The nuclear cytoplasmic interaction (NCI) hypothesis of genome evolution and speciation in plants states that newly formed allopolyploids pass through a bottleneck of sterility and the fertile plants that emerge are fixed for species-specific chromosome translocations. These translocations restore fertility and reduce negative effects of the maternal cytoplasm on an alien paternal genome. Using fluorescent in situ hybridization and genomic in situ hybridization and by reviewing published data, we test the NCI hypothesis using three natural Nicotiana allotetraploids (all 2n = 4x = 48, N. arentsii, N. rustica and several genotypes, including a feral plant and cultivars, of N. tabacum (tobacco)). We compare these data with three synthetic tobacco plants (Th37) that are F3 descendent progeny of an allotetraploid formed from ♀N. sylvestris (2n = 24) ×♂N. tomentosiformis (2n = 24). No intergenomic translocations were observed in N. arentsii and N. rustica. An analysis of subtelomeric tandem repeats in these allotetraploids and their putative parents shows minimal genetic changes; those that do occur may reflect evolution in the diploids or the polyploids subsequent to allopolyploidy. All natural N. tabacum genotypes have intergenomic translocations. This may reflect a large ‘genomic-shock’ generated by allopolyploidy involving widely diverged parental species. Two of three synthetic tobacco plants had a translocation similar to that found in all cultivars of tobacco. This translocation may be significant in tobacco fertility and may have been fixed early in tobacco's evolution. But it is lacking in the feral tobacco, which might indicate a polyphyletic origin or early divergence from all cultivars examined. Overall, only in tobacco is there any evidence that NCI may have influenced genome evolution, and here further data are required to verify chromosome identity. © 2004 The Linnean Society of London, Biological Journal of the Linnean Society, 2004, 82, 599–606.