Diets estimated from different proxies such as stable isotopes, stomach contents, and dental microwear often disagree, leading to nominally well-supported but greatly differing estimates of diet for both extinct and extant species that complicate our understanding of ecology. We show that these perceived incongruences can be caused by proxies recording diet over vastly different timescales. Field observations reveal a diet averaged over minutes or hours, whereas dental morphology may reflect the diet of a lineage over millions of years of evolution. Failing to explicitly consider the scale of proxies and the potentially large temporal variability in diet can cause erroneous predictions in any downstream analyses such as conservation planning or paleohabitat reconstructions. We propose a cross-scale framework for conceptualizing diet suitable for both modern ecologists and paleontologists and provide recommendations for any studies involving dietary data. Treating diet in this temporally explicit framework and matching the scale of our questions with the scale of our data will lead to a much richer and clearer understanding of ecological and evolutionary processes.
