Erect perennial herb or shrub, 2–6 m tall, with tuberous roots. Stems few, erect or drooping, ± woody, 35–100 mm diam., hollow, often unbranched or branched above at flowering, dying back to woody base in winter. Leaves opposite; lower cauline leaves shortly petiolate, 28–45 cm long, 2-pinnate with 3–7 primary pairs of pinnae; leaflets opposite, ovate, the base rounded or obtuse, serrate, acuminate, 9–11.5 cm long, 3.5–4.5 cm wide, with hairs on veins adaxially, and on lamina abaxially. Upper cauline leaves 1-pinnate or simple. Inflorescence a loose cyme of many capitula. Capitula large, 8–9 cm diam. (including rays), usually slightly nodding. Outer involucral bracts 5 or 6, lanceolate, narrow ovate to somewhat narrow fiddle-shaped, 10–12 mm long, glabrous, recurved at anthesis; innermost bracts broadly ovate, 20–22 mm long, many-lined, scarious. Ray florets 8 or 9; ligule 28–45 mm long, lilac/magenta. Disc florets numerous, golden yellow. Cypselas (not seen). Nov.-Mar. (based on two collections).
In Victoria, known from two recent collections (2020 and 2024) from beside the Barham River near Apollo Bay, where scattered for several 100 m in damp weedy remnant Eucalyptus forest.
Dahlia excelsa and D. imperialis Ortgies are treated as separate species by some authors, and the horticultural literature has long treated D. imperialis as the common Tree Dahlia (Lawson & Spencer 2002). Using the key in Sorenson (1969), the plants naturalised in Victoria key to D. excelsa based on having 3-7 primary pinnae whereas there are 9-15 primary pinnae in D. imperialis. Sorenson also used the shape of the outer involucral bracts as an additional character to separate the two species, with D. excelsa having “linear” bracts, and D. imperialis having “obovate or oblanceolate” bracts, however the interpretation of shape, and Sorenson’s concept of “linear” is difficult to know. The shape of the involucral bracts in the Barham River collections are lanceolate, narrow ovate to somewhat narrow fiddle-shaped. Sorenson notes that the only material of D. excelsa that he had available for his revision were the two photographs (housed at The Field Museum) of the type specimens; field searches by him for D. excelsa were unsuccessful. Sorenson also noted that it was likely that the type specimens of D. excelsa were collected from a botanic garden in Mexico City.
Sorenson’s recognition of D. excelsa as separate from D. imperialis was provisional, noting that that the two species are probably variants of the same species. D. imperialis is a later published name and is treated here as a synonym. However, the problem surrounding the identity and the place of origin of D. excelsa is recognised here as is the possibility that D. excelsa may not be the appropriate name for the Victorian plants.
The colour of the ligulate florets in plants cultivated in Victoria is mostly lilac/magenta but occasionally white or pink.
Bentham, G. in Maund, B. & Henslow, J.S. (1838). The Botanist 2: t. 88.
Lawson, L. & Spencer, R. (2002). Dahlia. In: Spencer, R.. Horticultural Flora of South-eastern Australia 4. Flowering plants. Dicotyledons. Part 3. The identification of garden and cultivated plants. University of New South Wales Press.
Sorenson, P.D. (1969). Revision of the the genus Dahlia (Compositae, Heliantheae — Coreopsidinae). Rhodora 71: 309–365, 367–416.
Webb, C.J., Sykes, W.R. & Garnock-Jones, P.J. (1988). Flora of New Zealand IV: Naturalised pteridophytes, gymnosperms, dicotyledons. Botany Division, D.S.I.R., Christchurch, New Zealand.
