Our understanding of how polyploidy influences gene evolution is limited by the fact there have been few molecular descriptions of particular genes and their expression in polyploid plants and their diploid progenitors. Here we use evidence from sequencing of genomic DNA and cDNA obtained by reverse transcriptase-polymerase chain reaction and 3' rapid amplification of cDNA ends to describe PgiC genes and their expression in two allotetraploid species of the wildflower genus Clarkia, C. delicata and C. similis. PgiC encodes the cytosolic isozyme of phosphoglucose isomerase (EC 5.3.1.9) and was duplicated in the ancestral stock of Clarkia, giving rise to paralogous genes PgiC1 and PgiC2. The active form of the PGIC enzyme is a dimer of like subunits. The electrophoretic patterns in the parent species show three bands of activity, representing two homodimers and a heterodimer of intermediate mobility, and are encoded by two genes. The electrophoretic patterns in the tetraploids also show three bands, but the tetraploids were expected to have multiple PGIC isozymes encoded by four genes. Our molecular studies demonstrated that each tetraploid has two PgiC1 and two PgiC2 genes, as predicted. One gene in each of them has been silenced by a single mutation, and a functional protein is no longer produced. In C. similis, PgiC2(mod) was silenced by a mutation of a single nucleotide in exon 5 that created a stop codon. In C. delicata, a polymorphism exists between a normal allele and a defective allele of PgiC2(epi) that has a deletion of a splice junction in intron 19 that results in the synthesis of a transcript lacking an entire exon, an example of exon skipping. The three-banded PGIC electrophoretic pattern of both tetraploid species arises because isozymes encoded by two or three of the genes comigrate. A very recent origin for both tetraploids is suggested by the near identity of several of their PgiC genes to their corresponding diploid orthologues and the absence of any acceleration in mutation rates. The problem of assessing genetic redundancy in tetraploids is discussed.