Sixteen 2-year-old “Fuyu” trees were used for this experiment. The trees were grafted on one year old D. kaki seedlings grown in 3 L pots in spring two years ago and grown under a rain shelter in Sweet Persimmon Research Institute, Gimhae, South Korea. Pots, spaced at 50 × 50 cm, contained sandy loam soil and they were not fertilized until the treatment. The chemical properties of the pot soil before treatments are shown in Table 1. LPM used in this experiment was collected from an aerobically-digesting facility for pig slurry. Chemical properties of the LPM are shown in Table 2. Trees were supplied with about 0.7 L・d⁻¹ water throughout the experiment with a sprinkler irrigation system.

With two nutrient sources, CF and LPM, and two levels of each, the treatments were arranged in a 2 × 2 factorial design with three single-tree replicates using

twelve trees. An additional four trees were used for analysis at start of the experiment before the first treatment on July 1. Each nutrient source at low and high levels was supplied to the pot on July 1, 3, and 5 in three equal aliquots. LPM of 300 (low) and 600 mL (high) was poured on soil surface of the pots on each treatment date. The treatment of LMP was established not to injure trees based on our preliminary study. For the CF treatment, urea and potassium chloride (KCl) were fertigated with 1% (w/v) solutions as was commonly done in the orchard. Urea and KCl were adjusted to have the equivalent amount of total N and K contained in the LPM. Uptake and partitioning of K were not of primary interest in this study, but it was nonetheless incorporated in this CF treatment to mimic the composition of the LPM. The total N and K supplied to a pot were 1.2 g N and 1.15 g K for the low and 2.4 g N and 2.3 g K for the high level.

Secondary shoot growth was monitored to assess the response of the trees to different nutrient sources and their levels. Four trees on July 1 and the remaining 12 trees on August 6 were destructively harvested and divided into various tree parts. N contents in July 1 samples served as the basis for comparison with those in August 6 samples. The trees were divided into leaves, aerial woods, and roots. Aerial woods were separated into shoots, secondary shoots, 1- to 2-year-old woods (old woods) while the roots were divided into small (≤2) and large ones (>2) according to their diameter in mm. After drying at 80˚C for 48 h, the samples were ground with a Wiley mill (3383-L10; Thomas Scientific, Swedesboro, NJ) to pass through a 20-mesh screen. To determine total N, 0.2-g sub-samples were analyzed with a Kjeldahl instrument (Kjeltec 2300, Foss Co., Höganäs, Sweden) using the micro-Kjeldahl method [15] . By taking the concentration of N in the different tree parts and their dry weights (DWs), N contents in each part were calculated. Differences in the amount of N between July 1 and August 6 are considered the amount absorbed and distributed to different tree parts.

Soil samples were collected from soils remained after excavating roots from four pots on July 1 before treatments and the remaining 12 pots on August 6. Organic matter and various elements were analyzed according to the Methods for Chemical Analysis of Soils and Plants [16] . Briefly, soil pH was measured with distilled water (1:5 with H₂O) with a pH meter (Orion 520A; Orion Research Inc., Boston, MA). Organic matter was measured by Tyurin method [17] . Total N content was determined after the Kjeldahl digestion. The available P was determined using the Lancaster method. Exchangeable K⁺, Ca²⁺, and Mg²⁺ were estimated by using inductively coupled plasma spectroscopy (Analyst 300; Perkin-Elmer, Norwalk, CT) after extracting with 1 M ammonium acetate (NH₄OAc).

Statistical analyses were performed by analysis of variance using SAS software (version 8 for Windows; SAS Institute, Cary, NC).

From 2 weeks after the treatments, unnecessary secondary shoots started to grow for the CF but none for the LPM. Treating with either the low or the high level of CF produced secondary shoots by August 6 (Figure 1), exhibiting the final DWs of 3.4 and 2.9 g per tree, respectively. Table 3 shows N concentrations in different tree parts measured on August 6. N concentration was not significantly affected in all tree parts by different nutrient sources. However, N concentration was higher especially in shoots, old woods, and small roots at the high level of both CF and LPM. No occurrence of secondary shoot growth of the trees receiving LPM suggested that N uptake by those trees was slower than that by the

^aCF: fertigated with urea and potassium chloride (KCl) solution; ^bLevels (g/pot): low, 1.2N - 1.15K; high, 2.4N - 2.3K. Data represented are means ± SE (n = 3). NS: non-significant, *: significant at P ≤ 0.05, **: significant at P ≤ 0.01.

trees receiving CF. Excessive N uptake during summer would cause vigorous growth of water sprouts and secondary shoots with the concurrent deterioration of fruit quality [11] [18] . It appeared that LPM could be more favorable to keep tree vigor stable than CF as long as the method for its application was carefully controlled in terms of the time and quantity.

Table 4 shows the amount of N absorbed and distributed to different tree parts between July 1 and August 6. Two nutrient sources did not significantly affect the amount of N increase in different tree parts. Of the total aerial wood, N in secondary shoots accounted for 28% at the low and 17% at the high CF. At the high level of both CF and LPM, N increase tended to be higher in all tree parts except secondary shoots. A significant N increase was observed in the leaves (P = 0.044) and large roots (P = 0.027) at the high level of application. Total N was 80% more for the CF (551 vs. 993 mg) and 31% more for the LPM (582 vs. 766 mg per tree) at the high level. Considering that the trees that received LPM absorbed a similar amount of N to the ones that received CF (Table 4), LPM could be an acceptable supplemental source of nutrients as has been previously reported for nectarine [8] and pear [9] . Total N content per tree increased more with the high level of both LPM and CF, however, the extent of the increase was less with LPM than with CF. It appeared that, unlike CF, N in LPM was not readily available for the tree since it mostly exists in organic forms and it would require some time to mineralize them [3] [19] .

It is of interest whether the LPM application changes chemical properties of the soil at the end of the experiment. The soil that received LPM contained more organic matter, available P, and exchangeable K⁺ and Mg²⁺ than the soil that received CF (Table 5). Organic matters in the LPM soils increased by 10 and 24%

^aCF: fertigated with urea and potassium chloride (KCl) solution; ^bLevels (g/pot): low, 1.2N - 1.15K; high, 2.4N - 2.3K. Data represented are means ± SE (n = 3). NS: non-significant, *: significant at P ≤ 0.05, **: significant at P ≤ 0.01.

^aCF: fertigated with urea and potassium chloride (KCl) solution; ^bLevels (g/pot): low, 1.2N - 1.15K; high, 2.4N - 2.3K. Data represented are means ± SE (n = 3). NS: non-significant, *: significant at P ≤ 0.05, **: significant at P ≤ 0.01.

at the low and high levels, respectively, compared with those of the respective CF. Contents of available P and exchangeable K⁺ for the LPM were 1.8 ~ 5-fold higher than those for the respective CF. Higher total N content was also observed in the LPM soils although not statistically significant. High contents of organic matter in the LPM soil on August 6 (Table 5) indicated that much of the organic matter still remained undecomposed in the soil by that time. Soil organic matters could slowly release inorganic nutrients through mineralization, the process of which helps to decrease the extent of nutrient leaching from irrigation [19] . Therefore, it is probable that the soil organic matter from the LPM could have contributed to significant increases of total N, available P, and exchangeable K⁺ and Mg²⁺ in the soil.