In an agricultural ecosystem, the redution of weed seed bank is of utmost importance. In lowland areas, red rice (Oryza sativa) is the most important weed. Therefore, an experiment was conducted to evaluate the evolution of the red rice seed bank under different seeding systems and alternatives of soil management. The experiment was conducted during two years (1996/97 and 1997/98 growing seasons) in Santa Maria, RS in Albaqualf soil with an average infestation of 556 viable seeds per square meter. The assessement of the seed bank was conducted by ten samplings using a cylinder with 10cm of diameter and 10cm depth. After each sampling the seeds were separated and the number of viable seeds was determinated by tetrazolium test. The sampling were carried out before establishing the experiment (September 1996) after the first harvest (May 1997) and after the second harvest (May 1998). The results showed that the traditional rice cropping method increases the red rice seed bank, whereas the systems in which the rice is seeded in flooded soil (no tillage water seeded, water seeded and transplanted seedlings) are as efficient in reducing the red rice seed bank as the system that applies crop rotation with sorghum or fallow without the presence of grazing animals or soil tillage during summer.
Weed; crop rotation; plant population dinamics; crop systems; Oryza sativa