« It is also a good rule not to put overmuch confidence in the observational results that are put forward until they are confirmed by theory. »