Simulations of fluid flows often involve ``open" boundaries where fluid enters or leaves the domain. Mathematically, boundary conditions need to be imposed at these boundaries, but from a physical point of view there is no boundary there. The problem thus arise how to avoid creating artifacts which result from the choice of these unphysical boundary conditions. In a paper of Papanastasiou et al, a procedure for treating the outflow boundary in finite element simulations of the Navier-Stokes equations was introduced which appears to introduce no boundary condition at the outflow boundary. This method has been shown to work quite well. This raises the question why such a procedure works, since mathematically a boundary condition is clearly necessary. In this paper, it is shown that the procedure advocated by Papanastasiou et al actually introduces a hidden boundary condition, which is a special case of an ``extrapolation" boundary condition. The advantages of such a boundary condition are discussed.