Supplement

Fig. S1.

Fig. S1: Means (± SD, n = 6) of the annual average temperature (a), the sum of the annual rainfall (b) and the altitude (c) at the origin of the six populations from North America (NA), West Europe (WE) and East Europe (EE), respectively. Different lower case letters indicate significant differences between regions with P < 0.05 (Kruskal-Wallis test), ns - not significant.

Fig. S2.

Fig. S2: Weather conditions for Bielefeld [Germany: 52°2'1.14"N, 8°29'43.49"E, 124 m N.N.] in June, July, August and September 2013. Shown are precipitation [mm, total per day] by black bars and temperature [°C, mean per day] by the line. Data were provided by Reinhard Fischer, Oberstufen-Kolleg, Bielefeld. Green triangles indicate the start of the experiment and the end of the first growing period, respectively. The red square marks the end of the second growing period (= experimental end).

Fig. S3.

Fig. S3:Soil moisture content [m³ m-³] of pots in the growth (a) and the regrowth period (b) of Tanacetum vulgarepopulations from North America (NA), West Europe (WE) and East Europe (EE) assigned to different irrigation treatments (control, dry). Weekly determined soil moisture contents are averaged over four to five weeks. The boxes represent the 25% and 75% percentiles, the whiskers extent to the 10% and 95% percentiles, and the dots represent the 5% and 95% percentiles. Medians (solid lines) and means (dashed lines) are shown. There were 35–36 replicates per region and treatment.

Fig. S4.

Fig. S4:Morphological variation of compound leaves of Tanacetum vulgare, which can occur across populations and drought treatments.

Table S1: Fixed and random effects included in linear mixed models for the relative distance plasticity index (RDPI) calculated for traits of the target plant Tanacetum vulgare for the first (growth) and second (regrowth) growing period grown in a common garden (Germany, Bielefeld). Models were computed with restricted maximum likelihood and tested by marginal F-testsa. Significant values are highlighted in bold. The seeds of T. vulgare originated from six seed families, each of six populations from each of the three regions North America (NA), West Europe (WE), and East Europe (EE). The RDPI was calculated for trait values per seed family of plants grown under regular irrigation (control) or a reduced watering regime (dry).

RDPI of traits / Latitude
F df
(Pvalue) / Region
F df
(Pvalue) / Latitude × Region
F df
(Pvalue) / Population²
%
First period (growth)
Shoot biomass
T. vulgare / - / 0.73 2,15
(0.50) / - / < 0.01
Height / 17.920 1,12
(0.001) / 0.360 2,12
(0.705) / 2.654 2,12
(0.111) / 16.56
Specific leaf area¹ / 1.209 2,15
(0.326) / - / < 0.01
Leaflet number³ / - / 0.87 2,15
(0.439) / - / 11.76
Leaflet ratio¹ / - / 0.417 2,15
(0.666) / - / < 0.01
Second period (regrowth)
Shoot biomass
T. vulgare¹ / - / 4.264 2,15
(0.034) / - / 25.06
Ramet distance / - / 0.916 2,15
(0.421) / - / 2.56

a There was a stepwise reduction of the previous full model (latitude, region and their interactions). Only terms concerning the covariate latitude and its interactions with the fixed factor were excluded if they did not significantly improve the model. Significant model terms are indicated by bold p-values. Excluded model terms are not listed or indicated by dashes.

¹ log x+1 transformed

² Random effect, population; explained variation of total residuals [%]

3 not normally distributed

1