1. Widespread plant species generally have high intraspecific variation in functional traits, which is reflected in their great variety of phenotypes. This variety can result from both genetic differences due to local adaptation and phenotypic plasticity. With high intraspecific variation and nearly global distribution, the common reed Phragmites australis is a suitable model species for studying the underlying mechanisms of intraspecific trait variation. 

2. In this study, 71 genotypes of P. australis from seven phylogeographic groups were transplanted into two replicate common gardens located in very different climates: northern Europe and mid-east Asia. We measured seven functional traits of all these genotypes over the growing season, including shoot height, maximum biomass per shoot, shoot density, node number per stem, leaf lifespan, flowering occurrence and flowering date. Our aim was to assess the relative effects of genetic (phylogeographic origin) and environmental (common garden) status, and interactions between them, on intraspecific variation in functional traits of P. australis. 

3. We found common garden having the strongest influence on most functional traits studied. All traits except flowering occurrence varied significantly across gardens, revealing the important role of phenotypic plasticity on trait variation of P. australis. We also found significant differences in trait variation among the different phylogeographic groups of P. australis and, thus, evidence for genetically determined intraspecific variation in the morphological and life-history traits addressed in this study. All functional traits showed significant (p≤0.0054), albeit minor to moderately explained (R² ≤0.57), latitudinal patterns in both gardens. Covariation of multiple traits was similar in the two gardens. Phenotypic plasticity was trait-specific, and the plasticity of shoot height and maximum biomass per shoot increased towards higher latitude of genotypic origin. Our results indicate that the latitude of origin affects the evolution of functional traits, as well as their phenotypic plasticity. 

4. Since phenotypic plasticity is a crucial mechanism for acclimation and evolution, our findings support the role of gene-based adaptive phenotypic plasticity in plant evolution. The intraspecific spatial variation of functional traits and their phenotypic plasticity can help predict species distribution, persistence and invasion under global climate change.

Common garden setup

This study was conducted outdoors in two experimental gardens: one in Denmark (Aarhus University, 56°13´N; 10°07´E) and the other in China (Shandong University, 36°25´N; 117°26´E). Both gardens were located in open spaces without shading. We selected 71 genotypes in total of P. australis from four different continents, covering seven phylogeographic groups: one European (EU, n=12), one Mediterranean (MED, n=12), one Far East Australian (FEAU, n=12) and four North American (NA) groups, comprising NA Introduced (NAint, n=12), NA Native (NAnat, n=6), LAND (n=11) and DELTA (n=6) (C. Lambertini et al., 2006; Carla Lambertini, Sorrell, et al., 2012) (Figure 1). The LAND-group,  was previously called the Gulf (or GC) type or P. australis var. berlandieri is Haplotype I and is mainly distributed along the Gulf Coast of North America, while the DELTA-group, is Haplotype M1, and is dominant in the Mississippi River Delta (Lambertini et al., 2012; Ward, 2010). EU, FEAU, MED and NAnat groups are native (representing historic adaptation and evolution), while DELTA, LAND and NAint are introduced (representing recent adaptation) in the areas they were sampled (Lambertini et al., 2012; Guo et al. 2013).

Rhizomes from natural stands of different P. australis populations have been collected over several years, genotyped and maintained in the Danish common garden since 2001. Because of potential pot-binding, all genotypes were replanted every third year. In 2014, the rhizomes of each genotype were split into two parts: one part was transplanted into the Danish garden, the other part was transplanted into the Chinese garden. In both gardens, each genotype was planted in a pot (0.65 m diameter and 0.5 m height) containing a substrate of 50% peat soil and 50% quartz-rich sand (granule diameter of 0-2 mm). The pots were spaced 2 m apart from each other and buried into the ground to keep the soil surface at the same level with the surrounding ground. Plants were weeded manually, but no other pest control was undertaken in either garden. An automatic irrigation system using local municipal water was used to water the plants. Small holes in each pot allowed drainage of excess water to keep the water level similar among genotypes. Each pot was fertilized every two weeks from March to December with 100 mL of fertilizer containing 0.1 kg L^-1 macro nutrients (NPK ratio 19:5:19 + MgO) and trace elements.

Functional trait measurements

All measurements were carried out from March 21^st to November 17^th, 2016, on the same dates in both gardens. Shoot height, which was measured from the top of the soil to the tallest shoot tip including the apical leaf, and node number per stem of the five tallest and healthy shoots of each genotype were measured every two weeks. To determine aboveground biomass per shoot, the five tallest and healthy shoots of each genotype were harvested at the end of the growing season when the shoot height had not changed over three consecutive measurement dates, and oven-dried at 80 °C for at least 48 hours. Shoot density, defined as the number of live shoots per pot, was measured once at the beginning and once at the end of the growing season. For leaf lifespan, one young green apical leaf was marked randomly on five different shoots per genotype in early June, recorded as the beginning of leaf development. The onset of senescence was noted as the day the marked leaves started to turn yellow and wither. The number of days from the beginning of leaf development until turning senescent was defined as leaf lifespan.

For flower occurrence, we noted for each genotype whether it had flowered or not during the growing season. Flowering date was expressed as the number of days that had elapsed in 2016 until the date of observation of the first developing panicle.