1. We evaluated the use of non-radioactive fluorescent-labelled microspheres (FM) for the measurement of regional myocardial blood flow (RMBF) in an ischaemic sheep model, 2, Injection of FM directly into the coronary artery was compared with left atrial injection. There was a good correlation in the measurement of RMBF between these two injection methods (r = 0.92; n = 107 data points), Injection into the coronary artery requires less FM (one twentieth of that required by atrial injection) and is more economical. 3, The use of a fluorescent technique without filtering myocardial tissue was investigated, Calibration curves from the fluorescence plus myocardial tissue samples were similar to those of the pure fluorescence samples and both showed a linear relationship between fluorescent intensity and the number of microspheres (r > 0,97), These results indicate that the extraction of six fluorescent dyes (blue-green, yellow-green, green, orange, red and crimson) directly from the aqueous solution using ethyl acetate is effective, 4, A subendocardial ischaemic model was produced by partially occluding the circumflex artery (CxA) with concomitant left atrium (LA) pacing, During ischaemia, the endocardium/epicardium (Endo/Epi) flow ratios in the ischaemic area changed from 1,04 +/- 0,12 to 0,47 +/- 0,17 (P < 0,05; CxA injection) and from 1.08 +/- 0.12 to 0,51 +/- 0,05 (P < 0,05; LA injection), The ratio in the non-ischaemic area remained unchanged (1.12 +/- 0.26 to 1,01 +/- 0.22; not significant), 5, RMBF calculation using coronary inflow as the reference flow was also compared with that using the traditional method, We found that, in this study in which a non-filtering technique was applied, using coronary inflow as the reference flow was superior to the conventional distal sampling method. [References: 23].