In spite of many reports on the toxicity of silver nanoparticles (AgNPs), the mechanisms underlying the toxicity are far from clear. A key question is whether the observed toxicity comes from the silver ions (Ag) released from the AgNPs or from the nanoparticles themselves. In this study, we explored the genotoxicity and the genotoxicity mechanisms of Ag and AgNPs. Human TK6 cells were treated with 5 nM AgNPs or silver nitrate (AgNO) to evaluate their genotoxicity and induction of oxidative stress. AgNPs and AgNO induced cytotoxicity and genotoxicity in a similar range of concentrations (1.00-1.75 µg/ml) when evaluated using the micronucleus assay, and both induced oxidative stress by measuring the gene expression and reactive oxygen species in the treated cells. Addition of N-acetylcysteine (NAC, an Ag chelator) to the treatments significantly decreased genotoxicity of Ag, but not AgNPs, while addition of Trolox (a free radical scavenger) to the treatment efficiently decreased the genotoxicity of both agents. In addition, the Ag released from the highest concentration of AgNPs used for the treatment was measured. Only 0.5 % of the AgNPs were ionized in the culture medium and the released silver ions were neither cytotoxic nor genotoxic at this concentration. Further analysis using electron spin resonance demonstrated that AgNPs produced hydroxyl radicals directly, while AgNO did not. These results indicated that although both AgNPs and Ag can cause genotoxicity via oxidative stress, the mechanisms are different, and the nanoparticles, but not the released ions, mainly contribute to the genotoxicity of AgNPs.