Weedy rice (Oryza sativa f. spontanea) is one of the three worst paddy weeds in most rice growing areas.  The unexpected heavy infestation is derived from a persistence of soil seed bank of weedy rice, which the shattered seeds chiefly feed back to.  Information on soil seed bank dynamics is imperative to predict the infestation of weeds.  In the present paper, the effect of rotary tillage on weedy rice seed bank structure was studied first, and a burial experiment of marked seeds was conducted to observe the overwintering survival, seed viability and seedling emergence of weedy rice.  The results showed that the proportion of weedy rice seeds in deeper soil increased but seedling emergence decreased with increasing plowing depth.  The viability of weedy rice seeds decreased as the burial duration time extended but more slowly in deeper soil layers.  Additionally, there was no significant effect of burial density on seed viability.  Moreover, the logistic model fitted well (R²≥0.95, P≤0.01) with the depressive trends of seed viability with increasing burial time under all burial depths and densities which can provide us further information about seed survival.  In field experiments, number of seedling emergence significantly decreased as seed burial depth increased, conversely, proportion of seedling emergence increased as seed burial density decreased.  This study has important implications for determining strategies for weedy rice management by exhausting its seed bank through the alteration of tillage practices.

Crop type is one of main factors influencing weed community structure.  However, the identity of weed communities associated with the cultivation of different crops in farmlands remains largely unclear.  A field survey of weed seed banks was conducted in 2 280 fields at 228 sites of 62 locations representing three different types of farmland (95 paddy, 73 summer-ripe, and 60 autumn-ripe farmlands) along the bank of the Yangtze River in Anhui Province, China.  A total of 43 families and 174 species of weeds were found in these weed seed banks.  A comparison of the composition of weed groups in the seed banks showed that the species number and density percentage of grass, sedge and broadleaf weed groups were similar among the different types of farmland.  The seed banks of all three farmland types shared 71 common weed species, accounting for 40.80% of the total number of species.  These common weeds, which were both associated and not associated with crops, accounted for 91.71% of the total dominance degree among all farmland types.  The crop-associated weed species were distributed in all soil layers of each farmland type.  The Shannon-Wiener index H´ (description of species diversity which is more sensitive to dense species) and Pielou’s evenness index J (description of species evenness) in summer-ripe farmland were similar to those in autumn-ripe farmland but differed from those in paddy farmland.  However, the Simpson’s index D (description of species diversity which is more sensitive to sparse species) was similar among all three farmland types.  The results of similarity comparison indicated that although the aboveground weed community differed among the different cropping patterns, the weed species composition in the soil seed bank was still similar.  Consequently, our results demonstrate that after the implementation of long-term monoculture patterns, weed species compositions in the soil seed bank in different farmlands become homogenized regardless of the crop type.