The composition of 23 concrete mixtures was varied in five separate series to evaluate the influence of porosity on the <sup>222</sup>Rn exhalation rate. In each series, a range in porosities is obtained by varying (1) the amount of cement, (2) type of cement (Portland or blast furnace slag cement), (3) the amount of water at a fixed cement level, (4) addition of an air entraining agent, or (5) the amount of recycled aggregates. The porosities ranged from 1% to 16%. The <sup>222</sup>Rn exhalation rate is normalized to the <sup>226</sup>Ra activity concentration and expressed as the <sup>222</sup>Rn release factor to eliminate the effect of differences in <sup>226</sup>Ra activity concentrations among the various concrete mixtures. Since most <sup>222</sup>Rn originates from the cement, a <sup>222</sup>Rn release factor based on the amount of <sup>226</sup>Ra introduced by the cements appeared to be more adequate. Although the methods to attain the porosities in the concrete mixtures differ widely, this cement-related factor corresponds well with the capillary porosity of the mixtures. Since the water-to-cement ratio of the fresh paste is a good indicator of the capillary porosity, this is the guiding factor in the fabrication of concretes low in <sup>222</sup>Rn exhalation. The lower the water-to-cement ratio, the less capillary pore area will be available from which <sup>222</sup>Rn can emanate from the mineral matrix into the pore system. The good correlation between the cement-based <sup>222</sup>Rn release factor and literature data on the internal capillary pore area support the results of this study. Copyright © 2011 Health Physics Society.