The taxonomy of cannas

The term "species" refers to those types of plants that have evolved in nature and are discovered naturally occurring in the wild. When new types of plants have been bred by people, then these types are usually referred to as "cultivars", "hybrids, or "varieties" (though strictly speaking "variety" is a taxonomic term that means sub-species). The study and classification of species is known as "taxonomy".

To begin at the "family" level, consider the question "To what plant family do cannas belong?" It's rather a trick question that can leave some horticultural experts looking blank. They wonder if cannas are related to gladiolus, because the flowers are similar, and in a way so are the leaves. Or they may think they are related to bananas, or orchids. The answer, of course, is none of these. Cannas have their own plant family, Cannaceae.

It is quite unusual for a garden plant to have its own plant family. It shows that cannas are radically different to all other plants. What sets cannas apart from all other plants is their flower. Almost all plants have flowers which have an axis of symmetry. Canna flowers are different. They have no apparent axis of symmetry. Each “petal” (there are usually 4) is of a different size, and shape to the others, and each flower head is different from its neighbours. Also, what the gardener assumes to be petals are overgrown sterile stamens (staminodes). The true petals are insignificant and are reduced to the role of sepals.

At the higher level of classification, the family Cannaceae is a member of the order Zingiberales, which includes 7 other plant families: Zingiberaceae and Costaceae (gingers), Musaceae (bananas), Heliconiaceae, Maranthaceae, Strelitziaceae (bird-of-paradise plants) and Lowiaceae. Cannas share with these plants the elegant leaves that open by unfurling. The closest relation to cannas is believed by taxonomists to be the maranta family, some of which are popular pot plants known as "Prayer Plants".

At the lower level of classification, within the Cannaceae family, there is only one genus, Canna. So although all cannas are evolutionarily quite distant from all other plants, they are all quite similar to each other.

Below the level of genus, ie at the level of species, the picture gets more complicated. Historically, over 200 species have been described and named. Most of these names are no longer valid according to current taxonomy. Yet some continue to be commonly found in horticultural literature and plant catalogues. Such common non-valid names include C. edulis, C. brasiliensis, C. lutea, C. limbata (the first 3 of which, incidentally, are currently listed in the RHS Plant Finder).

Fortunately, in recent years, the genus Canna has been the subject of 2 monographs which has imposed order where previously there was confusion. These are by N. Tanaka (2001), and H. Maas-van de Kamer (2008).

Tanaka initially reduced Cannaceae to 19 species plus 7 sub-species. This list includes 4 new species discovered by Tanaka and Koyama in northern Argentina in the 1980s (C. jacobiniflora, C. amabilis, C. plurituberosa, C. stenantha).

Maas reduced Cannaceae even further, to 10 species and no varieties. She claimed that many of the wild cannas listed as species by Tanaka were no more than different versions of C. indica.

When two taxonomists disagree, there is no way that we, as non experts, can decide between them. Until the botanical world sorts out its differences, you and I simply take our pick on who to follow, or we can choose names from either expert. However, what I think would be wrong would be for us to continue to use names that have been ruled as invalid by both these experts.

Since the publication of the monographs by Tanaka and Maas, the picture has changed even further. Tanaka has discovered yet another new species (in Honduras) which he named C. tulianensis. Also, an Argentine botanist, Ciciarelli, has more recently claimed to have discovered 3 more species (in Argentina) which she has named C. ascendens, C. variegatifolia, C. fuchsina.

The current list of species and sub-species is shown in the table below. The monographs of Tanaka and Maas are very bulky documents, some 70 pages each. They make very interesting reading. You may find the Maas monograph as a PDF file on-line. So far as I am aware the Tanaka monograph is only available in printed form. Some of the Ciciarelli monographs can be found by a Google search.

Taxonomy of cannas according to Tanaka and Maas:

According to Tanaka 2001

According to Maas 2008

C. paniculata

C. paniculata

C. discolor

C. compacta

C. iridiflora

C. indica

C. indica

C. indica ssp flava

C. indica ssp maculata

C. indica ssp warszewiczii

C. indica ssp sanctae-rosae

C. bangii

C. bangii

C. coccinea

C. jaegeriana

C. jaegeriana

C. latifolia

C. turckheimii

C. liliiflora

C. liliiflora

C. pedunculata

C. pedunculata

C. flaccida

C. flaccida

C. glauca

C. glauca

C. amabilis

C. speciosa

C. jacobiniflora

C. stenantha

C. patens

C. plurituberosa

Taxonomy unconfirmed

C. altensteinii

Note: The species which Tanaka calls C. latifolia, Maas calls C. turckheimii.

Recent Discoveries:

C. tulianensis (Tanaka)

C. ascendens (Ciciarelli)

C. fuchsina (Ciciarelli)

C. variegatifolia (Ciciarelli)

Distribution of Species

Canna are indigenous to the tropical and sub-tropical Americas, from the southern USA to northern Argentina and Chile, including the West Indies. C. indica and C. glauca are common throughout this range. Other species have more restricted distribution, for example, C. flaccida is restricted to the southern edge of the USA, C. iridiflora to parts of Peru, C. liliiflora to parts of Bolivia, etc. The Maas paper contains useful maps to show the distribution of the various species.

Habtat Preferences

Within their ranges, cannas occupy various ecological niches. The C. indica complex is cosmopolitan; C. flaccida, C. glauca, C. amabilis and C. stenantha are found in wetlands; C. iridiflora, C. liliiflora, C. jaegeriana, C. bangii in highlands over 2000m. Some species are forest plants, others are often found as roadside plants or weeds of cultivated ground.

Horticultural qualities of canna species

Some canna species have potential as garden plants. This is an area of horticulture which has been neglected in the past but which is becoming more of interest.

Most canna species are quite unknown in the world of horticulture, and some are rare or unknown even in botanical collections. Such obscure species offer tantalising prospects for horticultural application and hybridising.

All cannas have elegant foliage, and in some species the foliage is particularly impressive or attractive. Some species also have relatively large and/or attractive flowers, such as C. glauca and C. flaccida (although C. flaccida does not flower well in the UK climate).

As the species normally produce seed (except for C. discolor), and since canna virus is not carried in seed, cultivating species from seed has the potential for producing interesting, unusual and healthy plants at relatively low cost.

Some species, for example C. warszewiczii and C. patens, produce copious amounts of seed. They must be hand pollinated because we don't have humming birds in the UK. The seed can be saved indefinitely, and seed sown early in the year will produce plants that flower the same year. It is indeeed a better way of growing these species than from rhizomes.

Two canna species that have large and apparently beautiful flowers are not known in cultivation. These are “holy grail” of the canna world, and my fervent wish is that some time I will find them. One is C. iridiflora which grows in Peru, at altitudes of 2000m or more, and is found from Cuzco to Huanuco. The other is C. liliiflora which grows in the mountains of Bolivia, also at altitudes of around 2000m. It is found in the Zongo Valley north of La Paz, and in the mountains north of Cochabamba. In addition to having large white scented flowers, it is also a plant that grows 6m tall! Anyone visiting those parts might keep a look out for these rareties. Collecting seeds or plants in the wild is illegal. But in practice the local people often grow local wild flowers in their gardens, and for a smile or a few coins, a visiting tourist might acquire a few seeds that would make a canna enthusiast very happy!