taxonID	type	description	language	source
EF51B220FFD2FFFDFF24FDB01FDDF821.taxon	materials_examined	Lectotype (designated by Weber 1986: 216): In sepibus maritimis Hispaniae, sine dat., Schott s. n. (W). Ind. Loc: Mountains of Gibraltar.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD2FFFDFF24FDB01FDDF821.taxon	description	The name Rubus ulmifolius was published twice by Heinrich Wilhelm Schott, first in the Väterländische Blätter für die österreichische Kaiserstaat (Schott 1818 a: 42) and subsequently in Isis (Schott 1818 b: 821). It was typified by Weber (1986) from a specimen preserved at W. There is no doubt about its identity, it is the discolour blackberry, i. e. a blackberry with abaxially white tomentose leaves, with strongly pruinose stems and small leaves, which is the most common species in South-West Europe. As a diploid taxon it has a large variability in contrast to the apogamous taxa which form the bulk of Rubus species in Europe. Because earlier botanists were not aware of this difference they dealt with the variations of R. ulmifolius in the same way as with the apogamous taxa. This resulted in a large number of related taxa, sometimes ordered as infraspecific taxa. Sudre (1908 – 1913) recognized 8 subspecies, 20 microgenera (an unfortunate word for an infraspecific rank!) and 94 varieties. Next to these many synonyms are listed. Most of these do not have real taxonomic value. Monasterio-Huelin & Weber (1996) reduced the number of infraspecific taxa to 4 varieties. Rubus ulmifolius has also many hybrids, often with unknown other parents.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD2FFFDFF24FDB01FDDF821.taxon	discussion	Before Focke (1877) recuperated the name R. ulmifolius most authors used other, later names for the species, e. g. R. discolor Weihe & Nees (1824: 46) or R. rusticanus Mercier (1861: 279), but since Focke’s publication is has been in common use.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD2FFFDFF24FDB01FDDF821.taxon	description	There is no doubt about the identity of the type. It consists of both an inflorescence and a piece of a primocane. On the label is written ‘ In sepibus maritimis Hispaniae’ which corresponds well with the locality in the protologue: ‘ in montosis Gibraltariae’, or at least does not contradict it.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD2FFFEFF24FA88183DFC73.taxon	materials_examined	Lectotype (designated by Van de Beek 2016: 46): Greece, Crete, Tournefort 6073 (P-TRF, 2 - D code P 00680425).	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD2FFFEFF24FA88183DFC73.taxon	description	The same taxon was published again as R. sanctus Schreber (1766: 15) and as R. parviflorus Weston (1770: 258). The publication of R. parviflorus is based on the description by Tournefort and thus this name is homotypic with R. creticus. Because Schreber based his description on a plant at M (see Van de Beek 2016) Monasterio-Huelin & Weber (1996) indicated this as the “ holotype ” of R. sanctus, but because Schreber included an illustration in the protologue the specimen at M should rather be treated as lectotype. However, because Schreber included the phrase name (or nomen specificum legitimum) “ Rubus creticus, triphyllo, flore parvo ” of Tournefort (1703: 43) which is the validating description of R. creticus, is R. sanctus an illegitimate name under Art. 52.3 (Turland et al. 2018) and a superfluous homonym of the latter. So R. creticus, R. sanctus and R. parviflorus are homotypic. Some batologists argue that R. ulmifolius and R. creticus are subspecies of the same species (Focke 1902: 504 [as R. ulmifolius ssp. anatolicus Focke]; Sudre 1908 - 1913: 76; Juzepczuk 1941: 24; Parsa 1948: 105; Van de Beek 2016: 46). None of these authors has drawn the conclusion that according to the rules R. ulmifolius should be an infraspecific taxon of R. sanctus, probably because of the popularity of the former. Moreover, R. ulmifolius will become a subspecies of R. creticus. This will lead to new combinations if R. ulmifolius is divided in smaller unities as some authors have done, and to numerous new hybrid formulas because R. ulmifolius crosses frequently with other species.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD1FFFFFF24FE7E1933FFAE.taxon	materials_examined	Lectotype (designated by Van de Beek (2016: 36): [illustration] “ Rubus ” in Matthioli in Commentarii secundo aucti (1559: 507). Ind. Loc: not indicated.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD1FFFFFF24FE7E1933FFAE.taxon	description	Epitype (designated here): Italy, Valgrisanche (Aosta), ca. 900 m, 03 Jul 1961, Van Ooststroom 22933 (L, 2 - D code L. 1907626) (Fig. 1). Another name which is relevant for the nomenclature of R. ulmifolius is R. vulgaris J. de Vries bis (1779: 196). Matzke-Hajek (2016) argued that this name would be not validly published because it is not in a scientific publication and the text is only a translation of Valmont de Bomare (1765) which does not use binary nomenclature consistently and consequently is not validly published. However, De Vries presents his book as a scientific commentary on Martinet’s catechism, and moreover, though it is recommended to not publish in popular papers or books (ICN Art. 30 A. 4, Turland et al. 2018), it is not forbidden, and a translation of an invalidly published text does not cause that also this translation is invalid if accepted by the author who uses it; many invalidly published texts serve as validating description by later authors. Rubus vulgaris is validly published as such. It is also clear that neither De Vries (who refers to Tournefort and to Duhamel [1768] in his book) nor Valmont thought out the name R. vulgaris themselves but adapted to common use, as many authors of that time refer to R. vulgaris Bauh. (see Bauhin 1623: 479).	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD1FFFFFF24FE7E1933FFAE.taxon	discussion	The type of R. vulgaris was designated by Van de Beek (2016: 36) and it is an illustration published in the Matthioli’s Commentarii secundo aucti (1559: 507). Though this picture is not very precise, it can, because of the strong prickles and rigid inflorescences, hardly be anything else than R. ulmifolius, the most common blackberry in Italy where Matthioli lived. However, the drawing does not give precise details so that it is not of a quality to identify it without knowing the broader context. Consequently, for a precise identification of this name an epitype is selected according to ICN Art. 9.9 (see Turland et al. 2018). This material, a complete specimen, with leaves and welldeveloped flowers (Fig. 1), clearly represents the traditional concept (e. g., J. de Vries bis 1779: 196) and current use and application of R. vulgaris as a synonym of R. ulmifolius (Van de Beek 2016).	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD0FFFFFF24FD3A19D7F80A.taxon	materials_examined	Lectotype (designated here): [icon] “ Rubus non spinosus, mai. fruct. nigr. ” in Barrelier (1714: ic. 395) (Fig. 2). Ind. Loc: not indicated.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD0FFFFFF24FD3A19D7F80A.taxon	description	Ortega, in volume 6 of Quer’s Flora Española (1784: 524) published the name Rubus non-spinosus. In the index of the volume he lists the species in the Linnaean binary form. Obviously he considered these as the formal names of the plants. He added as author ‘ Barr. ’, which refers to Barrelier (1714). Because he also refers to p. 223 of his own volume the text about the taxon on that page is included in the protologue. He cites both Barrelier (1714: “ Obs. 1373. Icon. 353 ”) and Tournefort (1700: 614), so that both are part of the protologue. No specimen of Barrelier could be found, only an image (Barrelier 1714: icon 353 [recte 395]) (Fig. 2). On the other hand, in the herbarium of Tournefort is a good specimen of a thornless form of R. ulmifolius (Tournefort 6078 [at P-TRF]) (Fig. 3) but revised as R. fruticosus var. inermis (see this taxon below). However, this material was not cited in the protologue and neither can it be treated as original material used by Ortega to describe his species. We have not found any original material of this name in the herbaria consulted (e. g., BM, E, H, P, MA). Thus, the Barrelier’s illustration “ Rubus non spinosus, mai, fruct. nigr. ” (1714: 395) cited by Ortega (without the abbreviations) in the protologue of R. non-spinosus is designated as the lectotype of the name. This drawing illustrates a complete plant, with leaves, flowers and fruits, without prickles, and matches with the traditional concept and current use of the name as a thornless form of R. ulmifolius.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD6FFFBFF24FF791863FFE7.taxon	materials_examined	Lectotype (designated by Van de Beek 1979: 206): Pourret 3168 (MAF-POURR). Ind. Loc: near Barcelona. (Fig. 6)	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
EF51B220FFD6FFFBFF24FF791863FFE7.taxon	description	Van de Beek (1979) argued that R. inermis Pourret (1788: 326) is identical with R. ulmifolius. Because this name is earlier than R. ulmifolius it would be the correct name of the species. Monasterio-Huelin & Weber (1996) doubted the identity of both and claimed that R. inermis might be a hybrid of R. ulmifolius. The lectotype of R. inermis Pourret, published in Van de Beek (1979: 206), Pourret 3168 (MAF-POURR), looks like a normal R. ulmifolius without prickles, but hybrids can sometimes be very similar. However, the type does not show reduced fertility. This is confirmed by two other specimens of Pourret which were found after the publication of the type. Another specimen (probably a duplicate of the lectotype) was found at P (with barcode P 02521232, image available at http: // mediaphoto. mnhn. fr / media / 1441338624753 ZJhEBLGDiQxlEGfy). It consists of two unarmed inflorescences, with leaves and flowers, and clearly has young fruits that are not defective. At the base the sheet is annotated with “ Rubus inermis ” by Pourret, and also contains a label annotated as “ Rubus inermis P. / Caule fruticoso sub-rotundo. / Folis ternatis subtus tomentosis / an Canadensis. [handwritten by Pourret] / Pourret scripsit. [handwritten by Spach] / à Barcelona [handwritten by Pourret] ”. The best specimen is preserved in the Salvador herbarium at BC (BC-Salvador 3833). It consists of both an inflorescence and a young primocane with leaves (Fig. 4). It looks like a normal R. ulmifolius without prickles. The Salvador family was a lineage of apothecaries who settled in Barcelona from the early 17 th to the mid- 19 th century. The Salvador herbarium is the oldest and best-documented pre-Linnaean one in Spain. Its labels bear pre-Linnaean names used in the works of Tournefort, Bauhin, Lobel, Dodoens, Magnol or Clusius, among others (Ibáñez et al. 2006, 2008). Around 1782, the collection was revised for the first time by Pierre André Pourret (1754 – 1818), a French clergyman who lived in exile in Spain at Santiago de Compostela from 1789 until his death (Timbal-Lagrave 1875, Colmeiro 1891). Pourret added the Linnaean names of the species to the labels of the collection (Camarasa 1989, 2007) and, in some cases, took out duplicates for his own herbarium (Bolòs 1946), now conserved in MAF as an independent historical collection (Gutiérrez-Bustillo & Navarro Aranda 1989). Some specimens in MAF-POURRET coming from the Salvador herbarium are among the most interesting in this collection. They can be easily recognized through the word “ Salv ” on the label, indicating a new pre-Linnaean description and thus a new taxon proposed by Joan Salvador. Moreover, Pourret published some taxa in his Chloris Narbonensis (Pourret 1788) as new species under the Linnaean system. Also, he sent some of these specimens to Lamarck in Paris and to Willdenow in Berlin and they are now preserved in the P and B-W herbaria (Bonnet 1916). All three samples of Pourret’s plant show an unarmed, but for the rest normal R. ulmifolius. Van de Beek (1979, 2016) thought it was collected in a natural habitat, because Pourret in his protologue does not refer to a garden. However, on the label of the specimen in the Salvador collection is written: ‘ In hortis colitur’ (handwritten by Salvador). Salvador also wrote ‘ Romaguera de St. Francesc. ’ Because this sample is not different from the other ones, these too, may have been collected in gardens of monasteries. The so called St. Francis blackberry is often bred in Franciscan monastery gardens because of its unarmed character. This plant was already known to early authors such as Tournefort (1700) and Barrelier (1714) and validated on the level of a variety by Weston: R. fruticosus var. inermis Weston (1770: 258). The lectotype of this name was designated by Van de Beek (2016) from a specimen preserved at P (Tournefort 6078, P-TRF) (Fig. 3). The name R. inermis was used once again by Willdenow (1809: 549). It is not based on a specimen of Pourret, but on another sample in B-W: BW 09891010 (image available at https: // herbarium. bgbm. org / object / BW 09891010), from a plant in the hortus of Berlin, and so heterotypic with R. inermis Pourr., and consequently as a later homonym illegitimate. Monasterio-Huelin & Weber (1996) identified it as R. ulmifolius var. anoplosthyrsus Sudre (1909: 70) so as conspecific with R. ulmifolius. Van de Beek & Widrlechner (2021) accepted the identification by Monasterio-Huelin and Weber [Lectotype: B (BW 09891010) (selected by Monaterio-Huelin & Weber 1996: 316, pro holotype)]. However, the specimen is somewhat different from the samples of Pourret and Tournefort: it has usually 3 - foliate leaves of which the lower ones are often abaxially greenish grey (not white), with a more irregular serrature with long mucrons, small (almost) sessile lateral leaflets, ovate central leaflets, and often stipules with a large base. In the herbarium of Jussieu are three specimens of R. inermis. At one of the sheets (P-JUSS 14325) is a mixture of various species, partially with influence of R. caesius Linnaeus (1753: 493). The specimens on the other two sheets are more useful. One (P-JUSS 14333) looks like the specimens of Pourret and Tournefort, the other one (P-JUSS 14326) is similar to Willdenow’s plant. The labels do not provide further information, so that their origin is unknown. Later collections from the hortus in Paris are similar to the Willdenow plant (‘ Cult. Paris. 10 sept. 1905 ’, P 02972716; Delacourt s. n., cult. Paris, 29.08.1901, P 04173016; Delacourt s. n., cult. ormament, Paris, 07.1895, P 04181928 and P 04181929). So it seems that the form of Pourret was bred in gardens in Spain, the form of Willdenow in Berlin, and both forms also bred in Paris. Crane & Darlington (1927) researched the genetics of some Rubus taxa. They also checked what they called R. ulmifolius var. inermis. Though the specimen that they published in their paper is poor it seems to be a R. inermis Pourr., with the better developed higher leaf with stalked lateral leaflets, lateral leaflets of normal size and an obovate central leaflet. The chromosome number is 14, as with R. ulmifolius. Michal Sochor (Olomouc) was so kind to check a sample of the Willdenow form, which is cultivated in the Rubus garden in Veenendaal (The Netherlands). This one is tetraploid. In sum, as far as conclusions can be drawn from present information, it appears that R. inermis Pourret (= R. fruticosus var. inermis Weston) is an unarmed form of the diploid R. ulmifolius, while R. inermis Willd. is another, tetraploid taxon. Its precise status must be subject of further research. Thunberg (1813: 7) published once again a R. inermis. The lectotype of this name was designated by Van de Beek & Widrlechner (2021: 82) from a specimen kept at UPS (UPS-THUNB 12270) (Fig. 5). This material is identical, though heterotypic, with R. inermis Pourret (1788: 326). Both Willdenow and Thunberg related their plants to North America, where R. inermis is not found (Van de Beek & Widrlechner 2021). The confusion may be caused by a comparison with R. canadensis Linnaeus (1753: 494), which was already made by Pourret on his label (‘ an canadensis ’ – maybe canadensis?). Linnaeus (1753: 494) mentions that species as unarmed and this may have caused the suggestion that unarmed blackberries come from America.	en	Ferrer-Gallego, P. Pablo, Beek, Abraham Van De (2021): On Rubus ulmifolius (Rosaceae) and related taxa. Phytotaxa 523 (2): 155-166, DOI: 10.11646/phytotaxa.523.2.3
