Species concepts in lichenology (by O.W. Purvis)

A previous discussion forum (IAL Newsletter 30, 1, 1997) focussed on generic concepts in lichens, particularly the relative merits of splitting large and unwieldy genera. I would like to initiate discussion at a lower level, the species, arguably the most important rank as these are (or should be) the practical labels we give to organisms which everyone uses. There has been some interesting recent discussion on species concepts (see Lichenologist 30, 4-5, e.g. Clerc pp. 321-340; Tibell pp 439-453). I wish to briefly consider: (1) species delimitation; (2) species and the users of taxonomy.

1. Species delimitation - Lichen species do not exist in the same way as they do for other organisms. Lichens are composite organisms, mini-ecosystems and do not have an integrated genome. We should not worry about this too much, but how we define the entities we call lichen species is very important. Ove Almborn (Särtryck ur Botaniska Notiser, 4: 454-7, 1965) highlighted the need for taxonomy to be based on 'firm principles' not deviating from other plant groups. But we know little about sexual reproduction in lichenised ascomycetes, and lichens are difficult experimental organisms under laboratory conditions. Modern science demands formulation of testable hypotheses. Lichen taxonomy is heavily based on a morphological species concept relying on phenetic characters including secondary metabolites which are assumed to reflect underlying genetic differences. Monographers rarely discuss in any detail the criteria they use to delimit species which may of course vary considerably from genus to genus. This is a pity and should be encouraged to stimulate more proactive debate in the wider lichenological community. Characterising lichen species depends on understanding the sources of phenotypic variation. Ecological factors must therefore be considered as e.g. pointed out by D. Hawksworth in 1973: "Species concepts are currently based on sharp discontinuities in one or several morphological and anatomical characters, particularly where there is evidence that genotypic differences are involved (e.g. the two entities growing side by side in a uniform ecological situation and retaining their identities) or there are differences in either ecological requirements or geographical distribution, or both". Whilst we cannot rely on always finding similar species in the same environment to enable comparison we have the potential nowadays to investigate genotypic differences. Lichen transplants also open up intriguing possibilities to test species delimitations and this is an area which can benefit from a multidisciplinary approach. There is great potential for lichenologists to work with scientists in other disciplines.

2 Species and the users of Taxonomy - Following the Rio Summit, there has probably never been a greater awareness of the importance of biodiversity. Species lists are appearing with increasing frequency in a variety of media, biodiversity action plans, www, legislation etc. Whilst this is to be applauded, this has profound implications for us all. My concern lies particularly with the aplication of names for our commoner lichens as these are the ones beginners are most likely to first encounter. They are also often chosen for experimental purposes (e.g. Peltigera for elucidating carbohydrate transfer). Common species often have a wide ecological amplitude and usually exhibit the greatest range of morphological variation. Some variation may be significant but only discernible using sophisticated techniques (histochemistry, SEM, TEM, secondary metabolites, molecular). Lecanora dispersa is one case in point, there are many others. I predict we will find new, 'diagnostic' characters in lichens with every new technique we apply. The temptation will be to describe new entities at specific level. I suggest our taxonomy should also consider the end-user. At a practical level, compilers of checklists, etc., could usefully include the term 'aggregate' as well as indicating the species within it. For computer based keys this is no great problem. For many users this will be the level of precision we can expect. The term "aggregate" may also encompass a different range of species according to different concepts in different countries. This will give recorders an opportunity to record at an appropriate level and its use is often essential when reevaluating previous records where no specimen exists. Certainly, if our criteria for classifying lichens at a species level are too narrow and require too sophisticated techniques for interpretation, we run the risk of producing lists and floras which will only be comprehensible to other lichenologists with access to the necessary techniques. This will exclude many other potential end users and is not in the spirit of the Biodiversity Convention.

O. William Purvis, London

Responding to the considerations by Purvis, we would like to stress that phylogenetic concepts based on molecular (DNA) characters are increasingly being employed to recognise species in species complexes, to determine a posteriori which types of phenotypic characters are good predictors of phylogenetic species, and to demonstrate how these characters evolve in lichenized fungi. For this purpose, sequence data of the more rapidly evolving regions of nrDNA such as ITS have been used: Differences between ITS sequences were recently taken as support for the description of a new species (Peltigera phyllidiosa, Goffinet & Miadlikowska, Lichenologist 31, 1999), and to confirm separation of closely related species (Ramalina panizzei from R. fastigiata, Groner & LaGreca, Lichenologist 29, 1997; Lasallia rossica from L. papularis, Niu & Wei, Mycosystema 6, 1993; Teloschistes lacunosus from T. villosus, Martin & Llimona in: Grube & Wedin (eds): Progress in Molecular Studies of Lichens, Graz 1998). However, a single locus will consistently recognise species as discrete clades of alleles that correlate with fixed phenotypic characters only when species have been genetically isolated for a sufficient period of time. Molecular characters, like phenotypic characters, may be polymorphic within a phylogenetic species, and across sibling species of a species complex. If a locus has not coalesced to different alleles among species, that locus will not separate those closely related species. Therefore, in cases where a single locus does not clearly resolve species correlated with phenotypic characters or biogeography, the addition of independently sorting genetic loci in studies of these species complexes is needed to provide a better estimate of species boundaries. The mycobiont genome can be sampled with a sequence-based approach to provide a number of independent, unlinked genetic loci. With the analysis of a multilocus data set, the phylogenetic species concept becomes an empirical approach, as phylogenetic species are recognised by the concordance of gene genealogies. Species are ranked at the level where genetic reticulation ends and genetic isolation begins, indicated on the multilocus tree by poorly resolved clades that are supported by robust branches. The loci that are found to be polymorphic within a phylogenetic species are then used for population genetic analysis. The incongruence of gene genealogies within species indicates that recombination has occurred among individual mycobiont genotypes, suggesting that sexual reproduction occurs in lichenized fungi. Therefore, species concepts based on other eukaryotic groups in which sexual recombination is known and well characterized are also applicable to lichenized fungi. These data will also show which phenotypic characters are monomorphic within a phylogenetic species and are possibly good indicators of phylogenetic species in other related species complexes in which molecular data are not available. As for polymorphic phenotypic characters within a phylogenetic species, the role of genetics cf. environment can be inferred by ecological distribution of the variation.

Scott Kroken and Martin Grube, Graz

I am grateful to W.Purvis for having initiated this interesting discussion. I appreciated particularly point 2 (Species and the users of taxonomy), and I would like to contribute by adding a few thoughts and citations which might be of general interest. The first citation is from Claridge, Dawah & Wilson (1997 - Species: The Units of Diversity, p. 388): "It is not uncommon to find in discussions of species and species concepts researchers confusing empirical data used in the operation of recognizing a species with a conceptualization or definition of species. Empirical data can include such things as anatomy, morphology, genetics (DNA, proteins), behaviour, etc., all possibly evaluated and analysed in a variety of ways and with a variety of methods. Our abilities to gather these data are artificially constrained by technological advances; that is, we can only collect data that current technology permits. These artificial constraints on our ability to perceive variation in nature should not be confused with our desire, objectives, or attempts to illuminate natural variation". I agree with Purvis when he writes that an all too narrow species concept will exclude many other potential users, and is not in the spirit of the Biodiversity Convention. Taxonomists must have end-user in mind. I also agree with an earlier statement of Purvis (in Claridge et al., 1977: p. 129): "The phylogenetic species concept has yet to play an important role in defining species... The biological species concept is inappropriate for lichens owing to technical problems in studying breeding behaviour in culture. In conclusion, lichen species are based on clear discontinuities in one or more unrelated fungal characters". Traditionally, the "morphological species concept" is the most practical for checklists, and especially for identification keys which can be used also by non-lichenologists. The discussion of "too narrow criteria classifying lichens" may be similar to that which confronts "splitters" versus "lumpers". According to Stuessy (1990: Plant Taxonomy: the Systematic Evaluation of Comparative Data, Columbia Univ. Press): "Splitters (narrow concept) tend to believe that morphological variations of a "minor" nature should be documented formally by the description of new taxa, whereas lumpers may observe the same variations but believe that their formal recognition is neither necessary nor desirable".

Josefina Alvarez-Andrés, Vigo

Albeit not new, the potential conflict between providers and users is becoming a sensitive issue in taxonomy. Due to higher international interest in biodiversity, and to the rapid development of computerized databases, the pressure for nomenclatural stability is increasing, and taxonomists will have to consider this point more than they did in the past. I still maintain that generic concepts are more relevant in this context. As far as species are concerned, we still do not have a clear, operationally valid and widely accepted definition of what a lichen "species" is (for a review see also Kärnefelt, Abstracta Botanica, 1997). Because of this unfortunate and somehow embarassing situation, we cannot set any clear limit to the circumscription and description of new "species". Sometimes new species are described just to attract the attention of students to a peculiar morph, which is not always a bad thing. The case of Lecanora dispersa s.lat. is a good example: several "species" were described or resurrected in this clearly heterogeneous assemblage (see e.g. Poelt et al., Bibl. Lichenol., 58, 1995). This has rendered identifications more difficult, and perhaps not all of these "species" will survive DNA analysis. However, such a preliminary arrangement makes more justice to reality than it does "Lecanora dispersa s.lat.", and paves the way for further progress. Several "chemical" species are now considered as part of the natural variation of a taxon, but many proved to correlate with other characters, and are maintained as indipendent species. The point raised by Purvis is likely to provoke two types of reactions by two main groups of taxonomic providers: a) Taxonomists will rightly claim that they should be left free to describe "species", even if based on cryptic characters, because this is their job: nature cannot be always interpreted with a lens and a few colour reactions; b) Authors of floras could try to work out keys which are more user-friendly, accepting several degrees of approximation, and making a larger use of "aggregates". Some of the keys by Clauzade & Roux's Determinlibro (1985), e.g. those of Acarospora, already tried to solve this problem by a generous adoption of infraspecific taxa. Originally I did not like this solution, but on the light of this discussion, I have now the impression that it was a good idea.

Pier Luigi Nimis, Trieste