Two related problems drive this project. The first is: what properties of development make organisms evolvable? A number of properties have come under scrutiny in recent years, such as modularity and robustness. Recently, Kirschner and Gerhart have proposed a theory of ´facilitated variation´, which includes a number of very general properties that make organisms more evolvable. This raises the second question: how should we discover, define clearly, and explore the relations between such properties? One way, of course, is by looking closely at the actual biology, honing our sense of what these properties are by examining particular examples, such as Kirschner and Gerhart do. An alternative approach would be to explore these ideas using a series of simple, abstract, models. These kinds of models proved enormously useful for understanding complex ideas in other areas of biology. They help by clearly articulating and disambiguating the assumptions in verbal ideas, and they have been used effectively to clarify complex debates. The project I propose for KLI is a speculative (and modest) contribution to such a modelling endeavour. I will focus specifically on the ideas central Kirschner & Gerhart’s theory of facilitated variation and try to capture the operation of these principles in simple, abstract, models. Their theory is a good candidate for this kind of study, for although it provides the an extremely clear and systematic approach to subject, there has been little modelling performed concerning the properties they outline. My modelling proposal is partially inspired by Dawkin´s visually compelling biomorphs — a modelling project which, like Kirschner and Gerharts, focused on the production of variation, rather than selection. It was during his attempts to modify the algorithm for development, to produce a broader and more diverse set of shapes, that he found that adding some features of development made the creatures more evolvable. The aim of the modelling project is to keep the the visual and interactive aspects of the biomorphs, but to add a series of progressive variations in the algorithms that govern development, incorporating features capable of expressing the ideas that Kirschner and Gerhart’s present in their work. It will be by contrasting the generative power of different models that these effectiveness of these properties can be ascertained. In addition, trying to build these ideas into simple models will force assumptions into the open, perhaps identifying parts of their theory which are underspecified. As evidenced by Dawkins’ original thoughts about evolvability, the mere act of trying to build models can be revealing in itself.