Field experiments were conducted at Shimen State Farm, Jinhua (29°7′ N, 119°39′ E, 64 m altitude) of Zhejiang Province in 2001 to 2002. A widely grown indica hybrid Shanyou63 (single rice from late May to early October) and a conventional indica Jinzao22 (double early rice from early April to middle July) were used. Some basic properties of the soils are shown in Table 1
. For Shanyou63 transplanting spacing was 20 cm×20 cm with one seeding per hill. Plot size was 30 m2
. Phosphorus at 40 kg P/ha, potassium at 100 kg K/ha, and zinc at 5 kg Zn/ha were applied at basal fertilizer. For Jinzao22, direct sawn, plot size was also 30 m2
. Phosphorus at 15 kg P/ha was applied at basal fertilizer, potassium at 60 kg K/ha was applied at early tillering. The plots were kept flooded throughout the growing season. Pests, diseases, weeds were intensively controlled to avoid yield loss. Seven N treatments were arranged in randomized complete block design with four replicates.
The seven N treatments (Table 2
) were different fertilize N management strategies, including control (N1
) and three fixed-N split treatments. The three fixed-N split treatments with total N rates of 60 (N2
), 100 (N3
) and 140 (N4
) kg N/ha with 20% applied at three-leaf stage, 35% at early tillering, 45% at panicle initiation in Jinzao22 and total N rates of 60 (N2
), 120 (N3
) and 180 (N4
) kg N/ha with 35% applied at basal fertilizer, 20% at midtillering, 30% at panicle initiation and 15% at heading in Shanyou63. Control plots (N1
) received a full dose of phosphorus, potassium but no N.
FFP (N5) was based on the common practice of the farmers near the sites. Total N rate was 140 kg N/ha with 40% applied before sowing, 20% at three-leaf stage, 40% at early tillering in Jinzao22 and total N rate was 200 kg/ha with 50% applied before transplanting, 35% at tillering, 15% at flowering (FL) in Shanyou63.
Treatment N6 is SSNM, the timing of N application was fixed but the rate of N application was varied depending on leaf N status. N applications in Jinzao22 and Shanyou63 are described in Table 3, respectively. A chlorophyll meter (SPAD-502) was used to obtain SPAD values on ten uppermost fully expanded leaves in each plot.
Treatment N7 is RTNM, there was no basal N application. Weekly SPAD monitoring started at 10 d after transplanting (or 30 d after sown) and continued until flowering. If the SPAD reading was below 35, 45 kg N/ha was applied around the panicle initiation (PI) stage, 30 kg N/ha was applied for other growth stages in Shanyou63. If SPAD reading was below 34, 45 kg N/ha was applied from PI to booting stage and 35 kg N/ha was applied at other stages in Jinzao22.
Sampling and measurements
At transplanting (TP), midtillering (MT), panicle initiation (PI), flowering (FL) and maturity (MA) stages, five representative hills of the plants at each plot were separately sampled and divided into leaf blades, stems plus sheaths and grains. The samples were oven-dried at 70 °C till constant weight. Nitrogen concentration in plant tissues was determined by the Kjeldahl method (Lu, 1999
). At maturity, rice was harvested and grain yield was expressed as 14% water content.
The term NUE is used in different contexts by different workers (Craswell and Godwin, 1986
). The following definitions are used here.
Post-flowering dry matter redistribution (post-DMR)=DMA at flowering−DMA in stems and leaves at maturity,
Post-flowering nitrogen redistribution (post-NR)=NA at flowering−NA in stems and leaves at maturity,
where ΔTN was total aboveground plant N accumulation in the plot that received N fertilizer minus total aboveground plant N accumulation in the zero-N control, FN was the amount of N fertilizer applied, ΔGY was grain yield in the plot that received N fertilizer minus grain yield in the zero-N control.
Analyses were conducted on individual year data and then on 2001 and 2002 combined. Year was treated as the main unit with replication nested within years; the treatments were the subunits. Data were analyzed by the following analysis of variance (SPSS, version 11.5). Differences among treatment means were compared by least significant difference (LSD). For mean separations, LSD values were used at P