Vascular Plants from the Journey through Portugal (1797–1801) by Hoffmannsegg and Link at the Herbarium of the Real Jardín Botánico of Madrid

During the journey through Portugal by Hoffmannsegg and Link (1797–1801), these authors collected an appreciable number of specimens, most of which have been lost. Their collections are relevant since they were used by themselves or by other authors to describe numerous species. In the herbarium of the Real Jardín Botánico of Madrid, 70 specimens from this journey have been located. In the archive of this institution the letters that Hoffmannsegg and Link sent to Cavanilles accompanying these plants have also been located. The analysis of these letters, the herbarium labels and of the protologues has permitted to establish that 15 specimens are original material, four of which had already been proposed as lectotypes by other authors (Airochloa caudata Link, Silene fuscata Link ex Brot., Silene micrantha Link ex Otth and Silene pernoctans Link). The designation of a neotype for Stipa gigantea Link should be superseded, because an original material has been found. Thus, a lectotype for this taxon is proposed.


Introduction
The naturalist and botanist Johann Centurius Graf von Hoffmannsegg (Dresden, 1766-Dresden, 1849) [1] promoted two scientific expeditions to Portugal at his expense. The first one took place between the fall of 1795 and the spring of 1796, in the company of the German naturalist Wilhelm Gottfried Tilesius. This first expedition had little success, according to Gomes Oliveira [2], due to the difficulties they encountered in obtaining permission from the local authorities to travel through the country. On this occasion they barely toured the environs of Lisbon.
On the second trip, he was accompanied by Johann Heinrich Friederich Link (Hildesheim, 1767-Berlin, 1851) [3]. This journey began in Hamburg on 18 August 1797, heading for Lisbon, but due to poor sea and ship conditions, the sea voyage stopped at Dover, from where the expeditionaries went to Calais. They crossed France and Spain to enter Portugal through Elvas on 11 February 1798. During their trip they met Antonio José Cavanilles (Valencia, 1745-Madrid, 1804) in Madrid, from whom they received the gift of a " . . . grand nombre de plantes choisies de l'Espagne" [4] (p. 2). Hoffmannsegg and Link initially went to Lisbon where they obtain a generous safe-conduct that allowed them and their servants to travel freely throughout Portugal [2].
From this base camp they carried out movements and collections throughout the country, from the Algarve, in the extreme south, to Gerês in the vicinity of the Spanish border with Galicia. In Coimbra they meet Brotero with whom they later exchanged plants. They worked in Cabo da Roca, Alentejo, Setúbal, Serra da Arrábida, Coimbra, Aveiro, Porto, Serra da Estrela, etc. until the spring of 1799, at which point Link had to return because the license given to him by the University of Rostock was expiring. He did so via England, where he compared the Portuguese plants with those in J. Banks's herbarium. type specimens of Link plants, three of them accessible through JSTOR Global Plants [8] and the other one on its Data Portal website [20]. The JSTOR Global Plants website [8] offers information about a Crocus clusii J. Gay specimen housed in the Kew herbarium (K) that can clearly be attributed to this author, to add to another Verbascum specimen found in The Linnaean Society herbarium (LINN). The Herbarium Hamburgense (HBG) website [21] and the Finnish Museum of Natural History herbarium (H) database [15] give information with regard to a different specimen collected by Link. Stafleu & Cowan [16] (p. 655) indicate that there are also Link specimens in BR, C, FI, LIV, P, PH and W, which has not been confirmed, at least regarding Portuguese plants.
The numbers of plants of these authors obtained from the literature and from the available databases must be taken with some caution since the digitization process of the herbarium collections is still in progress in many of this herbaria. Link [22] (p. 238) notes that "Nous avons trouvé en Portugal 1532 espèces de plantes ordinaires, 572 espèces de plantes cryptogamiques", which indicates that they collected a higher number of specimens and seems to suggest that most of the collections that these authors gained in Portugal have been lost.
Hoffmannsegg and Link initially published articles in German journals to publicize their findings in Portugal as a previous step to their main work. Among the most relevant, since they contain taxonomic novelties, we could mention Link [23][24][25][26][27][28] and Hoffmannsegg and Link [29]. Brotero in 1804 published his Flora Lusitanica, where he recognized the plants communicated by Hoffmannsegg and Link to him, which apparently was not reciprocal [30,31].
The main work of Hoffmannsegg and Link, the Flore Portugaise, was published in two volumes with independent pagination. The first [4], to which fascicles 1 to 14 correspond, was published between 1 September 1809 and 1820. The second volume [32], to which fascicles 15-22 correspond, was published between 1813 and 1840 [14] (p. 246). The work was left unfinished. According to Gomes Oliveira [2] (p. 114), the editorial plan included a third volume with eight issues. This flora has detailed descriptions in French and Latin, includes 109 splendid plates and mentions precise localities of most of the species. Gomes Oliveira [2] (p. 17) points out that 659 species are described, that is, about a quarter of those known today in mainland Portugal [33] (p. 112). The Flora Lusitanica of Brotero [34] describes 1225 species [2] (p. 136), which is slightly less than half of those known today [33] (p. 112).
Parallel to the Flore Portugaise, Link [35][36][37][38] continued to publish taxonomic novelties based on the Portuguese collections. Other authors then had access to the plants collected by Hoffmannsegg and Link and also published new species based on them [39][40][41].
An indication of the importance of the work of Hoffmannsegg and Link is its contribution to the description of Iberian endemisms. According to Buira and Aedo [42], they are among the twelve most prolific authors in terms of the description of endemic species. In addition, they are, together with Brotero, the main responsible for the description of the endemisms of the western Iberian Peninsula.
The objective of this work is to make known the Portuguese plants of Hoffmannsegg and Link that are preserved in the Herbarium MA of the Real Jardín Botánico of Madrid, and to establish if they constitute original material of the names proposed by the aforementioned authors.

Materials and Methods
The archive of the Real Jardín Botánico of Madrid keeps six letters sent by Hoffmannsegg to Cavanilles from Lisbon between October 1798 and August 1801 [43][44][45][46][47][48], and two letters sent by Link also to Cavanilles from Lisbon, in October 1798 [49] and in January 1799 [50].
On the other hand, a review of the nomenclatural database of the Flora iberica project has been made to detect the names of the taxa described by Hoffmannsegg and Link. Only those heterotypic names validly published have been considered and, in the case of illegitimate names, those that are by homonymy.
From these two sources of information, the Portuguese plants that Hoffmannsegg and Link sent to Cavanilles have been located in the MA herbarium, with the additional help of the herbarium database. The currently accepted name is indicated if it is different from the original. In the corresponding section, it is discussed whether there is sufficient evidence to consider that these Hoffmannsegg and Link plants are original material of a validly published name. Proposing lectotypes has been avoided, except in exceptional cases of groups well known to the authors, in accordance with Recommendation 9A.2 from the International Code of Nomenclature [51] (p. 26): "Designation of a lectotype should be undertaken only in the light of an understanding of the group concerned", to avoid possible future confusion and further changes. It is necessary to keep in mind that original materials corresponding to the names of Hoffmannsegg and Link may still appear in various European herbaria. It is mentioned whether other authors have proposed any lectotype on these plants.
We have adopted the volume and fascicle numeration as well as the dates proposed by IPNI to properly and unequivocally cite the Journal für die Botanik.

Hoffmannsegg and Link's Letters to Cavanilles
Hoffmannsegg's first letter to Cavanilles is dated 14 October 1798 [43]. In it, he announces that he and Link are preparing to send a first set of 50 plants so that Cavanilles can give them his opinion about their identity. In this letter he establishes the working method through a numbered list of the specimens that would be included together with the shipment, with a copy in the hands of the German botanists: " . . . Comme j'ai gardée une copie de la liste-Ci-jointe, Vous n'avez qu'a nous rapporter seulement au no. de cells sur lesquelles il Vous plaira de nous instruire" (Figure 1). He also announces a second similar shipment coming four to six weeks later. The attached list is not included in this letter. However, in Link's letter to Cavanilles, dated 12 October 1798 [49], despite its earlier facial date, it refers to the aforementioned Hoffmannsegg letter and includes the announced list of 50 plants. Cavanilles notes his opinions in the margin of the letter received. For example, he adds: "Sit; venumtamen flores sunt avena elatioris, a qua foliis difert" to "35. Avena pallens nob.", "Verbenacam credo" to "43. Salvia clandestine" or "bene" to "44. Pinguicula lusitanica", to correct or accept the identifications of the German botanists. It goes without saying that Cavanilles would reply to the letter with this information.
In the following letter from Link, dated 28 January 1799 [50], a list with plants numbered from 51 to 102 is attached, to which Cavanilles adds his comments in the same way. In the remaining letters from Hoffmannsegg, dated 14 April 1799 [45], 1 February 1801 [46], 6 July 1801 [47], and 16 August 1801 [48], he is thankful for the determinations and diverse news that are given, but without lists of plants. The only other list available is the one received on 27 April 1801 [44], with no attached letter available in which another shipment of plants numbered from 1 to 45 is recorded. This correspondence mentioned above is obviously incomplete, but sufficient to understand the working method agreed upon by these authors. In all probability, an indeterminate part of the letters from German botanists to Cavanilles has not reached the present day, since part of the plants sent do not appear in any of the three available lists. It is interesting to indicate that the documentation on Cavanilles remained in the hands of his heirs for almost 200 years until 1992, when it was deposited in the archive of the Real Jardín Botánico of Madrid [52].

The Portuguese Plants of Hoffmannsegg & Link at the Herbarium MA
The current sheets where the Hoffmannsegg and Link plants have been mou have small labels of approximately 5-10 × 3-5 cm. A comparison between the data the letters and those from the numbered labels supports the idea that the sheets were with only a number annotated by the German botanists and that Cavanilles adde comments in clearly distinguishable handwriting and ink. This portion of the ori folded paper sheet was cut out at some point and used as a label. A small part o sheets that are not mentioned in the letters seem to have been totally or partially beled by a hand other than Cavanilles ( Figure 2). Labels usually lack a precise locality, since the lists sent by Hoffmannsegg and generally do not include this data. Cavanilles adds "ex Lusitania" or an equivalent plus the name of the plant. Exceptionally in MA00114705, Cavanilles writes "Nea Tagus in Portugal", translating the information from the letter [49] in which Link says Lathraea phelypaea In arenosis/trans Tagum". Other times Cavanilles does not transpos locality (22. Rhinanthus versicolor Lamark. Frequens circa Olyssiponem) from Link's letter to the label, and simply notes "Lusitania".
The collection's 70 vascular plants of Hoffmannsegg and Link located on the barium MA are listed below (Table 1), with the MA herbarium barcode number an indication of the current accepted name if it differs from the original.

The Portuguese Plants of Hoffmannsegg & Link at the Herbarium MA
The current sheets where the Hoffmannsegg and Link plants have been mounted have small labels of approximately 5-10 × 3-5 cm. A comparison between the data from the letters and those from the numbered labels supports the idea that the sheets were sent with only a number annotated by the German botanists and that Cavanilles added his comments in clearly distinguishable handwriting and ink. This portion of the original folded paper sheet was cut out at some point and used as a label. A small part of the sheets that are not mentioned in the letters seem to have been totally or partially relabeled by a hand other than Cavanilles (Figure 2). Labels usually lack a precise locality, since the lists sent by Hoffmannsegg and Link generally do not include this data. Cavanilles adds "ex Lusitania" or an equivalent text, plus the name of the plant. Exceptionally in MA00114705, Cavanilles writes "Near the Tagus in Portugal", translating the information from the letter [49] in which Link says "14. Lathraea phelypaea In arenosis/trans Tagum". Other times Cavanilles does not transpose the locality (22. Rhinanthus versicolor Lamark. Frequens circa Olyssiponem) from Link's letter [49] to the label, and simply notes "Lusitania".
The collection's 70 vascular plants of Hoffmannsegg and Link located on the Herbarium MA are listed below (Table 1) [43][44][45][46][47][48][49][50] in the archive of the Real Jardín Botánico of Madrid.  [23], when describing the species, indicates that it has "Labium trifidum, lacinia media trifida, cuius laciniae lateral longissimae, angustissimae, intermedia brevior", as corresponds to O. italica and indicates as locality "Häufig auf den Hügeln um Bellas". This geographical reference does not appear in the available letters. The label of the MA00023704 sheet reads "ex Lusitania/Link dedit". In the Berlin herbarium (B-W-16824-01 0) there is a specimen attributed to Link, from Portugal, ("Habitat in Lusitaniae collibus") that would also turn out to be original material.
Silene fuscata Link ex Brot., Fl. Lusit. 2: 187 (1804). Brotero [34] [23], when describing the species, indicates that it has "Labium trifidum, lacinia media trifida, cuius laciniae lateral longissimae, angustissimae, intermedia brevior", as corresponds to O. italica and indicates as locality "Häufig auf   [28] describes the species "circa Coina, Moita, Aldea galliga ut alibi trans Tagum prope Lisboa". In the letter sent by Link to Cavanilles in January 1799 [50] it reads "N. 73. Singularis Ulicis species, quam stauracanthum appellamus", but no locality is indicated. On the label of MA00060058, which bears the number "73" handwritten by Link, Cavanilles notes "Ulex stauracanthus Link/e Lusitania misit autor". This sheet is original material. 15. Stipa gigantea Link, J. Bot. (Schrader) 2(2): 313 (Apr. 1800). Link [23] describes the species from "An den Sandhügeln bei Setuval, Bellas, und am südlichen Abhange der Estrella fanden wir es nicht selten". This species does not appear in the available letters. On the label of the MA00005273 sheet, which bears the number "62" handwritten by Link, Cavanilles notes "Avena gigantea sp. n./ex Lusitania". This specimen is connected to Link species through the epithet "gigantea" and fits very well with the current concept of S. gigantea. Vázquez & Barkworth [62] (p. 491) assumed that the original material had been destroyed in B and designated a neotype (HSS-9002). According to Article 9.19 of the International Code of Nomenclature [52] (p. 25): " . . . that choice should be superseded if (a) the holotype or . . . any of the original material is found to exist". We designate here as lectotype the specimen MA00005273 (Figure 4), which contains a fragment with a well-developed inflorescence and allows us to unequivocally appreciate the distinctive characters of the species [63].

Types of the Taxa Described by Hoffmannsegg and Link in the Herbarium MA
As indicated above, the lectotypes of the names Airochloa caudata Link, Silene fuscata Link ex Brot., Silene micrantha Link ex Otth and Silene pernoctans Link were proposed by previous authors [53,58]. Vázquez & Barkworth [62] erroneusly proposed a neotype for

Types of the Taxa Described by Hoffmannsegg and Link in the Herbarium MA
As indicated above, the lectotypes of the names Airochloa caudata Link, Silene fuscata Link ex Brot., Silene micrantha Link ex Otth and Silene pernoctans Link were proposed by previous authors [53,58]. Vázquez & Barkworth [62] erroneusly proposed a neotype for Stipa gigantea Link. In this work, original material of S. gigantea has been located at MA, and consequently a lectotype has been proposed to replace the neotype.