Native and invasive species
The plant species pool consisted of eight native species and three invasive species that commonly grow on grasslands around Taizhou city, Zhejiang Province, China (Table S1). Six of the native species were perennials (Cirsium arvense var. integrifolium, Aster indicus, Inula japonica, Viola philippica, Plantago asiatica and Lysimachia fortunei) and two were annuals (Solanum nigrum and Polygonum posumbu). The three invasive species (Solidago canadensis, Erigeron canadensis and Symphyotrichum subulatum) were all asteraceae, as most invasive species in China are from this family (Ma 2013). Solidago canadensis is a perennial and can reproduce both sexually and clonally by producing rhizomes; E. canadensis and S. subulatum are non-clonal annuals. All three invasive species originated from North America and can produce plenty of viable seeds every year (Ma 2013).
Seeds from the native and invasive species were collected from field sites around Taizhou city in 2013 (Table S1). On March 18–19, 2014, seeds from each species were surface-sterilized and sown into three plastic containers (52 cm long × 35 cm wide × 15 cm high) filled with sterilized peat (Kuheng Co., Shanghai). The plastic containers were placed in a growth chamber at 25°C and a relative humidity of 70% with 16 h of daylight at 70 µmol·m− 2·s− 1. The seedlings 8–12 cm tall were used to construct native plant communities that were to be invaded or not invaded by each of the three invasive species. Height differences within the same species are limited to 2 cm.
Experimental Design
To test the hypotheses, we carried out two experiments. For the first experiment, we constructed native plant communities with three levels of species richness (1, 3 and 6 species) in pots (each 27.5 cm in diameter × 31 cm in height) filled with either sterilized soil (without soil microbes) or nonsterilized soil (with soil microbes) and allowed them to be invaded by each of the three alien invasive species (S. canadensis, E. canadensis and S. subulatum) or prevented invasion by these three species. For the treatments with soil microbes, 1.65 L soil was added to the middle layer of 14.85 L of a sterilized soil mixture, and for the treatment without soil microbes, 1.65 L sterilized soil was added to the middle layer of 14.85 L of a sterilized soil mixture. The sterilized soil mixture consisted of a mixture of soil, peat, sand and vermiculite at a volume ratio of 2:1:1:1, containing 1.60 ± 0.32 g kg− 1 total N, 0.58 ± 0.18 g kg− 1 total P and 13.2 ± 0.5 g kg− 1 organic matter (mean ± SE, n = 5). Soil was collected from the local plant communities in mountain areas near Taizhou city with the growth of the above-mentioned eight native plants. No invasive plants grew in these selected communities. The soil was mixed and sieved through a 1-cm mesh to remove larger roots and stones. The peat, sand and vermiculite were bought from Kuheng Co. (Shanghai).
We constructed three replicates of monocultures for each of the eight native species, five three-species mixtures with different species compositions and five six-species mixtures with different species compositions (Table S2). The species in each of the three- and six-species mixtures were randomly chosen from the native species pool. For the treatment without the invasive species, each pot contained six seedlings of the same native species (for monocultures), including two seedlings of each of the three species (for three-species mixtures) or one seedling of each of the six species (for six-species mixtures). The six seedlings in each pot were arranged in a circular pattern. For the treatment with the invasive species, one seedling of one of the three invasive species was grown in the center of a pot, surrounded by six seedlings of native species (Fig. 1). The experiment thus consisted of 34 species compositions ⋅ 2 soil microbe treatments ⋅ 4 invasion treatments, resulting in a total of 272 pots.
The experiment was started on May 26, 2014. The pots were placed randomly in a greenhouse (25°C during the daytime and 18°C at night) at Taizhou University in Taizhou city, Zhejiang Province, China. The aboveground parts of each plant species in each pot were harvested on August 25–28, 2014. All plant material was oven-dried at 70°C for 48 h and weighed. The soils in each of the 34 pots without the invasive species were used as inocula for the second experiment described below.
For the second experiment, 14.85 L of the same sterilized soil mixture as used in the first experiment was inoculated with 1.65 L of the soil from each of the 34 pots previously grown with the native species communities but without the invasive species. Three pots (27.5 cm in diameter × 31 cm in height) containing each of these 34 soil mixtures were prepared, and each pot contained one seedling of S. canadensis, S. subulatum or E. canadensis. There were 102 pots in total. The plant from each pot was harvested after three months, and the biomass was determined by drying the plant in an oven at 70°C for 48 h and weighing it.
Data analysis
The invasion resistance of a plant community was calculated as the ratio of the aboveground biomass of the community invaded by an alien species to that of the community comprising the same initial species composition but without the invasive species (Pfisterer and Schmid 2002; Tilman 1996; Wang et al. 2007). Complementarity and selection effects were calculated using the invasion resistance data and the additive partitioning method described by Loreau and Hector (2001). The complementarity effect of a mixture was calculated as , where N is the number of species in the mixture, is the mean value of the change in the relative invasion resistance across all species in the mixture and is the mean value of the invasion resistance of the monocultures across all species. The selection effect was calculated as Ncov(ΔRY, M), where N is the number of species, cov(ΔRY, M) is the covariance between the invasion resistance of species in monocultures (M) and their change in the relative invasion resistance in the mixture (ΔRY).
If complementarity between species plays a major role in the invasion resistance, then the invasion resistance of mixtures will be higher than that of the most resistant species in the mixtures. The over-invasion resistance index (OI) was calculated as follows (Hector et al. 2002): OI = Y/MAX(Mi), where Y is the invasion resistance of a mixture, and Mi is the invasion resistance of species i in the monoculture. If log (OI) > 0, the mixture should have a higher invasion resistance than that of the highest resistant species.
As the damage caused by soil pathogens to invasive species may be influenced by the phylogenetic distance between native and invasive species (Strauss et al. 2006; Zheng et al. 2018), we calculated the phylogenetic distances between the eight native and three invasive species using three commonly sequenced genes from the GenBank: rbcL, matK and ITS (Table S3). Of the 11 species, eight had three genes represented in GenBank. For three native species, C. arvense var. integrifolium, I. japonica and P. asiatica whose sequence data of the three genes were not available or incomplete in GenBank, we used the sequence data from their congeneric relatives (i.e., C. arvense, I. britannica and P. depressa) as proxies. Sequences were aligned for each region independently using MUSCLE (Edgar 2004) and combined into a single supermatrix. Analyses were conducted using the Maximum Composite Likelihood model (Tamura et al. 2004). The rate variation among sites was modeled using a gamma distribution (shape parameter = 1). All ambiguous positions were removed for each sequence pair (pairwise deletion option). Codon positions included were 1st + 2nd + 3rd + Noncoding. Evolutionary analyses were conducted in MEGA-X version 10.1.8 (Kumar et al. 2018; Nei and Kumar 2000). The weighted phylogenetic distance was calculated as follows:
where Ri is biomass ratio of native species i in the pot, and Di is phylogenetic distance between native species i and the invasive species.
Linear regressions were performed to test the relationships between native species richness and the biomass of the native species, the biomass of the invasive species and invasion resistance, and the diversity effects (i.e., complementarity effect and selection effect) and the invasion resistance. The regression slope difference between the sterile and non-sterile treatments was tested using ANCOVA. The differences between the sterile and non-sterile treatments at the three- and six-species levels were analyzed using the t-test. The effects of soil microbes and plant species richness on the biomass of native and invasive species were also analyzed by two-way ANOVA. Linear regressions were also performed to evaluate the soil legacy effect of native species richness on the biomass of the invasive species (for the second experiment). The relationships between the weighted phylogenetic distances and the biomass of invasive species were analyzed by linear regression. All analyses were carried out using SPSS 19.0 for Windows (IBM, Armonk, NY, USA).