Studies in the liverwort genus Fossombronia ( Metzgeriales ) from southern Africa . 4 . A re-examination of F . crispa , F . leucoxantha and F . tumida

The above three southern African species were described during the nineteenth century, but the descriptions are brief and mostly inadequate. The practice, albeit for justified financial reasons, of dividing collections (which were often mixed) and of distributing the parts to different herbaria, has led to some later authors applying the specific epithets to the wrong com­ ponents. This has caused a great deal o f confusion, and much time and effort have been expended in sorting out these mis­ applications. This paper aims to correct and to report on some past mistakes. Detailed descriptions, illustrations and a dis­ tribution map of these three species are accordingly provided. 1. Fossom bronia crispa Nees in Synopsis Hepaticarum, Gottsche et al.: 469 (1846). Type: Cap. de b. Spei, leg. presumably Zeyher, no collector’s name on label, only the number 3, probably added later (STR, lecto.! fide Perold 1997d). Jungermannia crispa Sprengel in schedule Herb. Zeyher. J. pusilla Lehm. : 369 no. 42, Ecklon (BM!). Fossombronia zeyheri Steph.: 32 (1900); Sim: 35 (1926); S.W. Amell: 79 (1963); Sergio: 191 (1985). Type.— Cap. b. spei, leg. Zeyher s.n. G 024669(G!) ex Herb. Rabenhorst, idem ex Herb. ( l)o p e , G024670(G\), portion(s) o f type, sub F. crispa. Plants in dense colonies, pale green or bright green to glaucous green; medium-sized to robust; shoots simple, (8—) 12—15 mm long, 1.5-2.0 mm high, 3.2-3.4 mm wide, or once/twice or repeatedly furcate, terminal seg­ ments moderately divergent, 3.0-5.0 mm long. Stems prostrate, chlorophyllose, sometimes outer cells purple, occasionally with a lateral bud, later developing into a side shoot, plano-convex in cross section, apically fre­ quently swollen, 320-350 pm (13 cell rows) high, 500-540 pm wide (Figure 1G), tapering proximally and toward base (Figure 1H), 220-230 pm high, 270-300 pm wide. Rhizoids purple, 12.5-22.5 pm wide. Leaves (Fig­ ure 1A-F) overlapping, widely spreading to suberect (Figure 2A), slightly undulate to crispate, succubously inserted, apically small, soon enlarging, oblong to irreg­ ularly rectangular, somewhat longer than wide or occa­ sionally as long as wide or even shorter than wide; 1450-1950 x 1250-2000 pm, often narrower below, 750-1375 pm; apex rounded or truncate, sometimes 2 or 3 times slightly notched and shallowly lobed, margins (Figure II) entire, but with 7-14 unicellular slime papil­ lae, generally well spaced, but at proximal (trailing) edge closer together and occasionally raised on a basal cell. Leaf cells thin-walled, at upper margins mostly rectangu­ * National Botanical Institute, Private Bag X101, Pretoria 0001. MS. received: 1996-11-05. l r across. 10.0-37.5 x 42.5-75.0 |jm, at lateral margins long-rectangular, 67.5-80.0 x 15.0-22.5 pm, upper 1aminal cells 5or 6-sided, 50.0-57.5 x 37.5-52.5 pm, mid­ dle laminal cells 87.5-107.5 x 35.0-45.0 pm, basal cells 120.0-137.5 x 30.0-47.5 pm. Oil bodies variable in number, 9^45 per cell, round, smooth or granular, up to 5 pm in diameter (Figure 1J); chloroplasts numerous, crowded, ± 5 pm in diameter (Figure 1J). Monoicous (Figure 2B. C), sometimes seemingly dioicous. Antheridia short-stalked, globose, dorsal on stem, sometimes crowded together near apex of shoot, interspersed with the younger archegonia and apparently naked (Figure 2B), at other times proximal to pseudope­ rianth (or on different plants) and subtended by perigonial bracts (Figure 2D), their shape irregular (Figure 1K-N), 400-850 x 250-450 pm wide across broadest part, mostly narrower toward apex and base, margins with up to 7 papillae or with 1 or 2 short, finger-like processes, cells in interior 5or 6-sided, 112.5-117.5 x 30.0-42.5 pm, marginally 32.550.0 x 20.0-50.0 pm. Archegonia in 1 or 2 irregular rows along stem, even ex­ tending toward base (Figure 2A), naked, sometimes 2 in close proximity becoming fertilized (Figure 2E). Pseudoperianth campanulate (Figure lO , P), at apex of branch or close to it, as tall as, or up to 375 pm taller than leaves, raised on a short stalk, 375-500 pm long, 550-600 pm wide, then widely flaring upwards, 1625-2150 pm long, 2500-2925 pm wide across mouth, margin with ± 5 main undulating lobes (Figure 2F), gen­ erally subdivided into smaller ones, at side mostly cleft once to base, often with lamellar outgrowths. 1125-1375 x 375 pm; cells not appreciably different from those of leaves. Capsules globose, ± 550 pm in diameter, wall bistratose, cells of inner layer irregularly shaped, 25.0-75.0 x 20.0-35.0 pm, each cell wall with 3 or 4 dark brown nodular and occasionally semi-annular thickenings (Figure 1R). Seta 2.9-4.'4 mm long, ± 150 pm in diameter, ± 7 cell rows across (Figure IQ). Spores light brown, 37.5-42.5 pm in diameter, including lamelFIGURE 1.— Fossombronia crispa\ A-F, leaves; G, cross section of stem apex; H, cross section of stem base; I, detail o f leaf margin; J, median leaf cells with oil bodies (solid lines) and chloroplasts (dotted lines); K-N, bracts; O, opened pseudoperianth; P, pseudoperianth from side; Q, cross section of seta; R, cells in capsule wall. A, C-F, K-P, S. Strauss 134a; B, I, J, S. Strauss CH13664\ G, H, R, Perold & Van Rooy 3558\ Q. J. Victor 1379. Scale bars: A-F, O, P, 500 pm; G, H, K-N, 250 pm; 1, J, R, 50 pm; Q, 100 pm. Artist; G. Condy. Bothalia 27,2 (1997) 107 FIGURE 2.— Fossombronia crispa. A, simple stem, leaves widely spreading; B, apex of segment with crowded antheridia and archegonia, more distally; C, apex of stem with naked antheridia, archegonia and a bract; D, pseudoperianth near apex of stem and bracts more proximally; E, two adjacent pseudoperianths; F, pseudoperianth from above. A, E, S.M. Perold 3444\ B, Perold & Koekemoer 3282\ C, S.M. Perold 3280 p.p.; D, F, S. Strauss 134a. A, x 7; B, x 13; C, x 41; D, x 15; E, x 8; F, x 19. lae projecting around periphery; hemispherical; distal face (Figure 3A, B) convex, ornamentation over polar area reticulate or partly reticulate, low, slightly wavy lamellae forming ± 6 complete and incomplete, irregular areolae across diameter of face (a total of 15-24), each 7.5-12.5 pm wide, sides of lamellae with a few faint buttressing striations that soon disappear, spore surface between lamellae granular (Figure 3C); proximal face lacking triradiate mark, flat, covered with fine, low ridges forming very small, irregular areolae, with papil­ lae here and there (Figure 3D), otherwise very coarsely granular, toward centre of face granules coalescing into large, irregular clumps (Figure 3E), around spore pe­ riphery 14-18 low ‘spines’, which are the ‘ends’ of the lamellae from the distal face extending over the sides and are connected by a low, much interrupted, membra­ nous wing or perispore. Elaters (Figure 3F) yellowbrown, 60.0-122.5 |jm long, 7.5-12.5 |im wide in cen­ FIGURE 3.— Fossombronia crispa. A -E , spores: A, distal face; B, side view of distal face; C, detail of lamellae and surface on distal face; D, E, proximal face. F, elater. A, D, S.M. Perold 3317\ B, S.M. Perold 3280 p.p.; C, S. W. Arnell 2201’, E, S.M. Perold 3444\ F, Perold & Van Rooy 3558. A, x 725; B, x 859; C, x 2596; D, x 705; E, x 685; F, x 852. 108 Bothalia 27,2 (1997) tre, tapering to one or both tips and ending in a 5 pm wide loop, or not tapering, ends blunt, with 2 or 3 loose spirals or tightly coiled. Fossombronia crispa grows on soil at river banks, seepage areas and at road sides. Although fairly fre­ quently sterile, it is, nevertheless, the most common southern African species of Fossombronia and its range extends from the winter rainfall area of Western Cape to the summer rainfall areas of Eastern Cape, KwaZuluNatal, Mpumalanga and Northern Province (Figure 4). When fertile, the species bears ripe sporangia during most of the year in the summer rainfall areas. Since F. crispa is so common, it is clear why it was first described by the mid 1800’s and not as late as 1900 by Stephani under the epithet, F. zeyheri. Fossombronia crispa is distinguished by its entire, crisped leaves, its frequently robust size, generally purple stem, pseudoperianths with lobed, undulating mouth, reticulate or incompletely reticulate spores with low lamellae and elaters with strong spirals. Although there is some similarity in their spores, the elaters of F. crispa are clearly different from those of typical F. capensis (see Perold 1997b), which are short, have weak spirals and collapse upon drying. It is shown in Perold (1997d) that F. crispa Nees, as described in the protologue, definitely has entire leaves. The spores, as was later found, are reticulate or incom­ pletely reticulate but certainly not spinose. This species was subsequently redescribed by Stephani (1900) as F. zeyheri (sub F. crispa), and was considered by Scott & Pike (1987a) to be part of the widely distributed and later described, F.foveolata complex, because of similarities in spore ornamentation. By inference F. crispa would also be part of this complex. It was, however, described 30 years before F. foveolata and the specific name, F. crispa, would therefore have priority. In general, F. crispa is a larger plant than F. foveolata and I continue to treat them as distinct taxa, because this study is as yet confined to southern African species. Moreover, the specific epithet, F. foveolata, has a long history of acceptance. The epithet, F. crispa, has, however, since Stephani (1900) generally been misapplied to plants with dentate FIGURE 4.— Distribution of F. crispa. • ; F. leucoxantha. □ ; and F. tu­ mida. O , in southern Africa. leaves and spinose spores (i.e. F. leucoxantha). For some reason Scott & Pike (1988) regarded the BM specimen, (top right comer of herbarium sheet, Perold 1997d: fig. 6C), as the type of F. crispa. They referred to it as ‘Cape of Good Hope, leg. Zeyher, sub ‘Jungermannia crispa’ Sprengel, BM (hb. Hampe. sin. num. 14/8/1825)’ and an­ notated it on the sheet as 1:3. As I learnt from the label, this specimen is most likely an Ecklon collection, no

Fossombronia crispa grows on soil at river banks, seepage areas and at road sides.Although fairly fre quently sterile, it is, nevertheless, the most common southern African species of Fossombronia and its range extends from the winter rainfall area of Western Cape to the summer rainfall areas of Eastern Cape, KwaZulu-Natal, M pumalanga and Northern Province (Figure 4).When fertile, the species bears ripe sporangia during most of the year in the summer rainfall areas.Since F. crispa is so common, it is clear why it was first described by the mid 1800's and not as late as 1900 by Stephani under the epithet, F. zeyheri.Fossombronia crispa is distinguished by its entire, crisped leaves, its frequently robust size, generally purple stem, pseudoperianths with lobed, undulating mouth, reticulate or incompletely reticulate spores with low lamellae and elaters with strong spirals.Although there is some similarity in their spores, the elaters of F. crispa are clearly different from those of typical F. capensis (see Perold 1997b), which are short, have weak spirals and collapse upon drying.
It is shown in Perold (1997d) that F. crispa Nees, as described in the protologue, definitely has entire leaves.The spores, as was later found, are reticulate or incom pletely reticulate but certainly not spinose.This species was subsequently redescribed by Stephani (1900) as F. zeyheri (sub F. crispa), and was considered by Scott & Pike (1987a) to be part of the widely distributed and later described, F.foveolata complex, because of similarities in spore ornamentation.By inference F. crispa would also be part of this complex.It was, however, described 30 years before F. foveolata and the specific name, F. crispa, would therefore have priority.In general, F. crispa is a larger plant than F. foveolata and I continue to treat them as distinct taxa, because this study is as yet confined to southern African species.Moreover, the specific epithet, F. foveolata, has a long history of acceptance.
Archegonia in 1 or 2 rows along stem, hidden by leaves, sometimes 2 in close proximity becoming fertilized.
Pseudoperianth (Figure 6E, F) campanulate, proximal to apex, as tall as leaves or projecting somewhat above them, raised on a short stalk, then widely flaring above, 1500-2250 pm long, 750-1500 pm wide at base, 2375-4375 pm wide across mouth, margin much 'ruched', with several 'folds' and up to 39 toothed processes (Figure 5P, Q), 3 -7 cells or 100-325 pm long, topped by a slime papilla, ± 12.5 x 15.0 pm and gradually widening below; cells comparable in shape and size to those of leaves.
Capsules globose, 800-1050 (jm in diameter, wall bistratose, cells of inner layer irregularly shaped (Figure 5R), 30.0-62.5 x 32.0-50.0pm, each cell wall with 2 or 3 dark brown, nodular, and sometimes semi-annular thickenings.Seta 0.2-4.0mm long, ± 220 (am in diame ter, 7 or 8 cells across (Figure 5S).Spores light brown, 42.5-55.0 in diameter, including spines projecting at margin (Figure 7D); hemispherical; distal face (Figure 7A -C ) convex, with 12-15 truncate or conical spines up to 5 pm long, in irregular rows across diameter of spore, some confluent to form short ridges and mostly intercon nected by small buttressing ridges that radiate from the bases of adjacent spines, often with several smallish papillae interspersed between spines and ridges; proxi mal face (Figure 7E) with triradiate mark distinct or lack ing, flat, ornamentation variable, from numerous, rather fine papillae to coarse tubercles, often interspersed with short, irregular ridges, around spore periphery with up to 30 or occasionally more spines seen in profile or broadon.Elaters yellow-brown, 130-202 pm long, 7.5 pm wide in centre, tapering to tips, 2.5-5.0 pm wide, some times surface sprinkled with fine papillae (Figure 7F), bispiral or trispiral.
Fossombronia leucoxantha grows on damp, rather coarse to clayey soil at various places in the Peninsula and Western Cape, i.e.Bakoven, Bot River, Cave Peak, C hapm an's Peak, Constantia Slopes, D evil's Peak, Genadendal, Karweiderskraal, Kirstenbosch Gardens, Kloofnek, between Kloofnek and Round House, Lion's Head, Round House, Signal Hill, Table Mountain and Wynberg, Cape Town (Figure 4).In this rather restricted area the species is quite common.It is distinguished by its 'ruched' leaves, of which the margins, as well as the rim of the mouth of the pseudoperianth are denticulate to incised-dentate.There are a few specimens that have almost entire leaves, notably Esterhuysen 24885 and Artiell 616, that probably do not belong here in spite of having spinose spores.The spores of F. leucoxantha are densely spinose, but often have papillae or short ridges between the spines.This species is distinguished from F. gletiii (Perold 1997a) which also has spinose spores, but is restricted to the summer rainfall area in Northern Province, Gauteng and Mpumalanga and its leaves are angular, and the rather small pseudoperianth is divided into deep lobes.
Fossombronia leucoxantha was described by Lehmann (1829) from an Ecklon collection.This Ecklon specimen was mixed with a tumid-leaved plant, later de scribed by Mitten (1878) as F. tumida from a collection by Rev. Eaton at the Cape.Stephani's leones (1985), nos.3044 and 3045, illustrate F. tumida, but at the bottom right corner they bear the epithet, F. leucoxantha.In a note under his description of F. leucoxantha, Stephani (1900) referred to its leaves as 'aufgeblasen', which could only apply to F. tumida.Subsequently, Sim (1926) complained that Stephani 'mentions it [meaning F. tumi da] near F. leucoxantha, which is not its place'.made no further reference to the former in his book, al though there are specimens held at BOL, PRE and S that he had identified as F. leucoxantha.Scott & Pike (1987b) have already drawn attention to the confusion between F. leucoxantha and F. tumida.They also state that 'A rnell's own specimens which he had identified as F. leucoxantha 'L.& L \ are mostly F. spinifolia St.'.This assumption is incorrect: they had no knowledge of the spore morphol ogy of F. spinifolia, as they had overlooked a capsule with ripe spores in the type specimen.Fossombronia spinifolia has spores with short irregular lamellae and very few spines.It also is a small plant and has been treated in a previous paper in the current series (Perold 1997c).Scott & Pike (1988)  Their spore micrographs of the lectotype of F. leucoxan tha (Scott & Pike 1987b: figs 3 & 4) depict spines and some ridges and are marginally different from mine (see Figure 7 A -E in this paper), but are nevertheless still within the acceptable range of variation in spore orna mentation that I found in the many spore micrographs that I took of this species.
Stephani (1900) recognized three southern African species with spinose spores, F. spinifolia, the so-called F. crispa and F. leucoxantha.He did not have the opportu nity to examine spore-bearing material of F. leucoxantha and F. tumida which he confused with each other; Sim (1926) recognized two species with spinous spores, F. crispa with 'long papillae' and F. leucoxantha with 'short papillae' in which he followed Stephani;Arnell (1963) placed F. leucoxantha (sensu) Steph. in synonymy under F. tumida, treating F. crispa as the only species with spin ous spores; Sergio (1985) accepted F. crispa as having spinose spores and F. zeyheri as having reticulate spores; Scott & Pike (1988) retained F. crispa as a species with spinose spores and thought that Arnell's earlier determi nations of F. leucoxantha were actually F. spinifolia.
Fossombronia tumida grows on damp, rather sandy soil at L ion's Head, the Round House, Kloofnek, Stellenbosch Flats and at Roman River, south of Wolseley (Figure 4).A specimen, Wager 39 (CH 3703) is said to be from East London, but this locality could be wrong.Sim (1926) observed that it is found in S.W. (meaning Western Cape) localities, 'but not seen east ward'.This species is by no means common.
Fossombronia tumida is distinguished by its inflated or tumid appearance and very concave leaves, with small, scattered red flecks containing round bodies.The spores are generally reticulate over the distal pole, and the outer areolar walls break down into lamellae which continue to the margin.Scott & Pike (1987b) regard F. tumida as very similar vegetatively to F. intestinalis Tayl.from Australia and Tasmania, but 'with experience, however, they can be separated.The former is larger, with flaccid leaves often tinged with purple-brown and with entire instead of slightly denticulate leaves.The spores are quite distinct'.Scott (1985) claims that F. intestinalis also occurs in southern Africa, but does not cite a particular specimen.I have not found it so far.The confusion between F. tu mida and F. leucoxantha is discussed above under the lat ter species.

SPECIMENS EXAMINED
With one exception, only fertile specimens are includ ed; held at PRE, unless otherwise indicated.