This paper presents the results of an evaluation of the probabilistic sampling methods applied to subsurface sounding (e.g. test pit, auger holes) as tools to measure the density distribution of subsurface archaeological remains. By means of the simulation of three types of sites with different density distributions of subsurface archaeological materials and the application of different sampling schemes and sampling fractions, the degree of error in representing the density distribution of subsurface materials achieved with different sampling alternatives is observed. It is concluded, on the one hand, that a sampling scheme scarcely used in field archeology, sampling by interval transects, produces fewer errors. On the other hand, beyond the expected relation between the size of the sample and the precision with which a universe is represented, it has been possible to estimate which range of sampling fraction produces acceptable errors.