The balance of forces in the sea ice model of Hibler <1979> is examined. The model predicts that internal stress gradients are an important force in much of the Arctic Ocean except in summer, when they are significant only off the northern coasts of Greenland and the Canadian Archipelago. A partition of the internal stress gradient between the pressure gradient and the viscous terms reveals that both are significant, although they operate on very different timescales. The acceleration term is generally negligible, while the sum of Coriolis plus sea surface tilt is small. Thus the seasonal average force balance in fall, winter, and spring is mostly between three terms of roughly equal magnitudes: air drag, water drag, and internal stress gradients. This is also true for the monthly average force balance. However, we find that there is a transition around the weekly timescale and that on a daily basis the force balance at a particular location and time is often between only two terms: either between air drag and water drag or between air drag and internal stress gradients. The model is in agreement with the observations of Thorndike and Colony <1982> in that the correlation between geostrophic wind forcing and the model's ice velocity field is high. This result is discussed in the context of the force balance; we show that the presence of significant internal stress gradients does not preclude high wind-ice correlation. A breakdown of the internal stress gradient into component parts reveals that the shear viscous force is far from negligible, which casts strong doubt on the theoretical validity of the cavitating fluid approximation (in which this component is neglected). Finally, the role of ice pressure is examined by varying the parameter P*. We find a strong sensitivity in terms of the force balance, as well as ice thickness and velocity.¿ 1997 American Geophysical Union