Introduction

Sulfur is reduced to a sulfide form under anoxic conditions in submerged paddy soils. Free hydrogen sulfide is generated and inhibits the rice root growth in paddy soils with low content of free iron oxide and high sulfate content. On the other hand, application of sulfate has been expected to decrease methane (global-warming gas) emissions from paddy soils (e.g. Van der Gon and Neue, 1994) because of sulfate's function as an electron acceptor. In addition, increased sulfide reduces concentrations of heavy metals such as Cd in pore water of anoxic soils through precipitation as insoluble sulfide (Carbonell-Barrachina et al. 2002). This precipitation might reduce uptake of heavy metals by rice plants. These effects can be beneficial for prevention of global warming and safe food production. For those reasons, it is very important to estimate sulfur levels without negative effects to rice root growth in paddy soils for environmentally conscious rice production with high yield levels. Some studies have reported no clear decrease in rice yields through application of sulfur (e.g., Van der Gon and Neue, 1994). Ferrous iron reacts with sulfide and alleviates toxicity of free hydrogen sulfide. Therefore, sulfur levels in paddy soils, which inhibit rice growth, should differ depending on the iron content of soils. In this study, we investigated the influences of sulfur levels on rice grown in paddy soils with different contents of iron.

Materials and methods

We used three alluvial soils (Entisols) and three volcanic ash soils (Andisols) with wide variation in reducible iron content (extraction with pH 2.8 CH₃COONa buffer of reduced soil) in this study. Soils named in this study are the following. Alluvial soils used were Tsuruoka, Ogata and Kiyosato soils with respective reducible iron contents (g Fe kg^-1) of 1.3, 6.6 and 15. Volcanic ash soils used were Naruko, Shikama and Moka with respective reducible iron contents (g Fe kg^-1) of 3.1, 5.9 and 6.5. These soils were filled into square frames (0.076 m²; 0.15 m deep) placed at an experimental paddy field of Tohoku University, Japan. Soil sulfur contents were varied by gypsum addition at rates of 0, 500 and 1000 g m^-2 (corresponding to 0, 93 and 186 g S m^-2) on 12 May 2004. Basal fertilizers (5 g N m^-2, 3.1 g P m^-2, 5.8 g K m^-2) and rice straw (500 g m^-2) were applied, and topdressing (2 g N m^-2) was conducted at the ear formation stage. Rice seedlings (Oryza sativa L. cv. Hitomebore) were transplanted to each frame on 14 May and harvested on 18 September in 2004. In 2005, the same soils were used without further addition of gypsum. Six soils with three sulfur levels were arranged by randomized block design with three replications. Sulfide (as total reduced sulfur) and free hydrogen sulfide in the soils were evolved with N₂ and trapped in a Zn-acetate solution with or without previous addition of sulfuric acid, respectively. Concentration of sulfide in the solution was determined by iodimetric procedure.

Results and discussion

In 2004, sulfur treatment stimulated increase in tiller number, but decreased brown rice yield by 0–12% compared to no sulfur treatment in six soils. In April 2005, amounts of sulfate remaining in soils, as estimated by extraction with 0.03 M Ca(HPO₄)₂ solution were different among the six soils. Tsuruoka and Kiyosato soils respectively retained 15% and 27% of sulfur applied in 2004 for high sulfur treatment of 186 g S m^-2. Soil with higher contents of reducible iron retained greater amounts of sulfur, probably because of more intensive formation of insoluble FeS. Moreover, greater amounts of sulfur remained in volcanic ash soils than in alluvial soils after one year.

In 2005, sulfur treatment did not increase the tiller number and decreased the brown rice yield by 7–18% compared to no sulfur treatment in Tsuruoka, Naruko and Shikama soils. With high sulfur treatment, free hydrogen sulfide measured at the ripening stage of rice increased significantly in these soils. A significant positive correlation was found between the amounts of free hydrogen sulfide and the molar ratios of sulfide to reduced iron of soils. Suppression of rice growth in high sulfur soil was more severe in 2005 than in 2004, although soil sulfur contents were lower in 2005 than 2004. The reason for more severe suppression of growth in 2005 would be that gypsum treatments increased the concentration of NH₄⁺ in soil solution through exchange of Ca²⁺ from gypsum for NH₄⁺ absorbed in soil and resulted in increasing tiller number in 2004. With sulfur treatment, grain yields decreased and the amount of free hydrogen sulfide increased to more than 0.5 mg S kg^-1 in Tsuruoka, Naruko and Shikama soils in 2005. This correspondence suggests that the hydrogen sulfide toxicity decreased the rice yields in these soils. Molar ratios of sulfide to reduced iron were more than 0.05 in these soils. It was concluded that the paddy soils with the ratio of more than 0.05 generated free hydrogen sulfide and decreased the rice yield.