Model-revision is where pupils, having learnt how to use a model, have to change it in some way. The intention is that it can represent a phenomenon in contexts other than those initially encountered or so that it can be used for purposes other than originally envisaged. Stewart, Hafner, Johnson and Finkel (1992) taught the skills of model-revision by having high school students use a simulation of scientific activity. First, a phenomenon was observed in groups; that experience was then shared between the groups; then the groups each designed an explanatory model; the groups then defended their models against the critique of other groups; and finally the groups revised their models until a degree of convergence was achieved. This emergent skill of model-revision was then used on standard problems in genetics, using historical papers and computer software. The historical papers showed them how the problem had been solved at the time of the invention of the solution. The computer software enabled them to simulate the process of change of the explanation and the emergence of an acceptable solution. In an interesting variant of the model-revision theme, Frederiksen, White and Gutwil (1999) used computers to support students in progressing through (revising) a sequence of models of higher abstraction/complexity. Students were taken through a sequence of models of electricity (the electricity-particle model, the aggregation model, the algebraic model) such that the entities and interactions of a simpler model provided the emergent properties of the next in the sequence through a process of revision.