Understanding the strengthening of small-scale materials and structures is one of the key issues in nanotechnology. Many theories exist, each addressing a small domain of experimentally observed size effects and invoking different mechanisms. Measurements of the stress–strain relationship of nickel foils in flexure by the load–unload method provide strikingly accurate data from the elastic region through the yield point and to high plastic strain. The data show that the effects on the rate of work-hardening due to crystallite size and sample size interact, whereas in existing theories they should be independent. Existing theories cannot be complete. The symmetry of the dependence of flow stress on grain size and structure size suggests that strengthening effects are due to a finite strained volume, however this is delimited.