Previous studies have shown that long thin asbestos fibres are more pathogenic in in vivo and more active in in vitro assays than short fibre samples. In the present study a long fibre amosite asbestos sample and a short fibre sample prepared from it were tested for ability to cause inflammation in the peritoneal cavity of the mouse; a UICC sample intermediate in fibre size and an inert compact dust, TiO2, were also tested. The ability of the dust samples to cause inflammation, as judged by macrophage and neutrophil recruitment, was ranked in the order long fibre greater than UICC greater than short fibre greater than TiO2. Ability of amosite samples to cause inflammation was therefore related to the proportion of long fibres. The enhanced ability of long fibres to cause inflammation and cause macrophage activation is probably a key factor in the ability of long fibres to cause pulmonary fibrosis and may also be important in fibre carcinogenesis.