The time between the infection of a primary case and one of its secondary cases is called a generation time. The distribution (and mean) of the generation times is derived for a rather general class of epidemic models. The relation to assumptions on distributions of latency times and infectious times or more generally on random time varying infectiousness, is investigated. Serial times, defined as the times between occurrence of observable events in the progress of an infectious disease (e.g., the onset of clinical symptoms), are also considered.