Study sites
We sampled six meadows just offshore from the cities of Calvi, Lumio, Saint-Florent, Nonza, Macinaggio, and Bastia along the northeastern and northwestern Corsican coasts (Table 8). The west coast, or "old Corsica", is dominated by granite rock, whereas the northeast coast, or "alpine Corsica", is dominated by shale. The six investigated meadows have different environmental characteristics that encompass the range of characteristics found in the Mediterranean Sea (Figure 2). Off the shore of Calvi, at a depth of 10 m, the meadow is dense and continuous and colonizes a sandy substratum with a 2% slope [62]. The meadow at Saint-Florent is patchy with rocks between the stands of P. oceanica. At Nonza, the meadow colonizes a pebbly substrate. The coasts of south Saint-Florent have a slight slope of sand or rock, whereas the slope is steeper and rocky around Nonza. These two meadows are well separated. At Macinaggio and Bastia, the coast is principally rock, and the meadow grows on a sandy plain that is damaged by trawling [34].
The French coasts have been divided into 50 zones ("zones homogènes") by the Agence de l'Eau. Each zone is characterized according to physical criteria (type of coast, currents, etc.), the quality and biodiversity of the ecosystems, human activities (fisheries, fish farming, urbanization, etc.) and anthropogenic input (sewage, dredged materials, etc.) (Table 1).
Sampling for trace metal measurements in living tissues
Ten orthotropic shoots were collected at the Calvi site from February 1988 to December 2004 next to the STARESO station (February and May 1988; March, June, September, and December 1993; February, April, and June 1994; November 2003; and December 2004). In the laboratory, epiphytes were removed with a shard of glass to avoid metal contamination. Leaves from samples collected between 1988 and 1994 were lyophilized and digested using a double boiler, whereas samples collected in 2003 and 2004 were subjected to microwave digestion. For the samples collected between 1988 and 1994, the levels of Cr, Ni, Cu, Zn, Cd, and Pb were measured with an inductively coupled plasma-atomic emission spectrometer (ARL-3510). Recovery ranges with certified QUASIMEME materials (QTM041BT, QTM042BT) were 88 ± 1.5 % to 99 ± 1% for these metals. Samples collected from 2003 and 2004 were analyzed by a certified laboratory (Institut Malvoz-Laboratoire Santé et Cadre de Vie, Liege, Belgium) with an inductively coupled plasma-mass spectrometer (Elan DCR II). A set of certified Material samples (DORM-2, National Research Council, Institute for National Measurement Standards, Ont; Canada) spiked with grade concentration of metals was analyzed to ensure the accuracy of metals. The metal concentrations were expressed as μg.gDW-1. The change of measuring apparatus is due to a breakdown of the apparatus which we use in routine (ICP AES ARL-3510). All the analyses were carried out by our technician and all the procedure of preparation of the samples was carried out in our laboratory. Rigorous protocols have been followed and intercalibrations have been made. For example, for the Zn concentrations, tests have been realized in the Calvi leaves (2003 and 2004); a correlation test of Spearman rank has been made which show a correlation coefficient of 0,997 at a Pvalue of 0,001. In spite of this breakdown which obliged us has to use another apparatus, our results are comparable and can be regarded as resulting from only one and single series.
Sampling for trace metal measurements in dead sheaths
At each site during November 2001 and February 2002, density estimates (n = 10) were made at a 10-m depth by a diver [63]. Fifteen shoots from P. oceanica plants were collected to measure biometric parameters, and 15 orthotropic rhizomes were collected at each site for measurement the trace metal levels.
Using lepidochronology, it is possible to use the dead sheaths from P. oceanica to examine the history of metal concentrations in the environment over several decades (Figure 1). The sheath is the basal part of the leaf that remains attached to the rhizome after abscission of the leave apex (blade). The technique of lepidochronology, which is analogous to dendrochronology, is derived from the life cycle of the phanerogam sheaths, which have an annual periodicity with a maximum and a minimum thickness. Cyclical patterns, therefore, allow each sheath to be assigned a chronological date. In this way, it is possible to measure metal concentrations in each sheath and recreate the temporal history of metal concentrations in the environment [6]. Throughout this work, when we use the word "sheath" alone, it refers to a dead sheath.
Biometry
In the laboratory, shoots were dissected as follows: the leaves were separated and measured, scraped to remove epiphytes [64], and the biomass (gDW shoot-1) was measured after lyophilization. The leaf area (cm2 of leaf area shoot-1) and leaf area index (m2 of leaves per m2 of substrate) were calculated according to Giraud [30].
Water samples
Sediment pore water and water column samples were collected with a syringe for analysis of nutrients (NH4+, NO3-, and PO4-) [62]. Nutrient concentrations were determined using a SAN SKALAR autoanalyzer. The results are expressed in μM, and the detection limit was 0.01 μM for the three nutrients.
Dead sheaths
The orthotropic rhizomes were dissected for lepidochronological analysis [19]. The sheaths of three shoots were pooled according to lepidochronological year, pulverized, and lyophilized. A portion of the sample (155 mg) was digested with a mixture of 1 ml of concentrated HNO3, 0.1 ml of H2O2, and 0.25 ml of deionized distilled H2O. The mixture was placed in a Teflon bomb and microwaved for 5 min at 300 W, 30 s at 600 W, and 4 min at 250 W. Concentrations of Cr, Ni, Cu, Zn, As, Se, Cd, and Pb were measured in sheaths using an inductively coupled plasma-mass spectrometer and are expressed in μg.gDW-1.
Statistical analysis
The trace metal concentrations in living tissues from the Calvi site samples were compared using Repeated Measures Analysis of Variance to determine the significance of differences between the means for different samples. When a difference was detected, a Tukey's HSD post hoc comparison was used to determine whether the difference in mean values was significant. Normality and homoscedasticity were tested using the Kolmogorov-Smirnof test and Bartlett's test, respectively. Data were log-transformed when necessary.
Different tests were used for the trace metal concentrations in dated sheaths from the northern part of Corsica. To test temporal evolution, Repeated Measures Analysis of Variance was used. When a significant difference between samples was detected, Tukey's HSD post hoc comparison was used. Normality and homoscedasticity were tested with the Kolmogorov-Smirnof test and Bartlett's test, respectively. Data was log-transformed when necessary. To test the significance of differences of trace metals between sites, the Kruskall-Wallis test followed by multiple comparisons based on the Kruskall-Wallis rank sums was used.
To compare relative metal concentrations between dated sheaths and living leaves at Calvi, a Mann-Whitney U Test was used. The results were considered statistically significant at P ≤ 0.01.
Answers to all your Biology Questions
