The stability of armour units is described by several formulae like Iribarren's, Hudson's and Van der Meer's. In these formulae the required weight of the armour elements depends strongly on the density of the used material. Therefore it may be useful to change this density. According to the stability formulae, higher density will result in lighter and smaller units. Smaller units may reduce the costs. Literature study: In all formulae the stability of armour elements is expressed by the dimensionless stability number Hs/Δ D. This stability number is a function of several parameters like unit shape, placing method, slope angle etc. In this function Δ or ρ are not present but there is a linear relation between Hs/D and Δ. Stability formulae for breakwater armour elements are based on model experiments with elements with densities up to approximately 3200 kg/m³. In the past several improvements on stability formulae and breakwater design were made. Several different shapes of armour units were developed and wave spectra were introduced in flume testing. By these improvements the stability formulae are refined and validated. Now it is possible to make concrete with a density up to 4000 kg/m³. According to the stability formulae this will result in very small and light armour units. But the stability formulae are until now not validated for densities up to 4000 kg/m³. It is not known whether the linear relation between Hs/D and Δ is still valid for these densities or not. In order to investigate the validity of stability formulae and the linear relation between Hs/D and Δ for high densities, experiments are recommended. Flume tests: During October 1999 flume tests on concrete elements with a density up to 4000 kg/m³ were performed in the Scheldt Basin of WL | Delft Hydraulics, location "de Voorst", The Netherlands. In order to investigate whether these high-density elements behave as expected and to compare the test results also tests with normal density (approximately 2300 kg/m³) were performed. The elements were attacked by irregular waves. Elements were placed by hand as well as dumped on the slope. It is concluded that the high density and normal density concrete elements do behave in a different way during placing of the elements. High-density concrete elements tend to lay more random on a slope. In order to compare test results in a correct way a new criterion, called "laying-roughness" is developed. Conclusions The following main conclusions are drawn: * The high-density concrete elements do behave as expected based on stability formulae. * The linear relation between Hs/D and Δ in the stability formulae for armour elements is still valid for densities up to 4000 kg/m³.