Biologists use models in two distinct ways that have not been clearly articulated. A model may be used either as an exemplar of a larger group, or as a surrogate for a specific target. Zebrafish serve as an exemplary model of vertebrates in developmental biology; rodents are both exemplary vertebrates and specific surrogates for humans in biomedical research. The distinction between exemplary and surrogate models is important, because the criteria for and implications of model choice diverge in significant ways, depending on which role the model is to serve. So, too, do the kinds of conclusions we can legitimately draw from model-based research. The divergence derives in part from the use of the two sorts of models to answer different kinds of questions: exemplary models most often serve basic research, while surrogate models are used when the target species we ultimately want to learn about is inaccessible or difficult to study, as in medical research. There are many reasons to consider exemplary and surrogate models separately: they are suited to different tasks and contexts, rest on different assumptions, and, finally, they have unique limitations.
