MEP Candollea 68-2_. 09.12.13 10:08 Page301

Notes on the flora of Madagascar, 30-34

Martin W. Callmander, Peter B. Phillipson, Thierry Deroin & Laurent Gautier (ed.)

Abstract Résumé CALLMAnder, M. W., P. B. PhiLLiPSon, T. deroin & L. gAuTier CALLMAnder, M. W., P. B. PhiLLiPSon, T. deroin & L. gAuTier (ed.) (2013). notes on the flora of Madagascar, 30-34. Candollea 68: 301- (ed.) (2013). notes sur la flore de Madagascar, 30-34. Candollea 68: 301-320. 320. in english and French, english and French abstracts. en anglais et français, résumés anglais et français. ongoing research on Madagascar’s flora is revealing numerous Les recherches en cours sur la flore de Madagascar révèlent taxonomic novelties and nomenclatural inconsistencies, and de nombreuses nouveautés taxonomiques, des problèmes de providing new data on species distribution. This is the seventh nomenclature et de nouvelles données sur la distribution des set of notes in a series that aims to provide the botanical com- espèces. Cette publication est la septième d’une série de notes munity working on the flora of Madagascar an opportunity to destinées à donner à la communauté botanique internationale publish short communications on these topics, and this issue travaillant sur Madagascar la possibilité de publier de courtes comprises five notes. contributions traitant de ces aspects et dans ce numéro, com- – Note 30. nomenclatural notes on R. Br. (Lau- prend cinq notes. raceae) from Madagascar, by henk van der Werff. Five species – Note 30. notes nomenclaturales sur Cryptocarya R. Br. (Lau- of Malagasy Sonn. are transferred to Cryptocarya raceae) à Madagascar, par henk van der Werff. Cinq espèces r. Br. following the merging of these two gener a and de Ravensara Sonn. de Madagascar sont transférées à Crypto - the accepted conservation of Cryptocarya over Ravensara. carya r. Br. suite à l’intégration de ces deux genres et la This necessitates two new combinations and two new names. conservation de Cryptocarya sur Ravensara. Ceci nécessite The fifth species is placed in synonymy. deux nouvelles combinaisons et deux nouveaux noms. La cin- – Note 31. Further nomenclatural notes on Malagasy Diospyros quième espèce est placée en synonymie. L. (Ebenaceae): goudot types in the geneva herbarium, by – Note 31. notes nomenclaturales additionnelles sur les Dios- george e. Schatz, Porter P. Lowry ii, Cyrille Mas & Martin pyros L. (Ebenaceae) de Madagascar: les types de goudot dans W. Callmander. discovery of several goudot collections of l’herbier de genève, par george e. Schatz, Porter P. Lowry ii, Malagasy Diospyros L. deposited in the geneva herbarium Cyrille Mas & Martin W. Callmander. La redécouverte de plu- clarifies the typifications of Diospyros leucocalyx hiern and sieurs échantillons, aujourd’hui déposés à l’herbier de genève, Maba madagascariensis A. dC. resolution of the type of de Diospyros L. récoltés à Madagascar par goudot clarifie le Diospyros leucocalyx requires that the long used name Diospyros statut nomenclatural de Diospyros leucocalyx hiern ainsi que megasepala Baker be placed into synonymy under Diospyros de Maba madagascariensis A. dC. Le suivi des règles relatives

Addresses of the editors: MWC: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and Conservatoire et Jardin botaniques de la ville de genève, ch. de l’impératrice 1, CP 60, 1292 Chambésy, Switzerland. e-mail: [email protected] PBP: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and Muséum national d’histoire naturelle, département Systématique et evolution, uMr 7205, oSeB, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. e-mail: [email protected] Td: Muséum national d’histoire naturelle, département Systématique et evolution, uMr 7205, oSeB, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. e-mail: [email protected] Lg: Conservatoire et Jardin botaniques de la ville de genève et université de genève, Laboratoire de botanique systématique et biodiversité, ch. de l’impératrice 1, CP 60, 1292 Chambésy, Switzerland. e-mail: [email protected]

ISSN: 0373-2967 – Online ISSN: 2235-3658 – Candollea 68(2): 301-320 (2013) © CONSERVATOIRE ET JARDIN BOTANIQUES DE GENÈVE 2013 MEP Candollea 68-2_. 09.12.13 10:08 Page302

302 – Candollea 68, 2013

leucocalyx hiern. Past confusion between the two Paris spec- à la procédure de typification nécessite que le nom Diospyros imens Perrier de la Bâthie 700 and Pervillé 700, the former megasepala Baker, longtemps utilisé, soit placé en synonymie of which is the type of Dichapetalum thouarsianum var. pubes- de Diospyros leucocalyx hiern. de même une vieille confusion cens desc. and the latter of which was previously cited as entre les spécimens parisiens Perrier de la Bâthie 700 et Per- the type of Maba madagascariensis, is also discussed and villé 700 est ici levée. Ce premier est le type de Dichapetalum clarified. thouarsianum var. pubescens desc. Alors que le dernier a été – Note 32. hybrid origin of Arthropteris boutoniana (hook.) précédemment cité comme type de Maba madagascarie nsis. Pic. Serm. (Arthropteridaceae) from Mauritius and Madaga - – Note 32. origine hybride de Arthropteris boutoniana (hook.) scar, by France rakotondrainibe & Catherine reeb. The study Pic. Serm. (Arthropteridaceae) de l’île Maurice et de of the Arthropteris hook. f. collections from the Malagasy Madagascar, par France rakotondrainibe & Catherine reeb. region undertaken as a part of the revision of Filicophyta L’examen des récoltes d’Arthropteris hook f. de la région from Madagascar currently in progress, has revealed the malgache, effectué dans le cadre de la révision en cours des intraspecific morphological heterogeneity of Arthtropteris bou- Filicophytes de Madagascar, a permis de mettre en évidence toniana (hook.) Pic. Serm. and demonstrated its hybrid origin. l’hétérogénéité morphologique intraspécifique de Arthropteris The putative parents are Arthropteris monocarpa (Cordem.) boutoniana (hook.) Pic. Serm. et son origine hybride. Les C. Chr. and Arthropteris orientalis (J. F. gmel.) Posth. var. parents présumés sont Arthropteris monocarpa (Cordem.) orientalis. The hybrid Arthropteris ×boutoniana is chara - C. Chr. et Arthropteris orientalis (J. F. gmel.) Posth. var. orien- cterized by its hairy indusia, and its reduced fertility with few talis. L’hybride Arthropteris ×boutoniana se caractérise par or totally aborted sporangia. it has been reported from des indusies hirsutes et une fertilité réduite à nulle qui se traduit Mauritius and Madagascar. par des sporanges peu nombreux ou avortés. Sa présence est – Note 33. Lectotypifications and a new combination in Poly- signalée à l’île Maurice et à Madagascar. gala L. (Polygalaceae) from Madagascar, by gregory A. – Note 33. Lectotypifications et nouvelle combination dans Wahlert, Peter B. Phillipson & george e. Schatz. Polygala Polygala L. (Polygalaceae) à Madagascar, par gregory A. longeracemosa var. retamoides h. Perrier, narrowly distributed Wahlert, Peter B. Phillipson & george e. Schatz. Polygala on soils derived from marble and quartzite in the Ambatofi- longeracemosa var. retamoides h. Perrier, distribuée étroite- nandrahana area, was found to be morphologically distinct ment sur des sols dérivant de marbre et de quartzite dans la from the nominate variety. The authors conclude that it merits zone d’Ambatofinandrahana, est considérée distincte morpho- recognition at the rank of species and provide the necessary logiquement de la variété nominative. Les auteurs concluent new combination; they also designate lectotypes for both taxa, que ce taxon mérite d’être reconnu au rang d’espèce et formu- and provide a key, preliminary conservation assessments and lent donc la nouvelle combinaison nécessaire; ils désignent a distribution map. également des lectotypes pour les deux taxons, et fournissent – Note 34. Adieu Adelosa Blume (Lamiaceae): Further obser- une clé, des évaluations préliminaires de conservation et une vations on Rotheca raf. for Madagascar, by Martin W. Callman- carte de répartition. der, Peter B. Phillipson, James A. Wearn & rogier P. J. de kok. – Note 34. Adieu Adelosa Blume (Lamiaceae): observations Adelosa microphylla Blume is equated with the widespread and additionnelles sur Rotheca raf. pour Madagascar, par Martin variable Rotheca incisa (klotzsch) Steane & Mabb., and the W. Callmander, Peter B. Phillipson, James A. Wearn & rogier status of the poorly-known Clerodendrum mirabile Baker and P. J. de kok. Adelosa microphylla Blume est incluse dans Rotheca myricoides (hochst.) Steane & Mabb. are examined. l’espèce largement répandue et morphologiquement variable Two new combinations in the Rotheca raf. are proposed. Rotheca incisa (klotzsch) Steane & Mabb. Le statut des espèces imparfaitement connues Clerodendrum mirabile Baker et Rotheca myricoides (hochst.) Steane & Mabb. est examiné. Key-words deux nouvelles combinaisons dans le genre Rotheca raf. sont ARTHROPTERIDACEAE – EBENACEAE – LAMIACEAE – proposées. – POLYGALACEAE – Adelosa – Arthropteris – Clerodendrum – Cryptocarya – Diospyros – Maba – Polygala – Ravensara – Rotheca – Madagascar – Mauritius – – hybrid – iuCn red List MEP Candollea 68-2_. 09.12.13 10:08 Page303

30. van der WERFF, Henk: Nomenclatural notes on Cryptocarya R. Br. (Lauraceae) from Madagascar

Introduction Nomenclature

koSTerMAnS (1950) recognized both Cryptocarya r. Br. Cryptocarya oblonga (kosterm.) van der Werff, comb. nova. and Ravensara Sonn. in his treatment of the Lauraceae for the ϵ Ravensara oblonga kosterm. in Bull. Jard. Bot. etat Flore de Madagascar et des Comores. he accepted 18 species Bruxelles 28: 184. 1958. in Ravensara, a genus endemic to Madagascar, and eight Typus: MADAGASCAR. Prov. Toliara: Mandena, Fort - species in Cryptocarya, a large, pantropical genus. This treat- dauphin, 1.v.1955, st., Service Forestier 13163 (holo-: P ment was based on his studies in the late 1930’s (koSTerMAnS, [P00853162]!; iso-: Bo, P [P00540966], TeF [TeF000298]). 1939). Subsequently kostermans published 15 additional species of Cryptocarya (koSTerMAnS, 1957) and nine addi- Observations. – The type (the only collection known when tional species of Ravensara (koSTerMAnS, 1958). The sole the species was described) is sterile. Four recent collections difference between Cryptocarya and Ravensara is found in the have broadened our knowledge of this species. Dumetz 696 fruits: fruits of Ravensara are ruminate, whereas Cryptocarya (st.) is also from Mandena, at an altitude of 10 m. Rajoharison fruits are not. Because ruminate fruits have also been reported & al. 222 is from Antsotso, and occurs in “forêt sur sable”. its from Cryptocarya species outside Madagascar, vAn der WerFF fruits are round, smooth, ca 1.3 ϫ 1.3 cm, with the tepals per- (1992) proposed to merge the two genera and conserve Cryp- sisting on top of the fruits, and its infructescences are short, tocarya over Ravensara. This proposal was accepted. The tax- up to 3 cm long. Ramison & Rabehevitra 557 is from Sainte onomy of Cryptocarya (including Ravensara) on Madagascar Luce in “forêt littorale”. The inflorescences are short, ca 2 cm is poorly known. koSTerMAnS (1957, 1958) did not include a long, clustered distally on the twigs and resemble a terminal key to species and described several based on few collections, inflorescence. The young growth is densely ferruginous pubes- with some known only in fruit, others only in flower or even cent although the indument wears off quickly. Leaves of this based on sterile collections. it seems prudent to transfer the species are to 10 cm long, with an obtuse to rounded base and Ravensara species to Cryptocarya only once they are better a rounded or emarginated apex. Lateral veins are immersed on known because it is possible (and likely) that some species the upper leaf surface and immersed or weakly raised on the may have been described twice, once in Cryptocarya and once lower surface. The fruits are ruminate. in Ravensara. vAn der WerFF (2008) started this process and Cryptocarya oblonga, a small to 10 m tall, is restricted transferred eight Ravensara species to Cryptocarya. in this to a small area of coastal forest on sand n of Tolagnaro (Fort contribution an additional five species are transferred which dauphin). necessitates two new combinations and two new names. due Recent collections seen. – MADAGASCAR. Prov. Toliara: Man- to a new synonymy, a third new name is not necessary for the dena, 24º57’S 47º00’e, 0-10 m, 17.iv.1989, st., Dumetz 696 (Mo); fifth species. Ampasy, 24º57’25’’S 47º00’04’’e, 0-10 m, 30.ix.2000, st., Rabenan- toandro & al. 296 (Mo); iabakoho, Antsotso, 24º35’S 47º12’e, 27.v.2007, fr., Rajoharison & al. 222 (Mo); Mahatalaky, Sainte Luce, 24º47’S 47º11’e, 2.ii.2008, fl., Ramison & Rabehevitra 557 (Mo).

Address of the author: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. e-mail: [email protected]

ISSN: 0373-2967 – Online ISSN: 2235-3658 – Candollea 68(2): 303-306 (2013) © CONSERVATOIRE ET JARDIN BOTANIQUES DE GENÈVE 2013 MEP Candollea 68-2_. 09.12.13 10:08 Page304

304 – Candollea 68, 2013

Cryptocarya polyneura (kosterm.) van der Werff, comb. nova. Cryptocarya ambrensis van der Werff, nom. nov. ϵ Ravensara polyneura kosterm. in Bull. Jard. Bot. etat ϵ Ravensara areolata kosterm. in not. Syst. (Paris) 8: Bruxelles 28: 188. 1958. 107. 1939 [non Cryptocarya areolata gamble in Bull. Typus: MADAGASCAR. Prov. Toamasina: Manakam bahiny Misc. inform. kew 1910: 144. 1910]. est, Ambatodrazaka, zahamena, 28.ii.1950, fl., Réserves Typus: MADAGASCAR. Prov. Antsiranana: roussettes, Naturelles 1936 (holo-: P [P00540957]; iso-: P [P00540955, diégo Suarez, s.d., fl., Ursch 187 (holo-: P [P00540982]; P00540956, P00853160]!, TeF [TeF000291]). iso-: Bo, P [P00853166]!). Observations. – The type of Ravensara polyneura is only ϭ Ravensara acutifolia kosterm. in Bull. Jard. Bot. etat in bud. kostermans cited four additional collections, with flow- Bruxelles 28: 175. 1958 [non Cryptocarya acutifolia ers or fruits, but did not include those in his description. This h. W. Li in Acta Phytotax. Sin. 17: 69. 1979]. Typus: species has long and narrow, coriaceous leaves, at least three MADAGASCAR. Prov. Antsiranana: J.B. 19, roussettes, times as long as wide, dark green when dry and with an obtuse diégo Suarez, 11.xi.1954, fl., Service Forestier 11014 to rounded base. Mature leaves are glabrous and the midrib (holo-: P [P00853165]!; iso-: P [P00540987]). and lateral veins are raised on the lower surface. The inflores- Observations. – Cryptocarya ambrensis is an inconspi - cences are short, to 3 cm, densely dark brown pubescent with cuous species with small (to 7 ϫ 3 cm), chartaceous, elliptic, nearly sessile flowers arranged in dense clusters. The young glabrous leaves. Terminal buds and inflorescences are minutely growth has a similar dark brown pubescence as the flowers brown puberulous; twigs are glabrous except for the very tips. and inflorescence, but the twigs become soon glabrous. The inflorescences are laxly flowered and up to 5 cm long; fruits fruits are ruminate. are ruminate. numerous small gland dots are visible on the recent collections all come from the area between lower leaf surface, but there is some variation in leaf shape Moramanga and the zahamena reserve at altitudes between 900 from distinctly acute apices to blunt apices. and 1160 m. i have not seen the two paratypes of Ravensara acutifolia This species has been confused with Cryptocarya pervillei cited by koSTerMAnS (1958) which were collected in “Tia- Baill., another species with long and narrow leaves. it differs narantsoa”, believed to be an error for Fianarantsoa, in south- in having a light brown to golden brown indument (not dark central Madagascar. The type and all recent collections are from brown), less coriaceous leaves with a more persistent indument. the Montagne d’Ambre in the far north of Madagascar, between i had previously identified Rakotomalaza 1334 and Andriat- 840 and 1240 m elevation. The collection Gautier & Chatelain siferana & al. 2109 as C. myristicoides Baker, but that species 4973, collected near daraina 85 km Se of Montagne d’Ambre has stiffly chartaceous (not coriaceous) leaves with immersed resembles Cryptocarya ambrensis vegetatively; however, it lacks veins, and longer, less condensed inflorescences. The collection the gland dots on the lower leaf surface and the fruit is not rumi- Antilahimena 5915 was previously identified as C. acuminata nate; it represents an unknown Cryptocarya species. Merr. (an error for Ravensara acuminata (Meisn.) Baill. Recent collections seen. – MADAGASCAR. Prov. Antsi ranana: ϭ Cryptocarya litoralis van der Werff), a coastal species also Montagne d’Ambre, 1.x.2005, buds, Acevedo-Rodriguez 14514 with less coriaceous leaves, brown pubescence (not dark (Mo); Montagne d’Ambre Pn, Lac Maudit, 12º34’S 49º09’e, 1100- brown) and often conduplicate (rather than flat) leaves. dupli- 1250 m, 3-10.vi.1993, fr., Andrianantoanina & Bezana 143 (Mo); cates may have been distributed under these identifications. ibid. loc., Andrianantoanina & Bezana 166 (Mo); Montagne d’Ambre, près d’Antsalaka, 12º27’S 49º13’e, 250-500 m, 19- Recent collections seen. – MADAGASCAR. Prov. Toamasina: Phelps dodge project site, 15 km ne of Moramanga, 18º51’02’’S 48º18’24’’e, 26.vii.1993, fr., Andrianantoanina & al. 229 (Mo); Parc national 950 m, 13.ii.1997, fl., Andriatsiferana & al. 2109 (Mo); Moramanga, de Montagne d’Ambre au grand Lac, 12º35’S 49º09’e, 1475 m, 9- Andasibe, Ambatovy forest, 18º48’29’’S 48º18’50’’e, 1060 m, fr., Anti- 17.ix.1993, fr., Andrianantoanina 327 (Mo); Montagne d’Ambre, à lahimena & al. 3004 (Mo); Moramanga, Andasibe, Menalamba, la Station des roussettes vers la piste d’Ankorefo, 12º31’30”S 49º10’ 18º51’06’’S 48º18’39’’e, 1119 m, 18.xii.2005, fr., Antilahimena & 17”e, 800-1000 m, 26.ix.1995, Andrianantoanina & Bezara 872 Edmond 4457 (Mo); Moramanga, Ambohibary, Ambatovy/Andranovery (Mo); Montagne d’Ambre national Park, at Cul de Sac, 12º31’36”S forest, 18º52’04’’S 48º18’15’’e, 1031 m, 24.x.2007, fr., Antilahimena 49º10’20”e, 840 m, fr., Harder & al. 1642 (Mo); Montagne d’Ambre, & Marcellin 5915 (Mo); Moramanga, Andasibe: Forêt d’Analamay, partie centrale, 12º34’S 49º10’e, 1249 m, 11.xi.2007, buds, Ranirison 18º50’51’’S 48º19’29’’e, 990 m, 8.vi.2007, fr., Bernard 559 (Mo); & al. 1167 (Mo); Montagne d’Ambre, partie nord, 12º28’S 49º10’e, route Analamay, 18º49’47’’S 48º18’48’’e, 1160 m, 16.v.1997, y. fr., 1015 m, 30.v.2008, fr., Trigui & al. 460 (Mo). Rakotomalaza & al. 1334 (Mo). réserve forestière Sandrangato, Mora- Three other Cryptocarya species have been reported from manga, 30.x.1964, fr., Service Forestier 21904 (Mo); Périnet, Anala- the north of Madagascar, including two others from Montagne mazaotra, Moramanga, 28.ix.1966, fr., Service Forestier 26105 (Mo). d’Ambre. Cryptocarya septentrionalis van der Werff (type Ursch 96, prope diego Suarez) differs from C. ambrensis by its lanceolate leaves; the eight recent collections are from lowlands up to 320 m. MEP Candollea 68-2_. 09.12.13 10:08 Page305

Uvaria lombardii L. Gaut. & Deroin (Annonaceae), une nouvelleNotes espèceon the floraendémique of Madagascar, de Madagascar 30-34 – 305

Cryptocarya ocoteifolia kosterm. is based on a type col- to determine if Ravensara coriacea is a synonym of Crypto- lected in the forêt d’Ambre (Service Forestier 7208) which carya rigidi folia or C. retusa or a distinct species. Both consists of a leafy twig and a detached fruit. koSTerMAnS C. rigidi folia and C. retusa have priority over Ravensara (1957) also cited four sterile paratypes. however, koSTerMAnS coriacea. (1962) noted that Capuron had pointed out that the type col- Recent collections seen. – MADAGASCAR. Prov. Antananarivo: lection was a mixture, consisting of Cryptocarya fruits and Antananarivo, Ankozobe, Ankafobe, 18º07’14”S 47º11’30”e, 1442 m, leafy Ocotea twigs. Because neither detached Cryptocarya 24.ii.2005, fr., Lehavana & al. 286 (Mo); 7 km e of Anjozorobe, fruits nor sterile Ocotea twigs can be identified with any con- 18º22’S 48º00’e, 1300 m, 12.v.1987, fr., Schatz & al. 1388 (Mo). Prov. fidence, i consider Cryptocarya ocoteifolia an incompletely Toamasina: Phelps dodge project site, 15 km ne of Moramanga, known species. 18º51’02”S 48º18’24”e, 950 m, 13.ii.1997, st., Andriatsife rana & al. Cryptocarya rotundifolia kosterm. is only known from the 2088 (Mo), Alaotra-Mangoro region, Ambohibary, Ambatovy, type collected in the Forêt d’Ambre near Antsiranana (Service 18º50’48”S 48º17’51”, 1050 m, 3.v.2007, fr., Antilahi mena & al. 5516 Forestier 7192). The specimen consists of leafy twigs and (Mo); Toamasina, 2 to 4 km e of Perinet, 1000 m, 24.iv.1974, fr., detached fruits; the leaves are large, broad (to 14 ϫ 8 cm) and Gentry 11251A (Mo); Antetezampandrana, pk 27-28, route Moramanga- chartaceous. The fruits are old and the fleshy outer layer has Anosibe, s.d., fl., Service Forestier 26842 (Mo). disappeared. The fruits are not ruminate. The large, rather thin leaves are unlike any other collection of Cryptocarya from the Cryptocarya spathulata kosterm. in Bull. Jard. Bot. etat Brux- north of Madagascar. it is possible that the type is also a mixed elles 27: 185. 1957. collection. More and better collections are needed to establish Typus: MADAGASCAR. Prov. Toamasina: Analamazaotra, C. rotundifolia as a good species. 18.i.1950, fl., Service Forestier 1476 (holo-: P [P0085 3164]!). Cryptocarya rigidifolia van der Werff, nom. nov. ϭ Ravensara laevis kosterm. in not. Syst. (Paris) 8: 100. 1939 ϵ Ravensara elliptica kosterm. in not. Syst. (Paris) 8: [non Cryptocarya laevis Mart. in Flora 21: Beibl. 64. 1838]. 110. 1939 [non Cryptocarya elliptica Schltr. in Bot. Typus: MADAGASCAR. Prov. Toamasina: Analamazaotra, Jahrb. Syst. 39. 108. 1906]. 20.vi.1919, buds, Thouvenot 98 (holo-: P [P00540973]; Typus: MADAGASCAR. Prov. Toamasina: Forêts montag- iso-: P [P00853163]!, P [P00540974]), syn. nov. neuses de l’est, fl., Louvel 229 (holo-: P [P00853161]!). Observations. – Although C. spathulata is a later heterotypic Observations. – Cryptocarya rigidifolia is very similar to synonym of Ravensara laevis, it is the earliest available name C. retusa (nees) van der Werff, but the type (Du Petit Thouars in Cryptocarya and should be used for this species. The types s.n.: holo-: B; iso-: P [P00540954]) of the latter lacks flowers, of both names came from the same locality and both were in and the slight vegetative differences have made it difficult to flower. Because the sole difference between Cryptocarya and decide whether or not they represent separate species, the type Ravensara is a fruit character, this example illustrates the dif- also lacks any indication of the collection locality. examination ficulty of assigning flowering specimens to either Cryptocarya of recent flowering collections that could be referred to one or or Ravensara, when these genera were considered distinct. other of these species has shown that lowland (5-20 m altitude) specimens have dorsally pubescent anthers and weakly raised Cryptocarya occidentalis van der Werff, nom. nov. secondary veins, while collections fr om higher altitude (950- ϵ Ravensara perrieri dubard & dop in Bull. Soc. Bot. 1440 m) have glabrous anthers (although the filaments are France 54: 156. 1907 [non Cryptocarya perrieri pubescent) and more strongly raised secondary veins. Leaves danguy in Bull. Mus. hist. nat. (Paris) 33: 523. 1927. of the type of C. retusa agree with the leaves of the lowland ≡ Aspidostemon perrieri (danguy) rohwer]. specimens, while the type of C. rigidifolia agrees with the spec- Typus: MADAGASCAR. Prov. Mahajunga: Ambongo, imens from higher altitude in leaf and flower characters. There- rives de la Mahavavy, viii.1904, fl., Perrier de la Bâthie fore i accept C. retusa as a lowland species characterized 1789 (holo-: P [P00540958] image seen; iso-: P [P00853 by its pubescent anthers and weakly raised lateral veins and 160]!, P [P00540959]). C. rigidifolia as an interior species from higher elevations with glabrous anthers and more pronounced raised lateral veins. Observations. – This species is widespread in the western Fruits of C. rigidifolia are ruminate. parts of Madagascar and can be recognized by its glabrous, Ravensara coriacea kosterm. is also very similar to Cryp- chartaceous, lanceolate to oblong-lanceolate leaves with incon- tocarya rigidifolia and C. retusa. The type specimen (Service spicuous venation (secondary veins poorly visible). The inflo- Forestier 12356) is from Menagisy-Brickaville and has young rescences and flowers are minutely brown puberulous. The inflorescences. As long as flowers are unknown, it is difficult fruits are ruminate. MEP Candollea 68-2_. 09.12.13 10:08 Page306

306 – Candollea 68, 2013

duBArd & doP (1907) recognized the difficulty of assign- References ing this species to either Cryptocarya or Ravensara because the type was in flower. They went ahead with the description duBArd, M. & P. doP (1907). description de quelques espèces nouvelles de Madagascar. Bull. Soc. Bot. France 54: 155-160. because “Dans tous les cas, l’espèce est nouvelle, quel que soit le genre auquel elle appartient”. koSTerMAnS, A. J. g. h. (1939). enumeratio Lauracearum Mada- gascariensium et ex insulis Mascarenis. (revisio Lauracearum dubard & dop cited as type collection Perrier de la Bâthie vi). Notul. Syst. (Paris) 8: 67-128. 1 and as type locality “rivière Mahoudedy supérieure”, while koSTerMAnS, A. J. g. h. (1950). Lauracées. In: huMBerT, h. (ed.), koSTerMAnS (1939) cited Perrier de la Bâthie 1789 and gave Fl. Madagascar Comores 81. as type locality “Mahomavy”. The scan of the holotype shows koSTerMAnS, A. J. g. h. (1957). Le genre Cryptocarya r. Br. that the collection number 1789 has been scratched out and (Lauracées) à Madagascar. Bull. Jard. Bot. Etat Bruxelles 27: 173- replaced with 1. The handwritten type locality seems to be 188. “Mahavavy”, not “Mahoudedy” or “Mahomavy”, but the writ- koSTerMAnS, A. J. g. h. (1958). Le genre ravensara Sonn. (Laura- ing is not entirely clear. cées) à Madagascar. Bull. Jard. Bot. Etat Bruxelles 28: 173-191. Recent collections seen. – MADAGASCAR. Prov. Antsiranana: Forêt d’Ampasindava, 7.vii.1956, fl., Service Forestier 16517 koSTerMAnS, A. J. g. h. (1962). Miscellaneous botanical notes 4. (Mo); Beravina, 18.viii.1955, fl., Service Forestier 15141 (Mo); Reinwardtia 6: 281-325. Beravina, 19.viii.1956, fl., Service Forestier 16355 (Mo). vAn der WerFF, h. (1992). Proposal to conserve 2813 Cryptocarya Prov. Mahajanga: Melaky, Mainti rano, au n de Belitsaky, le long against ravensara (Lauraceae). Taxon 41: 129-130. de la rivière de Manomba, 17º52’23’’S 44º28’52’’e, 140 m, vAn der WerFF, h. (2008). A new species and new combinations in 25.x.2009, fr., Andriamihajarivo & al. 1802 (Mo). Prov. Toliara: Beroroha à 4 km avant Antsoa, 21º15’43’’S 45º10’04’’e, 461 m, Cryptocarya from Madagascar. Adansonia ser. 3, 30: 41-46. 3.xii.2010, fr., Andriantiana & al. 1004 (Mo); Atsimo-Andrefana region, Makay Massif, along the Menampandaha river, 21º11’57’’S 45º20’15’’e, 510 m, 23.xi.2010, fr., Phillipson & al. 6201 (Mo); Beroroha, Betorabato, 21º34’04’’S 45º34’34’’e, 296 m, 13.i.2011, fr., Razakamalala 6001 (Mo). MEP Candollea 68-2_. 09.12.13 10:08 Page307

31. SCHATZ, George E., Porter P. LOWRY II, Cyrille MAS & Martin W. CALLMANDER: Further nomenclatural notes on Malagasy Diospyros L. (Ebenaceae): Goudot types in the Geneva herbarium.

Introduction Systematics in the course of evaluating Ebenaceae for the Catalogue 1. Diospyros leucocalyx hiern in Trans. Cambridge Philos. of the vascular of Madagascar (MAdAgASCAr CATA- Soc. 12: 267. 1873. Logue, 2013), and subsequently clarifying the nomenclature Typus: MADAGASCAR: Ambanivoules, 1833, Goudot s.n. of more than half of the currently recognized species (SChATz (holo-: g [g00368907]!). & LoWry, 2011), the identity of several names remained uncer- ϭ Diospyros megasepala Baker in J. Linn Soc., Bot. 21: tain because their types had not yet been located, and/or there 423. 1885. Typus: MADAGASCAR: Central Madagascar, was confusion about their typification. recent examination of s.d., Baron 2365 (holo-: k [k000350806]!; iso-: P several specimens collected in Madagascar by Jules Prosper [P00541783]!). goudot in the herbarium of the Conservatoire et Jardin botaniques de la ville de genève confirms that they include Observations. – Diospyros leucocalyx was described in material that should be considered the types of two names: hiern’s Monograph of Ebenaceae (1873) with only a single Diospyros leucocalyx hiern and Maba madagascariensis A. collection cited as “Madagascar, Ambanivoule, Goudot! A. d. dC. Jules P. goudot, not to be confused with his brother Justin, 1833”. Later, Perrier de LA BâThie (1952a, 1952b) placed who collected plants in South America, was born in the Jura D. leucocalyx into synonymy under D. gracilipes hiern, but region of France, and died at some unknown date in Madagas- failed to cite the goudot specimen in either publication. A spe- car, having probably returned there in 1861 after the death of cimen in the geneva herbarium with the label “Madagascar. queen ranavalona i, who had expelled him along with other M. goudot 1833.” [printed], and “petit arbrisseau. Ambani- French in 1858 (dorr, 1997). having married a Merina and voules (frts.) voir le fruit.” [handwritten] must certainly be the become fluent in Malagasy, goudot’s Malagasy nickname, type of D. leucocalyx hiern. Based on the extremely large Bibikely or insect, reflected his equally consuming passion for female calyx and large leaves, Laurent gautier (g) annotated collecting insects in addition to plants. in this note we provide the specimen in 1995 as the long recognized but later published information clarifying both the identity and typification of D. megasepala Baker, which must now be placed into syno- Diospyros leucocalyx and Maba madagascariensis. nymy under D. leucocalyx. With regard to the location of goudot’s collection, “Amba- nivoules” likely refers to an ethnic group of Malagasy, and cor- responds to “Antanbanivolo”, or “people living at the base of the mountains covered with bamboos”. in his compendium on the delessert herbarium and library, LASégue (1845: 188) includes an entry on goudot in which he discusses where gou- dot collected in Madagascar: “…chez les Ambanivoules, dans la chaîne des hautes montagnes qui s’étendent du nord au sud

Addresses of the authors: geS: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. e-mail: [email protected] PPL: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and département Systématique et evolution (uMr 7205), Muséum national d’histoire naturelle, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. CM: département Systématique et evolution (uMr 7205), Muséum national d’histoire naturelle, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. MWC: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and Conservatoire et Jardin botaniques de la ville de genève, ch. de l’impératrice 1, CP 60, 1292 Chambésy, genève, Switzerland.

ISSN: 0373-2967 – Online ISSN: 2235-3658 – Candollea 68(2): 307-309 (2013) © CONSERVATOIRE ET JARDIN BOTANIQUES DE GENÈVE 2013 MEP Candollea 68-2_. 09.12.13 10:08 Page308

308 – Candollea 68, 2013

de l’île, à 20 ou 25 lieues [approximately 80 to 100 km] de (the type of D. bernieri hiern from Tintingue, which we have Tamatave. il a herborisé, en 1832, dans les environs de la petite placed in synonymy under D. squamosa A. dC.), suggesting rivière de Manaarez, qui prend sa source dans les montagnes that hiern may have made an error when preparing or tran- des Ambanivoules, ainsi que sur les bords et dans les épaisses scribing his notes. The third collection hiern cited, Pervillé forêts qui suivent le cours d’yvondrou [ivondro], rivière qui 700, which was collected on nosy Be, i.e., not near Tamatave, sépare le territoire des Bétanimènes de celui des Ambani- is discussed further below. voules”. Some eighty years later, Perrier de LA BâThie, both in the precursor to (1952a) and treatment of Ebenaceae in the Flore 2. Diospyros ferrea (Willd.) Bakh. in gard. Bull. Straights de Madagascar et des Comores (1952b), accepted Maba mada- Settlem. ser. 3, 7: 162. 1933. gascariensis A. dC., stating that it was “toujours représentée dans l’herbier du Muséum de Paris que par le type (Pervillé ϵ Ehretia ferrea Willd., Phytographia 1: 4. 1794. 700), qui a été récolté sur l’île de nosy-be” (Perrier de LA Typus: INDIA: sine loc., s.d., König s.n. (holo-: B†; iso-: BâThie, 1952a: 104). Perrier de la Bâthie also accepted C [C10010825, C10010826, C10010827] images seen). M. buxifolia as a species of the eastern littoral forest, including ϭ Pisonia buxifolia rottb. in nye Saml. kongel. danske Bernier 112 from “ile Sainte Marie” among the material cited, vidensk. Selsk. Skr. 2: 536. 1783. ϵ Ferreola buxifolia apparently repeating hiern’s error. given its true provenance (rottb.) roxb., Pl. Coromandel 1: 35. 1795. ϵ Maba as cited in the protologue of near Tamatave, it is unclear why buxifolia (rottb.) Pers., Syn. Pl. 2: 606. 1807. ϵ Ebenus Perrier would have considered Pervillé 700 from nosy Be to buxifolia (rottb.) kuntze, revis. gen. Pl. 2: 408. 1891. be the type of M. madagascariensis. rather, a specimen at ϵ Diospyros ferrea var. buxifolia (rottb.) Bakh. in Bull. geneva, Goudot s.n., collected in 1831 at Tamatave, is certainly Jard. Bot. Buitenzorg ser. 3, 15: 57. 1937 [non Diospy- the type of this name. in addition to the collection label, it bears ros buxifolia hiern in Trans. Cambridge Philos. Soc. an original label identifying it as “Maba madagascariensis A. 12: 218. 1873]. Typus: INDIA: sine loc., s.d., König s.n. dC.” it also bears an annotation label affixed by Frank White (holo-: C [C10010827]; iso-: C [C10010825, C100108 in 1972 correctly identifying it as Diospyros ferrea (Willd.) 26] images seen). Bakh., a name based on Ehretia ferrea Willd., which provides the oldest available epithet for the taxon first described as Piso- ϭ Maba madagascariensis A. dC., Prodr. 8: 241. 1844. nia buxifolia rottb., but which cannot be transferred to Diospy- ϵ Diospyros ferrea var. madagascariensis (A. dC.) ros because the epithet “buxifolia” is already occupied in Bakh. in Bull. Jard. Bot. Buitenzorg ser. 3, 15: 432. Diospyros as D. buxifolia hiern (hiern, 1873). Pisonia buxi- 1941. Lectotypus (designated here): MADAGASCAR: folia and Ehretia ferrea are both based on specimens of J. G. env. de Tamatave, 15.ii.1831, Goudot s.n. (g [g0036 König (s.n.) collected in india, originally deposited in Copen- 8886]!; isolecto-: g [g00368906]!, g-dC [g001 hagen a nd Berlin respectively. We have been unable to locate 2323]!). any material that could be the König s.n. type of E. ferrea, hav- Observations. – Maba madagascariensis A. dC. was ing enquired at both B, where Willdenow was based, and BM, described by Alphonse de CAndoLLe (1844) without any indi- which has significant holdings of könig specimens. in Copen- cation of a collection other than “v. in h. deless.”, and a hagen, there is a könig collection [C10010827] that bears a location in Madagascar as “prope Tamatave”. At the time, Ben- handwritten label “irumbilli Tamulorum idest Lignum duritie jamin delessert’s herbarium was still at his home in Paris, fer. dioeca pentandr. baccifera Arbor interdum frutex where both his close friend A. P. de Candolle and de Candolle’s in sylvis kön.”, the “irumbilli Tamulorum”, which refers to son Alphonse were frequent visitors (STAFLeu, 1970). delessert the vernacular name (“irumbilli”) of the Tamil people and died in 1847, and his herbarium was eventually given to the derives from the word for iron (“irumpu”) in the Tamil lan- city of geneva in 1869 by the daughters of delessert’s brother guage, corresponding closely to “irrumbili Thamulorum” in François. in preparation for his monograph of Ebenaceae, W. the rottbøl protologue of Pisonia buxifolia. interestingly, the P. hiern visited various herbaria on the continent, including Willdenow protologue of Ehretia ferrea includes the citation the delessert herbarium “at geneva” (hiern, 1873: 28). in his “Frambelli Tamulis, i.e. lignum ferreum. Frambo significat monograph, hiern (1873) placed M. madagascariensis in syn- ferrum.” Therefore, it seems likely that rottbøl and Willdenow onymy under M. buxifolia (rottb.) Pers., citing three were studying duplicates of the same könig gathering, with collections from Madagascar: Gerard 28, Bernier 112, and Willdenow interpreting the “i” in “irumbilli” as an “F”. The Pervillé 700. regarding the first two of these collections, we “i” in “irumbilli” has also been interpreted as a “T”, both by have bee n unable to relocate Gerard 28, and Bernier 112 is a Copenhagen in their transcription of the label on [C10010827] Convolvulaceae, Evolvulus alsinoides (L.) L. from Ling vatou and by roxburgh; BAkhuizen vAn den Brink (1936-1941) [Lanivato], although Bernier 113 is a collection of Diospyros attributes the vernacular name “Trumbilli” to roxburgh, and MEP Candollea 68-2_. 09.12.13 10:08 Page309

Uvaria lombardii L. Gaut. & Deroin (Annonaceae), une nouvelleNotes espèceon the floraendémique of Madagascar, de Madagascar 30-34 – 309

also cites “iramballi” as another Tamil vernacular name. The Acknowledgments Goudot s.n. geneva type of Maba madagascariensis appears to be conspecific with the König s.n. type of Pisonia buxifolia We thank the curators at g and P for providing access to at Copenhagen, and thus Maba madagascariensis is confirmed their collections, and ranee Prakash and nick Turland for sear- as a synonym of Diospyros ferrea. Whether D. ferrea is inter- ching for the könig specimens in London and Berlin respec- preted to have an extremely broad distribution from tively. West Africa to Polynesia and hawaii, as circumscribed by BAkhuizen vAn den Brink (1936-1941), is highly ques- References tionable. in our opinion, with the exception of material from BAkhuizen vAn den Brink, r. C. (1936-1941). revisio ebenacearum coastal east Africa, collections from continental Africa do not Malayensium. Bull. Jard. Bot. Buitenzorg ser. 3, 15: 1-515. conform particularly well to the König s.n. type of Pisonia buxifolia and thus probably do not belong in D. ferrea. While BreTeLer, F. J. (1981). The African dichapetalaceae, vii. A taxo - nomic revision species m-q. Meded. Landbouwhogeschool 81(10): coastal collections from Southeast Asia and Polynesia may 1-94. represent D. ferrea, many of the infraspecific taxa described CAndoLLe, A. P. de (1844). ebenaceae. Prodromus Systematis Natu- by BAkhuizen vAn den Brink (1936-1941) may well deserve ralis Regni Vegetabilis 8: 209-243. recognition at the species level pending a comprehensive revi- sion of this complex throughout its range. deSCoingS, B. M. (1960). revision des dichapetalum de Madagascar. Mém. Inst. Sci. Madagascar, Sér. B, Biol. Vég. 9: 63-120. The enigma of Pervillé 700 [P00573496], cited initially by deSCoingS, B. M. (1961). dichapétalacées. In: huMBerT, h. (ed.), hiern as representing Maba buxifolia, the taxon under which Fl. Madagascar Comores 110. he placed M. madagascariensis as a synonym, and then cited by Perrier de la Bâthie as the type of M. madagascariensis, is deSCoingS, B. M. (1962). note complémentaire sur les dichapetalum Malgaches. Nat. Malg. 13: 47-51. further confounded by descoings’s misinterpretation of hand- writing on the label of a Perrier de la Bâthie collection. in his dorr, L. J. (1997). collectors in Madagascar and the Comoro Islands. royal Botanic gardens, kew. precursor to the treatment of Dichapetalaceae for the Flore de Madagascar et des Comores (deSCoingS, 1961), deSCoingS hiern, W. P. (1873). A monograph of ebenaceae. Trans. Cambridge (1960) described Dichapetalum thouarsianum var. pubescens Philos. Soc. 12: 27-300. desc., citing the type as “Pervillé 700”, but he then failed to LASégue, A. (1845). Musée Botanique de M. Benjamin Delessert. ie cite this specimen the following year in his Flore de Mada- Librairie de Fortin, Masson & C , Paris. gascar treatment. in fact, the specimen in question is certainly MAdAgASCAr CATALogue (2013). Catalogue of the Vascular Plants Perrier de la Bâthie 700 [P00298479], collected in 1898 at of Madagascar. Missouri Botanical garden, St. Louis & Antana- Tampolo and bearing a handwritten label with an abbreviated narivo [http://www.efloras.org/madagascar]. signature by Perrier that appears to read “M. PerrdelB”, which Perrier de LA BâThie, h. (1952a). révision des ebénacées de Mada- was misinterpreted by descoings as “Pervillé”, despite the gascar et des Comores. Mém. Inst. Sci. Madagascar, Sér. B, Biol. fact that Pervillé had died by 1868. in a supplemental note the Vég. 4: 93-154. following year, deSCoingS (1962) also failed to cite Pervillé Perrier de LA BâThie, h. (1952b). ebénacées. In: huMBerT, h. (ed.), 700, but did cite Perrier de la Bâthie 700, albeit not under Fl. Madagascar Comores 165. D. thouarsianum var. pubescens, but rather under the species SChATz, g. e. & P. P. LoWry ii (2011). nomenclatural notes on Mala- name D. thouarsianum with the indication “sans localité”. in gasy diospyros L. (ebenaceae). Adansonia ser. 3, 33: 271-281. placing D. thouarsianum var. pubescens in synonymy under STAFLeu, F. A. (1970). Benjamin delessert and Antoine Laségue. D. madagascariense Poir. var. madagascariense, BreTeLer Taxon 19: 920-936. (1981) perpetuated the incorrect citation of Perrier de la Bâthie 700 as Pervillé 700. in reality, Pervillé 700 from nosy Be is neither a type of Maba madagascariensis nor of Dichapetalum thouarsianum var. pubescens, nor is it a collection of Diospyros ferrea, which is restricted in Madagascar to the east coast. rather, it is indeed an as yet undescribed species of Diospyros with affinity to D. bernieriana (Baill.) h. Perrier. MEP Candollea 68-2_. 09.12.13 10:08 Page310

32. RAKOTONDRAINIBE, France & Catherine REEB Origine hybride de Arthropteris boutoniana (Hook.) Pic. Serm. (Arthropteridaceae) de l’île Maurice et de Madagascar

Introduction pinnatifide (ϭ bipinnatifide), des nervures libres terminées par des hydatodes et des sores indusiés. Le Tableau 1 résume les Le genre Arthropteris hook f. classé récemment dans la caractères morphologiques discriminants observés sur les spé- nouvelle famille des Arthropteridaceae (Liu & al., 2013) com- cimens de P, k et BM de la région malgache se rattachant à prend cinq espèces et une variété dans la région malgache: ce complexe d’espèces. Chez A. monocarpa, l’articulation du A. boutoniana (hook.) Pic. Serm., A. monocarpa (Cordem.) pétiole est située très près du rhizome; la face supérieure du C. Chr., A. orientalis (J. F. gmel.) Posth. var. orientalis, limbe et des nervures est hirsute; les hydatodes sont nus, sans A. orientalis var. subbiaurita (hook.) Bonap., A. palisotii dépôt calcaire; chaque segment du limbe ne porte qu’un seul (desv.) Alston et A. parallela (Baker) C. Chr. (TArdieu-BLoT, sore et les indusies sont glanduleuses et dépourvues de poils. 1958; BAdré, 2008; roux, 2009; rAkoTondrAiniBe, donnée A l’inverse, chez A. orientalis, l’articulation du pétiole est non publ.). Suite au réaménagement de l’herbier du Muséum située loin du rhizome, dans sa moitié supérieure; la face supé- national d’histoire naturelle de Paris (P), de nombreux spéci- rieure du limbe et des nervures est glabre à subglabre; les hyda- mens, auparavant stockés, ont été rendus accessibles. L’examen todes sont couverts d’un dépôt calcaire blanc; chaque segment de l’ensemble des récoltes d’Arthropteris de la région de limbe porte plusieurs sores et les indusies sont également malgache, effectué dans le cadre de la révision en cours des Filicophytes de Madagascar, a permis de mettre en glanduleuses et glabres. Chez l’holotype de A. boutoniana, évidence l’hétérogénéité morphologique intraspécifique de l’articulation du pétiole se situe près du rhizome comme chez A. boutoniana et son origine hybride. A. monocarpa mais, comme chez A. orientalis, la face supé- rieure du limbe et des nervures est glabre à subglabre, les hyda- todes sont couverts d’un dépôt calcaire blanc et chaque Morphologie comparée de Arthropteris boutoniana segment porte plusieurs sores; l’indusie hirsute constitue le et de ses deux parents présumés, A. monocarpa et seul caractère discriminant auquel il faut ajouter une fertilité A. orientalis var. orientalis réduite ou absente – les sporanges étant peu nombreux ou avor- dans la région malgache, Arthropteris monocarpa, tés (hooker, 1854; Florence, com. pers.) – conséquence pro- A. orientalis et A. boutoniana forment un complexe d’espèces bable d’une hybridation entre A. monocarpa et A. orientalis. affines au sein duquel l’identité de A. boutoniana est difficile deux autres spécimens de l’île Maurice et quatre de Madagas- à définir. Aspidium boutonianum hook. (Type: Bouton s.n. car présentent également des indusies hirsutes et une fertilité [k000435681] de l’île Maurice) a d’abord été traité comme réduite ou absente. Les autres caractères morphologiques qui une variété de A. orientalis par ChriSTenSen (1932) et TArdieu- opposent les parents présumés (longueur du phyllopode, pilosité BLoT (1958) puis élevé au rang d’espèce par PiChi SerMoLLi de la face supérieure du limbe et des nervures, présence ou (1966). Ce dernier souligne les affinités mais aussi les diffé- absence de dépôts calcaires blancs sur les hydatodes, nombre rences entre les trois espèces A. boutoniana, A. orientalis et de sores par segment fertile) sont transmis aux hybrides de A. monocarpa. elles possèdent toutes trois un rhizome longue- façon indépendante. L’ensemble de ces observations nous amè- ment rampant, des frondes articulées, un limbe 1-penné- nent à considérer A. boutoniana comme un hybride.

Adresses des auteurs: Fr: Muséum national d’histoire naturelle, département Systématique et evolution, uMr 7205, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. e-mail: [email protected] Cr: Muséum national d’histoire naturelle, département histoire de la Terre, uMr 7207, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France.

ISSN: 0373-2967 – Online ISSN: 2235-3658 – Candollea 68(2): 310-312 (2013) © CONSERVATOIRE ET JARDIN BOTANIQUES DE GENÈVE 2013 MEP Candollea 68-2_. 09.12.13 10:08 Page311

Uvaria lombardii L. Gaut. & Deroin (Annonaceae), une nouvelleNotes espèceon the floraendémique of Madagascar, de Madagascar 30-34 –– 311311

Tableau 1. – Morphologie comparée des parents présumés Arthropteris monocarpa (Cordem.) C. Chr., Arthropteris orientalis (J. F. Gmel.) Posth. var. orientalis et de leurs hybrides. Les caractères morphologiques discriminants chez les hybrides sont en gras dans le tableau. (FR = F. Rakotondrainibe; AFR = Afrique; MAD = Madagascar; MAU = Maurice; REU = Réunion; COM = Comores; SEY = Seychelles). * cf protologue et/ou PICHI SERMOLLI (1966).

Articulation du Limbe et nervures Dépôt calcaire Nombre de Indusie - Sporanges Région (localité) pétiole: position latérales sur hydatodes sores par de la récolte (longueur du (face sup.) segment - Spores phyllopode [cm]) A. monocarpa près du rhizome ± hirsute absent 1 glabre - bien formés AFR, MAD, REU, (Parent 1) (0,5-1,5) glanduleuse - nbreuses ?MAU Bouton s.n. près du rhizome glabre à pr ésent (1-)2-5 hirsute* - bien formés MAU [K000435681] (< 1) subglabre* - peu nbreuses (Montagne longue) (hybride) Anonyme s.n. près du rhizome glabre à présent 1(-2) hirsute* - peu nbreux MAU [BM000785531] (< 1) subglabre* - absentes (?Montagne (hybride) longue)* Lorence 1227 moitié inf. glabre présent 1 hirsute - nbreux avortés + MAU (hybride) (3,0- 4,5) glanduleuse - bien formés (Piton Cantin) - absentes FR 6501[a] médiane ou hirsute abs ent 3-6 hirsute - avortés MAD-Nord FR 6501[c] moitié inf. glanduleuse - absentes (Daraina) (hybrides) (5-12) FR 6501b près du rhizome hirsute absent 3-5 hirsute - avortés MAD-Nord (hybride) (1-2) glanduleuse - absentes (Daraina) Rasolohery 246 moitié sup. hirsute absent 2- 4 hirsute - avortés MAD-Centre (hybride) (9,5-10,5) glanduleuse - absentes (Zahamena) A. orientalis var. orientalis moitié sup. glabre à présent 2-9 glabre - bien formés AFR, MAD, REU, (parent 2) (8-30) subglabre* glanduleuse - nbreuses MAU, COM, SEY

Taxonomie Distribution et écologie. – A ce jour, l’hybride Arthropteris ×boutonania n’a été récolté à Madagascar que dans deux loca- Arthropteris ×boutoniana (hook.) Pic. Serm. in Webbia 21: lités, l’une dans la région de Sambava-daraina (nord-est), 508. 1966 (pro sp.). l’autre dans la réserve de zahamena, dans les environs ϵ Aspidium boutonianum hook. in icon. Pl.: tab. 931. d’Amba tondrazaka (Centre-est) (Fig. 1). il vit en épiphyte 1854. ϵ Arthropteris orientalis var. boutoniana (hook.) dans la forêt sempervirente ou semi-caducifoliée, entre 650 et C. Chr. in h. Perrier, Cat. Pl. Madag. Pterid.: 33. 1932. 1330 m d’altitude. Les deux parents présumés étant sympa- Typus: MAURITIUS: sine loc., s.d., Bouton s.n. (holo-: k triques et largement répandus dans l’île, il n’est pas exclus que, [k000435681] image vue). malgré le fait qu’il semble stérile, cet hybride soit plus fréquent dans la nature. en effet, son rhizome long, grimpant et vivace ϭ Arthropteris orientalis (J. F. gmel.) Posth. × A. mono- carpa (Cordem.) C. Chr., plusieurs formes intermé- assure sa pérennité sur plusieurs années. en dehors de la diaires entre les deux parents présumés. grande ile, A. ×boutonania n’est connu que de l’ile Maurice, dans la vallée des Prêtres, sur le sommet oriental de la Spécimens étudiés. – MADAGASCAR. Prov. Antsiranana: vohémar, daraina, forêt de Binara, 880 m, 4.xi.2001, Rakotondrainibe & al. Montagne longue et sur le Piton Cantin. notons que, à ce jour, 6501[a] (P [P00248603]), ibid loc., 6501[b] (P [P00248604]), ibid loc., la présence d’un seul des deux parents présumés, A. orientalis 6501[c] (P [P00717928]); Ambatondrazaka, réserve de zahamena, à var. orientalis, a été signalée à Maurice. La présence ou la 2 km d’Ankosy, 1100-1330 m, 28.i.2001, Rasolohery 246 (P [P0024 disparition récente de A. monocarpa reste à confirmer. 6410]). MAURICE: sine loc., Anonyme s.n. (BM [BM000785531]); vallée des Prêtres, Piton Cantin, 3.v.1975, Lorence 1227 (P [P00218612]). MEP Candollea 68-2_. 09.12.13 10:08 Page312

312 – Candollea 68, 2013

Références

BAdré, F. (2008). Ptéridophytes. In: AuTrey, J. C., J. BoSSer et i. k FerguSon (ed.), Fl. Mascareignes 1-26. CorneT, A. (1974). Essai de cartographie bioclimatique à Mada - ga s car. notice explicative no. 55, orSToM. hooker, W. J. (1854). Aspidium (Lastrea) Boutonianum. Icon. Pl.: tab. 931. Liu, h.-M., r.-h. JiAng, J. guo, P. hovenkAMP, L. r. Perrie, L. She- Pherd , S. hennequin & h. SChneider (2013). Towards a phylo - ge netic classification of the climbing fern genus Arthro pteris. Taxon 62: 688-700. PiChi-SerMoLLi, r. e. g. (1966). Taxonomic notes on the fern genus “Arthropteris”. Webbia 21: 507-516. roux, J. P. (2009). Synopsis of the Lycopodiophyta and Pteridophyta of Africa, Madagascar and neighbouring islands. Strelitzia 23: 1- 296. TArdieu-BLoT, M.-L. (1958). Polypodiacées (senso lato). In: huM- BerT, h. (ed.), Fl. Madagascar Comores 5(3).

Fig. 1. – Distribution de Arthropteris monocarpa (Cordem.) C. Chr. (carrés), Arthropteris orientalis (J. F. Gmel.) Posth. var. orientalis (triangles) et Arthropteris ×boutoniana (Hook.) Pic. Serm. (étoiles) sur la carte bioclimatique de Madagascar (CORNET, 1974) et la carte de Maurice. Chaque point représente une station dans laquelle le taxon est présent 1 à x-fois.

Remerciements nous adressons nos remerciements à J. Florence qui a vérifié à k et BM la stérilité des spécimens Bouton s.n. (k [k000435681]) et Anonyme s.n. (BM [BM000785531]) et nous a fait bénéficier de ses commentaires sur la première ver- sion du manuscrit. nous remercions aussi nicolas Turland pour ces précieux conseils nomenclaturaux. MEP Candollea 68-2_. 09.12.13 10:08 Page313

33. WAHLERT, Gregory A., Peter B. PHILLIPSON & George E. SCHATZ: Lectotypifications and a new combination in Polygala L. (Polygalaceae) from Madagascar

Introduction Taxonomy The genus Polygala L. (Polygalaceae) was last treated for 1. Polygala longeracemosa h. Perrier in Bull. Trimestriel Madagascar by Perrier de LA BâThie (1955) in the Flore de Acad. Malgache ser. 2, 14: 9. 1932. Madagascar et des Comores series, and later, PAivA (1998) Lectotypus (designated here [second step]): MADAGASCAR. provided a taxonomic conspectus of African and Malagasy Prov. Toliara: Bassin supérieur du Mandrare (Se): species. in the course of re-evaluating species diversity for the Mont Amboahangy près d’esira, [24°15’S 46°39’e], 1000- Catalogue of vascular Plants of Madagascar (MAdAgASCAr 1150 m, 25.xi.1928, fl., Humbert 6830 (P [P00375348]!; CATALogue, 2013), Polygala longeracemosa var. retamoi des isolecto-: P [P00375346, P00375347]!) (Lectotype [first h. Perrier stood out as a morphologically distinctive taxon and step] designated by PAivA, 1998: 258). Paralectotypus: was found to occupy a geographic range separate from that of MADAGASCAR. Prov. Toliara: Massif de Beampingaratra the nominate species. in addition, P. longeracemosa var. reta- (Se): Mont Papanga, versant nW, c. 1200 m, 3-4.xi.1928, moides is a narrowly distributed edaphic specialist restricted fl. & fr., Humbert 6405 (P [P00375343]!; isoparalecto-: P to thin, rocky soils derived from marbles and quartzites in the [P00375344, P00375345]!). vicinity of Ambatofinandrahana in the high Plateau of Observations. – PAivA (1998) cited Humbert 6830 (P) as the Fianarantsoa Prov. We disagree with PAivA (1998) who placed holotype of the name P. longeracemosa and effected lectotypi- P. longeracemosa var. retamoides in synonymy under fication in accordance with the iCBn (MCneiLL & al., 2012, P. longeracemosa h. Perrier. given the morphological and eco- Arts. 9.2, 9.11). however, the gathering of Humbert 6830 at P logical differentiation as well as the separate distributions of consists of three specimens. Consequently, we designate the sheet the two taxa, we propose to recognize P. longeracemosa var. [P00375348] as the [second-step] lectotype as allowed under Art. retamoides at the rank of species. 9.17 because it is the most complete specimen and the label was here, we designate lectotypes for P. longeracemosa and annotated with the word “type” in Perrier’s hand. P. longeracemosa var. retamoides, and we propose the neces- Distribution and habitat. – Polygala longeracemosa is dis- sary new nomenclatural combination and status for the latter tributed in the mountains of southeastern Madagascar, at the species level. We also provide a key to separate to sep- including the massifs of Andohahela, kalambatritra and arate the two species, a distribution map, and preliminary ivakoany and on the horombe Plateau at elevations from 900- iuCn conservation assessments for both taxa following iuCn 1750 m. The species occurs in secondary or degraded grass- red List Categories and Criteria (iuCn, 2012; calculations lands, eri coid shrublands, and forests edges, usually in rocky following CALLMAnder & al., 2007). areas (Fig. 1). Key to separate Polygala longeracemosa from P. retamoides Conservation status. – Polygala longeracemosa is known from 13 locations and subpopulations, four of which are 1. Stem glabrous to sparsely pubescent with simple white situated in Protected Areas (Andohahela national Park and hairs; leaves up to 3.3 cm long, 0.5 cm wide ...... kalambatritra Special reserve), but with a somewhat restricted ...... 1. P. longeracemosa Area of occupancy (minimum of 140 km2) and extent of 1a. Stem densely pubescent with simple white hairs; leaves occurrence (c. 14,000 km2), it is assigned a preliminary con- absent or up to 1.2 cm long, 0.2 cm wide...... servation status of “near Threatened” (nT)...... 2. P. retamoides

Addresses of the authors: gAW: university of utah, department of Biology, 257 South 1400 east, Salt Lake City, uT 84112 u.S.A. e-mail: [email protected] PBP: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and Muséum national d’histoire naturelle, département Systématique et evolution, uMr 7205 oSeB, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. geS: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A.

ISSN: 0373-2967 – Online ISSN: 2235-3658 – Candollea 68(2): 313-315 (2013) © CONSERVATOIRE ET JARDIN BOTANIQUES DE GENÈVE 2013 MEP Candollea 68-2_. 09.12.13 10:08 Page314

314 – Candollea 68, 2013

1100 m, xii.1933, fl., Humbert 13412 (P [P06241677]!); Massif de l’Andohahela, 1700 m, i.1934, fl., Humbert 13673 (P [P06241675]!); horombe, 1300 m, v.1919, fl., Perrier de la Bâthie 12680 (P [P03008 169]!); Mt. Trafonaomby, parcelle i d’Andohahela rni, 24°33’S, 46°43’e, 1000-1957 m, 7.iv.1994, fl., Randriamampionona 752 (P [P02998110]!); Forêt de Befarafara, kalambatritra AP, 23°24’44”S 46°27’58”e, 1575 m, 26.v.2005, fl., Razakamalala & al. 2046 (Mo!, P [P02998109]!, TAn!); à l’e d’Ampandrandava, entre Bekily et Tsivory, vers la source de la Menakompy, [24°07’S 45°48’e], 1000- 1200 m, iv.1943, fl. & fr., Se yrig 646 (P [P06241679]!).

2. Polygala retamoides (h. Perrier) Wahlert, Phillipson & g. e. Schatz, comb. et stat. nova (Fig. 2). ϵ Polygala longeracemosa var. retamoides h. Perrier in Bull. Trimestriel Acad. Malgache ser. 2, 14: 10. 1932. Lectotypus (designated here): MADAGASCAR. Prov. Fiana rantsoa: rocailles cipolins [marble outcrops], Ambatofinandrahana [20°35’S 46°49’e], ii.1919, fl., Perrier de la Bâthie 12441 (P [P00375349]!; isolecto-: P [P00375350]!). Observations. – Perrier de LA BâThie (1932) designated his own collection, Perrier de la Bâthie 12441, as the type of P. longeracemosa var. retamoides. The only known material of the collection are two specimens at P, both of which are annotated with this name in Perrier de la Bâthie’s handwriting and both bear “Type” labels. We designate the sheet [P00375349] as the lectotype because it includes complete stems and inflorescences, as well as the woody base of the stem - an important morphological feature lacking on the sheet [P00375350]. Polygala retamoides is similar to P. longeracemosa by its habit, racemose inflorescence, and morphology of the flowers and seeds, but it differs by its densely pubescent stems (vs. glabrous to sparsely pubescent in P. longeracemosa) and its smaller leaves (when present) up to 1.2 ϫ 0.2 cm (vs. leaves up to 3.3 ϫ 0.5 cm). Polygala retamoides also differs ecolog- Fig. 1. – Distribution of Polygala longeracemosa H. Perrier (circles) and Polygala reta- ically by being restricted to marble and quartzite substrates, moides (H. Perrier) Wahlert, G. E. Schatz & Phillipson (triangles) in Madagascar mapped on the biogeographic zones of CORNET (1974). whereas P. longeracemosa grows on soils derived from laterite and gneiss. Additional material examined. – MADAGASCAR. Prov. Fiana - Distribution and habitat. – Polygala retamoides is restricted rantsoa: iakora, Begogo, Bekora, forêt de Sahalava au S du village to high elevation grasslands on thin, rocky soils overlaying d’Androizaha, 23°32’09”S 46°32’21”e, 1404 m, 27.i.2005, fl. & fr., Andrianjafy & al. 704 (Mo!, P [P06241680, P06241681]!). Prov. marble and quartzite substrates from 1300-1800 m, near Amba - Toliara: Betroka, ivahona, Mt. kalambatritra, forêt d’Analamaro, tofinandrahana, Fianarantsoa Province in the eastern part of 23°27’38”S 46°24’40”e, 1540 m, 6.xi 2005, fl., Andrianjafy & al. the itremo Complex (Fig. 1). 536 (Mo!, P [P02998108]!); Ampandrandava, [24°05’S 45°42’e], [900 m], s.d., fl. & fr., Herb. Jardin Bot. Tananarive 6276 (P [P06241 Conservation status. – data on the occurrence of P. reta- 684]!); Massif du kalambatritra (Centre-Sud), [23°22’S 46°29’e], moides is very limited, with only two of the eight known gath- 1600-1750 m, xi.1933, fl, Humbert 11772 (Br, g!, k!, Mo!, P erings bearing precise geo-references. however all of these [P03008170, P06241683]!, Pre, TAn, WAg); haute vallée de la were apparently made near the town of Ambatofinandrahana, Manambolo, affluent de l’ionaivo, [24°30’S, 46°35’e], 900-1100 m, and none are known from the western part of the itremo xi.1933, fl. & fr., Humbert 12132 (P [P06241678]!); Massif de l’ivakoany, [23°50’30”S, 46°26’e], 1250 m, xi-xii.1933, fl. & fr., Complex which is now under temporary protection as a new Humbert 12253 (P [P06241685]!); vallée de la Sakamalio, affluent protected area. These gatherings probably represent a single de la Manambolo, bassin du Mandrare, [24°32’S 46°41’e], 900- location and sub-population with a highly restricted extent of MEP Candollea 68-2_. 09.12.13 10:08 Page315

Uvaria lombardii L. Gaut. & Deroin (Annonaceae), une nouvelleNotes espèceon the floraendémique of Madagascar, de Madagascar 30-34 –– 315315

(P!); Mahavanino, 7 km au S d’Ambatofinandrahana, sur la route de Fenoarivo, lieu-dit de Analalehibe, Ambatofinandrahana, 20°37’43”S 46°50’32”e, 1350 m, 19.xi.1993, fl., Labat & al. 2370 (k!, Mo!, P [P001718]!, TAn!); Ambatofinandrahana, 9 km from Ambatofinan- drahana, 20°36’47”S 46°49’56”e, 1398 m, 18.iii.2004, fl., Skema & al. 46 (Mo!, TAn).

Acknowledgements We are grateful to the curators of the herbaria in g, k, Mo, P, and TAn for providing access to their collections. We also acknowledge our late friend and colleague Jean-noël Labat who collected and photographed Polygala retamoides. The authors were supported by grants from the u.S. national Science Foundation (0743355) and the Andrew W. Mellon Foundation.

References

CALLMAnder, M. W., g. e. SChATz, P. P. LoWry ii, M. o. LAivAo, J. rAhAriMAMPiononA, S. AndriAMBoLoLonerA, T. rAMinoSoA & T. ConSigLio (2007). Application of iuCn red List criteria and assessment of Priority Areas for Plant Conservation in Mada- gascar: rare and threatened Pandanaceae indicate new sites in need of protection. Oryx 41: 168-176. CorneT, A. (1974). Essai de cartographie bioclimatique à Mada - ga scar. notice explicative 55. orSToM. LoWry, P. P. ii & g. e. SChATz (2006). A new restricted-range species of Buxus L. (Buxaceae) from central Madagascar. Adansonia ser. Fig. 2. – Living plant of Polygala retamoides (H. Perrier) Wahlert, G. E. Schatz & Phillipson, 3, 28: 67-70. growing in its natural habitat on the marble outcrops south of Ambatofinandrahana. MCneiLL, J., F. r. BArrie, W. r. BuCk, v. deMouLin, W. greuTer, [Labat & al. 2370] [Photo: J.-N. Labat] d. L. hAWkSWorTh, P. S. herendeen, S. knAPP, k. MArhoLd, J. PrAdo, W. F. Prud’hoMMe vAn reine, g. F. SMiTh, J. h. occurrence and Area of occupancy of no more than a 10 km2 WierSeMA & n. J. TurLAnd (2012). international Code of nomenclature for algae, fungi and plants (Melbourne Code) in an area that is under considerable human pressure from fire adopted by the eighteenth international Botanical Congress and mining; Polygala retamoides is therefore assigned a pre- Melbourne, Australia, July 2011. Regnum Veg. 154. liminary conservation status of “Critically endangered” [Cr PAivA, J. A. r. (1998). Polygalarum Africanarum et Madagascarien- B1ab(i,ii,iii,v)+2ab(i,ii,iii,v)]. it is interesting to note that other sium prodromus atque gerontogaei generis heterosamara kuntze, species also appear to be endemic to the marble outcrops near a genere Polygala L. segregati et a nobis denuo recepti, synopsis Ambatofinandrahana but are not known from elsewhere in the monographica. Fontqueria 50. itremo Complex, for example Buxus cipolinica Lowry & g. Perrier de LA BâThie, h. (1932). Les Polygala de Madagascar. Bull. e. Schatz (LoWry & SChATz, 2006), suggesting the importance Trimestriel Acad. Malgache ser. 2, 14: 1-30. of this area to be brought into the protected area network of Perrier de LA BâThie, h. (1955). Polygalacées. In: huMBerT, h. Madagascar. (ed.), Fl. Madagascar Comores 109. Additional material examined. – MADAGASCAR. Prov. Fia- iuCn (2012). IUCN Red List Categories and Criteria: version 3.1. narantsoa: Ambatofinandrahana, [20°35’S 46°49’e], x.1963, fl., 2nd edition. iuCn Species Survival Commission, gland & Cam- Bosser 18119 (P!); ibid. loc., 1964, fl., Bosser 19611 (P!); au S d’Am- bridge. batofinandrahana, [20°35’S 46°49’e], 13.xii.1976, fl., Cremers 3645 (P!, TAn!); env. d’Ambatofinandrahana [20°37’19”S 46°50’09”e], MAdAgASCAr CATALogue (2013). Catalogue of the vascular Plants 1600-1800 m, 21.ii.1938, fl., Decary 13174 (Mo!, P [2 sheets]!); of Madagascar. Missouri Botanical garden, St. Louis & Antana- Ambatofinandrahana [20°35’S 46°49’e], iii.1960, fl., Keraudren 219 narivo [http://www.efloras.org/madagascar]. MEP Candollea 68-2_. 09.12.13 10:08 Page316

34. CALLMANDER, Martin W., Peter B. PHILLIPSON, James A. WEARN & Rogier P. J. de KOK: Adieu Adelosa Blume (Lamiaceae): Further observations on Rotheca Raf. for Madagascar

Introduction in the context of our review of the genus Clerodendrum and its segregates for the MAdAgASCAr CATALogue (2013), The genus Adelosa Blume (Lamiaceae) and its single we (MWC and PBP) examined the isotype material of Adelosa species, Adelosa microphylla Blume, were first described by microphylla held at P. it comprises two separate sheets, each BLuMe (1850) based on a single gathering made by Auguste with several portions of Pervillé’s collection, and in contrast Pervillé in northern Madagascar. This plant has puzzled tax- with the holotype at L one of the sheets [P00091374] possesses onomists for more than a century and a half. BAiLLon (1891) a mature flower. The corolla, although not in a very good con- in his Histoire des Plantes did not accept the genus and placed dition, was observed to be strongly zygomorphic, and to it in synonymy under Clerodendrum L., a conclusion followed expand abruptly at the mouth of the tube, distinctive charac- by BriqueT (1895). More recently, MoLdenke (1981) resur- teristics of the genus Rotheca, and we realized that Adelosa rected Adelosa, placing it near Premna L. and Clerodendrum, microphylla was none other than the widespread and variable but nevertheless expressed uncertainty about this. one of us Rotheca incisa (klotzsch) Steane & Mabb., which is native to (rPJk), re-examined the holotype (Pervillé 629) in Leiden Madagascar and Africa (STeAne & MABBerLey, 1998). how- herbarium (L) in an attempt to resolve the problem. Careful ever Blume’s name predates Clerodendrum incisum klotzsch, dissection of the flower buds revealed the style to be attached the basionym of Rotheca incisa, and we therefore make the between the four unilocular lobes of the ovary, a condition that required new combination R. microphylla (Blume) Callm. & is found in Clerodendrum (sensu lato) and it was concluded Phillipson. We have reviewed the many names that have been that Adelosa microphylla Blume belongs in this genus (de kok, treated as synonyms of this species, and its morphological vari- 2001). however the specimen at L could not be identified to ation across Madagascar and on the African continent, as well any known species of Clerodendrum from Madagascar with any confidence, due to its fragmentary condition. as other apparently closely related species. Clerodendrum has generally been circumscribed to com- during this study, additional material of the poorly-known prise c. 400-500 species native to tropical and warm temperate Clerodendrum mirabile Baker was brought to light. This regions of the world with its centre of diversity in tropical enabled us to ascertain that it also belongs to the genus Rotheca and to elucidate its geographic distribution, and we provide regions of Africa and Asia (hArLey & al., 2004), but recent molecular phylogenetic studies have revealed that it is poly- the necessary new combination: R. mirabilis (Baker) Callm. phyletic (STeAne & al., 1997, 1999, 2004; yuAn & al., 2010), & Phillipson. and several genera have been segregated from Clerodendrum, We have also re-examined the status of Rotheca myricoides leaving it more narrowly circumscribed and limited to c. 180 (hochst.) Steane & Mabb., which is treated as native to Mada- described species (WeArn & MABBerLey, 2011a, 2011b). in gascar in the Flore de Madagascar et des Comores (as Clero- Madagascar, Clerodendrum sensu Wearn & Mabberley, as well dendrum myricoides hochst.), based on the collection “Warbur as the segregate genera Rotheca raf. and Volkameria L., are 553” [Warburg?] cited as from Madagascar without locality present (CALLMAnder & PhiLLiPSon, 2012). (MoLdenke, 1956). The specimen of this gathering at P is

Addresses of the authors: MWC: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and Conservatoire et Jardin botaniques de la ville de genève, ch. de l’impératrice 1, CP 60, 1292 Chambésy, genève. Switzerland. e-mail: [email protected] PBP: Missouri Botanical garden, P.o. Box 299, St. Louis, Mo, 63166-0299, u.S.A. and Muséum national d’histoire naturelle, département Systématique et evolution, uMr 7205 oSeB, CP 39, rue Cuvier 57, 75231 Paris, cedex 05, France. JAW, rPJk: herbarium, Library, Art & Archives, royal Botanic gardens, kew, richmond, Surrey, TW9 3dS, united kingdom.

ISSN: 0373-2967 – Online ISSN: 2235-3658 – Candollea 68(2): 316-320 (2013) © CONSERVATOIRE ET JARDIN BOTANIQUES DE GENÈVE 2013 MEP Candollea 68-2_. 09.12.13 10:08 Page317

Uvaria lombardii L. Gaut. & Deroin (Annonaceae), une nouvelleNotes espèceon the floraendémique of Madagascar, de Madagascar 30-34 –– 317317

clearly marked as having been collected in a garden. We have found no evidence of the species occurring naturally in Mada- gascar, nor of this commonly cultivated ornamental plant having become in any way naturalized. in this note we present preliminary threat assessments for Rotheca microphylla and R. mirabilis for Madagascar following iuCn red List Categories and Criteria (iuCn, 2012), and an updated species key to the native Malagasy species of Rotheca. Lists of representative collections can be found in the MAdAgASCAr CATALogue (2013).

Key to the Malagasy species of Rotheca 1. inflorescence developing on old wood early in the season, generally produced on leafless stems before vegetative growth has commenced; corolla tube shorter than the corolla lobes ...... Rotheca nudiflora 1a. inflorescence developing on young growth during the growing season, at the same time or after the leaves, stems with leaves at time of anthesis; corolla tube many times longer than corolla lobes ...... 2 2. Calyx 2-2.5 cm in length at anthesis ...... Rotheca mirabilis 2a. Calyx 0.5-1 cm in length at anthesis ...... Rotheca microphylla

Systematics 200 km Rotheca microphylla (Blume) Callm. & Phillipson, comb. nova. ϵ Adelosa microphylla Blume, Mus. Bot. 1: 176. 1850. Typus: MADAGASCAR. Prov. Antsiranana: Ambongo, croît au bord de la mer, 14.ii.1841, fl., Pervillé 626 (holo-: Fig. 1. – Distribution of Rotheca microphylla (Blume) Callm. & Phillipson (triangles) L [L0062225]!; iso-: P [P00091374, P00091375]!, ny and Rotheca mirabilis (Baker) Callm. & Phillipson (stars) in Madagascar mapped on the biogeographic zones of CORNET (1974). [ny00103724] image seen). ϭ Clerodendrum incisum klotzsch in Peters, naturw. reise Mossambique, Bot. 1: 257. 1861. ϵ Rotheca incisa (klotzsch) Steane & Mabb. in novon 8: 205. 1998. Typus: MOZAMBIQUE: rios de Sena, querimba ϭ Clerodendrum bernieri Briq. in Bull. herb. Boissier 4: & Boror, Peters s.n. (holo-: B; iso-: k [k000192940]!), 348. 1896. Typus: MADAGASCAR. Prov. Antsiranana: syn. nov. sur les rochers calcaire de Lingvatou, [12°26’S 49°30’e], e ϭ Clerodendrum macrosiphon hook f. in Curtis’s Bot. Mag.: 1835, fl., Bernier 169 (2 envoi) (holo-: g [g00366 tab. 6695. 1883. ϵ Clerodendrum incisum var. macrosi - 312]!; iso-: P [P00446637, P02865951, P02865955]!), phon (hook. f.) Baker in Fries, Fl. Scan. 5: 307. 1900. syn. nov. Typus: TANZANIA. Zanzibar: usaramo, hort. kew, ϭ Clerodendrum pusillum gürke in Bot. Jahrb. Syst. 30: 19.v.1882, Kirk s.n. (holo-: k [k000607526]!), syn. nov. 390. 1901. Typus: TANZANIA. MBEYA DISTR.: unyika, ϭ Clerodendrum lindemuthianum vatke in Linnea 4: 348. nsan gamales, s.d., Goetze 1393 (holo-: B), syn. nov. 1896. Typus: MADAGASCAR. Prov. Antsiranana: ϭ Clerodendrum incisum var. vinosum Chiov., Fl. Somala vavatobé, ii.1880, fl., Hildebrandt 3332 (holo-: B; iso-: 2: 364. 1932. Typus: SOMALIA: Licchitore, 22.vii. BM [BM000923603] image seen, k [k000192958]!; 1929, fl., Senni 542 (holo-: FT [FT002833] image P [P00446638, P00446639]!), syn. nov. seen), syn. nov. MEP Candollea 68-2_. 09.12.13 10:08 Page318

318 – Candollea 68, 2013

ϭ Clerodendrum incisum var. longepedunculatum B. with habitat and locality, the small-leaved specimens occurring Thomas in Bot. Jahrb. Syst. 68: 78. 1936. Typus: in exposed arid places such as isalo national Park, and the Kenya: coast opposite Lamu, s.d, Hildebrandt 1911 large-leaved specimens from more humid forest locations such (holo-: B), syn. nov. as Manongarivo Special reserve, but floral characteristics ϭ Clerodendrum incisum var. parvifolium Moldenke in appear to be quite constant. in Africa most collections have Phytologia 3: 407. 1951. Typus: MADAGASCAR. Prov. large leaves, and the marginal teeth tend to be fewer and more Antsiranana: vallée moyenne du Mandrare, près deeply incised, but again the floral morphology is not distin- d’Anda dabolava, Mont vohitrosy, 800-850 m, xii. guishable from the Malagasy plants. it is tempting to separate 1933, fl., Humbert 12741 (holo-: P [P00446640]!; iso-: the Malagasy and African plants as different species, but after ny [ny001377354] image seen, P [P00446641]!), syn. careful examinations of specimen held at g, k, P and Mo, we nov. believe that the differences in leaf margins are not clear-cut. in the absence of a better understanding of the causes of the ϭ Clerodendrum incisum var. afzelii Moldenke in Amer. variation, we conclude that it is currently impossible to recog- J. Bot. 38: 325. 1951. Typus: MADAGASCAR. Prov. nize meaningful separate taxa even at infraspecific rank, and Tulear: Manasoa Tanosy, 13.i.1913, fl., Afzelius s.n. we maintain the existing broad concept of the species. (holo-: k [k000607527]!; iso-: ny image seen), syn. nov. Conservation status. – With an “extent of occurrence” (eoo) of 436,663 km2, and an “Area of occupancy” (Aoo) ϭ Clerodendrum dalei Moldenke in Phytologia 4: 287. of 315 km2, and 29 known subpopulations, seven of which are 1953. Typus: KENYA: near Marjoreni, S. digo, ix. within protected areas (Ankarafantsika, Ankarana, Bemaraha, 1937, fl., Dale 3811 (holo-: Br [Br0000008715517] isalo, Loky Manambato, Manongarivo and Forêt de Mikea). image seen; iso-: eA [eA000001136, eA000001137] it is thus assigned a preliminary status of “Least Concern” images seen, k [k000192876]!, ny [ny00137334] [LC]. image seen), syn. nov. Observations. – Rotheca microphylla is a highly variable Rotheca mirabilis (Baker) Callm. & Phillipson, comb. nova. species that occurs throughout much of western, southern and northern Madagascar (Fig. 1), where it is largely confined to ϵ Clerodendrum mirabile Baker in J. Linn. Soc., Bot. 22: the dry Bioclimatic region (CorneT, 1974), and the western 513. 1887. slopes of the central plateau. it is also native to eastern and Typus: MADAGASCAR: Central, s.d., fl., Baron 4755 (holo-: southern parts of Africa in Angola, kenya, Malawi, Mozam- k [k000192959]!; iso-: BM [BM000923602] image seen, bique, Somalia, Tanzania (including zanzibar) and zambia. it ny [ny00137366] image seen, P [P00446657]!). occurs in dry forest and woodland, mostly from sea level to Observations. – The type collection has no indication about 800 m elevation, but occasionally higher, where it forms of origin other than “Central Madagascar”, but in the Flore a lianescent shrub typically scrambling up to c. 3 m high or de Madagascar et des Comores (MoLdenke, 1956), a second more in the under-storey and on the forest margins, but some- collection was cited for the species – Baron 6889 (k [k0006 times even higher. it also grows in open areas of degraded for- 07528]), collected at “Ankavandra” a small town in the Manam- est, bush and secondary anthropogenic grassland, where it bolo river valley in the Menabe region of Toliara Province. The tends to develop a low-growing spreading growth-form. it discovery of a third specimen, Du Puy MB733 (Mo-3846948, P appears to be somewhat tolerant of grazing, trampling and [P00075324]), collected recently and with precise locality just burning and can become locally abundant in disturbed 40 km SW of Ankavandra, helps to confirm the species as a likely with a somewhat weedy tendency. With its prolific flowers, highly-restricted endemic, possibly confined to the area just to distinctive erect crochet-like buds, and conspicuous long white the east of the Bemaraha protected area (Fig. 1). corolla tube and spreading lobes (see Fig. 2), it is a decorative Conservation statu s. – With only three collections known plant that has been introduced into cultivation as an ornamental (Aoo of 27 km2), two of which dating back to a century and in tropical and sub-tropical gardens in various parts of the only one recent, all of them falling outside the Protected Area world, sometimes marketed in north America as the “Musi- network, Rotheca mirabilis can be assessed as “Critically cal-note Plant”. endangered” [Cr 2ab(i,ii,iii,iv)] following the iuCn red List in addition to its variable habit, the diversity in Rotheca Categories and Criteria (iuCn, 2012) (calculation following microphylla is most marked with respect to leaf shape and size. CALLMAnder & al., 2007). in Madagascar leaves vary from lanceolate and no more than 25 ϫ 8 mm and almost entire with at most a few shallow mar- ginal teeth, to elliptic and as much as 100 ϫ 50 with coarsely serrate margins. To some extent this variation is also correlated MEP Candollea 68-2_. 09.12.13 10:08 Page319

Uvaria lombardii L. Gaut. & Deroin (Annonaceae), une nouvelleNotes espèceon the floraendémique of Madagascar, de Madagascar 30-34 –– 319319

Fig. 2. – Living plant of Rotheca microphylla (Blume) Callm. & Phillipson.

[Photo: J. Bosser] MEP Candollea 68-2_. 09.12.13 10:08 Page320

320 – Candollea 68, 2013

Acknowledgements iuCn (2012). IUCN Red List Categories and Criteria: Version 3.1. 2nd edition. iuCn Species Survival Commission, gland & Cam- The authors are grateful to the curators of the herbaria in bridge. geneva (g), kew (k), Paris (P) and Saint-Louis (Mo) for MAdAgASCAr CATALogue (2013). Catalogue of the Vascular Plants allowing access to specimens. We thank Jean Bosser, formally of Madagascar. Missouri Botanical garden, St. Louis & Antana- at Mnhn, for providing the photo of R. microphylla (Fig. 2). narivo [http://www.efloras.org/madagascar]. Financial support was provided by Andrew W. Mellon Foun- MoLdenke, h. n. (1956). verbénacées. In: huMBerT, h. (ed.), Fl. dation. Madagascar Comores 174. MoLdenke, h. n. (1981). notes on the genus Adelosa (verbenaceae). Phytologia 48: 392-393. STeAne, d. A. & d. J. MABBerLey (1998). rotheca (Lamiaceae) References revived. Novon 8: 204-206. BAiLLon, h. (1891). verbénacées. Hist. Plantes 11: 78-121. hachette STeAne, d. A., r. W. SCoTLAnd, d. J. MABBerLey, S. J. WAgSTAFF, & Co., Paris. P. A. reeveS & r. g. oLMSTeAd (1997). Phylogenetic relation- BLuMe, C. L. (1850). Mus. Bot. 1(11): 176. ships of Clerodendrum s.l. (Lamiaceae) inferred from chloroplast dnA. Syst. Bot. 22: 229-243. BriqueT, J. (1895). verbenaceae. In: engLer, A. & k. PrAnTL (ed.), Nat. Pflanzenfam. iv(3a): 132-182. Leipzig. STeAne, d. A., r. W. SCoTLAnd, d. J. MABBerLey & r. g. oLMSTeAd (1999). Molecular systematics of Clerodendrum (Lamiaceae): CALLMAnder, M. W. & P. B. PhiLLiPSon (2012). An endemic species iTS sequences and total evidence. Amer. J. Bot. 86: 98-107. of rotheca raf. (Lamiaceae) from Madagascar. Candollea 67: 370-372. STeAne, d. A., r. P. J. de kok & r. g. oLMSTeAd (2004). Phyloge- netic relationships between Clerodendrum (Lamiaceae) and other CALLMAnder, M. W., g. e. SChATz, P. P. LoWry ii, M. o. LAivAo, Ajugoid genera inferred from nuclear and chloroplast dnA J. rAhAriMAMPiononA, S. AndriAMBoLoLonerA, T. rAMinoSoA sequence data. Molec. Phylogen. Evol. 32: 39-45. & T. ConSigLio (2007). Application of iuCn red List criteria and assessment of Priority Areas for Plant Conservation in Mada- WeArn, J. A. & d. J. MABBerLey (2011a). Clerodendrum confusion gascar: rare and threatened Pandanaceae indicate new sites in - redefinition of, and new perspectives for, a large labiate genus. need of protection. Oryx 42: 168-176. Gard. Bull. Singapore 63: 119-124. CorneT, A. (1974). Essai de cartographie bioclimatique à Mada - WeArn, J. A. & d. J. MABBerLey (2011b). Clerodendrum (Lamia - ga scar. notice explicative 55. orSToM. ceae) in Borneo. Syst. Bot. 36: 1050-1061. de kok, r. P. J. (2001). A note on the status of the enigmatic mono- yuAn, y.-W., d. J. MABBerLey, d. A. STeAne & r. g. oLMSTeAd typic genus Adelosa (Labiatae). Blumea 46: 585-587. (2010). Further disintegration and redefinition of Clerodendrum (Lamiaceae): implications for the understanding of the evolution hArLey, r. M., S. ATkinS, A. L. BudAnTSev, P. d. CAnTino, B. J. of an intriguing breeding strategy. Taxon 59: 125-133. Conn, r. grAyer, M. M. hArLey, r. de kok, T. kreSTovSkAJA, r. MorALeS, A. J. PATon, o. ryding & T. uPSon (2004). Labiatae. In: kuBiTzki, k. & J. W. kAdereiT (ed.), The families and genera of vascular plants 7: 167-275. Springer, Berlin.