Particle emissions from twelve buses, operating alternately on low sulfur (LS; 500 ppm) and ultralow sulfur (ULS; 50 ppm) diesel fuel, were monitored. The buses were 1-19 years old and had no after-treatment devices fitted. Measurements were carried out at four steady-state operational modes on a chassis dynamometer using a mini dilution tunnel (PM mass measurement) and a Dekati ejector diluter as a secondary diluter (SMPS particle number). The mean particle number emission rate (s(-1)) of the buses, in the size range 8-400 nm, using ULS diesel was 31% to 59% lower than the rate using LS diesel in all four modes. The fractional reduction was highest in the newest buses and decreased with mileage upto about 500,000 km, after which no further decrease was apparent. However, the mean total suspended particle (TSP) mass emission rate did not show a systematic difference between the two fuel types. When the fuel was changed from LS to ULS diesel, the reduction in particle number was mainly in the nanoparticle size range. Over all operational modes, 58% of the particles were smaller than 50 nm with LS fuel as opposed to just 45% with ULS fuel, suggesting that sulfur in diesel fuel was playing a major role in the formation of nanoparticles. The greatest influence of the fuel sulfur content was observed at the highest engine load, where 74% of the particles were smaller than 50 nm with LS diesel compared to 43% with ULS diesel.