Continental convergence between Arabia and Eurasia is taken up by distributed deformation in Iran. At wavelengths large compared with the thickness of the lithosphere this deformation is best described by a continuous velocity field. The only quantitative source of information on the spatial distribution of strain rates within Iran is the record of earthquakes. We find that we can reproduce the style of deformation observed in the seismicity by simply minimizing the rate of work in a continuous viscous medium that has to accommodate the Arabia-Eurasia plate motion between the defined shapes of Iran's rigid borders. When, in addition, we specify central Iran, Azerbaijan, and the southern Caspian basin to be relatively rigid blocks within the deforming zone, then the fit to the style of the observed strain rate distribution is even better. We conclude that much of the pattern of deformation in Iran is predetermined by the shape of its rigid borders and by the disposition of relatively rigid blocks within it. This is likely also to have been a common occurrence in older orogenic belts. We confirm earlier suggestions that earthquakes between 1909 and 1992 can account for only a small part (~10--20%) of the total deformation required by the convergence between the Arabia and Eurasia plates. We then show that the whole plate motion can be accommodated by a velocity field with the same orientations and relative magnitudes of principal strain rates seen in the earthquakes but with larger absolute magnitudes. There is therefore no requirement that the large proportion of aseismic deformation in Iran is substantially different in style, orientation, or distribution from that released seismically in the earthquakes. ¿ American Geophysical Union 1995