Saxicolous species of the genus Rinodina ( lichenized Ascomycetes , Physci-aceae ) in southern Africa

A revision of saxicolous species of the genus Rinodina (Ach.) Gray (lichenized Ascomycetes. Phvsciaceae) in southern Africa is presented. The study area covers the following countries: Angola. Namibia. Republic of South Africa. Lesotho. Swaziland, Zimbabwe, and Mozambique. Fifteen Rinodina species are treated in detail, and three of them are new to science: Rinodina longisperma Matzer & H.Mayrhofer. R. scabridula Matzer & H.Mayrhofer, and R. striatitunicata Matzer & H.Mayrhofer. The general part of the study includes information on the most important characters of the species with a special focus on ascospore types and ontogeny, and biogeographical notes are also given. A key to the species is provided. All the species are described in detail with notes on their ascus characters, spermogonial apparatus, and chemical characteristics, and their substrate and general distribution are indicated. The descriptions are accompanied by illustrations of the ascospores. Except for Rinodina oxydata s.I., the distribution of all the species in southern Africa and other parts of Africa respectively, is mapped. Several names were referred into synonymy, including Rinodina almbornii H.Mayrhofer. a synonym of R. confragosula (Nyl. in Cromb.) Miill.Arg.. and R. depressa (Vain.) Zahlbr. and R. albicans H.Mayrhofer which are synonymous to R. huillensis Vain. A separate section encompasses nine excluded species.


INTRODUCTION
The cosmopolitan genus Rinodina (Ach.)Gray com prises about two hundred species (Hawksworth et al. 1983), and its species taxonomy is still not settled satis factorily, particularly in the southern hemisphere.This became obvious, once again, during the preparation of the present work, a re-assessment of saxicolous species occurring in southern Africa.Type material of all critical species and, to the best of our knowledge, almost all rele vant collections were examined.Nevertheless, the present study should be seen as an actual inventory rather than an ultimate revision.Given the vastness of the study area, there are relatively few collections of saxicolous Rino dina species, and most of them come from the Cape Province in the Republic of South Africa.Type speci mens of some species are scant or damaged, several spe cies are still known from a few collections only, and thus the extent of their variation is difficult to estimate.This is particularly problematical in cases where species which are separated by slight morphological differences only, have been described.Morphologically and chemically, many of the saxicolous Rinodina species in southern Af rica do not exhibit 'exciting characters', and great simi larities can be observed between several species.There fore, determination of specimens has often proved to be rather difficult.In the material studied, several Rinodina species were detected which are most probably unde scribed.Not all of them are included in the present study, as they are often known from a single locality only, and their relationships to already described and similar spe cies could not be resolved.
However, despite the problems indicated above, many new results are presented in this paper.Several names of taxa are referred into synonymy, and three new species are introduced.The importance of minute characters such as ascospore ornamentation and length of spermatia for the circumscription of single species is documented.The first data on the biogeographical affinities of saxicolous Rinodina species occurring in southern Africa are pre sented.

MATERIAL AND METHODS
Standard light microscope techniques and standard tests for thallus chemistry (tic.spot tests for colour reac tions) were employed.Descriptions of ascus apices are based on observations in dilute original Lugol's solution (MERCK 9261) without pretreatment with KOH (K).For the testing of the N ( = HN03)-reaction of the blue-green to blackish pigment, longitudinal sections of thalli, apothecia.and/or spermogonia respectively, were directly mounted in this medium.

THE STUDY AREA
The main focus of this study was on species occurring in extratropical southern Africa, mainly including Na mibia.the Republic of South Africa.Lesotho, and Swazi land.Some specimens from adjacent countries (Angola.Zimbabwe.Mozambique) were also studied.However, as already mentioned above, most of the specimens exam ined were from the Republic of South Africa, mainlv from the Cape Province.For comparative purposes it was also necessary to study saxicolous Rinodina species described or mentioned from other parts of Africa, mainly including the islands of Pagalu (Annobón) and Sáo Tomé & Principe in the Atlantic Ocean, and the island of Socotra in the Indian Ocean.Bothalia 26,1 (1996) MORPHOLOGY, ANATOMY AND CHEMISTRY A short survey is given here of the most important characters of the species treated in detail.

Thallus and prothallus
Thalli in Rinodina species are usually crustose or crustose to squamulose.The formation of blastidia is the most important character for the separation of R. scabridula.The presence or absence of a distinct prothallus should not be overestimated in the circumscription of species.

Chemistry'
Only a few species contain secondary lichen substances which are detectable by thin-layer chromatography (tic).Rinodina huillensis is the only species with gvrophoric acid (and, in part, other secondary compounds) and is well characterized by the C+ red reaction in the medulla of the thallus.The thalli of R. confragosa, R. oxydata s.l., and R. substellulata give a K+ yellow reaction.The main li chen compound in these taxa is atranorin.

Pigments
A blue-green to blackish pigment was often found to be located in parts of the thallus, apothecia (proper exciple, epihymenium), and/or in the spermogonia of six spe cies: R. confragosula, R. huillensis, R. oxydata s.l., R. striatitunicata, R. substellulata.and R. subtristis.The pig ment reacts N+ red, K-, and HC1+ blue after pretreatment with K.It is most probably identical to the pigment which was described as 'Lecideagriin' by Bachmann (1890), and it occurs also in other genera of the Phvsciaceae, e.g. in Buellia De Not.(Scheidegger 1993, as 'pigment A).
An interesting and unique character is the K+ yel lowish to orange reaction in the hypothecium, the epi hymenium (here not always), and in spermogonia (not always) of R. reagens, which is due to the presence of a green-vellow pigment.

Apothecia
In the genus Rinodina.a great variation in apothecial types can be observed, and even within single species or specimens, the formation of apothecia can be variable.Apothecia are crvptolecanorine, lecanorine (in addition to the thalline margin a proper margin may be developed), and/or lecideine.Apothecia which appear to be lecideine in surface view may contain (dead) algal cells in the exciple (e.g. in R. oxydata s.l.. R. subtristis, R teichophiloides).

Asci
Most of the species have asci which can be described as corresponding to the Lecanora type (compare e.g.Honegger 1978;Rambold et al. 1994), but some variation in the formation of ascus tips can be observed.In R. oxy data s.l.. the asci correspond to the Lecanora type but show tendencies towards the Bacidia type (see e.g.Hafellner 1984;Rambold et al. 1994).The axial body in several species is often more or less indistinct (e.g. in R. geesteranii, R. huillensis.R. longisperma, R. oxy data s.l., R. subtristis) or even not observable (e.g. in R. longisperma) under the light microscope.In R. huil lensis, the axial body may be surrounded by a tube-like structure, and the same phenomenon could be observed in several other species after pretreatment with KOH.

As cospores
The most important characters for the separation and identification of Rinodina species are those of the ascospores.For terms used in the description of ascospores occurring in Rinodina (and other genera of the Phvsciaceae) see Poelt & Mayrhofer (1979); revised compilations of ascospore types were published by Hafellner et al. (1979), Mayrhofer & Poelt (1979) and Mayrhofer (1982Mayrhofer ( , 1984a)).With the increasing knowl edge on possible formation types in the ascospores of Physciaceae, it became necessary to modify the classi fication systems for spores by subsequent authors.For a better understanding, the most important changes are briefly outlined here.
1.The presence vs. absence of a torus (i.e. a dark belt in the region of a spore septum) is not always a reliable character for the definition of spore types.For instance, within several species a torus can clearly be observed in some of the ascospores under the light microscope, but is indistinct or even not visible in others (e.g. in Rinodina fimbriata, R. microlepidea, R. striatitunicata, R. substellulata); the same has been reported for other genera and species of the Physciaceae (e.g.Amandinea petermannii; see Matzer et al. 1994b).Scheidegger (1993: 333) pointed out that within Buellia closely re lated species may differ considerably in their torus while, on the other hand, species with only slight affinities to each other may have a similar torus.Consequently, a distinction of spore types which is solely based on the presence or absence of a torus cannot be recommended.This mainly concerns the following ascospore types: Beltraminia type (incl.Buellia type) and Physconia type (incl.Dubyana and Sicula type); in both cases the terms which have been introduced first into the literature should be used (for dates of introduction see Mayrhofer 1982).Such a procedure has been applied already by Matzer & Mayrhofer (1993), Scheidegger (1993), and, in part, by Rambold et al. (1994).
2. Original definitions of single spore types have been modified or improved; the most important case in this context is the Dirinaria type.It was introduced by Mayrhofer (1982: 315) for ascospores lacking a torus and with internal wall thickenings of the Physcia, Mischoblastia, and/or intermediate types.In subsequent studies, the mode of spore ontogeny in this spore type (apical internal spore wall thickenings appear before the septum is inserted) was used more or less implicitly as the major character for the definition of the Dirinaria type (Giralt & Matzer 1994;Giralt et al. 1995: Rambold et al. 1994), or this peculiar ontogeny was stressed as being a very important character (Giralt 1994;Giralt & Mayrhofer 1994a. b, 1995).In the present study this concept of the Dirinaria type is adopted.
Within the species treated in detail in the present study, the following ascospore types were observed: Beltraniinia type (incl.Buellia type): spores without in ternal wall thickenings, Buellia distrata.
Physcia type: septal and apical wall thickenings well developed, spore lumina at their distal ends concave, torus usually developed, in young spores the septum is inserted before wall thickenings become distinct.R. confragosa.R. confragosula.R. scabridula, R. sp. 1.
Tunicata type: spores with a markedly thickened outer wall layer, R. striatitunicata.Dirinaria type: when mature similar to the Physcia type (septal and apical wall thickenings well developed, spore lumina at their distal ends concave), but usually lacking a torus and with a different spore ontogeny: in young spores the septum is laid down after the formation of apical wall thickenings, R. gennarii.(compare also R. geesteranii).
As can be seen from this compilation, in most of the species the ascospores could not clearly be assigned to a single spore type.Most of the spores encountered belong to the Pachvsporaria or Milvina type.In contrast to all other species treated in detail, the insertion of the septum in ascospores of R. geesteranii may occur before or after internal wall thickenings become distinct.

Spermogonial apparatus
The spermogonial apparatus in all species studied in this respect is typical of the genus Rinodina.Therefore.
we give here a general description which is not repeated in the species accounts: Spermogonia immersed in the thallus.with the upper parts prominent or not. in longitudinal section ± globose, ellip soid.cupuliform, or flask-shaped.Spermatiophores sep tate, composed of spermatiogenous cells.Spermatiogenous cells intercalar and terminal: intercalar cells ± cy lindrical and with a lateral projection below the upper transverse septum: terminal cells ± flask-shaped or ob long-conical and with an apical projection; spermatia are formed on the lateral and apical projections.Spermatia bacilliform.hyaline, non-septate.
The spermatiophores correspond to type VI of Vobis (1980, as 'Conidiophoren').For illustrations of sperma tiophores and spermatiogenous cells respectively, see Matzer et al. (1994a) and Mayrhofer et al. (1993).The spermogonial apparatus of R. longisperma is basically of the same type as described above, but it differs in the way that intercalar spermatiogenous cells could be observed only exceptionally.

BIOGEOGRAPHICAL NOTES
According to the present, rather incomplete data, saxicolous species of Rinodina from southern Africa show four different biogeographical elements or affini ties.Particularly noteworthy is the high incidence of endemism.
1. Endemic element: this is the dominant element and includes ?/?. confragosula (a record from New Caledonia seems to be dubious), R. geesteranii.R. huillensis, R. longisperma.R. microlepidea.R. scabridula.R. striati tunicata.R. subtristis.R. teichophiloides, and R. sp. 1.Most of these species have been found only in the Repub lic of South Africa and. in part, in adjacent countries, while R. huillensis and R. subtristis are known to extend further north to Angola.
2. African element: R. substellulata is known to occur in the Cape Province but also on some tropical islands adja cent to the African continent (Principe.Annobón.and Socotra).For synonyms and typification see Mayrhofer & Poelt (1979) and Mayrhofer (1984a).
Thallus thin to thick, crustose to squamulose, con tinuous and areolate or composed of discrete areolae which may be ± wart-like, whitish, pale grey to ochra ceous; prothallus absent or present and then it may be strongly developed, brown to black.Chemistry> : thallus K+ yellow, P ± yellowish; tic: atranorin, chloratranorin (not always), zeorin (not always), in the material from southern Africa an unidentified compound (rf-classes 6/6/6) was found.
Among the species treated in the present study, Ri nodina confragosa can be recognized easily by the K+ yellow thallus reaction and by the ascospores of Physcia type.As far as thallus formation is concerned, the species shows a wide range of variation in Europe, from thinly crustose to only slightly squamulose (these morphotvpes are also present in southern Africa) to very thick and then with ± wart-like to squamulose areolae; the thallus is continuous and rimose-cracked or com posed of discrete areolae, and a prothallus may be ab sent or strongly developed.Variability can also be ob served in chemistry, and different chemical races appear to exist (Hecklau et al. 1981;Mayrhofer & Leuckert 1985).
Thallus crustose, rimose-areolate, some areolae occa sionally in the form of small squamules, greyish, greybrown, brown, occasionally whitish due to presence of a pruina; in longitudinal section outermost layer of the phenocortex is brown or, particularly in areolae which bear an apothecium, (blackish to) blue-green; prothallus absent or present, black.Chemistry: no lichen substances detect able by tic.
Apothecia to 0.8 mm diam., crvptolecanorine, more rarelv lecanorine or becoming lecideine and then adnate, occasionally contiguous, disc black, rarely dark brown, occasionally whitish pruinose, the disc may be surrounded by a small edge of white pruina, disc slightly concave, plane, or convex.Exciple: where a proper exciple is de veloped it is brown and blackish to blue-green in parts or dominantly blue-green.Epihymenium 10-30(-35) //m tall, entirely brown to olivaceous, brown to olivaceous and blue-green in parts, or dominantly blue-green with brown to olivaceous spots.Hymenium 70-120(-130) um tall.Hypothecium to 250 um deep, hyaline; adjoining parts of the thallus to the hvpothecium may be blue-green.Paraphyses 1-5 um, apices 3-6 um wide.Asci corresponding to Le canora type, with 8 or less spores.Ascospores (Figure 1C, D) with internal wall thickenings of Pachvsporaria or Milvina type, or intermediate between these two types, occasionally also of Physcia type or spore lumina irregu larly biconical in shape, septum in young spores inserted before internal wall thickenings become distinct, torus de veloped, spores finely scabrid, without septal swellings in KOH, 14-32 x 7.5-16.0um.Spermogonia immersed in the thallus, with the upper parts prominent or not, often contiguous, ostiolar region black or brown; in longitudinal section at base and laterally hyaline, rarely pale brownish, above brown to olivaceous and/or blackish to blue-green.Spermatia (3-)4-5 x 1.0-1.5 pm.Chemical reaction of the blue-green pigment in the thallus, apothecia, and spermogonia: N+ red.
Rinodina confragosula grows on hard siliceous rocks such as granite, sandstone, and quartzite.It is known with certainty only from the Republic of South Africa and Leso tho where it was found to be widely distributed at altitudes between 450 m and 3 080 m (Figure 2).The specimen cited from Concepcion in Argentina (Muller Argoviensis 1889) represents a Rinodina species for which an appropriate name is not available at present.A record from New Caledonia was given by Muller Argoviensis (1887).R. confragosula is a quite variable species in several respects (see description).It can be recognized best by the usually black apothecia, the presence of a blue-green, N+ red pigment in the phenocortex, apothecia, and sper mogonia, and by the absence of lichen substances detect able by tic.Within a single specimen, the range of spore size is often significantly smaller than given in the de scription of the species above, but in other specimens the range was found to be 18-31 x 10-16 pm (PRE 2089b), 20.0-27.5 x 9-11 pm (Brusse 2547), and 22-32 x 10.5-16.0(Brusse 4575).
Other saxicolous species occurring in the study area and showing close similarities with R. confragosula in clude R. longisperma.R. microlepidea, and R. subtristis.Rinodina longisperma is distinguished from R. confrago sula by smaller apothecia and ascospores, and by longer spermatia.The separation of R. confragosula from R. mi crolepidea and R. subtristis respectively, is discussed in detail under both the latter species.
The differences used for the delimitation of Rinodina almbornii from R. confragosula by Mayrhofer (1984a) are considered to fall within the variability of R. confragosula.Mayrhofer (1984a) pointed out that R. confragosula is closely related to R. kozukensis (Vain.)Zahlbr.which was described from Japan by Vainio (1921, as Melanaspicilia kozukensis).The relationship of these two species will be discussed in a forthcoming paper.
Rinodina argentiniana Mull.Arg.(Argentina, Concep cion, 1882, Lorentz, G, holo.!) differs from R. confrago sula by having ascospores lacking a torus and by a different type of spore ontogeny (internal apical wall thickenings appear before the spore septum is inserted).The record of R. argentiniana from Lesotho (summit of Masiti Mountain, 6300 ft, 1929-30, Hewitt, TRH) by Mayrhofer (1984a) actually refers to R. confragosula.Ac cording to the material available, R. argentiniana does not occur in the study area.Bothalia 26,1 (1996) One of the specimens cited by Mayrhofer (1984a) un der Rinodina depressa could not be determined with cer tainty, and it is treated here under R. cf.confragosula: Western Cape, Van Rhvnspass.Van Rhynsdorp, Van der By I 766 (W).
Rinodina geesteranii is only known from the type lo cality on the Cape Peninsula where it was collected on hard quartzitic rocks (Figure 3).It is associated with other lichen species, most commonly with Peltula euploca (Ach.)Poelt in Pisút.suggesting that the locality is more or less regularly wetted by rain or dripping water (com pare notes on the ecology of Peltula euploca in Biidel 1987: 58).This view is also supported by the fact that the thallus of R. geesteranii is partly associated with cy anobacteria; green algae containing lichens which have connections with lree-living cyanobacteria occur exclu sively at localities which receive liquid water (Poelt & Mayrhofer 1988).R. geesteranii is characterized mainly by the thickly crus tose to slightly squamulose thallus.and by the ascospores lacking a torus and with internal wall thickenings interme diate between Pachvsporaria and Milvina type.Three other saxicolous Rinodina species w ith ascospores lacking a typi cal torus are known to occur in the study area, viz.R. gennani.R. reagens, and R. teichophiloides.In R. teicho philoides, torus-like structures may be developed (see ac count of R. teichophiloides in the present study).Rinodina gennarii is separated from R. geesteranii mainly by a thin  Mayrhofer & Poelt (1979).Mayrhofer (1984a).
The following description is based only on the speci mens cited below.
Rinodina gennarii is w idespread in temperate regions of the northern and southern hemisphere and is found on a wide variety of substrata (e.g.Fox & Purvis 1992;Mayrhofer 1983Mayrhofer , 1984a;;Mayrhofer & Poelt 1979;Nimis 1993;Santesson 1993).The species is here recorded for South Africa where it occurs on granitic and quartzitic rocks from low to high altitudes (Figure 3).This species is characterized mainly by ascospores of Dirinaria type and the numerous, often crowded, small apothecia.Giralt & Mayrhofer (1995) did not observe marked differences between R. gennarii and the corticolous/lignicolous R. oleae Bagl.However, we refrain here from putting R. gennarii into the synonymy of R. oleae.R. gennarii and R. oleae are tentatively treated as a pair of closely related, similar species of which one is saxi colous and the other corticolous/lignicolous.Further ob servations, including detailed data on distribution and ecology are necessary to decide whether this concept can be maintained or not.
Rinodina gennarii and R. oleae are not the only ex ample for vicarious species within the genus.Another cor responding pair of species is formed by the saxicolous R. beccariana Bagl.and the corticolous R. roboris (Duf.ex Nyl.) Arnold var.roboris (Mayrhofer et al. 1993) (BM, para.!).
Rinodina huillensis is a widespread species in south ern Africa and is so far known from Angola, the Re public of South Africa, and Swaziland.It occurs on granite, quartzitic rocks and sandstone at a range of altitudes (Figure 4).
Among the saxicolous Rinodina species known to oc cur in the study area, R. huillensis alone contains gyro phoric acid in the thallus (medulla C+ red), often accompanied by lecanoric and orsellinic acid.It is a wellknown phenomenon in many lichens that in addition to gyrophoric acid, at least traces of lecanoric and orsellinic acid can be found (e.g.Schreiner & Hafellner 1992: 22).Further saxicolous species of Rinodina containing gyro phoric acid occur outside southern Africa, but these are distinguished from R. huillensis by several characters (Matzer et al. 1994a).In a single specimen of R. huillensis (Brusse 4497, PRE) 5-O-methylhiascic acid was detected.This lichen compound was also reported from Rinodina tephraspis (Tuck.)Herre, a northern hemisphere species which differs chemically from R. huillensis by the pres ence of zeorin (Mayrhofer et al. 1992).
Re-examination of the types of Rinodina albicans and R. depressa revealed that these names are synonyms of R. huillensis.Vainio (1901) described R. huillensis and R. depressa (as Buellia depressa) in the same publication; R. huillensis is selected here as the correct name as it was mentioned first in Vainio's paper and appropriate generic classification was applied.The statement of Mayrhofer & Leuckert (1985) that R. depressa does not contain lichen substances is based on the following specimen: Republic of South Africa, Roggeveld Moun tains, Brusse 3265 (PRE).We studied this specimen, and it may represent a Rinodina species but cannot be iden tified with certainty as the ascospores are badly devel oped.From its external morphology it can be concluded that it is not R. huillensis.One of the specimens (Van der Byl 766, W) cited under R. depressa by Mayrhofer (1984a) is treated as Rinodina cf.confragosula in the present paper.Several specimens mentioned by Mayr hofer (1984a) under R. huillensis do not belong to this species: one from Angola (Humpata Plateau, 6-2-1960, Degelius, GZU) refers to R. subtristis whereas the other one from the same country was not examined.The ma terial from Namibia (Haifischbucht, Fincke.W) repre sents Rinodina longisperma, a species newly described in the present study.The specimen from the Northern Province (Transvaal, Louis Trichardt. 8-10-1953, Almborn.LD) was selected as the isotype of Rinodina scabridula in the present work.Another specimen from Mpu malanga (Eastern Transvaal) was also cited: Dist.Pil grims Rest, 4 miles SE of P. R., on rocks near the road.22-10-1953, Almborn (LD).We examined two specimens held in LD which are labelled as indicated above but with different collection numbers: Almborn 7831 actually represents R. huillensis whereas the crustose lichen species in Almborn 7836 has hyaline asco spores and does not belong to the Physciaceae.The collection of R. huillensis from (Orange) Free State (Ladvbrand.Maas Geesteranus 6534.L. LD) was cited under Rinodina microlepidea by Mayrhofer (1984a).
Rinodina longisperma is so far known from quartzitic rocks including sandstone from Namibia and the Republic of South Africa (Figure 6).
Rinodina longisperma can be recognized best by a set of characters including the small and cryptolecanor ine apothecia.the small ascospores similar to Pachysporaria or Physconia type.and.particularly, by the relatively long spermatia (hence the epithet 'longis perma').The bulk of Rinodina species in which sper matia have been observed posses spermatia with length measurements from 3 to 6 (to 7) um.whereas in R. longisperma they are (4-)6-9 um long.In addition, pe culiarities could also be observed in the spermatio genous apparatus of R. longisperma.M ost of the Rinodina species studied in this respect have many ter minal and intercalar spermatiogenous cells (e.g.see 'morphology and anatomy' in the general part of the present study; Matzer et al. 1994a: 108, fig. 8).In R. longisperma most of the spermatiogenous cells are ter minally arranged, whereas those with an intercalar po sition could be observed only exceptionally.
The following specimens were included in R. mi crolepidea by Mayrhofer (1984a), but for a range of rea sons a new settlement is proposed here: (Orange) Free State, Ladybrand.Maas Geesteranus 6534 (L, LD) be longs to Rinodina huillensis.The specimen from (Orange) Free State.Dist.Trompsburg.2-10-1953, Almborn 5809 (LD) which is cited by Mayrhofer (1984a) with the date '2-10-1954', collection number '5808'.and herbarium 'L' represents a Rinodina sp.The ascospores are usually 2celled but exceptionally also 3-celled spores could be ob served.The Rinodina species from the Western Cape, Calitzdorp-Kruisrivier, Maas Geesteranus 6711 (L) differs from R. microlepidea by the type of spore ontogeny in the way that internal wall thickenings appear before the septum is inserted.The ascospores of this specimen re semble those of R. geesteranii, but differences exist in the morphology of the thallus.
8. Rinodina oxydata s.l.Rinodina oxydata s.l. is distributed world-wide.The following description is based on material from southern Africa only.For further explanations see the discussion given below.
In its present circumscription, the species is known only from single localities in the Republic of South Africa and Zimbabwe where it grows on hard siliceous rocks (Figure 6).
The status of Rinodina microlepidea is rather vague.The description given above is based only on two rather scant specimens and thus, not all characters (e.g.ascus type, behaviour of ascospores in potassium hydroxide) could be studied in full detail.Unfortunately, the type collection is in a very poor condition, and the apothecia in particular are very old or damaged.Therefore, no statements can be given on the variability of R. mi crolepidea and only one further specimen which looks more or less similar to the type is accepted as belonging to that taxon.The distinction between Rinodina subtris tis and R. microlepidea is problematical.According to the material available, the former species is separated by larger apothecia (to 0.7 mm diam.) and the coloration of the thallus: brown or ochraceous instead of whitish, yel lowish to ochraceous.Like R. microlepidea and R. sub tristis, other Rinodina species in the study area are characterized by the lack of secondary lichen substances and ascospores of Pachvsporaria or similar types.This in cl udes R. confragosula.R. longisperma.a n d R. scahridula.Rinodina confragosula differs from R. mi crolepidea mainly by its larger apothecia (to 0.8 mm diam.) and by ascospores which are rather variable in the formation of internal wall thickenings (Pachvsporaria.Milvina, Physcia, or intermediate types or the spore lumina are irregularly biconical) and exhibit a broader size range (14-32 x 7.5-16.0pm).Rinodina longis perma is separated from R. microlepidea mainly by its often smaller ascospores, 1 ()-16(-18) x 6-1 ()(-12) «m. and longer spermatia, 5-9 p m. Rinodina scahridula can easily be distinguished by the formation of blastidia on the thallus.Thallus thin, crustose, areolate, continuous, yellowish, pale ochraceous, pale greyish; prothallus absent or present, blackish.Chemistry': thallus K+ yellow; tic: atranorin (not found in Rinodina minima; see below).
In the study area, Rinodina oxydata s.I. is known only from three localities (Mozambique, Transvaal, Table Mountain) where it grows on schistic and quartzitic rocks.
The name Rinodina oxydata is commonly applied to lichens exhibiting the following set of characters: thallus thinly crustose or thick and nearly squamulose, occasion ally evanescent, containing atranorin (K+ yellow), apothecia showing a wide variety from cryptolecanorine, lecanorine to lecideine, ascospores (usually) of Mis choblastia type.Particularly, in thallus formation, in the apothecial type, and, in part, in ascospore size, a wide range of variability can be observed.Many binomials have been introduced for lichens with characters mentioned above, but none of these names can be used unequivocally unless a thorough revision of the R. oxydata group on a world-wide scale has been carried out (see Table 2).Such a revision should also include R. substellulata, a species with ascospores of Pachysporaria type occasionally grad ing into Milvina type, but in all other characters displaying close similarities with R. oxxdata.All characters used for the delimitation of species (thallus formation, apothecial type, ascospore size) are not constant and show various transitional states, especially when many specimens are analyzed.On the other hand, it seems to be premature to put all these names into synonymy of R. oxydata (A.Massal.)A.Massal.As a consequence, the description given above includes all R. oxydata-Y\ke lichens known from southern Africa, and a survey on the 'species' known from that region is given in Table 1 below.Rambold et al. (1994) argued for a possible re-estab lishment of the generic name Mischoblastia A.Massal. which may be based on Rinodina oxydata.The most im portant characters mentioned to be characteristic for Mis choblastia are: asci corresponding to Bacidia type, asco spores of Mischoblastia type, presence of an aeruginose, N+ red pigment in the exciple.However, it should be stressed here that all these characters do not justify a separation of Mischoblastia from Rinodina (Ach.)Gray; this is evident from the following observations: 1, asci of R. oxydata (and of species which are considered to be closely related to R. oxydata) exhibit a considerable range of variation and correspond either to Bacidia type or to Lecanora type, or are more or less intermediate between these two types; 2, the presence of Mischoblastia type ascospores cannot be an argument as the formation of pronounced internal wall thickenings is one of the most important characters of the genus Rinodina; 3, an aerugi nose, N+ red pigment occurring in the exciple and/or in the epihymenium is known not only from R. oxydata and related species but also from other Rinodina species: e.g.R. beccariana Bagl.(Mayrhofer et al. 1993)    Icones: Matzer & Mayrhofer (1994).
Apothecia to 1.7 mm diam., cryptolecanorine or lecanorine and then adnate.disc reddish brown, dark brown to black, often with a narrow, pale reddish brown border line between disc and thalline margin, plane to convex, thalline margin persistent, entire or slightly crenate.Epihymenium 10-30 pm tall, brown or yellowish brown.Kor K+ yellowish to orange.Hymenium 100-150 pm tall.Hypothecium to 180 pm deep, entirely hyaline, hyaline and green-yellow in parts, or entirely green-yellow.K.+ yellowish to orange.Paraphyses 1-4 //m, apices to 6 pm wide.Asci corresponding to Lecanora type, 8-spored or with fewer spores.Ascospores (Figure 7D) with internal wall thickenings of Mischoblastia type, occasionally grad ing into Pachvsporaria type, septum in young spores in serted before internal wall thickenings become distinct, torus absent, spores finely scabrid.with septal swellings in KOH, 22-31 x 10-19 wm.Spermogonia immersed in thallus, with a reddish brown, dark brown or blackish ostiolar region; in longitudinal section reddish brown to brown above, otherwise hyaline to green-yellow.K-or K+ yellowish to orange.Spermatia (3.5-)4.0-4.5(-5.5)x 1.5-2.0pm.This species is hitherto known from the Western Cape in southern Africa (Figure 6) and a few inland locations in New Zealand (Matzer & Mayrhofer 1994).In the study area it is confined to calcareous sandstone.
A detailed discussion of the species was provided just recently by Matzer & Mayrhofer (1994) and thus, only a few notes are given here.Rinodina reagens is mainly characterized by the presence of a green-yellow pigment in the hypothecium.epihymenium.and in the spermogo nia, which reacts K+ yellowish to orange.This pigment is not known from any other Rinodina species.The large ascospores without a torus are also a distinctive character which separates the taxon from almost all other Rinodina species known to occur in the study area.Similar spores are found in Rinodina teichophiloides which is closely related to R. reagens.These sympatric species are con fined to different substrates.In South Africa R. reagens grows on calcareous sandstone, whereas R. teichophi loides is restricted to hard siliceous rocks.
Rinoilina scabridula is known hitherto from two localities in the Northern Province (northern Transvaal) where it oc curs on sandstone and granitic rocks (Figure 6).
Rinodina scabridula is characterized mainly by the for mation of blastidia (for this term see Poelt 1980 andGiralt et al. 1993a, as 'Blastidien') usually at the margins of thallus areolae, and by its ascospores which are of Mil vina, Pachysporaria, Physcia or intermediate type.To de termine the species, careful observations are necessary as blastidia are not formed by all thallus areolae and thus, parts of the thalli are smooth.Consequently, confusion with Rinodina subtristis (see present study) may be pos sible.
The specimen which has been selected as the isotvpe of R. scabridula in the present work was treated under R. huillensis bv Mayrhofer (1984a).
Rinodina striatitunicata is only known from two lo calities in the Western Cape near the Cango Caves.It grows on volcanic conglomerate and sandstone situated in Karoo vegetation (Figure 6).
As indicated by the epithet, Rinodina striatitunicata is characterized mainly by ascospores of Tunicata type with peculiar striate ornamentation.The sculpture on the spore surface is elongate and more or less parallel (compare Scheidegger 1993: 335).In order to recognize the orna mentation in ascospores of R. striatitunicata, more or less mature or overmature spores must be studied carefully under the light microscope at a magnification of x 1000 or more; young ascospores are more or less smooth.
Other saxicolous Rinodina species with Tunicata type ascospores are R. calcarea (Arnold) Arnold.R. tunicata H. Mavrhofer & Poelt (compare Mayrhofer & Poelt 1979), and R. ft I son ii H.Mavrhofer (Mayrhofer 1984b).They dif fer from R. striatitunicata in several respects and.as far as the ascospores are concerned, mainly by a microrugulate (sculpture less than 1 um in size, circular to elongate and irregularly arranged) instead of a striate ornamenta tion.
The thallus of R. striatitunicata varies in consistency and colour according to the rock mineralogy: a rough and rusty morphotvpe was found on volcanic conglomerate (Almborn 4320).whereas dark brown (to grey in small parts) and compact thalli occur on sandstone (Hafellner 30649 and 30650).

Specimens exam ined
Rinodina substellulata is known with certainty from some islands adjacent to the African continent (Principe and Annobón in the Atlantic Ocean: Socotra in the Indian Ocean), and from the Northern Province (northern Trans vaal) and the Western Cape in the Republic of South Af rica (Figure 9).It grows on hard siliceous rocks (including quartzitic rocks, sandstone, and basalt).Further records are given from Costa Rica (Stizenberger 1893), Indonesia (Mayrhofer 1984a), and Australia (Mayrhofer 1984b).R. substellulata is characterized mainly by the K+ yel low thallus reaction (indicating atranorin) and ascospores of Pachysporaria type which may show tendencies to wards Milvina type.The species is closely related to taxa of the Rinodina oxydata group.Rinodina oxydata and re lated species (see account of R. oxydata in the present studv) differ from R. substellulata mainly by having as cospores of Mischoblastia type.Rinodina beccariana var.lavicola (J.Steiner) Matzer & H. Mayrhofer (Mayrhofer et al. 1993) shares several similarities with R. substellulata (e.g.K+ yellow thallus reaction, lecideine apothecia.pres ence of a blue-green, N+ red pigment in the apothecia.Pachysporaria type ascospores) but is separated by chemi cal and morphological characters: the thallus is grey to brown and contains zeorin.the apothecia are up to 1 mm in diameter, and the ascospores exhibit a wider size range (14-25 x 7-14 um).In addition, the ecology and distri bution of both taxa is different, with R. beccariana var.lavicola occurring on volcanic rocks near the coast or on low coastal mountains in Macaronesia and in the Medi terranean region.
Rinodina subtristis was originally known only from the Cape of Good Hope (Crombie 1876a, b; as Lecidea sub tristis) but can now be reported also from other regions of the Republic of South Africa and from Angola (Figure 9).The species grows on hard quartzitic rocks (including quartzitic sandstone).
In addition to R. subtristis, there are several other Ri nodina species in the study area which are characterized by the lack of secondary lichen compounds detectable by thin-laver chromatography and by the presence of Pachvsporaria type ascospores with a torus, viz.R. longisperma, R. microlepidea, R. scabridula, and R. confragosula.The problematic separation of R. subtristis fro m R. mi crolepidea is discussed under the latter species in the pre sent work.Distinctive characters of R. longisperma include the small cryptolecanorine apothecia.the small ascospores in which the spore lumina may have small protrusions, and the comparatively long spermatia.Rino dina scabridula is mainly distinguished by the formation of blastidiate thalli.The separation of R. subtristis from R. confragosula is based on a set of characters: grev thalli are often found in R. confragosula but never in R. sub tristis; the apothecia of R. confragosula are usually purely black and cryptolecanorine but dark reddish brown to black and lecideine or lecanorine in R. subtristis-, the as-  (Mayrhofer 1984a) from Annobón ( = Pagalu).All these species and R. subtristis are known only from single or at best a few specimens, and thus statements whether these taxa are really well separated, cannot be made without some hesitation.Under this precondition, R. praefinita is dis tinguished from R. subtristis by smaller ascospores (± 13-17 x 7-10 /*m), and R. subanceps differs by the thick, more or less squamulose thallus.A well-devel oped thallus and ascospores with an indistinct torus are characteristic for R. newtonii.A detailed study of the relationships between these taxa is beyond the scope of the present work and was not carried out.Mayrhofer & Leuckert (1985) recorded the presence of zeorin in a specimen of R. subanceps from Annobón (H-NYL 28964), but lichen compounds could not be detected by recent tlc-analyses in the holotype from the island of Cabras (Sáo Tomé & Principe, H-NYL 28965) and in another specimen from Annobón (H-NYL 28514).
One of the specimens now regarded as belonging to R. subtristis was treated under R. huillensis by Mayrhofer (1984a) Thallus thinly crustose to somewhat squamulose, oc casionally evanescent, effuse to rimose-areolate, light greenish grey, grey-brown, dark brown, ochraceous, oc casionally associated with cyanobacteria; prothallus ab sent or present, brown.Chemistry: no lichen substances detectable by tic.
Rinodina teichophiloides is a characteristic species of maritime localities in the Cape Province (Figure 10).It grows on hard siliceous rocks such as quartzitic sandstone, on inclined, vertical and overhanging rock faces.Occa-  sionally cyanobacteria were found growing at the margin of well-developed thallus areolae.

J
Rinodina teichophiloides was discussed in detail re cently by Matzer & Mayrhofer (1994).Since then, addi tional material has become available and some annotations can be given on the morphology of the species.Apothecia with a lecideine appearance in surface view may contain (dead) algal cells in the proper exciple.Paraphyses with 'oil cells' (oil paraphyses, compare Poelt & Pelleter 1984) were observed in some specimens.The ascospores lack a typical torus but intense pigmentation in the septal region may simulate the presence of a torus.Rinodina teichophi loides can best be recognized by the often rather thin thal lus, by apothecia often appearing to be lecideine, the large ascospores with internal wall thickenings of Mischoblastia or Pachysporaria type, and by the often coloured hy pothecium which lacks a distinct colour change in potas sium hydroxide.It is closely related to R. reagens, a species which also occurs in the Cape Province.The latter can easily be distinguished by the presence of a K+ re active pigment in the hypothecium and-not always-in the epihymenium and in the spermogonia.Both the spe cies also differ in substrate ecology (see discussion of R. reagens in the present study).
The species is known only from a single locality in the Eastern Cape (Figure 10).
The above description is based on the single collection cited below.In Mayrhofer (1984a) this specimen was in cluded in Rinodina interpolata (Stirt.)Sheard.but we have some doubts whether it really belongs to that species.Ri nodina interpolata usually occurs in Europe on more or less vertical or overhanging surfaces of cliffs mainly in coastal sites but also in inland localities (e.g.Fox & Purvis 1992;Mayrhofer 1984a: Mayrhofer & Poelt 1979;Santesson 1993: Sheard 1973); it was also recorded from the Antarctic region (see Lindsay 1973).A definite statement on the occurrence of R. interpolata in southern Africa only seems possible after extensive field studies and col lecting of more material.Only a few significant differ ences between Rinodina sp. 1 and R. interpolata can be observed: the ascospores in R. interpolata are also of Phy scia type but, particularly when older, often grade into Physconia type; in several specimens of R. interpolata, zeorin could be detected by tic (hitherto unpublished data;Leuckert & Mayrhofer 1984).For a description of R. in terpolata see Fox & Purvis (1992), Mayrhofer &Poelt (1979), andSheard (1973).

EXCLUDED SPECIES
This section includes saxicolous species from southern Africa which were referred to the genus Rinodina by vari ous authors, and Rinodina species which were erroneously recorded from the study area.
Buellia distrata is known only from the type specimens collected on Table Mountain on quartzitic rocks.This species has ascospores which lack internal wall thickenings, therefore it cannot be placed within Rinodina as was done formerly (e.g.Dodge 1971;Mayrhofer 1984a).Alternative placements could be in either Atnandinea Choisy ex Scheid.& H. Mayrhofer or Buellia De Not., two genera which are separated mainly by their sper matia: long, filiform, and curved in Amandinea and short, bacilliform, and straight in Buellia (e.g.M atzer & Mayrhofer 1993;Matzer et al. 1994b;Scheidegger 1993).As spermogonia are not present in the material of Buellia distrata now available, the best solution seems to retain the existing name of this taxon.This species was not validly transferred into the genus Rinodina by Dodge (1971) as reference to the basionvm was omitted (1CBN, Art.33.2.).Judging by the small spores mentioned in the protologue the species probably does not belong to Rinodina.This taxon was described from Port Natal (Durban) in the Republic of South Africa.According to Mayrhofer (1984a) the type cannot be located.Crvptogamie Exotique 5: 272 (1932).Type: ?Swaziland, Mbabana, ad lapides siliceos, V. Lebzelter s.n.(W, holo.);not seen, citation after Zahlbruckner (1932: 273).
The records of this species from the Zambezi River (Stizenberger 1895;Doidge 1950) are based on material of Rinodina oxydata (Mayrhofer 1984a).
Lecanora deminutula Stizenb.: 210 (1890).Mayrhofer (1984a) pointed out that the scant type is in a p(X)r condition and a treatment of the species is not pos sible.This species was described from the Cape Province.Ac cording to the protologue it probably belongs to Buellia.The name Rinodina microphthalma was treated as a synonym of Lecanora microps Stizenb.by Stizenberger (1890), and of Buellia verruculosa Mudd bv Zahlbruckner (1931).9.
Modern taxonomic concepts clearly suggest that this species should be included in the genus Hafellia Kalb.H. Mayrhofer & Scheid. (Sheard 1992).

FIGURE 2 .
FIGURE 2.-Known distribution of Rinodina confragosa in southern Africa, □ ; and of R. confragosula.• .Two collection localities of/?, confragosula could not be located in detail and are therefore not mapped: Lesotho, Mamerthes; KwaZulu-Natal, Drakensberg, on ridge of the Sugar Loaf near The Cavern.

TABLE 1 .
-Rinodina oxydata-like lichens known to occur in southern Africa; for further explanations see the discussion in the text

TABLE 2 .
-Selection of "species' which have to be considered in the course of a revision of the Rinodina oxydata group Exs.: Plantae Graecenses Lichenes 433, as Rinodina teichophila.