Background and Purpose: The effect of laser pulse width on calculus retropulsion during ureteroscopic lithotripsy is poorly defined because of the limited availability of variable pulse-width lasers. We used an adjustable pulse-width Ho:YAG laser to test the effect of pulse width on in vitro phantom-stone retropulsion and fragmentation efficiency. Methods and Materials: An Odyssey 30 Ho:YAG laser (Convergent Laser Technologies, Oakland, CA) with adjustable pulse width (350 or 700 mu sec) was used to treat spherical 10-mm plaster calculi in a model ureter (N = 40) and calix (N = 16) utilizing 200- and 400-mu m fibers (10 Hz, 1.0 J). Calculi were placed in a water-filled clear polymer tube, and laser energy was applied continuously in near contact until the stone had moved 8 cm. The time (seconds) and energy (joules) needed to cause the stone to traverse this distance was recorded. Stones were also placed in a stainless-steel mesh calix model in which retropulsion was limited. Laser energy was applied for 5 minutes at each pulse width. A laser-energy meter (Molectron Detector Inc, Portland OR) was used to quantify fiber transmission efficiency after 1 minute of continuous lithotripsy for each fiber at each pulse width. Results: Retropulsion was greater for stones treated at 350 mu sec, indicated by a shorter time to traverse the model ureter. For the 200-mu m fiber at 350 mu sec, the average time was 11.5 seconds nu 20.3 seconds at 700 mu sec (P < 0.001). The average total energy delivered was 114.9 J at 350, mu sec nu 199.8 J at 700 mu sec (P < 0.001). For the 400-mu m fiber at 350 mu sec, the average time was 5.8 seconds nu 11.9 seconds at 700 mu sec (P < 0.001). The average total energy was 57.1 J at 350 mu sec nu 127.3 J at 700 mu sec (P < 0.001). In the caliceal model, at 350 and 700 mu sec with the 200- and 400-mu m fibers, mass loss was 34.9% and 33.4% (P = 0.8) and 14.6% and 21.6% (P = 0.04), respectively. The reduction in energy transmission at 350 mu sec and 700 mu sec with the 200-mu m fiber after 60 seconds of continuous lasing was 8.82% nu 9%, respectively (P = 0.95). For the 400-mu m fiber, the transmission loss was 18.4% at 350 mu sec nu 4.4% at 700 mu sec (P = 0.0002). Conclusion: When treating ureteral calculi, retropulsion can be reduced by using a longer pulse width without compromising fragmentation efficiency. For caliceal calculi, the longer pulse width in combination with a 400-mu m fiber provides more effective stone fragmentation.