RE: non-compressible, laminar flow, newtonain fluid. Viscosity is often defined by looking at a fluid between two plates with the upper plate moving at a small velocity relative to the lower plate. The plate moves in the x direction and a velocity gradiant is created. Viscosity is then defined as: Fx/Axz= Sxy = mu dVx/dy. Fx = force between adjaycent fluid layers in the x direction. Axz = area between the fluid layers Sxy = stress in the direction of the force (x) where y is the direction of the normal to the area. mu = viscosity dVx/dy = Velocity gradient in the y direction. This is the standard way texts define viscosity (unless I've messed it up!). Next this idea is generalized to multiple dimensions and the stress tensor is defined. This is where I need help. It is argued that for non rotating fluids that the stress tensor must be symmetric. In particular Sxy = Syx. I see this in one sense. Usually this idea is arrived at by noting that a small element would develop an infinite torque or moment if it was not symmetric....okay good. BUT: Finally my problem: In the simple example, how does a force in the Y direction develop. How does a normal force to the direction of fluid flow, Fy, exist. By definition of a fluid, if Fy exists, the fluid must shear....Any help or reference would be most appreciated.