Ramalina peruviana Ach. was described in the early days of lichen classification (Acharius Reference Acharius1810). In those days, only a limited set of distinguishing characters was known, resulting in short descriptions with limited information (Fig. 1). Today, microscopy, TLC and phylogenetic analyses provide much more data, allowing a much more detailed and phylogenetically supported classification (see e.g. Spjut et al. Reference Spjut, Simon, Guissard, Magain and Sérusiaux2020). However, as a result, it can be problematic finding the correct application of names such as Ramalina peruviana. Spjut et al. (Reference Spjut, Simon, Guissard, Magain and Sérusiaux2020) thus proposed the hypothesis that records of R. peruviana from Southern Europe and Macaronesia, far away from Peru, belong to a separate species with a type from the region, Ramalina crispans Werner, described from Morocco.

Figure 1. The first description of Ramalina peruviana (Acharius Reference Acharius1810). In colour online.

To test this hypothesis, we looked for a specimen from the locus classicus that would allow extraction of the ITS sequence. The age of the type specimen makes it unsuitable for simple sequencing and we needed a fresh specimen. Unfortunately, this locus classicus is not very precisely indicated in the original description (Fig. 1), just given as Peru (‘in Peruvia’). However, in the English translation the complete second paragraph of the first description says: ‘Lives in Peru, together growing with Borrera villosa and Borrera ephebea. Lagasca.’. These associated lichen species, now named Seirophora villosa (Ach.) Frödén and Tornabea scutellifera (With.) J. R. Laundon, restrict the locus classicus greatly because they occur in South America only along the Pacific coast from Peru to northern Chile. Here, the Humboldt Current in the Pacific Ocean causes frequent fog in the otherwise desert-like, cool, coastal area and provides a suitable climate for these species. The presence of these two species along the Pacific coast of Peru, and in particular near Lima (Peru), is confirmed in the revisions of Hillmann (Reference Hillmann1930) and Nimis & Tretiach (Reference Nimis and Tretiach1997).

A further, rather cryptic indication for the origin of the type specimen of Ramalina peruviana is the addition ‘Lagasca’. This refers to Mariano Lagasca y Segura (1776–1839), a leading botanist in Spain. Apparently Acharius received the specimen from him. Lagasca is not known to have visited Peru himself and must have obtained the specimen from somebody else. Krog & Swinscow (Reference Krog and Swinscow1976), who identified the holotype of R. peruviana in the Acharius herbarium in Helsinki (abbreviation H-ACH), give the name as ‘Lagasta’, apparently a misinterpretation of the unusual characters used in the book. Remarkably, the picture of the holotype on the internet (JSTOR 2022) lacks the name of this botanist and instead has ‘Cavanilles’ written in pencil and probably added later. This represents Antonio J. Cavanilles (1745–1804), another leading botanist in Spain at that time, who also did not visit Peru himself. In view of this contradictory information, it is believed that the specimen was collected during the Expedición Malaspina in 1789–1794, most probably by the botanist of the team, Thaddäus Haenke, a prolific plant collector in South America. His visit to Lima (Callao) during this expedition is certain because he left the expedition there to explore the Andes. In those days, a trained botanist travelling to Lima and being prepared to collect botanical specimens was still very unusual, so it is unlikely that somebody else did collect the material of Ramalina peruviana. After the expedition, Malaspina, the expedition leader, fell into disrepute and was put in jail (Wikipedia, accessed 4 October 2022; https://es.wikipedia.org/wiki/Alejandro_Malaspina). For this reason, the evaluation of his botanical collections was taken over by Spanish botanists. This would explain why the names of Lagasca and Cavanilles both show up in connection with the type specimen of R. peruviana. Summarizing all available evidence, the type specimen of R. peruviana most probably came from fog oasis-like vegetation near Lima in Peru. Such vegetation is still present there in the National Reserve of Lachay, and that seems to be a suitable place to obtain ITS sequences of Ramalina peruviana in the original sense (Fig. 2).

Figure 2. Specimens of Ramalina peruviana from the probable locus classicus, Lachay near Lima, Peru. A, well-developed specimen, collected in 1960 (Mattick 25, B 60 0016401). B, specimen used for sequencing, collected in 2018 (B 60 0176943), dwarfed probably by air pollution. Scales: A & B = 2 mm. In colour online.

For the sequencing, the following sample from the locus classicus was used: Peru, Prov. Lima, district Huacho, National Reserve of Lachay (Resolucion Jefatural 007-2018), elev. 350 m, 11°21.3ʹN, 77°21.85ʹW, 15 Dec. 2018, A. M. Ramírez Ordaya, 17 dic. 2018 (USM, B 60 0176943). A fragment was sent to the Botanical Museum in Berlin (B), where the chemistry was investigated by TLC (Orange et al. Reference Orange, James and White2001; solvents A, B and C). DNA sequencing of the ITS region was performed by P. Alvarado (Alvalab, Spain; http://www.alvalab.es/) using the extracted material remaining after TLC, and the methods outlined in van den Boom & Alvarado (Reference van den Boom and Alvarado2021). In addition, using the same methods, new sequences were obtained from other specimens of R. peruviana and five related species (Table 1).

Table 1. Newly generated ITS sequences from specimens of Ramalina species, including voucher information and GenBank Accession numbers.

For the phylogenetic analysis presented in Fig. 3, all available ITS sequences in GenBank of Ramalina peruviana and the probable relatives R. crispans, R. hyrcana Sipman and R. montagnei De Not. were used. We also added four sequences each of related species present in fig. 9 of Spjut et al. (Reference Spjut, Simon, Guissard, Magain and Sérusiaux2020). All sequences were trimmed according to the shortest sequence (MN811427), and sequences shorter than this were omitted. The sequences were aligned with Muscle (https://www.ebi.ac.uk/Tools/msa/muscle/; accessed 12 September 2022) (Madeira et al. Reference Madeira, Pearce, Tivey, Basutkar, Lee, Edbali, Madhusoodanan, Kolesnikov and Lopez2022) with some manual corrections. A maximum likelihood (ML) tree was calculated with IQ-TREE (Trifinopoulos et al. Reference Trifinopoulos, Nguyen, von Haeseler and Minh2016; http://iqtree.cibiv.univie.ac.at/) using default settings (number of bootstrap alignments 1000, maximum iterations 1000, minimum correlation coefficient 0.99, perturbation strength 0.5 and IQ TREE stopping role 100). The tree was visualized with FigTree (http://tree.bio.ed.ac.uk/software/figtree/) and rooted in accordance with fig. 9 of Spjut et al. (Reference Spjut, Simon, Guissard, Magain and Sérusiaux2020). Maximum parsimony and maximum likelihood values are presented in Fig. 3.

Figure 3. Maximum likelihood (ML) tree of all available ITS sequences of the Ramalina peruviana complex, compared with four selected sequences each of R. fraxinea (L.) Ach., R. polymorpha, R. fastigiata (Pers.) Ach., R. canariensis J. Steiner and R. farinacea (L.) Ach. The sequence from the locus classicus is in bold red font; the clades of R. peruviana and R. intermediella are indicated with grey boxes and R. montagnei is shown outside these clades. Bootstrap values (maximum parsimony/maximum likelihood) > 50 are indicated. In colour online.

Ramalina peruviana clusters with R. crispans, R. hyrcana and R. montagnei (Fig. 3) and forms a well-supported clade. Sister is R. polymorpha (Lilj.) Ach., as in fig. 9 of Spjut et al. (Reference Spjut, Simon, Guissard, Magain and Sérusiaux2020). Within the clade, all sequences except those from East Asia are on a well-supported branch. On this branch, R. montagnei holds a distinct position and appears to deserve recognition as a separate species. It differs morphologically because it is a non-sorediate, regularly fertile taxon. In the alignment it deviated by two base pairs from all other samples (Fig. 4). Otherwise it agrees with R. peruviana s. str., except for one base pair which it shares with the East Asiatic samples.

Figure 4. The variable base pairs in the ITS sequence of the Ramalina peruviana s. lat. cluster. In colour online.

Further samples on this branch, previously identified as Ramalina peruviana, R. crispans and R. hyrcana, have completely identical sequences, which supports their treatment as a single species. The earliest name available for this subclade is Ramalina peruviana. For R. crispans, no morphological difference from R. peruviana is known (Spjut et al. Reference Spjut, Simon, Guissard, Magain and Sérusiaux2020) and it is best considered as a synonym of R. peruviana. Ramalina hyrcana is said to be different from R. peruviana because the lobes are more flattened and the soralia are downturned which give the thallus lobes a dorsiventral aspect (Lumbsch et al. Reference Lumbsch, Ahti, Altermann, Amo de Paz, Aptroot, Arup, Bárcenas Peña, Bawingan, Benatti and Betancourt2011; fig. 21E & F). However, observation of specimens in herbarium B suggests that these differences are not constant, and the species is considered here to be also a synonym of R. peruviana.

The position of the nine East Asian specimens in Fig. 3 indicates that they deserve a separate classification. They deviate in the alignment from R. peruviana as defined above by six base pairs forming two groups of sequences containing three and six specimens (Fig. 4). Ramalina intermediella Vain., described from Japan and reported as a synonym of R. peruviana by Stevens & Kashiwadani (Reference Stevens and Kashiwadani1987), appears to be the earliest name available for East Asiatic samples, which are currently treated as R. peruviana, for example in Kashiwadani et al. (Reference Kashiwadani, Moon and Lai2006).