4 DISCUSSION
The plant N:P ratio not only is a link factor among different scales of molecular, cellular, individual, population, community and ecosystem (Elser et al., 2000; Sterner & Elser, 2002; Elser & Hamilton, 2007) but is also an important indicator of nutrient supply (Koerselman & Meuleman, 1996) and dynamic nutrient limitations of the environment (Güsewell, 2004; Vitousek et al., 2010; Hu et al., 2018). Since species richness is always associated with productivity (Grime, 2001), nutrient availability first affects productivity, then plant density over long time. Therefore, the vegetation N:P ratio affects plant density. In the present study of a successional series in a semi-arid area, as the vegetation N:P ratio increased, plant total density increased significantly in the A. ordosica community (early successional stage) but decreased significantly in the S. bungeana community (late successional stage). This suggests that the vegetation N:P ratio was the mechanism driving negative density dependence.
In the natural world, no organism is isolated; rather, each organism is a member of a population consisting of many homologous species. Individuals of the same species have similar requirements for growth, reproduction and survival. Thus, when their demand for resources exceeds the supply, conspecifics compete for resources. Plants cannot move on their own; thus, nutrient limitation will affect their physiology and biochemistry and limit plant biomass. Strong nutrient competition increases nutrient stress and affects individual reproduction and mortality, thereby affecting plant density. The vegetation N:P ratio is an important indicator to of plant nutrient limitation (Güsewell, 2004; Vitousek et al., 2010). When nutrient limitation is severe, the vegetation N:P ratio can affect not only productivity but also the plant density of the community, especially in arid and semi-arid areas.
In the present study of a successional series in a semi-arid area, as the vegetation N:P ratio increased in the early successional-stage community of A. ordosica , total density increased linearly and significantly, whereas biomass increased, but the relationship was not significant. For sand-dune plants in the early successional stage, soil nitrogen varies to a greater extent that does soil phosphorus (He, You, & Yu, 2016). This difference exists because of the deposition of atmospheric nitrogen and the transformation of organic matter via litter decomposition, which causes soil nitrogen to increase rapidly. Plant nitrogen then increases accordingly, which eventually leads to a larger vegetation N:P ratio. However, due to the extremely low nitrogen content in dune soil for the early successional stage community, the absolute nitrogen contents of plants were low. Therefore, the total density of the A. ordosica community increased with increasing vegetation N:P ratio. In the mid-successional stage, S. alopecuroides community, soil phosphorus began to accumulate, and soil nitrogen continued to accumulate. There was no significant statistical relationship between total density and vegetation N:P ratio; however, the biomass of the S. alopecuroides community increased significantly with increasing total density. In the late successional stage, phosphorus was consumed in large quantities due to its physiological and biochemical effects on plant drought resistance, and the plant communities were restricted by phosphorus. In the S. bungeana community, representing the late successional stage, total density decreased significantly with increasing vegetation N:P ratio, which indicated significant negative density restriction in this stage. At this stage, biomass decreased with increasing total density of the plant community, reflecting phosphorus limitation; however, the decrease was not significant.
Koerselam & Meuleman (1996) indicated that when the vegetation N:P is <14, plant growth at the community level is mainly restricted by N, whereas when N:P is 14-16, plant growth is restricted by N and P, and when N:P is >16, plant growth is mainly restricted by P. In the present study, we demonstrated that the vegetation N:P ratios of the A. ordosica , S. alopecuroides and S. bungeana communities were 11.27±0.97, 17.08±0.86 and 20.84±1.01. These results indicated that the A. ordosica community was limited by N, whereas the S. alopecuroides and S. bungeanacommunities were limited by P. Interestingly, in the early and middle successional stages, there was a significant quadratic function relationship between total density and vegetation N:P ratio in the communities, and the vertex coordinate was (16.6, 353.3). These findings revealed that the community would face negative density constraints when the vegetation N:P ratio was close to 16.6. In the middle and late successional stages, plant growth was mainly restricted by P, which led to negative density dependence.
For populations and communities, the vegetation N:P ratio is of core interest when predicting nutrient limitation, whether N limitation or P limitation (Vitousek et al., 2010). For instance, in our previous research we established an 80 × 80 m sample plot in an A. ordosica community on fixed sand-dunes in the Alxa Desert. We investigated plant density in 400 quadrats (4×4 m) in the plot and determined total N and available P in the 20 cm soil layer of each quadrat. Finally, we regressed the density of A. ordosica against the soil N:P ratio. We found that A. ordosica density had no significant regression relationship with soil N or P but had a significant, negative linear relationship with the soil N:P ratio. In previous work, regression results of soil nutrients and plant density were similar between a sampling scale of 40×20 m and a sampling scale of 4×4 m (Wu et al., 2009). The above observations suggest that the N:P ratio is more informative than N or P content alone.
Negative density dependence is a very common phenomenon in nature. The factors affecting negative density dependence are water, nutrients and salt. Regarding nutrients, density is not limited so much by N or P but by the vegetation N:P ratio. The results of the present study indicate that the vegetation N:P ratio in a semi-arid area is the factor driving negative density dependence.
With the development of a plant community, the vegetation N:P ratio continuously changes. From the early successional stage to the late successional stage, the vegetation N:P ratio gradually increases. In the early successional stage A. ordosica community, total plant density increased with increasing vegetation N:P ratio; however, in the late successional stage S. bungeana community, total plant density decreased with increasing vegetation N:P ratio. Thus, the vegetation N:P ratio is a driving factor of plant community variation.