SamplingIn order to compare the commonly used 15 N 2 tracer addition method to measure N 2 fixation 1 with the addition of 15 N 2 -enriched water as suggested by Mohr et al. (2010) 2 , seawater was sampled on two cruises in the Atlantic Ocean, the first on board R/V Meteor (M80/1) on a longitudinal transect (23°W) between 15°N and 5°S, the second on board R/V Polarstern (ANT-XXVI/1) on a transect between 54°N and 54°S (Bremerhaven, Germany to Punta Arenas, Chile). In total 39 triplicate incubations were conducted with both methods in parallel. On the M80/1 cruise, seawater was sampled at 11 stations from the surface (bucket), 20 m depth and the chlorophyll maximum (CTD rosette sampler) at 7:00 in the morning, whereas on the ANT-XXVI/1 cruise, seawater was sampled at 6 stations at 16:00 from the ship's clean seawater supply which is installed at 11 m depth (keel of the ship). IncubationsSeawater samples were filled headspace-free (bubble addition method) or with a 100-150 ml headspace (dissolved method) into 4.5 L polycarbonate bottles and closed with Teflon ® -coated butyl rubber septum caps. To determine N 2 fixation rates with the bubble-addition method, a 4.5 mL 15 N 2 gas bubble (Sigma-Aldrich, ≥98 atom%) was injected through the septa into each of triplicate bottles (yielding a theoretical enrichment of ~12 atom% assuming a rapid isotopic equilibration between the added 15 N 2 gas and the ambient dissolved N 2 of the water sample). After injection, bottles were gently inverted one hundred times. For comparison of N 2 fixation rates, we added 15 N 2 -enriched seawater to a second set of triplicate bottles (dissolution method). In detail, the preparation of the 15 N 2 -enriched seawater was started by degassing filtered seawater (0.2 µm filtered, Durapore) using a membrane flowthrough system (Mini-Module, Membrana) in which the seawater flowed on the inside of the membrane and a vacuum (-960 mbar, water jet pump) was applied to the outer side of the membrane. The seawater flow rate was about 400 -500 mL min -1 and seawater was recirculated for the first 10-15 min of the degassing step. Degassed seawater was then filled directly from the flow-through system into evacuated gas-tight 3L Tedlar® bags without a headspace. Addition of 15 N 2 gas was dependent on the amount of seawater in the Tedlar® bag and was added at a ratio of 10 ml 15 N 2 per 1L seawater. The volume of degassed seawater in SUPPLEMENTARY INFORMATION