RNA interference (RNAi), the process by which small RNAs (˜20-30 nt in length) derived from double-stranded RNA precursors can induce silencing of cognate sequences, was initially characterized in Caenorhabditis elegans. Since then distinct RNAi mechanisms and pathways have been described in diverse eukaryotes, suggesting that RNA-mediated silencing is an evolutionary conserved process in the eukaryotic domain of life. Core protein components of the RNAi machinery include Argonaute-PIWI polypeptides, RNAseIII-like endonucleases termed Dicers and RNA-dependent RNA polymerases. Although the archetypal domains of these proteins appear to have been assembled from prokaryotic sources, phylogenetic analyses indicate that the three components came together as a functional unit in the last common ancestor of eukaryotes. Consistent with this interpretation, core RNAi proteins are widely distributed among organisms in all major eukaryotic lineages, particularly Argonaute-PIWI polypeptides, which typify the key RNAi players. Nonetheless, the RNAi machinery has also been lost independently in multiple divergent species during evolution, suggesting that its ancestral function was not essential for unicellular life. The prevailing hypothesis is that the primeval RNAi machinery emerged as a defense system against parasitic genetic elements such as viruses and transposons. In contrast, a regulatory function of RNAi, through microRNAs and an assortment of other distinct small RNAs, may have evolved more recently, influencing newly arisen, lineage-specific processes such as cell differentiation and development in multicellular eukaryotes. However, defining the contribution of small RNA-mediated gene regulation to the evolution of organismal complexity remains a challenge for the future.