Studies in the liverwort genus Fossombronia (Metzgeriales) from southern Africa. 9. A new species from Mpumalanga and KwaZulu- Natal, with notes on other species

A new species of Fossombronia, F. renateae, from Mpumalanga and KwaZulu-Natal in southern Africa is described. It is distinguished by its reflexed, undulating leaves and often stipitate pseudoperianth with a highly convoluted mouth, as well as by completely or incompletely reticulate spores and usually by short, blunt elaters with two spirals which often form rings or coils. Notes are provided on two unnamed species from Mpumalanga which cannot be fully described as yet, since ripe spores are not available for study. Brief references are made to some tropical African species. Fossom bronia renateae Perold, sp. nov. Plantae repcntes, dense congestae vel cum muscis intercrescentes. Folia late patentia, undulata, oblonga vel breviora quam lata, supra leve lobata. Dioicae. Antheridia in plantis masculis bracteis obtecta. Pseudoperianthium campanulatum; ore perirregulariter lobato et maxime convoluto. Sporae 37.5-47.5 pm diametro, cum 8 vel 9 cristis irregularibus, aliquando anastomosantibus, aliquot areolas formantibus. Elateres plerumque bis spirales, annulares vel spirales, 25-75 x 12.5-17.5 pm. TYPE.— Mpumalanga, 2530 (Lydenburg): Lone Creek Falls, on soil, at margin of pond below falls, (-BB), Perold & Koekemoer 4073 (PRE. holo.). Plants smallish to medium-sized to rather large, creeping, in dense, crowded stands or intimately to loosely intergrown with mosses, green; shoots often sim­ ple, in male plants 5-13 x 3^4 mm. 1.3-2.5 mm high; in female plants 9-15 x 3.5-4.0 mm. 1.6-2.5 mm high, sometimes furcate or with short, lateral branches near apex and/or toward base. Stems prostrate, in some popu­ lations distally markedly fleshy, tapering proximally, in living material from above, hardly distinct from adjoin­ ing leaf bases, sometimes faintly longitudinally striate, in cross section plano-convex, in male plants at apex (Figure 1L) up to 600 pm (12-14 cell rows) high, ± 630 pm wide, at base (Figure 1M) ± 570 x 540 pm; in female plants at apex (Figure IN) up to 680 pm (15 cell rows) high. ± 1000 pm wide, at base (Figure lO) ± 420 x 700 pm. Rhizoids purple, 12.5-25.0 pm wide. Leaves over­ lapping by about 'A of their width, spreading widely, undulating, succubously inserted, often slightly decur­ rent, apically smaller, soon becoming larger, oblong or shorter than wide, shallowly lobed above; in male plants (Figure 1A-F) 1200-2500 x 1300-2375 pm; in female plants mostly larger (Figure 1G—I), 1250-2500 x 1350-3500 pm, sometimes narrower above or below; * National Botanical Institute. Private Bag X101, (XX) 1 Pretoria MS received: 1999-01-18 margins with up to 13 well-spaced, sessile papillae, but sometimes fewer, only 2 toward leaf base, occasionally raised on a basal cell. Leaf cells thin-walled, in male plants not appreciably different from those of female plants, at upper margins (Figure 1J) rectangular across, occasionally isodiametric, 25.0-37.5 x 27.5-52.5 pm, at lateral margins long-rectangular. 50.0-87.5 x 15.0-22.5 pm; upper laminal cells 5or 6-sided, 35.0-52.5 x 27.5-37.5 pm; middle laminal cells 47 .5-75.0 x 3 2 .5 ^5 .0 pm; basal cells 75.0-107.5 x 42.5-62.5 pm. Oil bodies few (Figure IK), only 4-6 per cell, 5 pm diam., or a little larger, round, but some slightly irregular and with a central indentation; chloroplasts numerous, mostly rounded. ± 5 pm diam. Dioicous. Antheridia dorsal on stem, in 1 or 2 crowd­ ed rows, short-stalked, globose, distal ones often larger, 220-270 pm diam., white, proximal ones smaller, ± 160 pm diam., yellow, each shielded by a perigonial bract (Figures 1P-S; 2A. B), 460-580 x 200-380 pm, some­ times 2 adjacent ones joined together and then 2or 3lobed, mostly with apical processes topped by a papilla, marginal cells ± rectangular, 40-75 x 27.5^40.0 pm, inner cells usually 5-sided. 65.0-87.5 x 22-35 pm. Archegonia in 1 or 2 rows dorsally along stem (Figure 2C), distally and proximally, naked, sometimes 2 per branch, at intervals or 2 adjacent or even 3 in a cluster, becoming fertilised. Pseudoperianth (Figures IT. U; 2D-F) campanulate, about as tall as leaves, raised on a short stalk, then widely flaring above, up to 1625 pm long, ± 3500 pm wide across mouth, very irregularly lobed and highly convoluted, lobes 300—400 x 400-450 pm, with papillae at angulations, some sessile, others raised on a basal cell; cells comparable in shape and size to those of leaves, except for basal ones which are larger. 100.0-112.5 x 37.5-45.0 pm. Capsules globose, ± 750 pm diam., walls bistratose, cells in inner layer irregularly shaped (Figure IV), subquadrate, long-rectangular or tri­ angular. 25-50 x 27.5-40.0 pm. each cell wall with 1 or 2 dark brown, nodular and sometimes semi-annular thick­ enings. Seta 1.1-2.5 mm long, in cross section (Figure 1W) 200-300 pm diam.. 8 cells across. Spores light brown, hemispherical, 37.5-47.5 pm diam., including 78 Bothalia 29,1 (1999) FIGURE 1.— Fossombronia renateae. A-F, male leaves; G-I, female leaves; J, upper leaf margin; K, median leaf cells with oil bodies (solid lines) and chloroplasts (dotted lines); L, c/s male stem apex; M, c/s male stem base; N, c/s female stem apex; O, c/s female stem base; P-S, perigonial bracts; T, pseudoperianth from side; U, opened pseudoperianth; V, cells in inner capsule wall; W, c/s seta. A-F, K-M , P-S, Perold & Koekemoer 4071\ G -I, N, O, Perold & Koekemoer 4073; J, Lubenau-Nestle SA265\ T, U, Lubenau-Nestle SA264 p.p.; V, Doidge CH3564W, Perold & Koekemoer 4078. Scale bars: A -I, T, U, 500 pm; J, W, 100 pm; K, V, 50 pm; L-S, 250 Mm. Artist: G. Condy Bothalia 29,1 (1999) 79 FIGURE 2.— Fossombronia renateae. A, male plant with perigonial bracts; B, close-up o f perigonial bracts; C, female plant with archegonia; D, female plant with pseudoperianths (indicated by arrows); E, female plant with pseudoperianth near apex o f branch; F, close-up of pseudoperianth. A, B, E, F, Perold & Koekemoer 4071\ C, Perold & Koekemoer 4073\ D, Liibenau-Nestle SA265. A, x 8.7; B, x 20; C, D, x 8; E, x 8; F, x 25.6. cidges projecting at margin; distal face (Figure 3A-D) convex, with a tendency to form 8 or 9 well-marked are­ olae across face, 5.0-7.5 pm wide, occasionally ridges running parallel to each other or else in different direc­ tions, interconnected by faint or distinct cross walls or anastomosing and forming incomplete areolae, some­ times with small inclusions, surrounding ridges 2.5-4.0 |im high, crenulate above; proximal face (Figure 3E) lacking triradiate mark, mostly flat, covered with coarse, very irregular spicules or spikes, 20 or 21 ‘spines’ pro­ jecting up to 2.5 pm high around spore periphery and completely or somewhat incompletely joined by a ± 5 pm wide, membranous wing or perispore, its margin crenu­ late. Elaters in most populations relatively few, with 2 FIGURE 3.— Fossombronia renateae. A-E, spores; F, elater. A, B, distal face; C, side view of distal face; D, detail of part o f distal face; E, prox­ imal face. A, Perold & Koekemoer 4073; B-E, Lubenau-Nestle SA265: F, Liibenau-Nestle SA264 p.p. A, x 1046; B, x 908; C, x 881; D, F, x 1744; E, x 752. 80 Bothalia 29,1 (1999) spirals (rarely with 3 spirals partly), light brown or green­ ish, often in rings or coils (Figure 3F), rather short and stout, blunt at both ends, or ± wedge-shaped, 25-75 x 12.5-17.5 pm, sometimes sticking to spores. Fossombronia renateae is so far known from only two localities, both in the Afromontane, summer rainfall region of southern Africa, from Lone Creek Falls, near Sabie in Mpumalanga and Xumeni Forest Reserve, near Donnybrook in KwaZulu-Natal (Figure 4), but it must surely be more widespread. At Lone Creek Falls, the plants grow on soil between rocks and are kept moist by spray from the falls. Unfortunately, there are no details about the substrate and conditions on the label of the Doidge collection from Xumeni Forest Reserve. Bryum alpinum Huds. ex With, and Funaria limbata (C.Miill.) Broth, are mosses that grow together with the Lone Creek plants. This newly isolated species has been named F. renateae in honour of Dr Renate Liibenau-Nestle of Kempten, Germany, who collected it when on a visit to South Africa in 1992 and kindly lent her specimens to PRE for exami­ nation. In 1998, accompanied by Ms M. Koekemoer, cura­ tor of PRE herbarium, I visited Lone Creek Falls (and the vicinity) three times. In April we failed to find fertile spec­ imens of this species, but in June we were successful, although the material only had a few dehisced capsules retained from the previous season, and fortunately still with some unshed spores. These spores and elaters are closely similar to those of the Liibenau-Nestle specimens. Samples of the fresh collections with young capsules were cultivated for two months (until August) before ripe spores were obtained. Many of these spores were malformed, but the ornamentation was fairly similar to that of the earlier collections. The elaters, however, were more numerous, usually with 2 (rarely 3) darker and thicker spirals, 45-110 (-132.5) x 10.0-17.5 pm; only a few elaters were identi­ cal to the Liibenau-Nestle and Doidge ones. Scott & Pike (1988) refer to the elaters of F. caledonica Steph. as "bispiral, with the spirals breaking down into rings’ (translated by them from Latin). They found similar elaters in F. papillata (and several other species), where capsules with elaters of this sort and capsules with normal elaters occurred in the same colony. Accordingly, they considered the production of such elaters a mere abnormality not worthy of specific recognition. However, in a genus where there are so few reliable tax­ onomic characters to discriminate between species, I would hesitate to discard this one entirely. The plants collected by Perold & Koekemoer are decidedly more robust and the stems tleshier than in the earlier collections. The species can be distinguished by its undulating leaves, by its often stipitate pseudoperi­ anths, of which the mouth is highly convoluted, by com­ pletely or incompletely reticulate spores and generally by short and relatively wide elaters, the two spirals often forming rings or coils. Fossombronia capensis S.W Arnell from the winter rainfall area in the southern Cape, also has short and stout elaters, usually with loose, pale-coloured spirals (Perold 1997a), but in spore ornamentation it differs completely from F. renateae. Africa. NOTES ON TWO UNNAMED SPECIES FROM MPUMALANGA A new, unnamed species was found beside the Malieveld River, 9 km East of Sabie on the road to Hazyview, Mpumalanga. It has been collected three times, Perold & Koekemoer 3573 , 4053, 4082, in the past few years. This species grows on the concrete sup­ ports of the bridge, just above or at the water level, as well as on the soil of the streambank. Unfortunately, ripe spores have not yet been obtained and it cannot be described at this stage. This species can be recognised by almost entire, rounded, slightly concave and rather stilt, imbricate leaves, 1250-1750 x 1500-2000 pm. They overlap by ± ’/} of their width and are not decurrently inserted on the stem. The shoots are up to 19 mm long, including the apical branches which are 5-7 mm long. The pseudope­ rianth is cupulate, also rather stiff and the margin is undulating. The perigonial bracts shielding the antheridia in the male plants, are conspicuous. Another species of Fossombronia collected at Lone Creek Falls, Perold & Koekemoer 4059, 4060, is very large, with up to 50 mm long shoots. It is an erect-grow­ ing, bottle-green, aquatic plant. Its leaves are sub­ quadrate to rectangular, 2800-3375 x 2725-3200 pm, and only very slightly decurrent at the ‘trailing’ edge, rather stiff, 4-stratose at the base and bistratose up to ± midlength. Unfortunately, only specimens with archego­ nia were found, and the plants kept in cultivation deteri­ orated rapidly. It cannot, therefore, be described at this stage, but it is still worth mentioning, in case fruiting material turns up later. Several more, large, aquatic specimens, without spores have been collected elsewhere in southern Africa. It would appear that fertilisation of these plants in quite rapidly flowing streams, seldom takes place. NOTES ON SOME TROPICAL AFRICAN SPECIES It is unlikely that the above-mentioned plants belong to F. pulvinata, a sterile species described by Stephani Bothalia 29,1 (1999) 81 (1911, 1917) from Rwanda, leg. M ildbraed 22182 (G!). The leaves of F. pulvinata , in contrast to the abovementioned South African plants, are strongly decurrent, much wider than long ( l7 mm lata, medio 4 mm longa’) and rather lax. Scott & Pike (1988) think that this species may be F. australis. Jones (1990) reports N oteroclada porphyrorhiza (Nees) Steph. with purple rhizoids, as once recorded from Kilimanjaro. The pseudoperianth is said to be later­ ally compressed and bilabiate, but spores are not men­ tioned. This species had previously been transferred to Fossombronia porhyrorhiza (Nees) Prosk. by Proskauer (1955). I have not seen this specimen and am unable to comment on it. I have, however, examined a sterile spec­ imen from Lydenburg, Wilms 024671(G), which had been referred to Fossombronia (= Androcryphia) porphyrorhiza Nees. This plant has surely been misidentified, as the type species is from South America, which Scott & Pike (1988) think may also be F australis. Another Fossombronia species described from tropi­ cal Africa on Mt Kilimanjaro in Kenya, is F. grandis Steph.. leg. Volkens 1226 (G 18524) holotype (G !). It is a large plant, as the specific epithet is intended to convey. The shoots are 20 mm long and simple, the leaves large and broadly rounded, the margins with some short, longi­ tudinal folds. Stephani (1900) reported the spores of this dioicous species to be 34 |jm diam.; my measurements are 32.5-37.5 fjm diam. The distal face is coarsely retic­ ulate, with rather thin. 2.5-5.0 pm high ridges delimiting 4 or 5(-6) angular areolae, 7.5-10.0 fjm wide, across the diameter and with ± 12 spinous projections connected by a perispore, at the periphery; the proximal face is gene­ rally concave in the centre, with fine, irregular, sinuating ridges, sometimes forming small areolae, partly or com­ pletely surrounded and often a little obscured by several large, angular areolae, seemingly 'folded' inward from the outer row on the distal face. The elaters are 242-280 x 7.5 pm and often 2-spiral. I have also examined a specimen, Adam & Jager 9211/X trom Mt Elgon National Park, which was kindly lent by Dr ChuahPctiot. It, and a sterile specimen, Pécs 9214/AC had been assigned to F grandis by Prof. T. Poes. Chuah-Petiot (1995) reports F. grandis (Pécs 9236/A A ) from Naro Moru. A specimen. Hedberg 1747h (BOL) from Mount Kenya (at the stream below Tyndall glacier), was deter­ mined by Arnell (1956) as F grandis. It is sterile and the determination is doubtlul. The spore ornamentation in F grandis is so closely similar to that of F. angulosa (Dicks.) Raddi. [said to be a salt-tolerant (Smith 1990). mediterranean-atlantic species (Jovet-Ast 1946)], and also large as well as dioicous, that I am surprised no one else has remarked on it. Scott & Pike (1988) regard both F. angulosa and F. grandis as good species and do not comment on the spore ornamen­ tation. I found the leaves in F. angulosa to be mostly longer than wide and shortly lobed above. In F. grandis the leaves are generally wider than long, the apex round­ ed and, at the base, several cell layers thick, the latter also found to be the case in F. angulosa. In both species the pseudoperianths are stipitate. Despite the similarities between these two species, it seems phytogeographically unlikely that F. grandis would be conspecific with F. angulosa. Moreover. I have examined only two sporulating specimens of F. grandis. the type specimen and the Adam & Jager collection, which makes me hesitate to place this species in synonymy under F. angulosa. Krauss (1846) collected a specimen of the so-called F. angulosa from southern Africa 'in rupibus rivulor natalensium', but this has not been confirmed and is thought to have been a misidentification of the commonly occurring F. crispa Nees (= F. zeyheri Steph.) (Perold 1997b, c).

Fossombronia renateae is so far known from only two localities, both in the Afromontane, summer rainfall region of southern Africa, from Lone Creek Falls, near Sabie in Mpumalanga and Xumeni Forest Reserve, near Donnybrook in KwaZulu-Natal (Figure 4), but it must surely be more widespread. At Lone Creek Falls, the plants grow on soil between rocks and are kept moist by spray from the falls. Unfortunately, there are no details about the substrate and conditions on the label of the Doidge collection from Xumeni Forest Reserve. Bryum alpinum Huds. ex With, and Funaria limbata (C.Miill.) Broth, are mosses that grow together with the Lone Creek plants.
This newly isolated species has been named F. renateae in honour of Dr Renate Liibenau-Nestle of Kempten, Germany, who collected it when on a visit to South Africa in 1992 and kindly lent her specimens to PRE for exami nation. In 1998, accompanied by Ms M. Koekemoer, cura tor of PRE herbarium, I visited Lone Creek Falls (and the vicinity) three times. In April we failed to find fertile spec imens of this species, but in June we were successful, although the material only had a few dehisced capsules retained from the previous season, and fortunately still with some unshed spores. These spores and elaters are closely similar to those of the Liibenau-Nestle specimens. Samples of the fresh collections with young capsules were cultivated for two months (until August) before ripe spores were obtained. Many of these spores were malformed, but the ornamentation was fairly similar to that of the earlier collections. The elaters, however, were more numerous, usually with 2 (rarely 3) darker and thicker spirals, 45-110 (-132.5) x 10.0-17.5 pm; only a few elaters were identi cal to the Liibenau-Nestle and Doidge ones. Scott & Pike (1988) refer to the elaters of F. caledonica Steph. as "bispiral, with the spirals breaking down into rings' (translated by them from Latin). They found similar elaters in F. p a pillata (and several other species), where capsules with elaters of this sort and capsules with normal elaters occurred in the same colony. Accordingly, they considered the production of such elaters a mere abnorm ality not worthy of specific recognition. However, in a genus where there are so few reliable tax onomic characters to discriminate between species, I would hesitate to discard this one entirely.
The plants collected by Perold & Koekemoer are decidedly more robust and the stems tleshier than in the earlier collections. The species can be distinguished by its undulating leaves, by its often stipitate pseudoperi anths, of which the mouth is highly convoluted, by com pletely or incompletely reticulate spores and generally by short and relatively wide elaters, the two spirals often forming rings or coils.
Fossombronia capensis S.W Arnell from the winter rainfall area in the southern Cape, also has short and stout elaters, usually with loose, pale-coloured spirals (Perold 1997a), but in spore ornamentation it differs completely from F. renateae.

NOTES ON TWO UNNAM ED SPECIES FROM MPUMALANGA
A new, unnamed species was found beside the Malieveld River, 9 km East of Sabie on the road to Hazyview, Mpumalanga. It has been collected three times, P erold & Koekem oer 357 3 , 4053, 4082, in the past few years. This species grows on the concrete sup ports of the bridge, just above or at the water level, as well as on the soil of the streambank. Unfortunately, ripe spores have not yet been obtained and it cannot be described at this stage.
This species can be recognised by almost entire, rounded, slightly concave and rather stilt, imbricate leaves, 1250-1750 x 1500-2000 pm. They overlap by ± '/} of their width and are not decurrently inserted on the stem. The shoots are up to 19 mm long, including the apical branches which are 5-7 mm long. The pseudope rianth is cupulate, also rather stiff and the margin is undulating. The perigonial bracts shielding the antheridia in the male plants, are conspicuous.
Another species of Fossombronia collected at Lone Creek Falls, P erold & K oekem oer 4059, 4060, is very large, with up to 50 mm long shoots. It is an erect-grow ing, bottle-green, aquatic plant. Its leaves are sub quadrate to rectangular, 2800-3375 x 2725-3200 pm, and only very slightly decurrent at the 'trailing' edge, rather stiff, 4-stratose at the base and bistratose up to ± midlength. Unfortunately, only specimens with archego nia were found, and the plants kept in cultivation deteri orated rapidly. It cannot, therefore, be described at this stage, but it is still worth mentioning, in case fruiting material turns up later.
Several more, large, aquatic specimens, without spores have been collected elsewhere in southern Africa. It would appear that fertilisation of these plants in quite rapidly flowing streams, seldom takes place.

NOTES ON SOME TROPICAL AFRICAN SPECIES
It is unlikely that the above-mentioned plants belong to F. pulvinata, a sterile species described by Stephani (1911Stephani ( , 1917 from Rwanda, leg. M ildbraed 22182 (G!). The leaves of F. p u lvin ata , in contrast to the abovementioned South African plants, are strongly decurrent, much wider than long ( l7 mm lata, medio 4 mm longa') and rather lax. Scott & Pike (1988) think that this species may be F. australis. Jones (1990) reports N oteroclada porph yrorh iza (Nees) Steph. with purple rhizoids, as once recorded from Kilimanjaro. The pseudoperianth is said to be later ally compressed and bilabiate, but spores are not men tioned. This species had previously been transferred to Fossombronia porhyrorhiza (Nees) Prosk. by Proskauer (1955). I have not seen this specimen and am unable to comment on it. I have, however, examined a sterile spec imen from Lydenburg, Wilms 024671(G ), which had been referred to Fossombronia (= Androcryphia) porphyrorhiza Nees. This plant has surely been misidentified, as the type species is from South America, which Scott & Pike (1988) think may also be F australis.
Another Fossombronia species described from tropi cal Africa on Mt Kilimanjaro in Kenya, is F. grandis Steph.. leg. Volkens 1226 (G 18524) holotype (G !). It is a large plant, as the specific epithet is intended to convey. The shoots are 20 mm long and simple, the leaves large and broadly rounded, the margins with some short, longi tudinal folds. Stephani (1900) reported the spores of this dioicous species to be 34 |jm diam.; my measurements are 32.5-37.5 fjm diam. The distal face is coarsely retic ulate, with rather thin. 2.5-5.0 pm high ridges delimiting 4 or 5(-6) angular areolae, 7.5-10.0 fjm wide, across the diameter and with ± 12 spinous projections connected by a perispore, at the periphery; the proximal face is gene rally concave in the centre, with fine, irregular, sinuating ridges, sometimes forming small areolae, partly or com pletely surrounded and often a little obscured by several large, angular areolae, seemingly 'folded' inward from the outer row on the distal face. The elaters are 242-280 x 7.5 pm and often 2-spiral. I have also examined a specim en, Adam & Ja g er 9211/X trom Mt Elgon National Park, which was kindly lent by Dr Chuah-Pctiot. It, and a sterile specimen, P écs 9214/AC had been assigned to F grandis by Prof. T. Poes. Chuah-Petiot (1995) reports F. grandis (Pécs 9236/A A ) from Naro Moru. A specimen. H edberg 1747h (BOL) from Mount Kenya (at the stream below Tyndall glacier), was deter mined by Arnell (1956) as F grandis. It is sterile and the determination is doubtlul.
The spore ornamentation in F grandis is so closely similar to that of F. angulosa (Dicks.) Raddi. [said to be a salt-tolerant (Smith 1990). mediterranean-atlantic species (Jovet-Ast 1946)], and also large as well as dioicous, that I am surprised no one else has remarked on it. Scott & Pike (1988) regard both F. angulosa and F. grandis as good species and do not comment on the spore ornamen tation. I found the leaves in F. angulosa to be mostly longer than wide and shortly lobed above. In F. grandis the leaves are generally wider than long, the apex round ed and, at the base, several cell layers thick, the latter also found to be the case in F. angulosa. In both species the pseudoperianths are stipitate. Despite the similarities between these two species, it seems phytogeographically unlikely that F. grandis would be conspecific with F. angulosa. Moreover. I have examined only two sporulating specimens of F. grandis. the type specimen and the Adam & Jager collection, which makes me hesitate to place this species in synonymy under F. angulosa. Krauss (1846) collected a specimen of the so-called F. angulosa from southern Africa 'in rupibus rivulor natalensium', but this has not been confirmed and is thought to have been a misidentification of the commonly occurring F. crispa Nees (= F. zeyheri Steph.) (Perold 1997b, c).