results in the production of two equivalents of alka-linity per mole.14. Sulfate is routinely accurately measured on ODP cruis-es. In contrast, because sediments outgas as cores arebrought up through the water column, concentrationsof CH4in deep-sea porewaters cannot routinely beaccurately estimated from ODP CH4concentration data(headspace analysis). This quality of the CH4data doesnot preclude using these SO42⫺and CH4data (i) tomap downhole profiles of relative CH4abundance, or(ii), as described in the text, to estimate steady-staterates of sulfate-reducing methane oxidation and (byinference) CH4production. The DSDP and ODP ship-board data used for these profiles and maps were editedto remove samples affected by seawater contaminationand sites where spot coring, poor core recovery, orintermittent porewater sampling caused large gaps thatrendered profiles difficult to interpret. Sites with fewerthan three samples in the zone of stable SO42⫺con-centrations were excluded from the SO42⫺map. Be-cause we limited our study to the diffusional realmtypical of open-ocean sediments, this analysis omittedstratigraphic intervals from downhole SO42⫺and CH4records in cases where porewater chemical profileswere unambiguously affected by hydrologic flow orpronounced lithologic breaks, subsurface anhydrite de-posits, or sulfate-enriched brines. Sites identified asprobably affected by CaSO4precipitation in the under-lying basement were also deleted from the SO42⫺map.15. P. N. Froelich et al., Geochim. Cosmochim. Acta 43,1075 (1979).16. M. E. Q. Pilson, An Introduction to the Chemistry ofthe Sea (Prentice Hall, Upper Saddle River, NJ, 1998).17. The one known exception is mud-volcano ODP Site779 in the western Pacific (32), where CH4may bediffusing up into the sediments after being created bythe heating of buried organic matter.18. T. Fenchel, G. M. King, T. H. Blackburn, Bacterial Biogeo-chemistry (Academic Press, San Diego, CA, 1998).19. G. M. King, Geomicrobiol. J. 3, 275 (1984).20. The rate of subsurface SO42⫺reduction at steadystate is equal to the downward flux of SO42⫺into thesediments. At any given site, the depth-integratedSO42⫺flux into the sediment and the reduction rateof SO42⫺as a function of depth can be estimatedfrom the porewater profile of dissolved SO42⫺, po-rosity, and core-based estimates of effective diffusiv-ity (33). Diffusivity depends on the porosity andtortuosity of the sediment (33).21. A. J. Spivack et al., Eos 81, F216 (2000).22. P. Wellsbury et al., Nature 388, 573 (1997).23. For this study, the top of the subsurface biosphere isdefined as 1.5 mbsf. Total subsurface respirationestimates assume SO42⫺⫹ C6H12O63 S2⫺⫹6CO2⫹ 6H2O. For the system where fermentationplus methanogenesis and anaerobic methane oxida-tion occurs, they assume (i) C6H12O63 3CH4⫹3CO2and (ii) SO42⫺⫹ CH43 S2⫺⫹ CO2⫹ 2H2O.Flux uncertainties are 2␴ estimates derived fromstandard deviations in the slopes of the sulfate con-centration profiles, as well as assumptions of 10%uncertainty in diffusivity and 30% relative uncertain-ty in porosity. The latter assumptions are based onstandard deviations of diffusivity and porosity mea-surements at representative sites. Photosynthesis es-timates are from (34).24. D. E. Canfield, Am. J. Sci. 291, 177 (1991).25. C. Niewo¨hner, C. Hensen, S. Kasten, M. Zabel, H. D.Schultz, Geochim. Cosmochim. Acta 62, 455 (1998).26. M. Adler, C. Hensen, S. Kasten, H. D. Schultz, Int. J.Earth Sci. 88, 641 (2000).27. All estimates of cell abundance are acridine orangedirect counts from (6, 35). Uncertainties in per-cellreduction rates are 2␴ estimates derived from the stan-dard deviations in flux estimates ( Table 1) and standarddeviations in the cell counts (typically 30 to 40%).Mean SO42⫺reduction per cell is integrated to thegreatest depth sampled for cell counts at Site 834 (100mbsf ) and Site 1149 (400 mbsf ). Mean SO42⫺reduc-tion per cell at Site 851 is limited to sediments shal-lower than 150 mbsf, because at greater depths at thissite, SO42⫺is introduced by diffusion from the under-lying oceanic crust.28. B. B. Jørgensen, Geomicrobiol. J. 1, 49 (1978).29. C. Knoblauch, B. B. Jørgensen, J. Harder, Appl. Environ.Microbiol. 65, 4230 (1999).30. K. Ravenschlag, K. Sahm, C. Knoblauch, B. B. Jørgensen,R. Amann, Appl. Environ. Microbiol. 66, 3592 (2000).31. Per-cell sulfate reduction rates in pure cultures ofsulfate-reducing bacteria are typically about an orderof magnitude higher than per-cell rates in naturalsulfate-reducing ecosystems because most cells innatural environments are not sulfate reducers. Innatural ecosystems where SO42⫺is the terminalelectron acceptor, sulfate-reducing bacteria reducethe terminal electron acceptor. Cells that undertakeother steps in the ecosystem’s collective metabolicpathways need not be sulfate reducers. For example,in the Arctic coastal sediments used to calculatethese near-surface per-cell reduction rates, sulfate-reducing bacteria constitute 4 to 16% of enumeratedcells and their RNA constitutes 11 to 29% of theprokaryotic RNA (30).32. P. Fryer, J. A. Pearce, L. B. Stokking, Shipboard Scien-tific Party, Proc. ODP Init. Rep. 125, 687 (1990).33. R. A. Berner, Early Diagenesis: A Theoretical Approach(Princeton Univ. Press, Princeton, NJ, 1980).34. M. J. Behrenfeld, P. G. Falkowski, Limnol. Oceanogr.42, 1 (1997).35. B. A. Cragg, R. J. Parkes, personal communication.36. Edited (14) methane and sulfate profile data repre-sented on the maps are available on Science Onlineat www.sciencemag.org/cgi/content/full/295/5562/2067/DC1.37. We thank D. C. Smith, A. Teske, B. B. Jørgensen, andthree anonymous reviewers for thought-provoking andhelpful discussions. Supported by the Joint Oceano-graphic Institutions U.S. Science Support Program andthe NASA Astrobiology Institute. All porewater chemi-cal data were provided by the Ocean Drilling Program.30 July 2001; accepted 8 February 2002Demographic Characteristicsand Population DynamicalPatterns of Solitary BirdsBernt-Erik Sæther,1* Steinar Engen,2Erik Matthysen3In birds and many other animals, there are large interspecific differences in themagnitude of annual variation in population size. Using time-series data on pop-ulations of solitary bird species, we found that fluctuations in population size ofsolitary birds were affected by the deterministic characteristics of the populationdynamics as well as the stochastic factors. In species with highly variable