This work employs five commercial carbon-black filled elastomers made from ethylene-propylene-diene, nitrile, butyl and natural rubber compounds to study different manifestations of the stress-softening phenomenon known as the Mullins effect. The structure of the materials was characterised using density, hardness, thermal analysis and microscopy. Cyclic tensile deformations under uniaxial, constant-width and equibiaxial models were used to study the mechanical response following different strain histories. The mechanical response and the energies stored and dissipated were both dependent on the strain history. Elastomers with no strain history exhibited a relatively constant dissipated energy fraction, which reduced to a lower constant fraction when the historical maximum strain was higher than the current maximum strain. A physically based modelling approach with a simple representation of the evolution of bound rubber was proposed to account for this. The approach was able to qualitatively reproduce the energy maps observed in the experimental data. A set of time-dependent test protocols were developed in both uniaxial and biaxial modes of deformation to investigate the effect of strain history on stress-relaxation and stress-memory following cyclic loading. All filled rubbers relaxed an approximately constant fraction of the stress after a given time when the applied strain during stress relaxation was smaller than the historical maximum. Under these circumstances the relaxed fraction was independent of both the strain history and the current strain, with important implications for scragging. An empirical function involving just five parameters was proposed to predict the fraction of unrelaxed stress for an arbitrary strain history. Stress memory was observed following unloading, with a recovered stress magnitude approximately independent of history. Anisotropy was investigated through linear dimensional swelling measurements. A degree of anisotropy was observed specific to the manufacturing process: compression moulding led to transverse isotropy, while sheet-rolling led to full anisotropy. Following a strain history, the anisotropy was shown to evolve further.