Accurate identification of small-spored Alternaria spp. is challenging because of morphological plasticity under non-standard conditions and the common misapplication of the name A. alternata to a variety of morphologically distinct taxa. A system used by some authors of naming phytotoxigenic alternarias as pathotypes of A. alternata has further clouded the meaning and usefulness of this specific epithet. Two hundred and sixty isolates of small-spored Alternaria, primarily from fruit substrates, were segregated into morphological groups and then subjected to RAPD-PCR analysis using total genomic DNA and three different primers. When cultured under defined conditions, the pattern of sporulation observed at 50 × magnification was predictive of genetic relatedness as determined by cluster analysis of RAPD fragment patterns. In RAPD analyses, morphological groups or species were resolved as distinct branches of the dendrogram: Alternaria gaisen (= A. kikuchiana, A. alternata Japanese pear pathotype, group 2), A. longipes (= A. alternata tobacco pathotype, group 5), the ‘tenuissima’ group (group 5), the arborescens group (group 3) and the ‘infectoria’ group (group 6). Isolates in groups for which there were only a few representatives clustered as a branch. Analysis of RAPD fragment patterns confirmed that when Alternaria isolates are cultured and observed under defined conditions, their phenotypic plasticity is minimized and valid taxonomic separations can be made upon morphological characteristics. We conclude that A. alternata, A. gaisen, A. longipes and other small-spored saprotrophic or perthotrophic taxa are recognizable as morphologically distinct taxa. We suggest that the ‘pathotype’ system of naming small-spored Alternaria taxa confers no predictive value relative to observable morphological and genetic characters, and should be abandoned.