Typification of names of plants described by Carlos Pau from Morocco (1908-1922)

The vascular plant names proposed by Carlos Pau from Morocco between 1908 and 1922 were studied and, when necessary and possible, typified. Ten lectotypes have been designated; most of them preserved in MA and BC herbaria (isolectotypes at BCN, G, LD, and P).


INTRODUCTION
Carlos Pau (1857Pau ( -1937) ) was one of the first Spanish botanists to study the flora of Northern Morocco (see González-Bueno & Gomis, 2005).His research began in 1908, when he examined some plants that he had received from other naturalists such as Benito Vicioso.During the next 28 years, and until his death, he proposed about 400 new taxa from this area, including species, subspecies, varieties and forms (90% validly published).Although Carrasco (1975) published an exhaustive list of taxa described by Pau, it is not complete and, moreover, most of Pau's names are still untypified.

MATERIAL AND METHODS
This work is based on the analysis of relevant literature (including protologues), and examination of herbarium specimens.The main source of original material for the names proposed by Pau is the herbarium of the Royal Botanic Gardens of Madrid (MA), where his collection (about 100,000 specimens) was deposited a few years after his death (see Mateo, 1995).Other original material can be found at the herbarium of the Botanical Institute of Barcelona (BC) owing to the relationship between Pau and Pius Font Quer (curator of the herbarium at that time) and with Fr.Sennen (see Mateo, 1995).Duplicates of Pau's collection were found at BCN, G, LD, MPU, and P (herbaria acronyms according to Thiers, 2017).Wherever possible, we have chosen to typify material of the Pau collection deposited in MA; whenever there was more than one specimen, we chose the most complete and well preserved, in accordance with the description given in the protologue.To evaluate the synonymy of these names, we have compared the original material with the type specimens of the accepted names, revised the original description of these accepted names and checked recent North African floras.Names typified here are listed chronologically.In some cases (Art.9.19 of ICN;McNeill & al., 2012), we have proposed that some previously designated types be superseded because they are based on specimens from the exsiccata "ITER MAROCCANUM" (1927)(1928)(1929)(1930)(1931)(1932) that were collected after the names had been published.

Convolvulus suffruticosus var. melliflorus
Carine & Robba (2010) reported a "holotype" for this taxon cited as "Convolvulus valentinus var.melliflorus Pau", but there are two sheets at MA collected by Pau in Selouane that are syntypes (Art.9.5), while Carine & Robba (2010) did not mention any particular sheet.MA No. 94100 has two original handwritten labels by Pau, one with the information about the gathering and another reporting the diagnosis.MA No. 94102 is labelled as "B.et C. Vicioso Herbarium Mauritanicum" but the gathering information is the same.The specimen here chosen as lectotype is the one labelled by Pau that bears a better and more complete plant.Dobignard & Chatelain (2010-2013) included this variety in Convolvulus valentinus subsp.suffruticosus (Desf.)Maire, and the MA specimens were redetermined as C. valentinus Cav. by Menemen in 1994-95.However, the molecular and morphological study by Carine & Robba (2010)  Pérez Camarero, durante la primavera pasada".Font Quer sent some of these plants to Pau for their determination because at that time he did not have African specimens for comparison (see González-Bueno & Gomis, 2007).In this paper all plants are numbered.Five new taxa are described, of which four are here typified.

Malcolmia patula var. longifolia
Number 23 is a new variety of Malcolmia patula Lag.ex DC. with the diagnosis "Folia siliqui et rostra longiores" made by Pau. Al-Shehbaz & al. (2014) mentioned a BC specimen as holotype (BC barcode BC-77766), and a MPU specimen as isotype, both collected in Larache in 1930 and included in the exsiccata "ITER MAROCCANUM, 1930"; thus they were collected 16 years after the description of the taxon in 1914.According to Art. 9.19 of ICN (McNeill & al., 2012) this typification has to be superseded because original material exists: the plants collected by Pérez Camarero in 1914.There are three sheets of this gathering, one at MA and two at BC, one conserved in the general collection and the other in the personal collection of Sennen.The BC specimens are labelled as "HERBARIO DE P. FONT QUER" with an indication of the taxon name and the protologue handwritten by Font Quer.The MA specimen has an original label also handwritten by Font Quer with the locality, collector, date and number, but the taxon name is "Malcolmia patula (Lag.)DC.?".The three sheets bear complete plants and we select as the lectotype that preserved at MA where the main collection of Pau is conserved.
All three specimens have a revision label by Nogueira in 1989 who re-determined them as Malcolmia lacera (L.) DC.López González (1986) demonstrated that the designated type of this taxon is a specimen of Raphanus L. and not Malcolmia L.; so, this is a misapplied name and, according to Al-Shehbaz & al. (2014), all the records of M. lacera from the Iberian Peninsula and north-western Africa have to be attributed to M. triloba (L.) Spreng.There are two sheets preserved at BC and labelled as N. perezii, both bearing the herbarium collection label of Font Quer.One sheet is conserved in the general collection and it also includes another label with the indication handwritten by Font Quer "Nonnea phaneranthera Viv.? No conozco la descripción de esta especie.Sin escamas en el tubo".The other is conserved in the Sennen collection (BC).In both of these sheets, the gathering information reported, as well as the plant, match the protologue.The specimen chosen as lectotype is the more complete plant.
On the other hand, all these authors recognize two species on the basis of the stamens length: N. calycina and N. heterostemon Murb.(a Moroccan endemism).Contrarily, Valdés (2011) and Valdés & Talavera (2012) only accept N. calycina due to its heterostyly.Pau (1914) did not mention the stamens length in his description of N. perezii, but examining the type specimens, we have observed that one stamen is much longer than the others.Murbeck (1898)

Ononis maweana var. fontqueri
Number 15 is a new variety of Ononis maweana Ball that, according to Pau (see Font Quer, 2014), would differ from the nominal variety by the leaves and calyx.We have found two sheets collected by F. Pérez Camarero in Larache in 1914.The BC specimen is labelled as "HERBARIO DE P. FONT QUER" with an indication of the variety handwritten by Pau.The MA specimen contains only one identification at species level but is numbered 15 as reported in the protologue.As a consequence, we have chosen this second sheet as the lectotype.On the basis of the examination of these sheets and following Dobignard & Chatelain (2010-2013)

Triglochin barrelieri var. maura
Number 5 is a new variety of Triglochin barrelieri Loisel.Although the name is cited as "var.nov.maurum Pau in litt.", the description of this taxon was made by Font Quer, who indicated that it differs from the nominal variety by its longer and almost straight pedicels and larger fruits ("Difieren del tipo por los pedúnculos mayores y casi rectos, con frutos doble mayores").Since the description was made by Font Quer, the authorship of this variety should be "Pau ex Font Quer".Köcke & al. (2010: 363) lectotypified this taxon on the basis of a specimen collected by Font Quer in "Rincón de Medik" from the exsiccata "ITER MAROCCANUM, 1930" conserved in Maire's collection (MPU barcode 006791).However, we have found original material with the same collector ("Pérez Camarero"), date ("Primavera de 1914") and locality ("Larache") indicated in Font Quer's publication in MA and BC herbaria (although they are labelled only as "Triglochin barrelieri").Being that the specimen proposed by Köcke & al. (2010) was collected 16 years after the taxon was published and original material exists, this lectotypification is not correct.It should be added that the BC specimen has another label handwritten by Pau, indicating "var.nova maura" and a diagnosis: "Frutos mayores, pedúnculos patentes no arqueados y más largos".However, we selected the MA specimen as lectotype because it includes a better preserved and complete plant.
The BC specimen also has a revision label by von Mering who re-determined it as Triglochin barrelieri Loisel.According to Köcke & al. (2010), this species is very variable in overall size and fruit and pedicel length, and includes also plants with long pedicels and big fruits, the distinctive characters of T. barrelieri var.maura.

Genista caballeroi
According to Pau (Caballero, 2017: 19), Genista caballeroi is similar to G. tricuspidata Desf.but without stipulary appendices.The locality indicated in the protologue is "Habitat copiosissima infra Hardú in latere septentrionale Montis Gurugú.Fl.Junio".There are two sheets in MA herbarium collected in Gurugú in June 1915 by Caballero. One (MA No. 58809) bears two original labels handwritten by Pau with the gathering information and the taxon name, whereas the other (MA No. 58812) is labelled "B.et C. Vicioso Herbarium Mauritanicum" and lacks information on the day of collecting.Both specimens bear complete plants and we have chosen as the lectotype the one labelled by Pau.
Genista caballeroi has been accepted by Talavera (2002) as an endemic taxon from N-Morocco and NW-Algeria, based on the absence of stipulary appendices to most of the leaves, the number of ribs on the stems and the size of the hairs on both stem and legume.However, neither Fennane (2007) nor Dobignard & Chatelain (2010-2013) accept it as a good species, and they synonymize it with G. tricuspidata, a species considered very variable from the morphological point of view.Desfontaines (1798), in the description of G. tricuspidata, did not mention the presence or absence of stipulary appendices, the character that Pau used to describe G. caballeroi.Neither the number of ribs non the hairs' size-characters used by Talavera (2002)-are included in this description.Therefore, although we have not seen the original material of G. tricuspidata, according to our criteria, G. caballeroi cannot be considered a different taxon.Pau in Caball. in Trab. Mus. Nac. Ci. Nat., ser. bot. 11: 19. 1917

Pau (1918): Plantas de Melilla:
This paper includes a list of plants with seven new taxa (five from Melilla and two from the Iberian Peninsula), of which one is here typified.The main collector (Angel Aterido) and the collection details are indicated in the introduction, where the locality and date are annotated "Tifasor; y fueron recogidas desde el día 10 al 30 de Julio del año 1915".It should be noted that Pau reviewed almost all the specimens mentioned in the paper when he was in Barcelona as he stated in the introduction "D.Ángel Aterido … recogió algunas plantas que he podido revisar durante mi corta estancia en la capital del Principado Catalán este año pasado".Pau (1918) provided a description for this new species specifying that it differs from B. balansae Boiss.& Reut.by its shorter leaves, more distant veins, shortly pedunculated umbels, shorter and wider involucral bracts, widened fruits and sharper ridges.Four herbarium sheets are indicated as possible types on the website of the MPU herbarium (https://collections.umontpellier.fr/herbier-mpu-presentation/base-dedonnees-botanique-herbier-mpu);however, none of them constitutes original material of this taxon, and therefore they are not eligible as lectotypes according to Arts.9.1 and 9.2 of ICN (McNeill & al., 2012).In fact, these plants were collected by Sennen in 1930 and 1931, some time later than the publication of the species in 1918.Moreover, they came from a different locality (Tizi-Assa) than that mentioned in the protologue (Tifasor).The appropriate original material is that from Tifasor collected by Aterido in July 1915, as indicated in the protologue.There are two sheets of this gathering in MA and BC herbaria.We select the one in MA as the lectotype because it has a note handwritten by Pau indicating "mi tipo" and the plant is more complete that the one in BC.

Bupleurum melillense
Although Pau considered the distinguishing characters of Bupleurum melillense as being enough to propose a new species, on the basis of the examination of the two types and attending to the revision label by Neves in 1998 in the BC specimen, this taxon should be synonimized to B. balansae, characterized by its slender rays and linear leaves (Fennane, 2007).We have designed the lectotype of B. balansae according to the PhD thesis of Neves (2000: 261).Pau in Bol. Soc. Aragonesa Ci. Nat. 17: 129. 1918

Celsia barnadesii var. mauritanica
This taxon was first mentioned by Pau (1922: 59), where he compared the Moroccan and Spanish specimens ("difieren ligeramente del tipo español … por sus cápsulas menores y oblongas"), and mentioned the locality of this variety as "en abundancia en las cercanías de Zeluán".Some years later, Pau (1929) subsequently proposed it as a species and indicated that the collection date was May 2, 1910.In the latter publication, Pau (1929) cited Celsia battandieri Murb.as a synonym since Murbeck included the original material of Selouane collected by Pau in the publicaction of this new species (Murbeck, 1925).
We have found three sheets at MA, all of them with a printed label of the Sennen exsiccata "1929 -PLANTES D'ESPAGNE -SENNEN Nº 6982" (Pau donated this material to Sennen for its distribution).The determination indicated on this printed label is Celsia mauritanica Pau.Among these three sheets, we propose the MA barcode MA108921 as the lectotype because it includes (1) an original label by Pau as C. barnadesii var.mauritanica, and (2) a large number of capsules whose size and shape are diagnostic for this variety, according to the protologue.Sennen distributed this exsiccata to different herbaria as evidenced by the isolectotypes found in BCN and P.
The specimen cited by Murbeck (1925) is conserved in LD herbarium.Although it is cited as "Tetuan, leg.Carlos Pau, 9/5 1910", Pau (1929) indicated that Murbeck had wrongly transcribed Pau's original handwritten label with the real locality ("Zeluan") and real date ("2.5.1910").The name indicated by Pau in the label, Celsia barnadesii, is crossed out by Murbeck, who determinated it as C. battandieri.Even through it is not labelled as var.mauritanica, the plant and the information of the gathering match that of Pau's publication and can be considered as an isolectotype.
There is also a specimen preserved in G herbarium [Charles Joseph-Marie Pitard's collection, who was a French correspondent of Pau (see Mateo, 1995)] collected by Pau from Selouane on May 2. Although the year is not indicated on the G specimen, we know that Pau only visited Selouane in 1910 (Mateo, l.c.).Instead of the printed Sennen label, it bears a label of "Plantas Españolas" (the first labels used by Pau), and it is only identified as Celsia barnadesii (without indicating variety); it seems that Pau donated this specimen before the description of the variety published in 1922 and used a label of his Spanish herbarium.However, the plant and the information on the locality, date and collector match the protologue and can be considered also as an isolectotype.Dobignard & Chatelain (2010-2013) included this variety in Verbascum battandieri (Murb.)Hub.-Mor.The G, P and BCN sheets (designated here as isolectotypes) have a revision label by Benedí in 2002 who re-determined them as V. battandieri.Moreover, Murbeck (1925), in his description of Celsia battandieri, cited the LD sheet as syntype of this species, so he considered that the characters of Pau's variety are included in the variability of this species.Centaurea fragilis var.zeluana Pau (1922: 52) mentioned this variety by indicating that in his herbarium there were some curious forms close to Centaurea malacitana Boiss.with pinnatifid basal leaves and entire upper leaves ("Forma intermedia inter genuinam et malacitanam; folia sup.integra ut in hispanica, basilaria pinnatifida").Although there is no indication of locality in the protologue, the variety name "zeluana" should be referred to Selouane (= Zeluan).
There are two sheets labelled as Centaurea fragilis var.zeluana and collected by Pau in Selouane, one in MA herbarium and another in BC herbarium.Besides, the BC specimen bears the indication "var.nova" handwritten by Pau.We select the one conserved in MA herbarium as the lectotype, since this herbarium includes the main collection of Pau.According to our examination of this MA specimen and also the revision label in the BC specimen added by Garcia-Jacas & Susanna in 2005, the features characterizing this variety are included in the variability of C. fragilis Durieu, an endemism from Morocco (El Oualidi & al., 2012).Although no original material of C. fragilis has been seen, according to the description made by Durieu (1847), this species includes plants with pinnatipartid basal leaves and entire upper leaves, the characters of Pau's variety.

Pau ( 1911 )
: Una visita botánica al Riff (Abril, Mayo 1910): This is a paper about Pau's first trip to Morocco in 1910, when he visited Melilla, Selouane, Puntilla and Corazón del Toro (Bugenzen).Eight new taxa for the Moroccan flora were described and one is typified here.
of this complex supports the treatment of the variety under C. supinus.
, this variety should be included in Ononis maweana Ball.
Triglochin barrelieri var.mauraPau ex Font Quer in Bol.Soc.Esp.Hist.Nat.14:425.Caballero (1917)published a list of plants collected by himself on a trip to Melilla in 1915, as he explains in the introduction of this paper: "El día 24 de Mayo de 1915 desembarcamos en el puerto de Melilla con el fin de realizar algunas exploraciones botánicas".On the sixth page he acknowledged Pau for his help in solving taxonomical problems ("nuestro sincero agradecimiento al sabio botánico español D. Carlos Pau, que ... ha resuelto todas las dudas que nos hemos permitido comunicarle.").This help resulted in the proposal of four new taxa, one typified here.

: Nueva contribución al estudio de la flora de Granada:
This paper is about plants of Granada (a Spanish province) but Pau described two varieties from Morocco by comparing African with Iberian samples, both typified here.