Genetic variation occurs at all levels across the genus Fusarium. In some cases such variation has been used to define species, and in others to describe populations or lineages. When amplified fragment length polymorphisms (AFLPs) are evaluated, strains in different species usually share at least 60% of the fragments and those in different species 40% of the fragments, or less, with isolates sharing between 40 and 60% of the fragments in an indeterminant situation. This gray area also is reflected in morphological characters, usually indistinguishable, and cross-fertility, usually some cross-fertility but often not as fertile as are strains that are more closely related. In terms of DNA sequence, the genes used for species diagnostics often have not been tested on large numbers of strains. For example, the TRI101 gene of F. graminearum contains at least 25 single nucleotide polymorphisms (SNPs) from 36 strains and yielded 17 alleles that have been proposed as a means to subdivide this species into at least nine. However these subdivisions fare poorly as more strains are analyzed, with the number of alleles increasing to >40 when approximately 500 strains from Korea and South America are sequenced. Some of the newly identified alleles cannot be correctly assigned to one of the nine subdivisions based on the proposed diagnostic SNPs. Before SNPs are proposed as characters to define species, it is important to verify their specificity based on a sufficiently large sample and to evaluate the genetic variation present in terms of an independent measure of genetic relationships. Only in such a manner can names that are meaningful in the context of trade and quarantine regulations be developed.