By changing the scale of measurement, molecular epidemiologyimproves on traditional epidemiology also in another importantrespect: it goes beyond associations and is in principle able toshed light on the mechanisms of disease causation, rather thanjust hypothesize them, because it can make appropriate tests