Notes on Segregation Distortion, Novel Traits, and Transgressive Segregation 1902-2002:
While Thinking About the Alfalfa Genome

Edwin T. Bingham
University of Wisconsin-Madison 53706

Prologue: This report is a work in progress that has changed my thinking about the alfalfa genome, its relationship to the genomes of the crossable perennial species, and about speciation in Medicago. Until I reviewed the literature cited in this report, I was frustrated about many things that I did not understand last century. Now, everything seems to be falling into place. The alfalfa genome is beginning to yield its secrets! Moreover, some new experiments using existing stocks to test new hypotheses, should put capstones on some important genomic relationships.


Segregation distortion (SD) is a common feature of papers on mapping alfalfa using molecular markers. This will be discussed in relation to the genome after considering the general area of segregation distortion in plants. Over the years, SD has been termed selective fertilization, unequal segregation, altered, irregular, and blurred segregation. It is discussed in terms of meiotic drive and even outlaw genes: and, novel variation in some interspecific hybrids is often associated with SD. I started reading the most recent literature, and worked back in time. This was a fascinating, but unsettling experience as I discovered and rediscovered literature that I had missed over the years or that I had not considered relevant to my alfalfa breeding and genetics research.

Early Reports of Segregation Distortion

These notes will begin with the earliest reports of SD, which so happen to coincide with the rediscovery of Mendelian genetics. A wealth of information was found in a delightful little book by Donald F. Jones entitled "Selective Fertilization" (1928). I thank Jerry Kermicle for calling the book to my attention. Jones (1928) pointed out that in the first detailed investigations of heredity following Mendel, Correns found that certain crosses of maize did not have the expected proportion of one-fourth recessives. Correns reported this in 1902 based on results from a cross of pointed popcorn with a sweet, wrinkled variety. In a large number of trials the sugary factor pair (Susu) usually segregated for 25% recessives except for one situation where Correns observed 16% recessives. This deficiency of recessives was confirmed by Jones and by other prominent researchers including Lock (1906) and East and Hayes (1911) (see Jones 1928 for refs). Not only did Jones confirm Correns, he also observed exactly 16% recessives based on 3,681 seeds classified for sugary. In the full study, Jones compared five situations involving combinations of heterozygous and homozygous sugary, and found that there was no deficiency in the percentage of recessives, i.e. no differential fertilization in those matings in which the male gametes were all alike. A selective action took place only when segregating pollen was used. Hence, the selective fertilization took place only in certain cross combinations and only in segregating pollen. Jones reviewed other examples of deficiency of recessives in maize, Melandrium, Lychnis Rumex, Datura and Qenothera Jones' general conclusion was that the transmission of heritable characters is dependent on many things:

Jones then went on to review and discuss the last point in terms of differences in the rate of pollen tube growth. His conclusion was that most of the deviations from expected ratios in his work were due to pollen tube factors.

The area of gametophytic factors in maize that affect fertilization is extensive and was reviewed by Oliver Nelson in 1994. Nelson reviews much research by Mangelsdorf, Jones, Emerson, Demerec, Schwartz and himself (see Nelson for refs.). Kermicle and Allen (1990) reported an extensive study of cross incompatibility between maize and teosinte, and discussed how the Gal -s system could function as a barrier to pollination. Recently, Kermicle has proposed utilizing the strategy to prevent genetically modified maize from crossing with normal maize. Thus, selective fertilization is just as relevant today as when Jones published his book in 1928. The SD in maize is controlled by genes already in place in the maize genome. In much of the review that follows, there is both this type of SD, but also SD that results from crossing over involving chromosomes that have various degrees of divergence.

Segregation of Novel Traits

An early study by Lotsy (1915, 1916) was cited by Stebbins (1950), Rick and Smith (1953) and Grant (1975) (all of whom will be reviewed later). Lotsy found new traits that segregated in the F2 of a hybrid of Antirrhinum glutinosum and A. majus The Lotsy references are available in the U.W. libraries. Both Antirrhinum species have the typical flower with a prominent two-lipped corolla limb. Some individuals in the F2 generation of the interspecific cross had tubular corollas with reduced limbs, as in the related genus Rhinanthus. This unusual flower type occurred sporadically in the F2 and Lotsy stabilized it in some F4 lines. Lotsy was confident that this was new variation and in fact wrote a book entitled: "Evolution by Means of Hybridization" (1916). His study remains controversial, however, because he did not provide evidence that the mutation was not carried by one of the parents.

Another study involving different strains of the same two species went to great lengths to show that a new variant had arisen in the F2 of the species hybrid, and was not hidden by heterozygosity in one of the parents. This research was by Kenneth Mather at Birmingham, England. His initial study involved making the species crosses between A. majus and A. glutinosum (Mather, 1947). Mather could find no mechanism isolating the two species once pollen had been successfully transferred to a stigma of the opposite type. Pollination by artificial means was easy between these species as within them. Furthermore, when stigmata of either species were pollinated simultaneously with pollen from the two, hybrids were produced as commonly as maternal types. (The discussion could just as well be about Medicago sativa and M. falcata. Mather's observations of bee behavior brought him to the conclusion that the mechanism isolating A. majus and A. glutinosum was to be found in bee preference for one type or the other. (Does this remind you of alfalfa, or what?)

In a follow-up to the original A. majus by A. glutinosum hybrids, Mather and Vines (1951) report discovering cleistogamy in derivatives of the hybrid. They studied large sample sizes of parent and hybrid populations, used statistics appropriate for Mather, and concluded that cleistogamy had not been seen in the parental species nor in the Fl generation. They suggested that 2 -3 genes of supplementary (complementary) effect are necessary for cleistogamous flowers. They do not attempt to suggest a mechanism for the origin of the novel variation; but, Grant 1975 suggests that the new variation seen by Mather and Vines represents "macrorecombination". Mather and Vines state: "Thus the two parental species, one an obligatory cross-breeder and the other regularly setting a considerable fraction of its seed by cross-pollination, contain between them all the genetical materials necessary for the production of an inbreeding type of plant, i.e. a type of a plant with a breeding system distinct from, and even opposed to those of its parents."

Stebbins (1950) has a section on page 279 that discusses hybridization and the origin of new types. Stebbins discusses how hybridization of species mainly produces convergence between previously distinct species. He goes on to point out that there is evidence that in some instances hybridization can result in the appearance of types which are actually new. The review of this evidence was thrilling to me. The original references were consulted whenever possible. Stebbins writes on page 285, "Some evidence at hand suggests that the recombination of genetic factors in the offspring of interspecific hybrids may sometimes lead to new types radically different from those found in either parent." He reviews Lotsy, 1915, indicating that the paper contains some striking examples of new traits in the progeny of Antirrhinum glutinosum crossed with a peloric form of A. majus. From Stebbins, I learned about Hagedoorn and Hagedoorn (1921) who cite the example of Vilmorin's hybrid between Argemone mexicana and A. platyceras, from which several strongly aberrant types segregated in the F2 generation.

Some of these were different from either parent in such fundamental characteristics as the number of sepals or of carpels. Similar aberrant types appeared in the F1 hybrids of Paeonia lactiflora ("P. albiflora") and various members of the complex of P. anomala (Saunders and Stebbins 1938). Less extreme new types are reported by Clausen (1926) as segregates from hybrids between Viola arvensis and V. tricolor.

At this point I realized that we had not understood the significance of the black seed trait that segregated in backcross generations of our haploid X M. falcata materials. The black seed trait is certainly novel because it does not occur in either M. sativa or M. falcata. It is complex in inheritance (Kimbeng and Bingham 1997) which is evidence of variation resulting from recombination, in my view, currently. To our knowledge, the only other Medicago species where black seeds occur is M. intertexta. There is some kind of an evolutionary statement being made in the segregation of black seeds in advanced hybrid generations of M. sativa and falcata. It is my recollection that the black seed Don Barnes worked with goes back to an old M. varia population, which descended from a hybrid of M. sativa and M. falcata.

Now, we turn to another book that was a rich source of references on novel traits: Grant (1975) Genetics of Flowering Plants. In his chapter on gene interaction, Grant has a section titled: Macrorecombinations. Grant defines macrorecombinations as: "Morphological characteristics unlike those found in either parent," and indicates that he coined the term (Grant 1956. He indicates that macrorecombinations occasionally appear in the progeny of interspecific hybrids.. Grant further states that, the phenomenon has long been known, and starts his review with Lotsy (1916). Lotsy 1916 is a follow-up to the 1915 paper reviewed earlier in this section.

Grant (1956) also reported on his work with Gilia, involving a case where new variation emerged in the F3 generation of a cross between G. achilleaefolia and G. millefoliata. The variants had altered flower morphology and variant frequency was greater in F4 families derived from single-plant selections. Other examples of what Grant terms macrorecombinations can be found in advanced generations of interspecific hybrids of Gossypium, Geum, Lycopersicon and Layia, as reviewed by Grant (1975).

Variants in Lycopersicon were reported by Rick and Smith (1953) in a paper titled "Novel Variation in Tomato Species Hybrids". This paper was pivotal in my thinking about variation that I have seen in advanced hybrid generations of Medicago sativa and M. falcata at both diploid and tetraploid levels. The crosses used L. esculentum as the female, L. peruvianum as the male, and involved different parents of each species in two different hybrids. L. esculentum is the familiar tomato and peruvianum has small green fruits, several forms of resistance, and is high in vitamin C.

Rick and Smith's Hybrid No. 1, as they referred to it in the paper, was between L. esculentum var. Michigan State "Forcing" and L. peruvianum var. dentatum Pl128657. It was a classic interspecific hybrid in that the hybrid was difficult to make and it was difficult to obtain the F2 by selfing. The first two variants to be discussed appeared among 46 F2 plants. The authors discussed the fact that the F2 segregated for such a wide array of recombinations of parental characters that no two individuals were alike. What they considered variants were entirely outside the normal range of segregation. The first variant was termed "entire leaf'; other features accompanying the leaf modification were reduction in vigor and slight asymmetry of the flower. (The description makes Bingham think of the cauliflower head-simple leaf trait in alfalfa, also known currently as uni, Brouwer and Osborn, 1996).

The second variant in Hybrid No. 1 was "compound inflorescence". Inflorescences are so greatly subdivided that each one bears several hundred flowers. This compares to tomato with 3-9 flowers, L. peruvianum with 15-40, and the Fl hybrid with 8-20 flowers. The authors point out that proliferations of flowers to this degree were never seen in any of their lines, and it is certain that this gene was not present in the L. esculentum. The L. peruvianum pare

Mutation. The possibility of increased mutability in species hybrids.
Complementary action of genes of the parent species. This will be discussed again later.
Recessive genes derived from the self-incompatible parent. Recessive traits seldom or never seen because of enforced out crossing.

Additional mechanisms that might be mentioned currently include:

Activation of transposable elements.
Altered patterns of methylation.
Cytogenetic divergence. Translocations, inversions and/or what Stebbins (1950) and Stephens (1950) term “cryptic structural difference”.

Notice that some of these mechanisms also could produce SD.

[More, see PDF]

Literature Cited
Citations include only essential information.