A decade-long rifting episode that began in the Krafla volcanic system, north Iceland, in 1975 caused substantial, regional, postevent anelastic deformation. This was modeled as viscous relaxation in an elastic/viscoelastic structure by Hofton and Foulger . The results from modeling the deformation detected in north Iceland have far-reaching implications both for local and regional processes and for the fundamental behavior of deformation around spreading plate boundaries in general. Tilt in the vicinity of the Krafla volcano fits the model well after 1988/1989 which suggests that the volcano magma chamber stopped inflating/deflating in 1988. A viscosity of 0.8¿1018 Pa s was required to match the local tilt data, less than that predicted for north Iceland as a whole. Vertical motion measured using the Global Positioning System (GPS) 1987--1992 around the ice cap Vatnaj¿kull is inconsistent with isostatic uplift. Using the elastic-viscoelastic model to predict motion in other regions of Iceland suggests that the deformation effects of the Krafla episode are significant in many parts of Iceland and should be taken into account when modeling deformation there. Though not a realistic plate boundary model, interesting complexities of the elastic-viscoelastic model are highlighted by deformation modeling of an infinitely long dike. This predicts that the amount of horizontal displacement close to the dike may exceed the amount of initial dike opening early in the spreading cycle. A more realistic approximation to the plate boundary in north Iceland, involving five overlapping segments that experience dike emplacements at discrete intervals, suggests that the width of the zone within which transient, time-dependent deformation occurs may be several hundred kilometers wide, considerably wider than the neovolcanic zone. A kinematic approach to describing plate motions is not appropriate close to spreading plate boundaries and elsewhere where the viscosity of the Earth is low. ¿ American Geophysical Union 1996