235

Lyndley A. CRAVENAustralian National Herbarium, CPBR, CSIRO Plant Industry,

GPO Box 1600, Canberra, ACT 2601 (Australia)lyn.craven@csiro.au

Bernard E. PFEILAustralian National Herbarium, CPBR, CSIRO Plant Industry,

GPO Box 1600, Canberra, ACT 2601 (Australia)Present address: Department of Plant Biology, Cornell University,

Ithaca, NY 14853 (USA)bep27@cornell.edu

Australian representatives of Macrosteliatransferred to Hibiscus (Malvaceae), with the description of a new species

ABSTRACTThe Australian species of Macrostelia fall within the circumscription ofHibiscus and the two taxa concerned are transferred to the latter genus. A newspecies is described. Three new names are proposed: H. tozerensis Craven &B.E. Pfeil (based on M. grandifolia Fryxell), H. macilwraithensis (Fryxell)Craven & B.E. Pfeil (based on M. grandifolia subsp. macilwraithensis Fryxell),and H. propulsator Craven & B.E. Pfeil (for the new species).

RÉSUMÉTransfert des représentants australiens de Macrostelia dans le genre Hibiscus(Malvaceae) et description d’une espèce nouvelle.L’espèce australienne de Macrostelia relève de la circonscription d’Hibiscus etles deux taxons concernés sont transférés dans ce dernier. Une nouvelle espèceest décrite. Trois noms nouveaux sont proposés : H. tozerensis Craven &B.E. Pfeil (basé sur M. grandifolia Fryxell), H. macilwraithensis (Fryxell)Craven & B.E. Pfeil (basé sur M. grandifolia subsp. macilwraithensis Fryxell)et H. propulsator Craven & B.E. Pfeil (pour l’espèce nouvelle).

ADANSONIA, sér. 3 • 2004 • 26 (2) : 235-240© Publications Scientifiques du Muséum national d’Histoire naturelle, Paris.

KEY WORDSMalvaceae,

Hibiscus, Macrostelia,

Australia.

MOTS CLÉSMalvaceae,

Hibiscus, Macrostelia,

Australie.

HOCHREUTINER erected the genus MacrosteliaHochr. to accommodate two new malvaceousspecies from Madagascar, M. calyculata Hochr.and M. involucrata Hochr. (HOCHREUTINER

1952). The new genus was primarily distin-guished by HOCHREUTINER from other genera ofMalvaceae in having the petals proximallyconnate into a long tube; in other genera thepetals are connate only at the very base. Sub-sequently a third Malagasy species was added tothe genus, M. laurina (Baill.) Hochr. & Humbert(HOCHREUTINER 1955). FRYXELL extended thegeographic range of Macrostelia to Australia anddescribed a new species, M. grandifolia Fryxell, inwhich two subspecies were recognised, M. grandi-folia subsp. macilwraithensis Fryxell and the auto-nymic subspecies (FRYXELL 1974).

The Australian macrostelias were included inan analysis of chloroplast sequence data from thendhF coding region and the non-coding rpl16intron from a comprehensive sample of species ofHibiscus L. and other genera of Hibisceae (PFEIL

et al. 2002). Macrostelia consistently nestedwithin Hibiscus and, in a combined analysis ofndhF and rpl16 data, grouped in a well supportedclade that also included species of H. sectionsBombicella DC., Hibiscus and Lilibiscus Hochr.,including the type of Hibiscus, H. syriacusL. Unfinished studies utilising ndhF data suggestthat the Malagasy species of Macrostelia also nestwithin the clade containing the AustralianMacrostelia species but are not grouped withthem (PFEIL and BAUM unpublished data). Inassigning the new Australian species to Macro-stelia, FRYXELL (1974) noted that the Australianspecies lacked the degree of adnation of petals tostaminal column that was characteristic of theMalagasy species. Our observations are in accord-ance with those of FRYXELL and we have furtherobserved that the degree of petal adnation mani-fest in the Australian Macrostelia species is com-parable to that in all the Hibiscus species we haveexamined. Taking their morphology into accountwith the results of the molecular sequenceresearch, it would seem that the two groups ofMacrostelia species are not as closely related aspreviously believed.

Of the other morphological features discus-sed by FRYXELL (1974) wherein Macrostelia dif-

fers from Hibiscus, the most distinctive is thatof leaf venation. Macrostelia has pinnatelyveined, unlobed leaves whereas Hibiscus fideFRYXELL (1974) has palmately veined andusual ly palmately lobed leaves (althoughHibiscus papuodendron Kosterm., at least, haspinnately veined leaves [pers. obs.]). Both typesof leaf venation may occur in other groups ofMalvaceae, however. FRYXELL (1999) recordsthe leaves of Pavonia subg. Goetheoides (Gürke)Ulbr. as being pinnately veined (the term pen-ninerved is used by FRYXELL) but the species ofthe other subgenera of Pavonia apparently havepalmate venation. Pinnate venation in factoccurs within three clades of a broadly definedHibiscus, namely Azanzae s.l. (H. papuodendronnests herein), Trionum s.l. (Pavonia nests here),and Hibiscus-Bombicella s.l. (Macrostelia nestshere) (PFEIL et al. 2002). The molecular sup-port for the basic structure and membership ofthese three clades is strong in both cpDNA(PFEIL et al. 2002) and nDNA (PFEIL et al.2004) and pinnate venation must be regardedas a character with poor predictive power as toclassification. It is pertinent that, even in spe-cies of Hibiscus s.l. with palmate venation, thepalmate condition applies only to the proximalveins (which often are much more developedthan the other veins); the distal veins are pin-nate. In this group of plants the difference bet-ween palmate and pinnate venation may bemore reflective of the relative degree of develop-ment of the veins than of a divergence in basicstructure per se.

It is our conclusion that Macrostelia includedwithin Hibiscus may be a better taxonomic place-ment than if it is kept separate. Given the sharedcharacter states of apically free styles, capitatestigmas, sterile staminal column apices, lack ofgossypol, and capsular fruit, pinnate venation isa weak character upon which to maintain agenus. However, given that the molecular syste-matics of the Malagasy species of Macrosteliahave not yet been studied to the same extent asthe Australian species, Macrostelia is being main-tained in the interim for its Malagasy representa-tives. The necessary nomenclatural transfers ofthe Australian taxa are effected below. A newAustralian macrostelioid species, H. propulsator

Craven L.A. & Pfeil B.E.

236 ADANSONIA, sér. 3 • 2004 • 26 (2)

Australian Macrostelia transferred to Hibiscus

237ADANSONIA, sér. 3 • 2004 • 26 (2)

Craven & B.E. Pfeil, is described also. AlthoughFRYXELL’s two subspecies of M. grandifolia aregenerally similar in their leaves, the flowers arevery different (Fig. 1). The epicalyx segments arefree in subsp. macilwraithensis and 6-7 in num-ber, whereas in subsp. grandifolia they areconnate for about half their length and 4-5 innumber. In addition, the petals of the former arespreading at anthesis and white, and of the latterare erect at anthesis and pale pink. These distinc-tive differences are maintained in wild andgreenhouse-grown material and no intermediateforms have been seen, thus we believe that subsp.macilwraithensis warrants recognition at speciesrank. The collector of two collections ofM. grandifolia, the late L.J. BRASS, indicatedthrough specimen label annotation (on Brass19352, 19353) that the species was dioecious.FRYXELL suggested that the species might best bedescribed as gynodioecious as the “staminate”herbarium specimen he examined had fullydeveloped styles and stigmas (FRYXELL 1974).The flowers of the Australian representativesdiffer in epicalyx and corolla form and are illus-trated in Figure 1. Observations by us of livingplants of each of the Australian species showsthat the species have hermaphrodite flowers butare strongly protandrous. At early anthesis, thestyle branches are small, erect and appressed toeach other, with the stigmatic hairs not readilyevident; at this stage the anthers have dehiscedand pollen is available for dispersal. Over thenext few days, the style branches elongate, sepa-rate and spread widely such that the stigmas,which by now have well developed stigmatichairs, may be brought into contact with the sta-mens. In nature, it is expected that insects pro-bably would have removed the pollen from aflower prior to the stigmas becoming receptive;thus ensuring that the flower outcrosses. In thegreenhouse, however, the pollen remains withthe anthers as insects that potentially could polli-nate flowers are excluded; selfing presumablyoccurs as we have noted that fruit is commonlyproduced on one accession of subsp. macilwrai-thensis. The Australian “Macrostelia” species dif-fer from the other Australian species of Hibiscusin that they commonly have several flowers ineach leaf axil.

1. Hibiscus macilwraithensis (Fryxell) Craven& B.E. Pfeil, comb. et stat. nov.

Macrostelia grandifolia subsp. macilwraithensis Fryxell,Austral. J. Bot. 22: 191 (1974), basionym. — Type:Australia: Queensland, McIlwraith Range, c. 18 kmNE by E of Coen, 12 Oct. 1962, Smith 11783(holo-, BRI n.v.).

FIG. 1. — Flowers of Australian Hibiscus: A, H. macilwraithensis(Fryxell) Craven & B.E. Pfeil; B, H. propulsator Craven &B.E. Pfeil; C, H. tozerensis Craven & B.E. Pfeil; a, flower atearly anthesis; b, flower at mid-anthesis; c, flower at late anthesis.From plants cultivated at Canberra, Australia.

2. Hibiscus propulsator Craven & B.E. Pfeil,sp. nov.

A H. macilwraithensi (Fryxell) Craven & B.E. Pfeilsegmentis epicalycis linearibus usque lineariellipticis etcorolla rubra, et H. tozerensi Craven & B.E. Pfeil epi-calyce 7-9-segmentatis et ad basim divisis et corolla effusarubraque differt.

TYPUS. — Australia: Queensland, Temple Bay,c. 2 km NW of Bolt Head, 12°15’S, 142°04’E, 12 July1991, Forster 8937 (holo-, BRI; iso-, A, B, BISH, BO,CANB, G, K, L, LAE, MEL, P, QRS).

Shrub to 3 m tall, multistemmed from belowground level; branchlets densely stellate-hairy, theinternodes 0.3-2.5 cm long; stipules filiform,deciduous, 0.25-1.2 cm long. Leaves alternate,4.5-28 cm long, 1.5-9 cm wide; lamina narrowlyobovate to narrowly subpanduriform, the baserounded to cordate, the apex acuminate to round-ed, the margin distantly serrate (often obscurelyso), the venation pinnate with 6-10 primary veinson each side of the midrib, weakly discolorous,the abaxial and adaxial surface sparsely and minute-ly stellate-hairy (excepting the midrib which ismore densely stellate-hairy than the remainder ofthe surface) with the hairs predominantly orexclusively inserted on the midrib and veins;petiole 0.2-0.5 cm long, stellate-hairy. Flowers 1-2(-3) per leaf axil; peduncle and pedicel com-bined 0.6-2.8 cm long, the articulation aboutone-third to half-way from the peduncle base,stellate-hairy; epicalyx segments 7-9, linear tolinear-elliptic, 0.7-1.2 cm long, free to the base,stellate-hairy; calyx narrowly ovoid, valvate, 5-lobed (2 or 3 lobes sometimes tardily separatingfrom the adjacent lobe), 1-2 cm long, the lobes0.3-0.6 cm long, stellate-hairy; petals 5, imbri-cate, adnate to the ovary/staminal column forc. 0.3 cm but otherwise free, spathulate with theapex obliquely truncate, 2.2-3.7 cm long, deeprose pink, wine-red or red, stellate-hairy on theabaxial surface; staminal column 2.4-3.2 cmlong, white, 5-toothed at the apex, stellate-hairyproximally and glandular-hairy for the remainder(especially so on the apical teeth); stamens 20, ina whorl of 10 pairs (the stamens of each pairsuperposed), the whorl 0.2-0.3 cm below the apexof the staminal column, the filaments spreading-

ascending, 0.3-0.6 cm long, white; style branches 5,erect and appressed to each other at anthesis andlater separating, elongating and recurving, 0.3-0.8 cm long, very pale pink, the stigmas capitate,hairy; ovary conical, 0.5 cm long, stellate-hairy,5-loculate with c. 6 ovules per locule. Fruit notseen. — Figs 1B; 2.

DISTRIBUTION AND ECOLOGY. — Hibiscus pro-pulsator occurs on the eastern side of Cape YorkPeninsula in the Olive River-Bolt Head region;recorded mainly in open forest (of Neofabricia,Thryptomene and Xanthostemon with emergentAraucaria cunninghamii), low closed forest, insemideciduous vine thicket, evergreen vine forest,and rainforest; recorded substrates are reddishsandy soil, and white sand; recorded landscapesare sand dunes, and a creekline in a narrow gorge.

ETYMOLOGY. — The specific epithet is derivedfrom the Latin propellere, propel, in reference tothe propeller-like form of the petals. The red,propeller-like petals immediately distinguish thisspecies from H. macilwraithensis and H. toze-rensis, which appear to be its closest relatives.

SPECIMENS EXAMINED. — AUSTRALIA: Webb &Tracey 13532, Queensland: N of Olive River mouth,12°07’S, 143°05’E (BRI); Webb & Tracey 13531,between Stoney Point and Mosquito Point, 12°25’S,143°16’E (BRI); Clarkson & Neldner 8808, TempleBay, 0-1 km W of Bolt Head, 12°15’S, 143°05’E(CANB, DNA, K, L, MBA, MEL, NY, PERTH,QRS); Forster 19355, same locality (BRI, MEL); Gray6853, same locality (CANB, QRS); Gray 6856, samelocality (CANB, QRS); Gray 6889, same locality(QRS); Sankowsky & Sankowsky 1144, same locality(BRI).

3. Hibiscus tozerensis Craven & B.E. Pfeil,nom. nov.

Macrostelia grandifolia Fryxell, Austral. J. Bot. 22: 189(1974), replaced synonym. — Type: Australia:Queensland, Cape York Peninsula, Tozer Range,N end, 29 June 1948, Brass 19353 (holo-, BRI n.v.;iso-, A n.v., CANB).

A new epithet is required as grandifolia is pre-empted in Hibiscus by H. grandifolius Hochst. exA.Rich.

Craven L.A. & Pfeil B.E.

238 ADANSONIA, sér. 3 • 2004 • 26 (2)

Australian Macrostelia transferred to Hibiscus

239ADANSONIA, sér. 3 • 2004 • 26 (2)

FIG. 2. — Holotype of Hibiscus propulsator Craven & B.E. Pfeil.

Acknowledgements

The directors and curators of the following herbariaare thanked for the opportunity of studying specimensin their care: BRI, CANB, P, QRS. Paul FORSTERkindly made special collections for our research; GarrySANKOWSKY provided living material for greenhousestudies; Alain CHANGY and Jean-Noël LABAT providedimages of types lodged in P; Laurent GAUTIER pro-vided information on type material lodged in G;Annette WILSON provided some bibliographic infor-mation; and Elizabeth MINCHIN gave advice on theformation of the epithet propulsator. The figures wereprepared by Carl DAVIES.

REFERENCES

BAILLON H.E. 1885. — Liste des plantes de Mada-gascar. Bull. Soc. Linn. Par. 1: 512.

FRYXELL P.A. 1974. — New species of Gossypium,Decaschistia, and Macrostelia (Malvaceae) fromAustralia. Austral. J. Bot. 22: 183-193.

FRYXELL P.A. 1999. — Pavonia Cavanilles (Mal-vaceae). New York Botanical Garden Press, NewYork (Flora Neotropica Monograph 76).

HOCHREUTINER B.P.G. 1952. — Macrostelia, unnouveau genre extraordinaire de Malvaces deMadagascar. Not. Syst. 14: 229-234.

HOCHREUTINER B.P.G. 1955. — Macrostelia: 118-121, in HUMBERT H. (ed.), Fl. Madagascar etComores 129. Typographie Firmin Didot, Paris.

PFEIL B.E., BRUBAKER C.L., CRAVEN L.A. & CRISPM.D. 2002. — Phylogeny of Hibiscus and the tribeHibisceae (Malvaceae) using chloroplast DNAsequences of ndhF and the rpl16 intron. Syst. Bot.27: 333-350.

PFEIL B.E., BRUBAKER C.L., CRAVEN L.A. & CRISPM.D. 2004. — Paralogy and orthology in the Mal-vaceae rpb2 gene family: investigation of gene dupli-cation in Hibiscus. Mol. Biol. Evol. 21: 1428-1437.

Submitted on 26 April 2004;accepted on 27 September 2004.

Craven L.A. & Pfeil B.E.

240 ADANSONIA, sér. 3 • 2004 • 26 (2)