Courtesy of Marcelle Sheppard

Canna cultivar groups

At Claines Canna Collection we have a collection with over 750 Canna accessions and we had a pressing need to have a method of categorising the collection. Therefore, we created a set of Canna cultivar groups, to satisfy our own requirement to produce manageable paper and web catalogue of this large collection. This article describes what we did, why we did it, and our own experiences.

The International Code of Nomenclature for Cultivated Plants (ICNCP) regulates the naming of cultivars and cultivar Groups, and we decided that this was a progressive way forward, rather than looking to the past and the categorisation methods that had already failed and been abandoned. There simply was no point in us trying to resurrect them. So out of the window went species-like names like hybrida, hortensis, generalis, orchiodes, most based on the genotype (the DNA makeup). Also, the recent attempt to call all cultivars x generalis is quite pointless, as just by enclosing the cultivar name with inverted commas means that it is a cultivar!

The modern Cultivar Groups are not concerned with the genotype of a plant, only what it looks like, i.e. the phenotype. We eventually decided to base the groups on Canna history, where different types and styles of cultivars were introduced at milestone points in time. The reasoning being that these differences are immediately visible to the experienced eye and meet our criteria of being sustainable and easy to use. So we specified the following groups:

Agriculture Group

 This group includes all Cannas that have been used in agriculture. Farmers have selectively bred Cannas over a period of thousands of years, by simply selecting which variants best produce what they want to grow. They probably gave the cultivars separate names, but the only name that has come down to us is C. achira. Achira is the generic name for Canna in the Andes region, where they were grown as a horticultural crop and not as ornamentals.

Normally these are C. discolor hybrids, and they have been raised for their ability to produce high quantities of starch. Triploids produce significantly more starch than diploids, and the research that has been done in the area of agricultural Cannas shows that they are nearly all triploids. Incidentally, they are nearly all seed sterile as well.

CannaQueensland Arrowroot

CannaQueensland Arrowroot’ has been an agricultural crop in Australia for two centuries, and it still continues to be grow there.

In his researches in Asia, Dr Tanaka has described several cultivar variants, and it is certain that many more agriculture group Cannas exist, but have not been brought into horticulture and catalogued, in other tropical and sub-tropical areas; India, Africa and Hawaii immediately spring to mind.

Most of the members of this group have fine large banana-like foliage and are also able to be described as members of the next group, the Foliage Group.

Foliage Group

Cultivars, F1 and F2 hybrids, normally with small species-like flowers, but grown principally for their fine, banana-like foliage. These cultivars were originally developed by Monsieur Théodore Année, the world’s first Canna hybridiser, who introduced many original cultivars and also acted as the inspiration for many more Canna (Foliage Group) introductions.

The hybridisers working on this group rapidly discovered a scientific fact, confirmed more recently in the 1960’s by Dr Khooshoo, that is, if the flower size is increased significantly, then the seedling will be less substantial than its seed parent. So, to produce more floriferous plants we have to accept that they will be smaller than the more architectural Foliage Group cultivars.

Crozy Group

CannaMadame CrozyPodgora Gardens

A cultivar group where the flower spikes are arranged close together on the stalk and have narrow to medium staminodes. There is always space between the staminodes when arranged formally, and the labellum (lip) is smaller than the staminodes, and is often twisted or curled.

CROZY Antoine

The pioneer of this group was Monsieur Antoine Crozy of Lyons, France, who started breeding Cannas as early as 1862, from stock originally developed by Monsieur Année, the world’s first Canna hybridiser, who introduced what we are now calling ‘Foliage Group’ specimens.

CannaAntoine Crozy

By careful selection, Crozy created hundreds of new floriferous cultivars and created a rage for new dwarf Cannas across the world. At this point, we have to put the word dwarf into perspective. They are smaller than most of the Foliage Group, but they were still mostly over 1 metre (3ft 3in) in height.

The cultivars in question were sometimes referred to as gladiolus flowering cannas, but describing flowers as similar to another genus is discouraged. In any event, they are gladiolus-style, not some sort of clone.

Also, in the past, they were sometimes called the “x generalis L.H. Bailey” garden species, although such “pretend” garden species are now deprecated in favour of Cultivar Groups. When first describing the generalis “pretend” species, Dr Bailey defined it as C. glauca L. x C. indica L. x C. iridiflora Ruiz and Pav. x C. warscwiczii A. Dietr. Incidentally, the International Code of Nomenclature for Cultivated Plants (1961) demanded that cultivars should be designated by their parents. The ICNCP eventually realised that this was unsustainable and revised their regulations to those we have today, and on which this effort is based.

Italian Group


A cultivar group with large, very fragile staminodes. Flowers are arranged somewhat loosely, in clusters with wide petals, so wide that there is no space between them when arranged formally. The labellum (lip) is larger, or at least as large, as the staminodes, unlike the other groups where it is smaller and sometimes curled. The stamen is also much wider than that in the other cultivar groups, so much so that early nurserymen referred to the stamen and anther as ‘inner petals’.

Canna flaccida

The Italian Group obtained its larger sized, but less sturdy, flowers from the introduction of Canna flaccida, crossed with C. ‘Madame Crozy’, in the early 1890’s by Dr Sprenger in Naples, Italy followed shortly afterwards by Luther Burbank in California, USA, with the same cross. Both used C. ‘Madame Crozy’ as the seed parent as it was considered to be the finest of the existing Crozy Group cultivars.


Specimens of this group used to be called the orchid flowering cannas, or x orchiodes L.H. Bailey, a garden species, which in the rules of the time he defined as being (C. glauca L. x C. indica L. x C. iridiflora Ruiz and Pav. x C. warscwiczii A. Dietr) x C. flaccida.

Those same rules also mean that when Sprenger crossed C. ‘Madame Crozy’ back with pollen from that specimen, then it needed at new garden species name, perhaps x generalis backcross L.H. Bailey, and its official definition was ((C. glauca L. x C. indica L. x C. iridiflora Ruiz and Pav. x C. warscwiczii A. Dietr) x (C. glauca L. x C. indica L. x C. iridiflora Ruiz and Pav. x C. warscwiczii A. Dietr.) x C. flaccida) ); really rolls of the tongue, doesn’t it?

It is easy to see why the ICNCP eventually changed its regulations to those we are describing here today. In any event, it is difficult to see the similarity between this group and orchids. Is it only me?

Variegated Group


This group covers all Cannas with variegated foliage, regardless of how the variegation has occurred. If the variegation is distinctive, then the cultivar is eligible for membership of this group.

CannaKansas City

The only natural and reproducible variegation in Cannas is the red/purple stripes on a green leaf that were introduced primarily through Canna indica f. warscwiczii.

CannaBengal Tiger

All other variegation is introduced through mutation of one sort or another, see C. ‘Kansas City’ (a natural sport) and C. ‘Bengal Tiger’ (radiation induced).  Such variegation, induced by mutation is not transmitted to seed offspring, and the only propagation method available is vegative, i.e. the division of rhizomes. Bear in mind that mutations can revert spontaneously, and then you are left with what is potentially another new cultivar.

Aquatic Group


This group is for cultivars that thrive as marginal water plants. Characteristically, they will have lance-shaped foliage and long, thin rhizomes that spread for long distances, and anchor the plant into the mud of the margins. Many cultivars that have the water species (C. glauca and C. flaccida) behind them, have these qualities and several breeders have introduced new aquatics in the last few decades.

Typically, aquatic cannas can be recognised both by the long, thin rhizomes but also by their lanceolate shaped foliage, some of which are coloured glaucous blue/green.

Conservatory Group

The growing conditions in a Conservatory are quite specialised and do not suit many cultivars, this group have been selected for thriving in this environment, required features being plant vigour, early flowering, foliar appearance, self-cleaning ability and good propagation qualities.

This cultivar group originated in Longwood Gardens, USA, who have vast areas of conservatories. Dr. Robert Armstrong, the newly appointed Geneticist, began a canna breeding program in 1967, when it was realized that cultivars available at that time were not suitable for use on display in the Conservatories. The new cultivars would also have to be suitable for growing in their immense garden complex.

New blood was required to create something new, and the project started with making crosses involving three cultivars, ‘Ambassadour’, ‘Moonlight’, and ‘Banner’, which were obtained from Yalta, Ukraine and South Africa.

As the project progressed, other cultivars and species were introduced from around the world. Seeds of five species were received from Lago Maggiore, Italy, Canna glauca was procured from Rio de Janeiro, Brazil, and in 1972, Dr. John Creech of the US National Arboretum, added three Russian cultivars, ‘K.A. Timirazov’, ‘Krimsky Riviera’, and ‘Soleznaya Krasavaya’. The scope of the project was extended to include new aquatic cultivars.


With the new additions to the breeding stock, red, orange and pink cannas were developed for the Conservatory along with the original white and yellow. Between 1972 and 1985, twelve cultivars were selected for release. Named for local historically significant places and events, they were:

The list above shows the original members of this group, but it will be extended as other cultivars prove to have the same qualities and capability of growing continuously 52 weeks of the year in the specialised conditions of a conservatory. If anbody has experiences of such growing conditions, then please let me know.

Miniature Group

CannaTom Thumb

This group covers cultivars growing under 0.5m (19″) in height, the flowers should be in scale to the rest of the plant. They should be suitable for being used as cut flowers for table decoration at formal dinners or in restaurants.

There has been interest in growing such small specimens in the past, the Seven Dwarves Series is one example, and the Alipore Canna Collection in India also had some success, referring to them as Pygmies. As that term is no longer considered politically correct, we coined the term miniature to describe these very special treasures.


CannaMadame Legris

The photographs above show typical diploid and tripoid cultivars, namely C. ‘Madame Legris’ and C. ‘Aida’.
The Crozy Group triploids have given us a slight problem, but not insurmountable. These are the Crozy Group cultivars who’s seedlings have changed from diploids to triploids. The outstanding example of this is C. ‘The President’, which produces spikes of large flowers, and many new Canna cultivars are of this type, including CannaAida’ (right).

Triplods have larger staminodes, and some even create that perfect circle where there are no gaps between the staminodes.  This also means that they are no longer complying fully with our definition for the Crozy Group, but they certainly are not Italian Group cultivars. We ignored this inconvenient fact and placed them in the Crozy Group, anyway. However, this is obviously not long-term correct, and we feel that we now need a new group to contain the Crozy Group cultivars which outgrow this group and become “super-crozy”.

Any respectable suggestions for a descriptive group name will be gratefully received.

The terms of reference for Cultivar Groups are rather loose, and if somebody wanted to start categorising cannas as Yellow Group or Red Group that would comply with the International Code of Nomenclature for Cultivated Plants, but would also seem to be a touch futile, and not provide the same terms of reference as the groups described above.

Other than that, the Cultivar Group categorisation has worked well for us and we have no regrets. It was a considerable effort to implement for hobby collectors, but we would commend our groupings to anyone else thinking along the same lines and please feel free to use what we have done.