A study was conducted to assess phenotypic, biochemical, and molecular diversity in coriander accessions from the North Central Regional Plant Introduction Station, in Ames, IA; initially, 139 accessions were characterized for phenological and morphological traits and for fatty-acid composition. A second year of data was collected on those traits from 60 accessions. Fruit essential-oil content and composition were determined, along with a headspace analysis from leaves and an analysis of amplified fragment length polymorphisms (AFLPs). Basic statistics and Pearson correlations were estimated. Matrices of geographical distances (GD), modified Rogers' (MD) distances for molecular data, and Euclidean distances for phenotypic (PD), essential-oil (EO) and fatty-acid (FA) profiles, and for the combination of the last two data sets (EOFA) were calculated. Analyses of variance (ANOVA), Mantel tests for distance matrices, cluster analyses (CA), principal components analyses (PCA), and analyses of molecular variance (AMOVA) were applied. All the characteristics, except for leaf volatiles, showed differences and wide variation among populations at the phenotypic, biochemical, and molecular levels. Significant correlations were identified for GD with PD and FA, for PD with FA and EOFA, and for MD with FA, with weaker correlations between MD and both EO and SOFA. CA yielded different dendrograms from PD, SOFA, and MD. PCA generally supported grouping patterns from CA. AMOVA supported the three grouping patterns established from PD, SOFA, and MD, but percentages of variation among groups, among populations and within populations were about 5, 25, and 70%, respectively. Fixation indices did not support clear differentiation among groups formed on the basis of phenotypic, biochemical, and molecular data. A shared genetic base and a recent human selection process affecting relatively few genes are discussed as possible explanations for the observed weak genetic differentiation among the studied coriander populations. However, it should be possible to use the concept of "Group" outlined by the International Code of Nomenclature for Cultivated Plants and combine the results from phenotypic, molecular, and biochemical characterizations to develop a utilitarian classification for coriander that could lead to a better understanding of the gene pool, and it should be helpful for researchers, growers and breeders.