Competitive interactions drive critical ecological processes in plant communities. Yet how competitive interactions are influenced by polyploidy which has a widespread incidence in plants remains largely unknown.

To evaluate the hypothesis of competitive asymmetry between polyploids and diploids, we set up competing tetraploid and diploid plants of perennial herbaceous Chrysanthemum indicum L. (Asteraceae) at different relative frequencies under contrasting soil water contents. We quantified the interaction intensity between competing plants of the same (intraploidy) and different (interploidy) ploidy levels, and measured functional traits related to gas exchange and plant water use to understand the underlying mechanisms.

The stronger competitive effect of tetraploids on diploids than that of diploids on tetraploids provided evidence for the competitive asymmetry. As a stronger competitor, tetraploids were limited more by individuals of their own than by diploids. Such competitive asymmetry was not only maintained under reduced soil water content, but also translated into higher aboveground biomass of tetraploids. Tetraploids showed more resource-acquisitive traits than diploids under high soil water content and more resource-conservative traits under reduced soil water content. As such, the higher trait plasticity in tetraploids than diploids likely explained the competitive asymmetry.

Synthesis. These results elucidate the nature and magnitude of species interactions between polyploid and diploid plants under changing environments and the underlying mechanisms, and provide important insights into the prevalence and persistence of polyploid plants under a changing climate.

Study system and sampling

The polyploid complex of Chrysanthemum indicum comprises two ploidy levels, diploid (2x = 18 chromosomes) and tetraploid (4x = 36) which likely have allopolyploid origins (Li et al., 2014). The diploids and tetraploids are morphologically similar, but tetraploids are relatively larger in plant and floral sizes (Figure 2). These perennial plants can reproduce both sexually by seed and asexually by rhizomes (Shi et al., 2011). Such asexual reproduction is often more pronounced in tetraploids as they form large clusters of individuals compared to diploids that often grow singly (based on field observations). The diploid and tetraploid populations grow mainly in parapatry in China and often occur in open and dry habitats, especially the tetraploids (Li et al., 2014). The tetraploids that likely experienced a geological period of drought during the Quaternary glaciation (Yang et al., 2006; Li et al., 2014) are more widespread than diploids (Figure 2). In their overlapping region in central China, the diploids and tetraploids are often spatially separated. Within this region in Shennongjia, China, we collected seeds from four diploid populations and six tetraploid populations (Figure 2 and Table S1) in November 2015. As C. indicum is a common species, we did not need permission for collection in this area. Specifically, 3–10 plants were chosen at random in each population (Table S1) considering variation in population size. Then, from each individual, 5–10 flower heads were collected to obtain seeds, and three fresh leaves were stored in moist plastic bags for verifying the ploidy levels using flow cytometry following a previous protocol by Guo et al., (2016).

Seedling cultivation

In March 2016, seeds from different flower heads of the same maternal plant were pooled and sowed in peat substrates (0–10 mm; Novarbo, Lauttakylantie, Finland) in 0.25-L pots. A total of 60 pots for the diploid populations and 72 pots for the tetraploid populations were used, with each pot containing c. 10 seeds from the same maternal family. Seedlings were grown under 25:15ºC day:night temperatures and watered every other day in a naturally lit glasshouse at the Germplasm Bank of Wild Species (Kunming Institute of Botany, China). After one month, we harvested leaves from three seedlings of each pot to confirm seedling ploidy level using flow cytometry. Pots with mixed ploidy levels (N = 8 pots) were excluded from this study, which was likely due to dispersal when populations of the two ploidy levels grow close (Figure 2). In May 2016, seedlings of similar size (c. 5 cm in height) were used for the competition experiment and functional traits experiment at the same time.

Competition experiment

To examine the effects of intraploidy and interploidy competitions, diploid and tetraploid seedlings were grown together at different relative frequencies (varying in 0, 2, 3, or 4 plants per ploidy level in each pot; Figure S3): 0:2, 0:3, 0:4, 2:0, 2:2, 2:3, 2:4, 3:0, 3:2, 3:3, 3:4, 4:0, 4:2, 4:3, 4:4. In this study, due to the concern of seedling mortality under a substantial reduction in soil water content (see watering treatments below), we did not include a single diploid or tetraploid plant growing alone during the experimental setup. These 15 density combinations constituted the basic unit of the competition experiment (Figure S3) and were replicated 10 times, with 720 total plants in 150 7-L pots filled with 1.5 kg custom potting mixture (1:1, peat:perlite). To reduce potential confounding effects of maternal family and/or population-specific responses (Rosche et al., 2018; Wei et al., 2020) on competitive interactions, seedlings from different maternal plants and populations were mixed at random and transplanted for the competition experiment. Soil water content was maintained at field capacity (‘FC’; 0.9 kg water per 1.5 kg potting mixture) by watering every two days for 3 weeks. On June 1, 2016, we started the watering treatments, with 5 replicates maintained at 80% field capacity (FC) and 5 replicates at 20% FC (referred to as ‘normal watering’ vs. ‘reduced watering’, respectively). To maintain the corresponding soil water content, pots were weighed every two days to record water loss and re-water accordingly, which was effective in maintaining the soil water content (Ma et al., 2010) when the weight of the matrix (soil and water) was much higher than plant biomass (kg vs. g). This experiment used a split plot design for the ease of watering, where watering treatments were assigned to whole plots with the replicated 15-density combinations randomized within the whole plots. Pots in the glasshouse were not rotated during the experiment due to the large size of the pots and fragile creeping stems under reduced watering. Thus pot positions (i.e. column and row unique to individual pots in the glasshouse) were considered in data analyses later to account for the potential influence of variation in microenvironmental conditions. The competition experiment was run for 2 months because (1) soil at 20% FC became very dry over the course of the experiment and plants already exhibited some leaf wilting (e.g. in diploids) and (2) we focused on growth rather than mortality-related responses in this experiment.

On July 30, 2016, we harvested the aboveground plant materials by ploidy level in each pot, but not the belowground due to challenges in separating roots between diploid and tetraploid plants. Trials with diploid and tetraploid plants growing separately indicated a similar root-to-shoot ratio between the ploidy levels (both ~1.0 under normal watering and 1.5 under reduced watering), and thus aboveground dry biomass is expected to correlate closely with whole plant dry biomass in both ploidy levels. The harvested plant materials were dried at 80ºC for 72 h to obtain the aboveground dry biomass of diploid and tetraploid plants in each pot. Due to occasional plant mortality during the competition experiment, pots that did not match their original density combination (N = 10) were excluded from data analyses.

Functional traits experiment

We measured a suite of functional traits that are relevant to plant growth and drought tolerance (Table 1) in a separate experiment, to avoid the potential influence of plant tissue collection on the competition experiment (Methods S1). In the functional traits experiment, seedlings of the same ploidy level (i.e. two diploid or two tetraploid seedlings in a pot; Methods S1) were transplanted with a total of 56 pots (112 seedlings) for each ploidy level. Ten ‘empty’ pots filled with the same potting mixture but no seedlings were included for measuring evaporation. As in the competition experiment, on June 1, 2016, half of the experimental (N = 28 pots per ploidy level) and ‘empty’ pots (N = 5) were assigned at random to the normal watering treatment (80% FC) and half to the reduced watering treatment (20% FC) for 2 months. No seedling mortality occurred during this experiment.

Functional traits were measured primarily at the end of the experiment during July 20–30, 2016. Instantaneous gas exchange (maximum photosynthetic rate, A_max; stomatal conductance, g_s; transpiration rate, E) was measured for a random subset of 13 pots per ploidy level per treatment using a LI-6400 portable photosynthesis system (LI-COR, Lincoln, NE, USA). Measurements were taken on a fully expanded green leaf from one plant of each pot between 09:30 and 12:00 h at a saturating irradiance (1200 μmol m^-2 s^-1). The cuvette CO2 concentration was maintained at 400 μmol mol^-1. Relative humidity and leaf temperature were maintained at 50–60% and 20–25ºC, respectively. Instantaneous water use efficiency (WUEi) was calculated as A_max/E.

From the same subset of pots used for gas exchange measurements, 10 pots per ploidy level per treatment were selected for determining leaf mass per unit area (LMA) and relative water content (RWC). Specifically, the largest, fully expanded green leaf from one plant in each pot was excised and weighed immediately for fresh mass. The leaf was scanned and the leaf area was estimated using ImageJ v1.45 (Schneider et al., 2012). The same leaf was then submerged in distilled water overnight to obtain the saturated mass, and dried at 80ºC for 48 h for dry mass. LMA was calculated as leaf dry mass/leaf area, and RWC was calculated as (fresh mass – dry mass)/(saturated mass – dry mass) × 100. In addition, one mature leaf was collected at 06:00 h from each pot (8 pots per ploidy level per treatment) for measuring predawn leaf water potential (Ψ_predawn) using a WP4C Dewpoint Potential Meter (METER Group, Inc., Pullman, WA, USA). Another mature leaf was collected and dried for measuring carbon isotope composition (δ¹³C, ‰) using an isotope mass spectrometer (Thermo Finnigan MAT GmbH, Bremen, Germany) at the Institute of Tibetan Plateau Research.

Plant total water use (TWU) and long-term water use efficiency (WUE_L) were measured in a different subset of 10 pots per ploidy level per treatment, to avoid the influence of destructive leaf collection. As the experimental and ‘empty’ pots were weighed every two days to record water loss (for maintaining soil water content), total water use in each pot was determined as the difference between accumulated water loss in each experimental pot (transpiration and evaporation) and the average accumulated water loss per empty pot (evaporation). Plant TWU was estimated as the average TWU in each pot. To determine WUE_L, plants in each pot were harvested for estimating dry biomass (including both belowground and aboveground). WUE_L was calculated as the ratio between plant dry biomass and TWU (Ma et al., 2010).