Dose-surface histograms are studied and compared with dose-volume histograms, as an evaluation tool for prostate treatment planning. For thin walled hollow organs, such as the rectum and bladder, the surface area irradiated is a more appropriate measure of the biological effect than the full volume. It is also more accurate and efficient to define the surface for a hollow structure and compute the surface area histograms. Application of the dose-surface histograms provide new insights into prostate treatment planning. A simple idealized geometry model demonstrates that the percentage surface area intersected by the geometric beam edge differs from the percentage volume intersected. For a group of prostate patients, it is shown that the dose-surface histograms yield substantially different results from the dose-volume histograms in ranking four-, six-, and, eight-field treatment plans and in calculating the fraction of the rectum irradiated to high dose. The difference in terms of surface area between these plans in the high-dose region is usually less than that in terms of the volume, and a reverse of plan ranking order can consequently occur. The percentage of organ surface irradiated to high dose is typically greater than the percentage volume by 5% to 10%. The use of the dose-surface histograms in analysis of organ motion and/or patient setup uncertainty, and analysis of rectal complications, is also discussed.