Fig. 1. Thirty-four sites broadly distributed across southern Quebec, supporting natural populations of Lemna minor . Map was produced with Datawrapper.
For each site we measured several environmental variables that we expected to be correlated with plant phenotype. Light availability, as percent transmittance of photosynthetically active radiation (PAR), was estimated in situ with the use of BF5 Sunshine Sensor (Delta-T, Burwell, Cambridge, UK) (Paquette et al. 2007). This instrument consists of an array of seven quantum sensors under a semi-shaded hemispherical dome to give estimates of diffused light under any meteorological condition. We took two simultaneous paired measurements, one at the sampling site in question, and a second reference point at a nearby open site (field or road) under full sun. Percent transmittance PAR was then estimated as the ratio of diffused light between the site and the reference measurement. This method has been demonstrated as a reliable and practical alternative to more standard measurement techniques including hemispherical image analysis (Rich et al. 1993, Paquette et al. 2007). Percent transmittance of PAR was estimated at each of the three microsites as well as the center of the site. All measurements were taken 1.5m above the water level, above any aquatic macrophytes or riparian herbaceous plants to obtain an estimate of shading from the canopy cover. These four measurements were then averaged to produce a single estimate of light availability for each site. To measure water nutrient content, we took eight water samples from the center of each site at a depth of 30cm. Total Nitrogen (TN), total Phosphorus (TP), dissolved Nitrogen (DN) and dissolved Phosphorus (DP) were estimated each from two replicate samples. Acid-washed tubes were first rinsed, and then filled with sample water, unfiltered for TN and TP samples, and sterile filtered at 0.45um for DN and DP samples. After sampling, all tubes were stored in a cooler on ice and brought back to the lab for analysis. Samples were then stored at 4oC and processed within 14 days. Water samples were analysed for TN and DN with a continuous flow analyser (OI Analytical Flow Solution 3100 ©) using an alkaline persulfate digestion method, coupled with a cadmium reactor (Patton and J.R. 2003) and for DP using a standard protocol (Wetzel and Likens 2000). TP was measured using colorimetric detection with a spectrophotometer at 890 nm, after digestion with potassium persulfate and the addition of an ammonium molybdate solution (Wetzel and Likens 2000). All samples were analysed at the GRIL, Université du Québec à Montréal (UQAM) analytical laboratory. We used a YSI probe (YSI professional plus, Xylem Inc., Yellow Sprigs, OH, USA.) to measure water temperature and pH at the centre of each site at a depth of 30cm. Full list of environmental correlates can be found in the supplementary information (Table S1).