Mycosphere

The problems of traditional and phylogenetic of fungi

Vasilyeva LN1* and Stephenson SL2

1Institute of Biology and Soil Science, Far East Branch of the Russian Academy of Sciences,Vladivostok 690022, Russia, [email protected] 2Department of Biological Sciences, University of Arkansas, Fayetteville, Arkansas 72701, USA, [email protected]

Vasilyeva LN, Stephenson SL 2010 – The problems of traditional and phylogenetic taxonomy of fungi. Mycosphere 1, 45 – 51.

The current taxonomic system, based on the hierarchy of Linnaean ranks, is said to be unsatisfactory. Many people insist that the very concept of rank should be eliminated, and the current taxonomic system should be replaced with a rankless system based on phylogenetic taxonomy. However, it is not the elimination of the concept of rank but rather the correct ranking of characters that will promote stability in taxonomy. The first problem of both traditional and phylogenetic taxonomies is the chaotic employment of differences that leads to the segregation of groups which are not comparable in rank. A proper rank coordination could be carried out by the special weighting of differences among tentative groups. Groups become equal in rank when they are distinguished by state combinations of the same character set. The second problem of both traditional and phylogenetic taxonomies is the employment of internal polymorphism of genera and families for the re-unification of into new taxa. This could be overcome if taxonomists take into consideration the law of homologous variation within closely related genera and families.

Key words – character weighting – ranks – taxonomic system

Article Information Received 10 February 2010 Accepted 24 March 2010 Published online 9 April 2010 *Corresponding author: Vasilyeva LN – e-mail – [email protected]

Introduction because of few available ranks, whereas “a Already it has been 15 years since the large number of ranks are necessary to classify statement was made that “in mycological complex phylogenetic trees” (l. c.). On the taxonomy, the rift between traditional mor- other hand, the phylogenetic taxonomists say phologists and molecular phylogeneticists that the very “concept of rank should be remains unbridged” (Seifert et al. 1995: S760). eliminated, and the current taxonomic system In this opposition, the contrast seems to exist should be replaced with a rankless system between morphology and phylogeny, as well as based on phylogenetic taxonomy” (l. c., italics between the words ‘traditional’ and ‘mole- added). These two contradictory demands— cular’. However, both morphology and many ranks vs. a rankless system—concern molecules can serve phylogeny, and one meets two different aspects of phylogenetic theory. with the real opposition between traditional and The notion of a ‘rankless system’ seems phylogenetic methodologies. Supposedly, “the to be the consequence of a world view called current taxonomic system, based on the the “great chain of being”. The latter held a hierarchy of Linnaean ranks is not satisfactory” central place in Western thought for centuries (Hibbett & Donoghue 1998: 347). On the one and saw the organic world as ordered in a hand, this hierarchy seems to be unsatisfactory linear sequence (Nee 2005). Such a sequence

45 does not entail the appearance of ranks, but and descendants). If one could establish the modern evolutionary views suggest a hierar- sequence of characters correctly, one would chical development of the biota. “When we find the only structure that retains stability in depict evolutionary relationships in the form of the continuing process of change. Thereby, one a tree, we acknowledge that genealogical would come to the most natural system relationships are hierarchical by nature” reflecting the hierarchical evolution. (Spatafora & Blackwell 1994: 233, italics Therefore, it is not the elimination of the added). very concept of rank that “promotes It should be emphasized, however, that nomenclatural stability” (Hibbett & Donoghue both world views do not exclude each other. 1998: 347) but rather it is the proper ranking of Living or extinct beings do comprise a se- characters for the groups being considered. quence, while relationships among organisms The problem of ranking is the gist of taxonomy, might be hierarchical. It is currently known that although both traditional and phylogenetic the original gene pool which once existed on taxonomy seem to have failed in resolving this the Earth split into smaller pools, and they are problem. It is true that the current taxonomic now called Bacteria, Archaea and Eucarya system is not satisfactory, but this is not (Woese 2000). The differences among them are because it is based on the hierarchy of the oldest fundamental differences that can be Linnaean ranks. Instead, the situation exists recognized for living organisms and have because there was no method for finding the persisted from their very appearance to the place of a character in the hierarchy. present day. Even if taxonomists could have a kind of Later, the domain Eucarya was split into a ‘rough’ hierarchy on the basis that some a number of kingdoms, and their differences— characters cover more extensive groups than once again—remained to be the same during others, the choice of characters for group the further differentiation of Eucarya into delimitation at each level is often haphazard. smaller groups. While almost all animal phyla The chaotic employment of any ‘unique were established during the ‘Cambrian character’ leads to groups that are similar to explosion’ (Marshall 2006), appearances of ones in the ancient Chinese encyclopedia new groups during the ‘Ordivician radiation’ described by the famous Argentine writer Jorge were manifested at lower taxonomic levels Luis Borges. In that encyclopedia, animals (Droser & Finnegan 2003). As a result, the were divided into many groups, including such organic world represents a nested hierarchy of examples as (1) belonging to the emperor, (2) groups (Fig. 1). embalmed, (3) fabulous, (4) stray dogs, (5) Thus, the same relationships exist among suckling pigs, (6) those that look like flies from taxonomic groups at the same level through a long way off, and so on. their evolution. These relationships could be Such a ‘classification’ is quite com- expressed in differences that have appeared in patible with the practice of constructing a key sequence during the differentiation of the for the identification of groups assigned to the organic world. When Linnaeus first introduced same level (species, genera, or families). For the very fruitful idea of the hierarchical example, the delimitation of genera in the ordering, he unintentionally created a taxo- family (, Ascomy- nomic model of hierarchical evolution. The cota) goes as follows: Trichomerium differs Linnaean hierarchy conveys both the nesting of from all other genera in having sessile fruit small groups within the larger ones and the bodies, Scorias is characterized by a ‘stromatal distribution of characters among levels in habit’ of those fruit bodies, Capnodium accordance with their appearance in time. possesses pigmented ascospores, Limacinula is The sequence of characters marking based on the presence of thin-walled hyphae different levels is the only evolutionary line below the ascomata, while Trichopeltheca has (‘natural scale’) in the hierarchical system, so no periphysoids (Reynolds 1986). there is no need to arrange organisms in a The dichotomous key for these genera is ‘chain of beings’ (i.e. the construction of similar to a cladogram and is considered to system does not require the tracing of ancestors represent ‘phylogenetic relationships’. In fact,

46 Mycosphere

Fig. 1. The organic world depicted as a nested hierarchy of groups. the first problem shared by traditional and considered to correspond to a different level of phylogenetic taxonomy is the ‘principle of hierarchy and also why phylogeneticists dichotomy’. The first hierarchical dichotomy require “a large number of ranks… to classify could be found in Plato's dialogue Sophiste, complex phylogenetic trees” (Hibbett & although its invention is attributed to the Donoghue 1998: 347). neoplatonic philosopher Porphyry. The famous One can understand the key-like and ‘Porphyrian tree’ is known to be a tree-like artificial nature of many ‘phylogenetic trees’ construction where points of branching are from the statement that “over 12,000 equally marked by the differences of successively parsimonious trees are possible for the 142 smaller groups. However many dichotomies do taxa” (Parmasto 1995: S843). Evidently, not convey hierarchical relationships at all. 12,000 different keys could be constructed, The keys of traditional taxonomy that are depending upon the choice of the first taxon based on chaotic employment of ‘unique’ which is pared off from all of the others, but character states for the segregation of species the true hierarchy of characters should be a or genera are clearly artificial but they, at least, single one. do not aim at the ‘phylogeny reconstruction’. It In the true hierarchy, many characters are is worse that phylogenetic taxonomy—using equal in level, and our task is to find such precisely the same approach to characters— character complexes. In his discussion of proclaims that a ‘dichotomy principle’ is the molecular characters and their significance for only one which should be used to show the real fungal systematics Hibbett (1992: 534) was and hierarchical relationships. This is why each quite right in stating that “molecular characters, point of branching in a cladogram is often though informative in many cases, are not the

47 ‘Rosetta stone’ of systematics”; in fact, they to as a clypeus. The presence or absence of a are subject to many of the same problems as clypeus distinguishes and morphological characters, and “the characters Mazzantia, and Melanopelta, themselves are only as good as the method of Apioplagiostoma and Hypospilina, analysis”. In the section that follows, the and . The further development of method of analysis aiming at the testing taxa stromatic tissue leads to the appearance of a for rank equality will be discussed. stromatic capsule that surrounds the fruit body. First, this method does not allow us to Moreover, this kind of fruit body combines employ chaotic differences for the segregation with different kinds of spores, as is the case for of taxa of equal rank. It is based on the very old Mamianiella (amerospores) and Mamiania rule discussed already by Linnaeus, who used (apiospores). such words as ‘unique idea’ for a set of The in two monographs characters distinguishing genera in the natural (Barr 1978, Monod 1983) also includes the order (Philosophia Botanica-187). That was his Plagiosphaera, which shares its place way of saying that taxonomic groups are with Pleuroceras by having the same state comparable in level, or rank, only when they combination of characters in question. One are defined by state combinations of the same treatment (Barr 1978) distinguishes these character set. When this rule is applied in genera on the basis that members of practice, taxa of the same level are usually Pleuroceras occur on dead leaves of deciduous arranged in the combinatorial—multidimen- trees, whereas members of Plagiosphaera sional - space of character states (Vasilyeva occur on herbaceous stalks. Such a difference 1999). does not allow us to consider these two taxa as One should not take any characters with separate genera, since many closely related states combining with each other. One should genera, such as , Gnomonia or weight delimitative characters in accordance Gnomoniella include members that occur both with the frequency of their participation in on leaves of trees and herbaceous stalks. As delimitation. When a taxonomist begins to such, this character displays variability inside study, for example, a family, he or she wants to genera (element of the internal polymorphism) know all genera that were placed into that in this group and cannot be used to divide some family but cannot always estimate the other genera. differences between them properly. A good In the other treatment (Monod 1983), one example is the family Gnomoniaceae (Dia- can see that and Pleuroceras porthales, ). Different authors have are distinguished from Plagiosphaera because assigned about 40 genera to this family and these two genera lack paraphyses, as do all have made use of a number of distinguishing other members of the Gnomoniaceae. In such a characters. case, Plagiosphaera should belong to a The most frequently used distinguishing different order, since the presence or absence character was the kind of spores, which of paraphyses is of value at that high level. The delimits Gnomoniella (amerospores) Apiogno- removal of Plagiosphaera from the Gnomo- monia (apiospores), Gnomonia (didymospores) niaceae would increase both the naturalness of and Ophiognomonia (scolecospores) (Fig. 2). the family composition and the naturalness of A second character often used for dis- character hierarchy, since one false ‘generic’ tinguishing genera in this family is the position character (occurrences on the leaves of of the elongated perithecial beak—it may be deciduous trees or herbaceous stalks) became central or lateral as in and lower in level, whereas another character Apioplagiostoma, Gnomonia and Plagiostoma, (presence or absence of paraphyses) used for Ophiognomonia and Pleuroceras. genera delimitation appears to be higher in A third character used most frequently to level. describe new genera in this family is the kind A similar investigation of other situations of fruit body. The latter could be simple in which two or more genera fall in the same without any stromatic development, or such a place in accordance with the distribution along stromatic development is present and referred the most frequent delimitative characters

48 Mycosphere

Fig. 2. Characters that have been used to distinguish different genera in the family Gnomoniaceae. allows us to go further in making the existing has a reticulate structure (de Hoog 1979) in system more natural without eliminating the opposition to a hierarchical system. However, Linnaean ranks. Moreover, in addition to the opposition is imaginary, since reticulations containing groups that are ranked properly in are particular phenomena at different levels of relation to each other, the natural system the hierarchy. What does really make a contrast should be highly prognostic (de Hoog 1981). is reticulate versus dichotomous arrangement Evidently, the empty places at each level of groups at the same level. can predict the possible—living, fossil, Since dichotomous constructions do not forthcoming, or yet unfound—organisms with convey real phylogenetic relationships among certain character state combinations. Some groups of the same rank (see above), how expected combinations in the family Gnomo- might these groups originate? There is no need niaceae really do exist, but - without the to talk about a special 'reticulate evolution', method of a posteriori estimation of delimi- since they appear as the result of Darwinian tative characters - these combinations are ‘indefinite’ variation, which simply means that buried in wrong places. Thus, the genus any character of an organism can change in a Chalcosphaeria (didymospores + clypeus + good or bad direction. lateral perithecial beaks) was described almost One can turn to a large group, for a century ago, but its name was reduced to example a genus containing numerous species, synonyms of either Plagiostoma (Barr 1978) or and expect that one of the species combinations Hypospilina (Monod 1983). Another example might be ancestral to some others. The species is Mamiania alni (didymospores + central that differs from many others at least in one beaks + stromatic capsule), which might character is the most suitable candidate for an deserve its own genus that is not yet described. ancestor, and one can designate some of its It might seem curious that the characters by capital letters (ABCD). All of improvement of a character hierarchy is carried these characters can change simultaneously but out with the help of combinatorial con- in different descendants, and one can find structions. Some authors consider such descendant combinations such as aBCD, AbCD, constructions as a special kind of a system that ABcD, ABCd.

49 Now, if only one character changes in What is it that phylogeneticists really do? each case, any descendant shares with an They take species, for example from the genera ancestor almost the same genetic program and Gnomonia and Plagiostoma, and involve them the ability to change in the same directions. in phylogenetic analysis. The same features of Therefore, already the third generation of these species could be apomorphic and descendants might represent more diverse plesiomorphic—at the species level—within combinations (i.e., abCD, abcD, aBCd, AbcD, both genera, but the employment of the AbCd, ABcd, etc.). After all, a combination ‘synapomophy principle’ will unite consisting of completely apomorphic states ‘apomorphic species’ from two genera into a (abcd) might originate, and the combinatorial new heterogeneous group. Thereby, space constructed with the help of four phylogeneticists make a change at the generic changing characters will be completed with level by using species characters. There are groups of the same level. even some drastic changes at much higher One should take into account that every levels—one can look at the treatment of such diversification includes one group that is erysiphaceous fungi among the leotioid characterized only by plesiomorphies and one discomycetes (Lutzoni et al. 2004: 1457). group that possesses exclusively apomorphies. The second blunder of the phylogenetic Character states of all other groups are approach is associated with the fact that mixtures of apomorphies and plesiomorphies in characters of internal polymorphism always different proportions. This makes the second outnumber the few differences that exist principle of phylogenetic theory, namely the between closely related genera or families. In ‘principle of synapomophy’, invalid. Similar to molecular taxonomy, which provides many the ‘dichotomy principle’, it merely creates more characters than morphology, the heterogeneous groups and increases the characters of internal polymorphism become an artificiality of systems. overwhelming majority. Of course, species of Except for cases in which synapo- many genera of the family Gnomoniaceae (and morphies are states of delimitative characters, even the whole order ) could they might occur in the internal polymorphism appear near each other on a ‘molecular tree’ of closely related taxa. It is very important to (Castlebury et al. 2002). There is nothing keep in mind the extent to which such ‘phylogenetic’ in such an approach; it is a polymorphism can exist. For example, the two purely 'numerical' taxonomy which makes genera Gnomonia and Plagiostoma differ in the unifications on the basis of 'overall similarity' position of the perithecial beak (central or and even does not envisage the testing of lateral) and are placed in different families by tentative groups by weighting their delimitative those authors who overestimate the level of this characters. character (Barr 1978; Lumbsch & Huhndorf 2007). However, these genera have exactly the References same internal polymorphism (i.e. repetitive characters that distinguish species within them). Barr ME 1978 – The Diaporthales in North Those repetitive characters are the basis America with emphasis on Gnomonia of the law of homologous variation (Vavilov and its segregates. Mycologia Memoir 7, 1922), which tells us that the more complete 1–232. the series of repetitive characters in two genera Castlebury LA, Rossman AY, Jaklitsch W, (or families), the more closely related these Vasilyeva LN 2002 – A preliminary genera (or families) are. Therefore, the position overview of the Diaporthales based on of Gnomonia and Plagiostoma within the same large subunit nuclear ribosomal DNA. family is supported by their similar internal Mycologia 94, 1017–1031. polymorphism. Unfortunately, the “phylogene- Droser ML, Finnegan S 2003 – The Ordovician tic” approach is known for two blunders when Radiation: a follow-up to the Cambrian that polymorphism is concerned. explosion? Integrative and Comparative Biology 43, 178–184.

50 Mycosphere

Hibbett DS 1992 – Ribosomal RNA and fungal Journal of Botany 73(Suppl.), S843– systematics. Transactions of the Mycolo- S852. gical Society of Japan 33, 533–556. Reynolds DR 1986 – Foliicolous ascomycetes. Hibbett DS, Donoghue MJ 1998 – Integrating 7. Phylogenetic systematics of the phylogenetic analysis and classification Capnodiaceae. Mycotaxon 27, 377–403. in fungi. Mycologia 90, 347–356. Seifert KA, Wingfield BD, Wingfield MJ 1995 Hoog GS de 1979 – Deductive classification— – A critique of DNA sequence analysis in worked examples using anamorph and the taxonomy of filamentous Ascomy- teleomorph data in the Ascomycetes. In: cetes and ascomycetous anamorphs. The whole . Vol. 1. (ed B Canadian Journal of Botany 73(Suppl.), Kendrick). Ottawa: National Museum of S760–S767. Canada, 215–239. Spatafora JW, Blackwell M 1994 – Cladistic Hoog GS de 1981 – Methodology of taxonomy. analysis of partial 35rDNA sequences Taxon 30, 779–783. among unitunicate perithecial Ascomy- Lumbsch T, Huhndorf SM. 2007 – Outline of cetes and its implications on the Ascomycota - 2007. Myconet 13, 1–58. evolution of centrum development. Lutzoni F, et al. 2004 – Assembling the fungal Ascomycete systematics (ed. DL tree of life: progress, classification, and Hawksworth). New York & London, evolution of subcellular traits. American 233–242. Journal of Botany 91, 1446–1480. Vasilyeva LN 1999 – Systematics in mycology. Marshall CR 2006. Explaining the Cambrian Bibliotheca Mycologica 178, 1–253. “explosion” of animals. Annual Review Vavilov NI 1922 – The law of homological of Earth and Planetary Sciences 34, 355– series in variation. Journal of Genetics 12, 384. 47–89. Monod M 1983 – Monographie taxonomique Woese CR 2000 – Interpreting the universal des Gnomoniaceae. Beiheft zur Sydowia phylogenetic tree. Proceedings of the 9, 1–315. National Academy of Sciences of the Nee S 2005 – The great chain of being. Nature USA 97, 8392–8396 435, 429. Parmasto E 1995 – Corticioid fungi: a cladistic study of a paraphyletic group. Canadian

51