sis. These mutants, hop2-2 and hop2-3 ,h ave short siliques and a reduced number of seeds but otherwise appear normal. On closer inspection, they exhibit strong abnormalities in meiotic progression. During male meiosis, normal chromosome pairing does not occur in prophase, and very few bivalents (but numerous univalents) are observed among hop2-2 metaphase cells (see figure). Segregation at meiosis II yields unbalanced pools of chromosomes at the end of meiosis; thus, four haploid daughter cells are not always produced, leading to abnormal pollen and, ultimately, fewer seeds. It turns out that in these mutants, RAD51 and DMC1 are loaded onto the nucleoprotein filament as usual but only the intersister DNA repair pathway is activated, making hop2-2 and hop2-3 excellent tools for the study of interhomologous recombination bias. The hop2-2 phenotype is correlated with a decrease in HOP2/MND1 complex abundance, while the hop2-3 mutant contains truncated HOP2 protein, which binds to MND1 in an inefficient manner compared with wild-type HOP2. Therefore, both mutants have less available functional HOP2/MND1 complex than the wild type, which makes it more difficult for DMC1 to enable repair via the homologous chromosome. By contrast, genetic analysis demonstrated that HOP2/MND1 is dispensable for RAD51-mediated intersister DNA repair. In addition, mutant analysis revealed that in the presence of DMC1 and in the absence of HOP2, RAD51-mediated repair is suppressed, suggesting that DMC1 is a negative regulator of RAD51 during meiosis. This negative regulation appears to be transient and is important for the establishment of DMC1-mediated, interhomolog connections during meiotic DNA repair. Thus, while RAD51 andDMC1bothformanucleoproteinfilament