This paper describes an experiment in which the performance of cane walkers with the traditional straight long cane and a redesigned cane, the curved cane, was compared. The curved cane has a curve where the tip touches the ground. Participants were 18 experienced cane walkers who were totally blind. The aspects of cane walking that were investigated included obstacle detection, drop-off (slope) detection and walking speed. The performance with both canes was investigated in two different ways: (1) by means of constructed courses in which objective measures of cane walking were derived; and (2) by means of more qualitative measures based on the participants' experiences with the curved cane during a 3-week try-out period. Results showed that obstacle-detection was significantly better with the curved cane, whereas drop-off detection and walking speed were comparable for the two canes. The participants' experiences mirrored these results.