Streamwise boundary undulations induce spanwise vorticity in flows. In Newtonian flow, vorticity decays exponentially with distance from the undulation. However, recent theory for inertioelastic polymeric flow predicts vorticity amplification in a "critical layer" far from the site of disturbance. Here we present the first experimental evidence for the existence of such critical layers, demonstrating their measurable role in real polymeric flows with inertia. These vorticity amplification effects should be considered in all evaluations of vorticity dynamics in inertioelastic flows with streamline curvature.The addition of even minute quantities (parts per million, ppm) of high-molecular-weight flexible polymers to a Newtonian solvent imparts a small but important degree of elasticity to the fluid, dramatically modifying its macroscopic flow behavior. Some well-known examples are the increase of the pressure drop measured across porous beds [1,2], the reduction of turbulent drag [3][4][5], and the modification of jet breakup, spray atomization, and drop impact behavior [6][7][8].In the context of turbulent drag reduction, dilute polymer additives dampen streamwise vorticity via a mechanism attributed to a resistive polymer torque (the curl of the polymer force) [9][10][11]. However, recent experiments and simulations indicate that such fluids can support an entirely new elastoinertial turbulence (EIT) dominated by spanwise vorticity rolls [12,13]. While EIT is thought to share mechanistic similarities with purely elastic instabilities and elastic turbulence [14,15], EIT clearly depends on inertia, and the details of the interactions underpinning the resulting self-sustaining chaotic motion remain to be fully explained.The influence of fluid elasticity on spanwise vorticity is less well understood than its effect on streamwise rolls. However, the dominance of spanwise-coherent structures in EIT suggests their important role in drag reduction. Spanwise vorticity can be generated by streamwise undulations in flow, which can be introduced in a controlled way by employing a regular wavy surface. Recent linear theory for polymer solutions in simple shear demonstrates that even small-amplitude waviness can cause significant vorticity amplification in critical layers located some distance away from the wavy surface where vorticity is injected [16]. This is fundamentally different