Introduction

Species of Eucalyptus are a primary source of fiber for the international paper and pulp industry (Turnbull 2000). The trees are propagated extensively as exotics in different parts of the world. The absence of pests and pathogens affecting trees in their non-native environments has, in part, contributed to the success of introduced Eucalyptus plantations (Burgess and Wingfield 2002b; Wingfield et al. 2008). However, the productivity of commercial Eucalyptus plantations is increasingly reduced worldwide by diseases caused by fungi and insects (Wingfield et al. 2008). Of these, foliar fungi have been implicated in reduced growth, reduced canopy leaf area, and delayed canopy closure in commercial plantations (Carnegie 2007b).

Teratosphaeria species are the most significant foliar pathogens of eucalypts (Carnegie et al. 1997; Park et al. 2000). They cause a complex of diseases, known as Teratosphaeria leaf disease (TLD), which has been reported to affect the eucalypt plantation industry in the subtropical and tropical areas of Australia (Andjic et al. 2010; Carnegie 2007a, b; Hunter et al. 2011). Species with Kirramyces anamorphs involved in this complex include T. suttonii, T. destructans, T. eucalypti, T. viscidus, and T. pseudoeucalypti. Previously placed in the genus Kirramyces and later in Readeriella, these species were transferred to the genus Teratosphaeria following the application of one genus name for anamorphs or teleomorphs in unambiguous monophyletic lineages in the families Mycosphaerellaceae and Teratosphaeraceae (Crous et al. 2009). The older name of the teleomorph state of the fungus T. suttonii took precedence over the epithet ‘epicoccoides.’ Although the origin of T. destructans (Wingfield et al. 1996), the most destructive of these pathogens, is unclear (Andjic et al. 2010; Burgess et al. 2007), T. suttonii, T. eucalypti, T. viscidus, and T. pseudoeucalypti are endemic to Australia (Hansford 1957). These species have morphologically similar spores, display similar symptoms on infected Eucalyptus leaves, and share common hosts. Unequivocal identification thus depends on DNA sequence comparisons (Andjic et al. 2007a, b; Burgess et al. 2006). Infection caused by these fungi results in the formation of irregular chlorotic or necrotic lesions on both surfaces of the leaves (Andjic et al. 2007a, b, 2010; Burgess et al. 2007). These lesions are usually associated with substomatal pycnidia that exude conidia. Spore masses appear as irregular brown-black clusters or smears, mostly on the lower leaf surfaces.

Among the species in this group, T. suttonii has a cosmopolitan distribution. It has been reported in native Eucalyptus forests and plantations in Australia (Carnegie 2007a, b; Walker et al. 1992) and other countries where these trees are grown as non-natives (Chipompha 1987; Chungu et al. 2010; Crous et al. 1988; Gardner and Hodges 1988; Padaganur and Hiremath 1973; Simpson et al. 2005) from a variety of Eucalyptus species (Crous et al. 1988; Sankaran et al. 1995) in the subgenera Corymbia, Monocalyptus, Symphyomyrtus, and Indiogenes. Carnegie et al. (1997) reported T. suttonii on three new hosts, namely, Eucalyptus argophloia, E. scias, and E. longirostrata.

Teratosphaeria suttonii infects seedlings, clonal hedges, and mature trees (Crous et al. 1988; Sankaran et al. 1995; Walker et al. 1992). It first manifests itself on aging leaves in the lower canopy or on leaves of stressed trees (Carnegie 2007a; Crous et al. 1988; Nichol et al. 1992). The disease can spread to younger leaves higher in the canopies of trees during continued epidemics (Park et al. 2000). It has led to extensive defoliation of plantations in Australia and Indonesia (Old et al. 2003) and caused the death of young plants in Malawi and South Africa (Chipompha 1987; Crous et al. 1989). It has recently caused substantial and comprehensive damage in young Eucalyptus plantations in New South Wales, Australia (Carnegie 2007b).

Leaf disease caused by T. suttonii is characterized by small, angular, irregular, purplish-red blotches on living leaves (Fig. 1b, c) (Andjic et al. 2007b). Individual lesions are delimited by veins and occur in large numbers on both sides of leaves (Hodges and Gardner 1984). Pycnidia exude spores in grey-brown to black cirri that form hair-like extrusions on the leaf surfaces or conidia can spread over the leaf surfaces, giving them a sooty appearance (Fig. 1c, d) (Walker et al. 1992). However, depending on the host, stage of development of infection, and climatic conditions, a variety of disease symptoms caused by infection with T. suttonii can be observed (Pegg et al. 2003; Walker 1962), thus leading to incorrect diagnosis of the pathogen.

Fig. 1
figure 1

Symptoms associated with infections by Teratosphaeria suttonii. a Trees severely defoliated. b Branch showing infections typically on the older leaves. c Sporulation of the fungus from stomata on the undersurface of a Eucalyptus leaf. d Infected leaf of a cutting during the rooting process where the infection by the fungus can reduce root strike

Full size image

A multiple genealogy study based on four gene regions grouped the isolates of T. suttonii from Australia, China, Indonesia, and South Africa into three well-supported subclades that could not be related to geographic origin (Andjic et al. 2007b). Several studies have demonstrated the significance of multilocus sequence typing (MLST) in resolving relationships among fungal strains, characterizing the genetic diversity, and identifying cryptic species (Taylor and Fisher 2003). However, lack of congruence among gene trees and species trees often present problems when interpreting species relationships. The genealogical sorting index (GSI) has been introduced to detect the progression from polyphyly to monophyly in diverging populations and approximate genealogical incongruence among taxa from which gene copies were sampled (Cummings et al. 2008). The GSI has recently been used to delineate a fungal species within the Neofusicoccum parvum–N. ribis species complex (Sakalidis et al. 2011). The level of monophyly of a group as represented by the GSI value can range from 0 to 1. GSI values at or close to 0 indicate the early stages of lineage sorting and thus the absence of exclusive ancestry. The GSI value of 1 marks the end of the divergence process. For a given tree topology, the statistical support for the observed monophyly is denoted by the corresponding P value. The overall GSI value for gene trees from multiple unlinked loci is provided by the GSIT, which indicates the degree of distinctiveness of the genome of a group, relative to that of another group (Cummings et al. 2008); thus, GSI is applicable for both single and multiple gene phylogenies.

The aims of this study were to use MLST to confirm the identity of the causal agent of TLD on eucalypts from various locations and hosts by comparing isolates collected from leaves of a variety of Eucalyptus species with known T. suttonii isolates and closely related species. A further aim was to establish the phylogenetic relationship of T. suttonii isolates from different parts of the world and, because differences have been observed in the morphology of conidia of this fungus, to consider whether collections do not encompass cryptic species.

Materials and methods

Origin and sampling of fungal isolates

A collection of isolates used in this study were obtained from diseased Eucalyptus leaves sampled from Australia, China, Indonesia, South Africa, United States of America, Uruguay, and Vietnam. Using a dissection microscope, conidia exuding from single pycnidia were collected at the tip of a sterile needle. Spores were placed on malt extract agar (MEA) (20 g l−1 Biolab malt extract, 15 g l−1 Biolab agar) and left to hydrate for 5 min. Under a dissecting microscope, a sterile needle was used to draw conidia across the agar surface, after which single spores were picked from the agar and transferred onto new MEA plates with 150 μg ml−1 streptomycin (Sigma-Aldrich). Cultures were grown at 25°C for 2 weeks, after which they were transferred to new MEA plates. Additional isolates from Australia, China, and Indonesia were obtained from the Murdoch University Culture Collection (MUCC), Perth, Western Australia and from the collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa. All cultures are maintained in the CMW collection. Leaf specimens were deposited in pretoria mycological (PREM, South African National Collection of Fungi).

Multilocus sequence typing

Total DNA was extracted from actively growing cultures by scraping mycelium from the surface of cultures into sterile 1.5-ml Eppendorf tubes and grinding freeze-dried mycelium to a fine powder. This step was followed by DNA extraction using the protocol of Cortinas et al. (2004), modified by adding Proteinase K (1 μg/μl) to the extraction buffer and 5 μl (1 mg/ml) RNAse to the final product (incubated overnight at room temperature) to digest RNA. The presence of DNA was confirmed by electrophoresis on 2% (w/v) agarose gels (Roche Diagnostics) stained with ethidium bromide and visualized under ultraviolet light. Subsequently, the concentrations of the extracted DNA were determined using a Nanodrop ND-1000 Spectrophotometer (Nanodrop Technologies, Rockland, DE, USA).

For the purpose of phylogenetic comparisons between isolates, polymerase chain reaction (PCR) was used for the partial amplification of three gene regions: these included the internal transcribed spacer regions ITS1, ITS2, and 5.8S of the rDNA operon using the primers ITS-1 and ITS-4 (White et al. 1990), part of the β-tubulin genes using primers BT2a and BT2b (Glass and Donaldson 1995), and the transcription elongation factor-1α (EF-1α) gene region using primers EF1-728F and EF1-986R (Carbone and Kohn 1999). All amplifications were carried out in 25-μl reaction mixtures, with cycling conditions as described previously (Cortinas et al. 2006) using the GeneAmp PCR system 2700 thermal cycler (Applied Biosystems). PCR products were purified and sequenced as described previously (Andjic et al. 2007b).

The resulting sequences of both strands were edited individually, and assembled into contigs using Vector NTI Advance 11. An initial identification of the isolates was done by performing a similarity search of the ITS rDNA sequences (standard nucleotide BLAST) against the GenBank database (http://www.ncbi.nlm.nih.gov/BLAST). Additional sequences of related species were retrieved from GenBank (Table 1). Sequences were aligned using Mafft version 6. Where required, adjustments were made manually by inserting gaps.

Table 1 Isolates of Teratosphaeria suttonii considered in the phylogenetic study
Full size table

Parsimony analysis was executed using (PAUP) version 4.0b10 (Swofford 2001) and its heuristic search component. The most parsimonious trees were obtained by employing heuristic searches with random stepwise addition in 100 replicates, with the tree bisection-reconnection branch swapping option (TBR) on and the steepest-descent option off. The analysis excluded all parsimony-uninformative and ambiguous characters, and gaps were considered as a fifth character. Approximate levels of homoplasy and phylogenetic signal, tree length (TL), retention index (RI), and the consistency index (CI) were determined. Trees were unlimited, tree branches of zero length were collapsed, and the multiple equally parsimonious trees generated were saved. The bootstrap proportions (Felsenstein 1985) were used to estimate confidence levels of the phylogenies. Related species, including T. destructans (CMW17919), T. eucalypti (CMW17917), T. cryptica (CMW3279), T. viscidus (MUCC452), T. zuluensis (CBS117262), T. nubilosa (CMW11560), and T. molleriana (CMW4940), were treated as outgroup taxa. All sequences generated in this study were deposited in GenBank (Table 1).

The aligned data sets used in the parsimony analysis were subjected to Bayesian analysis. The process involved determining the best nucleotide substitution model using MrModeltest version 25 (Nylander 2004) (Evolutionary Biology Centre, Uppsala University). Subsequently, the specifications of the general time-reversible (GTR) substitution model and a proportion of invariable site (I) were used to integrate variable rates across sites as part of performing phylogenetic analyses with MrBayes version 3.1 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). Beginning at random tree topology, the Markov chain Monte Carlo (MCMC) analysis of four chains went up 1,000,000 generations. Every 1,000 generations trees were saved, resulting in 1,001 saved trees. Burn-in was set at 50,000 generations (after the likelihood values converged to stationary); 950 trees were left, from which the consensus tree and posterior probabilities were calculated. The consensus tree was reconstructed in PAUP 4.0b10, and maximum posterior probabilities were allocated to branches following the construction of a 50% majority rule consensus tree from sampled trees. Sequence alignments have been deposited in TreeBASE, S11949 (http://www.treebase.org).

Bootstrap consensus trees and trees generated from Bayesian analysis from individual gene trees were uploaded into the online GSI program from http://www.geneologicalsortingindex.org. Each isolate was designated a corresponding ITS haplotype number previously obtained from a heuristic search of the ITS sequence data. Using 1,000 permutations, the GSI was calculated following the methods of Cummings et al. (2008).

Morphological characterization

For the purpose of comparing in vivo and in vitro spore characteristics (i.e., spores obtained from specimen and cultures, respectively), a sterile needle was used to obtain conidia from cultures or herbarium material. To induce sporulation, isolates were grown on 2% MEA and oatmeal agar (OMA) at 20°C in the dark for 30 days. Conidia from leaf material and squash mounts of fruiting structures were mounted in lactic acid and water (1:1 volume) and observed under a compound microscope. Characteristics determined for each isolate or specimen included conidial size, shape, number of septa, and pigmentation. Fifty measurements of spore length and width were recorded for each isolate or specimen at 1,000× magnification, using the Carl Zeiss microscope (Carl Zeiss, Mannheim, West Germany). The extremes of measurements were recorded in parentheses.

Results

Multilocus sequence typing

The three gene regions ITS, EF-1α, and β-tubulin were sequenced for a total of 100, 90, and 88 isolates, respectively, including the ex-type sequences obtained from GenBank (CMW5348) (Table 1). A few isolates to be sequenced were selected at random from each of the populations from each country sampled. Phylogenetic analysis of the generated DNA sequence data was initially performed on data for the three gene regions separately. For each gene region, isolates with identical sequence data were given the same haplotype number (Table 1). The final analysis performed on the three gene regions separately was performed on a reduced number of isolates, which included representatives of all polymorphisms observed (indicated by an asterisk in Table 1) and including the outgroups.

The aligned ITS sequence data consisted of 617 characters, 127 base pairs of which were from a large indel, in some isolates from Australia (47) and all isolates from South Africa, United States of America, Uruguay, and Vietnam. Two Australian isolates had a shorter indel of 77 base pairs. The indel was coded and treated as a single evolutionary event, thus leaving 106 parsimony informative characters that were used in the analysis. The data contained significant phylogenetic signal (P < 0.01; g1 = −1.01) to allow for meaningful analysis. Heuristic searches of unweighted characters in PAUP resulted in six most parsimonious trees of 167 steps (CI 0.82, RI 0.96), of similar topology, of which one is presented in Fig. 2a.

Fig. 2
figure 2

Phylogenetic relationships among Teratosphaeria suttonii isolates included in the study. The trees present a summary of the parsimony and Bayesian analyses of DNA sequence data obtained from the internal transcribed spacer (ITS) (a), elongation factor (EF)-1α (b), and β-tubulin (c) gene regions. Estimates of nodal support have been deduced from parsimony jackknifing (above nodes) and Bayesian inference analysis (below nodes). The different colors represent the 14 ITS haplotypes, showing how they have been reduced and intermixed in the EF and BT phylogenies

Full size image

There were 25 polymorphic loci for the T. suttonii isolates, resulting in 14 ITS haplotypes (Table 1; Fig. 2). These were distributed into two main groups, the first containing haplotypes 1–9 and the second containing haplotypes 10–14. Haplotype 4 comprised the highest number of isolates (30), followed by haplotypes 13 and 8. Haplotypes 1, 2, 3, 6, 8, 10, and 11 were found only in Australian isolates; haplotypes 1, 6, and 8 were restricted to Queensland isolates; and haplotypes 2, 3, 10, and 11 were found in isolates from different parts of Australia. Haplotypes 4, 5, 7, 13, and 14 included isolates from Australia and other countries. These isolates included the combinations Australia, South Africa, Uruguay, and United States (haplotype 4), South Africa, and Uruguay (haplotype 5), New South Wales and Vietnam (haplotype 7), New South Wales and Indonesia (haplotype 13), and Western Australia, Vietnam, and China (haplotype 14). Haplotype 12 occurred only in China isolates, and haplotype 9 was a single isolate from Queensland, Australia. ITS haplotypes 1, 8, and 12 resulted in equally high bootstrap values in parsimony and Bayesian analysis (Fig. 2a). ITS haplotypes 5, 6, and 11 had moderate parsimony and no posterior probability values, whereas ITS haplotypes 7, 10, and 14 had moderate parsimony and high posterior probability values. ITS haplotype 3 had no bootstrap support and a high posterior probability, and ITS haplotypes 2, 4, 9, and 13 were not supported. Isolates representing ITS haplotypes, marked with a superscript H in Table 1, were deposited with the mycological culture collection of the South African Plant Protection Research Institute.

The aligned EF-1α data set consisted of 290 characters, of which 112 were parsimony informative and included in the analysis. The data contained significant phylogenetic signal (P < 0.01; g1 = −2.97). Heuristic searches of unweighted characters in PAUP resulted in a single most parsimonious tree of 261 steps (CI 0.801, RI 0.881), as represented in Fig. 2b. Among isolates of T. suttonii there were three polymorphic loci, which resulted in four EF haplotypes (Table 1; Fig. 2b). The smallest supported group of isolates in the EF tree consisted of three isolates from China, corresponding to isolates from ITS haplotypes 12 and 14. The other three EF haplotypes were made up of ITS haplotypes 1, 2, 3, 4, 5, 6, 7, 8, 9, and 14, including isolates from Australia, South Africa, Uruguay, United States, and Vietnam; ITS haplotypes 4, 5, 7, 10, and 11, representing isolates from Australia and Uruguay; and ITS haplotypes 12, 13, and 14, representing isolates from Australia, China, and Indonesia.

The β-tubulin dataset consisted of 381 characters, of which 2 were parsimony informative and were used in the analysis. The data contained significant phylogenetic signal (P < 0.01; g1 = −2.25). Heuristic searches of unweighted characters in PAUP resulted in two trees of 140 steps (CI 0.79, RI 0.91). One of the trees was saved for representation (Fig. 2c). Among isolates of T. suttonii, there were two polymorphic loci, which resulted in 2 BT haplotypes (Table 1; Fig. 2c), thus reducing the 14 ITS haplotypes to 2 haplotypes. The smaller BT subclade had equally high bootstrap values in parsimony and Bayesian analyses; the other subclade was not supported by posterior probability values. The smaller BT subclade consisted of the ITS haplotypes 2, 4, 5, 9, and 11. The group included isolates from Australia, South Africa, Uruguay, and United States. The larger subclade included isolates from Australia, China, Indonesia, South Africa, and Vietnam, corresponding to ITS haplotypes 1, 2, 3, 4, 6, 7, 8, 10, 12, 13, and 14.

For the GSI analysis (Table 2), 31 isolates in 9 of the 14 ITS haplotypes, namely haplotypes 1, 5, 6, 7, 8, 10, 11, 12, and 14, had a genealogical divergence of 1,000. Haplotypes 2, 3, and 4 had low genealogical divergence values of 0.118, 0.087, and 0.369, respectively, whereas haplotype 13 scored a moderate GSI value of 0.688. Taxa in all haplotypes except 2 and 3 showed significant estimates of measures of exclusive ancestry. Statistics could not be produced for the ITS haplotype 9 because it consists of a single haplotype. Significant measures of exclusive ancestry were estimated for taxa of the ITS haplotypes 1, 11, and 13 in BT, but overall low to moderate genealogical divergence was found for all isolates at the BT and EF loci, scoring GSI values ranging from 0.002 to 0.824 and from 0.004 to 0.075, respectively. Failure to separate isolates on the BT and EF trees (Fig. 2b, c; Table 2) supported the low levels of genealogical divergence at these loci.

Table 2 Genealogical sorting index (GSI) and probability values for the consensus bootstrap trees from parsimony analysis of sequence data from the three gene regions
Full size table

Morphological characterization

Variable conidial shapes were observed among the isolates examined (Figs. 3, 4). Spores were straight, slightly curved, and curved. Spores from all isolates and herbarium material were generally brown, with a few that were faint brown at the narrow ends. A collection of spore measurements was obtained from representative isolates and herbarium material (Table 3). The lengths and widths of spores in vivo ranged from 28.49 to 77.49 μm and from 2.06 to 7.62 μm, respectively. In vitro spore length and width ranged from 16.07 to 66.57 μm and from 2.30 to 6.98 μm, respectively. The shortest spores were recorded in vitro from isolate CMW31916; the longest spores were recorded in vivo from specimen PREM60495. The narrowest and widest spores were recorded in vivo from leaf material (PREM60540 and PREM60536). Where both in vivo and in vitro measurements could be recorded, length of conidia from herbarium material was generally greater than for the in vitro measurements. Conidial widths did not differ significantly for material from culture or from leaf tissue. The average length of spores for isolates CMW28689, CMW29223, and CMW32939 and from corresponding leaf material was similar. The number of septa recorded in the conidia ranged from 1 to 10. Where spores were observed both in vitro and in vivo, often fewer septa were recorded for those produced in vitro. In a few instances, the number of septa was the same both in vivo and in vitro collections. Single septate spores were obtained only in vitro.

Fig. 3
figure 3

Average lengths and widths of spores obtained from cultures and leaves. Error bars represent the standard deviation

Full size image
Fig. 4
figure 4

Conidia obtained from isolates of Teratosphaeria suttonii. a CMW35813, b CMW35768, c CMW35791, d CMW29223, e CMW18629, f CMW31916, and leaf specimen g PREM60538, h PREM60488, i DC10, j PREM60484, k PREM60491, l PREM60495. Where two spore images are given, as for b, c, f, j, k, and l, the image on the right represents spores from culture and the one on the left from an infected leaf. The numbers in brackets represent those for ITS haplotypes

Full size image
Table 3 Characteristic features of conidia obtained from leaves and cultures
Full size table

Discussion

A multigene phylogeny was constructed for a large collection of T. suttonii isolates, including those from across the natural range of this species, as well as from many countries where it has been introduced. Initial examination of ITS data alone revealed several distinct and strongly supported terminal clades, suggesting that T. suttonii represents a species complex. This finding was consistent with early observations of considerabe variability in the morphology of the conidia. However, much lower variability was observed in BT and EF sequence data. Additionally, the trees obtained for the different gene regions were not congruent and, apart from the ITS region, there was no GSI support for monophyly for the other gene regions. The lack of clarity in the molecular support was also confounded by the wide range of conidial morphology in the fungus, with no relationships emerging consistent with the morphological and the molecular divisions. We therefore conclude that T. suttonii represents a single, highly variable species. Such considerable genetic instability has also been observed in Fusarium avenaceum and Fusarium heterosporum isolates from the same host and geographic origin (Benyon et al. 2000). The high level of genetic variability was considered to be a factor of genetic instability. Similar to T. suttonii in this study, Catenulostroma wingfieldii has been reported to display variable spore characteristics (Crous et al. 2008).

Isolates from Australia were scattered throughout the phylogenetic trees and included 12 of the 14 ITS haplotypes identified for T. suttonii in this study. This result was not surprising, given the broad natural distribution of the fungus in eastern Australia (Park et al. 2000; Walker et al. 1992). The two haplotypes not represented in Australia either represent unsampled haplotypes from Australia or they might be explained by a postintroduction mutation, as Perez (2010) has shown for the related T. nubilosa. Isolates obtained from other countries where T. suttonii is believed to be an introduced pathogen were considerably less variable, represented by only one or two ITS haplotypes. According to the results, the haplotypes introduced into Asia are different from those introduced into South Africa, United States, and Uruguay, and they thus represent independent introductions, probably directly from Australia. It was, however, not possible to determine whether isolates found in Florida and Uruguay originated from Australia or South Africa because both the South Africa and Australia isolates share similar haplotypes with isolates from the former two countries. Eucalyptus planting stock, particularly in the form of seed, has been widely traded around the world, and it is believed that this has been one of the important sources of pathogens moving into new areas (Andjic et al. 2011; Wingfield et al. 2008). For example, T. nubilosa, an important Eucalyptus pathogen and relative of T. suttonii, moved from Australia into South Africa, and the latter country provided a beachhead for movement of the pathogen into Africa and Europe (Hunter et al. 2008). Thus, the origin of populations of forest pathogens such as T. suttonii in countries can often be linked to global trade in forest products and not necessarily to introductions from countries where these pathogens are native, as has, for example, been established for T. nubilosa and D. pinea (Burgess and Wingfield 2002a; Hunter et al. 2008).

Interestingly, T. suttonii is not endemic to Western Australia (Jackson et al. 2008). In contrast, it was probably introduced in the early 2000s when the rapidly expanding eucalypt plantation industry was sourcing seedlings from nurseries all down the eastern seaboard of Australia. This idea was also reflected in the haplotypes of the pathogen observed in Western Australia in the current study, where they were representative of isolates commonly found over a wide geographic range in Australia.

A high level of variability in conidial size, number of septa, and shape was observed between isolates irrespective of their molecular aggregation. Conidial characteristics could not be related to origin, host, or whether they were produced in vivo or in vitro under controlled conditions. Variation in spore morphology has previously been reported in T. suttonii (Knipscheer et al. 1990), although these authors did not mention all the different forms of variation observed in the current study. Based on the species description, T. suttonii is characterized by straight to slightly curved spores that are brown, rough walled, with usually 3–5 septa, sinuate, narrowing to the rounded, slightly paler apex, sometimes slightly constricted at septa, and measuring (21–)36–57(–61) × 3–5.5(–7) μm (Crous et al. 1989; Hansford 1957; Padaganur and Hiremath 1973; Walker 1962). In general, conidia from different isolates observed in this study fit the original and subsequent descriptions of T. suttonii. This study has used nucleotide sequences for three genes and phenotypic characteristics to identify fungal isolates from leaves with typical TLD symptoms thought to be associated with T. suttonii and to infer their evolutionary history. The morphological and molecular data at hand, the indistinguishable symptoms of infection, and similar biological niches provide convincing evidence for a single but diverse species rather than a species complex. Repeated recombination, interrupted deviation over time, and incomplete geographic containment may have resulted in introgression between previously defined species (Taylor et al. 1999). Further studies using deep sequencing for the identification of single nucleotide polymorphisms (SNPs) (Perez 2010) may still elucidate species boundaries within this complex.