**2. Material and methods**

#### **2.1 Study site**

Two experiments (Exp.-1 and Exp.-2) were conducted in CETAPAR-JICA for soybean productivity with agropastoral system. CETAPAR-JICA is located in Colonia Yguazu (a Japanese settlement, 35°27'S, 55°04'W) in Alto Parana, Paraguay. Soil in this area is fertile and is known as "Terras Roxas" in Brazil (Igarashi 1997). Mean annual temperature and precipitation from 1972 to 2002 were 21.6°C and 1545 mm, respectively.

#### **2.2 Experimental design**

For Exp.-1, part of a field at CETAPAR-JICA, where soybean and wheat had been continuously cropped in a no-tillage system since 1993, was converted to Guinea grass (*Panicum maximum* cv. Tanzania) pasture in 1996. Established as a permanent pasture, it was maintained without fertilizer, cutting, or renovation for 7 years after establishment, and was used as a complementary pasture. In October 2003, the pasture was converted into an agropastoral plot where soybean and wheat were cultivated. the agropastoral plot was 2.97ha. In another part of the field adjacent to the agropastoral site, where soybean and wheat had been continuously cultivated in a no-tillage system since 1993, the non-converted treatment was replicated in three plots (control plots). Each plot was 0.68 ha.

For Exp.-2, 15 plots were arranged at the study site, each plot was 0.68 ha (124 m × 55 m) where soybean and wheat had been continuously cultivated in a no-tillage system since 1993. Twelve plots were randomly converted to Guinea grass (*Panicum maximum* cv. Monbasa) pasture in November 2003. These pastures were managed as intensive grazing pastures under high grazing pressure. The strip grazing was conducted in the pasture year round, and cattle were fed supplement during four months in dry season. Fertilization was also conducted (ammonium sulfate). The stocking rate was from 4.5 to 6.0 UA/ha for 3 years. Three plots of these pastures were also reconverted to soybeanwheat fields in October 2007 as no-tillage system (agropastoral plots). The non-converted treatment was replicated in three plots (control plots). Control plots in Exp.-1 and Exp.-2 were same plots.

#### **2.3 Chemical and physical properties of soil**

To investigate the chemical properties of the soil, samples from depths of 0-10, 10-20, 20-40, and 40-60 cm were collected independently from each plot, and the concentrations of phosphate, the percentage organic matter and pH, were measured. The concentrations of phosphate were analyzed using the Mehlich-III method, and percentage organic matter was analyzed using the Walkley-Black method. The pH of soils was measured using a pH meter (Horiba Co. Ltd.).

Moreover, soil was sampled from 0–5, 5–10, 10–20, 20–40, and 40–60 cm depths for measurement of physical properties, three phases of soil, bulk density, and soil aggregates (Only Exp.-1). Three phases of soils and soil aggregates were measured using a three-phase meter and an aggregate analyzer (Daiki Rika Kogyo Co., Ltd.), respectively.

We analyzed soybean and wheat production and soil chemical and physical data between the agropastoral plots and control plots using t-test, and the annual variation of chemical data in both plots using the Tukey-Kramer method. Details of the study methods were shown in Shimoda et al (2010, 2011).

#### **3. Results**

#### **3.1 Soybean production**

Since the yield of a soybean had a large change every year, the effect of agropastoral system was evaluated by using ratio of the soybean yield in agropastoral plots to that in control plots (Table 1). As a result, the ratios of the first year when reconverted into soybean field from the pasture were 2.35 in Exp-1 and 1.02 in Exp-2, respectively. In addition, the ratios of the second year were 1.86 in Exp-1 and 1.42 in Exp-2, respectively. The effect of Exp-1 was larger than that of Exp-2.

For Exp.-1, part of a field at CETAPAR-JICA, where soybean and wheat had been continuously cropped in a no-tillage system since 1993, was converted to Guinea grass (*Panicum maximum* cv. Tanzania) pasture in 1996. Established as a permanent pasture, it was maintained without fertilizer, cutting, or renovation for 7 years after establishment, and was used as a complementary pasture. In October 2003, the pasture was converted into an agropastoral plot where soybean and wheat were cultivated. the agropastoral plot was 2.97ha. In another part of the field adjacent to the agropastoral site, where soybean and wheat had been continuously cultivated in a no-tillage system since 1993, the non-converted

For Exp.-2, 15 plots were arranged at the study site, each plot was 0.68 ha (124 m × 55 m) where soybean and wheat had been continuously cultivated in a no-tillage system since 1993. Twelve plots were randomly converted to Guinea grass (*Panicum maximum* cv. Monbasa) pasture in November 2003. These pastures were managed as intensive grazing pastures under high grazing pressure. The strip grazing was conducted in the pasture year round, and cattle were fed supplement during four months in dry season. Fertilization was also conducted (ammonium sulfate). The stocking rate was from 4.5 to 6.0 UA/ha for 3 years. Three plots of these pastures were also reconverted to soybeanwheat fields in October 2007 as no-tillage system (agropastoral plots). The non-converted treatment was replicated in three plots (control plots). Control plots in Exp.-1 and Exp.-2

To investigate the chemical properties of the soil, samples from depths of 0-10, 10-20, 20-40, and 40-60 cm were collected independently from each plot, and the concentrations of phosphate, the percentage organic matter and pH, were measured. The concentrations of phosphate were analyzed using the Mehlich-III method, and percentage organic matter was analyzed using the Walkley-Black method. The pH of soils was measured using a pH meter

Moreover, soil was sampled from 0–5, 5–10, 10–20, 20–40, and 40–60 cm depths for measurement of physical properties, three phases of soil, bulk density, and soil aggregates (Only Exp.-1). Three phases of soils and soil aggregates were measured using a three-phase

We analyzed soybean and wheat production and soil chemical and physical data between the agropastoral plots and control plots using t-test, and the annual variation of chemical data in both plots using the Tukey-Kramer method. Details of the study methods were

Since the yield of a soybean had a large change every year, the effect of agropastoral system was evaluated by using ratio of the soybean yield in agropastoral plots to that in control plots (Table 1). As a result, the ratios of the first year when reconverted into soybean field from the pasture were 2.35 in Exp-1 and 1.02 in Exp-2, respectively. In addition, the ratios of the second year were 1.86 in Exp-1 and 1.42 in Exp-2, respectively. The effect of Exp-1 was

meter and an aggregate analyzer (Daiki Rika Kogyo Co., Ltd.), respectively.

treatment was replicated in three plots (control plots). Each plot was 0.68 ha.

**2.2 Experimental design** 

were same plots.

(Horiba Co. Ltd.).

**3. Results** 

**2.3 Chemical and physical properties of soil** 

shown in Shimoda et al (2010, 2011).

**3.1 Soybean production** 

larger than that of Exp-2.

The first, second and third year in Exp.-1 were drought years and the second year in Exp.-2 was drought year too. The ratios of drought years were larger than the ratio of normal year in both experiments.

Moreover, during the experimental period in Exp.-1, the relative yield of soybean decreased year by year from 2.31 t/ha in the first year to 1.11 t/ha in the fourth year.


Source: 1 from Shimoda et.al. (2010) and 2 from Shimoda et al. (2011). \*:Drought year.

Table 1. Study site profile and soybean production
