Long-term inhalation and intraperitoneal injection studies were undertaken with laboratory rats treated with a specially prepared short-fibre sample of Canadian chrysotile asbestos. This was compared, at an equal mass dose, to dust generated from the same chrysotile batch so as to contain the highest possible number of long fibres. The long-fibre cloud contained roughly five times more fibres greater than 5 micron in length as seen by phase contrast optical microscopy (PCOM). For increasing lengths, the ratio between the dust clouds increased progressively, reaching over 80: 1 for fibres greater than 30 microns in length. Rats treated with long-fibre chrysotile developed six times more advanced interstitial fibrosis (asbestosis) than animals treated with short-fibre chrysotile and three times more pulmonary tumours. At the end of the 12-month dusting period, three times more short chrysotile than long had been retained in the rat lung tissues. During the following 6 months, however, the short-fibre chrysotile was removed from the lungs much more rapidly than the long. Following intraperitoneal injection at a mass dose of 25mg of dust, both long and short chrysotile produced mesotheliomas in more than 90% of rats. At a dose level of 2.5mg of dust, the short-fibre chrysotile produced mesotheliomas in only one-third as many rats as the long-fibre dust which still produced mesotheliomas in more than 90% of animals injected. At a dose level of 0.25mg of dust, the short-fibre chrysotile produced no mesotheliomas while the long-fibre chrysotile still produced these tumours in 66% of rats. In the two highest doses, where short-fibre chrysotile produced mesotheliomas, the mean tumour induction period was significantly longer than for tumours produced by long chrysotile.