Molecular Speciation of Phosphorus in Soil Under Various Long-Term Fertilization Regimes

Aims The objectives of this study were to examine the long-term substitution of mineral phosphorus (P) fertilizers with manure (M) plus nitrogen (N) fertilizers and how they affect the forms of P that occur in soil, soil P distribution, and plant growth. Methods We used a solution of 31 P nuclear magnetic resonance ( 31 P-NMR) spectroscopy to study the correlations between long-term fertilization regimes and the forms of P that occur at different soil depths. Then we investigated yield, plant growth, and soil properties. Results A 40-year eld experiment showed that the use of M + N fertilizers can signicantly improve plant growth and yield. The proportion of organic P in the 20-40 cm soil layer was signicantly increased by long-term M fertilization. The concentrations of various forms of P (orthophosphate, pyrophosphate, diesters, monoesters, and total inositol hexakisphosphate, IHP) in topsoil increased signicantly with the combination of M with N + P mineral fertilization. The addition of M greatly increased the stereoisomers of IHP (myo-IHP, scyllo-IHP, neo-IHP, and D-chiro-IHP) and the proportion and concentration of corrected diesters. There were no signicant differences in the pyrophosphate contents of the 40-60 cm soil layer according to fertilization type and year of fertilization. There were also no signicant differences in IHP stereoisomers and diesters according to fertilization year. The P forms that contributed to corn yield were orthophosphate, diester, and IHP. Further, pyrophosphate made no signicant contribution to corn growth. Conclusions Over the long-term, pig manure can signicantly increase the amount of orthophosphate that is directly absorbed by crops and the amount of IHP stereoisomers that can be used by plants. Orthophosphate and IHP are the two key factors that have a positive effect on plant growth. method (Cade-Menun and Liu, 2014). In the present study, we used 31 P-NMR to study soil at depths of 0– 60 cm as part of a 40-year eld experiment. The aim of this study was to investigate the effects of long-term application of N or manure fertilizer, and their use as partial substitutes for mineral P, on the variety and contents of P forms in topsoil and deep soil, on plant growth and crop yield responses.


Introduction
The dynamics of phosphorus (P) in soils play an important role in the global biogeochemical P cycle, which is bene cial to sustainable agriculture (Gilbert, 2009). To ensure crop yields, farmers apply P fertilizer every year, as the utilization e ciency of P fertilizer is low in the current season (Zicker et al., 2018).
However, P resources are non-renewable, and excessive application of P fertilizer increases the consumption of P rock resources (Mayer et al., 2016). Additionally, nutrient losses into water systems cause water quality problems such as eutrophication (Cassidy et al., 2017). Additionally, improper disposal of livestock and poultry manure can cause environmental pollution (Hanifzadeh et al., 2017). Most studies on the effect of P fertilizer on topsoil have ignored the subsoil, which holds immense P reserves that constitute 25-50+% of the P in the soil pro le (Kautz et al., 2012;Koch et al., 2018). Therefore, using manure to replace mineral phosphate fertilizer can protect the environment. The forms of P available in the environment in uence the mobility of P and the ability of plants to obtain nutrients (Li et al., 2018). Thus, it is necessary to clarify the forms of P in soil and the effects of long-term fertilization on P.
China is the greatest consumer of P fertilizer in the world (Tian et al., 2016). In order to slow down the consumption of P resources, other fertilizers are often used as substitutes (e.g. compost), or as supplements to improve the utilization e ciency of P fertilizer (e.g. manure, nitrogen, N, and potassium, K, fertilizers; Yan et al., 2013). N fertilizer is applied in large amounts and plays important roles in enhancing the soil's fertilizer supply capacity and crop yields (Hountin et al., 2000). However, the long-term application of large amounts of N fertilizer degrades the balance of soil nutrients (Li et al., 2007) and affects the content of soil P. Many studies have shown that P fertilizer can stimulate plant growth more strongly in ecosystems with higher N application (Li et al., 2016). Some studies have found no effect of the application of fertilizers on the forms of P (Koch et al., 2018), but others have found that the application of mineral fertilizer or compost can improve the utilization e ciency of P in soil (Yan et al., 2013). Application of livestock and poultry manure to soil can increase soil P content (Ward et al., 2017). Thus, reducing the use of mineral P by partially replacing it with manure and N fertilizers can improve the utilization e ciency of P fertilizer, reduce the depletion of P resources, and bene t the environment.
The distributions of forms of P affect the transformation of P in soil and the absorption of P by plants (Farrell et al., 2015). In general, only soluble orthophosphate anions (H 2 PO 4 − or HPO 4 2− ) can be absorbed directly by plant roots, although there are many forms of P in the soil (inorganic and organic forms; Richardson et al., 2011). The main source of P in most organisms is orthophosphate, but the content of organic P in soil is higher than that of orthophosphate (Abdi et al., 2015). Organic P comprises 5-95% of total soil P and mainly comprises phosphate monoester, phosphate diester, and phosphate (Condron et al., 2005). To clarify the process of soil P cycling, it is necessary to identify and quantify the forms of P in soil. To ensure crop yields, fertilization is usually used to increase soil P (Tilman et al., 2002). Some studies have shown that fertilization for many years can affect the composition of P in the soil (Deiss et al., 2016); for example, mineral P fertilization can increase the proportion of soil orthophosphate (Mcfarland et al., 2013). Conversely, there have been no reports of extreme differences in organic P compounds in soils fertilized with mineral or manure fertilizers (Annaheim et al., 2015).
Analyzing 31 P in soil extracts by nuclear magnetic resonance is a very effective way of obtaining detailed information on soil P forms to understand the distribution of P species in different types of soils (Ward et al., 2017). The most extensively-used method for extracting soils and fertilizers is the NaOH ethylenediamine tetraacetic acid (NaOH-EDTA) method (Cade-Menun and Liu, 2014). In the present study, we used 31 P-NMR to study soil at depths of 0-60 cm as part of a 40-year eld experiment. The aim of this study was to investigate the effects of long-term application of N or manure fertilizer, and their use as partial substitutes for mineral P, on the variety and contents of P forms in topsoil and deep soil, on plant growth and crop yield responses.

Study sites and eld investigation
The study soil was a simply cultivated wet-leached soil developed from Quaternary loess parent material. This is one the main soils found in Liaoning Province, China. According to the United Nations FAO classi cation, the soil was haplic luvisol (FAO, 1998 The samples were collected as part of the long-term stationary brown soil fertilization experiment conducted at Shenyang Agricultural University, Northeast China. This experiment was established in 1979 and involves corn-corn-soybean crop rotation. The experiment was divided into three blocks: unfertilized, mineral-fertilized, and manure-fertilized. Each block was divided into nine plots, 160 m 2 in area (10 × 16 m). The treatments selected for the current experiment were 1) soil obtained before planting in 1979 (F0), 2) controls (unfertilized, CK), 3) mineral-fertilized (N fertilizer or N + P fertilizer; NF, NP), and 4) manurefertilized (manure + N fertilizer or manure + N + P fertilizer; MN, MNP). The mineral fertilizer was comprised of N as urea (46%), P as calcium superphosphate (5%), and K as potassium sulfate (42%). The application rate of N (P) fertilizer was 120 kg ha −  then centrifuged (1500 mg, 20 min) and transferred to NMR tubes of 10-mm diameter. They were stored at 4℃ and, within 24 h, a 10-mm wide-frequency probe was installed with a spectrometer (GE Omega 500 MHz), and the 31 P-NMR spectra were determined (Turner, 2003). A pulse of 90°, detection time of 0.68 s, pulse lag of 4.32 s, and temperature of 20℃ were used; 85% H 3 PO 4 was used as the standard sample, and each sample passed through 8000 scans. Turner et al. (2003) argued that NaOH-EDTA extraction and 31 P-NMR provide the most accurate estimation of soil organic P. In the spectra of all samples, orthophosphate peaks were identi ed at 6 ppm. Additionally, there were peaks of orthophosphate monoester, neo-inositol hexakisphosphate (neo-IHP), D-chiroinositol hexakisphosphate (D-chiro-IHP), myo-inositol hexakisphosphate (myo-IHP), scyllo-inositol hexakisphosphate (scylo-IHP), choline phosphate, αglycerophosphate, β-glycerophosphate, and glucose 6-phosphate. Further, α-glycerophosphate and β-glycerophosphate are products of diester degradation, so the phosphate monoesters were corrected with consideration for these (Schneider et al., 2016). The relative proportions of all compositions in a spectrum were determined by MestRec and ChemDraw software.

Statistical analysis
Statistical analysis involved two-way ANOVAs performed using SPSS 21.0 (SPSS Inc., Chicago, Illinois, USA). 31 P-NMR spectral analysis was conducted using MestReNova software version 11.0. Signi cant differences were determined in terms of pand t-values, and then linear discriminant analysis (LDA) was used to estimate the effects of the characteristics. A simple Pearson correlation analysis was performed to examine the associations between plant yield, soil properties, and P compounds. The 31 P-NMR data were centered-log-ratio-transformed prior to statistical analysis (Abdi et al., 2014).

Results
Soil chemical properties, plant growth, and yield As shown in Table 1, compared with CK, soil pH was signi cantly decreased in the other treatments. We found that compared with CK, the application of N fertilizers (NF and MN) greatly increased the soil total organic C and total N contents. However, application of P fertilizer (NP) did not, while the addition of manure fertilizer (MNP) improved the contents of total organic C and total N. Compared with the other treatments without P fertilizer, the application of P fertilizer (NP) greatly increased soil total P, Olsen-P, and NaOH-EDTA total P. Further, the addition of manure fertilizer (MNP) was extremely effective. There was no great difference in yield between the CK and NF treatments. However, the application of P fertilizer or manure fertilizer (NP, MN, or MNP) signi cantly increased yield. Moreover, the yield was highest in MNP, being seven times greater than in the CK group and ve times higher than in the N group ( Table 2). The soil total N varied between 0.20 g kg − 1 in F0 and 1.60 g kg − 1 in MNP; organic C varied between 8.82 g kg − 1 in CK and 14.10 g kg − 1 in MN; and total P was highest in MNP and lowest in CK. The soil characteristics (pH, total C, total N, total P, and Olsen-P) and NaOH-EDTA total P in the CK treatment were signi cantly lower than in the F0 treatment, except for total N. Compared with CK, the soil characteristics in the other treatments were extremely increased except for pH. We found that, compared with the application of mineral fertilizers (NF and NP), the soil characteristics, NaOH-EDTA total P, and yield were greatly increased by the application of manure fertilizers (MN and MNP; Table 1). The stem diameter, plant height, number of corn cobs, stem and leaf weight, and yield values were in uenced by the fertilizer treatments (Table 2). There were no statistically signi cant differences between the CK, NF, and NP treatments in terms of stem diameter and plant height, but these properties were signi cantly increased by the application of manure fertilizers (MN and MNP). Compared with CK, the number of corn cobs was signi cantly increased by fertilization, with the highest numbers observed in the MNP treatment. Furthermore, the trend for stem and leaf weight was similar to that observed for the number of corn cobs; that is, stem and leaf weight was highest in the MN treatment. 31 P-NMR spectroscopy: identi cation of P forms The 31 P-NMR spectra of total NaOH-EDTA soil extract solutions collected from the six samples at a depth of 0-20 cm is shown in Fig. 1. The 31 P-NMR peaks detected in this study fell between 20 and − 20 ppm. Furthermore, inorganic P forms were detected: orthophosphate (the main peak in all spectra), pyrophosphate, and polyphosphate; as were organic P forms: phosphate monoester, phosphate diester, and phosphates. Phosphate monoester was detected with a 3.5 to 6.7 ppm chemical shift; the 31 P-NMR spectra indicate that phosphate monoesters were dominated by stereoisomers of inositol hexakisphosphate (IHP) in all treatments (Fig. 2). Four peaks of stereoisomers were identi ed as neo-IHP, D-chiro-IHP, myo-IHP, and scyllo-IHP (Fig. 2). There was no peak of Dchiro-IHP in the F0 and CK treatments, and no peak of neo-IHP in the NF and NP treatments. Four peaks of IHP (myo-IHP, scylo-IHP, neo-IHP, and D-chiro-IHP) were found in the manure fertilizer treatments (MN and MNP). In addition, the myo-IHP (phytate inositol) peak was the most frequent IHP peak observed in the six treatments (Fig. 2). 31 P-NMR: distribution of P forms As shown in Table 3, the contents of different P compounds in soil extracts were analyzed by 31 P-NMR spectroscopy. For all treatments, the total inorganic P and total organic P concentrations in topsoil were lowest in the CK treatment and highest in the manure fertilizer treatments (MN and MNP). Compared with CK, the MN and MNP treatments had a 24.4-times greater total inorganic P concentration and 4.4-times greater total organic P concentration. We found that phosphate monoester was the most common form of P in the CK, NF, and NP treatments (mineral fertilization). One of the interesting ndings was that, compared with CK in the same year, the orthophosphate concentration was much higher in the NF and NP treatments, but signi cantly lower in the F0 soil, which was processed 40 years ago (Table 3-4). The concentrations of inorganic P and organic P in the fertilization treatments gradually decreased with increasing soil layer. However, as the soil depth increased, the concentration of inorganic P increased in CK. Interestingly, long-term non-fertilization and only planting resulted in the disappearance of organic P compounds in the 40-60 cm soil layer. The orthophosphate, phosphate monoester, and total IHP differed signi cantly among all treatments: in the 0-60 cm depth soil, the effect ranking was 0-20 cm > 20-40 cm > 40-60 cm; in all other treatments, the effect was MNP > MN > NP > NF > CK. However, the concentration of phosphate diester in the fertilization treatment also followed this pattern, but under the CK treatment, the concentration in the 20-40 cm soil layer was greater than in the 0-20 cm soil layer. Furthermore, the pattern of change in the corrected phosphate diester concentration changed became consistent with the patterns observed for the other P forms. In addition, fertilization greatly increased the concentration of IHP stereoisomers. The corrected M/D ratio (phosphate monoester/phosphate diester) varied greatly from the uncorrected phosphate diester degradation, and the concentrations of phosphate diester in all treatments were signi cantly higher (Table 3). After correction for phosphate diester degradation, the contents were signi cantly increased in all treatments except for the ones receiving manure fertilizer (MN and MNP). However, there were no signi cant differences in the corrected ratios of phosphate monoester/phosphate diester.
According to the 31P-NMR peaks of P compounds detected in this study, the range of variation in the concentration of soil P forms over time was 20 to -20 ppm (Fig. 1). Total inorganic P percentages ranged from 47.4-79.2%, mostly as orthophosphate (40.7-76.1%). The total organic P percentages ranged from 20.8-52.6%, mostly as phosphate monoester (20.2-48.6%). Hence, their concentrations showed the same trend. In conclusion, the percentages and concentrations of total inorganic P were higher than those of total organic P, and the P form with the highest percentage and concentration was orthophosphate. With time, the concentrations of all P forms increased signi cantly. Compared with 1979, the percentages of total inorganic P and orthophosphate were much lower in 2006, while the percentages of other P forms increased signi cantly with time. The percentages of total inorganic P and orthophosphate increased with soil depth, but their concentrations exhibited the opposite trend. The percentages and concentrations of total organic P, phosphate monoester, and total IHP decreased with soil layer depth. The total uncorrected phosphate diester concentration was low at the 0-40 cm soil depth and lowest at 40-60 cm. However, after correction, the percentage and concentration of diester increased when time was increased, then decreased when soil depth was further increased.

Variations in soil P forms with time and depth under different fertilization regimes
The concentrations of orthophosphate and phosphate monoester in the topsoil (0-20 cm) of the manure fertilizer treatments (MN and MNP) were signi cantly higher than those of the other treatments. The concentrations of orthophosphate and phosphate monoester decreased with soil depth but, at the 20-40 cm soil depth, were still signi cantly higher in the MN and MNP treatments than the other treatments (Fig. 3). In addition, orthophosphate was highest Compared with the MN treatment, phosphate monoester was lower in the treatment that also had P fertilizer (MNP; Fig. 3). There was an extreme interaction of year, fertilization, and year × fertilization for pyrophosphate at the 0-40 cm soil depth (P < 0.05; P < 0.01), but there was no signi cant difference at the 40-60 cm soil depth (Table 5). Compared with the other treatments, the concentration of pyrophosphate was much higher in the MNP treatment. For all treatments, the concentration of pyrophosphate decreased with soil depth between 0-40 cm, particularly for the MNP treatment, while the concentrations were similar for all treatments at 40-60 cm (Fig. 3).
The P-compound classes in the NaOH-EDTA topsoil (0-20 cm) extracts (as detected by solution 31P-NMR) included orthophosphate (5.8-6.0 ppm), and there were several resonances of soil phosphate monoester (3.5-6.7 ppm) and pyrophosphate (− 4.1 to − 4.7 ppm; Fig. 2). We found that the addition of manure fertilizer markedly changed the P compounds in the soils. Generally, phosphate monoester was the most abundant P component in the extracts of the CK, NF, and NP treatments in 2006, comprising 55.6-71.4% of the total extractable soil P. However, with manure fertilizer addition (MN and MNP) or in the original soil (F0), orthophosphate was the most abundant P compound, comprising 66.7-90.0% (Fig. 4). With time, the orthophosphate concentration increased greatly with the addition of N or P fertilizer, and the effect of manure fertilizer was even more signi cant. Compared with the CK treatment, N fertilizer addition greatly increased the concentration of phosphate monoester, but there was no signi cant difference in phosphate monoester between the NF and NP treatments. In addition, the concentrations of phosphate monoester in the MN and MNP treatments were signi cantly higher than in the NF and NP treatments. However, after manure fertilizer addition, there was no signi cant difference in phosphate monophosphate concentration with P fertilizer addition (MN and MNP). The pyrophosphate concentration was not affected by the year or by mineral fertilizer addition (N and P). The pyrophosphate concentration in the MNP treatment was highest in 2006 and much lower in 2018. In 2018, the concentrations of orthophosphate, phosphate monoester, and pyrophosphate in the CK treatment were much lower than those of the F0 soil in 1979 (Fig. 4).
Based on data from 2006 and 2018, the trends in Pi/Po were the same among the different treatments (Fig. 4) (Fig. 4). Soil total P, Olsen-P, orthophosphate, phosphate monoester, and pyrophosphate were signi cantly affected by soil, year, fertilization, soil depth, and their interaction (P < 0.001, Table 5). Only total IHP and phosphate diester were not affected by year, and phosphate diester was not signi cantly affected by the year × depth interaction. Soil total organic C was signi cantly affected by year and depth and their interaction (P < 0.001).
Correlations among P compounds and soil properties, plant growth, and yield The contents of total inorganic P, total organic P, orthophosphate, pyrophosphate, monoesters, diesters, and IHP were positively related to soil properties, plant growth, and yield (Table 6). Furthermore, all P compounds were signi cantly positively correlated with total N, Olsen-P, and stem diameter (P < 0.05; P < 0.01).
In addition, all P compounds were signi cantly positively correlated with total C, total P, plant height, and number of corn cobs, expect for pyrophosphate (P < 0.05; P < 0.01). Among all the P compounds, the contents of total organic P, monoester, and diester were signi cantly correlated with stem and leaf weight (P < 0.05). Among all the P compounds, the content of pyrophosphate and monoester were signi cantly correlated with yield (P < 0.05).

Discussion
Plant growth, yield, and soil characteristics In the present study, the soil pH decreased was decreased in all treatments, except for F0. Treatment NF had the lowest pH, and with the application of manure fertilizer, the pH decreased slowly (Table 1). This is consistent with other reports suggesting that the application of mineral fertilizer signi cantly reduces soil pH ( (Table 2). In the current study, higher plant height and thicker stem diameter were found after the manure fertilizer treatments (MN, MNP), but there was no signi cant difference between the unfertilized and mineral fertilizer (NF, NP) soils, which is inconsistent with previous research suggesting that fertilizer application increases the height and stem of plants  (Table 6). Because inorganic P (the main component of inorganic P is orthophosphate) is present primarily in the corn fruit, but there is only a small amount in the stems and leaves. However, pyrophosphate was signi cantly correlated with the stem diameter. Because pyrophosphate had a signi cant correlation with total N, and a larger N may be associated with an increased plant stem diameter (Puntel, 2012).

Orthophosphate in soils
There were signi cant differences in the concentrations of orthophosphate between fertilization treatments, soil depths, and fertilization years (Fig. 3-4). The trends observed in this study are related to the direct addition of orthophosphate to the soil as a result of fertilization. One interesting nding is that over the years of fertilization, the proportion of orthophosphate decreased gradually, while the concentration increased gradually (Table 3). Schneider et al. (2016) con rmed that chemical fertilizers increase the content of orthophosphate. The higher the concentration of orthophosphate, the higher the Olsen-P concentration and yield (Table 1-2). In addition, the content of orthophosphate is positively correlated with the concentration of soil P that plants can absorb and utilize, and with grain yield (Schneider et al., 2016; Xin et al., 2019). Thus, we conclude that the content of orthophosphate was greatly increased by the use of N fertilizer, and the addition of P may enhance this effect ( Table 3). The primary cause is a shift in the dominance of the major P compounds in the soil due to P addition. Speci cally, added P changes the dominant P form from phosphate monoester to orthophosphate. This is signi cant because it can be provided to plants in relatively small amounts for a long time. This is because inositol compounds are the main components of phosphate monoesters (Ahlgren et al., 2013), while orthophosphate is readily and immediately available. The effect of organic fertilizer was better than that of mineral fertilizer because there was a large amount of orthophosphate in the mineral fertilizer, and the content of orthophosphate is indirectly increased by the presence of a large number of microorganisms (Ahlgren et al., 2013). Therefore, orthophosphate concentrations in soil may also increase with high P-fertilization rates, whereas forms of organic P can increase at low P-fertilization rates (Barbara et al., 2017).

Pyrophosphate in soils
In the present study, the P compound with the third-highest content was pyrophosphate, which can be absorbed and utilized by plants and is also a reactive species (Barbara et al., 2017). In general, the concentration of pyrophosphate was low in all treatments, soil depths, and years ( Fig. 3-4 . Interestingly, we found that the content of pyrophosphate was signi cantly higher in the 0-40 cm soil layer than in the other layers, and that manure combined with P fertilizer had a signi cant effect on the pyrophosphate content (Table 4 and Fig. 3). The addition of P (organic or chemical P fertilizer) is an e cient way to increase the amount of P that is absorbed and utilized by plants. There were no signi cant differences in the composition of organic P in the soil according to the use of combined organic plus mineral fertilizer or mineral fertilizer alone (Ahlgren et al., 2013). Other studies have also found that pyrophosphate does not seem to be affected by P addition to soil. Additionally, there are many microorganisms in manure, and pyrophosphate is generally associated with microbes. The present study found that turning maize stubble from the previous year into the soil could increase oxygen and water retention and microbial activity; thus, rapid decomposition could explain the low overall level of pyrophosphate in the investigated soil. Several other studies have con rmed that to be readily used by plants, the composition of extractable P should be changed, such as by adding P to the soil (Watson et al., 1998).

Effects of long-term fertilization on the forms of organic P at different depths and on plant characteristics
The present study revealed that the main form of organic P in the soil was phosphate monoester. The low amounts of phosphate diester in the soil are consistent with the ndings of other studies (Table 2-3; Fig. 2; Yang et al., 2019). In addition, the content of total organic P increased each year, and the proportion of total organic P increased gradually at the 20-40 cm soil depth but decreased at the 40-60 cm soil depth (Table 3). It could be that long-term fertilization increased the proportion of organic P at the 20-40 cm soil depth. It has been previously reported that fertilization can cause organic P to accumulate in deep soil (Guardini et al., 2012; Ahlgren et al., 2013; and references therein). We observed that unstable inorganic P seems to accumulate in the topsoil and decreases with depth, while organic P occurred deeper in the soil pro le. As a result, losses of soluble inorganic P in surface runoff will increase . Furthermore, the results of this study indicate that the composition of the organic P compounds in soil is in uenced both by whether the soil is fertilized or unfertilized, and by fertilizer type. For example, the IHP stereoisomers in different fertilization treatments varied (Fig. 2). This is different from the results of Dodd et al. (2014), who found that the diversity of organic P compounds in the soil was not affected by the type of fertilization or farming method (Song et al., 2011). This may be related to different soil types and environments. However, some studies have reported a signi cant difference in the organic P content of soils with a lot of mineral or manure fertilizers (Motavalli and Miles, 2002). The addition of manure fertilizer (MN) greatly increased the concentration of total organic P in soil, but this decreased with the addition of P fertilizer (MNP) ( Table 2). This is because the combination of manure plus P fertilizer increased soil P, while organic matter has a xed ratio of C, N, and P. Therefore, the concentration of organic P decreased, but the present study was conducted in a temperate humid-semi-humid monsoon climate, so the decrease in organic P was relatively weak (Table 1- . Thus, it was necessary to reduce the single dose of chemical P fertilizer in soil and partly supplement it by applying N plus manure fertilizer. Soil available P, plant growth, and grain yield were increased when the forms of organic P that were applied became degraded (Table 1-2). Our results indicated that diester was signi cantly correlated with plant growth and grain yield. In contrast, monoester was signi cantly correlated with plant growth but was not signi cantly correlated with grain yield (Table 6). This indicates that monoester contributes more to the corn growth process and does not make much of a contribution to the corn fruit when the crop is mature. This result is in contrast to the ndings of other studies (Wei et al., 2014). Furthermore, IHP was signi cantly correlated with plant growth and yield, except for the weight of the stems and leaves. This may be because IHP is mainly related to the microorganisms in the soil, and an increase in organic matter and N will increase the microbial biomass, which can increase the content of hormones and humus in the soil, and increase the yield of crops relative to the weight of the stems and leaves; thus, IHP is a P source that crops can absorb and utilize (

Phosphate monoester in soils
Overall, long-term fertilization and soil depth had signi cant effects on the forms of P in soil, with inorganic P accumulating on the soil surface and decreasing in content with depth. However, organic P (total organic P content, total IHP, phosphate monoester, and diester) accumulated in the deep soil layer (20-40 cm) over the study period. We observed that the content of phosphate monoester was high in the topsoil (Table 3; Fig. 3), which seems reasonable because, compared with the subsoil, the concentration of bacteria and fungi in deep soil is higher and the predominating aerobic conditions facilitate rapid microbial turnover, thus supporting the degradation of organic fertilizer. Furthermore, the degradation products of the microbial biomass are important sources of phosphate monoester in soils (Bünemann et al., 2004).

Total IHP in soils
We found that IHP monoesters and IHP stereoisomers were the most identi ed forms of organic P, with four peaks being identi ed. There were no neo-IHP peaks in the NF and NP treatments, indicating that long-term single application of mineral fertilizer will lead to the disappearance of neo-IHP, which is contrary to Abdi's (2014) conclusion. There were four IHP peaks (myo-IHP, scyllo-IHP, neo-IHP, D-chiro-IHP) observed with manure fertilization (MN and MNP). In addition, compared with CK, the pig manure treatments had greatly increased contents of IHP stereoisomers (myo-IHP, scyllo-IHP, neo-IHP, D-chiro-IHP). Because pig manure promotes microorganisms and their activities (Wong et al., 2009), scoyllo-IHP and D-chiro-IHP may be derived from microorganisms, and D-chiro-IHP may also be derived from plants (Xin et al., 2019). myo-IHP was the most dominant form in all treatments (Liu et al., 2013b). The content may be overestimated from interpretation of the NMR spectra, or it may be realistic. Since crops have been planted, there will be plant tissue input, so the content of this compound in the soil is high (Giaveno et al., 2010). In many studies, it has been robustly con rmed that myo-IHP is mainly derived from seeds and is widespread in plants. (Noack et al., 2014). In the F0, CK, NF, and NP treatments, myo-IHP predominantly originated from the input of the crop (seeds, roots, and residues), while in the MN and MNP treatments, part of the myo-IHP came from the pig manure fertilizer. Part of the phosphate monoester in the soil comes from external inputs, which is good evidence that myo-IHP is related to animal manure (Xin et al., 2019). Further, the application of pig manure tends to increase myo-IHP concentrations. This is contrary to other research results (Annaheim et al., 2015). This difference may be because this previous study used mineral fertilizer and cow dung. Many studies have shown that IHP is a compound with great a nity for amorphous metals and soil particles (Yan et al., 2014). In particular, the a nity of myo-IHP towards goethite is higher than that of orthophosphate (Yan et al., 2015). This indicates that myo-IHP is available to plants (Celi et al., 1999). Recent research has shown that because many microorganisms play vital roles in the environment, these ubiquitous compounds can play a role in IHP (Turner et al., 2007). In the long run, IHP can serve as an effective and stable source of nutrients for plants, although there is a relatively stable P pool in the soil (Richardson et al., 2011).

Phosphate diester in soils
In the present study, there were low amounts of phosphate diester compared with other forms of soil P. These results are consistent with other studies. At pH > 5, the adsorption capacity of soil particles for DNA, which is one of the forms of phosphodiester with the highest content, decreases. Therefore, these ndings may be caused by soil pH (Khanna et al., 1998). This may mean that phosphate diester is being ushed out and depleted (Ahlgren et al., 2013). Additionally, previous studies have reported that the absence of phosphate diester in agricultural soils according to NMR spectra  may be because it is degraded into phosphate monoester during measurement and extraction (Turner et al., 2003). In this study, this phenomenon may have been signi cant because of the 18 h extraction step for chemical fertilizer, which may have promoted such degradation (Cade-Menun and Liu, 2014). In this study, the concentration of phosphate diester was higher in the topsoil (Table 3). This was due to the higher accumulation of organic matter on the soil surface (Condron et al., 2005), so phosphate diester synthesis was higher. It may also be present naturally in soil.
Generally, the M/D ratio is a measure of instability (Schneider et al., 2016). Some researchers believe that phosphate diesters mineralize faster in soils that are conducive to decomposition, while phosphate monoesters are more tightly adsorbed and do not easily mineralize. Therefore, a low M/D value can indicate a decreased mineralization rate, resulting in a high concentration of phosphate diester (Schneider et al., 2016). Furthermore, the M/D ratio may have been arti cially increased so that it limited the mineralization index of phosphate diester. The data in Table 3 show that the uncorrected phosphate diester content was obviously underestimated due to hydrolysis and degradation. Several studies have also shown that the content of phosphate diester on beans will increase signi cantly after correction; the crops used in this study were maize and soybean.

Conclusions
This experiment sampled soil at three depths under different fertilization regimes over 40 years. The samples were analyzed by spectroscopy. Different fertilization regimes were found to in uence soil IHP stereoisomers; manure fertilization greatly increased the stereoisomers of IHP (myo-IHP, scyllo-IHP, neo-IHP, and D-chiro-IHP). Long-term fertilization greatly increased the concentrations of soil P forms (orthophosphate, pyrophosphate, phosphate monoester, and diester) at the soil surface. A combination of pig manure plus N and P mineral fertilizers greatly increased the proportion of orthophosphate and the concentrations of all forms of P (except pyrophosphate) in the deep soil layer. There was no signi cant difference in pyrophosphate in the 40-60 cm soil layer according to fertilization treatment or fertilization year. The fertilization year had no signi cant effect on IHP stereoisomers and phosphate diester. The soil P cycle changes very rapidly but is very complex; therefore, in order to clarify the mechanism of interaction between P species, further research is needed.
Factors such as soil properties, tillage methods, environment, and time should be considered to minimize the loss of P. From an environmental standpoint, the present study demonstrates that using N plus manure fertilizer to replace some P fertilizer may increase the content of soil-available P that is absorbed and utilized by plants, increase the grain yield and plant growth, and reduce the consumption of P resources and water pollution. Tables Table 1 Soil parameters and NaOH-EDTA total P in topsoil as affected by different treatments.   Table 3 Distribution of P forms concentration by 31 P nuclear magnetic resonance ( 31 P-NMR) spectroscopy in 0 -60 cm depth of soils from different treatments.   Table 5 Analysis of variance for the effects of years, fertilizers, and depth on soil total C, total P, Olsen-P and soil P forms determined by 31 P NMR spectroscopy.