Thermal laser angioplasty uses constant laser power, producing widely variable tip temperatures in vivo. Results have been suboptimal. We studied the effect of 50400°C tip temperatures on depth of ablation at 192 sites on plaqued and normal human aorta in vitro, and the angiographic and histologic response in vivo of 300-400"C at probe/vessel ratios of 0.5-1.0, in 40 normal canine femoral artery segments. In vitro, there was a direct relationship between tip temperature and depth of ablation, r = 0.71 (all segments), r = 0.74 for fibrous plaque, but a poor correlation in fatty plaque r = 0.35. In fibrous plaque, there was proportionately more ablation at tip temperatures >300"C, mean depth 0.62 mm, than at 150-3OO0C, mean 0.37 mm, (P < .001). Ablation was similar in plaqued and normal aorta. In vivo, 300"C, 350"C, and 400°C produced similar effects. At probe/vessel ratios <0.8, only disruption of internal elastic lamina was observed. At ratios 20.8, spasm occurred in 39% (7/18), transmural damage in 28% (5/18), and perforation in one of 18. Ablation is not selective for plaque and is highly variable in fatty plaque. Tip temperatures above 300°C produce greater ablation than at lower temperatures. In clinical applications, probehessel rations 50.7 may be most appropriate, and it appears that thermal remodeling may contribute more to outcome than plaque ablation.