The role of polyploidy in the speciation of flowering plants

Lecture © Marc A. McPherson
BIOL 606 Session, University of Alberta, January 26, 2000

Polyploidy was first discovered by Winkler in 1916 during his observations of a spontaneous autopolyploid induced by mechanically damaged tissue (Grant, 1971). Winge (1917) proposed an explanation for polyploidy using an arithmetic series he had observed in Chrysanthemum (2n=18, 36, 54, 72, and 90) and Chenopodium (2n=18, 36) (Grant, 1971). Winge hypothesized that polyploidy occurred by successive increases in the original somatic chromosome number. The most comprehensive work on polyploidy and its relationship to plant evolution is the book "Plant Speciation" by Vern Grant (1981). It has been estimated that 30% to 70% of angiosperms are polyploids (Grant, 1971). The broad range in estimates for angiosperm polyploidy is due to a lack of knowledge about whether these plants are ancient polyploids, or whether more recent events have caused the polyploidy we observe today. Furthermore, no one really knows the true base number of the angiosperms, making estimates of polyploidy for this large and diverse group difficult (Grant, 1971; personal communications Sean Graham, 2000).

The mechanisms known to cause polyploidy include somatic doubling (endopolyploidy), polyspermy, and unreduced gametes (Grant, 1971; Ramsey and Schemske, 1998). There are several environmental and biological phenomena associated with polyploidy. Polyploidy is more prevalent at high altitudes, high latitudes and in areas of recent glaciation. Perennials are more frequently polyploids than are annuals. Frequency of polyploidy is also correlated with plants that have the ability to self-fertilize and / or enter into interspecific crosses (Grant, 1971; Ramsey and Schemske, 1998).

Allopolyploidy is the doubling of the somatic chromosome number of an interspecific hybrid plant; in contrast, autopolyploidy is the doubling of the same chromosome set. Various authors have argued about the precise usage of these terms. The arguments about these definitions involve the different methods of delineating a species. Thus, the frequency of autopolyploidy relative to allopolyploidy differs dramatically depending on the author. Grant (1971) suggested that the usage of these definitions was too narrow and that in nature a gradient between the extremes of allopolyploidy and autopolyploidy exist.

Ramsey and Schemske (1998) take a broad view of a species by using the biological species concept, and for this reason these authors suggest that autopolyploidy occurs more often than does allopolyploidy. This view disagrees with most of the work published to date (Grant, 1971; Soltis, 1993). Ramsey and Schemske propose triploids to be a major mechanism to facilitate the formation of autopolyploidy. They discuss the possible reasons for inviability of triploids (triploid block) and how these maybe overcome by plants to produce viable polyploids. One possible explanation for the triploid block may be the ratio of the ploidy level between the embryo and the endosperm. This may explain why plants in several families, which lack endosperm in their mature seeds, have a higher frequency of polyploidy than other taxa. Ramsey and Schemske (1998) use various estimates from the literature and convoluted arguments in an attempt to establish their hypothesis that triploid plants may act as a bridge to the formation of tetraploids and various other polyploid derivatives. In conclusion, the literature pertaining to the subject of polyploidy is voluminous, but our knowledge about the mechanisms involved in polyploid formation and establishment remains enigmatic.

Grant, V. 1971. Plant Speciation. Columbia Press, New York, N. Y.
Ramsey. J. & Schemske, D. W. 1998. Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annu. Rev. Ecol. Syst. 29: 467-501.
Soltis, D. E. & Soltis, P. S. 1993. Molecular data and the dynamic nature of polyploidy. Critical Reviews in Plant Science. 12 (3): 243-273.


Rapporteur: Rich Palmer

From the start of the discussion, it was clear most people (including, it would appear, the authors of the focal paper) seemed to have difficulty articulating what the central questions of biological interest relating to polyploidy in plants really were. The general sentiment of the discussion was that the focal paper assembled a significant amount of information, but never really succeeded in arriving at any synthesis, other than some plausible alternatives for how it might arise (Fig. 5). Unfortunately, these alternative models were developed based mostly on information from crop plants, so it's unclear how relevant they are in 'natural' populations. The web appendices to the paper, though, where all the primary information was collated, were seen as a very valuable contribution.

Polyploidy isn't unique to plants, as it is common in some insects, fish, amphibians, and reptiles. If polyploidy is actually rarer in both birds and mammals than other vertebrates, this suggests some connection to homeothermy.

One of the more intriguing observations that arose during the discussion was that taxa derived from polypoid ancestors tend, over time, to become diploidized as duplicate sets of chromosomes begin to diverge independently.

Several questions prompted some interesting discussion:

1) Why are hybrids more vulnerable to polyploidy?

One likely explanation is that chromosomes don't pair up properly during meiosis, so this generates odd numbers of chromosomes in the gametes.

2) Why were no phylogenetic character mapping studies done to reconstruct patterns of ploidy evolution?

A big problem lies with actually 'coding' ploidy, since, without a lot of work, it is quite difficult to determine which chromosome arms are associated with which following fusions or fissions. Also, ploidy levels are so variable within most plant families, that a more finely resolved phylogeny would be required.

3) Is the apparent high level of polyploidy in flowering plants possibly due to 'under-reporting' of plant taxa with normal ploidy levels?

The general sentiment seemed to be no, since ploidy levels are such a common descriptive element of alpha-taxonomic descriptions.

4) How would a study of the evolution of polyploidy actually have to be done to understand the mechanisms involved?

The only way to do it convincingly would be to find a suite of populations within a single species, perhaps along an environmental gradient, that illustrated variations or stages on the route to polyploidy. More detailed reviews of the literature would probably be fruitless since insufficient geographic and temporal information would be available for any particular case to actually reconstruct it properly. In other words, the basic information is probably not even available in the literature. Direct and detailed study of a lineage in transition would be required.