Studies in the Marchantiales (Hepaticae) from southern Africa. 1. The genus Dumortiera and D. hirsuta; the genus Lunularia and L. cruciata

The genera Dumortiera (Dumortieroideae, Marchantiaceae) and Lunularia (Lunulariaceae), are briefly discussed. Each genus is represented in southern Africa by only one subcosmopolitan species, D. hirsuta (Swartz) Nees and L. cruciata (L.) Dum. ex Lindberg respectively. Die genusse Dumortiera (Dumortieroideae, Marchantiaceae) en Lunularia (Lunulariaceae) word kortliks bespreek. In suidelike Afrika word elke genus verteenwoordig deur slegs een halfkosmopolitiese spesie, D. hirsuta (Swartz) Nees en L. cruciata (L.) Dum. ex Lindberg onderskeidelik.

Only a few of the specimens examined were fertile; most of these were dioicous, with male and female receptacles borne on separate plants. Monoicous plants were quite rarely found, but have frequently been reported in the literature and even bisexual receptacles are known (Evans 1919).
All southern African specimens have been identified as belonging to D. hirsuta: the dorsal face of the thallus is always ± smooth and lacks papilliform cells or traces of air chambers. D. nepalensis and D. velutina, which our material resembles, are considered to belong to the poly morphic species D. hirsuta (Grolle & Piippo 1984). Amell (1963) referred to material from southern Africa (Cape Province, Transvaal, Natal) as D. hirsuta var. nepa lensis. Schuster (1992) considers D. nepalensis (Taylor) Nees to be a distinct taxon under the name of D. hirsuta subsp. nepalensis (Taylor) Schuster which is not found in Africa. D. hirsuta is subcosmopolitan and is widespread in tropical and temperate regions, generally growing in sheltered, wooded, shaded and damp areas, i.e. it is hygrophilous. In southern Africa it is known from northern and eastern Transvaal, Swaziland, Natal, Zululand and eastern, southern and southwestern Cape (Figure 4). Further north wards in Africa, D. hirsuta is reported from (or has been collected in) Zimbabwe (Best 1990 On the basis of flavonoid data, Campbell et al. (1979) found chemical affinities between Dumortiera and Wiesnerella, suggesting that they belong in a common family, Wiesnerellaceae (Inoue 1976). This was accepted by Grolle (1983), but Schuster (1984) has retained Dumortiera in the Marchantiaceae, creating a new subfamily, Dumortieroideae Schust., for it and designating Dumortiera as the type genus. Dumortiera has also been classified in the Dumortieroideae by Bischler (1988), who regards Dumor tiera and Wiesnerella as morphologically remotely related, although having similar flavonoid patterns. The treatment of Schuster (1984) is followed here, even though the phylogenetic position of Dumortiera remains a matter of some controversy, as conceded by Schuster (1984). Sup porting the above treatment are terpene studies of the genera by Asakawa et al. (1979Asakawa et al. ( , 1980aAsakawa et al. ( , 1980bAsakawa et al. ( , 1981, who found that Wiesnerella denudata and Conocephalum conicum shared 17 of these terpene compounds, but only had three in common with Dumortiera hirsuta.

Thallus large, flat, somewhat glossy, green, in overlying patches, sometimes in extensive turfs; on damp soil in old gardens, nurseries and forested areas, may have been introduced into southern Africa. Branches dichotomously furcate toward apex, new growth by apical or lateral innovations; thickened over midrib, gradually thinning toward slightly undulate, hyaline margins; apex emarginate. Dorsal epidermis persistent, hyaline, cell walls sometimes thickened at corners or entirely; air pores simple, elevated and conspicuous, surrounded by several concentric rings of cells, leading below into individual air chambers, these in one layer and floored by chlorophyllose layer of 3 -5 -ce lled erect, branched filaments; storage tissue compact, cells colourless, sometimes with pitted walls; scattered cells throughout with single, large, brown oil body; rhizoids numerous, between ventral scales and on midrib, some smooth, others tuberculate; scales ventral, in curved parallel rows on either side o f midrib, with round or reniform appen d ages, several c e lls c o n taining oil bodies.
Dioicous. Antheridia sunken into slightly elevated, oval or kidney-shaped, disciform receptacles encircled by raised, membranous sheath, at sides of male thalli (although originally terminal). Archegonia enclosed in terminal receptacles, but by continued growth of thallus laterally situated, conical, white and bud-like when young, sheathed in layers of scales and slightly sunken into rounded depression with distinct rim; after fertilization and further growth, four tubular involucres in the form of a cross are formed, each enclosing one or two sporophytes, composed of foot, seta and capsule, raised on unfurrowed, hairy stalk; capsules eventually exposed by elongation of seta, the wall unistratose, lacking annular thickenings, dehiscing by 4 valves. Spores very small, green or brown, smooth. Elaters long, tapering and bispiral. Gemmae numerous, disc-shaped, inside crescentshaped ridge; wholly diagnostic. Sporophyte virtually unknown in southern Africa.  Muller (1951Muller ( -1958.

Dioicous. Male plants quite rare in southern Africa.
Antheridiophore on alternate sides, having originally developed in terminal sinus near apex ( Figure 6A), slighdy raised, ovate, flattish discs, 2 -3 mm in diameter, encircled by membranous sheath with crenate edges, containing numerous antheridia, individually sunken in flask-shaped cavities ( Figure 6B), opening above by pores. Archegoniophore originally also developing in terminal sinus near apex, with further growth of thallus leaving it behind, so that it appears lateral in position on alternate sides of thallus ( Figure 5B); commences as domed disc ± 200 x 375 /xm , bearing several archegonia in radiating rows and attached by very short stalk to floor of shallow, round hol low, 1100 /xm wide X 300 /xm deep, with crenate rim; sheathed in ± 3 layers of scales ( Figure 5K), outer layer formed by fused, shaped scales, bulging in centre and con stricted below and above, upper edge irregularly fringed with filiform cellular appendages of up to 7 rectangular cells, ± 35 x 15 /xm , cells in body of scale thick-walled, 5-or 6-sided, up to 57 x 25 /xm , in between numerous smaller cells, 4-or 5-sided, 25 X 15 /xm , almost entirely filled with single oil body; between scales and especially from base of archegoniophore and arching over it, numerous uniseriate, long hairs of 16-20 cells each. Further development not recorded as fertilization did not take place, nor are sporophytes available for study. Gemmae numerous, disc-shaped ( Figure 5M ), notched Since so many collections are from nurseries or city gardens, it would appear that L. cruciata may have been introduced into southern Africa. It is not frequently found here, most collections being from the southwestern Cape, a few from southern Transvaal and some from Natal, and then quite frequently from nurseries (Figure 4). Further north in Africa, L. cruciata is known from Zimbabwe (Best 1990 Lunularia cruciata is quite widespread in the southern hemisphere but regarded by Engel & Schuster (1982) as probably Laurasian in origin. Since sexual reproduction and subsequent spore production are so rare, its gemmae obviously present a highly effective means of ensuring its dispersal, which is most likely aided by human activities.
On the basis of its flavonoid chemistry, Campbell et al. (1979), include it in the Marchantiaceae, but in the present treatment Grolle (1983) is followed and L. cruciata is classified in the monotypic family, Lunulariaceae Klinggr. (1858).
Its phylogenetic position seems rather unclear, Schuster (1984b) arguing that he would place it 'low' in the Marchantiales on account of its high level of seta re tention, 2 -3 sporophytes per gynoecium and a capsule with 4 well-defined valves. On the other hand, he expresses the viewpoint (Schuster 1984b) that its archegoniophore, clearly formed from two dichotomies, each producing archegonia, is complex and therefore an advanced feature.