An earlier version of this article was published in Hibiscus International Vol. 4, No. 1 (alternatively issue 18) (Jan-Mar 2004), the e-journal of the International Hibiscus Society
As originally conceived by Linnaeus [1, 2] the genus Hibiscus encompassed all the capsular-fruited “mallows”, except for the cottons. Since his day the various species have become better known, and the number of species known has increased greatly. The genus has been divided into sections [3], and groups of species with distinctive flower or fruit characteristics have been separated out (segregated) as new genera. Some plants related to Hibiscus, and discovered in subsequent years, have been placed in separate genera from the beginning. Other plants, originally placed in Hibiscus, have turned out to belong elsewhere; for example, Hibiscus populneus is now Thespesia populnea, and belongs in the group of genera clustered around Gossypium (cotton).
The last complete treatment of Hibiscus was by Hochreutiner [4], and is over a century old. This covered 197 species [6]. Current estimates of the number of species known vary from about 200 [5] to in excess of 300 [6, 20]. There are as many more again in related genera.
Botanists have differently divided Hibiscus into sections and segregate genera. In the absence of a recent treatment, there is no division of the genus into sections which commands universal acceptance. The recognised sections vary in nature, from diverse groups like Ketmia, to species-rich groups like Furcaria, to small closely knit groups such as Muenchhusia (the North American rose-mallows). Prior to the application of DNA sequencing techniques to the taxonomy of Hibiscus and allies I expected that a few genera would be found to be nested in Hibiscus, and this would be resolved by removing parts of Hibiscus so that each genus has a common ancestor to the exclusion of other genera. Instead the work of Pfeil et al [6] found that some segregate genera were not only nested within Hibiscus, but nested within particular sections of Hibiscus. Additionally species from the tribes Kydieae, Decaschistieae and Malvavisceae (Ureneae) were also found to be deeply nested in Hibiscus (deeply so in the latter two cases) suggesting that the whole of these tribes are descended from ancestors which would be placed in Hibiscus. Subsequent work [24] has found a few stray species of Hibiscus within a clade principally composed of endemic genera from Madagascar. These results mean that the appropriate modifications to the classification of hibiscuses and allied plants are not obvious; merging them all into one genus would produce a genus with well in excess of 500 species, with few or no diagnostic characteristics other than DNA sequence traits; while retaining the currently recognised genera would lead to breaking up Hibiscus into many genera, with a significant number lacking morphological distinctiveness. There is as yet insufficient sequence data on which to base a complete treatment of Hibiscus and allies; not only have not all species been included within the various studies, but also not all sections and segregate and allied genera have been sampled, so even if we were to boldly assume that all sections and genera are correctly identified as natural groups and none are nested one within another, we are lacking information on some relationships. The general outline can however be perceived, and it is described by Pfeil & Crisp [20]
Other segregate genera, such as Lagunaria, Alyogyne and Radyera, have been found to be distantly related to Hibiscus [6].
The Hibiscus sensu lato clade can be characterised by the presence of 5-celled ovaries and fruits (6–10-celled in Decaschistia, 10-celled in Cenocentrum, and 2–3-celled in Kydia and allies), the fruits usually being capsules which split lengthwise along the middle of the cells, or schizocarps; by the presence of reduced staminodes in the form of an antherless, often toothed, apex to the staminal column (absent in Papuodendron); and by the presence of 5 or 10 style-branches (6–10 in Decaschistia, 10 in Cenocentrum, and 2–3 in Kydia and allies) each terminated by globe-like stigmas (peltate in Julostylis, discoid in Cenocentrum). When the fruit is a schizocarp each cell is single-seeded.
Within the data of Pfeil et al [6] 6 groups can be recognised within Hibiscus. Later work identifies the tribe Kydieae as a 7th group [20], and the unpublished work of Baum and Koopman, cited by Pfeil & Crisp [20] identifies a 8th group, which falls into the “Malagasy“ clade mentioned above. There follows an overview of the sections of Hibiscus, and related genera, based on these 8 groups. Note that further research may yet identify additional groups among unsampled species and sections, and among segregate and allied genera. Sections Ketmia, Bombicella and Solandra are not natural groups, and would be revised in a full treatment of Hibiscus. In particular the 4 sampled species of Ketmia fell into 3 different groups, and I don't know where the main body of Ketmia, if there is one, would fall.
Group 1: This group contains a single species, which is part of section Ketmia. This is Hibiscus pentaphyllus (Hibiscus caesius) (image gallery), from southern and eastern Africa, India, and northern Australia. Hibiscus pentaphyllus has long flower-stalks, large flowers with an flat epicalyx of 8 long narrow bracteoles, and a spreading calyx composed of long acuminate sepals, and deeply 3–5-lobed leaves [7].
Group 2: This group consists of section Calyphylli (image gallery) plus the species Hibiscus dongolensis, which is currently placed in section Ketmia. The raw DNA sequence data only weakly associates Hibiscus dongolensis with section Calyphylli, but the pattern of insertions and deletions in the sequences suggest a common ancestry, whether a single polytypic species or a reticulate hybrid grouping.
These species are found in eastern and southern Africa and in Madagascar. Hibiscus calyphyllus (also known as Hibiscus calycinus and Hibiscus rockii) is naturalised in Hawaii. The species of section Calyphylli have cordate, ovate, toothed, leaves, which are unlobed, or weakly-lobed. The flowers are borne on short stalks. Pfeil et al [6] sampled two species of section Calyphylli (Hibiscus calyphyllus, Hibiscus ludwigii), but I suspect from other sources [7, 8, 9, 10] that Hibiscus macranthus, Hibiscus lunariifolius and Hibiscus platycalyx also belong here.
Group 3: This group is a moderately species rich and diverse group consisting of section Bombicella, and nested within it, sections Hibiscus, Lilibiscus, Spatula and Trichospermum s.s., and also the Hibiscus tozerensis (formerly Macrostelia grandiflora) complex, the Malagasy species of Macrostelia, Hibiscus cravenii (formerly Alyogyne cravenii), and the Hibiscus lobatus group within section Solandra.
Section Bombicella (image gallery) has a pan-tropical distribution. There are over 70 species, including Hibiscus cardiophyllus of North America, Hibiscus phoeniceus of the Caribbean and Latin America, Hibiscus micranthus of East Africa, Hibiscus pedunculatus of South Africa, Hibiscus hirtus of India and Hibiscus sturtii of Australia. They are generally found in dryer regions, but Hibiscus peralbus of the Kimberley monsoon forests of Australia and Hibiscus insularis of Philip Island are found in wetter regions.
Section Hibiscus (image gallery) contains the hardy shrub hibiscuses (known as Rose of Sharon in North America). These are three species, Hibiscus paramutabilis, Hibiscus syriacus and Hibiscus sinosyriacus, from Korea and northern China. There are about 100 cultivars in all [11]. The flowers are medium sized. The leaves are of uniform shape.
Section Lilibiscus (image gallery) contains the cultivated tropical hibiscuses, including Hibiscus rosa-sinensis (the Rose of China). There are about a dozen species, found in humid parts of the Indo-Pacific region from the coast of East Africa to Hawaii. Being extensively cultivated and hybridised the natural distribution and validity of the various species is uncertain. The flowers are generally large and somewhat asymmetrical. The epicalyx typically has around 6-9 narrowly triangular epicalyx bracts. The leaves are usually of uniform shape. Hibiscus schizopetalus of East Africa is distinctive in its split petals and small (almost unnoticeable) epicalyx.
Section Spatula (image gallery) is named for its spoon-shaped bracteoles. It includes the Australian Hibiscus normanii. The South American Hibiscus spathulatus also has spoon-shaped bracteoles [12].
Section Trichospermum s.s. includes (at least) the African Hibiscus aethiopicus and Hibiscus malacospermus. (The broader section Trichospermum s.l. includes section Panduriformes, which falls in group 4, and section Aristovalves, which is unsampled.) Hibiscus aethiopicus is a low growing perennial herb with long elliptic-oblong, rarely 3-lobed, leaves and solitary yellow flowers.
Hibiscus cravenii (formerly Alyogyne craveni) is a localised species from northern Australia. It was found by Pfeil et al [6, 21] to be misplaced in the genus Alyogyne and to agree with section Bombicella in both DNA and morphological traits.
The Hibiscus tozerensis complex (image gallery) is a group of 3 taxa, formerly placed in Macrostelia, from northern Queensland [22]. These were found by Pfeil et al [6] to be closely related to Alyogyne cravenii and some species of section Bombicella. The petals of these species have long claws which are fused to the staminal column for half their length. The morphological evidence [13] for associating the Hibiscus tozerensis complex with the 3 Malagasy species of Macrostelia consists of more than the long petal claws and is convincing, but the distribution casts doubt on this. Unpublished sequence data indicates that the Malagasy species of Macrostelia fall into group 3, but are not the immediate relatives of the Hibiscus tozerensis complex [22].
Hibiscus lobatus is an annual herb from the Old World tropics. Its flowers lack epicalyces, and have white or yellow petals [25].
Group 4: This is the largest and most diverse group. The work of Pfeil et al [6, 19] and Small [14] shows it to contain elements of sections Ketmia and Solandra, sections Striati, Trionum, Panduriformes, Muenchhusia and Venusti; the segregate genera Abelmoschus, Fioria and Kosteletzkya (section Kosteletzkya only), and the genus Senra Additionally 4 species of the former tribe Malvavisceae (or Ureneae), but not Urena lobata, fall in here, from which we are led to the tentative conclusion that the whole of that 'tribe' (except Urena) belongs to this group, and close to Fioria.
The group is predominantly composed of herbs and small shrubs. The sole species of section Solandra sampled by Pfeil et al [6], the African Hibiscus schinzii, was found by them to be closely related to two African species currently placed in section Ketmia, these being Hibiscus engleri and Hibiscus physaloides. Hibiscus schinzii is an annual herb, whose flowers lack an epicalyx [25]. The Indian Hibiscus solandra usually lacks bracteoles [15].
Section Striati (image gallery) is a small group of American species, including Hibiscus striatus and Hibiscus trilobus.
Section Trionum (image gallery) may be a monotypic section. Its titular species, Hibiscus trionum, is an annual herb, native to the Old World tropics, but widely naturalised elsewhere.. The lower leaves are rounded; the upper leaves are deeply 3–5-lobed. In fruit the calyx has a distinctive inflated appearance, which is shared with sections Clypeati and Striati.
The titular species of section Panduriformes (image gallery), Hibiscus panduriformis has large yellow flowers, sometimes with spoon-shaped bracteoles (with longer 'handles' than in section Spatula) and scarcely divided or undivided, rounded, leaves. It is found in Africa, the Mascarenes, India and Australia.
Section Muenchhusia (image gallery) is a close group of 5 species, the American Rose-Mallows, mostly from the eastern USA, but extending into Canada's Ontario Peninsula, and to California in the western USA. These are large-flowered perennial herbs. They, and especially their hybrids and cultivars, are grown as ornamentals.
Section Venusti (image gallery) is a group of a half a dozen or so species of shrubby hibiscuses from east Asia. The best known is Hibiscus mutabilis, grown in North America as the Confederate Rose.
Genus (or section) Abelmoschus (image gallery) is a group of about 15 species from the Old World tropics. They have elongated fruits. The calyx partially encloses the corolla in flower, but splits asymmetrically to reveal the petals. The calyx, and usually the epicalyx, are caducuous, i.e. are not retained with the fruit. Okra, Abelmoschus esculentus, is the most important food crop within the mallow family. Other species are also cultivated as fruits, vegetables or for fibre.
Genus Fioria (section Pterocarpus) (image gallery) has winged fruits. This genus is exemplified by Fioria vitifolia, which has distinctive vine-shaped leaves. There is disagreement as to whether there is just the one, variable, species, or whether there are several species of the genus.
Genus Kosteletzkya (section Pentaspermum) (image gallery) approaches Fioria in possessing winged or less commonly angled fruits, and approaches Malvavisceae in the cells of the fruit being single-seeded. It consists of 17 species [5], mostly from tropical America, and to a lesser extent, Africa, but also with species in the eastern USA, southern Europe, south west Asia, and Malesia. However some species of Kosteletzkya fall into the 8th group.
Genus Senra is a genus of perhaps 3 species from East Africa, Socotra, Arabia and Pakistan. It also approaches Malvavisceae in being single-seeded, but has two ovules in each ovary locule, only one of which develops into a seed. The stigmas may tend towards the typical capitate form, or may approach an obliquely truncate shape.
The old Malvavisceae is characterised by fruits being 5-celled schizocarps or berries, each cell being single-seeded, and the flower having 10 style branches, i.e. 2 for each ovary-cell. It contains as many species (about 300) as the rest of Hibiscus sensu lato. Kubitzki & Bayer [5] recognises 8 genera here; Urena, Malachra, Peltaea, Phragmocarpidium, Rojasimalva, Pavonia, Malvaviscus and Anotea - but Urena falls into group 5. Pavonia is by far the most diverse and species-rich of these genera. This group (excluding Urena) is mostly found in the Americas, but Malachra extends to the Old World Tropics and some 25 species of Pavonia are found in the Old World Tropics, including 2 species in Australia. Malachra has bracts enclosing inflorescences, but lacks epicalyces. Peltaea (image gallery) has both bracts and epicalyces. Rojasimalva is similar to Peltaea, but has mericarps with a spine longer than the body. Phragmocarpidium has 2-celled mericarps, the lower cell being empty. Malvaviscus (image gallery) has red or white berries, and red flowers with basally auriculate petals. Anotea has blue-black berries and yellow flowers. Pavonia (image gallery) contains the species with none of these specialisations, and may well not be monophyletic. Groups within Pavonia have been given generic rank, such as Triplochlamys (with a double, petaloid, epicalyx), Goethea (with flowers borne on or close to the branches), and Blanchestiastrum and Codonochlamys (with fewer bracteoles) [16].
Group 5: Groups 5, 6 and 7 from a clade, characterised by abaxial foliar nectaries, to which Pfeil & Crisp [20] give the name Phylloglandula. Group 5 is species-rich, but relatively stereotyped. It consists of section Furcaria and the genera Decaschistia, Cenocentrum and Urena
Section Furcaria (image gallery) consists of comfortably over 100 species [17] of erect-, often prickly-stemmed herbs and shrubs found throughout the tropics. Nectaries are borne on veins on the calyx lobes and the lower portions of the leaves. The lower leaves are typically lobed, often deeply so, and the upper leaves less lobed or unlobed. The bracteoles of the epicalyx are usually forked or with a distinct terminal appendage. There are 10 diploid species (9 African, 1 American), and many polyploid species of hybrid origin, involving those 10, and other, now extinct, diploid species.
Decaschistia is a genus of approaching a score of species from India, South East Asia, southern China and northern Australia. They are not known from Indonesia. They are distinguished by an increase in the number of ovary and fruit cells and the number of style arms [5, 15]. Pfeil et al [6, 19] found that two sampled Australian species were jointly nested in section Furcaria, but a later analysis placed the genus as sister to section Furcaria. The similarities between the Asian and Australian species are sufficient [13] to tentatively include the whole genus within group 5.
Cenocentrum is a monotypic genus from Indochina and southern China. It is unsampled, but is morphologically very similar to Decaschistia, and falls within the area of distribution of that genus. The stigmas are discoid, rather than capitate.
Urena (image gallery) is a pantropical genus of 6 to 8 species of shrubs. It is convergent with Malvavisceae, possessing a fruit of 5 single-seeded mericarps, an ovary of 5 locules, and 10 style arms and stigmas. The American species have spiny, burr-like, fruits, armed with glochidate (hooked) spines. Like section Furcaria it has abaxial leaf nectaries.
Group 6: This group consists of the genus Talipariti [18] (image gallery), recruited out of section Azanzae, and the genus Papuodendron, which was originally placed in tribe Durioneae and family Bombacaceae. Historically some species of Thespesia were misplaced in section Azanzae.
This group is the sister group to group 5. It shares with it the characters of leaf and calyx nectaries. It differs in being composed mostly of arborescent species, the leaves being usually unlobed, and of uniform shape on a single plant.
Talipariti is a genus of 22 species with a distribution centred on south east Asia, but also found in tropical America, tropical Africa, India, Japan and Pacific Islands. They have a number of distinguishing traits, including the presence of stem-clasping stipules. Talipariti tiliaceum is widely distributed along tropical shores as a mangrove associate and is also grown as an ornamental. Talipariti elatum is grown in a small way as a timber tree.
Papuodendron is a genus of 2 species of tree from New Guinea. DNA sequence data places them within Talipariti [19]. They differ from Talipariti in the absence of stem-clasping stipules, in the surfaces of the plant being scaly rather than (usually) hairy, in possessing pinnately-veined leaves, and in lacking staminodes.
Group 7 (“Kydieae“): This group consists of the genera Kydia, Julostylis, Dicellostyles and Nayariphyton from tropical Asia, with a total of 9 species. These are trees. These genera share the trait of having two or three (not five) ovary locules, and two ovules placed side-by-side (collateral) in each locule, rather than vertically as is normal in Malvoideae. (Are the ovary locules homologous to pairs of locules in Hibiscus?). Kydia produces two seeds in each locule; the remaining genera have single-seeded locules. Julostylis has a flattened (peltate) stigma; those of Dicellostyles and Kydia are large (and globose).
Group 8 (The “Malagasy” clade): DNA data places this group as the sister group to the previous 7 groups, but not unambiguously so. It is composed of the Malagasy endemic genera Humbertiella, Megistostegium and Perrierophytum, and some Malagasy species from section Azanzae and from section Azanzoides of the segregate genus Kosteletzkya. Pfeil & Crisp [20] propose to exclude this group from Hibiscus.
Humbertiella (image gallery) is a genus of 6 species from south west Madagascar
Megistostegium (image gallery) is a genus of 3 species from Madagascar, with a distinctive, more or less petaloid, 4-lobed epicalyx.
Perrierophytum is a genus of 9 species from west Madagascar.
All the above 3 genera have capsules with a single seed in each cell. In Megistostegium this results from the development of only one of the two ovules present in each ovary cell; in the other two genera each ovary cell has has only a single ovule.
Wercklea is a genus of 13 species of herbs, shrubs and trees from Central America, the West Indies, and north west South America, with very large, somewhat orbicular leaves, and stem-clasping stipules. It had been speculated [18] that the segregate genus Wercklea might be closely allied to Talipariti, which also has stem-clasping stiples, but unpublished data [24] places it with this clade.
Other sections and species: Apart from the sections listed above there are also at least sections Aristivalvus, Clypeati, Columnaris, Bombycidendron, Parapavonia, Pentacalycinus, and Trionastrum, for each of which I am aware of a single species. (However, I have noted some 50 species for which I have seen either no sectional affiliation, or have only seen a sectional affiliation in a work so old that the affiliation can't be taken as correct for a current classification.) Furthermore, some species were placed in section Azanzae, but were not transferred to genus Talipariti [18]. Due to the rules of biological nomenclature they cannot be classified in section Azanzae, and may be currently in a classificational limbo.
Hibiscus sororius, the type species of section Trionastrum, has bracteoles with a kidney-shaped appendage at the apex.
Other genera: Other genera related to Hibiscus are the Hawaiian Hibiscadelphus, the east African Symphyochlamys, and the Malagasy Helicteropsis, Humbertianthus and Jumelleanthus. These have not been studied by DNA sequencing. Given the results obtained with other genera in Hibisceae we have to consider the possibility that any of these genera might be nested within Hibiscus. In particular Hibiscadelphus is in some ways reminiscent of section Lilibiscus.
© 2005, 2009 Stewart R. Hinsley