Understanding the mechanisms of evolution and the degree to which phylogenetic generalities exist requires information on the rate at which mutations arise and their effects at the molecular and phenotypic levels. Although procuring such data has been technically challenging, high-throughput genomic sequencing is rapidly expanding our knowledge in this area. Most notably, information on spontaneous mutations, now available in a wide variety of organisms, implies an inverse scaling of the mutation rate (per nucleotide site) with the effective population size of a lineage. This pattern appears to arise naturally as natural selection pushes the mutation rate down to a lower limit set by the power of random genetic drift rather than by intrinsic molecular limitations on repair mechanisms. Support for this idea derives from the relative levels of efficiency of DNA polymerases and mismatch-repair enzymes in eukaryotes relative to prokaryotes. This drift-barrier hypothesis has general implications for all aspects of evolution, including the performance of enzymes and the stability of proteins. The fundamental assumption is that as molecular adaptations become more and more refined, the room for subsequent improvement becomes diminishingly small. If this hypothesis is correct, the population-genetic environment imposes a fundamental constraint on the level of perfection that can be achieved by any molecular adaptation. It also implies that effective neutrality is the expected outcome of natural selection, an idea first suggested by Hartl et al. in 1985. Although generally viewed as an independent process, mutation also operates as a weak selective force, thereby playing a central role in “nearly neutral” hypotheses in evolution. Most notably, genes and proteins with more complex structures are subject to higher rates of mutational degeneration simply because they are larger mutational targets. However, because the mutation rate is very low at the nucleotide level, the efficiency of such mutation-associated selection becomes of diminishing significance in populations with small effective sizes. Thus, mutationally hazardous genomic and gene-structural features, which may or may not be adaptive, are expected to passively arise in lineages with small effective sizes. This general principle, the mutational-hazard theory, will be illustrated with examples including: 1) the differential expansion of intron numbers in various phylogenetic lineages; and 2) the diversification of protein-architectural features.