Abstract
Data from the literature (tables 1, 2; fig. 1) show that the relative probabilities of occurrence for the various categories of spontaneous chromosomal mutations do not match those predicted by both the random-contact-and-exchange model (basically equivalent to the exchange theory of Revell) and the random-breakage-and-reunion model (i.e., the breakage-and-reunion theory of Sax or of Muller). These models do not take into account recent findings that, during the meiotic prophase and especially at pachytene, eukaryote chromosomes are attached by each end to the nuclear membrane, leading to a configuration we term the "suspension-arch structure" (fig. 3). We recalculated the relative probabilities of the occurrence of spontaneous chromosomal mutations given the suspension-arch structure and assuming that these rearrangements arise from errors in the resolution of interlockings between bivalents and a special type of crossover that we call the "hetero-site" crossover (figs. 2, 3, 6, 7). From these calculations we found that the relative probability of the occurrence of reciprocal translocations (the most fitness-damaging rearrangement) declines with increases in chromosome number and in nuclear volume (fig. 5). We also found that paracentric inversions occur increasingly more often than pericentric ones as the centromere position becomes more terminal and the distance between attachments to the nuclear membrane becomes greater (this distance increases as nuclear volume increases; fig. 8). These results are in accord with the cytogenetic data from Drosophila, humans, and ants. A puzzling phenomenon is that the relative rates of radiation-induced chromosomal mutations differ greatly from those calculated under the suspension-archstructure model and from rates of spontaneous chromosomal mutations. We propose the testable hypotheses that (1) most spontaneous chromosomal mutations occur in synaptonemal complexes and involve crossovers and errors in the resolution of interlockings, and (2) radiation and chemical mutagens allow rearrangements when chromosomes intersect at any stage (fig. 10). These considerations lead to the "minimum-interaction hypothesis," which states that karyotype evolution has been in large part shaped by selection to reduce the occurrence of such fitness-reducing spontaneous chromosomal mutations as reciprocal translocations. Some of the response to this selection is the result of the improvement of DNA-repair mechanisms, increased contraction of the chromosomes caused by higher-order helix formation, and the development of sex-chromosome heteropycnosis. We concentrate on examining two further, interacting responses in the light of the hypothesis. One of these is increase in nuclear volume (fig. 9), but if the ratio of genome size to nuclear volume is high, then an increase in chromosome number, caused by such factors as centric fission, is adaptive because it reduces the occurrence of reciprocal translocation. Although chromosome number can be reduced by centric fusion, such instances seem to be "back eddies" in the mainstream of karyotype evolution.