3.3 Plant growth responses to the N and P addition treatments
Tiller number, plant density, and maximum plant height were significantly affected by N and P addition across the three surveyed years (all P < 0.05; Figure 3, Table 1). The low N addition increased the tiller number by 168%, 170%, and 177% under the ambient, low, and high P addition treatments, respectively (allP < 0.05; Figure 3a left insert), as did the high N addition, but almost twice as strongly, with corresponding percentages of 313%, 397%, and 446% under ambient, low, and high P addition conditions, respectively (all P < 0.05; Figure 3a left insert). Although the low and high P addition negligibly affected tiller number under the ambient or low N addition treatments (all P> 0.05; Figure 3a right insert), they did raise tiller number by 38% and 40% under the high N addition treatments, respectively (both P < 0.05; Figure 3a right insert).
The low N addition did not influence the density of S. kryloiiunder ambient and low P addition treatments (both P> 0.05), but suppressed it by 53% under high P addition (P < 0.05; Figure 3b left insert). By contrast, high N addition enhanced plant density by 35% and 64% under ambient and low P addition, respectively (both P < 0.05), but did not affect it under high P addition (P > 0.05; Figure 3b left insert). Across the three years, low P addition decreased the density, on average, by 49% under ambient N conditions (P< 0.05). While low and high P addition respectively reduced plant density by 46% and 71% under low N addition (both P< 0.05), the effect of high P addition was weakened by the high N addition, so that plant density decreased by 50% (P< 0.05; Figure 3b right insert).
The low N addition enhanced the maximum plant height by 36%, 52%, and 58% under ambient, low, and high P addition treatments, respectively (all P < 0.05), but the corresponding effects were stonger, at 69%, 72%, and 83% for the high N addition (Figure 3c left insert; all P < 0.05). Although the low and high P addition did not influence the maximum height under ambient or high N addition conditions (Figure 3c right insert; all P > 0.05), they did so under the low N addition treatments by 16% and 15%, respectively (Figure 3c right insert; P < 0.05).
Path analysis for effects of reproductive traits and plant growth on seed production
We used SEM to examine the direct and indirect factors affecting seed production. The results revealed that tiller number, plant density, and maximum plant height were indirectly responsible for 89%, 48%, and 90% of the variation in seed production, respectively, under the N and P addition treatments (χ215 = 24.47,P = 0.058, RMSEA = 0.134; Figure 4). Direct contributions to changes in seed production arose from the seed number per inflorescence (R2 = 0.88, P < 0.001) and inflorescence number (R2 = 0.12, P = 0.044). The N addition promoted seed production of S. kryloii mainly via enhanced tiller number and an accompanying enhancement in the plants’ inflorescence number (Figure 4). The P addition increased seed production differently, mainly by reducing the density of S. kryloii , thereby enabling individuals to a greater maximum height and consequently a greater seed number per inflorescence (Figure 4).