The Genus Hydnora

Recent research on the genus Hydnora is reviewed. Hydnora is tentatively considered to consist of four species: H. johannis, H. triceps, H. esculenta, and H. africana. Notes on the floral biology, distribution, nomenclature, economic importance, and uses are given with emphasis on H. triceps(1).

The Hydnoraceae, with only two genera Hydnora and Prosopanche, includes some of the strangest plants in the world. Hydnora is found in the semi-arid regions of Africa and southern Arabia while Prosopanche grows in similar habitats in South and Central America. The vegetative plant body is highly reduced consisting of only roots and flowers. In fact, the Hydnoraceae are the only known angiosperms with no leaves or scales of any sort. Both genera are holoparasites and are totally dependent upon their hosts for their existence. Only the flower emerges from the soil. The extreme reduction limits the number of taxonomic characters and has raised questions as to their phylogeny. The Hydnoraceae has been allied with several different groups, eg, the recent summary in Nickrent and Franchina (1990). Previous work on Hydnora is summarized in Musselman and Visser (1989) but due to their furtive nature and seasonal appearance, they remain poorly known plants. This paper is a review of some of my current work on the systematics and biology of the genus based on herbarium and fieldwork, with a key to species and notes on each species.

Hydnora Thunberg, Kongl. Vetensk. Acad. Handl. 36: 69. 1775. Hydnora Aphyteia Linnaeus, Pl. Aphyteia, p. 7. 1776. TYPE: H. africana Thunberg, 15452, UPS!(2)

Subterranean holoparasitic herbs with often massive root systems spreading laterally from the host. Roots up to 1 dm wide 4-5 angled, terete or sometimes flattened. Flowers 3, 4, or 5merous. Perianth lobes patent and resting on the soil, or lobes not reflexed and flower opening by a separation of the lobes. Flowers variable in size from 5-25 cm, the length depending on the depth of the root, pedicel 4-9 cm. Ovary inferior, with numerous infolded, pendant placentae; unilocular. Stigmas sessile, intricately grooved. Androecium complex, with stamens fused. Pollen monocolpate. Perianth lobes 6-8 cm; in some species "bait bodies" are present between the inner marginsof the lobes; in other species a well developed osmorphoric region ("cucullus") is present. Fruit fleshy.

AA. Flowers 5-merous, imperfect................... 1. H. esculenta.

AA. Flowers 3 or 4-merous (rarely 5), perfect..... BB.

BB.Flowers 3-merous (rarely 4), cucullus absent; roots strongly angled...............................CC.

BB. Flowers 4-merous (rarely 3 or 5), cucullus present; roots not strongly angled. 2. H. johannis.

CC. Flower 5-8 cm long; perianth lobes forming a hood.......................... 3. H. triceps.

CC. Flowers 1-15 cm long; hood absent....... 4. H. africana.

1. Hydnora esculenta Jumelle & Perrier de la Bathie. Restricted_to Madagascar, this species has not been collected since 1947. I have found only four collections of this unusual species. A deliberate search should be made to determine if it is still extant. Hydnora esculenta is perhaps the most specialized of all members of the genus as the flowers are unisexual. The pistillate plants apparently have well developed staminodes.

2. Hydnora johannis Beccari. I consider the following as synonyms: H. abyssinica A. Braun, H. abyssinica var. quinquefida Engler, H. angolensis Decaisne, H. bogoensis Beccari, H. cornii Vaccaneo, H. gigantea Chiovenda, H. hanningtonii Rendle, H. michaelis Peter, H. ruspolii Chiovenda, and H. solmsiana Dinter. In 1873, Decaisne also described H. aethiopica apparently based on a specimen from Sabatier's expedition (Decaisne, 1873 in Musselman and Visser, 1989). However, this name appears to be based on an incomplete specimen. This is the most widespread and frequently collected Hydnora. It is known from western Namibia, northern Botswana, Zimbabwe, Zaire, Tanzania, Kenya, Ethiopia, Somalia, along the Blue Nile to Khartoum in Sudan; and in the Arabian Peninsula in southern Saudia Arabia, northern Yemen, and only recently from the Dhofar region of Oman. Except for one report of Albizzia, (Shantz, 198, K, as H. hanningtonii) the hosts for H. johannis are always species of Acacia. This is the only Hydnora known to be of any economic importance. In Ethiopia it has been reported to break up the pavement of roads (Parker, 1988) and likewise to break up cement floors in Sudan (Musselman, unpublished).

3. Hydnora triceps Drege & Meyer. In 1988 Visser discovered a_large population of this very poorly known plant which was thought extinct because it had been collected only a few timessince its discovery in the 1830's. Like H. africana, it parasitizes shrubby species of Euphorbia. Visser was able to visit this population only once before his death and found that Hydnora triceps represents yet another pollination syndrome within the genus. His preliminary work (Visser, 1989 and personal communications) indicates that bait bodies are present and that they attract blowflies which lay their eggs within the flower where the larvae develop. The flowers of H. triceps apparently develop only underground and do not emerge from the soil. Insects enter the flowers through soil cracks.

This remarkable adaptation to underground pollination deserves verification and further investigation. In fact, many angiosperms have underground flowers but inevery instance the flowers are actually produced above ground and grow underground to develop fruits (as in the common peanut), or are cleistogamous and borne right at the soil surface (as in many violets), or flower under litter or duff. These are not truly subterranean, however, in that the flower does not open beneath the soil. Subterranean flowering is extremely rare in the angiosperms and the only other documented case is that of the orchid genus Rhizanthella, endemic to Australia with two species (Jones,1988). No subterranean flowering dicots are known. In a popular account of the biology of Rhizanthella Sherwin Carlquist makes several statements pertinent to the study of H. triceps (Carlquist, 1965). "How does a flower manage a completely underground existence?" And, "Rhizanthella's habits are not only mysterious, we have no idea how or why such an orchid has evolved at all".My examination of material of H. triceps suggest that its flower structure and vernation are well adapted to a subterranean existence. Unlike any other Hydnora, the perianth lobes are connate in such a way that the only opening into the flower isthrough slits on the side of the floral tube. This hood-like structure ensures that the orifice is not occluded by soil as the flower grows towards the surface. This may be one way that it manages "...a completely underground existence". Further field studies are necessary to discover other adaptations.

What are the plant-insect relationships in H. triceps? First, the insects which can be pollen vectors in H. triceps need to be determined. The pollinator fidelity of the insects also needs study. Put another way, does H. triceps have a highly specialized pollinator which is dependent upon this plant either for its food source as a place for reproduction? In other species of Hydnora, pollinators harvest pollen (eg, Musselman and Visser, 1989). The situation in H. triceps is not clear and can only be determined by field observations. What is the possible relationship between the pollination of H. triceps and its host, Euphorbia dregeana E. Mey. ex Boiss. and other nearby plants? The fruit of H. triceps has never been described.

The name Hydnora triceps was first used by Meyer in 1838 based on a specimen collected by Drege in Namaqualand in southern Africa. The proper citation of this name would be Hydnora triceps Drege et Meyer. However, an Englishman by the name of W. J.Burchell traveled in southern Africa earlier in the 1800s and published an account of his travels in 1822 (Burchell, 1822). In1826, K. Sprengel in his Systema Vegetabilium cites the name Aphyteia multiceps. According to the rules of botanical nomenclature, this is a validly published name because Sprengel based it on the description of Burchell in which case the name would be cited as Aphyteia multiceps Burchell & Sprengel. Plants examined by Sprengel were usually kept in the large herbarium at Berlin and since the herbarium at Berlin was largely destroyed during World War II. The correct application of the name could only be clarified by noting the original description of the plant in question by Burchell. In this book in a footnote to page 213 Burchell (1822) notes:"Aphyteia multiceps, B., is a new species, found in the more western parts of this Karro, and of which I received a specimen from my friend Hesse, to whom it had been sent from the district of Clan William or the Elephant's River. It is easily distinguished, by a subterraneous stem, about two inches long, clothed with a few large scales, as in all radical parasitic plants, and producing, in a close head, several flowers, (in my specimen, five,) which had not the appearance of being succeeded by a seed-vessel of the magnitude at all proportionable to that of A. Hydnora.

This description clearly indicates that it can NOT be a species of Hydnora as no scales or multiple flowers are present in Hydnora. It is not certain which plant Burchell was referring to but my guess would be that it is one of the Balanophoraceae, perhaps Sarcophyte sanguinea which occurs in South Africa (Visser, 1981).

4. Hydnora africana Thunb. Synonyms: H. longicollis Welw., H. tinctoria Welw. (each of these names apparently a nomen nuda), H. africana var. longicollis Welw. Known to parasitize only shrubby species of Euphorbia. This is the best known of all Hydnora species even though it is restricted in distribution to a small stretch of southern Africa from the Cape Province as fare east as Swaziland.

In a review of floral biology of H. africana (Visser, 1981) and H. johannis (Musselman, 1984; Musselman and Visser, 1989) it was found that beetles pollinate the flowers but in remarkably ifferent ways. Hydnora johannis is visited mainly by scarib eetles (Musselman, 1984) which are attracted by a distinct osmophore ("cucullus") on the adaxial surface the perianth lobe. Hydnora africana, on the other hand, lacks such cuculli but has "baitbodies", specialized perianth hairs which omit the odor of decaying leather and attract various dermestid beetles (Visser and Musselman, 1986 and Visser, personal communications).

Hydnora africana is the only member of Hydnoraceae which has been cultivated. The following account is adapted from Carlquist (1989).

In December, a visit to the Karoo Garden in Worcester revealed that on a nearby hillside, some of the older Euphorbia mauritanica shrubs were yellowish, and I thought their decline might be the result of parasitism. Hydnora africana was fruiting abundantly on the roots of one such shrub, and I collected fruits in the hope of cultivating them upon my return, the last week of December, to the United States. Shortly after arrival back in California, I bought several rooted cuttings of Euphorbia caputmedusae in the hope of germinating seeds of the Hydnora; the fruits were still in fresh condition and intact. Euphorbia mauritanica is not in cultivation in California, but E. caputmedusae, most commonly cultivated here, was judged suitable because it is the host of a small population of H. africana at Houtbaai on the Cape Peninsula. I tapped the Euphorbia plants from their pots, lined the pots thickly with masses of the seeds, and reinserted the plants so that Euphorbia roots would be in intimate contact with Hydnora seeds. About a year later, I removed the Euphorbia plants, noted no apparent infection, and planted them in a larger ceramic pot. Three years from the time of original sowing, removal of the plants revealed no apparent infection. I assumed I had been unsuccessful, and planted the E. caputmedusae plants in a convenient sunny corner of my garden as a way of using the Euphorbia merely as an ornamental. Early in July 1979, five and one-half years after the attempted inoculation of the Euphorbia plants, I was startled to see a single Hydnora flower emerge from the soil surface.

Field work- Considerable more field work is necessary. These plants are so unusual in being largely subterranean and also seasonal that they are no doubt frequently overlooked. The recent re-discovery of H. triceps, noted above should encourage botanists to be more aware of the possible presence of these plants. Additional, well prepared collections are necessary to determine the extent of variability especially in H. johannis. Many herbarium specimens are little more than shards of unusable material.

Geography- The recent collections of H. johannis from Oman add a tremendous range extension. The vegetation of this region of the Arabian peninsula, however, is strongly linked with the corresponding region of Africa.

Culture- The remarkable success of Carlquist in cultivating H. africana should encourage others to do likewise. To my knowledge, there is no botanical garden in the world which is attempting this.

Hosts- Hosts of all four species are known. In the context of holoparasites, Hydnora appears to be rather advanced in that species are generally host specific. (The exception may be H. esculenta but more data are needed). Two unrelated families are parasitized, the Fabaceae and the Euphorbiaceae and within these two large families parasitism is more or less restricted to only two genera, Acacia and Euphorbia.

Conservation- The status of H. esculenta requires immediate attention. So many of the endemic plants of Madagascar are either extinct or facing extinction that this species, unique within the family, may have met the same fate.

Sub-generic classification- Musselman and Visser (1989) include a classification-sub-generic classification but this need considerable further work before it can be considered meaningful. The typification of H. triceps is discussed above.

Familial relationships- The genetic distance between Prosopanche, restricted to the New World, and Hydnora deserves study. Modern molecular studies of plastids (if present) and other nucleic acid sequencing could be powerful tools in understanding the relationships between these two genera. Such a study of holoparasites would have value in other groups, as well. Most modern systems of phylogeny place the Hydnoraceae close to the Rafflesiaceae. I consider this an error as there is little in common between these two families other than being holoparasites. A relationship with the Aristolochiaceae seems more reasonable considering the inferior ovary, beetle pollination syndrome, and large conspicuous singly borne flowers.

Breeding systems- While floral biology is being studied, we know nothing about the breeding systems of these plants. Part of this is due to the inherent difficulty in simply determining what is a plant! The lack of any fruits in H. triceps is intriguing in this regard. Is it an outcrosser?

Morphology- A thorough study of the development of Hydnora is needed. The early stages of germination have not been described. The monocolpate pollen, unusual in dicotyledons, deserves further analysis. The anatomy of the root and its distinctive budding system should be carefully described.

Uses- The most frequently cited use is eating the fruits. Those who have had the privilege of doing so can readily attest to their delicacy. Several animals eat the vegetative parts. Because of the high tannin content, the plants have been used as an anti-diarrheal medicine and for tanning hides (Musselman and Visser, 1989). Less common but more dramatic uses are elephant and rhinoceros food (Musselman and Visser, 1989)!

Field work in southern Africa was organized by the late J. H.Visser. I have examined material from the following herbaria andwish to thank the curators for arranging visits and/or loans forstudy: BR, C, E, EA, FT, K, L, LE, LISC, M, MO, ODU, P, PRE, RO,S, SRGH, UPS, and WIND.

Burchell, W. J. 1822. Travels in the Interior of Southern Africa. Volume 1. London: Longman, Hurst, Rees, Orme, and Brown. (Reprint by Johnson Reprint, New York. 1967).

Carlquist, S. J. 1965. Island Life. Garden City, NY: The Natural History Press..

1989. Hydnora in my garden! Haustorium, Parasitic Plants Newsletter 21:2-3.

Jones, D. L. 1988. Native Orchids in Australia. Frenchs Forest, NSW: Reed Books.

Musselman, L. J. 1984. Parasitic angiosperms of Sudan: Orobanchaceae, Hydnoraceae, and Cuscuta. Notes Royal Botanic Garden, Edinburgh 42: 21-39.

Musselman, L. J. and J. H. Visser. 1987. Hydnora johannis in southern Africa. Dinteria 19: 77-82.

Musselman, L. J. and J. H. Visser. 1989. Taxonomy and natural history of Hydnora (Hydnoraceae). Aliso 12(2): 317-326.

Nickrent, D. L. and C. R. Franchina. 199. Phylogenetic relationships of the Santalales and relatives. Journal of Molecular Evolution 31: 294-31.

Parker, C. 1988. Parasitic plants in Ethiopia. Walia 11:21-27.

Visser, J. H. 1981. South African Parasitic Flowering Plants. Cape Town: Juta.

Visser, J. H. 1989. Hydnora triceps. The Flowering Plants of Africa 5(2). Pretoria: Department of Agriculture and Water Supply.

Visser, J. H. and L. J. Musselman. 1986. The strangest plant in the world. Veld and Flora. December 1986/January 1987: 19-111.

1. ADAPTED FROM: Musselman, L. J. 1991. The genus Hydnora (Hydnoraceae). pp 247-25 in Ransom, J. K., L. J. Musselman, A. D. Worsham and C. Parker, eds. 1991. Proceedings of the 5th International Symposium of Parasitic Weeds. 55 pp ix. Nairobi: The International Maize and Wheat Improvement Center (CIMMYT).

2. A large project to typify all Linnean specimens is centered at the British Museum (Natural History) under the direction of Dr Charlie Jarvis who asked me to designate a type specimen for the genus. I had examined the type specimen of Hydnora africana in the Thunberg collection at Uppsala (Sweden) and with the help of Dr Mats Thulin have chosen Thunberg 15452 as the type. This sheet (Thunberg 15452) is marked " Capite bonae spei C. P. Thunberg."