Hygiene measurement parameters for airborne asbestos are discussed from mineralogical and biological standpoints. It is concluded that at present there is no practicable alternative to the membrane filter/phase contrast optical microscope (MF/OM) method for routine use, despite its many limitations, but that some features, including the fibre size limits, should be revised. On present evidence fibre counts appear to provide a better index of hazard than respirable mass concentration. Considerations of respirability and biological activity together suggest that the counts should be of fibres in the size range: length greater than 5–10μm but less than 100 μm; width less than 1.5–2 μm (no minimum specified); aspect ratio greater than 5:1 to 10:1. Compared with present definitions ( l >5 μm, d 3:1) this reduces the maximum diameter and increases the minimum aspect ratio. The main drawback of the MF/OM method is that fibres thinner than ca. 0.4–0.2 μm (depending on the microscope, method of mounting the sample and the observer/s visual acuity) cannot be seen, although they contribute substantially to the number to an extent that varies with asbestos variety. This makes standardization of the method extremely difficult and is the main reason why different hygiene standards are needed for different asbestos varieties, especially between the amphiboles crocidolite and amosite. Further studies of the lung deposition of fibres thinner than 0.1–0.2 μm are needed. In animal experiments chrysotile has generally been more damaging than the amphiboles, in contrast to human epidemiological experience. This is probably due to the lesser durability of chrysotile when retained in the body over long periods, as evidenced by the surprisingly small amounts of chrysotile compared with amphibole found in human lungs at autopsy despite the much greater industrial usage of chrysotile. Many non-asbestiform amphiboles and other minerals give rise to elongated dust particles of aspect ratio>3:l. Discrimination between these and true asbestos particles is extremely difficult in individual cases, although overall size distributions differ. Recent codes for the MF/OM method indicate that discrimination is not to be attempted, but present practice in the U.K. is variable with marked effect on counting levels. Increasing the aspect ratio limit to > 5:1 or preferably > 10:1 would exclude many of the non-asbestiform particles but have relatively little effect on the true asbestos count. This supplements the biological evidence for a change from the present aspect ratio limit of 3:1 which was arbitrarily chosen by the ARC in 1958. Another problem for which there is little scientific guidance is the evaluation of clumps and other entities. The use of aerodynamic size selectors in sampling together with an upper size limit of length + width < 100 μm for ‘thistle-down’ particles should help. Other technical changes to the MF/OM method are suggested. The novel suggestion is made that total fibre length per unit volume (m/m 3 ) might reflect the hazard of asbestos exposure better than fibre number concentration and be easier to assess by man or machine—by counting crossing-points between fibres and a line drawn across the sample.