Effect of different application rates of organic fertilizer on soil enzyme activity and microbial population

Abstract

After cultivating 24 crops of vegetables for three consecutive years in a greenhouse, the effects of different application rates of compost (Rate 1, 270 kg N ha⁻¹ y⁻¹; Rate 2, 540 kg N ha⁻¹ y⁻¹; Rate 3, 810 kg N ha⁻¹ y⁻¹; Rate 4, 1,080 kg N ha⁻¹ y⁻¹) were compared with the effects of chemical fertilizer (CF) and no application of fertilizer treatments (CK) for some selected soil chemical properties, microbial populations and soil enzyme activities (dehydrogenase, cellulase, β-glucosidase, protease, urease, arysulphatase, and acid and alkaline phosphatases). The results show that the pH, electrical conductivity, concentrations of total nitrogen (N) and the organic matter received from compost treatment were generally higher than those received through CF treatment. The soil microbial biomass, populations of bacteria, fungi and actinomycetes, as well as soil enzyme activities increased significantly in the compost-treated soils compared with the CF-treated soil. In most instances, no significant increase was observed in the enzymatic activities studied for compost applications higher than a Rate 2 treatment. However, all enzymatic activities examined showed significant linear correlations with the organic matter contents of the soils. The vegetable yield reached its highest level at the Rate 2 treatment and declined or leveled off in the higher treatments, implying that a high application rate of compost cannot further increase the crop yield after the soil fertility has been established. High organic matter content in the soil was found to alleviate the adverse effect of soluble salts on vegetable growth. In conclusion, an application rate of compost at Rate 2, 540 kg N ha⁻¹ y⁻¹, is adequate on the basis of vegetable yields and soil chemical, biochemical and enzymatic properties in greenhouse cultivation under subtropical climatic conditions.

Key words:

• compost
• microbial biomass
• organic farming
• soil enzymes
• vegetable growth

INTRODUCTION

Soil organic matter affects crop growth and yield, either directly by supplying nutrients or indirectly by modifying the soil physical properties, thereby improving the root environment and, thus, stimulating plant growth (Avnimelech 1986; Cabrera et al. 1988; Liebig and Doran 1999). Bauer and Black (1994) indicated that the highest total aerial dry matter and grain yields were associated with the highest organic matter contents of the soils. In addition, crop production based on the use of organic manures rather than chemical fertilizers is assumed to be a more sustainable type of agriculture. Therefore, in recent years the application of organic fertilizer has received great attention from environmentalists, agriculturists and consumers alike.

Owing to the threats of heavy rains and typhoons during the summer months, cultivation of vegetables in greenhouses is popular in Taiwan, especially for organic farming. Organic farming has become very popular and the acreage under cultivation is steadily increasing. For vegetable cultivation, six to ten crops in a year, depending on the type of crop, is common. Vegetable crops require an adequate and continuous supply of nitrogen (N) for proper growth and maximal high quality yields. Therefore, a high rate of N fertilizer is applied in vegetable cultivation in Taiwan (Lian et al. 1996). The nutrients released after the biological breakdown of the soil organic matter supply the nutrients essential for plant growth in organic farming. In general, the mineralization rate of soil organic matter is slow (Fernandez et al. 2006; Li et al. 2005). Therefore, to establish and maintain soil organic matter content to a certain level through the initial application of a large quantity and the continuous application of compost are important in organic farming. However, climate and soil significantly affect the accumulation and storage of organic matter in the soil because of the interactions of temperature and moisture on plant productivity and the ability of the soil mineral components to retain organic matter. Under tropical and subtropical climatic conditions, high cultivation frequency, and a low input of the organic matter content in the farmland soil of Taiwan is generally low, and it is common for the soil organic matter content to be lower than 20 g kg⁻¹ (Huang 1994).

The biological component of soils usually responds more rapidly to changing soil conditions than either the chemical or physical properties (Anderson and Gray 1990; Powlson 1994; Powlson and Jenkinson 1981). Soil enzymatic activities have been used as indicators of soil fertility because they are a reflection of the effects of cultivation, soil properties and pedological amendments (Ceccanti et al. 1993; Skujins 1978).

The aim of this study was to examine the effects of different amounts of compost on selected chemical and enzymatic activities of an agricultural soil under intensive cultivation of vegetables for three consecutive years (24 crops), and to find out what an adequate rate of compost application is in the organic cultivation of vegetables.

MATERIALS AND METHODS

To study the effects of different application rates of compost on the soil characteristics and the growth of different vegetables, an experiment was conducted in a greenhouse in Chinan Branch Station, Kaohsiung District Agricultural Improvement Station, Kaohsiung, southern Taiwan, for a 3-year period. The mean annual precipitation of 2,000 mm y⁻¹ has a distinct seasonal pattern, with 90% falling from May to September (65% falling from June to August) and 10% from October to April. The mean daily temperature is 26.6°C, the highest monthly temperature is in July (30.5°C) and the lowest monthly temperature is in January (21.1°C). The experiment was carried out between August 1999 and September 2002. The soil used was hyperthermic, udic, haplaquept, mixed and calcareous, with a silty loam texture (0–15 cm). Some selected characteristics of the soil before planting (0–15 cm depth) were: pH, 7.54 (1:1 soil : water ratio); electrical conductivity (EC), 1.13 dS m⁻¹; organic matter, 25.8 g kg⁻¹; total N, 1.62 g kg⁻¹; Bray-1 P, 137 mg kg⁻¹; Mehlich-III extractable K, 80 mg kg⁻¹ and Mehlich-III extractable Cu, 1.8 mg kg⁻¹. There were six treatments, including four rates of compost (Rate 1, Rate 2, Rate 3 and Rate 4), a chemical fertilizer treated plot (CF), and a non-fertilized control plot (CK). The amounts of compost applied in Rate 1, Rate 2, Rate 3 and Rate 4 were based on the same amount (Rate 1), twofold (Rate 2), threefold (Rate 3) and fourfold (Rate 4) of N in the CF treatment. The locations of the individual research plots and their respective treatments remained unchanged during the 3-year period. Fertilizer applied in the CF treatment to each crop was based on the recommended rate, which varied depending on the soil test and the crop grown. A total of three types of commercial composts and a farm-compost (Table 1) were applied rotationally throughout the entire study. Compost A was composed of pig manure and sawdust; compost B was composed of wool;

Table 1 Major chemical properties of the composts used in the experiment

Item^†	Compost A	Compost B	Compost C	Compost D
pH (1:5)	7.17	5.50	7.29	8.00
EC^‡ (dS m^−1)	12.6	18.7	4.6	11.6
Organic matter (g kg^−1)	715	521	736	469
Total nitrogen (g kg^−1)	52.4	26.0	24.3	21.8
Organic N (g kg^−1)	50.7	19.8	22.9	21.3
Insoluble N (g kg^−1)	34.3	14.6	19.3	18.8
NO^− 3-N(g kg^−1)	0.05	0.05	0.03	0.45
NH^+ 4-N(g kg^−1)	1.63	6.15	1.34	0.10
Total P (g kg^−1)	8.93	8.69	6.41	3.55
Total K (g kg^−1)	12.0	19.3	9.8	30.9
Total Ca (g kg^−1)	54.6	18.4	43.9	24.3
Total Mg (g kg^−1)	17.6	10.1	10.6	6.0
Total Cu (mg kg^−1)	18.8	12.6	15.1	16.0
C/N	7.9	11.6	17.5	12.5
Moisture content (%)	32.2	25.8	30.5	38.5
^†On a dry weight basis except for the moisture content.
^‡EC, electric conductivity (1:5).

compost C was composed of soybean meal, peanut hull and sawdust; compost D (farm-compost) was composed of the plant residues of the farm. The total rates of chemical fertilizer applied in the CF treatment were 270, 133 and 150 kg ha⁻¹ y⁻¹ for N, P₂O₅ and K₂O, respectively. The total rates of compost applied for Rate 1, Rate 2, Rate 3 and Rate 4 were 270, 540, 810 and 1,080 kg N ha⁻¹ y⁻¹, respectively. Therefore, the total amounts of compost applied in the 3 years of the experiment were 25,330, 50,660, 75,990 and 101,320 kg ha⁻¹ (on a dry weight basis) for Rate 1, Rate 2, Rate 3 and Rate 4 treatments, respectively. On a N basis, the total amounts of A, B, C and D composts applied for the Rate 1 treatment were 360, 180, 180 and 90 kg ha⁻¹ per 3 years, respectively. On an organic matter basis, the amounts of compost applied for the Rate 1 treatment were 15,907, 31,814, 47,222 and 63,629 kg ha⁻¹ per 3 years for compost A, B, C and D, respectively. The chemical fertilizers used were ammonium sulfate for N, ordinary superphosphate for P and KCl for K. All fertilizers were applied to the soil surface and were incorporated by means of rototilling before transplanting the seedlings. Twenty-four vegetable crops, including cabbage (Brassica oleraceae L. Capitata Group) (one crop), kohlrabi (Brassica oleraceae L. Gongylodes Group) (one crop), pak-choi (Brassica campesttics L.) (fourteen crops), spinach (Spinacia oleracea L.) (one crop), lettuce (Latuca sativa L.) (three crops), amaranth (Amaranthus tricolor L.) (three crops) and water convolvulus (Ipomoea aquatica Forsk) (one crop) were planted over the entire 3-year study period.

For each crop, uniform seedlings were transplanted to each plot. All plots were irrigated when necessary with tap water to saturation and then drained freely to field capacity. Hand weeding was done to manage the weeds and no pesticides were applied. After harvesting, the plots were immediately rototilled to a depth of 15 cm.

All treatments were replicated four times in a randomized complete block design using 1.5 m × 3 m individual plots. Composite soil samples, which were composed of three randomly extracted soil cores for each plot, were collected from the surface layer (0–15 cm) after harvest of the last crop of vegetables (20 September 2002). Field-moist soil samples were divided into two sub-samples. One soil sub-sample was sieved to pass through a 2 mm mesh and was then stored at 2°C for biological and biochemical analysis, and the other soil sub-sample was allowed to air-dry at room temperature. The air-dried soil samples were also ground to pass through a 2 mm screen and were then stored in sealed plastic bags for chemical analysis at a later date.

Soil pH (1:1 soil : water; for compost, 1:5) (McLean 1982) and EC of the saturation extract (for compost, 1:5) were determined (Rhoades 1982). Total N, including nitrate and nitrite, was determined by distillation after digestion of the sample with a mixture of salicylic acid and sulfuric acid plus sodium thiosulfate (Bremner and Mulvaney 1982). Nitrate N (including nitrite-N) was extracted with 1.0 mol L⁻¹ N hydrochloric acid (HCl) solution (Blacquiere et al. 1987), reduced with zinc powder (Broaddus et al. 1965) and then determined using the Griess–Ilosvay nitrite method (Keeney and Nelson 1982). The organic matter contents of compost and soil were determined by the loss of weight at 450°C (Gagnon et al. 1997). Soil moisture was determined after drying at 105°C for 24 h.

Soil microbiological analysis was carried out as follows: 10 g of fresh soil was diluted in 10-fold series in sterile water. Viable bacterial numbers were determined in NA agar (Difco, Detroit, MI, USA) containing 50 mg L⁻¹ cycloheximide using plate colony counts (Parkinson et al. 1971). Viable fungus propagules (c.f.u.) and viable actinomycete propagules were determined in rose bengal agar (Difco) containing 100 mg L⁻¹ chloramphenicol (Martin 1950) and in actinomycetes isolation agar (Dfico), respectively, using the same dilution plate method (Parkinson et al. 1971). Soil microbial biomass was measured using the chloroform fumigation–extraction method (Vance et al. 1987). Organic C in the extract was measured by persulfate oxidation using a total organic C analyzer (Model 1010, O.I. Analytical, College Station TX, USA). The microbial C (Cmic) was calculated from Cmic = 2.22 × Ec, where Ec = (C extracted from fumigated soil) − (C extracted from non-fumigated soil) (Wu et al. 1990). Microbial N (ninhydrin-N) release from the biomass was determined colorimetrically at 560 nm, as described by Amato and Ladd (1988). Basal respiration was determined by placing 20 g of oven-dried equivalent of field-moist soil in a tube and incubating the sample in the dark in a 250 mL airtight Duran bottle along with 20 mL of 0.05 mol L⁻¹ NaOH to absorb CO₂. The CO₂-C was determined by titration (Jäggi 1976). The nitrification rate was determined using the procedure of Berg and Rosswall (1985). The levels of nine enzymatic activities in the soil were measured: dehydrogenase (Thalmann 1968), cellulase (Schinner and Von Mersi 1990), β-glucosidase (Tabatabai 1982), protease (Ladd and Bulter 1972), urease (Kandeler and Gerber 1988), arylsulphatase (Tabatabai and Bremner 1970), acid and alkaline phosphatases (Tabatabai and Bremner 1969).

Statistical analysis was carried out by multiple anova (Duncan's multiple range test at the 95% level of probability) to determine significant differences between treatments. The occurrence of relationships between different data was assessed by means of simple linear regressions. The symbols *, ** and *** indicate significance at 95%, 99% and 99.9% levels of probability, respectively.

RESULTS AND DISCUSSION

Chemical properties of soils after planting vegetables

Table 2 shows the effect of the soil treatments on the soil characteristics. Owing to the calcareous nature of the soil, the pH of the soils was greater than 7.0. However, the application of chemical N fertilizer (ammonium sulfate) resulted in a significant lowering of soil pH. This can be attributed to the acidification effect of ammonium ions during their transformation in the soil (Bolan et al. 1991; Kemmitt et al. 2006). The application of compost also lowered the soil pH. Mineralization of N is the major source of soil acidification (De Vries and Breeuwsma 1987). However, the acidification effect of compost was not as great as that of ammonium sulfate on the basis of the same amount of N applied. This is because of the neutralizing effect of compost associated with its alkalinity and content of basic cations (Kingery et al. 1994; Zaller and Koepke 2004). This result can be shown by comparing the pHs of the soils receiving the same amount of N (i.e. the pH of the CF treatment was lower than that of the treatment receiving Rate 1).

Although not significant in every instance, there was a tendency for the soil EC to increase with increasing application rates of compost (Table 2). This increase can be attributed to the solubilization of soluble ions such as chloride, sulfate, sodium and other inorganic ions from compost and organic species formed through organic matter mineralization. Compared with the Rate 1 and Rate 2 compost treatments, the Rate 4 treatment resulted in a significantly higher EC, which is higher than the critical soil EC (4.0 dS m⁻¹), and this will adversely affect the growth of most vegetables (Rhoades and Miyamoto 1990). Only high application rates (Rate 3 or Rate 4) of compost resulted in an increase in soil total NH⁺ ₄-N and Mehlich III extractable K, Ca, and Mg.

As expected, the application of compost in three consecutive years resulted in a significant increase in soil

Table 2 Effect of different application rates of compost on some selected chemical properties of the soil

Treatment	pH (1:1)	EC (dS m^−1)	OM (g kg^−1)	Total N (g kg^−1)	NH^+ 4-N (mg kg^−1)	NO^− 3-N (mg kg^−1)	P^c (mg kg^−1)	K^d (mg kg^−1)	Ca (mg kg^−1)	Mg (mg kg^−1)
CF	7.04 d^†	3.51 b	26.6 e	1.59 cd	29.2 a	51 c	167 d	114 d	1681 d	239 d
CK	7.83 a	1.56 d	25.1 e	152 d	26.3 ab	17 c	135 e	74 d	1634 d	251 d
Rate 1	7.77 ab	3.15 b	31.4 d	1.94 cd	26.1 ab	54 c	168 d	141 cd	2108 c	333 c
Rate 2	7.64 ab	3.53 b	38.0 c	1.89 cd	18.7 bc	89 bc	209 c	240 cd	2533 b	382 b
Rate 3	7.54 bc	4.58 ab	46.1 b	2.39 b	13.5 cd	159 ab	251 b	411 b	3063 a	482 a
Rate 4	7.39 c	5.17 a	53.0 a	3.13 a	6.10 d	200 a	294 a	613 a	2958 a	472 a
^†Within columns, means followed by the same letter are not significantly different (P = 0.05) using Duncan's multiple range test. CF, chemical fertilizer; CK, no application of fertilizer; EC, electric conductivity; OM, organic matter; P, Bray-1 extractable P; K, Mehlich-III extractable K; Ca, Mehlich-III extractable Ca; Mg, Mehlich-III extractable Mg.

organic matter and total P contents (Table 2). The increase in soil organic matter contents for the Rate 1, Rate 2, Rate 3 and Rate 4 treatments was 6.3, 12.9, 21.0 and 27.3 g kg⁻¹, respectively, compared with the CK plot. Taking the average soil bulk density as 1.34 kg L⁻¹, the increase in soil organic matter contents was 12,600, 25,800, 42,000 and 55,800 kg ha⁻¹ (calculated to 15 cm in depth) for the Rate 1, Rate 2, Rate 3 and Rate 4 treatments, respectively. Therefore, the percentages of applied organic matter that decomposed were 79%, 81%, 89% and 88% for the Rate 1, Rate 2, Rate 3 and Rate 4 treatments, respectively. This result indicates that most of the compost applied to the soil decomposed during the 3 years, and that the proportion of the applied organic matter that decomposed increased with the application rate of the compost.

Microbial population, biomass and activities

As shown in Table 3, the populations of bacteria, actinomycetes and fungi in the soil amended with compost were significantly higher than the populations in the CK and CF treatments. On average, the populations of bacteria in the soil amended with compost were 1.68-fold higher than those in the CF treatment and increased with increases in the compost application rates.

Table 3 Effect of different application rates of compost on microbial populations of the soils

Treatment	Bacteria (10^7 c.f.u. g^−1)	Actinomycetes (10^7 c.f.u. g^−1)	Fungi (10^4 c.f.u. g^−1)
CF	3.24 d^†	1.17 d	3.39 c
CK	3.72 cd	1.32 d	3.80 bc
Rate 1	4.47 cd	1.82 c	4.68 abc
Rate 2	4.68 bc	2.00 bc	5.25 ab
Rate 3	5.75 ab	2.24 ab	5.62 a
Rate 4	6.92 a	2.40 a	5.75 a
^†Within columns, means followed by the same letter are not significantly different (P = 0.05) using Duncan's multiple range test. CF, chemical fertilizer; CK, no application of fertilizer.

Similar observations have been reported by Brady and Weil (1999), where the introduction of organic matter into soil resulted in increasing soil microbial populations and soil biological activity and Weon et al. (1999), where the number of c.f.u. of bacteria and fungi increased when pig manure compost was added to soil. In addition, Alvarez et al. (1995) reported that the application of compost to soil increases the incidence of bacteria in the tomato rhizosphere.

Although microbial biomass usually makes up less than 5% of soil organic matter (Dalal 1998), it carries out many critical functions in the soil ecosystem. In addition, measurements of microbial biomass show a more rapid response than those of organic C to changes in organic matter or to the rate of decomposition of organic matter (Nannipieri et al. 1990). Differences in microbial-C and ninhydrin-N between the mineral fertilizer treatment (CF treatment) and the compost treatments are significant (Table 4). The CK treatment also resulted in low microbial-C and ninhydrin-N. Even if there is no significance in every case or no significant difference in microbial-C and ninhydrin-N among the treatments of different application rates of compost (Table 4), there were significant linear correlations between microbial-C (or ninhydrin-N) and soil organic matter contents (r = 0.9810** and 0.9298** for microbial-C and ninhydrin-N, respectively). No measurements was conducted for the available C in the composts applied, however, it can be assumed that the applied compost resulted in a greater availability of C and that it increased the possibility of microorganism growth and, therefore, greater microbial-C and ninhydrin-N values were found. Deng and Tabatabai (1996, 1997) found that the increase in hydro-soluble C and ninhydrin-N resulted from the organic soluble C derived from the application of sewage and compost.

Soil respiration rate has been widely used as a parameter for assessing the microbial activity of a system (Chander et al. 1998). A higher rate of CO₂ evolution in the soil was recorded with increasing compost application rate (Table 4). There was a good linear correlation between

Table 4 Effect of different application rates of compost on some microbial properties of the soil

Treatment	Microbial C (µg C g^−1)	Ninhydrin-N (µg Leu g^−1)	CO2 respiration rate (µg CO2 g^−1 h^−1)	Nitrification rate (µg NO2 g^−1 h^−1)
CF	301 d^†	6.32 bc	1.65 c	0.594 d
CK	316 d	4.99 c	1.59 c	0.340 d
Rate 1	421 c	8.39 ab	2.32 b	0.74 c
Rate 2	467 bc	9.61 a	2.63 b	0.94 b
Rate 3	549 ab	10.81 a	3.56 a	1.22 a
Rate 4	601 a	10.85 a	3.95 a	1.28 a
CF, chemical fertilizer; CK, no application of fertilizer. ^†Within columns, means followed by the same letter are not significantly different (P = 0.05) using Duncan's multiple range test.

the rates of CO₂ evolution and the application rates of compost (r = 0.9801*), although not every case was significant. A highly significant linear correlation between respiration rates and soil organic matter contents was also observed (r = 0.9928**). This means that the soil organic matter supplies the energy source for aerobic microorganism respiration. This result was also evident by the significantly linear correlation between microbial populations and rates of CO₂ evolution (r = 0.9784**, 0.9626** and 0.9405* for bacteria, actinomycetes and fungi, respectively).

Nitrification (ammonia oxidation by autotrophic ammonia-oxidizing bacteria) is a key process in agricultural and natural ecosystems and plays an important role in the global N cycle (Innerebner et al. 2006). There were significant differences in nitrification rates between the compost treatments and the CF treatment (Table 4). Nitrification rates also increased with the compost application rates until the Rate 3 treatment. This results from the continuous release of more NH₃ that equilibrates to NH⁺ ₄ when the compost applied to the soil decomposes. A highly linear correlation exists between soil organic matter contents and nitrification rates (r = 0.9684**); however, no such relationship exists between soil total N contents and nitrification rates.

Soil biochemical characteristics

According to Lal (1998), soil quality refers to a soil's ability to sustain productivity in terms of agricultural production. Using the activities of soil hydrolytic enzymes is a common approach for estimating soil quality (Gil-Sotres et al. 2005). All enzymatic activities increased with increases in the application rates of the compost up to a threefold application rate (Rate 3 treatment), but there was no significant difference between the threefold (Rate 3 treatment) and the fourfold application rate in most cases (Table 5). This result indicates that no further increases take place in the enzyme activities examined when the soil organic matter content reaches a certain level. The increase in soil organic matter content, resulting from the application of compost, in addition to the

Table 5 Effect of different application rates of compost on the enzymatic activities of the soils

Treatment	dHase	CMCase	Gase	Protease	Urease	aSase	aPase	Pase
	(µg TPF g^−1 h^−1)	(µg glu g^−1 h^−1)	(µg PNF g^−1 h^−1)	(µg tyr g^−1 h^−1)	(µg NH^+ 4 g^−1 h^−1)	(µg PNF g^−1 h^−1)
CF	87 d^‡	407 d	19.7 d	73.8 d	46.8 b	40.1 c	129 e	372 e
CK	61 e	430 cd	17.9 d	76.7 cd	50.1 b	43.0 c	133 e	407 de
Rate 1	102 bc	502 bc	26.7 c	122 bc	55.2 b	45.9 bc	173 d	434 cd
Rate 2	112 b	538 ab	27.8 c	131 bc	65.2 ab	53.8 ab	204 c	456 c
Rate 3	159 a	552 ab	36.1 b	123 b	67.7 ab	55.0 ab	238 b	463 b
Rate 4	151 a	594 a	43.9 a	187 a	84.4 a	59.4 a	272 a	588 a
^†Within columns, means followed by the same letter are not significantly different (P = 0.05) using Duncan's multiple range test. CF, chemical fertilizer; CK, no application of fertilizer; dHase, dehydrogenase; CMCase, cellulose; Gase, β-glucosidase; aSase, arylsulphatase; aPase, acid phosphatase; Pase, alkaline phosphatase.

incorporation of stable enzymes contained in the compost (Díaz-Marcote and Polo 1995), favors the formation of complexes with free enzymes and, therefore, the soil enzyme activities increase.

Soil dehydrogenase activity reflects the total range of oxidative activity of soil microflora and is consequently used as an indicator of microbial activity (Masciandaro et al. 1994; Perucci 1992). The dehydrogenase activity of compost-treated soils was significantly higher than that of the CF-treated soil and reached its maximum in the Rate 3 treatment (Table 5). Not applying any fertilizers for 24 crops in 3 years resulted in the lowest dehydrogenase activity (CK treatment). Marinari et al. (2000) reported that a higher level of dehydrogenase activity was observed in soil treated with vermicompost and manure compost compared with soil treated with mineral fertilizer. Martens et al. (1992) reported that the enzyme activities in organic matter amended soil increased an average of twofold to fourfold compared with unamended soil. These results are similar to our finding that dehydrogenase activities in soils with Rate 3 and Rate 4 treatments were approximately twofold higher than activity in the CK and CF soils. There is a significant linear correlation between dehydrogenase activities and soil organic matter contents (r = 0.9438*).

Significantly higher cellulase and β-glucosidase activities were observed in compost-treated soils compared with CF and CK soils (Table 5). Both cellulase and β-glucosidase activities showed a significant linear correlation with soil organic matter contents (r = 0.9409* and r = 0.9865** for cellulase and β-glucosidase, respectively). The cellulose in the compost comes mainly from sawdust and is the substrate for cellulase and β-glucosidase activities. The results in the present study agree with those observed in a study on soils amended with sludge containing cellulose (Hattori 1988) or manure with straw (Dick et al. 1988).

Nitrogen mineralization is an important reaction in soils because it is related to plant growth and supplies a sufficient amount of N. Proteases (proteinases, peptidases or proteolytic enzymes) are enzymes that break the peptide bonds between amino acids of proteins and production of free amino acids. The lowest protease activity was observed in the CF-treated and CK-treated soils and the highest was observed in the Rate 4 treatment (Table 5). Even if there was no statistically significant difference in the protease activity among CK, Rate 1 and Rate 2 treatments, a higher protease activity was observed in the plots receiving compost. A significant linear correlation existed between the organic matter content of the soils and the protease activities (r = 0.9048*), meaning that organic matter supplies the substrate for the enzyme. Loll and Bollag (1983) also showed that soil amended with organic compounds such as straw will increase protease activity.

The activity of urease, the enzyme that catalyzes the hydrolysis of urea and which is widely used in the evaluation of changes in soil quality for soil management (Díaz-Marcote and Polo 1995), increased with the application of compost (Table 5). The low urease activity found in the CF treatment (similar to that of the control treatment) was expected because the presence of inorganic forms of N makes the synthesis of the enzyme unnecessary. Higher urease activity probably resulted from an increase in soil organic matter content (Table 2) and microbial population (Table 3), resulting in the secretion of urease, although no urea was applied. Similar results have been reported (Chakrabarti et al. 2000; García et al. 1994; Pascual et al. 1999). A highly significantly linear correlation existed between urease activities and soil organic matter contents (r = 0.9713**).

Higher arylsulfatase activity was found in the soil treated with a higher rate (higher than Rate 2) of compost (Table 5). This result also showed that arylsulfatase activity was significantly and linearly correlated with the organic matter content (r = 0.9246**) of the soils. Deng and Tabatabai (1996, 1997) reported that the activities of glucosidase, phosphatase and arylsulfatase were highly correlated with soil organic C content and suggested that organic matter plays an important role in protecting soil enzymes.

Both acid and alkaline phosphatase activities of compost-treated soils were significantly higher than those in the CF treatment (Table 5). The acid phosphatase activities also showed a significant linear correlation with the application rates of compost (r = 0.9995**). Giuaquiani et al. (1994) reported that phosphatase activity increased when compost was added at rates up to 90 Mg ha⁻¹, and that the activity continued to show a linear increase with compost rates up to 270 Mg ha⁻¹ in a field experiment. Martens et al. (1992) also reported that the addition of organic matter maintained high levels of phosphatase activity in soil during a long-term study. Dodor and Tabatabai (2003) found that higher phosphatase activity in soil resulted from higher organic C contents in the soils. In the present study, there was a significant linear correlation between soil organic matter content and phosphatase activity of the soils (r = 0.9928** and r = 0.9636** for acid and alkaline phosphatases, respectively). Owing to the slight alkalinity of the soil pH in this study the activities of alkaline phosphatase were higher than those of acid phosphatase.

Significant correlations were found between microbial biomass and the measured enzymatic activities (data not shown). Significant positive correlations were also found between each type of enzymatic activity measured in the experiment (data not shown).

Plant growth as related to enzymatic activities

Table 6 shows the fresh yields of vegetables of the last four crops in the different treatments. In general, no fertilization (CK treatment) for 24 crops of vegetable cultivation during three consecutive years resulted in lower yields compared with the yields of the fertilized treatments. There were no significant differences in vegetable yields between the lowest rate of compost treatment (Rate 1) and the chemical fertilizer treatment. Application of compost at a rate higher than Rate 2

Table 6 Effect of different application rates of compost on the growth of vegetables of the last four crops

Treatments	Pak-choi (21st crop)	Lettuce (22nd crop)	Amaranth (23rd crop)	Amaranth (24th)
	Fresh weight (Mg ha^−1)
CF	25.8 b^†	4.1 c	2.4 b	4.3 bc
CK	13.0 c	3.1 c	2.0 b	3.1 c
Rate 1	22.0 b	5.8 b	2.9 ab	4.9 b
Rate 2	27.6 a	6.9 b	3.1 a	7.3 a
Rate 3	29.0 a	10.1 a	3.7 a	8.3 a
Rate 4	27.5 a	11.3 a	3.1 a	8.0 a
^†Within columns, means followed by the same letter are not significantly different (P = 0.05) using Duncan's multiple range test. CF, chemical fertilizer; CK, no application of fertilizer.

inevitably resulted in a significantly higher yield compared with the CF treatment. However, no significant difference was observed in vegetable yield among the Rate 2, Rate 3 and Rate 4 treatments, except for lettuce. Higher content of N, organic matter, available P and microbial activities were found with Rate 3 and Rate 4 compost-treated soils compared with the other treatments (Table 2); however, no further significant increase in vegetable yields could be observed for compost application rates higher than Rate 2. This result shows that the application rate of the Rate 2 treatment is adequate for plant growth once the fertility of the soil is established through the application of compost. However, the same result was not observed during the first year of this experiment, in which the yield of the CF plot was considerably higher than that of the Rate 1 and Rate 2 plots (data not shown). In addition, sufficient amounts of continuously applied compost also resulted in establishing soil fertility and in the enhancement of soil quality (as indicated by higher soil enzymatic activities) and it is not necessary to apply a higher amount of compost in further applications.

The growth of most vegetables is adversely affected in soil with an EC of saturation extraction higher than 4.0 dS m⁻¹ (Rhoades and Miyamoto 1990). However, in the present study no adverse effect was observed in the growth of vegetables under Rate 3 and Rate 4 treatments in which the EC of saturation extraction was higher than 4.0 dS m⁻¹. On the contrary, the plants grew vigorously and obtained a high yield similar to that of the Rate 2 treatment. Therefore, we suppose that a high organic matter content in the soil can alleviate the harmful effect of soluble salts in soil on plant growth.

Conclusions

Even if most of the applied composts decomposed during the 3 years under subtropical climatic conditions, the continuous application of large amounts of compost resulted in a general increase in soil organic matter content, soil microbial population and soil enzyme activities. An application rate of compost higher than a Rate 2 treatment, 540 kg N ha⁻¹ y⁻¹, not only provides no further enhancement of soil microorganism populations and soil enzyme activities, but it also provides no further increase in vegetable yield compared with the CF plot. A high organic matter content in the soil can alleviate the adverse effect of soluble salts on vegetable growth. In conclusion, a compost application rate of 540 kg N ha⁻¹ y⁻¹ (Rate 2 treatment) is adequate on the basis of vegetable yields and soil chemical, biochemical and enzymatic properties for greenhouse cultivation under subtropical climatic conditions.