A sampling experiment was conducted in order to determine the best sampling size to estimate the quantity of red rice seeds in soil. Four sampling sizes 10cm deep were collected in a lowland area infested with red rice located in the Campus of ESALQ/USP during 1988/89. There was great heterogeinity in seed distribution and there were necessary about 30 soil samples varying between 0,375 up to 0,5m² in order that differences above 20% of average to be significant at the 5% error probability. The method was neither adequate nor practical due amount of time and work required to do it.
red rice; seeds; buried seed bank; sampling
