AIM--To present a new non-invasive method of performing a high definition topography of perfused vessels of the retina and the optic nerve head with simultaneous evaluation of blood flow. METHOD--By a combination of a laser Doppler flowmeter with a scanning laser system the perfusion of the retina and the optic nerve head is visualised. The principles of measuring blood flow by laser Doppler flowmetry are based on the optical Doppler effect: laser light scattered by a moving particle is shifted in frequency by an amount delta f. Our data acquisition and evaluation system is a modified laser scanning tomograph. The technical data are retinal area of measurement 2.7 mm x 0.7 mm, 10 degrees field with 256 points x 64 lines, measurement accuracy 10 microns, wavelength 670 nm and 790 nm, light power 100 microW and 200 microW, data acquisition time 2.048 s. Every line is scanned 128 times by a line sampling rate of 4000 Hz. By performing a discrete fast Fourier transformation over 128 intensities of each retinal point the laser Doppler shift is calculated for each retinal point. With these data a two dimensional map with 256 x 64 points of the retinal perfusion is created. The brightness of the pixel is coded by the value of the Doppler shift. Offline capillary blood flow is estimated in arbitrary units according to the theory of laser Doppler flowmetry in every region of interest of the perfusion picture. We estimated the reliability and the validity of the method. Retinal blood flow was measured by scanning laser Doppler flowmetry (SLDF) while varying intraocular pressure by a suction cup of three healthy volunteers. Measurements of retinal blood flow performed in 47 eyes by the presented method (SLDF) were correlated with data gained by a commercially available laser Doppler flowmeter. Perfusion pictures of the superficial retinal layer and of deep prelaminar layers in the optic nerve head are presented. RESULTS--The reliability coefficients r1 of 'flow', 'volume', and 'velocity' were 0.84, 0.85, and 0.84 respectively. We found a significant linear relation between SLDF flow and the ocular perfusion pressure (r = 0.84, p < 0.001). Comparative measurements of the retinal blood flow by SLDF and a commercially available laser Doppler flowmeter showed a linear and significant relation (flow r = 0.6, p < 0.0001, volume r = 0.4, p < 0.01). Capillaries of the retinal superficial vasculature or deep ciliary sourced capillaries of the optic nerve head became visible with a high resolution by the confocal technique dependent on the focus. Offline, the blood flow variables of areas of 100 microns x 100 microns were calculated. CONCLUSION--SLDF enables the visualisation of perfused capillaries and vessels of the retina and the optic nerve head in high resolution by two dimensional mapping of perfusion variables which are encoded by the Doppler signal. This method achieves simultaneously qualitative and quantitative evaluation of capillary blood flow of distinct areas of the capillary meshwork.