Inorganic nitrogen levels were mostly low in both study areas, often below the sensitivity of field monitoring instruments (ammonium <19 μg/L and nitrite <0.6 μg/L), with the exception of nitrates,

which at the time of the experiment were <0.6 μg/L in Circeo 17-AAG molecular weight and 150.5 ± 3.5 μg/L on average in the Gulf of Gaeta. Average total nitrogen concentrations were 166.0 ± 2.1 μg/L in Circeo and 291 ± 7 μg/L in the Gulf of Gaeta (all chemical data from ARPA) and the average temperature was ca. 15 °C in the two areas. The sea bed in the two areas was characterized by variable proportions of mud, sand and rock. In each study area, several sampling sites were chosen at two bathymetries (5 m and 12 m) on inshore-offshore transects: 6 sampling sites on 3 transects

in Circeo and 16 sampling sites on 8 transects in the Gulf of Gaeta. Transect positioning in the Gulf was based on remote-sensing hydrological surveys to identify areas with a high probability of being affected by inputs from urban, agricultural and livestock-rearing outflows due to superficial run-off and river drainage (see Fig. 1 and Supplemental Material for Method details). This allowed us to identify the main outflow Bleomycin routes and, subsequently, four subareas in the Gulf, hereafter called (from north-west to south-east) Vendicio, Formia, Scauri and Garigliano (Fig. 1). Fronds of U. lactuca, widely present in less wave-exposed intertidal coastal areas of both locations, and C.amentacea var. stricta, occurring in the supralittoral fringe of the reference location, were both collected from the reference location on 18 March 2012 and randomly deployed in replicates

at all sampling sites on 19 March. A small fragment from each frond of each species was cut and conserved (at −80 °C) before deployment and was subsequently used to determine the natural intraspecific variability of the initial δ15N value (T0) and to allow the final value of each sample to be compared to its corresponding initial value. The remaining fragments were singly housed in rigid plastic cages (1 cm mesh), which were tagged and suspended in the water column at ∼70% light (Secchi DNA ligase disk depth = 2–6 m) about 50–90 cm below the water surface, using a combination of buoys, ropes and weights. In each sampling site three replicate plastic cages with U. lactuca and three with C. amentacea were submerged. Since the U. lactuca and C. amentacea cages deployed at the two sampling sites closest to the fish farm were removed twice by persons unknown, comparison in the northern Vendicio area relied only on samples from the other two sampling sites, which were located more than 1.2 km away from the farm. After 48 h of submersion (T1), time enough for complete turnover of N in U. lactuca according to literature ( Runcie et al., 2003) and for δ15N equilibrium according to our preliminary tests (see Supplemental Material), samples were collected and transported to the laboratory in an ice-box.