Phylogenetic trees, diagrams depicting hypothesized evolutionary relationships between taxa, are used in a widely across multiple biology subdisciplines. Being able to read these diagrams is therefore a crucial step in achieving expertise in biology. Tree-thinking, however, has been found to be quite difficult, especially given the abstract nature of the diagram and its reference to directly unobservable macroscopic biological phenomena. The biology education literature, therefore, has dedicated some effort into investigating the ways in which tree-thinking might be improved in undergraduates. While these investigations have been fruitful, they have not yet included a systematic examination of expertise in comparison to novices in order to fully characterize the skills and conceptions required for tree- thinking. They also have not yet examined the concept at the root of tree-thinking: relatedness understanding. To remedy these issues, I conducted a study employing expert and novice framing of relatedness understanding. I also draw upon cognitive metaphor and Prototype Theory to inform this study which has not yet been employed by the tree-thinking community. I attempted to characterize the everyday ways individuals categorize and identify organisms and how they might differ by experience. I conducted a series of clinical interviews with 29 individuals: 11 experts and 18 novice undergraduate students. Participants grouped organisms from photos and were asked to explain and draw their reasoning. I analyzed the resultant transcripts and artefacts using Grounded Theory. Three primary components of relatedness understanding were apparent in all participants: structural metaphor, grouping strategy, and grouping criteria. Structural metaphor refers largely to spatial conceptual metaphors describing the arrangement of taxa. Grouping strategy refers to group membership determination while grouping criteria refers to the properties used with strategies to determine group identity. Experts exhibited understandings consistent with novices and partially non-normative, suggesting that some repression of nascent relatedness understandings may be occurring or heuristics may be used in certain conditions. Further investigation of the impact of these understandings on tree-interpretation is necessary as well as whether or not learner generated representations of these understandings can be leveraged into normative interpretation of phylogenetic trees.

Boxfishes are a group of heavily armored Tetraodontiform fishes that are highly variable in shape. Disparification of shape could be driven by a simple performance trade-off between its two hypothesized primary functions: protection from predation and maneuverability. Alternatively, disparification could be driven by many-to-one mapping of shape to performance, where a relaxation in morphological constraint where many of morphologies have the same performance. We tested this by isolating the major features of the boxfish carapace shape and tested for their correlation to performance, as well as for a negative correlation between performances. We found that some features were correlated but very weakly, and that the two performances did trade-off but also weakly. This weak correlation primarily suggests that many-to-one mapping of shape to performance is driving disparification, which was unobserved in continuous 3D shape systems until this study.