Alternate bars in straight alluvial channels are migrating or steady. The currently accepted view is that they are steady only if the width-to-depth ratio is at the value of resonance or if the bars are forced by a steady local perturbation. Experimental observations, however, seem to indicate that steady bars are also present in cases of migrating bars in the absence of a persistent perturbation. The companion paper by Mosselman (2009) provides a theoretical explanation. We review some experimental observations as well as long-term numerical tests using a 2D depth-averaged morphological model of a straight channel with non-erodible banks. Small random variations in total discharge are imposed at the upstream boundary. Rapidly growing migrating bars are found to develop first, but slowly growing steady bars are found to evolve subsequently, starting either from upstream or from downstream. Since steady bars are seen as a prerequisite to explain meandering of alluvial rivers, our findings imply that neither resonant width-to-depth ratios nor steady local perturbations are necessary conditions for the onset of river meandering.