The small bodies of our solar system are the remnants of the early stages of planetary formation. A considerable amount of information regarding the processes that occurred during the accretion of the early planetesimals is still present among this population. A review of our current knowledge of the density of small bodies is presented here. Density is indeed a fundamental property for the understanding of their composition and internal structure. Intrinsic physical properties of small bodies are sought by searching for relationships between the dynamical and taxonomic classes, size, and density. Mass and volume estimates for 287 small bodies (asteroids, comets, and transneptunian objects) are collected from the literature. The accuracy and biases affecting the methods used to estimate these quantities are discussed and best-estimates are strictly selected. Bulk densities are subsequently computed and compared with meteorite density, allowing to estimate the macroporosity (i.e., amount of voids) within these bodies. Dwarf-planets apparently have no macroporosity, while smaller bodies (<400 km) can have large voids. This trend is apparently correlated with size: C- and S-complex asteroids tend to have larger density with increasing diameter. The average density of each Bus-DeMeo taxonomic classes is computed (DeMeo et al., 2009; Icarus 202). S-complex asteroids are more dense on average than those in the C-complex that in turn have a larger macroporosity, although both complexes partly overlap. Within the C-complex asteroids, B-types stand out in albedo, reflectance spectra, and density, indicating a unique composition and structure. Asteroids in the X-complex span a wide range of densities, suggesting that many compositions are included in the complex. Comets and TNOs have high macroporosity and low density, supporting the current models of internal structures made of icy aggregates. Although the number of density estimates sky-rocketed during last decade from a handful to 287, only a third of the estimates are more precise than 20%. Several lines of investigation to refine this statistic are contemplated, including observations of multiple systems, 3-D shape modeling, and orbital analysis from Gaia astrometry.