Let people who have to observe sickness and death look back and try to register in their observation the appearances which have preceded relapse, attack or death, and not assert that there were none, or that there were not the right ones. A want of the habit of observing conditions and an inveterate habit of taking averages are each of them often equally misleading