The use of laser Doppler flowmetry (LDF) to measure gastric blood flow was evaluated in the cat and man. The reproducibility of laser Doppler flowmetry recordings was studied in the feline stomach. In five cats flowmeter signals and venous outflow of the stomach were simultaneously recorded. The flowmeter recordings were made during operation and gastroscopy in 140 patients by means of two different (PF1-4-kHz and PF2-12-kHz) laser Doppler systems. The flowmeter recordings were highly reproducible during the cat experiments, with a coefficient of variation varying between 4% and 13%. Angulation of the probe within 60-120 degrees against the tissue under study did not affect the flowmeter signal. Pressure of the probe against the studied tissue attenuated the flowmeter signal 42 +/- 13% (n = 10). A significant correlation coefficient (r = 0.76; p = 0.01; n = 30) was obtained between flowmeter signal and venous outflow of the stomach. In man the PF2-12-kHz system yielded a higher flowmeter signal (14.5 +/- 6.9 V; 32 patients) than the PF1-4-kHz system (7.9 +/- 2.5 V; 108 patients). Flow dimensions were calculated by using results previously obtained in the intestines. The estimated blood flow values amounted to 31 +/- 10 and 57 +/- 27 ml min-1 100 g-1 for the PF1-4-kHz and PF2-12-kHz systems, respectively. In conclusion, laser Doppler flowmeter is a promising technique for studying gastric perfusion during operation and endoscopy. Motion artifacts are the major drawback of the present laser Doppler systems.