The Redomestication of Maize
John Doebley
Laboratory of Genetics
University of Wisconsin-Madison

How long did it take ancient peoples to domesticate maize from teosinte? We don't know the answer to this question and will likely never have a very precise answer, but there are ways we can make an educated guess to this and related questions. With this thought in mind, I began a "long term" selection study with teosinte (Zea mays ssp. parviglumis) to see if I can change the population to be more maize-like, i.e. to "redomesticate" maize. This study involves growing a large number of teosinte plants each year and harvesting seed from the most maize-like individuals for the next generation. I hope to continue this process for 30 generations. Here, I report how I am going about this, my future plans, and results to date.

The starting population: I used teosinte seeds that were from 1 mile South of Palo Blanco, Guerrero, Mexico (Beadle and Kato Site 4). Teosinte from this area is the least maize-like of the Mexican annual teosintes and it only rarely hybridizes with maize (Wilkes 1967, 1977). Being free of maize "contamination" is important for my selection experiment because I don't want to merely filter out maize genes that were introgressed into teosinte, but to select natural teosinte variants and thereby move the phenotype in the direction of maize.

Field site: My plan is to grow each generation on Molokai Island, Hawaii in an isolation plot at Hawaiian Research. Three thousand seeds will be or have been planted in September of each year. The plot will be about 9000 square feet (0.084 hectares) in size. The plants will be irrigated and the plot treated with a pre-emergence herbicide prior to planting. I will be able to harvest seed and make measurements on the plants during the following January. The isolation plot is located away from other corn fields to prevent contamination with maize pollen. If maize pollen does contaminate my teosinte plot, this will be obvious the following season since maize-teosinte hybrids are readily apparent.

Selection scheme: Maize differs from teosinte by a suite of traits including seed size, seed number per ear, ear disarticulation, inflorescence phyllotaxy, spikelet abortion, and the length of the lateral branches (Doebley et al. 1990). Since it would not be practical to apply selection to all traits simultaneously, I chose to apply selection to the length of the lateral branches since this trait is easily scored, the population contained much variation for the trait, and a cloned gene (teosinte branched1 or tb1) is known to affect this trait (Doebley, Stec and Hubbard 1997; Wang et al. 1999). In the future, I'll be able to examine whether the population is responding to selection by changes in tb1 allele frequencies as occurred many thousands of years ago when ancient Mexicans first domesticated maize.

Each generation, I will select plants with the shortest lateral branches and harvest seed form these plants for the subsequent generation. I plan to harvest seed from about 125 selected plants so that if all 3000 seeds germinate and develop, I will be applying a selection intensity of about 4%. I will measure and harvest seed from a similar number of random plants as controls. From the selected and random plants, I will measure branch length, tiller number, fruitcase weight and plant height.

Plant architectural types: During the first season, I noticed that there were distinct types of plant architectures in the population. While most plants had long branches tipped by fully formed tassels (left photo), a few plants had short branches tipped by ears (right photo). There were also some intermediate types between these two extremes (center-left photo). [This sort of polymorphism for teosinte populations has been previously described (Iltis 1987).] During the second season, I also noticed a substantial number of plants with mixed tassel-ear inflorescences on the tips of their lateral branches (center-right photo). As a result of the selection process, I expect to increase the frequency of the plants with short branches tipped by ears and tassel-ears.

Progress to date: Thus far, I have completed three growing seasons or two cycles of selection.  At this point, the data are only suggestive, however I show them in the table below. 

Branch length is reported as "relative branch length" or the proportion of the height of the plants to which the length of the branches are equivalent.  Relative branch length appears to be declining as expected, however the change is not statistically significant.  Plant height, which is not under direct selection, varies across cycles, probably as a result of environmental differences (e.g. plant density) between seasons (see below).  Seed weight increased considerably by cycle 2, and this change appears to be significant. This change could result from either environmental differences between cycles or the effects of selection.  If selection is causal, it may be either as a correlated response to selection on branch length, or more likely, inadvertent selection.  Teosinte seeds possess factors that inhibit fresh seed from germinating (Beadle 1977).  These factors likely explain why only a fraction of the planted seeds grow into adult plants.  If germination inhibition is correlated with seed size, then larger seeds may be more likely to germinate and contribute to the next generation.  Tiller number fluctuated across cycles with no trend, except that there is a correlation with the number of plants in the field (plant density), which suggests an environmental effect.  The percentage of plants with the different branch morphologies as described above appears to be changing.  Initially, only 1.6% of plants had short female tipped branches, but that value is 12.8% in cycle 2.  The percentage of plants with fully formed long branches has declined considerably.  Whether these changes are a result of selection or environmental differences across cycles remains to be determined.

The environmental differences from year-to-year are considerable.  For cycle 0, there was reasonably good germination and the plants were at a relatively high density.  For cycle 1, germination was poor and plant density was sparse.  Moreover, during cycle 1, many plants appeared to be damaged by herbicide.  For cycle 2, the grower changed from a 36" row spacing to a 30" spacing, increasing planting density.  Also, the project was moved to a different field several miles from the location for cycles 0 and 1.  Whether any selection gain is being made will only be determined once residual seeds for the different cycles are grown in a common garden experiment.

The table below gives some basic statistics on the plants selected each cycle to form the next generation.  If the environments were equivalent from year-to-year, the data in this table and the one above could be used to estimate heritability.  However, given the environmental differences this is inappropriate.  The numbers in the table below do provide a rough guide as to how far I am trying to move the phenotype each cycle.

Plans for the future: Over the next 28 years (the funding gods agreeable), I will grow out about 3000 seed of the selected progeny from the preceding generation and select the short-branch plants.  I will use the data to estimate heritability measured both as realized heritability and as the correlation among half-sibs. I'll be able to measure heritability and the amount of additive genetic variance in this population for branch length and other domestication traits. Knowing these values will help to answer the questions posed above and a few others as well.  I'll also be able to look at how the frequencies of alleles at relevant loci (e.g. tb1) have changed. Each year, I will plot the change in phenotype on the charts above.

I don't anticipate that this selection experiment will actually produce a maize replica. The plants should become more maize-like for branch length (and perhaps seed size and tillering), but for most other traits they should remain true to the teosinte condition. There is a chance that some other correlated traits such as the number of fruitcases per ear may change (Doebley et al. 1995), so I'll be keeping an eye on these traits as well.

Literature Cited

1. Beadle, G. W., 1977  The origin of Zea mays, pp. 615-635 in Origins of Agriculture, edited by C. E. Reed. Mouton, The Hague.
2. Beadle, G. W., 1978  Teosinte and the origin of maize, pp. 113-128 in Maize Breeding and Genetics, edited by D. B. Walden. John Wiley & Sons, New York, NY.
3. Doebley, J., A. Stec and C. Gustus, 1995  teosinte branched1 and the origin of maize: evidence for epistasis and the evolution of dominance. Genetics 141: 333-346.
4. Doebley, J., A. Stec and L. Hubbard, 1997  The evolution of apical dominance in maize. Nature 386: 485-488.
5. Doebley, J. F., A. Stec, J. Wendel and M. Edwards, 1990  Genetic and morphological analysis of a maize-teosinte F2 population: implications for the origin of maize. Proc. Natl. Acad. Sci. 87: 9888-9892.
6. Iltis, H., 1987  Maize evolution and agricultural origins, pp. 195-213 in Grass systematics and evolution, edited by T. Soderstrom, K. Hilu, C. Campbell, and M. Barkworth. Smithsonian Inst. Press, Washington, D. C.
7. Wang, R.-L., A. Stec, J. Hey, L. Lukens and J. Doebley, 1999  The limits of selection during maize domestication. Nature 398: 236-239.
8. Wilkes, H. G., 1967  Teosinte: the closest relative of maize. The Bussey Institute, Harvard University, Cambridge.
9. Wilkes, H. G., 1977  Hybridization of maize and teosinte in Mexico and Guatemala and the improvement of maize. Econ. Bot. 31: 254-293.

Home

Top of Page