Tanacetum funkii Sch. Bip. ex Willk. was described by Willkomm (1865) based on plant material from the Sierra Nevada in SE Spain collected by Michael Funk, a physician from Bamberg in Germany, during a journey to Andalusia in 1848. The collections of Funk are owned by the Natural History Museum in Bamberg but today incorporated in the herbarium of the Botanische Staatssammlung in Munich (M) as a permanent loan. Duplicates of T. funkii from Funk’s collection were acquired in Paris (P) with the herbarium of Carl Heinrich Schultz (-Bipontinus) and in Coimbra (COI) with the herbarium of Moriz Willkomm.

The specimens in Funk’s herbarium were first studied by Schultz in 1850. He recognized the new taxon by adding handwritten labels with the name Tanacetum funkii that was subsequently validated by Willkomm (1865) in the second volume of the Prodromus Florae Hispanicae. The species was characterised as a much-branched, pubescent annual herb with very small solitary capitula at the end of branches. The 15-20 cm long stems bear pinnatipartite to bipinnatisect leaves with linear-lanceolate, mucronate segments, and shortly pedunculate capitula with pale scariously bordered involucral bracts. Shape and morphology of the achenes are unknown.

To the present day T. funkii has not been collected and due to the lack of additional materials showing further relevant characters, like flowers and fruits, its taxonomic position remains unclear. Even the study of the syntype specimen kept in the herbarium of Coimbra (COI) by Blanca & al. (2004) did not yield any additional information. Referring to this issue and the need of further investigation, T. funkii is mentioned in all European and regional botanical inventories, e.g., in Flora Europaea (Heywood, 1976), the Flora of Spain and the Balearic Islands (Smythies, 1984), La Flora de Sierra Nevada (Molero-Mesa & Pérez Raya, 1987), the Flora Amenazada y Endémica de Sierra Nevada (Blanca & al., 2004), and the recently published Flora Vascular de Andalucía Oriental (Blanca & al., 2009, 2011).

In the 1997 IUCN Red List of Threatened Plants (Walter & Gillett, 1998), T. funkii was classified in category “I” (indeterminate) due to the lack of information. According to Blanca & al. (2004), the intense search for T. funkii in the course of the LIFE project “Recuperacion de áreas con flora amenazada de Sierra Nevada” failed and additionally the habitat conditions of the classical localities were considered as heavily degraded by agricultural use. This resulted in the classification of T. funkii in the IUCN category “EX” (extinct) in the Atlas y Libro Rojo de la Flora Vascular Amenazada de España (Blanca & al., 2004) and subsequently in the Lista Roja 2008 de la flora vascular española (Bañares & al., 2008).

In the course of molecular phylogenetic studies of the tribe Anthemideae Cass. in general (Oberprieler & al., 2006, 2007, 2009) and of the genus Tanacetum L. (Sonboli & Oberprieler, 2010, 2012; Sonboli & al., 2012) in particular, we also came across T. funkii, which was listed by Bremer & Humphries (1993) in their generic monograph of the tribe and continued to be assigned to the genus until present (Euro+Med, 2015; Aedo & al., 2015). Morphological studies carried out on the type material of T. funkii gave rise to considerable doubts concerning the generic position of this species and motivated the present study for determining the taxonomic position of T. funkii within the tribe Anthemideae.

DNA was extracted from the type specimen of T. funkii (M 0032588) by using the CTAB DNA extraction protocol described by Doyle & Doyle (1987) and Doyle & Dickson (1987). PCR amplification of nrDNA ITS regions was carried out with primers 18Sf (Rydin & al., 2004) and P2Br (White, 1990) for ITS1 and P3f (White, 1990) and 28Sr (Rydin & al., 2004) for ITS2 utilising the Taq RED Polymerase (Ampliqon A/S, Odense, Denmark) according to the manufacturer’s suggestions. Sanger sequencing was done with the CEQ DTCS Quick Start Kit (Beckman-Coulter, Krefeld, Germany). Electropherograms were checked for base call errors using Chromas Lite v2.0 (Technelysium Pty Ltd, South Brisbane, Australia) and the newly established sequences for T. funkii were combined with nrDNA ITS sequences from previous phylogenetic analysis in the Anthemideae (i.e., Oberprieler & Vogt, 2000; Oberprieler, 2001, 2004a, 2004b; Oberprieler & al., 2007; Lo Presti & Oberprieler, 2009; Sonboli & al., 2012) into an alignment using BioEdit v.7.2.5 (Hall, 1999). Gaps were coded as binary characters using the simple gap coding method of Simmons & Ochoterena (2000) implemented in the software programme GapCoder (Young & Healy, 2003). This resulted in a combined data set that was divided into two partitions (i.e., nrDNA ITS sequence data, coded gaps of nrDNA ITS).

The program Modeltest v3.06 (Posada & Crandall, 1998) was used to find the model (among the 56 models tested) that best fits the underlying sequence information according to the Akaike Information Criterion (AIC). This resulted in the acceptance of a General Time-Reversable (GTR) model with a parameter describing the proportion of invariate sites and a gamma distribution of substitution rates over the sites (GTR+I+Γ), the base frequencies being freqA = 0.2340, freqC = 0.2283, freqG = 0.2461, and freqT = 0.2916, the proportion of invariate sites I = 0.0975, a gamma distribution shape parameter of α = 0.9344 and a substitution rate matrix of R[A-C] = 1.1504, R[A-G] = 2.3286, R[A-T] = 1.5896, R[C-G] = 0.3966, R[C-T] = 5.1490, and R[G-T] = 1.0. Substitution models and rates of substitution were allowed to vary among the parameters (“unlink” command and “ratepr = variable”) and a binary model (“Lset coding = variable”) was applied to the coded gaps. These parameters were used in a Bayesian inference (BI) approach to phylogeny reconstruction with the software programme MrBayes v3.2.2 (Ronquist & al., 2012). Two runs with eight Metropolis-coupled Markov chain Monte Carlo (MCMC) chains with incremental heating temperature of 0.2 were run for 5,000,000 generations and sampled every 1,000^th generation. The burn-in period was set to 25% of the 5,000 sampled trees and the first 1,250 trees were discarded. Estimation of tree topology and posterior probabilities of clades were based on the remaining 3,750 trees.

In the phylogenetic analysis based on nrDNA ITS sequence variation (fig. 1), T. funkii is found in a well-supported monophyletic group (posterior probability PP = 1.0) together with Anthemis cotula L. and Anthemis tigreensis J. Gay ex A. Rich. While in nrDNA ITS1 we observed two substitutions distinguishing between T. funkii and Anthemis tigreensis, there is no sequence difference detectable between the former and Anthemis cotula. In the nrDNA ITS2 region, a single nucleotide difference between T. funkii and Anthemis cotula was found (at position 512 of the total alignment), while three indels and four single nucleotide substitutions discriminate between T. funkii and Anthemis tigreensis. These results overwhelmingly argue for the inclusion of T. funkii in the genus Anthemis and its close relationship with Anthemis cotula; the single nucleotide difference in nrDNA ITS2 (position 512) between the two accessions being easily attributable to a plesiomorphic condition (G) in T. funkii and an apomorphic one (A) in the widespread and more likely genetically polymorphic Anthemis cotula, for which the accession for the present analysis came from an Italian population [Toscana: Grosseto, Padule di Punta Ala, Romi & Marchetti s.n., 15 June 1992 (Hort. Bot. Siena); cultivated in Hort. Bot. Berol. 103-03-93-10 (B 10 0603656)]. However, the observation made in many molecular phylogenetic studies at or below the species level for the nrDNA ITS marker showing intraspecific variation on the one hand (e.g., Lorenz-Lemke & al., 2005) or interspecific identity on the other hand (e.g., Kolarčik & al., 2010) is a justified argument for not basing taxonomic decisions solely on molecular markers and for considering other sources of evidence like morphology, anatomy, cytology, etc.

With regards to these other sources of evidence, morphologically the close relationship of T. funkii and Anthemis cotula is especially supported by the presence of linear-subulate receptacular scales (fig. 2a, b) observed in the capitula of the type material of the former species. While receptacular scales are very rarely observed in Tanacetum species (Oberprieler & al., 2006), i.e., in T. paleaceum Podl. and the former members of genera Gonospermum Less. (4 spp.) and Lugoa DC. (1 sp.) nested within Tanacetum (Sonboli & al., 2012), they are the rule in Anthemis representatives (again with some exceptions, i.e., the four species of the genus Ammanthus Boiss. & Heldr., nowadays included in Anthemis; Lo Presti & al., 2010). However, crossing experiments carried out by Mitsuoka & Ehrendorfer (1972) have shown that the presence/absence of receptacular scales is under oligogenic control in the tribe and, therefore, may be of minor taxonomic significance. Nevertheless, the presence of receptacular scales in T. funkii together with its solitary, small, and shortly pedunculated capitula (fig. 3) with conical receptacle (fig. 2a), its short, female ray florets (observed by C. Benedí, pers. comm.), involucral bracts with hyaline margin (fig. 2d), and its finely dissected leaves (fig. 2c) argue for its proximity to Anthemis cotula rather than to any Tanacetum species. The lack of well-developed tubular florets (fig. 2b) and achenes (presumably due to either the late collection date in August and September 1848 or some developmental irregularities) hampers evaluation of further and even more diagnostic characters.

As noted above, Anthemis cotula is an annual species of Anthemis sect. Maruta (Cass.) Griseb. that is wide-spread in the whole Mediterranean and Eurasian region (and introduced into other continents due to human activities). The species was studied morphologically quite extensively in its distribution range on the Iberian Peninsula by Benedí (1987) who described it as being an annual, foetid Anthemis species, with (10)20-35(45) cm long, apically ramified shoots, with 2-3-pinnatisect leaves (1.5)2.5-5.5(6) cm long and (0.5)1-3 cm wide, and having 3-5 mm long and mucronate lobes, and with capitula furnished with white, sterile ray-florets and a conical receptacle equipped with 3.5 mm long and 0.5 mm wide, subulate receptacluar scales either in its apical half or throughout (but then caducous in the lower half). In all these morphological characters, the Iberian populations correspond both to species circumscriptions given in other parts of the distribution range of the species (e.g., Fedorov, 1961; Yavin, 1970; Grierson & Yavin, 1975; Fernandes, 1976) and to the type specimens of T. funkii. The lack of well-developed florets and achenes on the type specimens is regrettable but not an impediment to the assumption of conspecifity of Anthemis cotula and T. funkii, because it appears more conservative and parsimonious, and therefore scientifically sound, to us to base this decision on observable characters than on mere speculations about alleged conditions of missing ones. Therefore, together with the above-mentioned proximity in molecular respects, we feel justified to propose the treatment of T. funkii as a synonym of Anthemis cotula:

Anthemis cotula L., Sp. Pl.: 894 (1753)

Tanacetum funkii Sch. Bip. ex Willk. in Willk. & Lange, Prodr. Fl. Hispan. 2: 102 (1865), syn. nov. Ind. loc.: “In Sierra Nevada in jugo argilloso el Puche et prope Cortijo de mimbres ad alt. circ. 3000´, FK.! – ○ Sept. (v.s.)”. TYPE: [Spain. Granada:] Sierra Nevada, auf dem Plateau “el Puche” [37º8‘19‘‘N, 3º29‘30‘‘W], 4000´ [feet], M. Funk s.n., Aug. 1848 (lectotype, here designated: P 00729970!, fig. 3; isolectotypes: M 0032589!, COI 00037339 icon!).

Additional material: [Spain. Granada:] Sierra Nevada, Cortijo de mimbres [37º7‘27‘‘N, 3º28‘8‘‘W], M. Funk s.n., “Anf. Septbr.”, 1848 (syntype: M0032588!).

The specimens from the locality El Purche (“el Puche”) kept in the herbaria COI and P, and the one from the locality Cortijo de las Mimbres (“Cortijo de mimbres”) in M, bear determination slips in Schultz’s handwriting with the indication “Tanacetum Funkii, C. H. Schultz Bip. 19.10.[18]50”. The specimen from the herbarium of Schultz kept in P is selected as lectotype (fig. 3).