787. Cross-labeling indicated that they were a single specimen (see Art. 8.3. of the ICN, Turland et al. 2018). The herbarium sheet P00083087 and those kept at Herb. MPU bore original labels handwritten by Battandier "Université d'Alger / Herbier de l'Afrique du Nord / Verbascum atlanticum Batt.! / Type! / Djebel Aïssa, rocailles grèseuses près de l'Aïn-Aïssa, 1600 m, / Leg J. A. Battandier". Referring to Art. 9.6 of the ICN (Turland et al. 2018), all these specimens should be considered as syntypes. We selected sheet P00083087 as the lectotype, because it was in a better condition than the other Herb. MPU specimens, and part of its features closely agreed with the original description.

Verbascum atlanticum
In his contributions for the Atlantic Flora, Battandier (1919) described another new species, V. pseudoblattaria, thirty years after the description of V. atlanticum. The original material of this new species, as referred to by Battandier (1919) in the protologue, was collected at Djebel-M'zi, south of Oran (Algeria). We traced two sheets at Herb. MPU (MPU006498, MPU006497), which are in complete agreement with the protologue and can be considered as original material. The sheet MPU006498 is selected here as the lectotype of the name V. pseudoblattaria, since it is in a better condition. (1) specimens MPU007789 and MP007790 were collected after the protologue was published, and (2) the collection locality indicated on the label of sheet MPU009629 did not match with the one cited in the protologue. Therefore, the "type" notation by Maire on Celsia ballii sheets was mistaken. Since no specimen from the original gathering in any institution was traced, a neotype should be designated according to Arts 9.8 of the ICN (Turland et al. 2018). We selected sheet MPU007789 stored at Herb. MPU as a neotype since it was collected and identified by Battandier and display all the morphological features described in the protologue.

Notes.
In the protologue of this species, Murbeck (1925) cited more than 30 gatherings, and he indicated the herbaria (B, WU, JE) that housed the specimens he examined. In the absence of any indication of a single specimen as the type, all the specimens cited in the protologue can be considered as syntypes according to Art. 9.6 of the ICN (Turland et al. 2018). Following the indications of Murbeck (1925), we only traced five sheets (JE00013700, JE00013701, JE00013702, JE00013703, JE00013725) at Herb. JE. These specimens completely agree with the protologue and can be safely considered as original material. However, as the other sheets stored at Herb. B mentioned by Murbeck were not found, we can assume that they were all destroyed by the fire following bombing by the Allied forces during World War II (see Hiepko, 1987). A detailed examination of the specimen and photos of JE00013725, presumably examined by Murbeck, matched with all the criteria and the description provided in the protologue, so we select it here as the lectotype of the name Verbascum battandieri. Notes. Original material is kept in the Linnaean Herbarium at the Linnean Society of London; an image of the lectotype is available at http://linnean-online.org/1836/.
(2) No specimens were cited in the protologue (Halácsy 1892) of Verbascum blattaria L. var. brevipedicellatum. However, according to Murbeck (1933) the description of this taxon was based on material collected by Sintenis and Bornmüller from 1883 to1892 in Greece. We have located five sheets that have been lodged at the following four Herbs: LD (LD1393506), JE (JE00012328), K (K000806399), and B (B100278360, B100278359), the labels' data of which correspond to those reported by Murbeck (1933). Hence, these specimens are eligible and can be considered as original material. The sheet LD1393506 preserved at Herb. LD is here selected as the lectotype of the name Verbascum blattaria L. var. brevipedicellatum.
(3) The name of Verbascum carduifolium was first published by Murbeck (1925: 168), but without a description. Four years later, Hayek (1929) gave the description of V. carduifolium Murb. ex. Hayek and mentioned a locality "The." [as an abbreviation of Thessalía]. Four years later, the taxonomic status of this species has been assessed by Murbeck (1933: 567) and he regarded it as a variety under Verbascum blattaria L. In addition to the morphological description, Murbeck (1933) cited an element collected by P. Sintenis viz. "In monte pindo Malakasi, in pratis supra pagum. determ. Leg. P. Sintenis 17 /6/ 1896, n. 632" and he indicated the herbaria (B, WU, LD) that housed the specimens he surveyed. We located six sheets, all of them from the locality given in the protologue, belonging to the Sintenis collections: B (Barcode B100278358), WU (WU0126534) and LD (LD1364911, LD1395023, LD1392908, LD1394623) at Herb. The sheet B100278358 conserved at Herb. B is here designated as lectotype for the name V. carduifolium.
(4) In the protologue of Verbascum blattaria var. gracilipes, Murbeck (1933) cited 11 gatherings, but he does not provide the name of the herbarium where the original material has been deposited. Six sheets from three different Herbs (B, JE, WU) were traced only: B (Barcode B100278357), JE (JE00012330, JE00012331) and WU (WU0126492, WU0126493, WU0126494). All of these specimens are consistent with the location and diagnosis given in the protologue and can be considered as original material. Since no specimen-type has been designated, all of these specimens can be considered syntypes under Art. 9.6 of the ICN (Turland et al. 2018). The specimen B100278357 preserved at Herb. P is selected here as the lectotype for the name of this variety.
(5) In describing Verbascum blattaria var. brachycalyx, Murbeck (1933) referred to eight collections made in five different regions in Greece, but he did not designate a single type for the name. We traced eight specimens in Herbs. B (B100278364, B100278363, B100278367, B100278366), S (S10-26942, S10-26943, S12-12718) and JE (JE00012329) which can be considered original material. We here designate the specimen B100278364 preserved at Herb. B as the lectotype of the name V. blattaria var. brachycalyx since it is well preserved. Notes. Original material is kept in the Linnaean Herbarium at the Linnean Society of London; an image of the lectotype is available at http://linneanonline.org/1832/.

Verbascum boerhavii
(2) Maire and Murbeck (in Murbeck 1927) described V. hookerianum var. pseudocalycinum on the basis of specimens collected from Reraya region in the Great Atlas, Morocco. However, the protologue does not give the name of the herbarium where the original material was deposited. The search for type material brought us to four specimens housed at Herb. RAB (RAB030509) and at Herb. MPU (MPU010269; MPU010270, MPU010268). The sheet RAB030509 stored at herb RAB is here selected as the lectotype for the name V. hookerianum var. pseudocalycinum. Notes. In the protologue of Verbascum calycinum, Ball (1875) cited one gathering and he mentioned the collection locality reading "Legimus in regione inferiori Atlantis Majoris, convalle Ait Mesanet prope Sketana". However, he did not quote any herbarium that houses the original material or accession numbers specified, nor the date of collection. According to Stafleu andCowan (1976-1988), the main collections of Ball are preserved at Herbs E and K, and also some at Herbs OXF and GL. During searches of Ball's collections from Morocco, three specimens were traced. One of them is kept at Herb. P (P03420188) and another is stored at Herb. K (K000410937) and is mounted on the same sheet with K000410936. Both specimens K000410937 and P03420188 bear a label reading: "Iter Maroccanum. 1871 / Verbascum calycinum, Ball / Ex regione inferiori Atlantis Majoris, prope Sketana, alt. 1300-1400 met / Majo 18-19 / J. A. ball /". However, the specimen K000410936 bears a label reading "Ex rupibus arenaceis Atlantis Majoris in convalle Ait Mesan, alt. 1400-2000 met / Majo 13-16 / J. A. ball /". All of those sheets were collected by Ball in the High Atlas, Morocco. As the collection locality of both specimens K000410937 and P03420188 agree closely with the protologue, they can be safely considered as original material under Art. 9.4. of the ICN (Turland et al. 2018 Among the original material of the genus Celsia, preserved in Linnaean herbaria, three sheets relevant to Celsia cretica were found: the first (Herb. Linn. No. 774.3, image available at: http://linnean-online.org/7174/) is annotated at the bottom "celsia cretica" and "774.3" at above the right hand of the sheet by Linnaeus. The second specimen (Herb. Linn. No. 774.4, image available at: http://linnean-online.org/7175/) is annotated "cretica" at the base and "774.4" at above the right corner of the sheet. Both specimens are clearly showing the characters mentioned in the Linnaeus [son] (1781) diagnosis. After careful examination of the available collections and consideration of all elements in the protologue, the sheet No. 774.3 is the most complete and well conserved, and it is designated here as the lectotype of the name Verbascum creticum. However, the sheet No. 774.4 is selected here as the isolectotype.

Verbascum calycinum
(2) In the protologue of Celsia cretica f. pallenscens, Maire (1931) provided the following locality "Bòne, pentes du Mont Edough, 300-700 m" but they did not specify the name of the herbarium where the type material has been stored. We traced three sheets (MPU002670, MPU002671), preserved at Her. MPU which is in complete accord with the protologue. The specimen MPU002670 is here chosen as the lectotype for the name Celsia cretica f. pallenscens. ( Notes. This taxon was firstly published by Maire and Murberk (1923) as Celsia sinuata Cav. var. demnatensis Maire & Murb., and they referred to a plant collected by R. Maire in 1921 near Demnat, Morocco (Maire, 1938). In the protologue, the locality, collector, and collection date were indicated as follows: "Rocailles calcaires au N. d'El-Arba, près Demnat, Dr R. Maire, 3 avril 1921". As mentioned above, the main Réne Maire herbarium of plants from North Africa is now in Herb. MPU (Stafleu andCowan 1976-1988 Notes. Verbascum dentifolium was described by Delile (1836: 28;1837: 287) based on material growing in the Montpellier botanical garden from collections made in Port-Juvenal at Montpellier. Port-Juvenal was a place where lots of wool bales were coming to France from the Middle East, introducing many species from seeds trapped in the wool. In his monograph of the genus Verbascum, Murbeck (1933) discussed in some detail the distribution and synonyms of a species to which this name has generally been applied, but he did not designate a lectotype for the name V. dentifolium. According to Murbeck (1933) the original material is kept at herbarium of Montpellier (MPU), in France. Research in Herb. MPU enabled us to locate two specimens bearing Delile's handwritten labels and agreeing closely with the protologue's description. The first one is mounted on two sheets. These two preparations, bearing the Herb. MPU barcodes MPU020144 and MPU020145, have a label reading "Verbascum, h. m. Julio 1826, N:10" that is accompanied by two other labels in which Delile gave a description of the species. According to Art. 8.3 of the ICN (Turland et al. 2018), the sheets (MPU020144 and MPU020145) must be considered as a single specimen. The second specimen (MPU020146) bears a label reading "Verbascum dentifolium, h. m. Julio 1831". Since the type is not specified, the two specimens are to be considered as syntypes according to Art. 9.6 of the ICN (Turland et al. 2018). The sheet corresponding to Herb. MPU barcodes MPU020144 and MPU020145 and with the collection date "Julio 1826" is selected here as the lectotype for the name Verbascum dentifolium.  (Benedí and Montserrat 1998). When the genus Celsia was subsumed in the genus Verbascum the epithet sinuata could not be used for this species as it already exists in the combination Verbascum sinuatum L. for another species. We concur with the conclusion of Benedí and Montserrat (1985) that the correct name for this plant in question is V. erosum Cav. Notes. Murbeck (1921) described Celsia faurei based on collections done by A. Faure in western Algeria. The author in the protologue includes an explicit reference to the locality (Algeria, Oran: Between Imbert River and the railway embankment), with several years for the collection (1911, 1916, 1918, and 1921) and the herbarium where the original material are kept (LD). Benedí and Montserrat (1997)  Notes. Font Quer (1927) based his species on a collection of plants from the Rif region that he has gathered during his 1927 campaign in Morocco. Font Quer gave the following diagnosis "Folia Acanthi mollis L., superiora oblonga, sinuato-dentata, floribus longe pedunculatis, pedicellis glanduliferis, racemosis; bractae lanceolatae brevissimae, calycis laciniae lanceolatae, margine integro, capsula globosa, obtusa calyce triplo longior" with the collection locality "hab. in saxosis, pr. Badù (Atlante rhiphaeo), 1500 m. alt., solo schistoso, 6 julii". This protologue has been validly published by Gonzâlez-Bueno (1988). According to Nualart (2017), Benedí and Montserrat (1997)   Notes. Benedí (2003) indicated as a lectotype the Desfontaines collection kept at Herb. P [P-DESF] but he omitted to mention another sheet of this collection kept at Herb. G (image available at https://plants.jstor.org/stable/viewer/10.5555/al.ap.specimen. g00439692). According to Lasègue (1845) and Stafleu andCowan (1976-1988), around 600 Flora Atlantica plants was given by Desfontaines to Lemonnier and was acquired by Delessert with the Lemonnier herbarium at Herb. G. In Herb. G we have traced one sheet, with barcode G00439692, that is stamped (Typus). This sample matches the collection locality in the protologue and it is the one that morphologically agrees best with the original description. This specimen bears two labels. The first has "Celsia betonicifolia Desf., Fl. Atl. [Desfontaines' handwriting]"; the second is a printed label including, name, locality, reference, and a short historical French comment, and is annotated "R-L Desfontaines, Herbier de Barbarie / Celsia betonicifolia Desf. Loc: in arvis ìnculus Algeriae / Desf.  Notes. Maire (1918) published the name Verbascum thapsus subsp. gaetulum based on a collection made by himself at the foothills of Araïra Mountain, Morocco. He also observed that this plant was very common in this locality. In the protologue, the author indicated that this plant has an intermediate morphology between V. thapsus L. and V. simplex Hoffmanns. and Link, but differed from both by its spatulate stigma and its tetrandrous or sub-tetrandrous flowers. Murbeck (1927) raised V. thapsus subsp. gaetulum to species rank. However, a few years later, Maire (in Jahandiez and Maire 1934) classified this taxon with the rank of variety in V. simplex, but without any justification for this new recombination. According to African Plant Database (APD) (2022), the accepted name should be V. gaetulum (Maire) Murb. despite Ibn Tattou (2007) also considering V. thapsus subsp. gaetulum an accepted taxon.

Verbascum erosum
During our research, we have found two sheets (MPU000364, MPU000365) corresponding to the original material of this taxon preserved at Herb. MPU, where the main collection of Maire is kept. According to Stafleu andCowan (1976-1988), if the original material of Dr. R. Maire are preserved at Herb. MPU, there are important sets of duplicates at Herbs Al, CAIM, P, and RAB. Since no holotype was specified, the cited specimens are all syntypes under the Art. 9.6 of the ICN (Turland et al. 2018). Therefore, the specimen MPU000364 is selected here as the lectotype, because it is the best original material, more complete and informative than the specimen MPU000365, even if this latter is accompanied by a handwritten detailed morphological description by Maire.  (1878) based on a sample collected by Hooker in the district of Ourika at the foothills of the High Atlas Mountains in Morocco. Ball (1878) in his protologue did not use the term type or mention the herbarium that houses the type specimens. However, in his monograph of genus Verbascum, Murbeck (1933) indicated that the original material was kept at Herb. K. During our extensive research, we have traced only one sheet (K000410848) that is totally in agreement with the protologue. This sheet bears a label reading "Verbascum hookerianum Nob., South Morocco Greater Atlas, Coll. Dr. Hooker, May 1871" and includes a part of the inflorescence with basal leaf. Since the Herb. K has a single specimen of the Hooker collection, one may argue that the relevant specimen is the holotype. Nevertheless, Ball did not use the term type or mention the name of the herbarium housing the type. Therefore, we here designate the sheet K000410848 as lectotype of the name.

Verbascum hookerianum
(2) The protologue of V. tagadirtense comprises a complete description in Latin, followed by the provenance "Région du Grand Atlas-. Pentes broussailleuses à Tagadirt N'Bourd, c. 900 m", but no indication about the name of the herbarium where the type is preserved (Murbeck 1923). Ten years later, Murbeck (1933), in his monograph of the genus Verbascum, indicated that the original material can be found at Herb. LD. During the course of the pursuit, two sheets were traced at Herb. LD which were in accordance with the protologue. The specimen LD1216036 is here selected as a lectotype of V. tagadirtense. Notes. Ascherson (in Ascherson and Schweinfurth 1887) described Verbascum letourneuxii on the basis of specimens separately collected by Ehrenberg and A. Letourneux from the Alexandria region in the north of Egypt. Within the protologue Ascherson and Schweinfurth (1887) confess that the specimen brought by Ehrenberg only constitutes a skeleton of a plant in fruit, while specimens collected by Letourneux, from a few kilometers from the collection locality of Ehrenberg, are well-developed specimens. However, the authors did not indicate the name of the herbarium housing the type specimen. According to Stafleu andCowan (1976-1988), the original material of Ehrenberg, Letourneux, Ascherson and Schweinfurth were kept at Herbs B, C, G, K, L, LE, P, S, and W. Nevertheless, the Berlin (B) herbarium was bombed during World War II on the night of 1-2 March 1943 (Hiepko 1987); hence, a good part of the authors' collection has been lost (Stafleu andCowan 1976-1988).

Verbascum letourneuxii
Based on the type specification given in the protologue (locality, collector, and collection date), eleven sheets were traced in different herbaria belonging to the Letourneux collections: W (W1889-0043225), G (G00015111, G00015112, G00015113, G00015114), S (S10-27120), K (K000975868), and P (P03417358, P03417357, P03417360, P03417361). Since the type has not been specified, all of the specimens are to be recognized as syntypes according to Art. 9.6 of the ICN (Turland et al. 2018). The sheet G00015113 preserved at Herb. G is here designated as lectotype for the name V. letourneuxii, since it is the specimen that shows the best quality of preservation of the important diagnostic features. Notes. Murbeck (1925) described this species based on a collection made by A.S. Mardochée, a Moroccan plant specimen collector on account of French botanist Ernest Cosson. In the protologue, Murbeck (1925) Stafleu andCowan (1976-1988), the main collections of Cosson are preserved at Herb. P. During our searches we have traced ten specimens collected by Mardochée housed at Herb. P (P03425553, P03425555, P03425557, P03425558, P03425559, P03425560, P03425563, P03425565) and two at Herb. K (K000410860, K000410861). Detailed examination of all of them revealed that the plants and the information on the labels of the samples match well with the protologue description. Since no holotype is indicated in the protologue the ten specimens traced are all syntypes according to Art. 9.6 of the ICN (Turland et al. 2018). Therefore, we select here as the lectotype the sheet P03425558 because it is more complete and agrees best with the original description. This specimen bears two handwritten labels, one by Cosson: "Celsia maroccana Coss. sp. nov.! Oudjan, Sud-Ouest du Maroc. Mardochée 1875" and the other bears a 1924 annotation by Murbeck who correctly identified it as C. longirostris Murb.
Moreover, as mentioned by Murbeck (1925), Cosson (in sched.) have described this plant in question as a new species beneath the name "Ceslia maroccana" accompanied by a handwritten diagnosis written by him on 24 August 1876. However, rendering to Murbeck (1925), Cosson has misinterpreted the plant in question as a new species because at the time mentioned he did not seem to have known that a related species was published a year earlier by Ball (1875: 172) (Maire 1940), the protologue listed nine collections, but no particular herbarium sheet was chosen as the holotype. We have traced eleven sheets (MPU004304, MPU004305, MPU004306, MPU004307, MPU004308, MPU059306, MPU059318, MPU059319, MPU059320, MPU059321, and MPU05932) preserved at Herb MPU that fit the protologue perfectly. The specimen MPU004308 appears to be the best preserved one and it shows more of the diagnostic features that are described in the protologue Therefore, it is the best candidate for typification and is chosen here as the lectotype for the name C. longirostris var. hoggarica according to Arts. 9.3 and 9.4 of the ICN (Turland et al. 2018 Notes. Original material is conserved in the Clifford Herbarium at the Natural History Museum of London and an image of the lectotype is available at http://data. nhm.ac.uk/dataset/clifford-herbarium. (2) Maire (1924) described Verbascum lychnitis var. giganteum on the basis of plant material that he had collected himself in the Azrou cedar forest, Central Middle Atlas, Morocco. However, in the protologue, he did not identify any herbarium sheet as holotype, nor did he give the name of the herbarium where the original material was stored. Six specimens were located through our research at Herbrs RAB (RAB30546), MPU (MPU006910, MPU006911, MPU006912, MPU006913), and P (P00083084) which can be considered original material. The sheet RAB30546 preserved at Herb. RAB is selected here as the lectotype for the name V. lychnitis var. giganteum since it is in a better condition. ≡ Celsia mairei Murb., Lunds Univ. Arsskrift, 2 n. f., 35 (1) Murbeck (1939) cited two gatherings: "Imper. Maroccanum: In rupestribus calcareis Atlantis Medii ad Tizi-n-Ouria (supra Ksiba), 1600 m, R. Maire, 21/6/1936" and "Tafriat, 3000 ft., some 50 or 60 km out of Marrakesh on the road to Ouarzazat across the Great Atlas by Tadert and the Tizi n' Tichka, E.K. Balls, 10 /7/ 1936, n. 2792". However, he did not give any information about herbaria housing the original samples, nor did he indicate which one is the holotype. The search for the original material of Celsia mairei led us to discover three sheets kept at Herbs MPU, RAB, and S. The sheet kept at Herb. MPU was collected by R. Maire (Barcode MPU059317), but those found at Herb. S (Barcode S10-270499) and at Herb. RAB (RAB30643) were collected by E.K. Balls. All three specimens found are morphologically in agreement with the original description (Murbeck, 1939). Moreover, collection locations indicated on the labels match those indicated in the protologue. Following Art. 9.6 of the ICN (Turland et al. 2018), these specimens must be considered as syntypes. According to its quality of conservation, the sheet S10-27049 preserved at Herb. S is here selected as the lectotype, because it is well-conserved and shows more diagnostic features described in the protologue. The specimen housed at Herb. RAB is recognized here as an isolectotype. Notes. This species was described by Ball (1875: 172) from material gathered by himself, J.D. Hooker, and G. Maw during their botanical expedition in Morocco in 1871 (Hooker and Ball 1878;Stafleu andCowan 1976-1988). Murbeck (1925) indicated that the original material is stored at Herbs K and B. During our research, we have traced five specimens kept at Herb. K and bearing K barcodes as follows: K000410853, K000410855, K000410856, K000410857, K000410858. However, no relevant specimens could be traced at Herb. B, so we have some reasons to think that this material was destroyed during the bombing raid at Berlin in 1943 (Hiepko, 1987). In all the specimens found the collecting localities on labels match those mentioned in the protologue and material morphologically agrees with the original description. Since no holotype was indicated, they are all syntypes according to Art. 9.6 of the ICN (Turland et al. 2018). Hence, among all the original material kept at Herb. K, the specimen bearing the K barcode K000410855 is selected here as the lectotype. Notes. The name of this species was first published by Mas Guindal (1928: 102), but without a description. A year later, Pau (1929) gave the description of Celsia masguindalii Pau based on the specimens collected by Mas Guindal in Río Martín near Tetuán. In their 'Taxonomic and nomenclatural notes on some species of the genus Verbascum L. (incl. Celsia L.)', Benedí and Montserrat (1985) have wrongly cited the lectotype number (MA108989) for Verbascum masguindalii (Pau) Benedí & J.M.Monts. instead of the number MA108916 as shown by Nualart (2017). Besides, Nualart et al. (2021) have traced two other eligible sheets (BC89918, BC141503) in Herb. BC. Those two specimens were also collected by Mas Guindal in Río Martín, but without a collection date on the labels. According to Gonzâlez-Bueno and Gomis (2007) and Nualart (2017), Mas Guindal during his visit in Tetouan (Morocco) between 1926 and 1931, has collected many times at this site (Río Martín). Notes. Murbeck (1933), in his monograph of Verbascum L., indicated that original material of Verbascum maurum was kept at Maire's herbarium at Herb. MPU. We have only found two sheets of this taxon in Herb. MPU (MPU008015, MPU008183). Both have a handwritten label by Maire bearing the information about the gathering, collector, date, and locality that match the protologue. The specimen MPU008183 is designated here as the lectotype, because it shows more of the diagnostic features described in the protologue. Notes. Pursuant to the Art. 9.17 (Turland et al. 2018), the lectotypification of Benedí (2003) could be considered a first-step lectotypification, since he has cited wrongly the number of the lectotype and did not specify exactly which of the two sheets MPU005282 and MPU006805 kept at Herb. MPU as lectotype. In addition to these two sheets stored at Herb. MPU, we have traced another of this gathering kept at Herb. P (barcode P03425758). Hence, the sheet preserved at Herb. MPU that bears barcode "MPU006805" is here designated as the (second-step) lectotype. Notes. According to Stafleu andCowan (1976-1988), the original material of the Italian botanist, Michele Tenore are preserved at Herb. NAP, with further material in Herbs AWH, B-Willd., BASSA, BM, BR, C, CGE, DWG, E, FI, G, H, K, M, MPU, OXF, P, PH, REG, and UPS. However, we could not trace any specimen in any herbarium for this taxon. Notes. This taxon was established by Murbeck (1933) under the recombination V. rotundifolium subsp. eu-rotundifolium. According to Ferguson (1972) and Benedí (2009), this subspecies is distributed in southern Italy, Sicily, Algeria, and Tunisia. However, for Fennane and Ibn Tattou (2005), Ibn Tattou (2007), and Dobignard and Chatelain (2013), this taxon is spontaneous in Morocco. Under its current recombination, V. rotundifolium subsp. rotundifolium it is accepted in the Partical flora of Morocco Ibn Tattou 2005, Ibn Tattou, 2007), in the index of synonyms for the flora of North Africa (Dobignard & Chatelain, 2013), and by the APD (African Plant Database 2022).

Verbascum pseudocreticum
(2) Pomel (1874) described on Verbascum numidicum based on the collections of Choulette exs. [exsiccata] 369, but did not provide collection date nor the name of the herbarium where the original material were housed. After a thorough search, two specimens were found, one at Herb. MPU (MPU004892) and the other at Herb. JE (JE00013694), both of which matched the data provided in the protologue. Thus, the sheet MPU004892 is chosen here as the lectotype of the name V. numidicum, because it is better preserved and shows more diagnostic features described in the protologue.
(3) The protologue of Verbascum rotundifolium var. castellorum (Maire 1938) is composed of a Latin diagnosis, and the details of the collection locality on limestone rocks below Ksiba region at an altitude of about 900 m in the Middle Atlas, Morocco, but there is no mention of the name of the herbarium that houses the original material. Four herbarium sheets deposited in Herbs MPU (MPU004005, MPU004006) and P (P01167373, P00083085) were traced. These specimens can be considered as the original material according to the type details provided in the protologue. The specimen P00083085 persevered at Herb. P is better conserved and best represents the diagnostic features of the taxon given in the protologue; hence it is selected here as the lectotype for the name V. rotundifolium var. castellorum. (2) Coincy (1895) in the protologue of Verbascum aurantiacum mentioned the collection locality as "Les pentes de la Sierra de Aquila près Baza (prov. de Grenade) le 8 juin 1895, à une hauteur que j'évalue à 1,200 m. environ [The slopes of the Sierra de Aquila near Baza, in the province of Granada, on June 8, 1895, at an altitude that I estimate roughly 1200 m]". However, no specific herbarium specimen was identified as a holotype, nor was where the original material was housed indicated. According to Stafleu andCowan (1976-1988), the plants collected by Coincy are preserved at Herb. P, and further material can be found at Herbs B and LY. After conducting exhaustive research in the three different Herbrs (P, B and LY), we located one sheet at Herb. P (P03808542) that is a perfect match with the protologue. Hence, we designate this specimen (P03808542) as the lectotype for the name Verbascum aurantiacum.

Verbascum rotundifolium
(3) When he described Verbascum rotundifolium subsp. castellanum, Murbeck (1933) included in the protologue 14 gatherings that were collected in Spain by diverse botanists, but he did not indicate any herbarium specimen as a holotype. Only four specimens related to the gatherings mentioned in the protologue have been traced. Three of these specimens were found at Herb. MA (MA108791, MA108794, MA108789), and the other specimen was found at Herb. JE (MA108792). (JE00007506). The specimen MA108794 preserved Herb. MA is here designated as lectotype for the name V. rotundifolium subsp. castellanum because it is complete and matches the information in the protologue. Notes. Barratte and Murbeck's (in Murbeck 1905) description of Verbascum tetrandrum was based on specimens gathered by Rabbi Mardochée near Tazeroualt in the southwest of Morocco. Within the protologue, Barratte and Murbeck (in Murbeck 1905) did not designate any herbarium sheet as holotype, nor the name of the herbarium where the type material was housed. Moreover, the original description was accompanied by a photography of the type with a legend that reads "1-Specimen fructiferum ad Tazeroualt lectum / 2 & 3-Pars caulis foliumque basilare speciminis in Gallia anno 1877 a cl. Cosson culti / 4-Corollae; speciminis culti". Following this information mentioned in the legend, the type material is formed by the specimens bearing fruits collected by Mardochée and specimens from cultivated grains by Cosson. Murbeck (1933) indicates that the original material was preserved in the Cosson herbarium. During our extensive search, two sheets were found at Herb. P (P03429046, P03429051). All of these specimens agree with the protologue and the accompanied iconography.

Verbascum sinuatum
The sheet P03429046 contains a fruiting part of an inflorescence and has three labels: an original label handwritten by Cosson "Verbascum pycnostachyum nov. sp. / Tazeroualt / Sous independant (Maroc) Robbin Mardochée"; the second is a revision label handwritten by R. Maire in 1921 who re-determined it as "Verbascusm tetrandrum Barr & Murb", and the last label is handwritten by Murbeck in 1926 who put just an exclamation mark (!) indicating that the sheet was verified by him. The sheet P03429051 has an apical part of stem and a basal leaf and two handwritten labels: the first one reads "Verbascum (Celsia) pycnostachyum / Cult. Thurelles / 1877", and the second is a revision label handwritten by Murbeck in 1924 who re-determined it as "Verbascusm tetrandrum Barr & Murb". All of these specimens should be considered as syntypes as stated by Art. 9.6 of the ICN (Turland et al. 2018). The specimen with barcode P03429051 is here chosen as the lectotype, because it is well preserved and it is the one that shows most of the morphological features in agreement with the original description. ≡ Celsia tibestica Quézel, in Bull. Soc. Hist. Nat. Afrique N. 48: 95. 1957 Notes. This species was described by Boissier (1853) from material collected at the foot of Trobol mountain in Syria. In the protologue, Boissier (1853) gave information about the collection locality, but he does not provide the name of the herbarium where the original material has been deposited. Boissier (1853) notes that V. tripolitanum starts flowering at the beginning of June. According to Stafleu andCowan (1976-1988) and Jacquemoud (2011), the Boissier herbarium related to the Flora Orientalis account is preserved at Herb. G. However, duplicates of Boissier's collections could be found in many other herbaria. During our research, we have traced one specimen (barcode P03287124) conserved in Herb. P that is in accordance with the protologue. The sheet P03287124 bears a part of an inflorescence, a basal leaf, and two handwritten labels, the left one reads "Verbascum tripolitanum Boiss /Syria Mar: / [Jul] 1846 / Tripoli / (donné par Mr. Boissier / Ed Boiss.)" and the second label on the right corner reads "Herbier Mus. Paris/ herbier tripolinatum / boiss / Syria /". So this sheet is designated here as the lectotype. ≡ Celsia zaianensis Murb., in Lunds Univ. Arsskrift, 2 n.f., 22 (1)

Notes.
In the protologue of Celsia zaianensis, Murbeck (1923) cited an element collected by Benoist from Zaïan region, between Aït Lais and Ain-leuh, Morocco, and he indicated that type material is preserved at the Cosson Herbarium. The latter collection is housed at Herb. P according to Stafleu andCowan (1976-1988). The protologue also includes citation of the locality "Entre Aït Lias et Aïn Leuh", collection date "5 Juin 1918", and name and number of collector "P. Benoist, 525". At Herb. P, we have found only one specimen with barcode P03425567 bearing a single plant, and determined by Murbeck as "Celsia zaianensis Murb.". This specimen is morphologically close to the original description. Furthermore, there is an indication on the label indicating a new species ("n. sp.") and the label has the collection date and collection locality, and collector number matching the information given in the protologue. Since only a single specimen is mentioned by Murbeck (1925), one may argue that the relevant specimen is the holotype. However, Murbeck (1925) did not use the term type. Therefore, we herein designate the sheet P03425567 as lectotype.