I've thought of an extension to the idea of landscape reading in water dowsing. Someone who has worked with an older dowser (OK, there would have to be a beginning somewhere) would have a knowledge of successful wells. Someone with a good three-dimensional model of the land in their head could have built up, consciously or not, a model of the local aquifers, and be able to predict the hydrology by interpolating between known points. This would not help so much with hard rock areas where water may be flowing in fissures. It would mean that a dowser away from his area of practice would be less effective.

Penny