The intensity of joints varies spatially throughout any rock mass, therefore the description of the whole rock mass quality by a single conventional rock quality designation(RQD) value is rather insufficient. In practice, the rock mass quality changes along any line through a rock mass depending on the direction of measurement. A new statistical method is proposed for expressing rock quality data which involves deriving an empirical RQD distribution from scanline data. The method is demonstrated with reference to three different quality rock masses in Saudi Arabia. The distribution can be used in the assessment of the level of risk entailed in adopting a particular single design value of RQD.