Despite fire being one of the oldest and most important ecological disturbance processes on Earth, many aspects of fire–vegetation feedbacks are poorly understood, limiting their accurate representation in predictive models. Translating plant flammability traits to fire behavior and fire effects on ecosystems has proven a challenge with different disciplines approaching the problem at widely different scales. One approach has been a top-down assessment of ecosystem-level effects of vegetation structural characteristics and plant physiology on fuel properties such as fuel moisture. This approach has had some success, but is often forced to collapse species-specific variation into a small number of functional types and, as a practical necessity, usually focuses on highly plastic traits (e.g., moisture content) that can be modeled across an ecosystem without the need to characterize species-specific characteristics. The other approach grew out of trait-centric comparative ecology and focused on how traits might influence individual plant flammability. However, the degree to which such lab-based flammability trials reflect real species-specific differences maintained during wildland fires has been questioned. We review the history of these approaches, discuss where each has succeeded, and identify areas of research aimed at closing the apparent gap in scale.