THE AFRICAN SPECIES OF DREPANOLEJEUNEA VESICULOSA GROUP WITH DESCRIPTION OF DREPANOLEJEUNEA VANDERPOORTENII SPEC. NOVA (JUNGERMANNIOPSIDA) FROM MADAGASCAR

Drepanolejeunea clavicornis and D. friesii were previously synonymised with D. physaefolia or all of them with D. vesiculosa. In the meantime, Drepanolejeunea vandenberghenii was described from the same species group, as new. In this paper many African specimens are compared with the original descriptions of the above species. Morphological investiga-tions of these and their distributional patterns suggested that the former synonymisation was not justified. In addition, a new, rheophytic species from the same group: Drepanolejeunea vanderpoortenii , is described, as new to science. As a result, from the taxa related to Drepanolejeunea vesiculosa , now six species are recognised from Africa, including its Indian Ocean islands. For these 6 morphotaxa an identification key is provided. The results need confirmation by a future molecular analysis.


INTRODUCTION
I already observed during our first collecting trip to Madagascar (March 1990) that most specimens identified as Drepanolejeunea physaefolia (Gottsche) Steph. or more properly D. physifolia (Gottsche) Pearson differ from the specimens collected in continental East Africa under the name of D. clavicornis Steph. and D. friesii Vanden Berghen, which were synonymised later with D. vesiculosa (Mitt.) Steph. (Tixier 1995, Wigginton andGrolle 1996) or with D. physifolia (Jones ex Wigginton 2004, formally not yet done). The main difference of Madagascan specimens is their dentate, sometimes serrulate leaf border and their ciliato-dentate perichaetial leaves.
I came to the conclusion, that the former synonymisations of the above Drepanolejeunea species were not well-founded. To clarify the matter, I started to study the more than 60 related African Drepanolejeunea specimens from our herbarium (EGR) and compare them with the original descriptions, illustrations and type specimens. I could successfully classify them into the species originally distinguished as Drepanolejeunea clavicornis, D. friesii, D. physifolia, D. vesiculosa and to the not long ago described D. vandenberghenii. At the same material they are still visible, in all species of this group 2-3 (to -5) suprabasal ocelli can form a short vitta. But the perianth shape remains as a useful distinguishing feature, as it is so strikingly different. Jones & Harrington (1983) mention another difference, apart from the perianth, in the shape of female bracts, which has only a few coarse teeth in D. friesii and laciniately dentate in D. clavicornis. They only tentatively identified their specimens from Sierra Leone and from Ghana, as D. friesii. The specimens examined by me from a great part of East Africa confirm their opinion. I have found on several specimens the same perianth type and perichaetial leaf shape, as described and illustrated by them. Buchbender and Fischer (2004: 279, fig. 5A, B, reproduced on our Figs 9-11 and 14) also published good illustrations of Drepeanolejeunea friesii (under the name of D. physaefolia). The differences from the real D. physifolia will be discussed under the latter species.
Drepanolejeunea friesii is the most widespread species of the group related to D. vesicaria all over the mountainous areas of continental Africa, and sporadically also on the Indian Ocean Islands (Fig. 15). It occurs on very different substrates, but prefers living leaves, ericaceous bark and twigs. It is published from Sierra Leone: Loma Mts., on sheltered boles, 1,680 m, Jones 1474p.p., 1477p.p., 1650m, Jones 1491(Jones and Harrington 1983   Lejeunea physifolia (originally described, as "physaefolia") sensu stricto is a species different from the previous taxa described from the African mainland and Sao Tomé Island. As the type material would have been destroyed during the 2nd World War in B (Grolle 1995) we can only relay on Stephani's Icones (1985, figs 002450 and 002451), which depict both Drepanolejeunea physaefolia and D. securifolia based most probably on Gottsche's types, as in Stephani's time no other specimens were known. The illustration of D. intorta (Stephani 1985, fig. 002448, copied to our Figs 16 and 18) confirms Grolle's concept synonymising it with D. physifolia. Both the Stephani Icones (1985) and the pictures by Arnell (1963: 181, fig. 132) illustrate the same species: a plant with obviously dentate lobe margin and with pinnately ciliato-laciniate perichaetial leaves. In fact, the acute papillae are turning outwards at the lobe margin forming acuminate teeth. Even the perianth keels are finely dentate as can be seen both on Arnell's picture and on our microphoto (Figs 19-21 and 23-24). On the contrary, the lobe margins of Drepanolejeunea friesii (and of D. vesiculosa) are smooth or only minutely crenulate and the perichaetial leaves and perianth keels are almost entire or have a few coarse teeth . Concerning its distribution, D. physiformis s. str. in the original sense was published only from Southeast Africa and the Indian Ocean islands: Mozambique, Ngorongosa Mts. (Arnell 1957). South Africa: Transvaal, Pietersburg (now Polokwane); Basutoland, Woodbush; Natal, Ngome Forest (Arnell 1963). Comoros, Ndzouani island (Johanna or Anjouan, Herzog 1947); Ngazidja and Dzouani (Pócs 1993(Pócs , 1995. Madagascar: type of Lejeunea physaefolia Gottsche (1882), Pearson (1892), and of L. securifolia (Grolle 1995), Pócs (1997). Réunion: Bescherelle and Spruce (1890) as type of Lejeunea intorta, W of Plaine des Palmistes, forest with tree ferns, above St. Paul, ca 2,000 m (Arnell 1965); Mauritius: Perrier Nature Reserve near Mare aux Vacoas; Le Pouce, rocks on old way between St. Pierre and Port Louis, ca 1,800 ft; Mt Cocotte, the top; the plateau N of Mt Cocotte (Arnell 1965). Seychelles: Mahé Island 100-800 m, not rare (Arnell 1957, Norkett, Onraedt in Grolle 1978; Praslin Island (Norkett in Grolle 1978).
Specimens seen and revised in our herbarium (EGR, duplicates of Madagascar specimens are deposited in TAN): I have found sporadic occurrence also in the mountains of East Africa, similarly to many other Lemurian flora This species can be separated from Drepanolejeunea friesii and of D. vesicaria by its widely obtusely rounded leaf and underleaf lobe apices and by its fully edentate bracts. It was described from dripping rocks, growing among Sphagnum and Breutelia, accompanied by detailed illustrations. The detailed and well-illustrated protologue describes 5 other paratypes near to the type locality collected from different substrates in Nyungwe Forest by Buchbender, Fischer, Solga and Pócs. Later on it was reported also from Malawi (Pócs 2008 Wigginton and Grolle (1996: 73), insufficiently justified, based on Tixier's synonymisation (Tixier 1995). Although only Drepanolejeunea friesii is similar in all other aspects, Drepanolejeunea vesiculosa has a character, the caduceus leaves, which do not occur in the majority of species described from Africa. But it is still a problem, whether this character is stable or not by the type and by all Asian specimens of D. vesicu- losa. If any, then only Drepanolejeunea friesii of the African species can be synonymous with it. This finally can be decided after molecular investigation will have been done on many populations (not on one or two specimens).  Gradstein et Pócs, Close to D. vandenberghenii, but well different by its blackish pigmentation, julaceous habit, thick-walled stems and, especially, the large preapical tooth almost touching the apical (= first) lobule tooth. The blackish pigmentation of the plants and the thick-walled stems may be interpreted as adaptations to the rheophytic habitat (Gradstein andVidal 1975, Pócs 2010).

DISCUSSION
The group of taxa I summed up under the name of 'Drepanolejeunea vesiculosa group' exemplifies very well the problem arising from the difficulty to distinguish infraspecific variability (often influenced by environmental conditions) from the species level of genetical isolation. We have to take in account the morphological malleability, the distribution pattern and the cryp- tic species formation. Doubtless, molecular investigation is an answer, but only, if before the DNA investigation and its statistical analysis we try to define well the units to be investigated. I tried to do this, by the help of careful morpho-taxonomic and area-geographical investigation of the primarily described species before their premature and unjustified synonymisation. A future molecular analysis can prove the rightness of the formerly defined species concepts.
In case of Drepanolejeunea clavicornis I am quite sure about the good definition of this species, being allopatric and having a very different perianth shape. The latter property is not so much exposed to the environmental conditions than the vegetative body, therefore I attach great significance to it.
Drepanolejeunea physifolia in strict sense is more or less an allopatric vicariant of D. vesiculosa or D. friesii. It probably evolved when Madagascar was still united to mainland Africa or isolated later in the Indian Ocean islands and reached the continent by air dispersal. It is an interesting fact, that most plants are female with perichaetium, very rarely fertilised, as perianths and sporophytes are extremely rare.
Drepanolejeunea vanderpoortenii is an allopatric and at the same time a peripatric species. The niche isolation seems to act very well, because rheophytic life form is not known among other Drepanolejeunea species. At the same time this taxon occurs only in Madagascar and has two morphological properties (lobule pre-tooth and blackish pigmentation) unknown by its relatives.
The cases of Drepanolejeunea friesii and Drepanolejeunea vandenberghenii are not so clearcut. They are sympatric and their morphological differences against the very widespread Drepanolejeunea vesiculosa are not so sharp and some transitions seem to occur between them. In their case indeed only molecular investigation can clarify if they can be distinguished at the species level or not. The distribution of these three species, if they are really distinct, in Africa is still inadequately known.
Identification key Based on the above morphological characters, I composed an identification key for the concerned taxa: 1a Perianth inflated, terete with 4-5 globose horns. Leaf margin crenulated. West African species. In all other aspects similar to Drepanolejeunea friesii D. clavicornis 1b Perianth in its upper half with five equal keels each ending up in an apiculate or rounded rostrum 2 2a Lobe margin with leaf margin dentate by the outward turning, high, acute papillae on (almost) each marginal cells. Female perichaetium leaves pinnately ciliato-laciniate, with teeth in length reaching or exceeding the half width of the bract or bracteole and ending in 2-3 uniseriate cells. Perianth keels regulately dentate at least on their rostrum.

D. physifolia
2b Lobe margin entire or finely crenulated, acute papillae restricted to the lobe surface. Female perichaetium leaves almost entire or coarsely dentate with teeth shorter than half width of the bracts or bracteoles and not ending in uniseriate ciliae. Perianth keels entire or coarsely dentate 3 5a Plants yellowish to light brownish-green, without blackish pigmentation. Stem with moderately thicked, colourless cell walls. Leaves imbricate, at apex often incurved, but the shoot never julaceous. Underleaves trapezoid in shape. Only the first lobular tooth is well developed, acute, falcate, one celled. There is no pretooth present. Although prefers wet habitat, never occurs in running water D. vandenberghenii 5b Plants with blackish pigmentation, sepia or dark rust brown. Stem with incrassate (2 µm), brownish pigmented walls. Leaves densely imbricate, fully incurved towards the stem, forming julaceous habit. The apical (first) lobule tooth 1 celled, falcate. On each lobule a 20-25 µm large, almost globose pretooth is attached to the last cell at the distal end of lobule margin, almost touching the apical tooth. Rheophytic species, occurring on stones in running water or on irrigated streambank D. vanderpoortenii *