Weedy rice infests rice fields worldwide. It is a particularly insidious weed due to its similarity to the domesticated varieties. Many of the easiest ways to kill weedy rice are likely to also harm the crop, so weed management is problematic. Moreover, if even a small fraction of the weedy plants survive and reproduce, weedy rice is so productive that it can spread and cause major economic damage (Ferrero 2003). There are some barriers to gene flow between domesticated Oryza sativa and its wild ancestor, Oryza rufipogon and other species of weedy rice, but gene flow can still occur (Chu & Oka 1970; Chen et al. 2004). Thus, one factor that may make this weed particularly problematic is that improvements introduced into domesticated rice can also spread into its weedy relatives (Langevin et al. 1990). Adding to this difficult situation, there are numerous different types of weedy rice, with very different morphologies (Fig. 1), and possibly different origins (Vaughan et al. 2001).
These different morphologies have lead to much speculation about the source of these weeds. There are three main hypotheses, each of which was considered by Londo & Schaal (2007) as follows: (i) weedy rice may have evolved from wild rice, which adapted to take advantage of cultivated rice habitats; (ii) weedy rice may have originated from escaped domesticated rice seeds, which then evolved weedy traits; or (iii) interbreeding between cultivated and wild rice may have lead to viable hybrids, which could have combined traits from both to form weedy forms. Using 29 weedy, 50 domesticated, and 63 wild rice accessions, including the two closest relatives to domesticated rice, O. rufipogon and Oryza nivara, Londo and Schaal investigated these hypotheses. They genotyped each individual at 16 microsatellite loci, and used the resulting data to analyse population clustering and hybridization. They also sequenced a nuclear pseudogene, which was useful for phylogeographical analyses and for distinguishing between the two domesticated subspecies O. sativa japonica and O. sativa indica.
Londo and Schaal's results showed that weedy rice has likely evolved multiple times, and in several different ways. A weedy rice population from California appears to be mainly derived from wild O. rufipogon, thus supporting the hypothesis that weeds evolved from wild rice relatives. However, another set of populations cluster with the ‘Aus’ variety of domesticated O. sativa indica grown in India and Bangladesh, and show no or very limited recent gene flow from wild O. rufipogon. From this evidence, it appears that these populations evolved mainly from escaped domesticated plants, consistent with the second hypothesis. Still another major lineage of weedy rice appears to have been derived from hybridization between domesticated O. sativa indica and wild O. rufipogon accessions in Thailand or Myanmar, and then introduced into the USA. This lineage provides support for the third, or ‘hybrid origin’, hypothesis. The few remaining populations have even more complex histories, involving hybridization between the main lineages of weedy rice, or between one of the weedy lineages and some locally cultivated domesticated rice. Thus, it turns out that all three proposed scenarios are likely to have occurred in at least one of the varieties.
Although the weedy varieties differ in some traits, such as seed colour, the multiple independent origins share some common phenotypes, including rapid growth, aggressive tillering, prolonged reproduction, seed dispersal at maturity and seed dormancy. Additionally, although the wild rice species most similar to weedy rice is the perennial, O. rufipogon, not the annual O. nivara, all of the weedy forms have an annual life history. These patterns raise several interesting questions for future studies involving the relative roles of selection and contingency during the parallel evolution of these weedy forms.
The fact that so many of the weedy rice accessions clearly originated in Asia emphasizes the importance of careful screening for seed contaminants when introducing new crop varieties, as this handful of introductions costs farmers millions of dollars annually in losses. Also of interest is the fact that several different types of hybridization played a major role in the history of the weedy populations examined. Introgression from domesticated rice may allow the weeds to gain advantageous traits, such as herbicide resistance from imidazolinone-tolerant strains now widely cultivated (Gealy 2005), or seed colour or size that mimic those of domesticated varieties, enabling weedy rice to escape screening (Ellstrand et al. 1999). The high potential for introgression into weedy varieties should be carefully considered when new transgenic genotypes of rice are generated, as problematic transgenes can escape even in early experimental stages of new crop production in some cases (e.g. Reichman et al. 2006). Gene flow between different weedy varieties may allow these advantageous traits to spread and combine into potentially more problematic phenotypes. Additionally, gene flow from wild rice may be important for introducing new traits into weedy rice. These studies highlight the importance of hybridization as a creative force in evolution, bringing new trait combinations together as well as leading to novel phenotypes. This may be of little comfort to rice farmers, but it is quite interesting to evolutionary biologists. Nonetheless, as this study and others (e.g. Kane & Rieseberg 2007) begin to uncover the origins of agricultural weeds, efforts to prevent the evolution of new weeds will have better odds of success.