The plastic strain controlled fatigue behaviour of polycrystalline Cu-15%Zn and Cu-30%Zn has been investigated with the aim of studying the effect of slip mode modification by the addition of zinc to copper. It has been clearly demonstrated, that true cyclic saturation does not occur in the plastic strain controlled fatigue of brass. This complicates the construction of a cyclic stress-strain (CSS) curve and thus the comparison with copper. A method to over-come this complication has been suggested. Surface observations on fatigued brass specimens show that individual grains tend to deform by Sachs type single slip. This behaviour has been described by the self-consistent Sachs-Eshelby model, which provides estimates of the CSS curve for brass polycrystals. Successive stages of primary hardening, softening and secondary hardening has been observed in the plastic strain controlled fatigue of brass. It has been found that the primary hardening is attributed to an increase of intergranular stresses whereas the sec-ondary hardening apparently is attributed to an increase of friction stresses. Investigations of the structural evolution show that the softening behaviour can be explained by the presence of short-range order (SRO). SRO promote the formation of extended dipole arrays which hardens the material. The formation of intense shear bands destroy the dipole arrays, which explains the cyclic softening. The present results reveal that Cu-30%Zn is a pure planar slip alloy, while Cu-15%Zn displays both planar and wavy slip. The mechanical and structural behaviour observed in brass resembles recent observations in 316L austenitic stainless steels, and the present results reveal that Cu-30%Zn and 316L have approximately the same fatigue life curve. This empha-sizes brass as being a convenient model system for the industrially important austenitic steels. A quantitative fatigue damage characterization has been carried out using a classification of sur-face cracks based on their length and growth behaviour. This has provided the basis for using a numerical Monte Carlo type model, which has been further developed to account for the ob-served intergranular damage evolution on Cu-30%Zn. With these modifications the model pre-dicts the fatigue life curve of Cu-30%Zn and 316L.