Among grasses, sugarcane is probably the crop that will most benefit from comparative genome analysis. Modern sugarcane cultivars display a considerable complexity related to polyploidy, aneuploidy and multispecific origin. Application of molecular techniques on a particular variety (R570, 2n=ca 115) improved our basic knowledge on this genome structure. Two genome origins coexist, that are derived from the ancestral species Saccharum officinarum and S. spontaneum and can be differentiated by GISH. They represent about 85% and 15% of the genome in R570, respectively. Their likely basic numbers are x=10 and x=8, as indicated by FISH with rDNA probes. RFLP mapping supports general collinearity between the two basic genomes. The S. officinarum part is poorly covered, whereas the S. spontaneum part of the genome of R570 is easily accessible to marking due to a lower representation (15% ) and a higher polymorphism. S. spontaneum-derived chromosomes tend to pair preferentially with each other in some homo(eo)logy groups, but not all. Several cases of interspecific cross-over were uncovered. A tentative consensus map enabled us to demonstrate genome collinearity between sugarcane and related crops such as sorghum and maize. Information obtained in other crops will be essential for targetting mapping efforts to genome regions of potential interest in this crop, where tagging all (homoeo)alleles at a particular locus requires considerable resolution.