The reproducibility and bias associated with the measurement of mean fibre diameter of fleece samples was evaluated for five methods of measurement. The instruments used were the airflow (AF), sonic fineness tester (SFT), liquid scintillation spectrometer (LSS), fibre fineness distribution analyser (FFDA), and projection microscope (PM). Duplicate fleece samples from 48 sheep within each of the Merino, Corriedale, Perendale, Romney, and Coopworth breeds were measured using each method. Overall mean diameters for these breeds were 23.0, 30.1, 37.0, 38.0, and 38.3 Itm, respectively. The variability of diameter measurement (betweenduplicate, within-sample variance) was 0.06,0.09, 0.48,0.64, and 0.96 Itm 2 for AF, SFT, LSS, FFDA, and PM methods, respectively. The deviation of each method from the mean of all five methods was examined. AF and SFT methods agreed closely and FFDA gave results similar to those of AF and SFT. These three methods gave lower readings than LSS and PM, with the negative bias becoming more pronounced as diameter increased. Relative to PM, LSS gave lower readings at fine diameters and higher readings at coarse diameters.