POLYPLOIDY AND DIPLOIDY: NEW PERSPECTIVES ON CHROMOSOME PAIRING AND ITS EVOLUTIONARY IMPLICATIONS

The last widely accepted classification of polyploidy recognized three major categories: autoploidy, segmental alloploidy, and true alloploidy. Criteria for recognizing these types were based largely on chromosome pairing in F 1 hybrids, but until very recently there were no quantitative criteria to distinguish autoploids from segmental alloploids. This is theoretically possible, and Jackson and Casey (1982) and Jackson and Hauber (1982) have presented models and methods to quantitatively predict the frequencies of the different meiotic configurations in autotriploids through autooctoploids. The expected numbers of configurations then can be statistically tested for goodness of fit with observed data. Synthetic and natural autotetraploids have been shown to fit the models well. However, the exceptions that do not fit are of particular interest because they indicate the presence of Ph ‐like (Pairing homoeologous) genes that cause fewer multivalents and more bivalents than expected in the autoploid models. The effect of Ph ‐like genes at the nuclear level apparently involves placement of chromosome attachment sites on the nuclear membrane such that different genomes occupy slightly different sites. This would be far enough apart in true alloploids so that no pairing occurred in the original F 1 hybrid. In true autoploids and normal diploids, homologues are equally close so that pairing obeys probability laws that make it possible to describe expected pairing events with appropriate models and equations. From the foregoing statements, one can deduce that the terms autoploid, segmental alloploid, and true alloploid need not and indeed should not be equated with results of crosses between any taxonomic levels. Rather, they represent conditions that can be attained by one or very few genes acting on nuclear membrane attachment sites. It is possible to deduce from hybrids at the diploid level the type of polyploid that can be obtained, and such data can have great agronomic value. A generally held dogma is that the inability of chromosomes to pair in F 1 hybrids is due to considerable genetic divergence and perhaps numerous structural changes at the light microscope and/or ultrastructural level. Data from various sources show that this need not be true, and there is no reason to believe this is a factor in what has been called genome divergence in the classical sense. Genome divergence in terms of pairing ability can occur at the population level due to one or a few genes.