A continuation of the series guest-written and photographed by Martin Deasy, who is a British horticulturist based in Oxford, England. Martin writes:
The startlingly furry appearance of Trichocladus crinitus, or black witch-hazel, is caused by the presence of rusty-brown stellate trichomes (or hairs). These are a feature of practically all Hamamelidaceae species, but here are closely packed to form an unusually dense indumentum. The resemblance of the unexpanded opposite (or sub-opposite) leaves to pairs of rabbits’ ears only emphasizes the impression of animal fur.
Trichocladus crinitus is endemic to the moist Afromontane forests of South Africa, where it is locally dominant as a constituent of the understory vegetation–the Afrikaans name onderbos translates literally as “undergrowth.” The plant forms a large, rather open shrub or small tree up to 3m in height, and its hard, white wood gives it its local Xhosa name iThambo (“bone”).
The tiny flowers are packed together in dense spherical heads. The photo shows flowers before and after the anthers have dehisced. The floral features can be easily made out: 5 green sepals with rusty-brown stellate trichomes on the outer (abaxial) surfaces, and 5 pale-pink petals with involute margins and stellate pubescence, again on the abaxial surface. The 5 pale pink stamens, and twin styles (green) can also be seen. The anthers release their pollen in an unusual manner: although the anthers are tetrasporangiate, each theca opens with a single valve.
Trichocladus crinitus is one of five species in a genus distributed from Ethiopia, down through East Africa to the Cape, and offers an interesting case study in biogeography and floristics. Early accounts of Trichocladus (e.g. Hutchinson 1933) noted its longitudinal north-south distribution on the continent, inferring dispersal into Africa southwards from an ancestral range in Eurasia.
However, it was subsequently noticed that Trichocladus‘s highly unusual mode of anther dehiscence–rare even among the rest of the angiosperms–is shared with a handful of Hamamelidaceae species restricted to Australia, Madagascar and Africa. Dubbed the “Southern Hamamelidaceae” (Endress 1989), the close genetic relationship of this group was subsequently confirmed by molecular studies, and circumscribed as tribe Dicorypheae (Li & Bogle 2001). The five constituent taxa are Noahdendron, Ostrearia, Neostrearia (Australia), Dicoryphe (Madagascar) and Trichocladus (Africa).
The disjunct distribution of the Dicorypheae aroused some interest, since the locations in which they occur map onto the remnants of the supercontinent Gondwanaland, which after an existence of several hundred million years, broke up in stages between 165mya and 50mya. The existence of a disjunct subgroup of closely-related and highly distinctive hamamelid taxa on vestigial Gondwanan landmasses suggested that the Dicorypheae may have originated in Gondwana, the surviving members arriving in their present situations by continental drift (or vicariance, to use the technical term).
The Hamamelidaceae is known to be an ancient clade (the fossil record dates back to at least the late Cretaceous, ca. 85mya), so this posited Gondwanan lineage was quite plausible. However, it would have made the Hamamelidaceae an exceptionally ancient angiosperm lineage, pushing the original diversification of the family well back into the mid-Cretaceous, prior to the disintegration of Eastern Gondwana.
Subsequent work has shown that the apparent Gondwanan distribution is a red herring. Tellingly, the Australian taxa occur in the extreme northeastern corner of the continent closest to Asia, where the relict Gondwanan rainforest was infiltrated by Asian taxa during the Miocene (5.4mya), when the collision of the Australian and Southeast Asian continental plates facilitated significant floristic exchange. Furthermore, molecular evidence has revealed the Dicorypheae to be embedded within a larger clade of overwhelmingly pan-Asian distribution, strongly indicating that the “Southern Hamamelidaceae” differentiated in Asia, arriving at their present positions by a process of dispersal.
Fossil evidence of Hamamelidaceae species from Antarctica–another Gondwanan remnant–might seem to throw a spanner in the works. But reports of fossil Hamamelidaceae pollen at Antarctic sites invariably turn out to refer to Altingiaceae, now split off as a separate family. In fact, the story of Trichocladus‘s evaporating Gondwanan history recapitulates an increasingly familiar narrative in which molecular dating techniques demonstrate that putative Gondwanan distributions actually result from much later dispersals out of the northern hemisphere (cf. Davis et al. 2002).
The plant shown grew for many years in the remarkable Temperate House at the Royal Botanic Gardens, Kew–the world’s largest surviving nineteenth-century glasshouse. The House is currently undergoing a major restoration, and most plants (with the exception of a few large palms) have been moved to temporary quarters for the duration of the five-year renovation project due to be completed in 2018.