| A laser-diffraction tensile tester and a balance-beam creep apparatus were improved and applied to the study of free standing polycrystalline thin films with a strong $<$111$>$ texture. Studied are electron beam deposited Ag, Cu, Al films, and Ag/Cu multilayers consisting of alternating Ag and Cu layers with 1:1 thickness ratio. All films have a total thickness around 3 $/mu$m. In tensile testing, a thin polymeric two-dimensional diffraction grid was deposited on the film surface by microlithographic techniques. Local strains were measured from the relative displacements of two diffracted laser spots. This allows determination of Young's modulus, Poisson's ratio and, since large strains can be measured, the yield stress, ultimate tensile strength and fracture strain. The average values of the Young moduli and Poisson ratios, determined from hundreds of measurements, are 63 GPa and 0.42 for Ag, 102 GPa and 0.37 for Cu, 57 GPa and 0.41 for Al, and 87.5 GPa and 0.38 for Ag/Cu multilayers. In all cases, the Young moduli are about 20% lower than the values calculated from the literature data and are independent of the bilayer repeat length, $/lambda ,$ in the Ag/Cu multilayers. No 'supermodulus' effect was observed at small values of $/lambda .$ An anelastic model was proposed to explain the low Young moduli, the hysteresis loops on the stress-strain curves, and a 4.3 $/pm$ 0.2 GPa/decade strain rate dependence of the Young modulus in Al. The ductility of the Ag/Cu multilayers decreases when $/lambda$ is reduced. For $/lambda <80$ nm, the films are brittle at room temperature: they break without macroscopic plastic flow. For $/lambda >80$ nm, the yield stress increases linearly with $/lambda/sp[[-]/alpha]$ where $/alpha$ = 0.244. The results are compared to the predictions of Hall-Petch-type models. In creep testing, steady-state creep rates were measured on Cu films as a function of stress and temperature. In the high temperature-low stress region (100-650$/sp/circ$C, 5-90 MPa), the creep rate is described by $/dot/varepsilon =A[/cdot]/sigma/sp[n]$ exp$/[[-]Q/kT/].$ A core-diffusion controlled dislocation climb model was proposed to explain the stress exponent $n=3.32$ and the activation energy $Q=1.0$ eV. In the low temperature-high stress region (25-100$/sp/circ$C, 90-290 MPa), the creep rate is described by $/dot/varepsilon/propto[/rm exp]/[B/sigma/],$ which is consistent with power law breakdown. A thin film creeps 1000 times faster than the bulk material at low stress but behaves similar to the bulk material at high stress. We attribute the difference to the release of dislocations at the film surface at low stress and the blocking of dislocations by cell walls at high stress. At low stress, the film is too thin to contain dislocation cells, whose size scales inversely with stress. |