This Article Abstract Full Text (PDF) Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Huhndorf, S. M. Articles by Fernández, F. A. Search for Related Content PubMed Articles by Huhndorf, S. M. Articles by Fernández, F. A. Agricola Articles by Huhndorf, S. M. Articles by Fernández, F. A.

Molecular evolution and systematics

Molecular systematics of the Sordariales: the order and the family Lasiosphaeriaceae redefined

     Botany Department, The Field Museum, 1400 S. Lake Shore Drive, Chicago, Illinois 60605-2496

Andrew N. Miller

     Botany Department, The Field Museum, 1400 S. Lake Shore Drive, Chicago, Illinois 60605-2496 University of Illinois at Chicago, Department of Biological Sciences, Chicago, Illinois 60607-7060

Fernando A. Fernández

     Botany Department, The Field Museum, 1400 S. Lake Shore Drive, Chicago, Illinois 60605-2496

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 

The Sordariales is a taxonomically diverse group that has contained from seven to 14 families in recent years. The largest family is the Lasiosphaer-iaceae, which has contained between 33 and 53 genera, depending on the chosen classification. To determine the affinities and taxonomic placement of the Lasiosphaeriaceae and other families in the Sordariales, taxa representing every family in the Sordariales and most of the genera in the Lasiosphaeriaceae were targeted for phylogenetic analysis using partial sequences of the large-subunit (LSU) nrDNA. Based on molecular data, only genera within the families Chaetomiaceae, Lasiosphaeriaceae and Sordariaceae are retained within the redefined Sordariales. The order is a coherent group with morphologies that vary along well-defined lines, including large ascomata with large-celled membraneous or coriaceous walls and ascospores that show variation on a distinctive developmental theme, often with appendages or sheaths. The paraphyletic chaetomiaceous complex and the strongly supported Sordariaceae are nested among taxa traditionally placed in the Lasiosphaeriaceae. Analyses also indicate that 11 genera belong in the paraphyletic lasiosphaeriaceous complex. These taxa share a similar developmental pattern in their ascospore morphology that extends to the Sordariales as a whole. Based on these similarities in morphology, 13 additional genera are retained within the lasiosphaeriaceous complex and more than 35 genera have relationships in the order overall. Based on LSU data, 17 genera that have been assigned to the Lasiosphaeriaceae sensu lato are transferred to other families outside the Sordariales and 22 additional genera with differing morphologies subsequently are transferred out of the order. Two new orders, Coniochaetales and Chaetosphaeriales, are recognized for the families Coniochaetaceae and Chaetosphaeriaceae respectively. The Boliniaceae is accepted in the Boliniales, and the Nitschkiaceae is accepted in the Coronophorales. Annulatascaceae and Cephalothecaceae are placed in Sordariomycetidae inc. sed., and Batistiaceae is placed in the Euascomycetes inc. sed.

Key words: Annulatascaceae, Batistiaceae, Boliniaceae, Catabotrydaceae, Cephalothecaceae, Ceratostomataceae, Chaetomiaceae, Coniochaetaceae, Helminthosphaeriaceae, LSU nrDNA, Nitschkiaceae, Sordariaceae

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 
The Sordariales is one of the most taxonomically diverse groups within the Class Sordariomycetes (Phylum Ascomycota, Subphylum Pezizomycotina, fide Eriksson et al 2001). The order contains lignicolous, herbicolous and coprophilous taxa currently arranged in approximately 115 genera divided among seven families but recently has contained up to 14 families (TABLE I). These families have been placed in the Sordariales based on the putative ascomal ontogenetic characters involved in centrum development (Luttrell 1951). Two of the families (Batistiaceae, Catabotrydaceae) are monotypic, and nine others (Annulatascaceae, Boliniaceae, Cephalothecaceae, Ceratostomataceae, Chaetomiaceae, Coniochaetaceae, Helminthosphaeriaceae, Nitschkiaceae, Sordariaceae) contain fewer than 15 genera each (Eriksson et al 2001). The Lasiosphaeriaceae is by far the largest and most morphologically diverse family in the order (TABLE II). Although it contained up to 53 genera (Eriksson and Hawksworth 1998), it currently is more narrowly circumscribed with 33 genera (Kirk et al 2001).

Selected members representing the Sordariales have been included in several molecular analyses, where they often have resolved as a monophyletic group (Spatafora and Blackwell 1994, Andersson et al 1995, Lee and Hanlin 1999, Suh and Blackwell 1999, Zhang and Blackwell 2002). In some of these analyses, taxa formerly thought to have relationships in the Sordariales were found to have their affinities elsewhere (i.e., Ceratocystis in Spatafora and Blackwell 1994, Melanospora Corda in Zhang and Blackwell 2002). In a number of these analyses, taxa under investigation were found to have close relationships to the Sordariales based on very limited taxon sampling (Fallah et al 1997, Chen et al 1999). Analyses never have been performed that have included representatives from all of the putative families in the order and a majority of genera in the Lasiosphaeriaceae. To determine the affinities and taxonomic placement of the Lasiosphaeriaceae and other families in the Sordariales, taxa representing every family in the Sordariales and a majority of the genera in the Lasiosphaeriaceae were targeted for phylogenetic analysis using partial sequences of the large subunit (LSU) nrDNA. Some questions we considered were: (i) Which families belong within the Sordariales? (ii) What is the placement of families excluded from the order? (iii) Which genera belong within the Lasiosphaeriaceae? (iv) What is the placement of genera excluded from the Lasiosphaeriaceae? and (v) How is the Lasiosphaeriaceae related to other families in the Sordariales and other groups of pyrenomycetes?

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 
Taxon sampling. – Taxa used in this study are listed in TABLES III and IV, along with their geographical locality, collector, voucher specimen and/or isolate number, and GenBank accession number. Cultures of multispore isolates were obtained by spreading centrum material from air-dried specimens onto 1% water agar (Difco) in 60 mm diam plastic Petri plates. After 24–48 h of incubation at room temperature, germinated asci and ascospores were transferred to 60 mm diam plastic Petri plates containing 1% cornmeal agar (Difco). Cultures were maintained at 10 C on 1% potato-dextrose agar (Difco) slants in 6 dram screw-cap tubes. All voucher specimens are deposited in the Field Museum Mycology Herbarium (F).

PCR reaction mixtures contained: 2.5 µL of 10x reaction buffer (100 µm Tris, 500 µm KCl) with 25 mM MgCl₂, 2.5 µL of 8 mM d-NTPs, 2.5 µL each of 10 µM primers, 3 units of Taq DNA polymerase, 0.5–2 µL of undiluted DNA extract and double-distilled sterile water to bring the total volume to 25 µL. Occasionally 2.5 uL of 10 ng/µL BSA (bovine serum albumin) and/or 2.5 µL of 50% DMSO were added to the PCR reactions to increase efficiency. The primers LROR and LR6 or LR7 (Vilgalys and Hester 1990, Rehner and Samuels 1995; www.biology.duke.edu/fungi/mycolab/primers.htm) were used to amplify 1100 bp or 1400 bp, respectively, of the 5' end of the LSU gene. All PCR reactions were performed on either a MJ Research PTC 200 or PTC 220 Dyad thermocycler using these thermocycling parameters: initial denaturation at 94 C for 2 min, followed by 30 or 40 cycles of 94 C for 30 s, 47 C for 15 s and 72 C for 1 min with a final extension step of 72 C for 10 min. PCR products were visualized on a 1% TBE agarose gel stained with ethidiumbromide. If these methods failed to produce any amplification product, PCR beads (Amersham Pharmacia Biotech, Piscataway, New Jersey) were used according to the manufacturer’s instructions. Thermocycling parameters were as above except initial denaturation was 5 min and annealing temperature occasionally was decreased to 41 C. Further amplification failure required the gene to be amplified in two overlapping fragments using the primer pairs LROR-LR3 and LRAM1-LR6 (LRAM1 = GGAGGTAA-ATTYCTTCTAAAGC). In rare cases of weak amplification, a punch of the PCR product was taken from the gel, suspended in 50–150 µL double distilled sterile water, melted at 70 C, and 1 µL of this dilution was reamplified using the thermocycling parameters above, except annealing temperature was increased to 50 C. PCR products were purified with GELase Agarose Gel-Digesting Preparation (Epicentre Technologies, Madison, Wisconsin) according to manufacturers instructions, and both strands were sequenced using a combination of these primers: LROR, LRFF1, LR3, LRAM1, LR3R, LR5, LR6, LR7 (LRFF1 = CTTTTCATCTT-TCGATCACTCTAC) (Vilgalys and Hester 1990, Rehner and Samuels 1995; www.biology.duke.edu/fungi/mycolab/primers.htm). Sequencing reactions were performed in a 10 µL total volume using dRhodamine Terminator or BigDye version II Terminator (ABI PRISM, Perkin-Elmer Biosystems, Foster City, California) fluorescent dyes following manufacturers protocols and precipitated with 10 µL of deionized sterile water, 2 µL of 3 M NaOAC and 50 µL of 95% EtOH. Sequences were generated on an Applied Bio-systems 377A or 3100 automated DNA sequencer. Each sequence fragment was subjected to a BLAST search to verify its identity. Sequences were assembled and aligned with Sequencher 4.1 (Gene Codes Corp., Ann Arbor, Michigan), optimized by eye and manually corrected when necessary.

Phylogenetic analyses. – Twelve ambiguously aligned regions were delimited, and characters in these regions along with two introns and portions of the 5' and 3' ends were excluded from all analyses. Equally weighted (MP1) and unequally weighted (MP2, MP3) maximum-parsimony analyses were performed using PAUP* 4.0b10 (Swofford 2002) as follows: Constant characters were excluded, unambiguously aligned characters were equally weighted and unordered, gaps were treated as missing, 1000 random-addition replicates were implemented with TBR branch-swapping, MUL-TREES option was in effect, and zero-length branches were collapsed. All remaining unambiguously aligned characters and changes between character states were equally weighted in the MP1 analyses. In the MP2 analyses, changes among transitions, transversions and gaps were subjected to a symmetric stepmatrix generated with the program STMatrix version 2.2 (François Lutzoni and Stefan Zoller, Depart ment of Biology, Duke University). The cost for changes among these character states was based on the negative natural logarithm of the percentages of reciprocal changes between any two character states. Five of the 12 ambiguous regions were recoded as five unequivocally coded characters using INAASE (Lutzoni et al 2000). These five characters along with the stepmatrix were used in the MP3 analyses. The remaining seven ambiguous regions were excluded from the MP3 analyses because their recoded characters contained more than 32 character states, which is not allowed in PAUP*. Branch support for all parsimony analyses was estimated by performing 1000 bootstrap replicates (Felsenstein 1985) with a heuristic search consisting of 10 random-addition replicates for each bootstrap replicate using the above settings, except the number of trees saved per random-addition replicate was limited to 100. Representatives from several additional families and orders were included to determine the phylogenetic position of polyphy-letic genera currently included in the Lasiosphaeriaceae. Four loculoascomycetes, Botryosphaeria ribis, Capnodium citri, Capronia mansonii and Mycopepon smithii, were used as outgroups.

The best-fit model of evolution was determined by MODELTEST 3.06 (Posada and Crandall 1998). MODELTEST 3.06 determines the best-fit model by comparing different nested models of DNA substitution in a hierarchical hypothesis-testing framework using the likelihood ratio test. Maximum-likelihood analyses were not performed due to the extreme computational efforts required by a dataset of this size. Bayesian analyses were performed with MrBayes 3.0b3 (Huelsenbeck and Ronquist 2001; http://morphbank.ebc.uu.se/mrbayes3). Constant characters were included, the above model of evolution was implemented and 10 000 000 generations were sampled every 1000th generation resulting in 10 000 trees. The first 2000 trees, which represented the burn-in phase of the analysis, were discarded, and the remaining 8000 trees were used for calculating posterior probabilities in the consensus tree. This analysis was repeated five times starting from different random trees to ensure trees from the same tree space were being sampled during each analysis.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 
Sequence alignment. – The final alignment included 85 taxa and 1337 bp after the introduction of gaps and is deposited in TreeBase (S947). The first 61 bp of the 5' end and the last 38 bp of the 3' end were excluded from all analyses due to missing data in most taxa. Twelve ambiguous regions representing 337 characters also were excluded from all analyses due to uncertainty in the alignment. Two taxa (Annulatascus triseptatus, Linocarpon appendiculatum) possessed single introns (67 bp and 65 bp, respectively), which subsequently were excluded from all analyses. These introns were identified as spliceosomal introns because they contained the highly conserved donor, branch and acceptor sites known to occur in ascomycete spliceosomal introns (Bhattacharya et al 2000). Of the remaining 769 characters, 449 were constant and 88 were parsimony uniformative for a total of 232 parsimony-informative characters in the MP1 and MP2 analyses. Five additional parsimony-informative characters derived from the unequivocally coded ambiguous regions also were included in the MP3 analyses.

Phylogenetic analyses. – The equally weighted MP1 analysis generated 110 equally most-parsimonious trees (MPT), all within a single island of trees. A strict consensus of the 110 MPT revealed only minor topological differences among genera in the Sordariales and Coniochaetales (data not shown). The inclusion of the stepmatrix in the MP2 analysis reduced the number of MPT to 12. A strict consensus of these trees revealed only minor topological differences within the Chaetomiaceae and the paraphyletic placement of Ascovaginospora with the Boliniales and Chaetosphaeriales (data not shown). The addition of the five recoded characters along with the stepmatrix in the MP3 analysis further reduced the number of MPT to three, which differed only in the paraphyly of Aporothielavia leptoderma, Chaetomium globosum and Chaetomium sp. One of the three MPT from this analysis is shown in FIG. 1. While minor differences in overall tree topology were found in the MPT among the three analyses, the topologies among and within the well-supported clades remained the same. As is typical in most large-scale phylogenetic analyses employing LSU sequence data, the backbone of the MPT received little bootstrap support. The effect of subjecting characters to a stepmatrix (MP2) and recovering the phylogenetic signal in five of the ambiguously aligned regions (MP3) was a reduction in the number of MPT along with a slight increase in bootstrap support for some clades.

View larger version (56K): [in this window] [in a new window]   FIG. 1. Phylogram of one of the three most-parsimonous trees generated from the MP3 analysis based on 1337 bp of the 5' end of the LSU nrDNA for 85 ascomycete sequences; length = 2629.36 steps, CI = 0.34, RI = 0.642, RC = 0.218. Bootstrap values 50% are shown above or below branches. Thickened branches indicate posterior probabilities 95%. Taxa placed in the Lasiosphaeriaceae sensu lato that occur outside the family are in bold. The 13 families recently placed in the Sordariales by at least one of the five classifications (TABLE I) are in shaded boxes or outlined.

Order Sordariales in a broad sense was found to be highly polyphyletic in that families segregated into 10 distinct monophyletic clades dispersed throughout several orders (FIG. 1). The circumscription of Sordariales is considerably narrowed for the clade containing genera within the families Chaetomiaceae, Lasiosphaeriaceae and Sordariaceae. Because the Chaetomiaceae and Lasiosphaeriaceae are paraphyletic, they here are recognized as the "chaetomiaceous complex" and the "lasiosphaeriaceous complex" respectively. Two new orders are proposed for the clades containing the Chaetosphaeriaceae and Coniochaetaceae respectively, while the order Coronophorales is accepted for the Nitschkiaceae and Boliniales for the Boliniaceae. Four families (Annulatascaceae, Cephalothecaceae, Catabotrydaceae, Helminthosphaeriaceae) are accepted or placed in the Sordariomycetidae inc. sed. The Batistiaceae is placed in the Euascomycetes inc. sed., and the Ceratostomataceae is accepted in the Hypocreomycetidae inc. sed.

The family Lasiosphaeriaceae in a broad sense was found to be highly polyphyletic in that genera segregated into numerous clades dispersed throughout several orders (FIG. 1). Eleven genera are retained in the redefined lasiosphaeriaceous complex with an additional 13 genera added based on morphological data, while 39 genera are excluded (TABLE II).

Taxonomic changes. – Chaetosphaeriales Huhndorf, A.N. Mill. & F.A. Fernández, ord. nov.

Ascomata superficialia vel semi-immersa, subglobosa, globosa vel obpyriformia, ostiolata; pagina glabra, aspera vel hirsuta; subiculum praesens vel absens. Paraphyses sparsae vel copiosae, simplices, septatae. Asci clavati vel cylindracei, unitunicati, cum vel sine annulo. Ascosporae ellipsoidae vel fusiformes, septatae, hyalinae vel fuscatae, laeves. Conidiogenesis phialidicae enteroblasticae.

Ascomata superficial or semi-immersed, subglobose, globose or obpyriform, ostiolate, glabrous, roughened or setose; subiculum present or absent; paraphyses sparse or abundant, simple, septate; asci clavate or cylindrical, unitunicate, with or without apical ring; ascospores ellipsoid to fusiform, septate, hyaline or pigmented, smooth; conidiogenesis enteroblastic phialidic.

Typus: Chaetosphaeria Tul. & C. Tul.

Coniochaetales Huhndorf, A.N. Mill. & F.A. Fernández, ord. nov.

Ascomata superficialia vel semi-immersa, subglobosa, globosa vel obpyriformia, ostiolata vel nonostiolata; pagina glabra vel hirsuta. Paraphyses filiformes, septatae. Asci subglobosi, clavati vel cylindracei, unitunicati. Ascosporae ellipsoidae, fusiformes vel discoidae, unicellulae, hyalinae vel pigmentiferi, laeves vel foveolatae, rimaris germinationibus. Conidiogenesis phialidicae, conidia ex phialidibus discretis vel cellulis intercalaribus, phialidicis genita.

Ascomata superficial or immersed, subglobose, globose or obpyriform, ostiolate or nonostiolate, glabrous or setose; paraphyses filiform, septate; asci subglobose, clavate or cylindrical, unitunicate; ascospores ellipsoid, ellipsoidfusoid or discoid, one-celled, hyaline or pigmented, smooth or finely pitted, with germ slit; conidiogenesis phialidic, conidia produced on discrete phialides or from phialidic intercalary cells.

Typus: Coniochaeta (Sacc.) Cooke

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 
Sordariales redefined. – Based on molecular data from partial LSU nrDNA genera within the families Chaetomiaceae, Lasiosphaeriaceae and Sordariaceae are retained within the redefined Sordariales. The Trip-terosporaceae is not distinguished from the Lasiosphaeriaceae at this time. The other families (sensu Eriksson et al 2001) find their placement elsewhere (see below and TABLE I).

There are several reasons for recognizing the Sordariales at this internode even though it currently lacks support. Although Bayesian inference is more likely than bootstrapping to provide support for short internodes (Alfaro et al 2002), neither method supported the Sordariales internode in these analyses. However, this internode is supported through significant Bayesian posterior probabilities in subsequent analyses using additional genes (ß-tubulin, RPB2) with similar taxon sampling (Miller unpubl data). The length of this internode is quite short, suggesting a rapid radiation of this group, and, under these circumstances, it is possible that even with additional data it might always lack bootstrap support (Berbee et al 2000). The order is also a coherent group with morphologies that vary along well-defined lines. Several putative synapomorphic morphological characters help to strengthen this ordinal clade, such as: (i) relatively large ascomata (approximately 400–600 µm diam), which are erumpent to superficial and never immersed; (ii) ascomal walls that are large-celled and membraneous or coriaceous but rarely carbonized; and (iii) ascospores that show variation on a distinctive developmental theme, often with appendages or sheaths. Taxa from three families are included in this overall clade, but, except for the Sordariaceae, lineages that fit the traditional families cannot be distinguished yet.

The Lasiosphaeriaceae as circumscribed by our molecular data is paraphyletic, and its members are described in more detail below. The Chaetomiaceae is recognized here as the "chaetomiaceous complex" and is represented by Chaetomium Kunze, Farrowia D. Hawksw., Thielavia Zopf, Aporothielavia Malloch & Cain and here including Zopfiella ebriosa. This group is also paraphyletic, nested within the clade that contains all sampled members of the Lasiosphaeriaceae. There is bootstrap support for the clade containing Z. ebriosa and Thielavia cephalothecoides (FIG. 1), and consequently Zopfiella is accepted in the chaetomiaceous complex. Chaetomium and other members of the Chaetomiaceae recently have been studied using 18S and 28S nrDNA sequences (Untereiner et al 2001, Lee and Hanlin 1999). In both cases the monophyly of the Sordariales was confirmed and with the limited taxon sampling of the Lasiosphaeriaceae that was included; the Chaetomiaceae formed a sister group to the other taxa, although without bootstrap support in the 28S data (Untereiner et al 2001). Our larger sampling of the Lasiosphaeriaceae suggests that relationships among taxa in the Chaetomiaceae and the Lasiosphaeriaceae are intricate and need further work.

The Sordariaceae is represented by Sordaria Ces. & De Not., Neurospora Shear & B.O. Dodge, Copromyces N. Lundq. and Gelasinospora Dowding. The clade containing these four genera is strongly supported but never resolves as a sister group of Lasiosphaeriaceae, as expected according to traditional morphological hypotheses. The Sordariaceae always is nested among the taxa traditionally found within the Lasiosphaeriaceae. Lundqvist (1972) provides an extensive review of the history of the Sordariaceae in its various forms (sensu lato and sensu stricto) and concludes that the family can be upheld for selected dark-spored genera that have similarities in perithecial, ascal and ascospore morphologies. He discusses the phylogenetic aspects within the group and notes the close relationship between Sordaria, Neurospora and Gelasinospora that has been shown repeatedly by various authors. He believed that the family was separate from taxa in the Lasiosphaeriaceae but that its origin (ancestor) probably was close to the hyaline-spored Lasiosphaeriaceae.

Our molecular data indicate that 11 genera belong in the paraphyletic Lasiosphaeriaceae, here recognized as the "lasiosphaeriaceous complex". These taxa share a similar pattern of ascospore morphology that extends to the Sordariales as a whole. Based on these similarities in morphology, 13 additional genera have putative relationships within the lasiosphaeriaceous complex (TABLE II) and more than 35 genera have relationships in the order. We believe that ascospore morphology is phylogenetically informative for determining whether a taxon belongs in the Sordariales. Lundqvist (1972) provides an excellent description of the ascospore morphologies that are found in his Sordariaceae sensu lato. If the chaetomiaceous complex is included with a few exclusions, the taxa in his group are essentially the same as our circumscription of the order. In the Sordariales, ascospore morphology exists as a continuum from a hyaline, cylindrical ascospore in Lasiosphaeria to a one-celled, brown, ellipsoid ascospore in Sordaria. Intermixed between these two extremes are many genera with two-celled ascospores that possess a brown, ellipsoid cell and different degrees of a hyaline (often basal) cell. There are numerous variations on this theme (see Lundqvist 1972 for excellent illustrations).

Our molecular analyses of the LSU also indicate that 17 genera that have been assigned to the Lasiosphaeriaceae sensu lato should be transferred to other families outside the Sordariales. By extrapolation, 22 additional genera with differing morphologies also should be removed from the order; some however have uncertain placement at this time (TABLE II). None of these genera possess ascospores on the same developmental theme as those included in the Sordariales.

The removal of the disparate elements at the family and genus level makes the Sordariales a more homogeneous assemblage. Sordariales sensu lato that includes also the families Boliniaceae, Helminthosphaeriacae, Chaetosphaeriaceae and possibly Catabotrydaceae theoretically could be recognized. However, circumscribing the order along these lines would let the group remain highly heterogeneous and permit further inclusions of taxa with dubious relationships.

Families removed from the Sordariales. – The position of the Boliniaceae historically has been uncertain. Members, such as Camarops with large conspicuous stromata, have been allied with the Xylariaceae (Munk 1953, Dennis 1960). Nannfeldt (1972) did not believe in this close relationship and accepted the family Boliniaceae for Camarops. In recent years the Boliniaceae has been placed in the Xylariales (Barr 1990) or been given incertae sedis status (Eriksson 1982). Andersson et al (1995) placed the Boliniaceae in the Sordariales based on 18s nrDNA sequence data. They suggested that the stromatic Catabotrydaceae, an already accepted member of the order, was closely related to the Boliniaceae and thus the order Boliniales was erected to include both families (Kirk et al 2001).

In our analyses, the Boliniaceae is represented by three species of Camarops that come together in a clade with strong bootstrap support. It comes as an unsupported sister group of the taxa representing the Chaetosphaeriaceae, Helminthosphaeriaceae and Catabotrydaceae, and all four families form an unsupported sister clade of the Sordariales. We accept the placement of the Boliniaceae in the Boliniales.

Catabotrydaceae.. The Catabotrydaceae was validated and placed in the Sordariales by Barr (1990). The family is monotypic based on a single species, Catabotrys deciduum, which forms large, dark stromata on dead tissues of large monocots. The species appears to be pantropical and rather common. Its relationships have not been understood clearly. In addition to finding a placement in the Sordariales, it has been placed in the Xylariales (Hyde et al 2000) and most recently in the Boliniales (Kirk et al 2001). While Catabotrys is resolved as an unsupported sister taxon to the Chaetosphaeriaceae in these analyses, it shows affinities to the Diaporthales and Hypocreales in additional analyses (data not shown). Because our analyses show its phylogenetic relationships as unresolved, we therefore place the Catabotrydaceae in the Sordariomycetes inc. sed.

Chaetosphaeriaceae and Helminthosphaeriaceae.. The Chaetosphaeriaceae recently was established for Chaetosphaeria and its relatives (Réblová et al 1999). The family was placed in the Sordariales due to similarities in ascomal wall and anamorph morphologies. Samuels et al (1997a) established the Helminthosphaeriaceae for Helminthosphaeria and, with reservations, placed the family in the "probably polyphyletic" Sordariales based on the anatomical characters of the ascomata and asci. Helminthosphaeria species generally have porate ascospores, a feature they share with some members of the Sordariaceae, Lasiosphaeriaceae sensu lato, Ceratostomataceae and Boliniaceae. Samuels et al (1997a) did not believe however that the Helminthosphaeriaceae had a close relationship with these families. Réblová (1999) placed the family in the Trichosphaeriales and included several additional genera.

The Helminthosphaeriaceae, which is represented by three species of Helminthosphaeria, is well supported by 100% bootstrap support and Bayesian posterior probabilities. Our data show the Chaetosphaeriaceae, Helminthosphaeriaceae and Catabotrydaceae clustering together as an unsupported sister group of the Boliniaceae. Four additional individual taxa formerly in the Lasiosphaeriaceae (Caudatispora biapiculata, Linocarpon appendiculatum, Lasiosphaeriella nitida and Ascovaginospora stellipala) also find their placement in this overall unsupported group. With limited taxon sampling from other Sordariomycetes orders, certain members of the Chaetosphaeriaceae resolved as a monophyletic group with a high level of bootstrap support separate from the taxa representing the Sordariales (Réblová and Winka 2000). Réblová and Winka (2001) suggested that the family was a likely candidate for removal to its own order from the broadly conceived Sordariales. Their data show the Chaetosphaeriaceae separated with high bootstrap support of 98% from other families in the order. Based on these data and our own analyses we believe that the Chaetosphaeriaceae represents a broad and morphologically diverse lineage distinct from the Sordariales that warrants separation at the ordinal level. We therefore choose to establish a new order Chaetosphaeriales for the Chaetosphaeriaceae.

Although it is evident that the Chaetosphaeriaceae and the Helminthosphaeriaceae are separate from the redefined Sordariales, the relationships between the two families are not clear. A few taxa formerly in the Lasiosphaeriaceae also come into an overall unsupported clade, and a number of additional taxa not included in these analyses also appear to have their placement in this group (data not shown). This group may represent a series of evolutionary events where there was a large, rapid radiation of taxa with numerous morphological characters that provide only cryptic information about relationships.

Réblová (1999) removed the Helminthosphaeriaceae from the Sordariales and placed it in the Trichosphaeriales based on ascomal wall morphology and conidiogenesis of associated anamorphs. The placement and relationships of the Trichosphaeriaceae within the Sordariomycetes were suggested by Réblová and Winka (2001). In their analysis of several new ascomycete taxa, they included Rhamphoria delicatula Niessl as a single taxon that they believed represented the Trichosphaeriaceae. Based on partial LSU nrDNA data, they found that R. delicatula was related to several taxa in the Annulatascaceae. They concluded that some characteristics of certain taxa in the Annulatascaceae match well with the circumscription of the Trichosphaeriaceae. However, without sequence data from the type species of Trichosphaeria Fuckel they could not confirm whether the clade containing taxa of both families is the true Trichosphaeriaceae. If the placement of the Trichosphaeriaceae based on the R. delicatula sequence data is accepted, then the Helminthosphaeriaceae would not be accepted in the Trichosphaeriales. In our analyses the Helminthosphaericeae does not appear closely related to the two species of Annulatascus that represent the Annulatascaceae. Therefore, Helminthosphaeriaceae currently is placed in the Sordariomycetes inc. sed.

Coniochaetaceae.. The Coniochaetaceae was established for two genera, Coniochaeta (Sacc.) Cooke and Coniochaetidium Malloch & Cain differing from Sordariaceae in ascospores having elongated germ slits. Coniochaeta has been allied variously with the Sordariaceae and the Xylariaceae, depending on the emphasis given to either the ascomal or ascospore morphologies. Most recent compilations have accepted the family in the Sordariales (TABLE I). Rogers (1994) considered it to be an isolated family whose relationship with other ascomycetes is obscure. Lee and Hanlin (1999) found the family to be strongly supported in the Sordariales with 18S nrDNA sequence data, where it came out as a well-supported sister group of the Chaetomiaceae/Sordariaceae clade. In analyses of 28S data with wider taxon sampling, Weber et al (2002) showed the strongly supported Coniochaetaceae separate from the Sordariales and clustering as a sister group to the Xylariales and Hypocreales. They declined to introduce a separate order for the family due to lack of bootstrap support.

Based on our data, we agree that the family does not belong in the Sordariales. The family represented by Coniochaeta, Coniochaetidium, Poroconiochaeta and herein Barrina (see below) has strong bootstrap and Bayesian support. Our data strongly support the Coniochaetaceae as being monophyletic and a new order, Coniochaetales, is established for the Coniochaetaceae.

Cephalothecaceae.. Suh and Blackwell (1999) discuss the history of the Cephalothecaceae and the relationships of cephalothecoid and other cleistothecial taxa. In their study, Cephalotheca Fuckel and two members of the Pseudoeurotiaceae formed a strongly supported clade that clustered as a sister group of the Sordariaceae/Lasiosphaeriaceae/Chaetomiaceae clade. This was supported by SSU data but not LSU data.

The Cephalothecaceae is represented in our analyses by Cephalotheca sulfurea and Cryptendoxyla hypophloia, along with two representatives of Albertiniella Kirschst. The family was well supported with 78% bootstrap support and 100% Bayesian posterior probability. The overall clade representing the Coniochaetaceae and Cephalothecaceae lacks bootstrap support but does show significant Bayesian support. We choose to place the family in the Sordariomycetidae inc. sed. until additional data are generated, while recognizing that it could be placed in the Coniochaetales.

Nitschkiaceae.. Over the years the Nitschkiaceae (= Coronophoraceae) has been variously accepted in the Coronophorales or placed in the Sordariales where it currently stands. Nannfeldt (1975) suggested the Nitschkiaceae be regarded as "a highly specialized satellite group derived from the Lasiosphaeriaceae" and be included in the Sordariales. The concept of the Lasiosphaeriaceae at that time contained taxa with characters that led to the misinterpretation of relationships between the two families. Taxa that herein are shown to be outside the Lasiosphaeriaceae were cited as having characteristics that united the Lasiosphaeriaceae with the Nitschkiaceae. Carroll and Munk (1964) suggested that the Coronophoraceae were nonostiolate relatives of the Lasiosphaeriaceae, closely related because of the putative intermediate nature of one species, Lasiosphaeria noonaedaniae Carroll & Munk. This is a species of Lasiosphaeriella, and the genus is represented here by L. nitida and shown to be outside the Lasiosphaeriaceae. Sivanesan (1978) suggested close affinities of the two groups when he described the genus Lasiobertia Sivan. (herein confirmed to be xylariaceous). Nannfeldt (1975) was quite correct in placing the Nitschkiaceae firmly in the ascohymenial and unitunicate pyrenomycetes but not correct in interpreting the closer affinities of the family.

Our data show Nitschkia grevillii and Chaetosphaerella phaeostroma (formerly a member of the Lasiosphaeriaceae) forming a strongly supported clade, which occurs as an unsupported sister group of the Hypocreales within the well-supported subclass Hypocreomycetidae. This clade is accepted as the order Coronophorales, which finds its placement in the Hypocreomycetidae.

Ceratostomataceae.. The Ceratostomataceae had its placement in the Sordariales until phylogenies based on nrDNA and RPB2 sequences placed Melanospora zamiae, the type, and several other species in the Hypocreales (Zhang and Blackwell 2002). In our analyses, Melanospora maintains this placement.

Annulatascaceae.. The Annulatascaceae originally was placed in the Sordariales (Wong et al 1998), but based on molecular data it was found to have affinities outside the order (Réblová and Winka 2001). Kirk et al (2001) place the Annulatascaceae in the Sordariomycetidae inc. sed. The family, which is represented here by two species of Annulatascus, A. velatispora, the type and A. triseptatus, is highly supported by bootstrap and Bayesian support. The clade that contains the Annulatascaceae also includes the Ophiostomatales and three taxa formerly in the Lasiosphaeriaceae, Pseudohalonectria, Plagiosphaera and Ceratosphaeria (see below). This overall clade occurs as an unsupported clade that is a sister group of the well-supported Diaporthales. We accept the placement of the Annulatascaceae in the Sordariomycetidae inc. sed. by Kirk et al (2001).

Batistiaceae.. The Batistiaceae was established for a single species, Batistia annulipes, and was placed in the Sordariales based on its phialidic conidiogenesis (Samuels and Rodrigues 1989). The genus had been linked at various times to other groups, for example, to the Xylariaceae based on its stipitate ascomata, carbonaceous texture and brown ascospores (Ciferri 1958) and to the Cephalothecaceae based on its cephalothecoid peridium (Samuels and Rodrigues 1989). Kirk et al (2001) removed the Batistiaceae from the Sordariales and placed it in the Sordariomycetidae inc. sed. Our data not only support its removal from the Sordariales but also its removal from the Sordariomycetidae. Its affinities outside the perithecial ascomycetes were not expected and, thus, were not tested with this dataset. Therefore the family is placed in the Pezizomycotina inc. sed. until further studies can be conducted.

Taxa included in the lasiosphaeriaceous complex. – The Lasiosphaeriaceae as circumscribed here is paraphyletic. Based on molecular data, 11 genera are retained in the redefined lasiosphaeriaceous complex: Apiosordaria, Bombardia, Bombardioidea, Cercophora, Jugulospora, Lasiosphaeria, Podospora, Schizothecium, Strattonia, Triangularia and Zygopleurage. In addition, based on morphological data, 13 genera also belong in the group: Anopodium, Apodospora, Arniella, Arnium, Camptosphaeria, Diffractella, Emblemospora, Eosphaeria, Fimetariella, Periamphispora, Pseudocercophora, Tripterosporella and Zygospermella (TABLE II). A number of taxa in synonymy with these genera also must be re-assessed.

As discussed previously, ascospore morphology in the Sordariales exists as a continuum from hyaline and cylindrical to brown and ellipsoid. Many genera in the lasiosphaeriaceous complex possess ascospores that represent the intermediate stage in this continuum by having a brown, ellipsoid cell and different degrees of a hyaline cell. The 11 sampled genera possess ascospores that represent this continuum. Lasiosphaeria ovina possesses hyaline, cylindrical ascospores, while Cercophora mirabilis and Bombardia bombarda ascospores are cylindrical with a brown, ellipsoid apical cell and a wide, elongate, hyaline basal cell. In Podospora fimiseda the basal, hyaline cell is long and narrow and has appendages with a complex microstructure. In Schizothecium curvisporum the basal cell is short and narrow and the brown cell is strongly curved. Zygopleurage zygospora has unmistakeable elongate ascospores with swollen brown cells at each end joined by a long, hyaline, collapsing intercalary cell. In Apiosordaria verruculosa, Jugulospora rotula and Strattonia carbonaria, the variously ornamented, brown cell is ellipsoid and the basal hyaline cell is small and conical. In Triangularia mangenotii, the brown cell is triangular, tapered at the top, wider at the bottom and the basal hyaline cell is short, semicircular and wide. Bombardioidea anartia has ascospores that are one-celled, brown and ellipsoid, similar to those of Sordaria. The other genera that are included in the Lasiosphaeriaceae for which no molecular data is available find their placement here because of ascospore morphology. All members have ascospores that vary along the brown cell/hyaline cell continuum.

Taxa excluded from the lasiosphaeriaceous complex. – All other genera included in the Lasiosphaeriaceae by Eriksson and Hawksworth (1993, 1998), Eriksson et al (2001) or Kirk et al (2001) have their placement elsewhere (TABLE II).

Taxa with relationships within the Sordariomycetidae.. Chaetosphaeria, Melanochaeta and Striatosphaeria have been transferred to the Chaetosphaeriaceae (Reblova et al 1999), and our molecular data concur with that finding. Four other taxa, Ascovaginospora, Lasiosphaeriella, Linocarpon and Caudatispora, find their placement in a larger unsupported clade containing the Helminthosphaeriaceae. Their relationships to each other and to the well-supported Chaetosphaeriaceae remain unresolved. Lasiosphaeriella was believed by Carroll and Munk (1964) to be intermediate between Coronophoraceae and Lasiosphaeriaceae because of its tuberculate peridium, radiating (diverging vertical) cells in the basal stroma, stalked ascus and allantoid ascospores. The germination of the ascospores by phialides suggested it belonged in the Lasiosphaeriaceae. This combination of characters made it difficult to place confidently it in any family, and it remained in the Lasiosphaeriaceae pending molecular information (Huhndorf and Fernández 1999). Linocarpon was placed in the Lasiosphaeriaceae by Hyde (1992a) because of ascus and ascospore morphology. It currently resides in the Hyponectriaceae (Eriksson et al 2001) with a question mark, and Kirk et al (2001) places it in the Xylariales. In our parsimony analyses, Lasiosphaeriella nitida occurs outside the Lasiosphaeriaceae on a branch with Linocarpon appendiculatum as a sister group to the Helminthosphaeriaceae. Caudatispora contains two species that are characterized by superficial clusters of ascomata and hyaline, ellipsoid ascospores with apiculate ends. Our data place C. biapiculata outside the Lasiosphaeriaceae on a lone branch as a sister taxon to the Helminthosphaeriaceae and Chaetosphaeriaceae.

Ascovaginospora was described for a single species from submerged Carex stems whose unicellular ascospores have a striking tetraradiate sheath (Fallah et al 1997). Using SSU nrDNA sequence data the species grouped with Podospora anserina and Fallah et al (1997) placed it in the Lasiosphaeriaceae. This placement was reiterated by Chen et al (1999). With our larger dataset, the genus is found to be outside the Lasiosphaeriaceae on a lone branch as a sister taxon of the Boliniales.

Barrina was not placed in the Lasiosphaeriaceae in the compiled lists of fungi (TABLE II) but based on teleomorph morphology it was thought to have a relationship with some members of the family, differing however from Lasiosphaeria and Cercophora (Ramaley 1997). The ascospores of Barrina are hyaline, ellipsoid, one-celled and lack a germ slit. They also are distinctive in that they readily germinate to form branched filaments that produce ascoconidia within the ascus. Ramaley (1997) suggested that the Phialophora-like anamorph most closely resembled anamorphs reported for some Coniochaeta species. The conidia may be formed on terminal or intercalary conidiogenous cells, which are reduced phialides that are not delimited from the subtending hyphal cell by a basal septum. Coniochaeta species with Lecythophora Nannf. anamorphs also form conidia on intercalary phialides (Gams 2000, Weber 2002). Our data strongly support the transfer of Barrina to the Coniochaetaceae.

Porosphaerella originally was described in the Trichosphaeriaceae (Müller and Samuels 1982) but later moved to the Chaetosphaeriaceae by Réblová et al (1999). In LSU analyses, Réblová and Winka (2000) found P. cordanophora to be outside the Chaetosphaeriaceae as a basal taxon in the Sordariales. It never was placed in the Lasiosphaeriaceae but was thought to have affinities to Chaetosphaeria and Porosphaerellopsis, putative members of the family. In our analyses it appears as an unsupported sister taxon of the Coniochaetaceae and might belong in the Coniochaetales sensu lato.

Annulatascus originally was placed in the Lasiosphaeriaceae and subsequently moved to its own family (Hyde 1992c, Wong et al 1998). The Annulatascaceae since has been shown to have affinities outside the Sordariales (Réblová and Winka 2001) with our data indicating possible affinities to the Ophiostomatales and the Diaporthales.

Pseudohalonectria and Ophioceras also had a place in the Lasiosphaeriaceae at one time but were found to have affinities to Gaeumannomyces (Chen et al 1999), and both were moved to the Magnaporthaceae (Shearer et al 1999). Our representatives grouped together, however, without support. We included a sequence of P. lignicola and a taxon we called O. tenuisporum because of its long, very thin ascospores (75–85 x 1.0–1.5 µm) and its tropical distribution. This collection differs from other collections of the species in having pale-colored ascomata occurring on a soft-textured petiole. Pale-colored ascomata are found mostly in Pseudohalonectria species, whereas most Ophioceras have darker ascomata. These characteristics also are reminiscent of the genus Plagiosphaera occurring on herbaceous stalks (Barr 1978). Most of the known species of Pseudohalonectria and Ophioceras occur on wood, and the relationships of those taxa with morphologically similar herbicolous ones have not been explored fully. Sequences for additional taxa would be helpful in understanding the delimitation of character states in these genera.

Our data suggest that Ceratosphaeria lampadophora also has affinities to this unsupported group. Ceratosphaeria lampadophora was shown to have a distinctive ascomal wall with a middle hyaline layer between two darker layers (Hyde et al 1997). Hyde et al (1999) recently described Pseudohalonectria eubenangeensis K.D. Hyde, Joanne E. Taylor & J. Fröhl. as a new species on palm with an ascomal wall structure very similar to C. lampadophora. Pseudohalonectria eubenangeensis spores are long, filiform and fasciculate, resembling some species of Ophioceras. Ceratosphaeria lampadophora spores are shorter, slightly curved at the ends and resemble some species of Pseudohalonectria. It also should be noted that the ascus apical rings in a number of these taxa are very similar (viz. refractive and thimble-shaped). However, ascal rings were found to be phylogenetically uniformative in the putative members of the Annulatascaceae (Ranghoo et al 1999, Réblová and Winka 2001).

Ordinal placement of these taxa is not clear with this dataset. They again come close to the Ophiostomatales clade, as was shown previously by Chen et al (1999). Overall the clade containing the Annulatascaceae, Ophiostomatales and Magnaporthaceae comes closest to the Diaporthales, although with no support. Relationships to the Diaporthales have been speculated for Ophioceras and Pseudohalonectria based on the morphological characteristics of long-beaked ascomata and deciduous asci (Shearer et al 1999).

Taxa with relationships within the Hypocreomycetidae.. Ceratocystis long since has been removed from the Lasiosphaeriaceae and is thought to have its relationships in the Microascales (Spatafora and Blackwell 1994, Eriksson et al 2001). Our analyses show C. virescens clustering with Plectosphaerella as a sister group to a clade containing Glomerella (placed in Phyllachorales and Glomerellaceae respectively by Kirk et al 2001; both Sordariomycetes inc. sed. fide Eriksson et al 2001). This concurs with 18S data (Zhang and Blackwell 2002) showing C. fimbriata and C. virescens grouping with Plectosphaerella and Glomerella. Rehner and Samuels (1995) likewise show similar relationships using 28S data. Our molecular data suggest that Porosphaerellopsis and Ascocodinaea have affinities to Glomerella and possibly Ceratocystis and Plectosphaerella but without support. This assemblage forms a sister group of a larger clade containing Hypocreales, Microascales and Coronophorales, which is highly supported with 86% bootstrap and 95% Bayesian support.

The affinities of Porosphaerellopsis have never been clear, with P. sporoschismophora having characteristics such as net-like paraphyses that are uncommon in the Sordariomycetes. The anamorph is suggestive of that of Melanochaeta in the Chaetosphaeriaceae, but it cannot be unequivocally referred to Sporoschisma Berk. & Broome because of the terminal pores in the conidia (Samuels and Müller 1978). The fusiform, septate ascospores also resemble those in Melanochaeta but again differing in having germ pores. Our data show that P. sporoschismophora is not related to M. hemipsila. The affinities of Ascocodinaea were thought to be clearer. Its small, dark, setose, pseudoparenchymatic ascomata and dematiaceous phialidic anamorphs placed it firmly in the Lasiosphaeriaceae (Samuels et al 1997b). Its simple teleomorph morphology is very similar to Chaetosphaeria and is distinguished by its stout tapering setae and versicolorous ascospores. Its Codinaea anamorph is similar to those found in a number of Chaetosphaeria species. Our data indicate that A. stereicola is not related to Chaetosphaeria or taxa in the Lasiosphaeriaceae. Both Porosphaerellopsis and Ascocodinaea have similar net-like paraphyses, but these are not seen in the other taxa in this clade. Additional molecular analyses with other taxa are necessary to refine the placement of P. sporoschismophora and Ascocodinaea, but at present they both are placed in the Hypocreomycetidae inc. sed.

Based on molecular data, Chaetosphaerella finds its placement in the Coronophorales. Chaetosphaerella phaeostroma clusters with Nitschkia grevillii with strong support. Chaetosphaerella shares the characteristics of munk pores in the ascomal wall cells, sterile ascomal base and inflated paraphyses with some members of the Nitschkiaceae. Based on morphology, two additional genera formerly in the Lasiosphaeriaceae also belong in the Coronophorales. Spinulosphaeria is very similar to Chaetosphaerella in having obpyriform to clavate ascomata with a stout sterile base, seated on a dense hyphal subiculum. These taxa also share other characteristics, such as broadly cellular, inflated paraphyses and asci without any distinct apical apparatus. Spinulosphaeria differs from Chaetosphaerella in its tooth-like spines on the ascomata, long-stipitate asci and ellipsoid, one-septate, brown ascospores. All attempts to sequence DNA from Spinulosphaeria failed. The morphology of Thaxteria has been used in the past to infer relationships to the Lasiosphaeriaceae or the Nitschkiaceae. Direct sequencing from PCR products of Thaxteria was problematic, so attempts were made to sequence from cloned copies. However, three different LSU copies were found from seven cloned sequences. Analyses including these copies placed Thaxteria in either the Lasiosphaeriaceae or the Nitschkiaceae (data not shown). Thaxteria has the Nitschkiaceae characteristics of munk pores in the ascomal wall cells, long stipitate asci and collapsing ascomata. Subramanian and Sekar (1990) accepted both Spinulosphaeria and Thaxteria in the Coronophorales, a disposition we also accept.

Taxa with relationships within the Xylariomycetidae.. Based on molecular data, two taxa, Apiospora and Lasiobertia, find their placement in the Xylariales. Apiospora was included in the Lasiosphaeriaceae in Eriksson and Hawksworth (1993) but since had been moved to its own family, Apiosporaceae (Hyde et al 1998). Lasiobertia was described by Sivanesan as differing from the Coronophorales and Lasiosphaeriaceae only in the presence of an amyloid ring. Hyde (1993) emphasized the amyloid nature of the ascus ring and the morphology of the ascospores and paraphyses, which he thought suggested a relationship to Oxydothis Penz. & Sacc. Wang and Hyde (1999) suggested the placement of Lasiobertia in the Clypeosphaeriaceae, while Kirk et al (2001) place the genus in the Apiosporaceae. Our representative is a new species that will be described shortly. Iodosphaeria was not sampled, but the presence of the amyloid ring suggests its affinities are also in the Xylariales and we accept its placement in the Amphisphaeriaceae (Eriksson et al 2001) although it recently was placed in its own family Iodosphaeriaceae (Hilber and Hilber 2002).

Taxa not sampled or with uncertain relationships.. For many genera currently included in the Lasiosphaeriaceae no molecular data are available. Characteristics of some of these genera suggest relationships to other families in the order or even to other orders, however their placement is uncertain. Taxa with long, narrow ascospores, such as Acrospermoides, Neolinocarpon, Palmicola, Plagiosphaera and Pulmosphaeria, may show affinities to Linocarpon, Ophioceras or the Diaporthales. Mycomedusiospora, with light-colored ascomata, may show a relationship with the Hypocreales. Myelosperma was placed in its own family but the affinities of it were not certain (Hyde and Wong 1999, Kang et al 1998). Merugia, with its filiform ascospores, has a Chalara-like anamorph associated with it, suggesting it might be allied with the Chaetosphaeriaceae. Phragmodiscus has fusiform multiseptate ascospores and an ascus apical ring that is suggestive of Ceratosphaeria. It differs in its ascomata immersed in a large, soft, spongy subiculum. With the Lasiosphaeriaceae redefined to include only those taxa with a characteristic morphology, these other taxa must be removed and for now are placed in the Sordariomycetidae inc. sed. (TABLE II).

	CONCLUSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 
In this work we have modified the circumscription of the Sordariales to better reflect phylogenetic relationships of the true members of the order. Dissimilar elements at the family and genus level have been transferred to groups that reflect their relationships. In doing so, we considered it necessary to make the nomenclatural changes for some taxa removed from the order. Two clades, Chaetosphaeriales and Coniochaetales, were natural candidates for elevation to ordinal rank. These clades received strong support, have distinctive teleomorph and anamorph morphologies and are considered places in the tree where large radiations of taxa occurred. The use of the family complex designation in place of the family name for the Lasiosphaeriaceae and Chaetomiaceae emphasizes the paraphyletic nature of these groups and thus serves to distinguish them from the monophyletic groups constituting families in the tree.

In this paper, traditional morphological-based hypotheses of relationships were tested using molecular data. In many cases these hypotheses conflicted with molecular-based hypotheses and new morphological-based hypotheses have been suggested that correspond to the molecular framework. The morphological data from numerous previous workers in various groups provided a good framework from which to test old hypotheses and form new ones. The use of a combination of molecular and morphological data herein have provided a more natural classification of the Sordariales and the lasiosphaeriaceous complex that leads to clear directions for additional work in these groups.

	ACKNOWLEDGMENTS

 
This work was supported in part by NSF PEET grants (Partnerships for Enhancing Expertise in Taxonomy, DEB-9521926 and DEB-0118695) to SMH, in part by a NSF BS&I grant (Biotic Surveys and Inventories, DEB-0072684) to SMH and FAF through The Field Museum of Natural History and in part by a NSF DDIG (Doctoral Dissertation Improvement Grant, DEB-0105077) to ANM through the University of Illinois at Chicago. ANM also was supported during this study by a Lester Armour Graduate Fellowship from The Field Museum of Natural History. Fieldwork for SMH was supported in part by the National Research Council Resident Research Associate Postdoctoral Program in cooperation with the U.S.D.A. Forest Service, Madison, Wisconsin. The authors are most grateful to Drs. Gerald Bills, William Buck, Hal Burdsall, Misericordia Calduch, Jacques Fournier, John Krug, Thomas Læssøe, Annette Ramaley, Amy Rossman, Gary Samuels and Alberto Stchigel for providing specimens or cultures. The New York Botanical Garden is thanked for letting us use the Batistia collections for DNA extraction. Sequences were generated in the Pritzker Laboratory for Molecular Systematics and Evolution at The Field Museum. This work represents a portion of a thesis in partial fulfillment of the requirements for the doctoral degree at the Graduate College of the University of Illinois at Chicago.

	FOOTNOTES

 
Accepted for publication September 13, 2003.

¹ Corresponding author. E-mail: shuhndorf{at}fieldmuseum.org

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION LITERATURE CITED

 
Alfaro ME, Zoller S, Lutzoni F. 2003. Bayes or bootstrap? A simulation study comparing the performance of Bayesian Markov chain Monte Carlo sampling and bootstraping in assessing phylogenetic confidence. Mol Biol Evol 20(2):255–266.[Abstract/Free Full Text]

Andersson K, Eriksson OE, Landvik S. 1995. Boliniaceae transferred to the Sordariales (Ascomycota). Systema Ascomycetum 14:1–16.

Barr ME. 1978. The Diaporthales in North America. Mycologia Memoir 7. J. Cramer, 232 p.

Barr ME. 1990. Prodromus to nonlichenized pyrenomycetous members of Class Hymenoascomycetes. Mycotaxon 39:43–184.

Berbee ML, Carmean DA, Winka K. 2000. Ribosomal DNA and resolution of branching order among the Ascomycota: how many nucleotides are enough? Molecular Phylogenetics and Evolution 17:337–344.[Medline]

Bhattacharya D, Lutzoni F, Reeb V, Simon D, Nason J, Fernandez F. 2000. Widespread occurrence of spliceosomal introns in the rDNA genes of ascomycetes. Mol Biol Evol 17(12):1971–1984.[Abstract/Free Full Text]

Cain RF. 1956. Studies of coprophilous Ascomycetes IV. Tripterospora, a new cleistocarpous genus in a new family. Can J Bot 34:699–710.

Carroll G, Munk A. 1964. Studies on lignicolous Sordariaceae. Mycologia 56:77–98.

Chen W, Shearer CA, Crane JL. 1999. Phylogeny of Ophioceras spp. based on morphological and molecular data. Mycologia 91:84–94.

Ciferri R. 1958. Batistia, a new genus of the Cephalothecaceae family. Atti Ist Bot Univ Pavia, Ser 5 , 15:160–170.

Dennis RWG. 1960. British Cup Fungi and their Allies. London: Ray Society. 585 p.

Dennis RWG. 1968. British Ascomycetes. Vaduz: J. Cramer. 333 p.

Eriksson OE. 1982. Outline of the ascomycetes—1982. Mycotaxon 15:203–248.

Eriksson OE, Hawksworth DL. 1993. Outline of the ascomycetes. Systema Ascomycetum 12:51–257.

Eriksson OE, Hawksworth DL. 1998. Outline of the ascomycetes—1998. Systema Ascomycetum 16:83–296.

Eriksson OE, Baral H-O, Currah RS, Hansen K, Kurtzman CP, Rambold G, Laessøe T. (Eds). 2001. Outline of Ascomycota—2001. Myconet 7:1–88.

Eriksson OE, Baral H-O, Currah RS, Hansen K, Kurtzman CP, Rambold G, Laessøe T. (Eds). 2003. Outline of Ascomycota—2003. Myconet 7:1–89.

Fallah PM, Shearer CA, Chen W. 1997. Ascovaginospora stellipala gen. et sp. nov. from sphagnum bogs. Mycologia 89:812–818.

Felsenstein, J. 1985. Confidence intervals on phylogenies: an approach using the bootstrap. Evolution 39:783–791.

Gams W. 2000. Phialophora and some similar morphologically little-differentiated anamorphs of divergent ascomycetes. Stud Mycol 45:187–199.

Hilber R, Hilber O. 2002. The genus Lasiosphaeria and allied taxa. Germany: Publ. By the author. 9 p., 7 pl.

Höhnel F. 1918. Mycologische Fragmente CCXVI. Über die gattung Bizzozeria Sacc. et Berl. Ann Mycol 16:76–77.

Huelsenbeck JP, Ronquist FR. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Biometrics 17:754–755.

Huhndorf SM, Fernández FA. 1998 Neotropical Ascomycetes 7. Caudatispora biapiculatis sp. nov. from Puerto Rico. Sydowia 50:200–204.

Huhndorf SM, Fernández FA. 1999. Neotropical Ascomycetes 8. New species of Lasiosphaeriella. Mycologia 91:544–552.

Hyde KD. 1992a. Fungi from palms I. The genus Linocarpon, a revision. Sydowia 44:32–54.

Hyde KD. 1992b. Fungi from decaying intertidal fronds of Nypa fruticans, including three new genera and four new species. Bot J Linnaean Soc 110:95–110.

Hyde KD. 1992c. Tropical Australian freshwater fungi. II. Annulatascus velatispora gen. et sp. nov., A. bipolaris sp. nov. and Nais aquatica sp. nov. (Ascomycetes). Australian Systematic Botany 5:117–124.

Hyde KD. 1993. Fungi from palms. IV. Reflections on Oxydothis and related genera. Sydowia 45:204–225.

Hyde KD. 1995. Tropical Australian freshwater fungi. VII. New genera and species of Ascomycetes. Nova Hedwigia 61:119–140.

Hyde KD. 1996. Tropical Australian freshwater fungi. X. Submersisphaeria aquatica gen. et sp. nov. Nova Hedwigia 62:171–175.

Hyde KD, Wong SW. 1999. Ultrastructural studies on the Myelospermaceae fam. nov., with a new species of Myelosperma. Mycol Res 103:347–352.

Hyde KD, Taylor JE, Fröhlich J. 1999. Two new species of Pseudohalonectria from palms. Mycologia 91:520–524.

Hyde KD, Taylor JE, Fröhlich J. 2000. Genera of Ascomycetes from Palms. Hong Kong: Fungal Diversity Press. 247 pp.

Hyde KD, Read SJ, Jones EBG, Moss ST. 1997. Tropical Australian freshwater fungi. XII. Rivulicola incrustata gen. et sp. nov. and notes on Ceratosphaeria lampadophora. Nova Hedwigia 64:185–196.

Hyde KD, Fröhlich J, Taylor JE. 1998. Fungi from palms. XXXVI. Reflections on unitunicate ascomycetes with apiospores. Sydowia 50:21–80.

Kang JC, Kong RYC, Hyde KD. 1998. Studies on the Amphisphaeriales 1. Amphisphaeriaceae (sensu stricto) and its phylogenetic relationships inferred from 5.8S rDNA and ITS2 sequences. Fungal Divers 1:147–157.

Kirk PM, Cannon PF, David JC, Stalpers JA. 2001. Ainsworth and Bisby’s Dictionary of the Fungi, Ninth Edition. UK: CABI International. 655 p.

Lee S, Hanlin RT. 1999. Phylogenetic relationships of Chaetomium and similar genera based on ribosomal DNA sequences. Mycologia 91:434–442.

Lundqvist N. 1972. Nordic Sordariaceae s. lat. Symbolae Botanicae Upsalienses 20:1–374.

Luttrell ES. 1951. Taxonomy of the pyrenomycetes. University of Missouri Studies . 24:1–120.

Lutzoni F, Wagner P, Reeb V, Zoller S. 2000. Integrating ambiguously aligned regions of DNA sequences in phylogenetic analyses without violating positional homology. Syst Biol 49:628–651.[Medline]

Müller E, Samuels GJ. 1982. Anamorphs of pyrenomycetous ascomycetes II. Porosphaerella gen. nov. and its Cordana anamorph. Sydowia 35:150–154.

Munk A. 1953. The system of the pyrenomycetes. Dansk Botanisk Arkiv 15:1–163.

Munk A. 1957. Danish pyrenomycetes. Dansk Botanisk Arkiv 17:1–491.

Nannfeldt JA. 1972. Camarops Karst. (Sphaeriales-Boliniaceae) with special regard to its European species. Svensk Botanisk Tidskrift 66:335–376.

Nannfeldt JA. 1975. Stray studies in the Coronophorales (Pyrenomycetes) 4–8. Svensk Botanisk Tidskrift 69:289–335.

Posada D, Crandall KA. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 49:817–818.

Ramaley AW. 1997. Barrina, a new genus with polysporous asci. Mycologia 89:962–966.

Ranghoo VM, Hyde KD, Liew ECY, Spatafora JW. 1999. Family placement of Ascotaiwania and Ascolacicola based on DNA sequences from the large subunit rRNA gene. Fungal Divers 2:159–168.

Réblová M. 1999. Studies in Chaetosphaeria sensu lato I. The genera Chaetosphaerella and Tengiomyces gen. nov. of the Helminthosphaeriaceae. Mycotaxon 70:387–420.

Réblová M, Winka K. 2000. Phylogeny of Chaetosphaeria and its anamorphs based on morphological and molecular data. Mycologia 92:939–954.

Réblová M, Winka K. 2001. Generic concepts and correlations in ascomycetes based on molecular and morphological data: Lecythothecium duriligni gen. et sp. nov. with a Sporodesmium anamorph, and Ascolacicola austriaca sp. nov. Mycologia 93:478–493.

Réblová M, Barr ME, Samuels GJ. 1999. Chaetosphaeriaceae, a new family for Chaetosphaeria and its relatives. Sydowia 51:49–70.

Rehner SA, Samuels GJ. 1995. Molecular systematics of the Hypocreales: a teleomorph gene phylogeny and the status of their anamorphs. Can J Bot 73 (Suppl 1) S816–S823.

Rogers JD. 1994. Problem genera and family interfaces in the Eupyrenomycetes. In: Hawksworth DL, ed. Ascomycete Systematics: Problems and Perspectives in the Nineties. New York: Plenum Press. p 321–331.

Samuels GJ, Müller E. 1978. Life History Studies of Brazilian Asco