To investigate possible fatigue damage caused by wind loading, a variable-amplitude loading distribution was developed and applied to a two-span laboratory model of screw-fixed corrugated roof-cladding. It was found that the Miner cycle-ratio summation did not provide a reliable estimate of the fatigue life of the cladding-fastener assembly. More consistent results were obtained by summing cycle-ratios separately over crack initiation and crack propagation periods. The order of application of various load ranges was found to be a major parameter affecting the duration to failure. The laboratory results confirmed that the magnitude and duration of cyclonic wind loading, typical of that produced by cyclone Tracy, could cause fatigue failure of screw-fixed corrugated roof cladding. It is also concluded that the Australian wind-loading code may underestimate wind speeds within suburban areas. In an attempt to simulate the effects of random wind loadings on roof claddings, a variable-amplitude loading distribution and a sequence of load application have been recommended.