Nordic Journal of Botany 33: 513–517, 2015 doi: 10.1111/njb.00688, ISSN 1756-1051 © 2015 he Author. Nordic Journal of Botany © 2015 Nordic Society Oikos Subject Editor and Editor-in-Chief: Torbjörn Tyler. Accepted 27 January 2015 Memecylon trunciflorum sp. nov. (Melastomataceae–Olisbeoideae) from the Udzungwa Mountains, southern Tanzania Robert Douglas Stone R. D. Stone (stonerd@ukzn.ac.za), School of Life Sciences, Univ. of KwaZulu-Natal, Private Bag X01, Scottsville 3209, South Africa. Described and illustrated is Memecylon trunciflorum R. D. Stone, an evidently localized endemic of the Udzungwa Mountains in southern Tanzania. he new species was previously confused with the vegetatively similar but distantly related M. erythranthum Gilg and M. semseii A. Fern. & R. Fern., from which it is distinguished by its anther connectives bearing a dorsal oil-gland and by its ellipsoid to obovoid fruits (vs anther connective gland absent and fruits globose in M. erythranthum and M. semseii). he new species is placed in M. sect. Magnifoliata R. D. Stone together with M. magnifoliatum A. Fern. & R. Fern., from which it difers by its smaller leaves mostly 9.5–15.0  3.5–6.0 cm (vs 18–35  8–13 cm), transverse veins 8–18 pairs (vs 25–28 pairs), short-pedunculate inlorescences with secondary axes well developed (vs peduncles and secondary axes absent), white lowers (vs bluish purple), and smaller fruits mostly 11.5– 14.5  9–11 mm on longer fruiting pedicels 8.0–13.5 mm (vs fruits 17–20  12–14 mm on pedicels 5.0–7.5 mm). Despite its local endemism, Memecylon trunciflorum has been assessed as ‘Least Concern’ according to IUCN criteria, although this assessment is dependent on the continued safeguarding of the Udzungwa Mountains National Park. he woody paleotropical genus Memecylon L. (Melastomataceae;  350 spp.) is now circumscribed to exclude Spathandra Guill. & Perr., Lijndenia Zoll. & Moritzi and Warneckea Gilg (Jacques-Félix 1978, Bremer 1982, Stone 2006, 2014, Stone and Andreasen 2010). In the treatment of Memecylon sensu lato for the ‘Flora of Tropical East Africa’ (Wickens 1975), the 10 species with strongly trinervate to multinervate leaves are properly placed in Warneckea (leaf-lamina smooth) or Lijndenia (leaf-lamina papillose-muricate in dried material), whereas the 13 species with apparently uninervate leaves belong to Memecylon sensu stricto. Additional characteristics distinguishing Memecylon sensu stricto from the other memecyloid genera include its leaf surface often appearing roughened (not granular) on drying and ibrous when torn (owing to the presence of iliform to columnar sclereids in the mesophyll); lowers mostly with a  truncate or sinuate calyx margin, the corolla in bud thus well exposed; uni-locular (rarely 4-locular) ovary; and embryo with foliaceous, extensively convoluted cotyledons (Jacques-Félix 1977, 1978, Jacques-Félix et al. 1978, Rao and Jacques-Félix 1978, Jacques-Félix and Mouton 1980, Bremer 1981, 1982, Rao et al. 1983). Borhidi (1993) transferred three East African Memecylon species (i.e. M. greenwayi Brenan, M. fragrans A. Fern. & R. Fern., and M. semseii A. Fern. & R. Fern.) to Lijndenia, but recent molecular analyses (Stone 2014) clearly show that these taxa should be placed in Memecylon sensu stricto. Borhidi's (1993) transfers of M. jasminoides Gilg and M. bequaertii De Wild. to Lijndenia are also inappropriate, and these species must be included in Warneckea where they were previously placed by Jacques-Félix (1978). Within Memecylon sensu stricto, molecular results have supported a sister-group relationship between M. subg. Mouririoidea (Jacq.-Fél.) R. D. Stone, representing a small group of western and central African species with 4-loculed ovary, and M. subg. Memecylon representing the remaining species which have 1-loculed ovary (Stone 2014). In tropical East Africa, seven morphologically well-deined, monophyletic species-groups are currently recognized as sections within M. subg. Memecylon: sect. Magnifoliata R. D. Stone, sect. Obtusifolia Engl., sect. Polyanthema Engl. sensu stricto, sect. Tenuipedunculata Engl., sect. Cauliflora Engl., sect. Montana R. D. Stone, and sect. Buxifolia R. D. Stone (Stone 2014). his paper concerns Memecylon sect. Magnifoliata, which evidently occupies an evolutionarily isolated position within M. subg. Memecylon, and is distinguished by its caulilorous habit, anther connectives bearing a dorsal oil-gland, and ellipsoid to obovoid fruits (Stone 2014). his section was previously treated as including a single species, i.e. the endangered Tanzanian endemic M. magnifoliatum A. Fern. & R. Fern. (Fernandes and Fernandes 1969, Wickens 1975, Vollesen 1980, IUCN 2013a). For this species there is still some ambiguity about the overall geographic range, but all of the material I have seen was collected in or near the Magombera Forest Reserve, an isolated fragment of lowland forest (approximate area 10 km2), situated at 220–300 m 513 a.s.l. between the Udzungwa Mountains National Park to the west and the Selous Game Reserve to the east (Rodgers and Homewood 1982, Marshall 2008). he species has also been reported from the Rubeho Mountains (Ukwiva Forest Reserve), based on the collection Mhoro BM16529 cited in Tropicos (2014), although this cannot be conirmed as I have not seen the specimen. Reports of M. magnifoliatum from the Selous Game Reserve would seem to be in error, as all of the specimens cited in Vollesen’s (1980) checklist are from the Magombera Forest. Here I describe a distinctive, smaller-leaved species in Memecylon sect. Magnifoliata from the Mwanihana Forest (now part of Udzungwa Mountains National Park), situated on the steep, east-facing escarpment of Tanzania’s Udzungwa Mountains (Lovett and Pócs 1993). his species was irst collected in the early 1980s and cited as “Memecylon sp. nov.” in Lovett et al. (1988), but has been confused with the vegetatively similar but distantly related M. erythranthum Gilg and M. semseii, both of which are now placed in M. sect. Cauliflora (Stone 2014). he new species had not been previously compared with M. magnifoliatum, but is strongly supported as its sister-taxon in phylogenetic analyses based on 5’ ETS and ITS sequences of nuclear ribosomal DNA (Stone 2014). All specimens cited have been seen unless otherwise indicated. Memecylon trunciflorum R. D. Stone sp. nov. (Fig. 1) Previously referred to as: “Memecylon sp. nov. 2” (IUCN 2013b). Type: Tanzania, Sanje, Uzungwa [Udzungwa] Mts, 960 m a.s.l., in forest near a clif bottom, on rocky ground, 3 Nov 1983, Lovett 215 (holotype: K). Description Shrub or tree to 15 m high; bark pale to dark brown, longitudinally issured; young branchlets terete in cross section; internodes (1.5–)3.0–6.0(–8.0) cm long. Leaves subcoriaceous; young leaves white to pale purple becoming dark green adaxially when mature, paler abaxially; petioles stout, 2–6 mm long, sulcate adaxially, convex abaxially; blades elliptic to ovate, (8.0–)9.5–15.0(–18.0) cm long, (2.7–)3.5–6.0(–6.7) cm wide, cuneate to rounded at base, acuminate at apex (the acumen varying from 5–11 mm long,  triangular in outline and rounded at apex, to 10–20 mm long, narrow and sharply acute); mid-nerve inely impressed adaxially, prominent abaxially; transverse veins 8–18 pairs, much weaker than the mid-nerve and oriented at a somewhat oblique angle (varying to subperpendicular), spaced 4–9(–12) mm apart, obscure adaxially,  prominent abaxially in dried material; lateral nerves situated 1–2 mm from the margins, of about the same strength as the transverse veins. Cymules subumbellate, densely aggregated in swollen patches on the trunk and branches; peduncles stout, 2–3 mm long; secondary axes 1–6(–8) mm long; pedicels slender, up to 5 mm long (but perhaps not yet fully elongated, see below); bracts lanceolate, scarious, ∼1/2 mm long, channeled adaxially, keeled abaxially, deciduous. Flowers at anthesis not seen; hypantho-calyx  obconic in bud but with rounded base, 2 mm long, 2 mm wide, minutely granular on 514 the outer surface, its margin shallowly 4-lobed, its lobes scarious-margined. Corolla well exposed in bud, reported to be white, ∼2 mm long, twisted, acute at the apex. Anthers with conspicuous oil-gland on dorsal side (the gland ellipsoid, 1/2 mm long). Ovary uni-locular; ovules 8. Fruits ellipsoid to obovoid, described by collectors as blue, glaucous blue, or blue-green, (9.0–)11.5–14.5 mm long, (7–)9–11 mm wide, with a persistent calycinal crown  1 mm high, appearing truncate but actually shallowly sinuate and remotely 4microdentate; fruiting pedicels slender but irm, 8.0–13.5 mm long. Distribution, habitat, abundance and conservation status Memecylon trunciflorum is evidently endemic to the Mwanihana Forest, located along the steep, east-facing escarpment of the Udzungwa Mountains in southern Tanzania (Fig. 2). his area of approximately 180 km2, irst gazetted as a Forest Reserve in 1958, was included within the boundaries of the larger Udzungwa Mountains National Park in 1992. Habitats in the area cover a wide elevational range (between 300–2080 m a.s.l.) and include lowland, transitional (submontane), montane and upper montane forest types (Rodgers and Homewood 1982, Lovett et al. 1988, Lovett and Pócs 1993). According to data provided on specimen labels, the habitat of M. trunciflorum is in submontane forest (rarely riverine forest in wooded grassland) at 820– 1500 m elevation. Associated tree species include Cephalosphaera usambarensis (Warb.) Warb. (Myristicaceae), Englerophytum magalismontanum (Sond.) T. D. Penn. (Sapotaceae), Maytenus acuminata (L. f.) Loes. (Celastraceae), Myrianthus holstii Engl. (Urticaceae), Newtonia buchananii (Baker f.) G. C. C. Gilbert & Boutique (Fabaceae), Parinari excelsa Sabine (Chrysobalanaceae), Sibangea pleioneura Radcl.-Sm. (Putranjivaceae), Synsepalum cerasiferum (Welw.) T. D. Penn. (Sapotaceae), and Uapaca paludosa Aubrév. & Leandri (Phyllanthaceae). he label of the collection Lovett 317 (K) described Memecylon trunciflorum as “common”, but it is unclear whether this meant common locally or perhaps throughout the area of submontane forest. Despite the evident restriction of M. trunciflorum to a single population and its very limited area of occurrence, the IUCN (2013b) has assessed it (as Memecylon sp. nov. 2) as being of ‘Least Concern’, citing an absence of documented or observed threats. It was noted however that this assessment is contingent on the continued efectiveness of conservation measures (safeguarding of the Udzungwa Mountains National Park). Similar species Memecylon trunciflorum is placed in M. subg. Memecylon due to its shallowly lobed calyx (corolla well exposed in bud) and uni-locular ovary. It is further placed close to M. magnifoliatum in M. sect. Magnifoliata on account of its combination of young branchlets terete (not quadrangular) in cross-section, inlorescences in dense clusters on the trunk and older branches, anther connectives bearing a dorsal oil-gland, and ellipsoid to obovoid fruits (Stone 2014). When M. magnifoliatum was irst described (Fernandes and Fernandes 1969), it was known from a single fruiting collection (Semsei 3375, holotype: COI, isotypes: EA, K); Figure 1. Memecylon trunciflorum. (A) leafy branch, (B) leaf, (C) detail of lower leaf surface, (D)–(F) variation in leaf apices, (G) lowering branch, (H), inlorescence, (I) lower before anthesis, (J) anther with connective bearing a dorsal oil-gland, (K) infructescence, (L) top view of fruit showing calycinal crown with scars marking points of attachment of stamens and petals, (M) bract. Voucher specimens: (A)–(E), (G)–(J) Lovett 215, K; (F), (K)–(M) Lovett 317, K. Scale bars: (A)–(B), (D)–(G) 20 mm; (K) 10 mm; (C), (H), (L) 5 mm; (I) 2 mm; (J), (M) 1 mm. Drawing by Sandie Burrows. its lowers were later described by Vollesen (1980), and his Fig. 8 (illustration of the loral parts of M. magnifoliatum) clearly shows the presence of an anther-gland. I have conirmed the presence of this anther-gland, and of a uni-locular ovary with 7–9 ovules, in material of the collection Vollesen MRC 4430 sent on loan from C to NU. Amongst other East African members of Memecylon, an anther gland is also found in the distantly related sections Obtusifolia (M. flavovirens Baker), Polyanthema (M. myrianthum Gilg), and Buxifolia, but the gland is completely absent in the sections Tenuipedunculata, Cauliflora and Montana (Stone 2014). he new species M. trunciflorum difers from M. magnifoliatum by its smaller leaf dimensions (mostly 9.5–15.0  3.5–6.0 cm vs 18–35  8–13 cm), leaf-lamina with fewer transverse veins (8–18 vs 25–28 pairs), pedunculate inlorescences with secondary axes well developed 515 Figure 2. Map of the northern Udzungwa Mountains and adjacent Msolwa river valley, Tanzania, showing known localities for Memecylon magnifoliatum and M. trunciflorum. Major forests are named. he inset at upper left shows the location of the Udzungwa Mountains within the Eastern Arc Mountains (in grey, with forest patches shown in black) of Tanzania and Kenya. Geographic coordinates of M. magnifoliatum and M. trunciflorum furnished by Roy Gereau. Limits of forested areas in the northern Udzungwas according to a land-use classiication by the GIS services branch of the International Livestock Research Institute (2002). Limits of the Eastern Arc Mountains (inset) are from Platts et al. (2011a, 2011b). (peduncles 2–3 mm, secondary axes 1–8 mm long vs peduncles and secondary axes absent), and smaller fruits on longer fruiting pedicels (fruits mostly 11.5–14.5  9–11 mm on pedicels 8.0–13.5 mm vs fruits up to 17–20  12–14 mm on pedicels 5.0–7.5 mm). here may also be a diference in lower-color, as the label of the holotype specimen of M. trunciflorum indicates “lowers white” while the calyx and petals of M. magnifoliatum were described by Vollesen (1980) as “bluish purple”. In the ‘Flora of Tropical East Africa’ (Wickens 1975), the new species M. trunciflorum would key to M. semseii on account of its apparently uni-nervate, elliptic to ovate leaves mostly 9.5–15.0  3.5–6.0 cm, and by its inlorescences with peduncle present but less than 10 mm long. Previous authors have also determined the new species as M. af. semseii, either on specimen labels or in the literature (Lovett et al. 1988, in reference to the collection homas 3702). However, M. trunciflorum can be distinguished from M. semseii by its caulilorous habit (vs inlorescences in the leaf axils and at the defoliated nodes of older branchlets), anther connectives bearing a dorsal oil-gland (vs gland absent), and ellipsoid to obovoid fruits (vs fruits globose). In addition, the geographic range of M. semseii is evidently limited to Tanzania’s West and East Usambara Mountains (Wickens 1975, R. D. Stone 516 unpubl.), and the species is not known to occur in or near the Udzungwas. Memecylon erythranthum is another species that resembles M. trunciflorum in its vegetative characteristics and caulilorous habit, and some previous collections of M. trunciflorum have indeed been misidentiied as M. af. erythranthum. he new species difers from M. erythranthum by its short-pedunculate inlorescences (vs sessile), anther connectives bearing a dorsal oil-gland (vs gland absent), and ellipsoid to obovoid fruits (vs fruits globose). Memecylon erythranthum occurs in forests of the East Usambara, Nguru and Uluguru mountains of Tanzania and in the Shimba Hills of southeastern Kenya (Wickens 1975, Luke 2005, R. D. Stone unpubl.); it is not known from in or near the Udzungwa Mountains, although there is a report of this species from the eastern highlands of Zimbabwe (Fernandes and Fernandes 1978, based on Swynnerton 190, Oct 1912, BM). I have not seen this specimen but doubt it is correctly identiied. Also resembling M. trunciflorum is the specimen Ede 67 (K), collected from the Iringa District, locality Nyambanitu [Nyumbanitu] which is in the western part of the Udzungwa Mountains in Kilombero Nature Reserve (7°48′S, 36°21′E). his collection, described as Memecylon sp. A in the ‘Flora of Tropical East Africa’ (Wickens 1975), is from a tree 15 m high with elliptic-ovate, short-acuminate leaves 8–11  4–6 cm. It is not included in the circumscription of M. trunciflorum although I think the two species may be closely related. Ede 67 difers from M. trunciflorum in having leaves somewhat shorter relative to their width, longer petioles 5–8 mm (vs 2–6 mm), secondary leaf venation quite obscure (vs 8–18 pairs of transverse veins prominent abaxially in dried material), shorter fruiting pedicels 5 mm (vs 8.0–13.5 mm), and fruits larger and subglobose (15.5– 17.0 mm  15–16 mm) with calycinal crown  absent (vs fruits ellipsoid to obovoid, mostly 11.5–14.5  9– 11 mm with calycinal crown persistent,  1 mm high). he lowers of Memecylon sp. A sensu Wickens (1975) are still unknown, and the species remains formally undescribed and unnamed. Additional specimens examined (paratypes) Tanzania, Uzungwa [Udzungwa] Mountains, Mwanihana Forest Reserve, Sanje River area, 3000–4500 ft a.s.l., moist intermediate forest, Nov 1981, Rodgers and Hall 1361 (DSM, K); Sanje, Uzungwa [Udzungwa] Mts, 1000– 1500 m a.s.l., in forest, 15 Jun 1984, Lovett 317 (K); Morogoro Region, Kilombero District, Mwanihana Forest Reserve above Sanje village, 7°50′S, 36°55′E, 1300 m a.s.l., forest on steep mountainside with small streams, 13 Sep 1984, homas 3673 (MO); same locality, 1100–1400 m a.s.l., ridgetop forest, 22 Sep 1984, homas 3702 (MO, P); Iringa Region, Udzungwa Mountains National Park, 7°46′S, 36°49′E, 1100 m a.s.l., riverine patch in wooded grassland, 27 Sep 2001, Luke et al. 7868 (EA); same locality, 1250 m a.s.l., submontane forest, 29 Sep 2001, Luke et al. 7989 (EA, K, NHT n.v.); Udzungwa Mountains National Park, Sonjo route above Mzimu, 7°48′S, 36°51′E, 820 m a.s.l., submontane forest, 20 Oct 2002, Luke et al. 9263 (EA n.v., K). Acknowledgements – he curators of the following herbaria are thanked for providing loans of or access to specimens: BR, C, CAS, DSM, EA, K, MO, P, PRE, UPS. Special thanks to Quentin Luke for providing study material of this new species and the closely related M. magnifoliatum. Sandie Burrows rendered the line drawing of M. trunciflorum. Roy Gereau furnished valuable comments on a previous version of the manuscript, and provided geographic coordinates of M. magnifoliatum and M. trunciflorum. References Borhidi, A. 1993. Warneckea hedbergorum sp. nov. (Memecylaceae) and a short review of the East-African Memecylon s.l. – Opera Bot. 121: 149–151. Bremer, K. 1981. Seeds and embryos in Sri Lanka (Ceylonese) species of Memecylon, with notes on Spathandra (Melastomataceae). – Nord. J. Bot. 1: 62–65. Bremer, K. 1982. Lijndenia, a re-established paleotropical genus of the Melastomataceae – Memecyleae. – Nord. J. Bot. 2: 121–124. Fernandes, A. and Fernandes, R. 1969. Melastomataceae africanae novae vel minus cognitae. V. – Bol. Soc. Brot., sér. 2, 43: 285–306 ( 17 plates). Fernandes, R. and Fernandes, A. 1978. Melastomataceae. – In: Launert, E. (ed.), Flora Zambesiaca. Vol. 4. R. Bot. Gard. Kew, pp. 220–276. International Livestock Research Institute, GIS Services 2002. Tanzania land use. – < http://192.156.137.110/gis/search. asp?id=543>, accessed 25 Jan 2015. IUCN SSC East African Plants Red List Authority 2013a. Memecylon magnifoliatum. he IUCN red list of threatened species, ver. 2014.2. –  www.iucnredlist.org , accessed 14 Aug 2014. IUCN SSC East African Plants Red List Authority 2013b. Memecylon sp. nov. 2. he IUCN red list of threatened species, ver. 2014.2. –  www.iucnredlist.org , accessed 14 Aug 2014. Jacques-Félix, H. 1977. La graine et l’embryon chez les Memecylon (Mélastomatacées) africains. – Adansonia, sér. 2, 17: 193–200. Jacques-Félix, H. 1978. Les genres de Memecyleae (Melastomataceae) en Afrique, Madagascar et Mascareignes. – Adansonia, sér. 2, 18: 221–235. Jacques-Félix, H. and Mouton, J. A. 1980. Identiication des Memecyleae (Melastomataceae) de l’Ouest-Africain d'après leurs caractères végétatifs. – Bull. Mus. Natl Hist. Nat. B Adansonia, sér. 4, 2: 3–19. Jacques-Félix, H. et al. 1978. Nervation et types foliares chez les Memecylon (Melast.) africains. – Adansonia, sér. 2, 18: 67–81. Lovett, J. C. and Pócs, T. 1993. Assessment of the condition of the Catchment Forest Reserves, a botanical appraisal. Morogoro Region. – Report prepared for the Catchment Forestry Project (under the Forestry and Beekeeping Division of the Ministry of Tourism, Natural Resources and Environment, Tanzania). Lovett, J. C. et al. 1988. A preliminary list of the moist forest angiosperm lora of Mwanihana Forest Reserve, Tanzania. – Ann. Miss. Bot. Gard. 75: 874–885. Luke, Q. 2005. Annotated checklist of the plants of the Shimba Hills, Kwale District, Kenya. – J. East Afr. Nat. Hist. 94: 5–120. Marshall, A. R. 2008. Ecological report on Magombera forest. – Commissioned by World Wide Fund for Nature, Tanzania Programme Oice, Dar es Salaam. Platts, P. J. et al. 2011a. Delimiting tropical mountain ecoregions for conservation. – Environ. Conserv. 38: 312–324. Platts, P. J. et al. 2011b. Shape iles for the eastern Arc Mountains in Tanzania and Kenya. – < http://dx.doi.org/10.5061/dryad. c5310 >, accessed 25 Jan 2015. Rao, T. A. and Jacques-Félix, H. 1978. Les types de sclérites foliaires et la classiication des Memecylon africains. – Adansonia, sér. 2, 18: 59–66. Rao, T. A. et al. 1983. Foliar sclereids in Memecylon and Lijndenia (Melastomataceae) from Borneo, Java, Malaya and Sumatra. – Nord. J. Bot. 3: 343–345. Rodgers, W. A. and Homewood, K. M. 1982. Biological values and conservation prospects for the forests and primate populations of the Uzungwa Mountains, Tanzania. – Biol. Conserv. 24: 285–304. Stone, R. D. 2006. Phylogeny of major lineages in Melastomataceae, subfamily Olisbeoideae: utility of nuclear glyceraldehyde 3-phosphate dehydrogenase (GapC) gene sequences. – Syst. Bot. 31: 107–121. Stone, R. D. 2014. he species-rich, paleotropical genus Memecylon (Melastomataceae): molecular phylogenetics and revised infrageneric classiication of the African species. – Taxon 63: 539–561. Stone, R. D. and Andreasen, K. 2010. he Afro-Madagascan genus Warneckea (Melastomataceae): molecular systematics and revised infrageneric classiication. – Taxon 59: 83–92. Tropicos 2014. Missouri Botanical Garden, Saint Louis, USA. –  http://www.tropicos.org , accessed 29 Jun 2014. Vollesen, K. 1980. Annotated check-list of the vascular plants of the Selous Game Reserve, Tanzania. – Opera Bot. 59: 1–117. Wickens, G. E. 1975. Melastomataceae. – In: Polhill, R. M. (ed.), Flora of Tropical East Africa. Crown Agents for Oversea Governments and Administrations, London. 517