Experimental Approach Alters N and P Addition Effects on Leaf Traits and Growth Rate of Subtropical Schima superba (Reinw. ex Blume) Seedlings

Abstract

Nitrogen (N) and/or phosphorus (P) addition has controversial effects on tree functional traits and growth; however, this experimental approach may clarify these controversial results. In this study, field and pot experiments were designed with +N (100 kg N ha⁻¹ yr⁻¹), +P (50 kg P ha⁻¹ yr⁻¹), +NP (100 kg N plus 50 kg P ha⁻¹ yr⁻¹), and a control (no N or P addition) to comparatively investigate the effects of N and P addition on 24 leaf traits and the growth rate of Schima superba (Reinw. ex Blume ) seedlings in subtropical China. We found that the experimental approach alters N and P addition effects on leaf traits and tree growth. Nitrogen addition strongly altered leaf biochemical and physiological traits and limited tree growth compared to P addition in the pot experiment, while the effects of N and P addition on leaf traits and tree growth were weaker in the field, since the seedlings might be mainly limited by light availability rather than nutrient supplies. The inference from the pot experiment might amplify the impact of N deposition on forest plants in complicated natural systems. These findings will help guide refining pot fertilization experiments to simulate trees in the field under environmental change. Future directions should consider reducing the confounding effects of biotic and abiotic factors on fertilization in the field, and refinement of the control seedlings’ genetic diversity, mycorrhizal symbiont, and root competition for long-term fertilization experiments are required.

1. Introduction

Nitrogen (N) and phosphorus (P) are among the most limited nutrients which play an important role in regulating the carbon (C) cycle [1,2,3]. Recently, anthropogenic atmospheric N deposition (33 kg N ha⁻¹ y⁻¹) has dramatically increased in subtropical forests [4,5], intensifying the imbalance of N and P input [6,7], and influencing traits related to growth, development, defense, and life history. For instance, unchanged leaf morphological traits that respond to simulated N deposition or P input provide an important mechanism for adapting to N-rich and P-poor conditions [8,9]. In a 10 year N and P addition experiment, the leaf photosynthesis rate, P, and chlorophyll of two understory species were altered by P addition, suggesting that subtropical forests might be P limited [10]. Moreover, no response of soluble protein, nonstructural carbohydrates, or free amino acids to N addition indicated that understory plants in reforested forests may be limited by light availability rather than N [11]. However, plants evolved to have various leaf traits (morphological, biochemical, and physiological traits) for use, allowing them to adapt to a wide range of ecological gradients [12,13], but few studies reported that multiple leaf traits responded to N and P addition simultaneously. Therefore, assessing the effects of N and P inputs on leaf traits and tree growth has been a central issue for testing the limitation of net primary productivity (NPP) in the context of global change.

Thus far, many studies have tested the hypothesis that P rather than N limits NPP in the highly weathered soils of humid subtropical or tropical forests. First, P limitation is an important basis for biogeochemical niche partitioning [14], which is widespread at the level of individual species in tropical forests [15]. Second, many studies reported that P addition increased small tree growth in subtropical or tropical forests, which is consistent with the P-limitation hypothesis [10,16]. Third, P concentrations in plant tissues (leaf, branch, and root) and the light-saturated photosynthesis rate generally increased with P addition [17,18,19] but decreased with N addition [8]. However, plant responses that were stronger with P addition and weaker with N addition were not supported in some cases. In tropical or subtropical forests, a meta-analysis showed that P addition did not affect the relative growth rate (RGR) in old growth forests, while leaf N and P concentrations and RGR were significantly affected by N and P addition in secondary tropical forests [20]. Furthermore, a global meta-analysis revealed that N addition might increase NPP via affecting traits related to photosynthesis, including total leaf area, leaf photosynthetic rate per area, and transpiration rate. These results indicated that the effects of N and P may vary with the forest environment, because tropical and subtropical forests are characterized by extreme biogeochemical heterogeneity due to large variations in soil type and plant species composition across the biome [21]. Therefore, much more work is still needed to examine the effects of N addition, P addition, and their interaction on subtropical forest plants.

Various biological (tree species, size classes, and mycorrhizal type) factors might influence fertilization effects in NP addition experiments. However, the experimental approach (pot or field condition) is often ignored. Pot and field experiments are both important for exploring the ecological responses of terrestrial plants to environmental changes [22,23]. Field experiments, as the bridge between the laboratory and natural conditions, can quantify the functions of the changes in natural ecosystems. By contrast, pot experiments are very enlightening to the field experiment, because they have great advantages in controlling exogenous conditions, such as light, soil, water, and genetic uniformity, but these experimental conclusions often ignore the interaction between canopy and understory plants and their adaptation feedback. Recently, there has been accumulating evidence that the effect of fertilization on plant functional traits differs in the field [24,25] and in artificial controlled (pot) experiments [26,27] because there are unavoidable differences in the abiotic and biotic factors between lab and field experiments [28]. In addition, some evidence for strong responses in lab experiments and unapparent responses in field experiments under fertilization has been verified in many manipulative experiments [8,29,30]. Therefore, these divergent responses in experimental approaches increase uncertainty in predicting the impact of global change on forest vegetation. It is necessary to conduct simultaneous field and pot experiments to evaluate the effects of N and P.

To understand the fundamental mechanism that drives variation in leaf traits and tree growth under N and P additions, two-year field and pot experiments with N and P addition were initiated using a random block design. As one of the dominant species in the subtropical forest of south China, Schima superba was selected as the model species, and its seedlings were used to intensively study leaf morphological, biochemical, and physiological traits. We hypothesized that (1) the P addition effect on leaf traits and seedling growth may be greater than that of the N addition effect, since subtropical forests are generally a poor P and rich N ecosystem; (2) N and P addition effects will be weaker in the field than in pot experiments. This study also explored the relationships among morphological, biochemical, and physiological traits, especially between leaf traits and tree growth. Overall, the manipulative experiments of N and P addition might help us to understand and forecast the adaption of subtropical trees under atmospheric N deposition and soil P limitation.

2. Materials and Methods

2.1. Experimental Design

2.1.1. Experiment 1 with a Field Approach

A field N and P fertilization experiment was conducted on a secondary broadleaf evergreen forest (>40 years of age) with a vertical structure, including a canopy tree layer (diameter at breast height, DBH > 5 cm and height > 5 m), an understory layer of saplings, shrubs, and seedlings (DBH < 5 cm and height < 5 m), and a ground-cover layer with ferns and herbs, which was located at Jiulianshan Ecological Station (JLSES, 24°29′18″–38′55″ N, 114°22′50″–31′32″ E, mean elevation 430 m), a part of the Chinese Forest Ecosystem Research Network in Jiangxi Province, China [31]. The average density and basal area of the trees were about 1100 trees ha⁻¹ and 25 m² ha⁻¹, respectively. S. superba (33%), Castanopsis fargesii (22%), C. carlesii (6%), and Machilus breviflora (5%) were the dominant species and accounted for about 70% of the total aboveground biomass of trees. In the understory layer, the seedlings of S. superba were abundant. Most local soils are red earth (Hapludult Ultisols, the U.S. soil taxonomy system), which forms form sandstone and slate regolith [4]. The soil (0–20 cm) had a total C content of 23.1–45.0 g kg⁻¹, N content of 1.7–3.8 g kg⁻¹, and total P content of 0.60–0.96 g kg⁻¹. The average of leaf N, P and N/P in these forests were 18.8 mg g⁻¹, 0.96 mg g⁻¹ and 19.6, respectively [31].

In 2015 and 2016, twelve 20 m × 20 m plots were established and randomly assigned to N, P, and N plus P fertilizer treatments (+N, +P, and +NP, respectively) and a control (CK) using a randomized complete block design (n = 3 repetitions × 4 treatments = 12 plots). All measurements were restricted to the center of each plot to reduce edge effects; each plot was at least 20 m from other blocks to avoid interference and contamination. In addition to the three CK plots, the +N, +P, and +NP treatments were applied with 100 kg N ha⁻¹ yr⁻¹ as NH₄NO₃, 50 kg P ha⁻¹ yr⁻¹ as NaH₂PO₄, and both together, respectively. Every six months, all fertilization treatments used weighed fertilizer mixed with 16 kg of clean and dry fine sand, which was applied to each plot after initial investigations of the corresponding plots in 2016. The CK plots received fine sand without N and P addition. Each plot was divided into 4 (10 m × 10 m) subplots to ensure a more uniform application of N and/or P within a plot.

Fertilizer amounts were determined by our previously simulated N deposition and P addition study which belongs to the core distribution of Chinese fir (100 kg N ha⁻¹ year⁻¹ and 50 kg P ha⁻¹ year⁻¹) and the center of N deposition in China with 49 kg N ha⁻¹ year⁻¹ [4], the ratio of nitrogen fertilizer to phosphate fertilizer is 2:1. We simulated a doubling of natural N deposition; meanwhile, a high dose of P input was used to understand whether exogenous P input would change the effect of nitrogen deposition on forest plants.

In 2018, soil pH and NO₃⁻-N were not affected by fertilization, while soil NH₄⁺-N was higher in plots treated by N addition than in plots treated with P addition. Available P was higher in plots treated with P addition (our unpublished data). Seedling (1 × 1 m) quadrates were conducted at the central point of each subplot. The relative growth rate (RGR, mm yr⁻¹) was measured as the change in the ground diameter of S. superba in two years (2017–2019). In the understory, the average understory shade was 20.42 ± 13.88 µmol m⁻² s⁻¹ calculated by diffuse-light photosynthetic photon flux (PPF) on an overcast sky in August (9:00–11:00 am) [32]. In August 2019, at least three seedlings of individuals of S. superba with a height of c. 1.5–2 m and ground diameter (c. 10–40 mm) were selected for the measurement of growth performance and leaf traits in each plot.

2.1.2. Experiment 2 with a Pot Approach

To control the external conditions, including water, light, nutrients, and uniform competition, a pot N and P fertilization experiment was conducted in an open-air greenhouse at the Jiangxi Agricultural University Garden in Nanchang, Jiangxi Province, China (28°45′ N, 115°49′ E). To maintain consistency in the soil environment, the soil was collected from the same stand in the Jiulianshan Ecological Station, Jiangxi, China. Stones were removed and sieved to pass through a 2 mm mesh. The soil had a total C content of 45.5 g kg⁻¹, N content of 2.86 g kg⁻¹, total P content of 0.81 g kg⁻¹, soil NO₃⁻-N of 8.82 mg kg⁻¹, soil NH₄⁺-N of 0.49 mg kg⁻¹, available P of 7.42 mg kg⁻¹, and pH of 4.31. In November 2017, 1-year-old seedlings of S. superba were rooted and planted in plastic plots (10 L). The plants were cultivated in the greenhouse for 4 weeks. Plants with a similar height (c. 20–25 cm) and ground diameter (c. 8–10 mm) were selected and assigned to 4 treatments with 6 plants in each group (n = 6 repetitions × 4 treatments = 24 pots). The 4 treatments were CK, +N, +P, and +NP, and similar dosages to our field N and P fertilization experiment were used. NH₄NO₃ or NaH₂PO₄ was dissolved in tap water and then slowly added to the pot each month; N and P fertilizers were applied 24 times over two years. All plants were irrigated with the same amount of water. The fertilization treatments were maintained throughout the growing season until August 2019. After measuring photosynthesis, plants were harvested and separated into above-ground and below-ground parts and washed carefully to remove the soil. Then, RGR of S. superba was measured as the change in the ground diameter of S. superba in two years (2017–2019). In pot conditions, the average PPF was 405.42 ± 73.88 µmol m⁻² s⁻¹ on an overcast sky in August (9:00–11:00 am). In August 2019, each seedling was selected for the measurement of photosynthetic performance and leaf traits.

The soil organic carbon (C) was determined by the Walkley–Black wet oxidation method following the removal of carbonates by acid pretreatment. Soil N and P were determined by a Smartchem140 analyzer (WESTCO Scientific, Brookfield, USA) after the samples were digested with an 18.4 mol L⁻¹ H₂SO₄ solution. Concentrations of mineral N (NH₄⁺-N and NO₃⁻-N, KCl-extractable) and available P (NaHCO₃-extractable) were colorimetrically analyzed with a Smartchem140 analyzer (WESTCO Scientific, Brookfield, USA). Soil pH was measured with a digital pH meter in a suspension with soil to water mass ratio of 1:2.5.

The dosage (100 kg N ha⁻¹ yr⁻¹, 50 kg P ha⁻¹ yr⁻¹) and duration (two years) of fertilization were fixed in both experiments, but addition frequencies were not the same (field, twice a year; pot, 12 times in one year). In fact, the above-ground net primary productivity (ANPP) was minimally affected by different N forms and addition frequencies but was affected by the critical N dosage [33]. In addition, high fertilization frequencies were widely applied in pot experiments due to their advantages in diminishing nutrient leaching caused by frequent watering.

2.2. Measurements of Leaf Photosynthetic Parameters

In August 2019, photosynthetic measurements were recorded with a Li-Cor 6400 portable gas exchange system (Li-Cor, Lincoln, NE, USA). The standard environmental conditions of the chamber block were established as a temperature of 25 °C and ambient CO₂ concentration of 400 μmol mol⁻¹. Light response curves were developed for S. superba by changing the PPF in 12 steps (1500, 1200, 1000, 800,500, 300, 200, 100, 50, 20, and 0 µmol m⁻² s⁻¹). The light compensation point (LCP) and light saturation point (LSP; µmol m⁻² s⁻¹) were calculated by fitting a rectangular hyperbolic model to the relationship between irradiance and net photosynthetic rate [34]. CO₂ response curves were measured after the measurement of light response curves and quantum flux density at 1000 and 600 μmol m⁻² s⁻¹, which were saturated to photosynthesis for field and pot seedlings, respectively. Ambient CO₂ levels were prepared in a sequence of 400, 300, 200, 100, 400, 600, 800, and 1000 μmol mol⁻¹, and steady-state rates of net assimilation were recorded at each CO₂ concentration. All measurements were made on healthy, fully expanded, intact leaves near the top of the seedlings. The leaves were exposed to saturating light for 8–10 min before measurements were taken to ensure photosynthetic induction and stomatal opening [10,35]. The light-saturated photosynthetic rate (P_max) was defined as an ambient CO₂ concentration of 400 μmol mol⁻¹ at saturated light intensity, which was measured between 8:00 and 11:30 am. Meanwhile, water use efficiency (WUEi) was calculated as the ratio of the rate of carbon assimilation (net photosynthesis) to the transpiration rate. The photosynthesis model of Farquhar et al. (1980) was fitted to the net assimilation vs. intercellular CO₂ (C_i) response curves [36], the maximum rate of carboxylation (V_cmax). The maximum rate of electron transport (J_max) was estimated as described in Niinemets et al. (2005) [37].

2.3. Measurements of Leaf Other Traits

After measurements of photosynthetic performance, leaves were put into a sealed plastic bag, placed in an ice chest, and then transported to the laboratory. Samples were divided into four parts. One part of each sample was used to determine leaf length, width, area, and leaf mass per unit area (g m⁻²) by calculating the ratio of leaf dry weight divided by the leaf area [38]. Then, the leaf samples were dried to a constant weight at 65 °C for 72 h and were then ground to a fine and homogeneous powder. The other parts of the samples were divided into three equal parts; each part of each sample was cut into discs, approximately 0.2 g of fresh leaf disc samples per tube, and placed in 2 mL microfuge tubes, which were stored at −20 °C before the measurement of starch plus soluble sugar (NSC) and photosynthetic pigment content (chlorophyll).

Leaf N and P were determined by a Smartchem140 analyzer (WESTCO Scientific, Brookfield, CT, USA) after the samples were digested with an 18.4 mol L⁻¹ H₂SO₄ solution. Leaf concentrations of chlorophyll were extracted by 80% acetone and determined by absorbance with a SpectraMax M3 spectrophotometer (Molecular Devices, San Jose, CA, USA) measured at 470, 663, and 646 nm wavelengths, respectively [39]. Starch and soluble sugar were extracted using a reagent kit (Solarbio, Beijing, China), and determinations were based on absorbance at 625 nm using the SpectraMax M3. Amino acids were extracted using a ninhydrin colorimetry reagent kit (Solarbio, Beijing, China), and determinations were based on absorbance at 570 nm using the SpectraMax M3.

2.4. Data Analysis

Modeling of linear mixed effects was conducted to test the effects of N addition, P addition, and their interaction on seedling traits. Fixed effects included N addition, P addition, and their interaction, while different sites were considered random effects. Linear mixed effects model analyses were performed in R version 3.5.2, using the lme, lme4, lmerTest and MuMIn packages.

One-way analysis of variance, followed by multiple least significant difference comparisons, was used to compare the indicators above for different fertilization treatments in each site. An independent t-test was used to compare the indicators above between the field and pot conditions. Linear regression analyses were performed to establish the correlation and significance of relationships between Pmax and leaf traits of S. superba—including LMA, C/N, N/P, photosynthetic N use efficiency (PNUE), photosynthetic N use efficiency (PPUE), LSP, Vcmax, Jmax, and RGR—under field and pot experimental N and P addition, respectively. The statistical analyses were performed in SPSS software; the normality and homoscedasticity of variables was assessed by using quantile–quantile plots and Bartlett’s test. All the results reported were significant at p < 0.05. Multivariate associations of 24 leaf traits and relative growth rates were analyzed with principal component analysis (PCA), using standardized values of S. superba under field and pot experimental N and P addition.

3. Results

3.1. Leaf Photosynthetic Performance

With the pooled data of the pot and field experiments, the linear mixed model indicated that Pmax was significantly affected by N addition. Vcmax was significantly altered by N addition and P addition. Jmax was significantly affected by N addition, P addition and their interaction. LSP was affected by the interaction between N and P addition (

Table 1. Significance (p-values) and model statistics from linear mixed-model analysis for light-saturated photosynthetic rate (Pmax); water use efficiency (WUEi); the maximum rate of carboxylation (Vcmax); the maximum rate of electron transport (Jmax); light saturation point (LSP); the light compensation point (LCP); leaf length, width, area, width:length (W:L), thickness; leaf mass per area (LMA), thickness; leaf C, N, P concentrations (TC, TN and TP); N/P, photosynthetic N use efficiency (PNUE); photosynthetic P use efficiency (PPUE); chlorophyll a + b (chl a + b); carotenoids and amino acid (AA); starch; soluble sugar; and nonstructural carbohydrates (NSC) of Schima superba under field and pot experimental nitrogen and phosphorus addition. Note: R²_c = Percentage of leaf traits explained by all the model (fixed + random). R²_m = Percentage of leaf traits explained by fixed factors.

Leaf Traits	Fixed Variables Statistics (p-Values)			Model Statistics
	N Addition	P Addition	Interaction	R2m	R2c
Pmax	0.0036	0.0618	0.1194	0.1138	0.7547
WUEi	0.9033	0.1439	0.8554	0.0360	0.4515
Vcmax	<0.0001	0.0168	0.2443	0.2541	0.7015
Jmax	<0.0001	0.0065	0.0403	0.2810	0.7395
LSP	0.1012	0.7244	0.0564	0.0231	0.8829
LCP	0.6180	0.8295	0.3300	0.0039	0.8923
Length	0.0785	0.5621	0.3996	0.0042	0.9666
Width	0.5848	0.8406	0.7979	0.0004	0.9701
Area	0.1822	0.6346	0.7232	0.0046	0.9282
W:L	0.5930	0.5141	0.2629	0.0398	0.3134
Thickness	0.6434	0.3313	0.3840	0.0036	0.9358
LMA	0.2299	0.1068	0.1348	0.0301	0.8409
TC	0.0082	0.4626	0.1165	0.0261	0.9182
TN	<0.0001	0.0915	0.6351	0.5930	0.8877
TP	0.0011	0.0084	0.0939	0.2461	0.6373
N/P	<0.0001	0.3230	0.8098	0.3593	0.7790
PNUE	0.0500	0.0081	0.5526	0.0904	0.7470
PPUE	0.0002	0.0067	0.0680	0.3421	0.6038
Chla + b	<0.0001	0.9034	0.3704	0.3161	0.7143
Carotenoids	0.1473	0.9882	0.0076	0.1163	0.6069
Starch	0.0783	0.1472	0.4545	0.1152	0.3388
Sugar	0.8572	0.8128	0.5850	0.0090	0.1979
NSC	0.8518	0.6049	0.5036	0.0159	0.2738
AA	0.0147	0.9518	0.2013	0.1557	0.3499

). WUEi was not affected by any treatment. Pmax and WUEi showed a significant increase of 2.69- and 1.57-fold in the pot condition compared to field conditions, respectively (Figure 1a,b). Vcmax, Jmax, LSP, and LCP also indicated a significant increase of 1.53-, 1.61-, 3.38-, and 4.41-fold, respectively, in pot conditions compared to field conditions (Figure 1c–f).

For pot conditions, Pmax, Vcmax, Jmax, and LSP increased with N addition (Figure 1a,c–e); WUEi was significantly decreased by P addition (Figure 1b). LCP did not respond to N addition compared with the CK (Figure 2f). By contrast, for field growth of seedlings, all photosynthesis parameters (Figure 1 and Figure 2) did not significantly respond to nutrient addition, except Jmax, which was increased by N addition (Figure 1d).

3.2. Leaf Morphological Traits

With the pooled data from the pot and field experiments, the linear mixed model revealed that leaf morphological traits were not affected by fertilization (Table 1). Leaf length, width, and area revealed no significant differences in pot conditions compared to field conditions (Figure 2a–c); whereas leaf width:length (W:L), thickness, and LMA showed a significant increase of 1.11-, 2.11-, and 1.44-fold, respectively, in pot conditions compared to field conditions (Figure 2d–f).

In pot conditions, no morphological traits were affected by fertilization treatment (Figure 2a–e), except LMA, which was increased by N addition (Figure 2f). However, for field growth of S. superba, LMA was significantly decreased by +NP compared to the CK (Figure 3f).

3.3. Leaf Biochemical Traits

With the pooled data from the pot and field experiments, the linear mixed model indicated that leaf C, N, and N/P concentrations were affected by N addition, while leaf P, PNUE, and PPUE were significantly altered by N and P addition (Table 1). All indicators—namely, leaf C, N, P, N/P, PNUE, and PPUE—revealed significant differences in pot conditions compared to field conditions (Figure 3). Leaf P, PNUE, and PPUE showed significant increases of 1.44-, 8.14-, and 2.22-fold, respectively, in pot conditions compared to field conditions (Figure 3c,e,f), while leaf C, N, and N/P were significantly decreased by 0.91-, 0.58-, and 0.26-fold, respectively, in pot conditions compared to field conditions (Figure 3a,b,d).

In pot conditions, compared with the CK, leaf C decreased with N addition (Figure 3a). Leaf N and N/P were higher with N and N-plus-P addition (Figure 3b,d); however, leaf P increased with P addition (Figure 3c). PNUE was lower in +NP compared to the CK (Figure 3e). Similar to leaf N, PPUE increased with N addition (Figure 3f). In field conditions, leaf C, N, P, N/P, PNUE, and PPUE were not altered by nutrient addition (Figure 3).

3.4. Leaf Physiological Traits

With the pooled data from the pot and field experiments, the linear mixed model revealed that leaf chl a + b and amino acids (AAs) were generally altered by N addition, while carotenoids were affected by P addition; starch, sugar, and NSC were not significant for N addition, P addition, or their interaction (Table 1). Leaf chl a + b revealed a significant increase of 1.49-fold in field conditions compared to pot conditions (Figure 4a). The carotenoid content was higher in +N and +NP plots compared to the CK (Figure 4b); however, leaf AA content, sugar, and NSC were significantly increased by 1.62-, 1.40-, and 1.35-fold, respectively, in pot conditions compared to field conditions (Figure 4c,e,f). The carotenoid and starch content were not significantly different between pot and field growth of seedlings (Figure 4b,d).

In pot conditions, leaf chl a + b increased with N addition and N plus P addition (Figure 4a), but carotenoids increased with N addition and P addition (Figure 4b). Furthermore, AAs increased with N addition (Figure 4c). Sugar and NSC were not altered by any treatment (Figure 4e,f), and starch was higher in the +NP plot compared to the CK (Figure 4d). As expected, chl a + b increased with N addition in the field experiment (Figure 4a). Consistently, for field growth of seedlings, carotenoids, sugar, and NSC did not significantly respond to fertilization (Figure 4b,e,f), while starch increased with P compared to N-plus-P addition (Figure 4d). In contrast to pot conditions, carotenoids and AAs were not significantly affected by any fertilization treatment (Figure 4b,c).

3.5. Relationship between Leaf Traits and Seedling Growth

RGR increased with N addition in the pot, while it was not affected by fertilization in the field (

Table 2. Relative growth rates (RGR) of Schima superba under field and pot experimental approaches, respectively treated by N and P addition. Note: The small letters indicate the differences (p < 0.05) among the four treatments. (CK = control, +N = nitrogen addition, +P = phosphorus addition, +NP = nitrogen plus phosphorus addition).

Variables	CK	+N	+P	+NP
Pot (RGR, mm yr−1)	24.54 ± 3.24 b	33.57 ± 2.50 a	24.10 ± 1.56 b	30.62 ± 3.38 ab
Field (RGR, mm yr−1)	1.55 ± 0.18 a	1.37 ± 0.44 a	1.67 ± 0.45 a	2.13 ± 0.69 a

). As indicated in Figure 5, there was a significantly positive correlation between Pmax and leaf traits, including LMA, N/P, PNUE, PPUE, Vcmax, Jmax, and RGR, in pot conditions. Moreover, there was a significant positive relationship between Pmax and leaf traits, including nutrient use efficiencies (PNUE and PPUE) and LSP, in field conditions. Pearson correlation coefficients among all leaf functional traits and growth rate for S. superba seedlings under field and pot experimental approaches are shown in Appendix A (

Table A1. Pearson correlation coefficients among leaf functional traits and growth rate for S. superba seedlings under field and pot experimental approaches.

	RGR	Pmax	WUEi	Vcmax	Jmax	LSP	LCP	Length	Width	Area	W:L	Thickness	LMA
RGR		−0.36	−0.04	0.30	0.10	−0.12	−0.35	0.10	0.31	0.22	−0.03	−0.10	−0.31
Pmax	0.45 *		0.47	0.34	0.28	0.59 *	0.25	0.42	0.36	0.38	−0.29	−0.30	−0.17
WUEi	0.15	−0.05		0.01	0.02	0.32	0.18	0.54	0.32	0.42	−0.50	−0.30	0.05
Vcmax	0.46 *	0.72 **	0.04		0.60 *	0.37	−0.24	−0.02	0.26	0.09	0.09	0.18	−0.10
Jmax	0.43 *	0.80 **	−0.02	0.95 **		0.46	−0.11	0.29	0.33	0.29	0.17	−0.14	−0.35
LSP	0.18	0.22	−0.16	0.04	0.09		0.41	0.48	0.66 *	0.55	−0.48	0.22	−0.13
LCP	−0.21	−0.11	−0.31	−0.01	−0.06	0.37		−0.11	−0.17	−0.20	−0.33	−0.18	0.12
Length	0.43 *	−0.23	−0.12	0.08	0.02	0.08	−0.16		0.86 **	0.95 **	−0.59 *	−0.23	−0.34
Width	0.12	−0.19	0.15	−0.29	−0.30	0.20	−0.15	0.55 **		0.96 **	−0.60 *	0.09	−0.34
Area	0.30	−0.26	−0.01	−0.10	−0.13	0.17	−0.16	0.91 **	0.84 **		−0.61 *	−0.06	−0.34
W:L	−0.18	0.10	0.44 *	−0.23	−0.22	0.02	−0.14	−0.39	0.45 *	−0.03		0.06	−0.34
Thickness	0.14	0.12	−0.04	−0.11	0.02	0.24	0.22	0.11	0.50 *	0.33	0.46 *		0.24
LMA	0.57 **	0.45 *	−0.18	0.37	0.40	0.38	0.42 *	0.17	0.08	0.16	−0.07	0.52 **
TC	−0.21	−0.46 *	−0.08	−0.51 *	−0.51 *	−0.20	0.21	−0.12	0.07	−0.07	−0.04	0.02	−0.20
TN	0.28	0.31	−0.02	0.46 *	0.39	0.40	−0.20	0.26	0.10	0.22	−0.12	−0.24	0.02
TP	−0.27	−0.34	−0.36	−0.52 *	−0.44 *	0.06	0.28	−0.22	−0.41 *	−0.35	−0.34	−0.14	−0.16
PNUE	0.18	0.42 *	−0.02	0.04	0.14	−0.15	0.10	−0.34	−0.10	−0.30	0.21	0.33	0.41 *
PPUE	0.47 *	0.84 *	0.01	0.74 *	0.77 *	0.21	−0.01	−0.04	0.08	0.01	0.26	0.36	0.62 *
C/N	−0.29	−0.44 *	0.15	−0.57 *	−0.56 *	−0.37	0.09	−0.21	0.05	−0.13	0.20	0.18	−0.08
C/P	0.26	0.37	0.19	0.55 **	0.47 *	0.09	0.05	0.18	0.41 *	0.33	0.38	0.38	0.36
N/P	0.34	0.43 *	0.05	0.61 **	0.53 **	0.37	−0.05	0.27	0.28	0.32	0.11	0.08	0.25
Chl a + b	0.55 **	0.25	−0.04	0.45 *	0.37	0.27	−0.19	0.33	0.10	0.26	−0.12	−0.11	0.20
Car	0.18	0.25	−0.39	0.20	0.26	0.48 *	0.00	0.11	−0.04	0.05	−0.14	0.06	0.25
AAs	0.31	0.24	−0.20	0.35	0.34	0.35	0.34	0.07	0.02	0.06	−0.06	0.26	0.55 **
Starch	0.39	0.04	−0.04	0.11	−0.01	0.33	0.10	0.37	0.36	0.40	0.04	−0.06	0.26
Sugar	0.04	−0.19	−0.11	0.05	−0.09	−0.15	0.32	0.36	0.27	0.36	−0.07	0.07	0.27
NSC	0.12	−0.17	−0.11	0.07	−0.09	−0.08	0.32	0.42 *	0.33	0.43 *	−0.06	0.05	0.31
	TC	TN	TP	PNUE	PPUE	C/N	C/P	N/P	Chl a + b	Car	AA	Starch	Sugar	NSC
RGR	−0.30	−0.10	−0.49	−0.38	−0.25	0.04	0.56	0.46	0.01	−0.02	−0.28	0.06	−0.01	0.01
Pmax	0.02	0.09	0.22	0.94 **	0.90 **	−0.12	−0.30	−0.23	0.10	−0.21	−0.06	0.14	−0.17	−0.09
WUEi	0.36	0.13	−0.24	0.41	0.62 *	−0.14	0.17	0.22	0.10	0.12	0.03	0.14	−0.02	0.03
Vcmax	−0.22	−0.33	−0.40	0.39	0.43	0.32	0.35	0.14	0.09	−0.06	−0.34	−0.08	−0.49	−0.43
Jmax	0.01	0.39	−0.34	0.11	0.41	−0.41	0.40	0.55	0.60 *	0.14	0.20	−0.11	−0.11	−0.13
LSP	0.03	0.02	−0.27	0.50	0.66 *	−0.08	0.16	0.15	0.10	−0.40	−0.36	0.30	−0.34	−0.17
LCP	0.53	0.08	0.12	0.23	0.19	−0.04	−0.11	−0.06	−0.11	−0.40	−0.33	−0.19	−0.33	−0.34
Length	0.08	0.41	−0.28	0.22	0.58 *	−0.51	0.23	0.41	0.44	0.09	0.35	0.24	−0.04	0.06
Width	−0.15	0.10	−0.44	0.23	0.52	−0.22	0.34	0.37	0.34	−0.06	0.02	0.39	−0.21	−0.03
Area	−0.07	0.24	−0.35	0.21	0.54	−0.36	0.26	0.37	0.39	0.02	0.19	0.42	−0.06	0.10
W:L	−0.52	0.28	0.50	−0.34	−0.53	−0.20	−0.38	−0.29	0.01	0.29	0.10	−0.42	0.44	0.20
Thickness	−0.34	−0.50	−0.18	−0.18	−0.26	0.48	0.02	−0.23	−0.31	−0.06	−0.35	0.15	−0.22	−0.12
LMA	0.55	−0.68 *	−0.32	0.11	0.02	0.71 **	0.24	−0.02	−0.49	−0.28	−0.11	0.32	−0.18	−0.03
TC		−0.03	−0.39	0.08	0.28	0.08	0.41	0.45	0.21	0.05	0.28	−0.14	−0.39	−0.37
TN	−0.41 *		0.33	−0.23	−0.01	−0.99 **	−0.19	0.24	0.55	0.30	0.52	−0.32	0.43	0.23
TP	0.40	−0.39		0.16	−0.20	−0.28	−0.97 **	−0.81 **	−0.18	−0.03	0.05	−0.14	0.47	0.33
PNUE	0.16	−0.64 *	0.07		0.86 **	0.21	−0.26	−0.33	−0.09	−0.28	−0.18	0.21	−0.27	−0.15
PPUE	−0.50 *	0.33	−0.61 *	0.30		−0.04	0.10	0.13	0.24	−0.12	0.05	0.20	−0.34	−0.21
C/N	0.53 *	−0.87 **	0.30	0.58 **	−0.44 *		0.15	−0.27	−0.57	−0.28	−0.51	0.22	−0.44	−0.28
C/P	−0.34	0.37	−0.89 **	−0.05	0.73 **	−0.36		0.90 **	0.23	0.05	0.03	0.02	−0.38	−0.30
N/P	−0.45 *	0.86 **	−0.68 **	−0.45 *	0.64 **	−0.80 **	0.77 **		0.50	0.21	0.33	−0.09	−0.18	−0.18
Chl a + b	−0.26	0.77 **	−0.38	−0.42 *	0.40	−0.67 **	0.41 *	0.73 **		0.70 *	0.66 *	−0.29	−0.14	−0.22
Car	−0.10	0.33	0.17	−0.13	0.27	−0.36	−0.05	0.23	0.47 *		0.68 *	−0.44	0.03	−0.14
AA	−0.10	0.27	−0.18	−0.11	0.55 **	−0.40	0.48 *	0.50 *	0.50 *	0.44 *		−0.25	0.21	0.08
Starch	−0.08	0.49 *	−0.20	−0.28	0.22	−0.41 *	0.31	0.54 **	0.47 *	0.13	0.42 *		0.36	0.66 *
Sugar	0.07	−0.17	−0.20	0.03	−0.01	0.12	0.20	−0.02	−0.14	−0.07	−0.10	0.12		0.94 **
NSC	0.05	−0.06	−0.23	−0.03	0.04	0.03	0.25	0.09	−0.03	−0.04	−0.01	0.33	0.98 **

Note: Black font represented pot experiment and blue font represented field experiment (p < 0.01, **; p < 0.05, *); Traits are abbreviated as follows: relative growth rates (RGR), leaf light-saturated photosynthetic rate (P_max), water use efficiency (WUEi), the maximum rate of carboxylation (V_cmax), the maximum rate of electron transport (J_max), light saturate point (LSP), and light compensation point (LCP), leaf length, width, area, width:length ratio (W:L), thickness, leaf mass per area (LMA), Leaf organic carbon (TC), nitrogen (TN), phosphorus (TP), stoichiometric ratio (C/N, C/P, N/P), photosynthetic nitrogen use efficiency (PNUE), photosynthetic phosphorus use efficiency (PPUE), chlorophyll a + b (Chl a + b), carotenoids (Car), amino acids (AA), starch, sugar, and nonstructural carbohydrates (NSC).

). However, significantly negative relationships were observed between Pmax and C/N in pot conditions.

PCA was employed to evaluate how leaf traits and relative growth rates were associated for Schima superba under field and pot experimental nitrogen and phosphorus addition (Figure 6). PCA axis 1 showed strong positive loadings for P_max, WUEi, Jmax, Vcmax, LSP, LCP, PNUE, PPUE, LMA, thickness, and RGR, and had negative loadings for leaf TC, length, area, width, and chlorophyll. PCA axis 2 showed strong positive loadings for leaf TN, N/P, and had negative loadings for leaf TP (Figure 6a). Seedlings of S. superba under field and pot experiments were well separated along the first PCA axis. In addition, +N was separated from CK, +P and +NP along the second PCA axis in the field, while +N and +NP were separated from CK and +P along the second PCA axis in the pot (Figure 6b).

4. Discussion

4.1. N Addition Strongly Altered Leaf Traits and Limited Tree Growth Compared to P Addition

Our comparative experiments provide unique insight into nutrient limitations in a subtropical forest based on field and pot conditions. It is well understood that P-limited trees grow in subtropical and tropical China, especially at young ages [40,41]. In contrast to our first hypothesis, the tree growth rate and Pmax of S. superba were not affected by any fertilization in our field experiment. Meanwhile, Pmax and RGR were not affected by P addition in pot experiments. Those findings revealed that our seedlings would not be limited by P, which indicated the forest was not a rich N and poor P ecosystem as expected, although soil data support that this forest would be limited by P because the soil-available P and available P/mineral N ratio increased more with P addition than N addition (unpublished data), suggesting that the effect of P addition might be more advantageous to sustain the soil N:P stoichiometry balance than the available P addition. Inconsistently, our results showed several lines of evidence of N limitation for a specific species (S. superba). First, leaf N generally increased by N addition and N + P addition in both experiments, although the leaf N in the field condition was only slightly higher in N addition and N + P addition plots. These results directly reflected the positive effects in response to N supply. Second, the mixed model suggested that leaf C, N/P, and Pmax were only significantly affected by N addition, which are used as indicators related to carbon accumulation, nutrient balance, light capture, and N metabolism. Third, the Pmax of S. superba in field conditions was not changed by either N or P addition, but Jmax and PPUE increased with N addition, indicating a higher N investment in electron transport to maintain photosynthesis rates. Last, PCA showed that seedlings treated by +N were separated from those of CK, +P and +NP along the second PCA axis in the field, while seedlings treated by +N and +NP were separated from those of CK and +P along the second PCA axis in the pot, which suggests that N addition significantly altered leaf traits and tree growth. Consequently, N and P limitations based on the soil level may not be well applied in a specific species. It is necessary to comprehensively explore the effect of N and P availability on leaf functional traits at the community level over the long term.

Interestingly, we found that in the pot experiments, +NP treatment did not significantly effect RGR, Pmax, Vcmax, or Jmax but did significantly alter leaf N, N/P, PNUE, chlorophyll a + b, and starch. Similarly, previous studies have shown that +NP can have positive, negative, and no effect on various leaf traits and basal area in different species [42], and negative effects of long-term P additions on understory plants in tropical forests [18]. One mechanical explanation for the negative effects of P addition on seedling growth may be related to the nutrient requirements for different plants. Previous studies found that S. superba has a higher requirement for N [26]. Another mechanical explanation may be due to the restriction of other trace elements such as potassium [K], magnesium [Mg], and calcium [Ca] caused by excessive P, as supported by long-term P addition experiments [18].

Our study did not support the hypothesis that P limitation is stronger and N limitation is weaker in these subtropical forests. Nonetheless, it would be premature to discard the hypothesis that P limitation is stronger than N limitation in lowland tropical forests for several reasons. First, our results found that S. superba can adapt to low-P conditions (low leaf P concentration and high PPUE) and has a higher demand for N than P. Thus far, a few species have been tested in limited fertilization experiments in these diverse subtropical forests, and little is known about the growth response of those species adapted to low-P conditions in response to increased resource levels. In our field experiment, effect sizes tended to be larger for plant responses to N addition than to P addition; however, the difference was small and insignificant. Second, RGR was not affected by N addition in the field experiments, suggesting that N did not limit seedling growth in this subtropical forest. In our study, N addition increased the photosynthetic capacity but also increased the risk of being browsed (e.g., leaf Chl a + b and AA concentration increased significantly with N addition). Plants fertilized with N were more attractive to herbivores and possibly other pests. Third, the negative effect of P addition on non-growth function (e.g., survivor rates, disease resistance, and plant regeneration) may offset the positive P effect on photosynthesis [18]. Collectively, plant responses to fertilization are complicated, and the impact intensity of N addition may be stronger than P addition in a specific species even in a rich N and poor P ecosystem.

4.2. Experimental Approach Altered the Effect of Fertilization on Leaf Traits and Tree Growth

In the present study, the effects of fertilization on most leaf traits and tree growth were weaker in the field than in the pot. Similar to our results, Durand et al. (2020) reported that genotype and water treatment had a much smaller effect on stomatal dynamics in the field than in the greenhouse [43]. Likewise, Zhu et al. (2014) reported that three years of N addition had no significant effect on Pmax in two dominant understory species in tropical regions [10], while many studies have found that S. superba seedlings responded positively to N supplementation in artificial conditions [26,29]. In our study, we also observed that the intensity of fertilization effects varied with different traits. For example, contrary to our second hypothesis, leaf morphological traits were generally not affected by fertilization regardless of the type of experiment, suggesting that unchanged leaf morphological traits might be an important mechanism under fertilization, similar to previous observations. Inconsistent with morphological traits, the effects of fertilization on biochemical (C, N, P, N/P, PNUE, PPUE) and physiological (chl a + b, carotenoids and amino acid) traits were weaker in the field than in the pot experiment. In addition, +N was slightly separated from CK, +P and +NP along the second PCA axis in the field, while +N and +NP were strongly separated from CK and +P along the second PCA axis in the pot, which is similar to the second hypothesis. Therefore, responses in pot experiments (laboratory, pot, and greenhouse) might not reflect the responses of forest plants in the field experiment under global change.

A possible mechanism for the weak response to fertilization in the field but a strong response in the pot experiment might concern light availability. First, higher light intensity, LMA, light compensation point (LCP), W:L, and thicker leaves in the pot as compared to the field experiment suggested that seedlings develop mechanisms in adapting to high-light conditions, such as photoprotection, increasing the respiratory rate, and so on [44,45]. Second, Chl a + b peaked in the field, while it was lower in pot conditions, and sugar and non-structural NSC showed a significant increase of 1.40- and 1.35-fold, respectively, in pot conditions compared to field conditions. This finding indicated that seedlings need to allocate much more N to capture light in low-light environments, which is consistent with previous studies that the storage of NSC is generally lower in trees growing in the shade than those in the sun because carbohydrate synthesis is often limited by lower levels of light availability [46]. Furthermore, unchanged Vcmax and light saturation point (LSP) in field conditions also indicated that seedlings may not have access to enough light for photosystem II activity [47]. Collectively, those results highlighting N additions might regulate the primary production of S. superba in pot conditions with abundant light availability, while understory tree seedlings might not be limited by nutrient availability in low-light conditions.

Another mechanism for the dependence of fertilization effects on experimental approaches might be due to differential abiotic (i.e., environmental conditions) and biotic (i.e., interspecific interaction, age, and competition) factors between the pot and the field experiments. In our study, we examined fertilization effects compared with the control within each experimental approach to reduce the exogenous difference, and these within-group patterns were our primary interest here. Nevertheless, there are several factors that can affect N and P addition effects: plant size (age), nutrient availability to the individual plant (plant density, soil depth, soil nutrient concentrations, soil decomposition rate, soil moisture, temperature, and so on). Thus, those factors were uncontrolled, and the strength of N and P addition effects cannot be attributed to light availability alone in this study. Clearly, future directions should consider reducing the confounding effects of biotic and abiotic factors on fertilization in the field, and refinement controls seedlings’ genetic diversity, mycorrhizal symbiont and root competition for long-term fertilization experiments are required.

4.3. Relationships between Tree Growth and Leaf Traits Were Weaker in the Field Than in the Pot

The effects of fertilization on leaf traits are tightly associated with photosynthesis rates because leaves are the most important tree organ used for photosynthesis and carbohydrate production [48,49]. However, similar to our hypothesis, the relationship between leaf traits and tree growth was weaker in the field experiment than in the pot experiment, and no link was found between RGR and leaf N:P or C:N in the field, indicating that plants adjust their stoichiometric ratio to favor either speed or efficiency of protein synthesis in different environments, but this leaf stoichiometry does not dictate tree growth alone. These results were consistent with the previous study, because plants may take up N and P in excess of growth requirements when nutrients are abundant, sustaining steady growth through periods of nutrient scarcity [50]. Similarly, the relationship between RGR and nutrient ratios was not always present for each of the marsh herbaceous plants [51]. Thus, existing relationships between leaf stoichiometry and RGR might depend on experimental approaches, and species-specific and experimental approaches should be carefully considered when using the theoretical association of tree growth rate with leaf stoichiometry.

In addition, a significant positive relationship was observed between Pmax and PNUE/PPUE in both experiments. Decoupling can be attributed to a weak relationship between intraspecific leaf traits and photosynthesis, because the leaf biochemical traits are based on interspecies variation. Interestingly, the function slopes of Pmax and PNUE/PPUE differed based on experiment type, suggesting that Pmax was more sensitive to resource use efficiency in the field than in the pot condition. Therefore, we speculated that S. superba would have a maximum rate of photosynthesis in response to N addition by investing the N chemical compound in chlorophyll for light harvesting rather than by increasing N use efficiency in pot conditions, while having conservative strategies associated with increased PNUE under field conditions. Similarly, Mao et al. (2018) found that PNUE decreased with leaf N in an understory species treated by N addition in a tropical reforested ecosystem [8]. Therefore, the relationship between Pmax and PNUE/PPUE could be used as a sensitive indicator to assess plant adaptive strategies under N deposition and P limitation. However, physiological traits were not related to tree growth, giving no support for the hypothesis that species with fast RGR should possess smaller or faster Chl, NSC, and AAs to facilitate growth under nutrient alteration conditions [52]. Furthermore, photosynthetic capacity stimulation was consistent with increased RGR, suggesting that the change in Pmax contributed to increasing tree growth.

5. Conclusions

Field and pot experiments were designed with +N, +P, +NP, and a control (100 kg N ha⁻¹ yr⁻¹ and/or 50 kg P ha⁻¹ yr⁻¹) to comparatively investigate the effects of N and P addition on leaf functional traits and relative growth rate (RGR) of Schima superba seedlings in subtropical China. We found that the impact intensity of N addition on leaf traits and growth rate was stronger than that with P addition in a specific dominant tree seedling in a subtropical forest. Specifically, N addition, P addition and their interaction had no effect on leaf morphological traits in both experiments, except leaf mass per area (LMA). The effects of N addition on leaf biochemical and physiological traits were generally greater than those with P addition, which was not consistent with the P limitation hypothesis in these rich N and poor P ecosystems; only leaf P increased with P addition and water use efficiency (WUEi) decreased with P addition in the pot experiment.

Furthermore, the experimental approach altered N and P addition effects. RGR was not affected by N and P addition in the field, while it increased with N addition in the pot experiment. These divergent fertilization effects likely depend on light availability rather than N and P supply. Therefore, long-term fertilization in the field experiment is still needed, and more refinements for pot fertilization experiments are required to simulate plants in the field, particularly considering more parameters defining the factors that can influence leaf traits and growth rate, including genetic, water, light, and vapor pressure deficit conditions, among others.