Retroposon-like elements constitute >20% of the mammalian genome. Most of the elements belong to the "long interspersed nuclear elements 1" (L1) class, which are incomplete retroposon elements at the 5′ end and carry multiple stop codons in the internal reading frames (20-21). In contrast, the VL30 retrotransposons contain LTRs but, like the L1 retrotransposons, are hobbled functionally by stop codons in the internal coding regions (1, 7). A minority of the elements from both classes present in rodent DNA is transcribed, and recent data suggest that the expression and movement of these elements may be playing a role in normal development (1-4, 22). DNA coding for "micro-RNAs" or noncoding regulatory RNAs has been found at chromosomal breakpoints associated with B cell neoplasms (23, 24) and to be amplified in B cell non-Hodgkin lymphomas as well (25). The work of Garen and colleagues (1) has decisively delineated how a noncoding RNA transcript arising from a retrotransposon could contribute to the evolution of cancer or leukemia through two mechanisms: integration mutagenesis of genes in cis to the integration site, and the coordinate up-regulation in trans of classes of growth promoting protoconcogenes at multiple sites throughout the genome. The seminal work of Garen and colleagues (1) has also underscored the potential importance of the VL30 elements for normal development. Their experiments have provided a model system in which to study the role of retrotransposons in early mammalian development, their action in germ cells and somatic cells that leads to human disease, and the potential risk they may confer on gene therapy trials based on retroviral vectors produced in mouse cells. Clearly, these important results have raised questions that are eminently testable and the work that will be stimulated in many laboratories should produce exciting discoveries over the next few years. © 2005 by The National Academy of Sciences of the USA.