The neodymium-YAG laser was used to treat bleeding experimental canine gastric ulcers in four separate studies: (a) Extensive investigation by using the 1.06 micron YAG laser wavelength was performed to determine optimal YAG laser treatment parameters to maximize hemostatic efficacy while minimizing associated tissue injury. Eight out of 43 combinations of power, pulse duration, and spot size, with and without coaxial CO2 were found to be best. The range of settings tested were 30-90 W for power; 0.2-1 sec pulse duration; and 2.0- and 3.4-mm spot sizes, with and without coaxial CO2. (b) Five of these treatment combinations were studied in more detail to assess the respective effect of spot size, pulse duration, and presence or absence of coaxial CO2 on tissue injury. In spite of optimizing treatment parameters and conditions, each combination caused significant damage to the gastric wall. No acute perforations occurred. (c) Endoscopic YAG laser treatments were applied successfully. Efficacy and total energy requirements were compared with similar treatments previously done at laparotomy. (d) In an attempt to reduce tissue damage, the YAG laser was modified to produce a 1.3-microm wavelength. YAG laser treatment using this alternative wavelength caused more immediate damage than the 1.06 micron wavelength. From these studies we conclude: (a) YAG laser photocoagulation effectively stopped experimental gastric ulcer bleeding when applied both at laparotomy and endoscopy; (b) compared to untreated ulcers, YAG laser treatment caused significant tissue damage despite control of several important variables (power output, spot size, pulse duration, and coaxial CO2); (c) increased total energy, larger spot size, and longer pulse duration appeared related to increased tissue injury with YAG laser treatment; total energy was the single most important factor. (d) Successful hemostasis with endoscopic application of the YAG laser required more total energy than treatment of laparotomy. (e) Modification of the YAG laser wavelength to 1.3 microns did not reduce tissue injury.