> I vaguely seem to remember the "fastcar" library having a cyl/tri > test. I could be wrong, but it's worth checking out. Don't have a > link handy but google should find it. Are you sure ? http://www.oxforddynamics.co.uk/collision_detection_routines.htm Cylinder-triangle is another kettle of fish anyway. You can find code for this in a JGT paper about particles or something. Pierre