Consider the situation where some evidence e has been entered to a Bayesian network. When performing conflict analysis, sensitiv ity analysis, or when answering questions like "What if the finding on X had been y instead of x?", you need probabilities P(e' I h) where e' is a subset of e, and his a configuration of a (possibly empty) set of variables. Cautious propagation is a modification of HUGIN propagation into a Shafer-Shenoy like architecture. It is less efficient than HUG IN propagation; however, it provides easy access to P(e' I h) for a great deal of relevant subsets e'.