We describe a simple, precise, and sensitive experimental protocol for direct measurement of N 2 fixation using the conversion of 15 N 2 to organic N. Our protocol greatly reduces the limit of detection for N 2 fixation by taking advantage of the high sensitivity of a modern, multiple-collector isotope ratio mass spectrometer. This instrument allowed measurement of N 2 fixation by natural assemblages of plankton in incubations lasting several hours in the presence of relatively low-level (ca. 10 atom%) tracer additions of 15 N 2 to the ambient pool of N 2. The sensitivity and precision of this tracer method are comparable to or better than those associated with the C 2 H 2 reduction assay. Data obtained in a series of experiments in the Gotland Basin of the Baltic Sea showed excellent agreement between 15 N 2 tracer and C 2 H 2 reduction measurements, with the largest discrepancies between the methods occurring at very low fixation rates. The ratio of C 2 H 2 reduced to N 2 fixed was 4.68 ؎ 0.11 (mean ؎ standard error, n ‫؍‬ 39). In these experiments, the rate of C 2 H 2 reduction was relatively insensitive to assay volume. Our results, the first for planktonic diazotroph populations of the Baltic, confirm the validity of the C 2 H 2 reduction method as a quantitative measure of N 2 fixation in this system. Our 15 N 2 protocols are comparable to standard C 2 H 2 reduction procedures, which should promote use of direct 15 N 2 fixation measurements in other systems.

Abstract. The surface water of the marine environment has traditionally been viewed as a nitrogen (N) limited habitat, and this has guided the development of conceptual biogeochemical models focusing largely on the reservoir of nitrate as the critical source of N to sustain primary productivity. However, selected groups of Bacteria, inc1uding cyanobacteria, and Archaea can utilize dinitrogen (N2) as an alternative N source. In the marine environment, these microorganisms can have profound effects on net community production processes and can impact the coupling of C-N-P cyc1es as weil as the net oceanic sequestration of atmospheric carbon dioxide. As one component of an integrated 'Nitrogen Transport and Transformations' project, we have begun to re-assess our understanding of (I) the biotic sources and rates of N2 fixation in the world's oceans, (2) the major controls on rates of oceanic N2 fixation, (3) the significance of this N2 fixation for the global carbon cyc1e and (4) the role of human activities in the alteration of oceanic N2 fixation. Preliminary resuIts indicate that rates of N2 fixation, especially in subtropical and tropical open oeean habitats, have a major role in the global marine N budget. Iron (Fe) bioavailability appears to be an important control and is, therefore, critical in extrapolation to global rates of N2 fixation. Anthropogenic perturbations may alter N2 fixation in coastal environments through habitat destruetion and eutrophication, and open ocean N2 fixation may be enhaneed by warming and inereased stratification of the upper water column. Global anthropogenie and c1imatie ehanges mayaiso affeet N2 fixation rates, for example by altering dust inputs (i.e. Fe) or by expansion of subtropieal boundaries. Some recent estimates of global ocean N2 fixation are in the range of 100--200 Tg N (1-2 x 10 14 g N) yr-I , but have large uncertainties. These estimates are nearly an order of magnitude greater than historieal, pre-I 980 estimates, but approach modern estimates of oceanic denitrification. 48

Seagrasses, along with all other marine primary producers, are generally considered to be nitrogen limited. Now experimental enrichments of the tropical seagrass Synngodium fihformeKutz. show that phosphorus, rather than nitrogen, can be the primary limiting nutrient in a marine carbonate environment. Phosphorus enrichment of carbonate sediments resulted in dramatic increases in seagrass growth, biomass, and tissue phosphonis con~position. Additionally, rhizosphere nitrogeil fixation increased in response to phosphorus enrichment, potentially making more nitrogen available to the plants.

Major nitrogen (N) pools and bacterial transformations of N were examined in carbonate sediments of 3 reefs in the central area of the Great Barrier Reef, Australia. Depth distributions of nitrate (NO3-) and ammonium (NH,+) and rates of NH4+ production, N2 fixation (nitrogenase activity by CzHz reduction) and denitrification were measured in muddy sediments of an inshore reef and in fine-. medium-and coarse-grained sediments at a midshelf and shelf edge reef. Ammonium efflux was estimated from pore water profiles. Estimates of potential rates of NH,' and NOs-utilization were made in the upper 2 cm of sediments at the midshelf and shelf edge reefs. Highest concentrations of NH,' (up to 70 pM at 8 cm) were observed in muddy carbonate sediments of inshore Pandora Reef, with somewhat lower concentrations (up to 20 vM) in fine-grained sands of the other 2 sites. Relatively small NH4+ pools, usually less than 10 btM, typified coarse-grained sediments. Nitrate was generally undetectable in these sediments. Rates of NH,+ efflux among sites ranged from 0 to 4 pm01 N m-2 h-', with highest fluxes associated with muds and fine-grained sands. Ammonification rates in the upper 2 cm ranged from 6 to 26 pm01 N m-' h-' among sites, generally increasing with depth. Nitrogenase activity was detected in all sediments examined, with hlghest rates near the surface. N2 fixation could account for more than 50 % of NH,+ production in the upper 2 cm of sediment at 3 of 4 sites. The potential in the upper 2 cm for NH4+ consumption (nitrification and assimilation) ranged from 10 to 60 pm01 N m-' h-', while NO3-reduction potenhal ranged from 10 to 80 pm01 N m-' h-' suggesting these may be quantitatively important pathways. Inhibitor experiments indicated that much of the NH,' uthzation might be by nitrification. Very high nitrification rates [up to 3.8 nmol N (g dry sed.)-' h-' or 70 pm01 N m-' h-'] were confirmed at 1 site by a "N isotope dilution method. Low denitrification rates were also detected in these environments, and in many cases under apparently oxic conditions. However, highest rates noted were less than 5 % of the rate of N o 3 -reduction. While shallow carbonate sands may be poor in organic material, they are active sites of bacterial N transformations. The NH,' and NOs-pools in the upper few cm appear to be highly dynamic, with estimated turnover times of substantially less than 1 d. It is also noteworthy that bacterial N:, fixation appears to account for a much larger fraction of NH4+ turnover than in shallow temperate zone sediments.

The 2 tropical cyanobactenal specles Trichodesmium thiebautii and 7: erythraeum had simlar photosynthetic characteristics in the southwestern Sargasso Sea and Canbbean Sea, with mean rates of hght saturated photosynthesis (uslng O2 electrode) of 42 (SD = 21.3) and 37 (SD = 18.4) mg 0, mg chl a -' h-' at 1410 FE m-2 S-', respectively over a 1300 n mlle cruise track. Rates of dark respiration were high, and the compensation point for both specles was 150 FE m-2 S-' (ca 55 m, m d d a y ) . Estimates of carbon doubllng times (using photosynthetic quotient) were from 3.0 to 3.8 d based on expected photosynthetic rates in the water column. The mean rate of nltrogenase actlvlty at 300 pE m-2 S-' by T thiebautu averaged 0.45 nmol ethylene colony-' h -' , 1.6 times that of 7: erythraeum (p < 0.01) as observed from samples collected on 3 cruises (64 paired observations). Furthermore, in a comparison of nltrogenase actlvitles, at light intensities between ca 500 and 2500 FE m-2 S-', 7: thiebautii was about t w~c e as actlve as T erythraeum. The phycoerythrin content of 7: erythraeum averaged 260 ng colony-', 4.4 times that of T thiebautii, and the mean PE : chl a ratios were 3.2 and 1.2, respectivelyOther pigments: (p-carotene, zeaxanthin, myxoxanthophyll, echinenone, and trace pigments) were s i d a r between the 2 specles. The organization of subcellular inclusions was distinctly different in these 2 species. The high abundance of T thiebautii relative to T erythraeum in many tropical seas may be due to higher rates of N2 fixation and a previously reported neurotoxin in the former specles.