Transfer of Cuban Marsdenia to Ruehssia (Apocynaceae–Asclepiadoideae), and two new species in Ruehssia

The cosmopolitan genus Marsdenia R. Br. in its conventional circumscription has been shown to be strongly polyphyletic, with all New World species grouping in a single clade. Therefore, the genus Ruehssia H. Karst. was recently reinstated to encompass the 42 Brazilian species, initially. In the present study, the nine Cuban species currently classified under Marsdenia are transferred to Ruehssia, and two species, R. lindenii and R. yamaniguey­ ensis, both endemic to Cuba, are described as new. Ruehssia yamanigueyensis is characterized by very thin, wiry stems and small leaves and flowers, while R. lindenii is like R. clausa in habit, but differs strongly in the shape of the staminal corona. Illustrations and a distribution map for two new species are provided, as well as a key to all Ruehssia species on Cuba. Lectotypification is performed where necessary.


Introduction
All Cuban species of Asclepiadoideae (Apocynaceae) possessing erect pollinia were always treated under Marsdenia R. Br., tribe Marsdenieae. Marsdenia was originally described by Brown (1810a) for five Australian species that are equipped with an urceolate corolla, a staminal corona of five flat, undivided lobes, anthers with appendages and erect, basifixed pollinia and smooth follicles. In the quasi-simultaneously published paper "On the Asclepiadeae", Brown (1810b) added two further species from outside Australia, M. tinctoria R. Br. from Indonesia and M. clausa R. Br. from Jamaica. Bullock (1957: 510) was well aware that Marsdenia in the sense of Brown (1810a, 1810b) did not constitute a homogeneous assemblage. He designated M. tinctoria as the type of the generic name, arguing that this species is most closely associated with the patron of the genus, William Marsden, a pioneer of the scientific study of Indonesia. Bullock (1957), however, was not aware that the same choice, albeit without explicit reasoning, had been made earlier by Britton & Wilson (1925).
Over the last 200 years, Marsdenia has grown to a heterogeneous genus comprising around 300 of the c. 740 species of the tribe Marsdenieae (Endress & al. 2018). Marsdenia is considered pantropical, with some 130 species occurring in the New World (Espírito Santo & al. 2019). For Cuba, it was Richard (1850) who described the first three species: M. affinis A. Rich., M. longiflora A. Rich. and M. saturejifolia A. Rich. The first revision of Marsdenia of Cuba (and the Antilles) dates back to Schlechter (1899), who accepted eight species on the island of a total of thirteen in the West Indies. However, Schlechter (1899) overlooked M. cubensis, published earlier by Turczaninow (1853), for which he cited the type (Linden 1845) under M. umbellata Griseb. Instead, he listed M. elliptica Decne., today considered a Puerto Rican endemic, for Cuba, citing Rugel 381, a specimen now identified as M. fusca Griseb. Later, only M. micran tha Alain was added as a new, then ninth species to the flora (Alain 1956). Lastly, Greuter & Rankin Rodríguez (2017) listed nine indigenous and one introduced species (M. floribunda (Brongn.) Schltr.).
On the infrageneric level, it was Rothe (1915) who proposed the first and last sectional and subsectional subdivision of New World Marsdenia (now Ruehssia H. Karst.). However, his classification, mainly based on corolla and corona characters, is artificial, as our phylogeny shows (Liede-Schumann & al. in prep.). Instead, the Antillean species analysed form a single, well-supported subclade, whereas Rothe (1915) spread the Cuban species alone over four of his nine sections. Molecular studies (Espírito Santo & al. 2019;Liede-Schumann & al. in prep.) have recovered several independent clades within Marsdenia s.l. showing that the genus is highly polyphyletic. All species of the New World are retrieved as a monophyletic group, and Espírito Santo & al. (2019)  Mangelsdorff discovered an undescribed species of "Marsdenia s.l.", which differs from all other known species by an exceptionally tiny habit and wiry stems. Later, while comparing his collections with the herbarium material available, he realized that three earlier gatherings in JE (HFC 11992, HFC 11988 and HFC 11872a), also from the vicinity of Yamanigüey, represent the new species here named R. yamanigueyensis. A second undescribed species, preliminarily labelled "Marsdenia lindenii" (in sched.), came to light during our herbarium studies and is described here.
The present work summarizes the treatment prepared for the account of the Asclepiadoideae in the Flora de la República de Cuba (Liede-Schumann & al. in prep.).

Material and methods
For our study, we analysed c. 200 specimens from the herbaria B, FR, GOET, HAC, HAJB and JE (herbarium codes according to Index herbariorum; http://sweetgum .nybg.org/science/ih/). The terminology for morphological descriptions follows Endress & al. (2018). Remarks -Karsten (1849) placed four Venezuelan species in Ruehssia, characterized by broadly elliptic leaves and shortly pedunculate, many-flowered sciadioids. The corolla tube of these species is short and rotate and the carnose, lanceolate-subulate staminal corona is almost completely adnate to the gynostegium. The follicles are exceptionally large and the pericarp is coriaceous. Presently these four species are considered synonyms of the widespread R. macrophylla (Bernal & al. 2015). Espírito Santo & al. (2019) broadened the concept of Ruehssia to accommodate all Brazilian species, so that the genus now comprises a wide range of growth forms, leaf shapes and floral structures. All species possess white latex, a corona of five lobes that are fused at least partially to the dorsal side of the anthers and mostly not exceeding the apex of the style-head, erect pollinia attached to the caudicles by their bases, and solitary follicles.  (1853) cited a single gathering, J. Linden 1846, in the respective protologues. However, neither author indicated a herbarium in which a specimen (syntype) of this gathering was housed. The KW specimen (KW 001001590), which Turczaninow probably saw, is richer and better preserved than the P specimen (P 00645929), which probably served Richard for his description, and is therefore designated as the lectotype for both names.

Key to the Cuban species of Ruehssia
For lectotypification of the Grisebach names, remarks of Gray (1860) are to be considered. In his preface to the Plantae Wrightianae, Gray (1860) stated that Grisebach examined the specimens of Wright (except lost numbers added by Gray in brackets) sent by him to Göttingen. Therefore, selection of lectotypes should use GOET specimens, as suggested by Howard (1988). Of Marsdenia campanulata, two isotypes are found in GOET, GOET 000294 and GOET 000734. Of these, GOET 000734 bears a dissected flower, a drawing and several handwritten notes, and is therefore designated as the lectotype. Finally, of the two syntypes of M. cam panulata var. bifida, C. Wright 2973 and F. I. X. Rugel 393, the specimen of Wright GOET 000295 is designated as the lectotype, again following a suggestion by Howard (in sched.). This specimen is furnished with a flower drawing and handwritten notes of Grisebach, and isolectotypes are extant in several herbaria, whereas F. I. X. Rugel 393 has only been located in MO (MO 2381344).

Reuss & al.: Ruehssia in Cuba
Remarks -Because in the protologue Turczaninow (1853) did not indicate a herbarium in which any specimen was housed, lectotypification is necessary. We designate the specimen in KW, because it is certainly the one with which Turczaninow worked.
Schlechter (1899)   Remarks -Of the two specimens of C. Wright 2976 in GOET, we designate GOET 000731 as the lectotype because it displays, besides flowers and a follicle, a flower drawing and handwritten notes, whereas GOET 00296 is a sterile specimen. Remarks -Of the three syntypes of Marsdenia linearis cited in the protologue, we designate the specimen collected by H. Nectoux and housed in P (P 00645951) as the lectotype because J. Decaisne worked in Paris and therefore his description of M. linearis is most likely based on this specimen. The second syntype, C. G. Ber tero s.n., is extant in G-DC (G 00137092), while the third syntype, P. A. Poiteau s.n., could not be traced. Marsdenia stenophylla, collected by B. Jaeger in Haiti, is here synonymized with M. linearis. Although the type was not seen, the protologue of the former is almost identical to the description of M. linearis. In particular, the linear leaves, the urceolate corolla with adaxially densely pubescent lobes and the ovate lobes of the staminal corona clearly point to this species. Remarks -In the protologue of Marsdenia longiflora, neither a specimen nor a housing institution is cited, only a note "Crescit circa Canasi". However, a single specimen of the Herbarium Richard in P (P 00645952) is encountered, bearing in La Sagra's own handwriting the inscription: "Marsdenia longiflora nob. sp. nov." as well as a short description. This specimen is designated as the lectotype of M. longiflora, according to the recommendation of McNeill (2014). This specimen obviously belongs to the original material seen by Richard, and not Wright 2974 (see Gómez 1894), actually housed in GOET and K, which was collected ten years after the publication of M. longiflora. Remarks -Richard (1850) and Turczaninow (1853) described Marsdenia saturejifolia and M. pauciflora, respectively, based on duplicates of the same gathering, J. Linden 2165. The specimen in Herb. Richard in P (P 04055872), which bears a handwritten description, is the richer one, and is therefore designated as the lectotype of M. saturejifolia. Remarks -Grisebach (1866) cited "Wr. 2972; conferatur Wr. 582" under Marsdenia umbellata. GOET 000732 bears a label stating: Coll. C. Wright, 1860 -1864, No. 2972 with "2972" heavily crossed out and the note "sent … as : 582. …" exactly corresponding to the protologue remark of Grisebach (1866). This specimen, used and annotated by Grisebach, can be regarded as the holotype of M. umbellata. The gathering Wright 2972 includes two different species: R. cubensis and R. umbellata (see above under R. cubensis). Remarks -Howard (in sched., 17 Jun 1988) considered GOET 000299, bearing a long, handwritten description and an analytical flower drawing, as the holotype of Marsdenia vinciflora. Considering also that no second specimen of C. Wright 2975 was found in GOET, and that this was the only gathering cited in the protologue, Howard's conclusion seems justified. Diagnosis -With lanceolate leaves and urceolate corolla, this new species resembles Ruehssia clausa in habit, but the staminal corona lobes in R. lindenii have a deltate shape, whereas they are almost completely reduced to scales in R. clausa.
Distribution and ecology -Ruehssia lindenii occurs all over Cuba. As far as reported, it can be found in drier to humid, semi-deciduous, sublittoral to hilly bushland near the coast, or in cliffs (Fig. 3).
Conservation status -Occurring all over Cuba, this species could be categorized as Least Concern (LC). However, because the species seems to be nevertheless rare, and actual data on flourishing populations are missing, the category Data Deficient (DD) according to IUCN Red List categories and criteria (IUCN 2012) seems appropriate.
Etymology -This new species is named Ruehssia lindenii to honour the heritage of Jean Jules Linden (1817 -1898), a Belgian botanist, who made many collections not only in Cuba (especially eastern Cuba), picking up a misnomer / erroneous labelling present in the Berlin herbarium ("Marsdenia lindenii?",in sched.,B 10 0374603;H. Manitz,pers. comm.). Linden is known as the father of modern orchid horticulture. He initiated the construction of greenhouses in Gent and Brussels to cultivate tropical plants that were sold even to the Russian Tsar. Linden also collected the type specimens of M. affinis, M. cubensis and M. saturejifolia.
Phenology -Flowering from January to April.
Distribution and ecology -To date, Ruehssia yama nigueyensis has been collected only in the province of Holguín (Moa), where it seems to be limited in distribution to the surroundings of Yamanigüey. It typically grows in charrascal (bosque esclerófilo humedo), a moist, sclerophyllous bushland on ultramafic soil at low elevations of up to 400 m (Fig. 3).
Conservation status -With an area of occupancy of less than 500 km² and just two gatherings from one (or two?) locations known, Ruehssia yamanigueyensis could be categorized as Endangered (EN). However, because the data are so poor, the category Data Deficient (DD) according to IUCN Red List categories and criteria (IUCN 2012) seems appropriate.
Etymology -This species is named after the type locality in the province of Holguín, near the village of Yamani güey.
Remarks -Ruehssia yamanigueyensis is clearly separated from all other species of Ruehssia in Cuba by its wiry habit and adaxially scabridulous leaves ( Fig. 2A) of just about 10 × 2.5 mm (making it the tiniest of all Cuban species of Ruehssia). Likewise, the flowers are diminutive, hardly reaching 3 mm in length (Fig. 2).