We report the rediscovery of Rubus pendulus Rusby, “Mora India”, described in 1933 from Colombia and not mentioned again until the present study. We also update its distribution with eight new localities in Colombia, seven in Ecuador and one in Peru, being a new record for the flora of the latter two countries. This is the first time that R. pendulus’ stipules and flowers are found and detailed through a botanical description, illustrations and photographs. Rubus pendulus is morphologically differentiated from R. bogotensis Benth., R. mollifrons Focke, R. porphyromallos Focke and R. urticifolius Poir., with whom it was previously confused and we give a brief explanation on the type specimen status of R. mollifrons and R. porphyromallos.

Andes, Ecuadorian, Rubeae, taxonomy

Rubus L. presents ca 836 species classified in 10 subgenera; thus, being one of the most diverse genera of the Rosaceae family (Huang et al. 2023). Despite the cosmopolitan distribution of the genus, it is more abundant in Asia, where at least 208 species (139 endemic) were reported in China alone (Lingdi and Boufford 2003). South America contains a low Rubus diversity with fewer than 60 species classified in the native subgenus Rubus L., and the introduced subgenera Batothamnus (Focke) E.H.L.Krause and Idaeobatus Focke (Macbride 1938; Romoleroux 1996; Forzza et al. 2010; Romoleroux et al. 2014; Bernal et al. 2020; Espinel-Ortiz and Romoleroux 2020, 2021; Moreno-Medina et al. 2020; Huang et al. 2023). Thus, more focalized taxonomic studies will help to uncover more species of Rubus in the Neotropics.

The “Mora India”, Rubus pendulus Rusby (subgenus Rubus), is among the few Rubus’ vine species in South America. Its holotype is the only collection, and thus, the only locality known for this species, and its flowers’ description is non-existent (Rusby 1933). Despite the type locality in Colombia, this name was excluded from the Colombian Plant Catalogue (Bernal et al. 2020), and has not been mentioned for over 90 years. Curiously, this species is not the only one that should be revised for the Colombian flora. Rubus mollifrons Focke and R. porphyromallos Focke also should be reviewed, as both do not have any type collection. In fact, during their original description Focke (1911b) did not cite any sample, but affirmed that R. mollifrons is found in Colombia and R. porphyromallos could inhabit the South American Andes.

After careful examination of more than 3000 Rubus samples from different herbaria, specimens representing this species showed only a few samples and were often annotated as R. bogotensis Benth. R. mollifrons, R. porphyromallos, R. urticifolius Poir. or were unidentified. However, R. pendulus vegetative and reproductive characters differ greatly from those of the species reported for Colombia, Ecuador and Peru (Macbride 1938; Romoleroux 1996; Bernal et al. 2020; Moreno-Medina et al. 2020; Espinel-Ortiz and Romoleroux 2020, 2021). Here, we provide information to support Rubus pendulus as a valid species, its differences from other Rubus species and a brief explanation of the type status of R. mollifrons and R. porphyromallos. We also updated the key for the Ecuadorian Rubus species.

The Rubus collections of the Herbaria HA, HUTI, LOJA, NY, Q, QAP, QCA, QCNE, QPLS and QUSF were revised, and samples not fitting with the species reported for Ecuador were studied. Additional samples from AAU, COL, F, MO and US were revised from online images to cover the original distribution of this species in Colombia and see if it reached Peru. In total, ca 2500 samples of Ecuador, ca 700 samples of Colombia and one of Peru were revised. During 2021, we recollected more material near Quito (Ecuador) in several field trips in order to complete the species descriptions of its flowers and update the ecological data. A taxonomic key for the Ecuadorian Rubus’ species is provided as supplementary material (See Suppl. material 1).

To categorize R. pendulus as a valid species, we used the Rubus species definition proposed by Weber (1977) and Haveman and Ronde (2013). These proposals suggest that a widely-distributed biotype, whose diameter of distribution area goes from 500 km to more than 1000 km, can be considered a species.

The botanical terms used in the descriptions followed those used by Stearn (1986), and the pubescence types were based on the terminology of Hickey and King (2000), and Wilhelm and Rericha (2020). Some specimens examined for the description (e.g. D. Espinel-Ortiz & H.G. Abad 300) were mounted in more than one herbarium sheet, and/or have additional dry or alcohol material; therefore, each part had its own herbarium barcode (QAP and QCA). For these samples, we wrote all the herbarium barcodes for each part in examined specimens when available. We used QGIS (QGIS.org 2022) for the distribution map, using the geographic coordinates from the samples. Additionally, geographic reference coordinates based on locality description were selected for the samples from Colombia that lacked this information.

Rubus pendulus was described by Rusby (1933) with a sample collected at Balsilla’s river, Balsillas, Huila Department in Colombia. All the revised material agrees with the original description and resembles the holotype collection. The most conspicuous characteristics from both, the holotype and the material examined here, are the setose hairs and hirsute pubescence referred in by Rusby as ferrugineous-tomentose, long petioles (4.8–10.6 cm long), bullate leaves, long and slightly thin leafletes (5.3–15.1 × 2.5–7.3 cm) with subcordate base, the secondary veins pattern (at an angle of about 45 degrees with the main vein), concave sepals with a mucronulate apex referred by Rusby as apiculate, and compact (crowded) and nearly subsessile fruits with small drupelets (2.1–4.3 × 1.1–2.8 mm).

Since its description (Rusby 1933), no other records of R. pendulus were reported until Romoleroux (1996) suggested that sample Lægaard 55178 (AAU) may belong to this species. However, the available material could not be properly identified as it lacked flowers and fruits. During this revision, several samples annotate here as R. pendulus, were previously identified as R. bogotensis, R. mollifrons, R. porphyromallos or R. urticifolius. From these, only the latter two show similar characteristics to R. pendulus, whereas R. bogotensis and R. mollifrons have almost nothing in common with it.

Rubus bogotensis is characterized for is abundant shortly stipitate glands covering all the plant, absence of setose hairs, long pedicels (5–20 mm long) and big fruits (15–20 × 10–20 mm) with only a few drupelets (10–35) per fruit (Romoleroux, 1996). On the other hand, R. pendulus has red setose hairs covering all the plant, glands only in some of the setose hairs, shorter pedicels (2.4–7.5 mm long) and smaller fruits (7.8–15.4 × 6.6–11 mm) with more drupelets (66–115) per fruit. In addition, R. pendulus’ bullate leaves differentiate it from R. bogotensis and the other species, as this is only a characteristic previously found in R. azuayensis Romol. and R. betonicifolius Focke, both simple-leaf species (See Suppl. material 2).

Rubus pendulus may resemble R. urticifolius by its red setose hairs, mostly eglandular trifoliate to 5-foliate leaves, and ovate to elliptic leaflets, but it differs from the latter by its bullate leaves, few flowered inflorescences (up to 60 flowers), and mucronulate sepals in contrast with the non-bullate leaves, many flowered inflorescences (60–150 flowers), and apiculate or acuminate sepals of R. urticifolius. Furthermore, R. pendulus has bigger fruits (7.8–15.4 × 6.6–11 mm) with more (66–115) and bigger drupelets (2.1–4.3 × 1.1–2.8 mm), whereas R. urticifolius has smaller fruits (7–10 × 6–9 mm) with fewer (30–50) and smaller drupelets (1.5–3 × 1–2 mm) (See Suppl. material 2).

The two species mentioned before were registered in Colombia, Ecuador, Peru and Bolivia (Romoleroux 1996). However, R. mollifrons was recorded only from Colombia and R. porphyromallos was said to inhabit the South American Andes (Focke 1911a, 1911b). These two species lack a holotype as Focke did not mention any sample during their original description (Focke 1911b). Even in his monograph (Focke 1911a), no sample was cited for either species. Following the Shenzen Code’s art. 9, as no sample or illustration was presented in the protologue, neotypes should be selected for both species (Turland et al. 2018). Luckily, Focke (1911a) included a photograph of Rubus mollifrons collection in his monograph; should this sample be found, it could be designated as the neotype for this species. On the other hand, no illustration of R. porphyromallos was included; therefore, a neotype following the species description should be selected. However, for both cases, before the designation of any neotype, an extensive revision of historic Colombian Rubus samples is necessary. For this reason, here, we proceed to differentiate R. pendulus from both species based on their descriptions (Focke 1911a, 1911b). Also, we give a brief explanation of the identity of the samples identified as R. porphyromallos in COL and NY.

Rubus mollifrons is described as a climbing shrub with tomentose stems; short, tomentose petioles, lateral petioles ca 1 cm long, and terminal petiolule ca 2 cm long; linear-lanceolate stipules; trifoliate leaves, with leaflets oblong-ovate, base subcordate, apex acuminate, 6–8 × 4–5 cm and 10–12 secondary veins; leaf adaxial surface densely pubescent and abaxially grayish-pannose (“canescent-velutina”); the inflorescences are grayish-tomentose, subarmed, pauciflora or uniflora; the flowers are short-peduncled, ca 5 cm; sepals ovate and grayish-tomentose; petals elliptic, white or slightly pink on the outside, and the petals are shorter than the sepals; no fruits observed (Focke 1911a, 1911b).

Focke (1911a, 1911b) described R. porphyromallos as a shrub covered in two kinds of pubescences: the first one is said to be “rufous-villous” or reddish-villous, and the other one tomentose. The stems are eglandular, with prickles; long petioles with prickles, basal petiolules 2–2.5 cm long, lateral petiolules ca 4 cm long, terminal petiolules 5–6 cm long; palmate-compound 5-foliolate leaves, leaflets oblong-ovate or ovate, base emarginate or subcordate, apex pointed, margin unevenly serrated, 15 × 10 cm, leathery, with 12–15 secondary veins; adaxial surface strigose; abaxial surface softly grayish-pannose (“canescenti-velutina”), young leaflets white; broad compound inflorescences, tomentose-villous, with prickles and trifoliate leaves; flowers shortly pedicellate, ca 1.5 cm diam.; sepals ovate, apex acute or minute, greyish-tomentose, not villous; petals obovate; stamens shorter than sepals; no fruits observed.

In Focke’s original description, the latin word “velutina” is literally translated to velvet; the equivalent pubescence is pannose, as he used the same word to describe the pubescence of R. boliviensis holotype which is pannose (Romoleroux, 1996). It is also worth mentioning that Focke already used the term bullate in R. betonicifolius, and red-setose pubescence as “rufo-setosi” in R. urticifolius (annotated as “R. urticaefolius”) (Focke 1911a), as both are among the most conspicuous characteristics of R. pendulus and were not mentioned for either R. mollifrons or R. porphyromallos (Focke 1911a, 1911b). Rubus pendulus differs from R. mollifrons by having red setose hairs all over the plant, longer lateral (0.7–2.97 cm long) and terminal (2.8–5.1 cm long) petiolules, bullate leaves, ovate to elliptic and bigger (5.3–15.1 × 2.5–7.3 cm) leaflets, with more secondary veins (10–19), leaf abaxial surface glabrous with red setose hairs, and villous only on the veins, more flowers (6–56) per inflorescence, shorter peduncles (0.8–2.0 mm long), sericeous-villous and tomentose sepals (See Suppl. material 2).

In the case of R. porphyromallos, its description has similar characteristics to that of R. pendulus, but it presents some differences such as eglandular stems, longer basal petiolules (2–2.5 cm long), non-bullate leaves, broader leaflets (15 × 10 cm), leaf abaxial surface pannose, ovate and greyish-tomentose, not villous sepals. Whereas Rubus pendulus has some red setose hairs ending in glands, shorter basal petiolules (0.3–0.7 cm long), bullate leaves, thinner leaflets (5.3–15.1 × 2.5–7.3 cm), leaf abaxial surface glabrous with red setose hairs, and villous only on the veins, obovate to elliptic or slightly lanceolate, sericeous-villous and tomentose sepals (See Suppl. material 2). Another difference could be the absence of setose hairs in R. porphyromallos as this term was not mentioned in its original description (Focke 1911b).

Rubus pendulus is a widely spread species from the north of South America that has been poorly collected before and thus confused with different species. However, morphologically it is different from other similar species. More collection efforts are necessary to have an assessment of this species’ complete distribution. Additionally, as R. porphyromallos showed the closest resemblance to R. pendulus, it is fundamental to designate a neotype for R. porphyromallos and study both of them genetically to understand their evolutive history.

We would like to thank Pontificia Universidad Católica del Ecuador (PUCE) for funding this research under the investigation projects: “Caracterización de la diversidad genética y fenología de Rubus ellipticus Sm. (Rosaceae), especie introducida en Ecuador”, “Morfometría de las especies del género Rubus L. de Ecuador (PUCE)” and “Evaluación de la distribución de las especies introducidas en Ecuador del género Rubus L. en base a información de colecciones botánicas de los herbarios ecuatorianos (PUCE)”. Ministerio de Ambiente, Agua y Transición Ecológica (MAATE) for research permit MAAE-DBI-CM–2021–0171. Gonzalo Abad and Martín Morocho for their assistance during field trips. We would also like to thank Carlos Cerón, and Carmita Reyes for collecting additional material of R. pendulus in Yunguilla and Jorge A. Pérez for the suggestions to improve this work.