Utilization of therapeutic ultrasound in the brain has been seriously limited by the commonly accepted view that these exposures would require that a piece of the skull bone be removed to allow the ultrasound beam to propagate into the brain. In this paper, the feasibility of delivering ultrasound therapy through the intact skull was studied. Sonications were performed through a piece of human skull with focused transducers at 0.248, 0.559, 1.0 and 1.68 MHz. The skull attenuated and distorted the field; however, a sharp focal spot was created at frequencies of 1 MHz or lower. At the higher frequency, the focus was destroyed. To investigate the feasibility of compensating for the ultrasound field distortion caused by the bone, phased array experiments were performed. Two arrays with 64 elements, operating at 0.6 MHz and 1.58 MHz, were used in these experiments. The phase shifts caused by the skull were measured for each element of the arrays and then compensated for by phase-control circuitry. These phase corrections allowed a sharp focus to be generated at both frequencies. Finally, tissue destruction was induced by using pulsed sonication through a piece of human skull in a rabbit brain in vivo at the frequency of 0.559 MHz. In summary, the results showed that transcranial delivery of therapeutic ultrasound into the brain may be feasible.