Muscle relaxation therapy assumes that generalized anxiety disorder (GAD) patients lack the ability to relax but can learn this in therapy. We tested this by randomizing 49 GAD patients to 12 weeks of Applied Relaxation (AR) or waiting. Before, during, and after treatment participants underwent relaxation tests. Before treatment, GAD patients were more worried than healthy controls (n=21) and had higher heart rates and lower end-tidal pCO2, but not higher muscle tension (A. Conrad, L. Isaac, & W.T. Roth, 2008). AR resulted in greater symptomatic improvement than waiting. However, 28% of the AR group dropped out of treatment and some patients relapsed at the 6-week follow-up. There was little evidence that AR participants learned to relax in therapy or that a reduction in anxiety was associated with a decrease in activation. We conclude that the clinical effects of AR in improving GAD symptoms are moderate at most and cannot be attributed to reducing muscle tension or autonomic activation.