Eight millimetre diameter angioplasty balloon catheters of both the Gruntzig and Olbert types from five manufacturers have been tested in vitro to establish bursting pressures and the changes in maximum and deflated diameters following repeated inflations, both when free and within a restraining sleeve. Maximum inflated diameters were within 10% of that stated and all types of balloon except one burst at a pressure greater than the recommended value. Deflated diameters were approximately 1 mm greater than insertion diameters, which are much smaller in the Olbert type. All balloons became a little larger with each of the first few distensions, and became stiffer. The maximum diameter was reached and remained constant after 10-15 distensions. Distension within a latex sleeve did not change bursting pressures, and it is considered that results from unconstrained testing can be extrapolated to behaviour in vivo. Computer modelling and calculation of maximum stress resultants also showed that calculated longitudinal and circumferential stresses are unaffected by applied restrictions. It is concluded that balloon technology is steadily improving and it is suggested that British Standards should be established for dilatation balloon catheters. Amongst other factors these Standards should include maximum recommended inflation pressures that are at least 2 atm less than bursting pressures, whilst the stated maximum diameter should be for fully extended balloons, and should have a tolerance of more than +/- 10%.