For chromatin lesions to interact to form exchanges of any sort, it is obvious that contact between them must be made. However, the probability of such interaction is conditioned by other factors like time, initial separation, metabolic activity, and, in the case of chemically induced lesions, scheduled DNA synthesis. The irradiated nucleus was, for a long time, regarded as a "bag of broken chromosomes" with the severed ends free to move around and find partners with which to form illegitimate reunions. Many of these would be seen at following metaphase as intra- and interchanges. Evidence is rapidly accumulating which indicates that this picture of the nucleus is false. We know now that chromosomes occupy highly localised domains with limited movement, and that there is no massive intermingling; that much of the chromatin is compacted and splinted with proteins and so precluded from exchange-type contact; that most of the chromatin is looped and "fixed" into an intra-nuclear protein scaffold or skeleton; that some chromatin is spun-out and associated with the nuclear envelope in the vicinity of the pore-complexes. Thus it would appear that movement, in the sense envisaged by early workers, is curtailed, and that only a proportion (probably a small proportion) of the chromatin is actually "at-risk" with respect to interchange formation. Where then does interchange take place? Are the "sites" pre-existent, or can proximity requirements be realised after radiation exposure? In what ways will the intra-nuclear architecture influence exchange? These are some of the questions which are considered in this paper.