Ceratostigma Bunge

More than herbs, less than shrubs, Ceratostigma species position themselves at the margin of woodiness. While some behave as shrubs in the wild, they may become subshrubby or effectively herbaceous when grown in colder gardens, so we include even the least woody species here to aid comparison. Quickly making garden impact, easily moved and propagated, several species became garden standards within a few decades of their introduction. Some can develop red tints towards the end of the season, but their flowers, usefully late in summer and into autumn, are the major attraction. In the nurseryman’s truism, ‘people like blue flowers’.

A member of the Plumbaginaceae, Ceratostigma belongs (with Plumbago and some small genera of little horticultural significance) to Subfamily Plumbaginoideae, an uncontroversial grouping of both woody and herbaceous plants well supported by morphological, biochemical and molecular evidence (Hernández‐Ledesma et al. 2015; Koutroumpa et al. 2018). Ceratostigma is distinguished by the combination of a non‐glandular, tubular, 5‐ribbed calyx (10-nerved at the base), and glabrous style. In practice though, anybody who knows one Ceratostigma is unlikely to mistake another for anything else.

The genus was first described scientifically (Bunge 1833) by the Russian botanist Alexander von Bunge, from his own specimen of C. plumbaginoides collected near Beijing at the end of a marathon collecting trip through Siberia, Mongolia and northern China. The name (from the Greek keras, a horn) refers to the glandular protruberances on the stigmas of most species. The later synonym Valoradia Hochst. is rarely used today, but the horticultural habit of informally referring to them as plumbagos sometimes slips into their being listed under Plumbago in the garden literature. The nearest we have to a modern revision is David Prain’s (1906) brief but workmanlike herbarium-based study, written before C. willmottianum had been recognized. The Flora of China treatment (Peng & Kamelin 1996) covering 5 species, gives descriptions which are in places immensely detailed, sometimes at odds with previous accounts, and occasionally demonstrably over-precise. This injects a degree of uncertainty, possibly warranted, into our concept of some species.

Of the seven species, C. abyssinicum (marginally hardy in our area) is East African; the remainder are Asiatic, with ranges not or scarcely overlapping. Three apparently closely related species are found in western China and the eastern Himalaya. Two are easily distinguished, the more shrubby, more western C. griffithii and the subshrubby C. willmottianum; between them geographically and morphologically sits the less well understood C. minus. The strongly rhizomatous, essentially herbaceous C. plumbaginoides is native to north eastern China. Two further species are at most barely in cultivation in our area, and are not discussed in full here. C. asperrimum Stapf ex Prain has a southerly range in Thailand and Myanmar. It differs from C. plumbaginoides in both leaf surfaces being roughly hairy, and is unlikely to prove hardy in our area. Collections from northern Thailand (BSWJ 7260 & 7274) have been offered by Crûg Farm Plants, UK but not tested outdoors (Crûg Farm Plants 2021). C. ulicinum Prain is a small-leaved, deciduous shrublet from high altitudes in Nepal and southern Xizang, probably cold-hardy, and potentially of interest to alpine-house growers, but apparently not in cultivation.

The flat-faced blue flowers, producing nectar at the bottom of a long, narrow tube, suggest pollination by long-tongued insects, perhaps butterflies or day-flying moths. There have been no scientific studies of Ceratostigma pollination, although Gao et al. (2021) imply that hawk moths may pollinate C. willmottianum. Heterostyly is very common in the Plumbaginaceae, but only well studied in Subfamily Limonioideae (Limonium, Armeria etc. – Baker 1966; Costa et al. 2019). This is the occurrence of two or more floral morphs with different style lengths, and usually with anthers positioned to match the style length of another morph, an arrangement encouraging outcrossing, familiar to gardeners in the ‘pin’ and ‘thrum’ morphs of most Primula and Pulmonaria. Morphs are genetically determined, each individual producing flowers of only one type. Two morphs have been demonstrated in C. plumbaginoides (Peng & Kamelin 1996), C. minus (Costa et al. 2019), and very thoroughly investigated in a wild population of C. willmottianum (Gao et al. 2021). In the last species, it is coupled with partial self-incompatibility, and different but not fully reciprocal anther positions. Significantly for gardeners, the short-styled morph had longer corolla lobes, making a broader-faced flower, perhaps because the hidden stigma picks up pollen from insects less readily, warranting a greater investment in attractants (Gao et al. 2021).

As plants of open habitats, Ceratostigma species suit sunny places in the garden; reasonably well drained soil is advisable, especially where hardiness is in question. Poor, dry soils are generally tolerated. While the very hardy C. plumbaginoides routinely dies back to woody stem bases at ground level, the western Chinese species are quite variably shrubby according to the severity of the winter. Even when woody stems survive above ground, gardeners often trim them back to near the base in early spring to encourage luxuriant growth and a better show of flowers (Huxley et al. 1992; Havlis 2021). Given warmth and adequate summer moisture they can put on prodigious annual growth. The more tender C. abyssinicum is in the Californian nursery trade, and might be tried in a warm corner in other very mild parts of our area; otherwise it is a conservatory plant.

Significant pests and diseases are rare. Propagation is conventionally by division in the rhizomatous C. plumbaginoides, in the others by softwood or semi-ripe cuttings in summer, with bottom heat and humidity; layering is also feasible (Huxley et al. 1992). These methods are so satisfactory that seed-raising is rare, except when introducing wild material or in developing new cultivars. Perhaps this is why no garden hybrids have been noted. Apart from Peter Catt’s small-scale but commercially successful breeding program in C. willmottianum, plant breeders seem to have left the genus alone.