Analysis of stress distribution along imbricate surfaces within a sand-box model of an accretionary wedge is used to calculate the condition for slip along an active frontal fault and the basal decollement. The same theory is applied for finding the stress limit that allows formation of a new fault in front of the sand wedge. When this stress limit is reached, the sole thrust propagates rapidly forward along the basal decollement and a new fault is created. Comparison between theory and experiment shows that stresses would be underestimated, if compaction of the shortened layer is not accounted for in the calculations. ¿ 2001 American Geophysical Union