Classically, geneticists are accustomed to thinking of missense mutations as relatively benign, or at least less hazardous than clear null alleles such as large genomic deletions or nonsense mutations. At the cellular level, however, the opposite is often the case; missense mutants that are synthesized but fail to mature along the normal folding pathway may prove more troublesome than simple loss of expression. Cells possess a robust machinery to degrade mRNAs that, as a consequence of genetic lesions or errors in RNA processing, display premature-termination codons. The cues by which inappropriate breaks in open reading frames (ORFs) are recognized are uncertain (Maquat 1996)-even the subcellular location of nonsense-mediated decay is controversial-but the consequence is unambiguous: thanks to this monitoring process, many prematurely terminating alleles become true nulls. Experience with genetically altered mice confirms the wisdom of this regulatory strategy. Transgenic animals that express genes either ectopically or with abnormal sequences are often manifestly sick. On the other hand, gene targeting, which typically creates null alleles of the gene of interest, often leads to normal or nearly normal phenotypes, probably because of functional re-dundancy within the genome (Wanget al. 1996). In mice as well as humans, lack of gene expression may be more benign than abnormal gene expression. No general mechanism exists, at the nucleic-acid level, to convert missense mutations to null alleles, but the cell is far from helpless to restrict the expression of abnormal proteins. The machinery that assists the folding of nascent proteins to their final native conformation also identifies unfolded or misfolded polypeptides. When things go right, this same system leads to the efficient removal of these abnormal proteins. Genetic disease,