Blade erosion in wet steam turbines is considered to be preceded by the collection of large water droplets on the blade concave surfaces; this is followed by steam breaking up the resulting water film at the stationary blade trailing edges into fairly coarse, slow-moving droplets, and culminates in their collision with the backs of the following moving blade row. Experimental information on droplet collection and on the breakup of water film at the trailing edges has been obtained on turbine blade cascades in a wet air tunnel, and suggestions have been made for relating these results to low pressure steam turbines. The erosion of tool steel and Stellite 6 blade shield materials has been measured in a contrarotating test rig for calculated impact speeds of 1730 ft./s; the best erosion rates to compare appear to be those in a tertiary zone where the rate of weight loss has fallen from a maximum to a lower constant value. The influence of vacuum on specimen weight loss is also considerable, being second only in severity to that of speed over the ranges considered, and this has been attributed to aerodynamic influences on the number of droplets reaching the specimen surface in addition to a reduction in the true velocity of impact. The possibility of reducing erosion in turbines by direct water extraction from hollow stationary blades has been demonstrated on an experimental arcuate cascade. Trailing edge slots have been found to be more effective than face slots if the depression in the hollow centre can be maintained below the trailing edge pressure.