What impact does epigenetics have on evolution by gene duplication? This question was of interest to Susumu Ohno, and was addressed by Sergei Rodin and I, with the finding that epigenetic silencing soon after gene duplication is likely to be quite important, even essential, for evolution by gene duplication [Rodin and Riggs, 2003, J. Mol Evol, 56:718]. These studies led to the proposal by Branciamore, et al [2014, Proc Natl Acad Sci, USA, 111:6353] that stochastic epigenetic changes during a developmental window should enhance the rate of evolution. Extending this study to a treatment of transposons [Branciamore et al., 2015, AIMS Genetics 2:148] led to a surprising prediction that one should find evidence for balancing selection. This prediction is supported by our recently published studies on hybrid cell lines [Branciamore et al., 2018, Proc Natl Acad Sci, USA, 115: E10379] where we do, indeed, find evidence for balancing selection, and we also find that a high percentage of developmentally-specific autosomal genes are expressed from only one allele in any given astrocyte-like cell, i.e., they show monoallelic expression. It is thus likely that no two astrocyte cells have the same epigenetic identity. Whether this is true in vivo and for other cell types remains to be determined. The implications of stochastic changes during relatively brief developmental windows, leading to frequent cell-lineage-specific monoallelic expression, will be considered. It is proposed that mammals are cellular mosaics for many autosomal genes, as they are for X-linked genes.