Genomic DNA was isolated from fungal mycelium grown on the agar plates following the protocol of Lee & Taylor (1990 ) or the UltraClean™ Microbial DNA Isolation Kit (Mo Bio Laboratories, Inc., Solana Beach, CA, USA). All isolates were sequenced with five genomic loci. The primers ITS5 or ITS1 and ITS4 (White et al. 1990 ) were used to amplify the internal transcribed spacers areas as well as the 5.8S rRNA gene (ITS) of the nrDNA operon. Part of the actin gene (ACT) was amplified using the primer set ACT-512F and ACT-783R (Carbone & Kohn 1999 ) and part of the translation elongation factor 1-a gene (EF) using the primer set EF1-728F and EF1-986R (Carbone & Kohn 1999 ). The primer set CAL-228F and CAL-737R (Carbone & Kohn 1999 ) was used to amplify part of the calmodulin gene (CAL) whereas the primer set CylH3F and CylH3R (Crous et al. 2004c ) was used to amplify part of the histone H3 gene (HIS). Additional degenerate primers were developed from sequences obtained from GenBank as alternative forward and reverse primers for some of the loci during the course of the study (Table 2 ); however, these were rarely used but based on their degenerate design could be of use to the broader scientific community. The protocols and conditions outlined by Groenewald et al. (2005 (link)) were followed for standard amplification and subsequent sequencing of the loci.
Sequences of Septoria provencialis (isolate CPC 12226) were used as outgroup based on availability and phylogenetic relationship with Cercospora (Crous et al. 2004b , 2006b (link)). The Cercospora sequences were assembled and added to the outgroup sequences using Sequence Alignment Editor v. 2.0a11 (Rambaut 2002 ), and manual adjustments for improvement were made by eye where necessary. Gaps present in the ingroup taxa and longer than 10 characters were coded as a single event for all analyses (see TreeBASE).
Neighbour-joining analyses using the HKY85 substitution model were applied to each data partition individually to check the stability and robustness of each species clade under each data set using PAUP v. 4.0b10 (Swofford 2003 ) (data not shown, discussed under the species notes where applicable). Alignment gaps were treated as missing data and all characters were unordered and of equal weight. Any ties were broken randomly when encountered. The robustness of the trees obtained was evaluated by 1 000 bootstrap replications (Hillis & Bull 1993 ).
MrModeltest v. 2.2 (Nylander 2004 ) was used to determine the best nucleotide substitution model settings for each data partition. Based on the results of the MrModeltest, a model-optimised phylogenetic re-construction was performed for the aligned combined data set to determine species relationships using MrBayes v. 3.2.0 (Ronquist & Huelsenbeck 2003 (link)). The heating parameter was set at 0.3 and the Markov Chain Monte Carlo (MCMC) analysis of four chains was started in parallel from a random tree topology and lasted until the average standard deviation of split frequencies came below 0.05. Trees were saved each 1 000 generations and the resulting phylogenetic tree was printed with Geneious v. 5.5.4 (Drummond et al. 2011 ). New sequences generated in this study were deposited in NCBI’s GenBank nucleotide database (www.ncbi.nlm.nih.gov ; Table 1 ) and the alignment and phylogenetic tree in TreeBASE (www.treebase.org ).
Isolates of Cercospora sp. Q were screened with five more loci to test whether additional loci could distinguish cryptic taxa within this species. This species was selected based on the intraspecific variation present inFig. 2 (part 5) and also the range of host species and countries represented. The primer set GDF1 and GDR1 (Guerber et al. 2003 (link)) was used to amplify part of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene, primer set NMS1 and NMS2 (Li et al. 1994 (link)) for part of the mitochondrial small subunit rRNA gene and part of the chitin synthase (CHS) gene was amplified using the primers CHS-79F and CHS-354R (Carbone & Kohn 1999 ). Part of the gene encoding for a mini-chromosome maintenance protein (MCM7) was amplified using primers Mcm7-709for, Mcm7-1348rev, Mcm7-1447rev (Schmitt et al. 2009 (link)) and part of the beta-tubulin gene using mainly the primers T1, Bt2b and TUB3Rd (see Table 2 for references).
Sequences of Septoria provencialis (isolate CPC 12226) were used as outgroup based on availability and phylogenetic relationship with Cercospora (Crous et al. 2004b , 2006b (link)). The Cercospora sequences were assembled and added to the outgroup sequences using Sequence Alignment Editor v. 2.0a11 (Rambaut 2002 ), and manual adjustments for improvement were made by eye where necessary. Gaps present in the ingroup taxa and longer than 10 characters were coded as a single event for all analyses (see TreeBASE).
Neighbour-joining analyses using the HKY85 substitution model were applied to each data partition individually to check the stability and robustness of each species clade under each data set using PAUP v. 4.0b10 (Swofford 2003 ) (data not shown, discussed under the species notes where applicable). Alignment gaps were treated as missing data and all characters were unordered and of equal weight. Any ties were broken randomly when encountered. The robustness of the trees obtained was evaluated by 1 000 bootstrap replications (Hillis & Bull 1993 ).
MrModeltest v. 2.2 (Nylander 2004 ) was used to determine the best nucleotide substitution model settings for each data partition. Based on the results of the MrModeltest, a model-optimised phylogenetic re-construction was performed for the aligned combined data set to determine species relationships using MrBayes v. 3.2.0 (Ronquist & Huelsenbeck 2003 (link)). The heating parameter was set at 0.3 and the Markov Chain Monte Carlo (MCMC) analysis of four chains was started in parallel from a random tree topology and lasted until the average standard deviation of split frequencies came below 0.05. Trees were saved each 1 000 generations and the resulting phylogenetic tree was printed with Geneious v. 5.5.4 (Drummond et al. 2011 ). New sequences generated in this study were deposited in NCBI’s GenBank nucleotide database (
Isolates of Cercospora sp. Q were screened with five more loci to test whether additional loci could distinguish cryptic taxa within this species. This species was selected based on the intraspecific variation present in
