A new approach for identifying starch granules of wild food plants from arid western North America

Taxonomic identification of starch granules is critical to understanding dietary patterns from archaeological contexts, but achieving high levels of confidence around such identifications has been problematic. This study highlights a statistical approach to morphometric and morphological characteristics that reduces uncertainties when making comparisons to reference material. Emphasizing starch granules with eccentric hila (those produced by underground storage organs), four taxa with ethnobotanical significance were examined from arid western North America. It was determined that the frequency distributions of granule size were bimodal or otherwise non‐normal, thus, requiring log‐transformation prior to making comparisons of means and measures of variance. Granule sizes in the upper 10 and 20% provided the most confident differentiation of starch granules among taxa. Significant and positive correlations of starch granule size to morphological characteristics (lamellae, aggregation, longitudinal fissure, shape) suggested that small granules are less likely to exhibit diagnostic features. Using this approach, the Four Corners potato ( Solanum jamesii ) could be readily distinguished from the three liliaceous taxa ( Erythronium grandiflorum , Calochortus nuttallii , and Fritillaria pudica ) by its large granule size and maximum fissure width. Identification of starch granules from archaeological contexts, therefore, requires improved recovery methods that increase yields (sample sizes) from artifacts and the probability of including larger granules that display diagnostic characteristics.