As much research has demonstrated, novel scientific concepts do not arise fully formed in the head of a scientist but are created in problem-solving processes, which can extend for considerable periods and even span generations of scientists. To understand concept formation and conceptual change it is important to investigate these processes in the contexts of scientific practice, that is, "in the wild." This article addresses "concept formation in the wild" through examining the relations between concept formation and physical and computational simulation modeling practices in two research laboratories in the bioengineering sciences. It argues that processes of concept formation and of building distributed cognitive systems are deeply entwined. (Contains 5 figures and 4 footnotes.)