A new method for creating a visual plant identification key (Dellinger-Johnston)

Abstract: Taxonomic keys are essential tools for species identification, used by students and professional biologists. In recent years, advancements in photography have allowed these keys to host high-quality photographs for aid in identification. However, most modern keys still rely heavily on text rather than images. Using text alone limits the user to a discrete number of characters, often described in esoteric terms. In order to create more effective keys, we developed a new method for constructing image-based taxonomic keys. These keys replace written characters with images - allowing the user to identify species using visual pattern recognition, rather than interpreting written text. In addition, we constructed our visual key using data on how different users assess the visual similarities between plant species. To ensure the strength of this methodology, our key focuses on the morphologically diverse genus, Quercus. A set of standardized photographs was taken of forty-three species of oak native or naturalized in the Southeast. These photographs were used to create a survey on how botanical experts and botanical novices rate the pair-wise similarity of different oak leaves. The mean of each rating was summarized into a distance matrix, which was then converted into a dendrogram. From the resulting dendrogram, a visual key was constructed using the standardized photographs of oak leaves. The key was then tested on against an existing dichotomous key using botanical novices and botanical experts. The resulting two-sample t-tests between the two identification keys demonstrated that users with our visual key produced between 22-30% more correct answers than users with the traditional key. Using this method of key creation, innovative keys could be constructed for other fields of biology.