Previously, the highest producing (HP) recombinant CHO subclones isolated at various methotrexate (MTX) levels showed different antibody production stability during long-term culture, although they were clonally derived from CS13 transformant. In this study, genetic basis for their difference in antibody production stability was investigated using southern blot hybridization and fluorescence in situ hybridization (FISH) techniques. Southern analysis of HP subclones revealed that light-chain (LC) and heavy- chain (HC) cDNAs were located closely within 23 kb on an amplification unit, and the configuration of LC and HC cDNAs within this amplification unit was not disrupted during long-term culture in the absence of MTX. However, when LC and HC genes were localized on the metaphase chromosomes of HP subclones using FISH, the amplified sequences were present as an extended array on diverse marker chromosomes. HP subclones selected at higher MTX level had more kinds of marker chromosomes. CS13*-002 isolated at 0.02 μM MTX had only one marker chromosome (m002), whereas CS13*-1.0 isolated at 1 μM MTX had five different ones (m10A, m10B, m10C, m10D, and m10E). Each marker chromosome showed different fate during long-term culture of HP subclones in the absence of MTX, resulting in different degrees of stability among the HP subclones. The m10A and m10B remained unchanged, whereas the others disappeared or evolved to variants with shortened amplified arrays. The cells containing stable marker chromosomes constituted dominant subpopulations in CS13*-1.0, and thereby CS13*-1.0 became most stable in regard to antibody production during long-term culture. Furthermore, our dual-color FISH showed that the telomeric ends of amplified arrays on the stable marker chromosomes were always surrounded by (TTAGGG)(n) sequences, indicating that (TTAGGG)(n) sequences are closely related to the stability and evolution of amplified sequences. Taken together, our data show that the assessment of genotypic stability of amplified CHO cells is a prerequisite for understanding their production stability during long-term culture in the absence of selection pressure.