Wheat grain yield, phosphorus uptake and soil phosphorus fraction after 23 years of annual fertilizer application to an Andosol Shigeru Takahashi * , Muhuddin R. Anwar Integrated Soil Fertility Management Research Team, National Agricultural Research Center, 3-1-1 Kannondai, Tsukuba, Ibaraki 305-8666, Japan Received 15 December 2005; received in revised form 2 November 2006; accepted 3 November 2006 Abstract A field experiment was conducted on an Andosol to evaluate wheat (Triticum aestivum L.) yield, P and N uptake and soil P fraction after long- term fertilization (no fertilizer, NPK, NP, NK and PK treatments). Application rates of N, P and K fertilizers were 100, 65 and 83 kg ha 1 year 1 by ammonium sulfate, superphosphate, and potassium chloride, respectively. Phosphorus fertilization was critical for grain yield since the NK treatment did not increase yield compared with no fertilizer treatment. Agronomic efficiency of P was greater than agronomic efficiency of N, although apparent recovery of P and N were 17 and 53%, respectively. Combination application of fertilizer P and N resulted in the greatest grain yield over 23-year cultivation. Interaction impact on grain yield between P and N ranged from 71 to 109%, and was greater than the values for cereals in the earlier works. The N/P ratios of wheat decreased by P application and increased by N application. The N/P ratios in NPK and NP treatments were higher than the values attaining maximum yield for cereal crops reported by other works. Increase in soil available P in the treatments with P application was modest after 23-year fertilization. Total inorganic P (P i ), Ca-P i + Al-P i + Fe- P i , increased in the treatments with P application at 0–15 cm. Total P i was greater at 0–15 cm depth than at 30–50 cm depth. Although apparent recovery of fertilizer P (Ca-P i as superphosphate) was less than 20%, soil Ca-P i was very low even in the treatments with P application. This meant that unutilized fertilizer P did not remain in the form of Ca-P i . In contrast to inorganic P, there was no significant difference in total organic P (P o ), Ca-P o + Al-P o + Fe-P o , among the treatments and soil depths. Regardless of fertilizer treatments, Al-P i was the predominant form at 0–15 cm depth and Al-P i concentrations were similar to Fe-P i concentrations at 30–50 cm depth. On the other hand, Fe-P o was greater than Al-P o at 0–15 cm depth. Difference in inorganic P at 0–15 cm depth demonstrated that unutilized fertilizer P was transformed mainly to Al-P i followed by Fe-P i . However, wheat seemed to absorb P from Al-P i and Fe-P i modestly. # 2006 Elsevier B.V. All rights reserved. Keywords: Fractionation; Long-term experiment; Phosphorus; Wheat 1. Introduction Andosols cover more than 50% of upland areas in Japan (Hirata et al., 1999). Phosphorus is one of the limiting nutrients in Andosols which have high P fixing capacity due to their high Al and Fe oxide concentrations. Thus, large amounts of P fertilizer are applied to Andosol fields continuously. This may be the reason for the high P/N ratio of fertilizer consumption in Japan (Aulakh and Malhi, 2005). Long-term experiments provide one of the means to measure sustainable management systems in agriculture (Rasmussen et al., 1998). Several long-term fertilizer experiments have been conducted on Andosols on changes in major and trace elements in soil (Hirata et al., 1999; Takeda et al., 2005), soil organic C (Mathews et al., 2001; Gudmundsson et al., 2004) and N leaching (Nakamura et al., 2004). Few long-term fertilizer experiments on wheat in Andosols are available. The low efficiency of plant P uptake is the main problem associated with P application. Nitrogen is also the key element in achieving consistently high yield in wheat. The interaction of N with P can be termed the single most important nutrient interaction of practical significance (Aulakh and Malhi, 2005). Recently, Sadras (2006) investigated the N/P ratio of cereals (n = 759) and indicated over 40% of crops attaining maximum yield had N/P ratio in a relatively narrow range between 4 and 6. When P input from fertilizer exceeds P output in crop, P accumulation in the soil gradually increase over time (Kuo et al., 2005). The information obtained from P fractionation www.elsevier.com/locate/fcr Field Crops Research 101 (2007) 160–171 * Corresponding author. Tel.: +81 298 38 8829; fax: +81 298 38 8837. E-mail address: shigeru@affrc.go.jp (S. Takahashi). 0378-4290/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.fcr.2006.11.003