Plant Evolution & Ecology (1)
Purugganan, M.D. and Fuller, D.Q. 2010. Archaeological Data Reveal Slow Rates of Evolution During Plant Domestication. Evolution 65-1: 171-183.
About 10,000 to 13, 000 years ago, humans began to cultivate food plant, which is called domestication in evolutionary biology. Domestication results from the earlier stages of wild plant food production and systematic cultivation, making crops more dependent on humans for survival but also more productive (Fuller 2007), and is one of the most important technological innovations in human history (Diamond 2002; Puruggannan and Fuller 2009). Domestication is a complex evolutionary process (Hancock 2005; Puruggannan and Fuller 2009), in which human involvement, including manipulation of the soil, vegetational environment, and cycles of harvesting-storage, exerts selection pressure on plant species and caused changes on their forms, structures and functions, and finally species diversification. Commonly, selection under domestication is believed to be strong and lead to rapid-to-very rapid evolution of domesticated species. It provided spatial isolation from wild populations, and rapid evolution became possible (de Wet 1975).
However, by definition, domestication is actually a form of interaction between animal and plant, which leads to coevolution observed in wild species under natural selection. Then, there is possibility that the rates of evolution during domestication is similar to or even lower than those in wild species. In order to measure the evolutionary rates, to estimate the degree of phenotypic changes is necessary. Phenotypic change is the evolutionary change on observable traits of organisms. And two metrics are generally used to measure the phenotypic rates of evolution: "darwin" to measure the increase of a phenotypic trait every one million years, and "haldane" to measure the change of a trait from mean situation in one generation.
According to archaeological record, two traits in particular---rachis nonshattering in cereal crops and grain/seed size in seed crops---provide quantitative information on phenotypic evolution during domestication. Shattering rachis makes seeds readily disperse naturally in wild stands when harvesting, whereas seed non-shattering plants reduce this ability, then seeds have a better chance of being gathered by people. Therefore, the evolution of nonshattering trait is regarded as a hallmark of domestication. Seed size increase in cereal crops generally occurs along with the rise of nonshattering. There is some experimental evidence to indicate that larger seeds give more vigorous seedlings than do smaller seeds and that larger seeds will germinate better after deeper planting (Heiser 1987). Thus prehistoric gardeners originally selected seeds with this trait unconsciously. So, seed size is also widely regarded as associated with human cultivation. Archaeological studies suggested that the two traits evolve rapidly during domestication.
In this study, authors compiled, analysed and compared archaeological data about evolutionary rates in both domesticated species and wild species, and then tried to address the question of whether domestication is indeed a case of rapid evolutionary diversification or has a similar pace to the evolution of wild species under natural selection.
l Authors compiled data on seed nonshattering and seed size from several archaeological studies on 11 annual crop species. Archaeological data are from 60 sites in five regions in Asia, Africa, and North America.
l Archaeological data were from primary reports with recent reviews in journal articles or archaeological monographs.
l Evolutionary rates both in darwins and in haldanes are given by general mathematical equations, and calculation is based on data for phenotypic traits of nonshattering and seed size.
l To compare evolutionary rates of domesticated species with wild species.
Nonshattering in cereal crops evolves slowly and rises to fixation over several thousand years (based on three cereal crops: einkorn wheat, rice and barley)
l Einkorn wheat in Europe-Asia took more than 3,000 years to get fixed basically in nonshattering.
l Rice in China took 300 years from about 27% to 39% , i.e. only about 12% to be fixed in nonshattering trait in a 300-year period, which means a fixation time span for nonshattering in rice more than 2,500 years.
l Barley in Near Eastern sites is the most rapid one in nonshattering evloving, but it still took about 2,700 years to get all fixed.
l High darwins for both barley and rice, and half lower for einkorn wheat; similar haldanes.
The evolution of seed size during domestication
l 18 crop species were studied. Almost all showed evolutionary increases in seed size during domestication, some earlier and some others later.
l Darwins range from about 50 to 350, and haldanes from 0.3x10-3 to 2.3x10-3
l This study model for legume species showed poor fit to the data, but more fit in a shorter domestication period (2,000 years instead of 4,000 years).
Comparison of rates for nonshattering versus grain/seed size increase
u In general, plants evolves more slowly in seed size increase than in nonshattering trait.
l darwins of evolution of nonshattering were much higher than darwins of evolution of seed size increase.
l nonshattering is mainly affected by individual loci, while seed size, including length, width, breadth and thickness, is controlled by multiple genes.
l haldanes did not reveal significantly different between nonshattering and seed size evolution.
Evolutionary rates during domestication are similar to those experienced by wild species
l The mean evolutionary rates during domestication are significantly lower than mean rates of phenotypic evolution of wild species.
l Both darwins and haldanes under domestication are significantly lower than those under natural conditions with human disturbance in the wild.
l With nonshattering and seed size increase examined separately, domestication rates still remains lower than those for species in the wild.
Time dependence of rate estimates
l With time, both darwins and haldanes decrease, i.e. a negative correlation of evolutionary rates during domestication with time.
l Fossil data appear that the negative correlation is not general, but a major trend.
Estimating selection coefficients for domestication traits
l Based on contemporary studies on genetic data, heritability (h2) for both nonshattering and seed size appear considerably high.
l Heritability values led to selection coefficients for favored phenotypes: s=2.00~3.25 x 10-3 for nonshattering; s=0.3~2.6 x 10-3 for seed size increase.
l For the exception of legume species, selection coefficients are also lower than those in wild species even if with a shorter domestication period and under the assumption of low heritabilities.
The results do not support the widespread presumption that human domestication resulted in rapid evolution in crop and live-stock species. There are several reasons: First, selection under domestication fall into two categories: conscious selection and unconscious selection. It can be imagined that most phenotypic traits were firstly selected by early farmers in a state of unconsciousness. Unconscious selection could have been responsible for most of the changes (Heiser 1987). Non-shattering means the loss of natural dispersal mechanisms. Grains with non-brittle rachis, i.e. non-shattering, have less chances to travel away from the harvested wild stands, and thus are more readily to be gathered by people. On the other hand, larger seeds give more vigorous seedlings than do smaller seeds and that larger seeds will germinate better after deeper planting (Heiser 1987). So, large size became a favored phenotypic trait under the unconscious selection in the first place. With time and with the increase of agricultural knowledge of early people, grains with large size were selected consciously. It may be said that unconscious selection is actually a form of natural selection under human activity. From this, evolutionary pace under domestication should be similar to that observed in wild species.
Secondly, it is possible that mutation genes under domestication influence multiple phenotypic traits. Domestication leads to the form of new species from wild progenitors. At the origin of new species, the population size was often highly reduced, which is termed "domestication bottleneck", and would have resulted in a loss of genetic diversity (Allaby 2008), in which deleterious mutations segregated or were eliminated. If the segregation occurred at higher frequencies, multiple phenotypic traits would be affected, which may result in decreased selection efficiency and lower evolutionary rates.
Third, it is almost certain that early farmers continued to cultivate and gather wild plants in the early stage of domestication, which necessarily led to interbreeding between domesticated species and wild species. The interbreeding caused gene flow, which could slow the fixation of favored genes by selection. As Tufto pointed out in 2009, many domesticated species, through interbreeding, spread their genes back into populations of wild relatives from which they originate. Besides, except nonshattering and seed size increase, none of other phenotypic evolution was observed in this study. However, the two traits are considered as hallmarks of domestication, whose evolutionary rates under domestication can be representative of those of other phenotypic traits, especially for seed crops, which are the most successful domesticated plant species so far.
By comparison, nonshattering is more exposed to direct interaction between human and plant, such as harvesting with human tools like sickle, while seed size increase probably originate from selection for increased seedling vigor, which is closely associated with ecological conditions under human activities. Based on the result that nonshattering appears to evolve faster, selection associated with direct human/plant interaction may be stronger than ecological selection (more non-human or unconscious?). In general, evolutionary rates between the two traits are comparable, and for each of them, evolutionary rates are also comparable across times and sites, which suggests that selection pressures may be similar in different species and human cultures.
This study has some limitations. First, archaeological data are limited, so may not provide representative enough sample sizes. Second, whether the same one time period fits to both traits or not is unknown. Finally, the authors used evolutionary rate estimates from contemporary microevolutionary studies as base comparison, which may not be appropriate. Even so, the study appears that plant domestication is a regular form of evolutionary speciation with a similar pace to those observed in wild species, which proves an appropriate model for evolutionary study.
Five sentences summary
Human domestication is an evolutionary process by artificial selection involving in the strength of human activities, and so previous assumptions commonly reveal that selection under domestication is strong, and leads to more rapid evolution of cultivated species than that of wild species under natural selection. In this study, authors tried to make clear that whether domestication is indeed a case of rapid evolutionary diversification or has a similar evolutionary pace to cases of natural selection. The authors compiled and re-calculated archaeological data about 11 crop species from 60 sites, and introduced two important metrics of evolutionary rates to measure evolutionary efficiency of two phenotypic traits (nonshattering and seed size increase). The results indicate slow rates of evolution during domestication. The authors concluded that domestication may be driven by unconscious artificial selection, which is a similar evolutionary phenomenon to that in wild species by natural selection, while some genetic changes during domestication may also help the slow pace.
Allaby, R.G. 2008. The rise of plant domestication: life in the slow lane. Biologist 55:
De Wet, J.M.J. 1975. Evolutionary dynamics of cereal domestication. Bulletin of The
Torrey Botanical Club 102: 307-312.
Fuller, D.Q. 2007. Contrasting patterns in crop domestication and domestication rates:
Recent archaeobotanical insights from the old world. Annals of Botany 100: 903-
Heiser, C.B. 1987. Aspects of unconscious selection and the evolution of domesticat-
ed plant. Euphytica 37: 77-81.
Purugganan, M.D. and Fuller, D.Q. 2009. The nature of selection during plant domes-
tication. Nature 457: 843-848.
Tufto, J. 2009. Gene flow from domesticated species to wild relatives: migration load
in a model of multivariate selection. Evolution 64-1: 180-192.