These notes summarise my conclusions on some confused names found in the European horticultural trade, and are based on my published paper, Barnes (1998) detailed in the list of references. There are only sixteen recognised species of Filipendula (formerly included in the genus Spiraea), so it is remarkable that great confusion has arisen from the persistent misapplication of just one name - Filipendula palmata (Pall.) Maxim.. This species is very rarely to be found in gardens, yet the name is commonly listed by nurseries and garden centres, almost invariably applied to other, quite distinct, species. In fact the name has been used for five distinct species!
Most often, plants grown (and illustrated, vide Bot. Mag. t.5726) under the name prove to be the probable hybrid, F. × purpurea, which originates in Japan, whereas F. palmata is confined to China, Korea, Sakhalin, Kamtchatka and Manchuria. This appears to be true also of cultivars grown as F. palmata 'Elegans', 'Elegantissima', 'Alba' and 'Rubra', all of which I believe should be referred to F. × purpurea.
One attractive plant has been grown in British gardens at least since 1914, when it was described by the famous English gardener and plantsman, Edward Bowles. This is a superb plant, usually less than 30cm tall, with beautifully-cut leaves and fluffy heads of deep pink flowers. Bowles was uncertain what to call it and tentatively referred to it as "Spiraea digitata or lobata". The first is a synonym of F. palmata; the second, of the North American F. rubra. Reginald Farrer, noted for his graphic descriptive powers, thought it was F. palmata. However, his description of the flowers as "pure raspberry-and-red-currant-tart-juice colour" makes it clear that he was referring to F. × purpurea and not to the white-flowered F. palmata (Pall.) Maxim..
Now F. × purpurea is closely related to, or more probably a hybrid of F. multijuga, and it is to that species that I believe these dwarf plants should be referred.
Filipendula multijuga (right, basal leaf) grows wild in central Honshu, Japan, at altitudes over 1000m. It is variable in stature but generally smaller than other species, ranging between 20 and 100cm in height. Its most distinctive characteristics are the hairless, pinnate basal leaves, which have several pairs of well-developed lateral leaflets; the petals, which have finely-toothed (erose) margins and no stalk or claw at the base.
Filipendula palmata (left, upper part of basal leaf) is native to northern China, Korea, Kamtchatka and Sakhalin. It is often a robust plant up to 1.8m tall, its basal leaves with only one pair (occasionally two pairs) of strongly-lobed lateral leaflet, and a 7- or 9-lobed terminal leaflet. The white flowers have distinctly clawed or stalked petals with smooth margins. Also the leaves are usually finely woolly or white-hairy beneath. There really is no excuse for confusing this species with any other, and especially not the deep pink-flowered F. × purpurea which is the most common impostor.
I referred above to the cultivars 'Elegans' and 'Elegantissima', often listed under F. palmata but actually belonging to F. × purpurea. Some books list both names, but my studies indicate that they are the same thing, and that F. × purpurea 'Elegans' is the correct name. Since F. palmata has white flowers, the names F. palmata 'Alba' and f. albiflora would seem to be redundant. However, this is another instance of confusion between that species and F. × purpurea, and I believe that plants so labelled are likely to be F. × purpurea f. albiflora. Likewise, the epithet 'Rubra' seems improbable in a white-flowered species (so far as I can establish, there is no variation in flower colour in wild populations). The one illustration I have seen without doubt shows Filipendula purpurea, once again.
One final puzzle is the correct spelling of the specific epithet of Filipendula camtschatica / kamtschatica. The great majority of references, both old and more recent, spell it 'kamtschatica'. The name was first published in this genus by Maximowicz, who was transferring the epithet used by Pallas when he described Spiraea camtschatica, but altering the spelling to a 'k'. The rule of priority determines that the original spelling of the epithet is to be retained, so this fine plant should be known as Filipendula camtschatica.
F. occidentalis (S. Watson) Howell: NW USA
F. camtschatica (Pall.) Maxim.: Kamtchatka, Kuril Is., N. Japan.
F. glaberrima Nakai: Korea
F. yezoensis Hara: Korea, N. Japan
F. × purpurea Maxim.: Japan (cultivated)
F. auriculata Maxim.: western Japan
F. multijuga Maxim. south & west Japan
F. palmata (Pall.) Maxim.: northern China; Mongolia, North Korea, Siberia, Manchuria, Sakhalin
F. rubra (Hill) B.L. Rob.: eastern North America
F. angustiloba (Turcz.) Maxim.: northern Asia: Amur, Ussuri, Manchuria
F. vestita (Wall,) Maxim.: Afghanistan, Himalaya
F. kiraishiensis Hayata: Taiwan
F. tsuguwoi Ohwi: southern Japan
F. formosa Nakai: South Korea
F. ulmaria (L.) Maxim.: Europe, West & Central Asia, Mongolia
F. vulgaris Moench: Europe, North Africa, Asia Minor, Siberia
Barnes, P.G. (1998) Confusion in cultivated meadowsweets, Filipendula
Miller The New Plantsman 5(3):145-153
Bowles, E.A. (1914) My Garden in Summer. T.C. & E.C. Jack, London
Curtis's Botanical Magazine, t.5726 (1868) "Spiraea palmata" [=Filipendula × purpurea]
Farrer, R. (1919) The English Rock Garden. T.C. & E.C. Jack, London