This paper concentrates on the problem of fault interaction and earthquake triggering through the 1980 Irpinia, Italy, sequence. More specifically, this paper deals with the problem of the triggering of the second subevent by the mainshock. The interaction between the two segments is modeled through a dynamic Coulomb failure function. The aims of this paper are, first, to discriminate between the dynamic and the static stress effects on the triggering, if these effects exist, second, to estimate the fault strength relative to the initial state of stress, third, to determine the parameters of a slip-dependent friction law that lead to the observed delay of 20 s. Numerical simulations show that the critical slip Dc may range from 0.03 m up to 1.7 m, and that the initial slope of the friction law μ'(0) must be lower than 0.04 m-1. We show that the relative magnitude of the fault strength and the initial state of stress govern the existence and value of a Dc lower threshold under which the fault always ruptures before 13 s. A close to failure fault is not consistent with a critical slip Dc less than 0.8 m, whereas small values of Dc, typically 0.03 m, imply a far from failure fault. General results concern the effect of a dynamic stress pulse. We show that an event can be triggered by a transient stress pulse and that in this case the event can have an initiation duration much longer than the pulse duration. We show that it is possible to explain both the triggering and the time delay only with the effect of the transient stress pulse. This may explain aftershock triggering even in regions of negative Coulomb failure function or long distance triggering of earthquakes by propagating waves. ¿ 2000 American Geophysical Union