1) Introduction Andisols is widely distributed in Japan, and is regarded as one of the most important soil types in Japanese agriculture, in particular for the upland crop production due to the high distribution percentage of approximately 47 % of upland field. In upland crop production, physiological disorder due to Ca deficiency often occur in crops grown in vegetable fields despite the much presence of exchangeable Ca measured by 1M ammonium acetate extraction. Hence, effective fertilization technique is required for Ca supply to vegetable crops. Toma et. al (1995) reported the possibility of the use of phosphogypsum (PG) as a effective calcium (Ca) source in non-allophanic Andisols. However, it is predicted that the effectiveness of PG application for Ca supply was differed among Andisols with different colloidal component. In this regard, Saigusa et al. (1990) reported that greater amount of Ca for crop growth was required in allophonic Andisol rather than in non-allophanic Andisols. Therefore, the objective of this study was to clarify the effect of PG application on the chemical properties of Andisols, and to estimate the ability of PG to supply Ca in Andisols with different colloidal components with special reference to water-soluble Ca.

2) Materials and methods Five virgin Andisols with different inorganic and organic colloidal components were used in this experiment. Kawatabi A and Bw horizon soils were classified into non-allophanic Andisols, and Kameoka A1, A2, and B horizon soils were classified into allophanic Andisols. We also divided along total carbon content into humic and non-humic Andisols; soils with greater than or equal to 5% of total carbon content were categorized into humic soils, and soils with less 5% of it were categorized into non-humic soils. The PG byproduct of the Japanese fertilizer industry was used in this experiment. The PG was applied to the soil samples at the rates of 0, 0.3, 0.6, and 1.2 g kg^-1 air-dry soil (< 2 mm) as a CaO and then mixed uniformly. The pH (1:5 H₂O) and electrical conductivity (EC) of the soils were measured by a glass electrode and EC meter, respectively. Acid oxalate extractable Al, Fe, and Si and pyrophosphate extractable Al and Fe were determined by ICP-AES (SPS3100, SII). The water-soluble Ca was determined by the water extractable methods. Soil to distilled water ratio was 1:5 by weight in water-soluble CaO.

3) Results and discussion (1) Changes in pH and EC of soils with different colloid component The change of soil pH differed among soil samples; Kameoka A2 and B horizon exhibited almost the same pH regardless of the PG application rate. Kawatabi A horizon and Kameoka A1 horizon soils exhibited a decrease by PG application. In particular, the soil pH of Kawatabi A horizon soil was sharply decreased when PG was applied at the rate of 0.3 CaO g kg^-1, and subsequently, it decreased gradually from 4.9 to 4.4 with increasing application rate. On the other hand, the soil pH gradually increased in the case of Kawatabi B horizon soil. Saigusa et al. (1992) reported that 8-15cmol_c kg^-1 of constant negative charge derived from 2:1−2:1:1 minerals was found in non-allophanic Andisols. It was suggested that Kawatabi A horizon soil possessed a high constant negative charge and that H⁺ ions were easily released by the isomorphous substitution of Al³⁺ ions held by a constant negative charge, resulting in a sharp decrease of soil pH. On the other hand, it was considered that in Kawatabi B horizon, the number of OH^- ions released could be exceed that of H⁺ ions released, by the ligand exchange of SO₄^2- ions. The soil EC was clearly increased in proportion to the application rate of PG for all the soil samples. This may reflect the relatively high solubility of PG in water, i.e., with 2.1 g L^-1. However, the increment in the soil EC was differed among soil samples; the soil EC of non-allophanic Andisols was relatively higher than that of allophonic Andisols. (2) Changes in water-soluble calcium in Andisols with different colloid component With increasing in the PG application rate, the water-soluble Ca content increased in all the soil samples, ranging from 0.57 to 3.17 cmol_c kg^-1 when 1.2 CaO g kg^-1 of PG was applied. However, the increase amount of water-soluble Ca was differed among Andisols with different colloidal component. The relationship between the allophane content and the apparent increment of water-soluble Ca in soils was examined. The apparent increment of water-soluble Ca tended to decrease exponentially with increasing of allophone content in the soils. It was indicated that the increment in water-soluble Ca with PG application was depressed by binding alloplhane with Ca. Furthermore, on comparing with between Kawatabi A and Kawatabi Bw horizon soils that have less than 50g kg^-1 of allophone content, it was deduced that the apparent increment of water-soluble Ca of humic Kawatabi A horizon soil was lower than that of Kawatabi B horizon by 0.1 cmol_c kg^-1. Therefore, it was suggested that the apparent increment of water-soluble Ca was more affected by allophone content than humus content, although the humus content also affected the Ca availability in Andisols.