Shadows are of great importance to computer graphics, especially for interactive animation, due to the visual cues which they provide. Most research in the field of dynamic shadowing has focussed on shadows from point light sources, since they are much easier to compute than those from area lights. This means that such shadows have sharp boundaries, and so resultant visualisations often appear sterile and unrealistic. We use discontinuity meshing as the underlying mechanism for shadow construction, and develop ideas which allow the information calculated by discontinuity meshing to be f updated, rather than re-computed, each frame. This has led to an algorithm which can update shadows interactively, and more importantly an algorithm which displays good scalability characteristics.