**3. Results**

### *3.1. Availability of Soil Nutrients and Carbon, Nitrogen, and Phosphorus (CNP) Concentration of Plant Litter and Soil*

Seven years after the annual addition of N and P to the plantation, we found that the soil of *L. gracile* has a higher NO3−-N content compared with the other three understory species. Moreover, the AP content in the soil of *Dicranopteris dichotoma* is significantly lower than that of the other plants (*p* < 0.01). DOC, NH4<sup>+</sup>-N, AN, and pH do not change among the four understory plant species (Table 1). Among the four species, *L. gracile* has higher concentrations of C and N in litter and N and P in associated soil (Table A1). The litter C, N, and P and soil C concentrations of *W. japonica* are all low (Table A1). The litter and soil C concentrations of *Dryopteris atrata* are all higher than the other species (*p* < 0.01). The litter and soil P concentrations of *Dicranopteris dichotoma* are all low (Table 1).

**Table 1.** Litter and soil nutrients of four understory plant species (means ± SE). Superscript letters in the same line indicate a significant difference at *p* = 0.05 based on LSD (least significant difference) comparison. Lines without letters indicate that the difference is not significant. LC: litter total organic carbon (C); LN: litter total nitrogen (N); LP: litter total phosphorus (P); SC: soil total organic C; SN: soil total N; SP: soil total P.


### *3.2. E*ff*ects of N and P Addition on Carbon, Nitrogen, and Phosphorus (C, N, and P) Concentrations and Stoichiometric Ratios of Di*ff*erent Functional Groups*

Litter C, N, and the C:N ratio were significantly affected by N addition, the plant functional group, understory species, and the randomized complete block design. Litter P and the C:P ratio depended on P addition, understory species, and the randomized complete block design (Table 2). Litter P also depended on understory species × P addition ("S × P[f]"). The N:P ratio of litter depended on P, N × P, and the plant functional group. However, the results varied with blocks except L N/P (Table 2).


**Table 2.** ANOVAs on the effects of nitrogen (N) and phosphorus (P) deposition, plant functional group (F), understory species (Sp; nested within plant functional group, f), and their interactions on litter (raw decomposed leaves) C, N, P, and C/N/P. B&R indicates randomized complete block design. The six indicators include

### *3.3. Correlation between the Litter of Di*ff*erent Functional Groups and Soil Carbon, Nitrogen, and Phosphorus (C, N, and P) with the Addition of N and P*

Based on the unitary regression analysis, the litter N of the fern functional group was positively correlated with soil P (*p* = 0.051, Figure A1f) compared with the CK. In addition, for the fern functional group, the litter P was positively correlated with soil C (*p* = 0.002, Figure A1g) and negatively correlated with soil P (*p* = 0.019, Figure A1i). With the addition of N, the litter C of Gramineae was positively correlated with soil P (*p* = 0.029, Figure A2c); the litter N was positively correlated with soil N (*p* = 0.015, Figure A2e). The litter P and soil P were positively correlated in the Gramineae group but negatively correlated in the fern group (Figure A2i). With the addition of P, there was a negative correlation between the C and soil P of the litter of the fern group (*p* = 0.051, Figure A3c); the litter N and soil N were negatively correlated (*p* < 0.001, Figure A3e) in ferns but positively correlated (*p* = 0.008, Figure A3f) in Gramineae. Litter P is positively correlated with soil P in both the Gramineae and fern groups (Figure A3i). With the application of both N and P, the change of the litter C concentration in the fern group was affected by the change of soil C, soil N, and soil P concentration. The litter C concentration increased with the increase of soil C but decreased with the increase of soil N and soil P concentration (Figure 1a–c). The litter C concentration was negatively correlated with soil N (*p* = 0.059, Figure 1b) in the fern group, whereas soil N was positively correlated with litter C (*p* = 0.003, Figure 1b) in the Gramineae group. The litter N concentration of the fern group decreased with the increase of the soil N concentration (*p* < 0.001, Figure 1e). There was no correlation between litter N, soil C, and soil P of the two different plant functional groups (Figure 1d, f). Moreover, the litter P concentration was positively correlated with both soil N and soil P (Figure 1h, i) in the fern functional group.

### *3.4. E*ff*ects of N and P Addition on Carbon, Nitrogen, and Phosphorus (C, N, and P) Concentrations and Stoichiometric Ratios of the Soils Associated with Di*ff*erent Plant Functional Groups*

A nested ANOVA showed that soil C was significantly affected by the understory species. Soil N and soil P were significantly affected by N addition, plant functional group, understory species, and randomized complete block design. The soil N also depended on P addition\*plant functional group (Table 3). The soil C/N ratio and C:P ratio depended on the plant functional group, understory species, understory species\*P addition, and randomized complete block design (Table 3; Figure 2). The soil C/P ratio depended on P addition\*plant functional group. The soil N:P ratio depended on N addition, understory species, and randomized complete block design (Figure 2). Similarly, results of SN, SP, S C/N, SC/P, and SN/P varied with blocks (Table 3).


**Figure 1.** Correlations between litter and soil carbon, nitrogen, and phosphorus (C, N, and P) of four dominate species with two different functional types (Gramineae *Lophatherum gracile* and fern *Woodwardia japonica*, *Dryopteris atrata* and *Dicranopteris dichotoma*) with N (100 kg ha−<sup>1</sup> year<sup>−</sup>1) and P (50 kg ha−<sup>1</sup> year<sup>−</sup>1) addition.
