INTRODUCTION This Item gives the results of fatigue tests on titanium and titanium alloy specimens@ excited by narrow-band loading of random amplitude with zero mean load in order to simulate the stress response to acoustic loading. The specimens@ also called coupons@ are small structural elements which are typical of those used in aircraft and space structures. The coupons are either simple cantilevers or built-up T-pieces. The specimens were vibrated until failure occurred. For a narrow band Gaussian process the number of upward zero crossings per unit time is nearly equal to the centre frequency of the process@ thus the number of cycles to failure@ or the equivalent endurance@ is equal to the product of the centre frequency (in Hz) and the time to failure (in s) (Reference 17). The equation relating Srms@ the root mean-squared value of alternating stress@ to N@ the number of cycles to failure@ is assumed to have a form similar to the Basquin equation. where is the fatigue strength exponent or the Basquin exponent and is a constant. The aim of the tests was to produce plots of stress against number of cycles to failure in a logarithmic coordinate system@ and to find a line which is best fitted to the data. The common logarithm of the stress is plotted against the common logarithm of the endurance and a straight line is fitted to data by the method of least squares. The slope of the line is a. The constant k=Srms(1)refers to the value of Srmsat 1 cycle. Although the S -Nline is obtained from coupon tests@ the results could be used to predict the life of more realistic stiffened panel structures. Extrapolation of the lines beyond 108 cycles@ where no data is available@ is expected to lead to an underestimate of endurance. The Item was originally published as ESDU 73010 (see also Reference 6). Failure is characterised by a decrease in the resonance frequency of the coupon and an increase in damping. Cracks are sometimes visible at failure. The stress is measured by a reference strain gauge or transducer@ located approximately at a position where a crack is expected. The materials used are commercially pure titanium@ a titanium alloy with 2% copper and a titanium alloy with 6% aluminium and 4% vanadium. Further details of the materials are given in Section 5. In one series of tests superplastic forming and diffusion bonding was used in the preparation of the coupons. Some tests were carried out at elevated temperatures.