Molecular biological techniques now provide many opportunities for determining relationships between organisms. In lichen-forming fungi, polymerase chain reaction (PCR) amplification can be undertaken with primers that have enhanced specificity for fungi, allowing the recovery of precise regions of the genome of the fungal partner. Such PCR techniques are suitable for the analysis of both fresh and dried specimens of great age, and as the techniques require only minimal amounts of starting material, they are suitable for the analysis of large numbers of specimens. PCR can be used to examine both conserved and variable regions of the fungal genome that may be taxonomically significant, and variability in these sequences can be used to describe species boundaries. The ease of analysis of large numbers of specimens without the need to purify or culture the fungal partner makes analysis and comparison of infra- and inter-specific relationships possible, and provides valuable information into their genetic background. Most molecular analyses undertaken with lichen-forming fungi have considered variation within ribosomal RNA clusters, and in this paper we consider the application and interpretation of such PCR-based analysis in the circumscription of species and in determining concepts for individuality and infra-specific groupings.

Practically most of the taxonomists working on vascular plants and fungi are using the so-called taxonomic species concept, relying largely on morphological criteria or other observable patterns of discontinuity and assuming that these patterns are reflecting the underlying genetic integrity of species. Such a concept is described in detail here for the genus Usnea. The typologicalview versus the populational view of species are discussed. The diagnostic value of characters used by previous and recent taxonomists of the genus is analysed. The fact that Usnea has such a bad taxonomic reputation is discussed. The use by J. Motyka, author of the world monograph published in the thirties, of characters that are strongly modified by environmental factors together with a typological view of species are thought to be mainly responsible for this. More than 770 names have been published worldwide so far and it is suggested that around 50% are to be considered as synonyms.

Morphological, anatomical, chemical and ecological evidence show that the South African lichen Roccellina capensis (Nyl. ex Stizenb.) Tehler has developed a dispersal strategy with either sexually dispersed individuals, or vegetatively dispersed, sorediate individuals, hence a putative species-pair. The concept ‘species-pair’ is well-known and widely used in lichenology. It refers to closely related, morphologically indistinguishable lichens that differ from each other by their dispersal strategies only. The so-called ‘ primary species ’ produces fruiting bodies and sexual spores, while its counterpart, the ‘ secondary species ’ is vegetatively dispersed by soredia, isidia or fragmentation. Our study uses molecular information to investigate if R. capensis should be treated as a species-pair or not. The internal transcribed spacer (ITS) of the nuclear ribosomal DNA was sequenced for 17 specimens of R. capensis in order to test if sexual and asexual (sorediate) specimens would group into two distinct, monophyletic groups. Such grouping would support their recognition as separate taxa in a species-pair. We also investigated whether the randomly amplified polymorphic DNA (RAPD) method would be a suitable procedure in population studies of lichens. The analysis of the ITS region produced a single most parsimonious tree. The sexual and asexual specimens were grouped according to their geographical distribution, rather than according to their dispersal strategy. This result clearly shows that the sexual and asexual specimens ofR. capensis do not represent a species-pair. The RAPD data set produced an almost completely unresolved tree due to several homoplasious characters, possibly a result of algal contamination.

Lichens are troublesome organisms for taxonomists because of their special ‘ double nature ’ i.e. being composed (normally) of two partners. Only recently has it been understood properly that the same fungus can take different photosynthetic partners and develop into quite different-looking organisms, for example Lobaria amplissima and Dendriscocaulon umhausense. The taxonomic problem is to show that two totally different-looking lichens in fact contain the same fungus. This is possible to demonstrate when mixed stands are available, but is now best done with molecular methods. Since the international code of nomenclature rules that the name of a lichen species is the name of the fungus, two different organisms with the same fungus must under the Code carry the same name, which is most impractical. To remove this unintentional complication, one must either make an exception in the Code for these cases, or establish an informal system to take care of them. The latter seems to be preferable.

The different views on the taxonomic use of chemical characters in lichenology are reviewed using the opposing opinions of W. L. Culberson and Rogers and the somewhat intermediate position of Hawksworth as examples. The differing significance of secondary metabolites is illustrated with two examples in the genus Lecanora. In L. campestris and similar species, the chemistry has been useful to differentiate morphologically similar species, which subsequently were found to differ in minute anatomical differences as well. However, in L. caesiorubella, a remarkable chemical variation can be observed that is not correlated with other differences and is thus interpreted as intraspecific variation. An a posteriori valuation of the taxonomic significance of metabolic data is preferred.

New sequences have been produced of the ITS rDNA region of twenty-nine species of Parmeliaceae and analysed by neighbour-joining maximum likelihood distance analysis, and maximum parsimony analysis. Twelve monophyletic groups have been identified, most of which correspond to proposed generic segregated from Parmelia s. lat. Parmelia and Parmelina, Melanelia, and Punctelia appear as independent monophyletic groups; the type species of the two latter genera were, however, not included in the analysis. The parsimony analysis does not support Xanthoparmelia as a monophyletic group separate from Neofuscelia, and Rimelia reticulate is nested within Parmotrema. Parmelinopsis horrescens is more closely related to Hypotrachyna revolute than to Parmelina. Genetic distance values suggest that Flavoparmelia caperata is closer to Parmotrema than to other taxa included, and that Platismatia glauca and Hypogymnia tubulosa are closer to Parmelia s. lat. than to the Physciaceae. However, there is no conclusive support for any of thelatter groupings in the parsimony analysis. We conclude that the ITS region contains valuable information for studying generic delimitations within the family Parmeliaceae, but our study suggests that this region will not conclusively resolve within-family relationships in this group.

A phylogenetic analysis is presented, based on 57 morphological, anatomical, chemical and ecological characters, including 19 species from the genera Arthonia, Arthothelium and Syncesia. The aim of the study was to test the monophylyof some of the groups within the genus Arthonia suggested, for example, by Redinger (e.g. species with reddish ascomata, with grey-pruinose ascomata and with brown-black hypothecia). The results strongly support that Arthonia and Arthothelium are paraphyletic genera. The best-supported node contains all Arthonia species together with A. crozalsiana de Lesd., A. ruana A. Massal. (both hitherto placed in Arthothelium) and Syncesia myrticola (Fée) Tehler. A well-supported clade is formed by a group of pioneer species, often non-lichenized or poorly lichenized, including the type species of Arthonia, A. radiata. The species with reddish and/or K+ reddish ascomata form one clade and the species with more or less brownish or blackish hypothecia form another clade with Syncesia myrticola, the sister group to the Opegraphaceae and Roccellaceae. The results are discussed and compared with Redinger' grouping. Relationships to other genera within Arthoniaceae are briefly discussed. Arthothelium scandinavicumi Th. Fr., Arthonia dispersa (Schrad.) Nyl., A. punctiformis Ach. and A. mediella Nyl. are lectotypified.

A first hypothesis for the phylogeny of Lecanora subgen. Placodium is presented by using molecular data. Previous evolutionary ideas and classification concepts for this group with non-molecular data are re-investigated using DNA sequence data from the nuclear ITS and 5.8S regions. Using Protoparmelia as an outgroup, the Lecanora subfusca group together with the L. rupicola group appear as a sister group to assemblages with lobate species. Subgen. Placodium as currently accepted, is not monophyletic. Molecular data suggest that the Lecanora dispersa group and the L. polytropa group are widened by lobate species and there is evidence from the ITS data that the monotypic genus Arctopeltis Poelt is closely related to the L. dispersa group. A congruence between molecular data and secondary chemistry supports the broader concept of the L. dispersa group suggested by the molecular data.

After a brief historical survey, some current trends in the delimitation of lichen genera are discussed. A widespread tendency is that of elevating any supposedly monophyletic group of species to genus rank. As the term ‘ monophyletic ’ has no lower limit, this is likely to result in an explosive inflation of new genera, in a severe loss of the information carried by generic names and in a high degree of nomenclatural disorder. Five criteria are proposed for the acceptance or rejection of new generic segregates: (1) DNA testing for monophyly, (2) phylogenetic analysis, (3) number of characters used, (4) number of species considered, (5) information content of the new splittings. Upon a critical analysis of several recent generic segregations, a more flexible approach to taxonomic ranks is recommended, and particularly, when most of the suggested criteria are not fulfilled, a more frequent use of the subgeneric rank, which does not imply name changes.

Organismal species exist as historical entities, and essentialistic thinking about ‘ defining ’ species and higher taxa should be avoided. Lichens are biological phenomena that are best understood as processes, and they behave very differently from the model organisms of most systematic theories. Classification and phylogenetic reconstructions are seen as different endeavours, having different aims and languages. A hierarchical classification is supported, and recognition of superspecific taxa is seen as a matter of convenience. Superspecific taxa may be recognized as groups having correlated features, whereas the common recognition of such taxa by a priori cardinal characters should be abandoned. Superspecific taxa having correlated features are distinctive and are likely to be monophyletic. Distinctive and phylogenetically coherent grades may need naming to avoid nomenclatural instability. In classifications, the hypothesized monophyletic status of taxa may be indicated by some convention. Phylogenetic reconstruction by cladistic parsimony analyses includes assumptions that need scrutinizing and modifying to improve the methodology. Analyses should be examined with respect to robustness for changes in the frequently applied equal weighting assumption. For some types of data an a priori weight assignment might be possible, or equal weighting may be justified.

Ostropales, in a wide sense, currently encompasses Ostropales s. str., with a few lichenized genera, and Graphidales, most of which are lichenized. Conotrema is the only lichenized genus undoubtedly placed in Ostropales s. str. We have sequenced the 18S rRNA gene from some key taxa to answer the questions as to whether one ortwo orders should be recognized and whether Conotrema should be kept in Ostropales if Graphidales is recognized. We have new sequences from Conotrema populorum (Stictidaceae, Ostropales), Cyanodermella viridula (Stictidaceae Ostropales), Graphis scripta (Graphidaceae Graphidales) and Diploschistes ocellatus var. almeriensis (Thelotremataceae Graphidales). In an analysis with 41 taxa, thelichenized D. ocellatus var. almeriensis and G. scripta are the sister group of the lichenized C. populorum and the non-lichenized C. viridula and Stictis radiata. Together they form a monophyletic cluster supported by a bootstrap value of 82%. Therefore, it is possible to use the ordinal name Ostropales in a wide sense to include the Graphidales. If Ostropales are conceived in a narrow sense, Conotrema should be kept in the Stictidaceae. Eight new PCR primer sequences are reported.

In order to identify monophyletic groups within the family Parmeliaceae, eleven taxa (Bryoria capillaris, Cetraria islandica, Evernia pruniastri, Hypogymnia physodes, Parmelia saxatilis, Platismatia glauca, Pleurosticta acetabulum, Usneaflorida, Vulpicida juniperina, V. pinastri, and Xanthoparmelia conspersa) were studied using newly produced nuclear rDNA sequence data from the ITS and 5·8S regions. The resulting evolutionary hypothesis was compared with results from previous phylogenetic analyses based on anatomy, morphology, and chemistry. The outcome of this comparison does not support the earlier proposed phylogenies but is not stable enough for identifying monophyletic groups, with one exception. The results indicate a close relationship between Cetraria and Vulpicida, which is contradictory to previous published analyses. The variation in ascus structures in the Parmeliaceae is discussed and it is questioned whether the earlier distinguished ‘ forms ’ of ascus types represent synapomorphies, if they are based on poorly supported analyses, or if they are exaggerations of relatively slight variation in shape. Further interpretations of the results are discussed and areas of future studies based on DNA-data are suggested.

Traditional classification concepts in lichenology are often, but not always, supported by molecular results. Molecular data should be compared and correlated with micro-morphological and ultrastructural information before systematic rearrangements are undertaken. Visualization of the distribution of morphological and other characters in specified groups is considered as a desirable result per se, but it is also important to discover whether correlating characters are dependent on each other or not; and if not, whether their distribution in a group might support existing classification concepts. A data set for lecanoralean and other lichenized and lichenicolous genera, comprising 90—mostly multi-state— characters was used to store morphological, chemical and ecological data, and to test character correlations. Several examples of such analyses are presented. The following pairs of characters show some degree of dependence: ascospore septation and number per ascus, ascospore wall type and pigmentation, ascospore and epihymenium pigmentation. Several authors postulated that ascus types are good phylogenetic markers. Ascus types have been widely used for classification concepts of the Lecanorales. Two-dimensional correlation queries of ascus types with the following morphologcal characters were made: substratum preference, thallus growth form and ascospore septation. These correlations supply characteristic profiles for the various ascus types, which have to be compared with forthcoming phylogenetic hypotheses based on molecular data.

The use of ascoma ontogeny in the systematics of lichenized ascomycetes is reviewed briefly. Two recently studied examples are discussed to illustrate the use of ontogenetic studies to discover phylogenetic relationships. As one example, the Agyriaceae was mainly circumscribed by the ascus type and included diverse genera. Xylographa is characterized by lirelliform apothecia. However, the ascoma development shares the ontogenetic pattern observed in Trapelia and its placement in the family was confirmed. In contrast, Anamylopsora exhibited a substantially different apothecial development and was separated as a family of its own. As a second example, the Candelariaceae is based on chemical characters. With the detection of pulvinic acid derivatives also in the genus Lecanora, the distinction of the Candelariaceae and Lecanoraceae became doubtful. Therefore, the ascoma ontogeny was employed to evaluate this classification and is illustrated here in detail for Candelaria fibrosa. The substantial similarity of the ascoma development found in both families suggests that both might be better united. Finally, a general discussion of the significance of ontogenetic characters in lichen taxonomy is included.

The integration of the taxonomy of lichenized and non-lichenized pyrenocarpous ascomycetes is still in progress. Within this group, at least 12 lichenization events, three re-lichenizations and at least 24 de-lichenization events, are supposed to have occurred during evolution. This is evident from the increasing amount of information regarding phylogeny and classification of pyrenocarpous fungi. The total number of lichenization events, including re- and de-lichenization, within all fungi is estimated to be at least 100. However, the vast majority of the pyrenocarpous lichens are concentrated in a few relatively large, presumably monophyletic, groups. In some cases, proposed integrations have proven to be untenable, for example Microtheliopsis and with the Herpotrichiellaceae, Flavobathelium with the Dimeriaceae and most notably the Normandina thallus with the basidiomycetes and its ascocarps with Sphaendina in the Mycosphaerellaceae. The same true for many cases where lichenized taxa were unwittingly included in otherwise non-lichenized groups, such as Didymosphaeria and Massarina, or even described new, presumably non-lichenized, genera like Holstiella, Starbaeckiella and Titanella.