Cleaning by a horizontal water jet, impinging onto a soiled Perspex vertical plate, is described. The plate, the substrate, was coated with PVA or petroleum jelly, the soil. The substrate was either. (i) fixed, for batch tests in which the cleaned area, roughly circular, grew with time, or (ii) the substrate moved vertically up or down in its own plane, the water jet remaining fixed; this reproduced the effect of a jet moving across a surface for cleaning, as found in real tank cleaning operations. In the batch experiments, growth of the radius a of the cleaning area is well described, at early times t, by a(5) - a(o)(5) = K-5 (t - t(o)), a(o) being the initial radius of the cleaned area at time t(o); K is a constant. At later times with petroleum jelly, the cleaning front reached a maximum value, when the outward momentum of the radially flowing water film balanced the strength of the soil. This maximum value is modelled as a ramp of viscoplastic soil inclined at angle chi to the substrate surface, where chi was found to vary from 7 degrees to 25 degrees. In the tests of continuous cleaning of petroleum jelly, a lengthening cleaned area, of width w(c), was observed on the moving substrate. Near the jet was a stationary clean front, whose shape looked like half an ellipse. This shape, and the width wc, are well described by theory (Wilson et al., 2015, 123, 450-459) using parameters from the above-mentioned batch experiments. This establishes a good link between batch and continuous cleaning experiments. (C) 2016 Elsevier Ltd. All rights reserved.