Studies in the liverwort family Aneuraceae (Metzgeriales) from south ern Africa. 3. Riccardia compacta

Riccardia compacta (Steph.) S.W.Amell is described and illustrated. Its presently known distribution in southern Africa is indicated on an accompanying map. Some differences between local plants of the species and those from Tanzania are discussed.


INTRODUCTION
Riccardia compacta (Steph.) S.W.Amell is the third species to be treated in the current revision of southern African taxa of this family. It was originally collected on Table M ountain by Jelinek. while on a N ovara Expedition, and described as Aneura compacta by Stephani in 1893. Sim (1926 referred to it briefly, stat ing that it frequently occurred in forest districts through out South Africa, which is doubtful, however, as it is apparently rare. Later on. two more specimens were col lected on Table Mountain: one by Pillans. and the second by Arnell. w ho also collected it at Knysna. Arnell (1952) then transferred it to the genus Riccardia. During the 1950s Esterhuysen collected more specimens on high altitude peaks in Western Cape, but. regrettably, there have been no local collections since then. Arnell (1959) described an Esterhuysen specimen from Tanganyika (Tanzania) as R. kilimandjarica S.W.Amell. which was placed in synonymy under R. compacta by Meenks & Pócs (1985). Pócs, either alone or w ith Ochyra. collected more material on Mt Kilimanjaro and Mt Meru during the 1980s. It was concluded by Meenks & Poes that there were a few differences between Tanzanian plants of the species and the type specimen collected by Jelinek on Table Mountain. They also state that R. compacta is an Afro-alpine species which occurs at higher altitudes in East Africa than in South Africa. Thalli prostrate, in compact cushions or patches, up to 3 or 4 mm thick, in several overlying lasers of dense and intricately intertwined, slight!) rounded, narrow bands, quite fleshy and firm, but rather brittle, bright green, with an ill-defined, somewhat darker, narrow strip occupying median 7» or more of branches; when dry. brown to dark brown. Main axis up to 15 mm long, not markedly dif- ferentiated. though often retaining its dom inance, branching generally crowded and irregular ( Figure 1A). sometimes bifurcate ( Figure IB), branches narrowing som ewhat toward tips or maintaining same width throughout, rarely slightly dilated distally. apically ± truncate, shallow ly notched, margins obtuse, not w inged. Primary' hranches/pinnae arising laterally from axis, very close together or at intervals of 0.30-0.95 mm between successive ones, opposite or subopposite, obliquely spreading at angles up to 30° w ith axis, occa sionally suberect. often relatively well developed. 2.0-3.5 mm long, up to 0.5 mm wide, usually linear, some bifurcate, rarely trifurcate. others poorly devel oped. almost rudimentary. Secondary branches/pinnules quite rare, if present, mostly remaining small. Stolons unbranehed. generally arising on basal part of axis, occa sionally developing more distally from apices of some primary branches. Dorsal epidermal cells in median part of apical segment of main axis ( Figure 1C Figure 1G) plano-convex. ± 200 pm or 5/6 cell layers thick medianly. up to 475 pm wide, medullary cell walls slightly thickened; cross section of stolon ( Figure IH) oval. ± 160 pm or 6 cell layers thick medianly. ± 210 pm wide, cell walls somewhat thick ened. Mucilage papillae ventral, clustered at emarginate apex ol pinnae and then in two. spaced, acropetal rows, one on either side ot midline, reddish brown, clubshaped. 62.5-75.0 x 20-25 pm abo\e. narrower below, rarely persistent. Rltizoids in widely scattered, irregular patches along ventral surface of pinnae, hyaline, up to 15 pm wide. Asexual reproduction b\ gemmae not seen. Dioicous. Antheridial branches solitary or occasional ly in sympodial pairs, lateral on main axis opposite or subopposite prim ary branch, otherw ise on primary branch close to its base ( Figure II), stipitate, ± oblong ( Figure 1J), up to 600 pm long, ± 320 pm wide, mostly bearing 3 or 4, rarely up to 8 pairs of antheridia, cavities 70-100 pm diam., encircled by 2 or 3 rows of cells, 25-30 x 30 pm ; in cross section ± 250 pm high ( Figure  1L), margins winged ( Figure IK), with single, incurved row of large, obliquely orientated cells, 75.0-112.5 x 40-45 pm, apically rounded and free, otherwise adjoin ing; ventral papillae up to 30 pm long, in lateral pairs, but rare. Gynoecial branches ( Figure IN) short and oblique ly ventro-lateral on main axis, opposite or at base of pri mary branch, ± sessile, not seen with archegonia, only with very young calyptra, up to 500 pm long, including corona, the latter with apical row of radiating cells, width across widest part below 480-525 pm , including sur rounding paraphyses, which are scale-like and dentate, with cells 40.0-52.5 x 27.5-30.0 pm. Calyptra ( Figure  1M, O) clavate, up to 2.4 x 0.7 mm, cross section of wall ± 180 pm or 5/6 cell layers thick ( Figure IP), some out ermost ones with low protrusions, at apex corona smoothly rounded above, cells thick-walled, dislodged at maturity by emerging capsule. Seta ± 6.5 mm long, ± 280 pm wide, with 4 central and an outer row of 12 cells, i.e. 4 cells diam. (Figure IQ ). Capsule ± ellipsoidal, 790-850 pm long, elaterophore internally attached to apex; valves 280-350 pm or 21-24 cell rows wide, cells of epidermal layer in external longitudinal view ( Figure  2A), 55-110 x 10.0-12.5 pm , with strong vertical (radi al) thickenings, evident as nodular brown bulges; in cross section ( Figure 1R) cells rectangular, ± 7.5 pm thick, thickenings on adaxial radial and abaxial radial walls extending slightly onto outer and inner tangential walls, resulting in a crescentic ring of thickening, bands on one side of median wall alternating with those on the other side in a mirror image; inner cells in internal longitudinal view ( Figure 2B) 67.5-85.0 x 12.5-17.5 (-20.0) pm, with nodular thickenings and with faint semiannular bands; in cross section ( Figure 1R) cells ± 10 pm thick, thickened on adaxial radial and abaxial radial walls, extending slightly onto outer tangential walls and further across inner tangential walls. Spores (Figure 1Ti_3 Hewson (1970), that the 'normal' number of n = 10 may be associated with the dioicous state and the monoicous state with double that number, i.e. 20].

DISCUSSION
The occurrence of Riccardia compacta in southern Africa appears to be restricted to Western Cape ( Figure  3), generally at altitudes between 1 085 and 1 220 m, but it has also been collected close to sea level. It usually grows in wet, shady places, with low light intensity, on rotten wood, soil banks or on Table Mountain Sandstone.
The oil bodies were initially and briefly described by Amell (1952) as 'rare, 3-8 x 8-22 pm. brownish* and he also illustrated a cell containing one (Amell 1952: fig.  2h). In a much more detailed account, Meenks & Pócs (1985) observed that the oil bodies were 'rare or even absent in the epidermal cells and never in the cells of the w in g \ They were, however, 'frequently present in the inner cells of the th a llu s\ and were described by the above authors as 'light brown, 1 or 2 per c e ll\ as well as 'globose to ellipsoid or bean-like, 8 x 8-30 x 12 p m '.  we were unable to find fresh material in order to study the oil bodies. All the collections that I examined, had long ago lost their oil bodies. The colour of fresh plants could also not be observed, but Pillans had noted on the label of his collection from Table Mountain, that they were bright green and Meenks & Pócs (1985) stated that they were emerald green to dark brown.

During a visit to
In a comparison between plants from Tanzania and those from South Africa (with information reported by Arnell 1963). Meenks & Pócs (1985) found that the for mer specimens were generally longer at 20 mm versus 7 mm; in transverse section 6 or 7 cell layers thick, versus 4 or 5 cell layers; spore diameter 12-22 pm versus 12-14 pm; wing of main axis 1 cell wide, cells elongated, up to 75 x 40 pm. versus wing absent.
In my findings the thalli of our plants are up to 15 mm long and in cross section the ultimate segment of the main axis is 6 or 7 cell layers thick, the spore diameter is 12.5-15.0 pm . and the male branches are shorter at ± 600 pm long, with 3 or 4. rarely up to 8 pairs of antheridia; the w ing is formed by a row of elongated, apically round ed cells that are laterally adjoined, but apically tree. Male branches in the Tanzanian specimen that I studied are mostly longer than ours. i.e. up to 16(H) pm long, with 4-1 4 pairs of antheridia. the w ing consisting ot a row ot quadrate to rectangular cells, 4 2 .5 -4 5 .0 pm high. 37.5-42.5 pm wide, their sides joined together right up to the edge; ventrally there are 2 rows of conspicuous redbrown mucilage papillae, up to 60 pm long, rarely seen in South A frican specim ens. As M eenks & Pécs observed, there are a few differences between Tanzanian and South African plants. The general facies o f the plants from the two areas is, however, closely similar and there is no doubt that they belong to the same species.
Meenks & Pócs expressed the opinion that all conti nental A frican records o f R. caespitans (Steph.) E.W. Jones (nom. inval.) probably belong to R. compacta. Apparently. Jones (1956) had identified Hedberg speci mens from Mt Muhawara. Uganda, as possibly Aneura caespitans\ but had added that R. caespitans is 'known with certainty only from the M ascareignes' [Stephani (1892) Bourbon, leg. Rodriguez]. I have examined Dusén G010680 from Cameroon, incorrectly labelled as 'TY PU S' with a red sticker, and also figure no. 236 in Stephani's leones hepaticarum (1985), which was prob ably drawn from the aforesaid Dusén specimen (accord ing to a note with it). This plant is decidedly different from R. compacta specimens: in cross section the ulti mate segments of the main axis are 4 cell layers thick and winged, with 3 unistratose cell rows at both margins. The lower axis in cross section is 6 cell layers thick median ly and not 9 cells as in Stephani's leones hepaticarum. W hether Dusén's specimen belongs to the same species as Rodriguez's type collection from Réunion, is very doubtful, as the latter species was reported by Stephani to be 10 cell layers thick medianly.
For several reasons I also hesitate to accept R. kili mandjarica as a synonym of R. compacta, as was done by Meenks & Pócs. In R. kilimandjarica the thalli appear to be lighter-coloured, lacking the central dark line, and more delicate than in R. compacta. Cross sections of the ultimate segment of the main axis are ± 500 pm w ide and 5 or 6 cell layers rows (or 150-170 pm) thick, with the medullary cells clearly thin-walled, not thick-walled as in R. compacta. Schuster (1963) suggested that a number of species, among them R. compacta, may be assigned to the sub genus Phycaneura, which is, however, characterized by relatively large, thin-walled dorsal epidermal cells. Although they had not studied specimens and illustra tions of R. compacta. Brown & Braggins (1989) think that the species that Schuster (1963) grouped together, appear to be members of section Alcicornia, subgenus Riccardia, rather than of subgenus Phycaneura. Stephani (1899) had placed together some of these species, e.g. A. alcicom is, A. compacta, A. diminuta, A. suhsimplex, A. sumatrana and others, as 'Plantae filiformes, subteretes'. R. compacta appears to bear some similarities to R. subalpina Furuki (1991) from Japan, as it also grows (most ly) at higher altitudes, and has small thalli and thickwalled cells.
R. compacta specimens are distinguished by the fol lowing characters: 1, the plants grow in dense, thick cushions of several layers of intricately intertwining branches, mostly at high altitudes; 2. the thalli are narrow and small, only up to ± 15 mm long, with irregularly branched, lateral branches, their margins generally paral lel and unwinged; 3, from above the dorsal cells of the ultimate segment of the main axis are thick-walled, and in cross section, so are the medullary cells; 4, the antheridial branches in South African plants arc usually short and have conspicuous marginal cells; and 5. the corona of the calyptra is smooth and the cells thickwalled. Jelinek I9560G (holotype).