Study species
For the production of the plant leachates, we selected 12 herbaceous plant species that co-occurred in a field in Taizhou, China (Table S1). We refer to those resident species as the allelopath species. Nine of the allelopaths are native and three of them are naturalized alien resident species. To avoid a possible bias by preferentially selecting resident species with known allelopathic effect, we did not consider a priori information on phytotoxic or allelopathic effects of the allelopath species. However, a posteriori literature searches showed that at least five of the species have documented allelopathic effects (Table S1). To test if the intensity of the allelopathic effect on a plant depends on its origin, we selected as test species six species that are native and six species that are alien to China. None of these 12 test species were included among our 12 allelopath species, and they were selected on the basis that they co-occur across a wide range of habitats in China (Zhang and Ding, 1993). In order to avoid a confounding of status with phylogeny, the native and alien species came from the same five families (Table S2). All six alien species are naturalized in more than 19 of the 34 Chinese regions (23 provinces, 5 autonomous regions, 4 municipalities, 2 special administrative regions) (Yan et al., 2019), and they vary in their degree of invasiveness in China, according to the invasiveness rankings of Yan et al. (2014). Seeds of the 12 allelopaths and the 12 test species were collected from fields in Taizhou (Zhejiang province, China) or obtained from commercial seed companies (Table S2).
Experimental Set-up And Measurements
On the 15th of June 2020, to produce aqueous leachates of the 12 allelopath species, we sowed seeds of each of allelopath species in plastic trays (l × w × h = 54 × 28 × 5 cm) filled with a 1:1 (v:v) mixture of sand and vermiculite (both purchased from Xiaoxuan Horticulture, Hebei province, China). We placed the trays in a growth chamber (day-time temperature: 18–21°C, night-time temperature: 16–20°C, day length: 14 hours, relative humidity: 60%). We transplanted the seedlings, two weeks after emergence, into plastic pots (2 L) filled with a 1:1 (v:v) mixture of sand and vermiculite. We had one seedling per pot, and for each species, we had 10 pots, which we placed in a greenhouse (day-time temperature: 22–25°C, night-time temperature: 18–21°C, day length: 14 hours, relative humidity: 60%). To ensure that we would have sufficient plant material for the production of leachates, we fertilized each pot once a week with 50 ml of a liquid fertilizer for 14 weeks in total. The fertilizer (Woshibao® liquid fertilizer, Woshibao fertilizer sales Co., Ltd, Lu’an, China; N: 50 g/L, P: 30 g/L, K: 50 g/L, Mg: 1.8 g/L, S: 2.2 g/L, micronutrients: 0.2 g/L) was diluted 1:200 with distilled water. The plants were watered ad libitum twice a week.
To increase comparability to most other allelopathy studies (Zhang et al., 2021), we used plant leachates to test for alleopathic effects. On the 17th of September 2020 (i.e. after 14 weeks of growth), we harvested for each of the 12 allelopath species at least five plants (sometimes more when the plants were small). We harvested both the shoot and roots of each plant, and then cut each plant into pieces of 2 cm. The root and shoot pieces of all plants per species were mixed, and we then took for each species a random sample of 300 g of plant material. The sampled plant material was then steeped in 900 mL of distilled water for 24 h at room temperature. The resulting leachates were filtered in two steps. We first removed large plant debris using a layer of Whatman No. 1 filter paper, and after that we used a 0.8 µm filter membrane (25 mm in diameter) to remove fungal spores. We used a new filter membrane for each species. The filtrates were collected into separate autoclaved Falcon tubes, and stored at 20°C until use.
To test whether the pure aqueous leachates or their mixtures had allelopathic effects on germination of the seeds of the 12 test species, we did a large germination experiment at Taizhou University, China. A total of 1824 Petri-dishes (diameter: 6 cm) were filled with agar substrate made with the aqueous plant leachates. From the twelve plant leachates, 37 different agar-plant leachate-mixtures were produced at four levels of diversity: 12 single species leachates, 12 mixtures of three species, 12 mixtures of six species and one mixture of all 12 species. In addition, we had one control agar gel without any plant leachate. For the 3-species and 6-species mixtures, the species were randomly selected, with the restriction that each species is part of an equal number of mixtures per diversity category (see Table S1).
For each resident allelopath species and each level of diversity, we mixed 300 ml of pure leachate or 300 ml of a leachate mixture into 600 ml of liquid agar gel. The latter was made by mixing 12 g high-strength agar, 30 g sucrose and 3.225 g Murashige and Skoog culture medium in 1 L distilled water. We adjusted the pH of the agar solution to 6.0 with NaOH or HCL. We poured the agar solution into 250 ml Schott glass bottles, and then autoclaved it for 15 minutes at 120°C and a pressure of 100 kPa. Thereafter, hot agar medium was cooled down without solidifying by placing it for 10 minutes in a 40°C water bath. We thawed the frozen plant leachates by placing them for 30 mintues in a heating cabinet at 40°C. We did not use a higher temperature to minimize the likelihood that some of the allelopathic substances would denature. We once more passed the leachates under sterile conditions through a 0.8 µm filter membrane. We then used the 12 pure species leachates to create 12 three-species, 12 six-species and one 12-species mixture (Fig. S1). So, in total there were 37 leachates. In addition, we made a water control of which the osmolality was adjusted with PEG (see below). For each of the leachates and the control, we mixed 300 ml into 600 ml of the liquid agar solution. We then poured for each of the 38 agar mixtures 10 ml into 48 Petri-dishes (totalling 1824 Petri dishes), after which the agar mixture was allowed to cool down and solidify.
As osmolality might have an effect on the germination of the seeds (Inderjit & Nilsen, 2003; Oduor et al., 2020), it was measured with an osmometer (Wescor 5600, Shanghai Pengqi Scientific Instrument Co., Ltd) for each of the 37 leachate mixtures three times, after the filtration step. Variation in the osmolality among leachate mixtures was not related to diversity (i.e. the number of species used to make the leachate mixtures; Fig. S1). For the control agar gel without plant leachates, we therefore adjusted the osmolality to the mean value calculated across all 37 plant leachates (0.049 Osmol/kg). To achieve this, the agar without plant leachate was mixed with a 8000-polyethylenglycol solution (8000-PEG; Sigma-Aldrich, Steinheim, Germany), which had a concentration of 0.091 g PEG/ml. We used PEG, because it changes the osmolality, but cannot be absorbed by plant cells, due to its large molecular size, and is therefore unlikely to have phytotoxic effects (Michel & Kaufman, 1973; Paparella et al., 2015).
In the period from the 6th to the 12th of December 2020, we sowed the seeds of the 12 test species. Before sowing the seeds, we sterilized them by putting them for 30 seconds in 75% alcohol, followed by five minutes in a 5% sodium hypochlorite solution. We then rinsed the seeds with distilled water. In each Petri-dish, we put 10 seeds of one of the 12 test species on the agar gel. For each combination of leachate type (37 plant leachates and 1 PEG control) and test species (n = 12), we had four replicates, resulting in 1824 Petri-dishes. Each Petri-dish was sealed with a strip of parafilm to prevent evaporative water loss. We then randomly assigned the Petri-dishes to positions within a growth chamber (day-time temperature: 21°C, night-time temperature: 16°C, day length: 15 hours, relative humidity: 60%). Each day, we recorded how many seeds had germinated. About two weeks after the last seedling had emerged (on the 16th of January 2021), the experiment was stopped. The collected data were used to calculate the proportion of germinated seeds and the days to first germination. We disgarded 60 Petri-dishes that were strongly infected with a fungus (PEG control: n = 1, diversity 1: n = 17, diversity 3: n = 26, diversity 6: n = 13, diversity 12: n = 3), resulting in a total of 1784 Petri-dishes with germination data.
Statistical analysis
We did all analyses and data visualization in R version 3.6.1 (R Development Core Team 2019). For the analysis of the proportion of germinated seeds (a binomial error distribution) and the number of days to first germination (a Poisson error distribution), we used generalized linear mixed models (GLMMs) as implemented in the glmer function of the lme4 package (Bates, 2014).
The fixed terms in our models were ‘Origin’ (alien vs native) of the test species, ‘Leachate presence’ (control vs plant leachates), ‘Diversity’ of allelopath species in the leachate’ (1, 3, 6 or 12 species) and the interactions of Origin with Leachate presence and Diversity. We initially also included osmolality of the leachate mixture as a covariable in the model, but as its effect was not significant, we removed it from the final model. As we expected, the germination parameters to either increase or decrease with an increasing number of species in the leachate mixture, we included Diversity as a continuous covariable. Diversity was fitted after the effect of Leachate presence, so that it effectively only considered the Petri-dishes with plant leachates (i.e. not the control). We accounted for variation among test species by including species identiy as a random term. We initially also included family of the test species as a random term, but as the variance explained by family was very low and its inclusion resulted in singularity warnings, we removed it from the final models. Furthermore, as each of the 38 leachate mixtures (including the control) was used multiple times, we accounted for non-independence among Petri-dishes with the same leachate mixture by including leachate mixture as a random term. To test which of the allelopath species significantly affect the proportion of germinated seeds and the days to first germination, we also analysed the subset of data including only the control and the single-species leachates. In this analysis, allelopath species was included as a fixed term, and test species as a random term. In all GLMMs, we assessed the significance of the fixed terms with log-likelihood-ratio tests (Zuur et al., 2009), in which a model with the term of interest was compared to a model without that term. Log-likelihood ratios are approximately χ2-distributed (Zuur et al., 2009).