How can epidemiologist differentiate between common-source and host-to-host epidemics based only on the incidence of disease?