The detection of void and/or partially or fully filled caves has been attempted using the resistivity method. This method has been used for this proposal by several authors, Spiegel et al. (1980), Smith (1986), Andrade Afonso (1987), Andrade Afonso et al. (1988 e 1989). Spiegel et al. (1980) considered the caves as void zones using the integral equation method. to obtain the solution for the pole-dipole array. Andrade Afonso (1987) and Andrade Afonso et al. (1988 e 1989), generalized this method for the dipole-dipole array. However, the method is unstable for voids close to the terrain surface. We tested the hypothesis of approaching, for numerical proposals, the void zones to areas of high resistivities in relation to the surrounding environment. The results obtained in an automated electrical analog test tank (2,1x1,25x1,25 m), filled with a 26 Ohmxm copper sulphate aqueous solution and in the field (caves with known geometry) were compared with results obtained using numerical techniques (Smith et al., 1984 and Sazaki, 1989) suitable for continuous media. The experimental pseudo-sections for the pole-dipole, as well as, for the dipole-dipole arrays fit very well with the numerical ones. On the other hand, we used the dipole-dipole array in several "young" karstic zones of (Algarve) in order to test the ability of the method to detect groundwater circulation. The models obtained fit very well with the known groundwater circulation zones and proved to be a powerful tool in the evaluation of groundwater resources.