IntroductionThe Southern Ocean (SO) is the largest upwelling region inthe world ocean but it is also the only one where the bulk ofthe upwelling macronutrients nitrate and phosphate isreturned with downwelling surface water to the oceaninterior. In striking contrast, the bulk of the upwelling silicicacid is taken up by diatoms in the surface mixed layer acrossthe Antarctic Circumpolar Current (ACC) and sinks to thedeep ocean as biogenic silica (Falkowski et al. 1998). Mostof it dissolves underway and accumulates as silicic acid inCircumpolar Deep Water (CDW). A fraction is buried asdiatom ooze in the underlying sediments. This opal belt isthe largest single sink for Si in the ocean (Tréguer et al.1995). The decoupling between N and P cycles from that ofSi is reflected in declining silicic acid concentrations, butnot nitrate, in the winter-mixed upper layer across the ACCwith Si:N ratios ranging from about 2 in the Antarctic Zone(AZ) in the south, to 0.3 in the Polar Frontal Zone (PFZ)between the Antarctic Polar Front (APF) and the sub-Antarctic Front (SAF) (Falkowski et al. 1998). So thecontemporary SO is a strong silicon sink (Boyle 1998) withsignificant effect on nutrient inventories and ratios in theentire world ocean (Brzezinski et al. 2003, Sarmiento et al.2004). This implies that the ecology of the surface layer ofthe ACC has a strong effect on the global silicon cycle butlittle effect on carbon, nitrogen and phosphorus cycles. The inefficient utilisation of N and P is due to the lowproductivity of Antarctic phytoplankton for which threeindependent but interacting factors are held responsible:growth limitation due to unfavourable light climate becauseof deep mixing or trace element deficiency, and highmortality due to grazing pressure by proto- andmetazooplankton in excess of the growth rate ofphytoplankton (Hart 1942, Chisholm & Morel 1991). Therelative importance of these mutually inclusive explanationsis still under debate but in recent years evidence hasaccumulated indicating that iron availability is a key factor.Extensive measurements of iron concentrations (de Baar &Boyd 2000) and the results of all in situ iron fertilizationexperiments (Boyd et al. 2000, Gervais et al. 2002, Coaleet al. 2004) have confirmed that phytoplankton growth ratesin the areas of the SO remote from land are limited by ironavailability. Similarly, the higher productivity of coastal andshelf waters can be attributed to iron input from land run-offand contact with the sediments (Martin 1990).The Si:N ratio is reported to be up to twofold higher iniron-deficient as compared to iron-sufficient diatoms(Takeda 1998, Hutchins & Bruland 1998) which impliesthat the former make thicker frustules (Boyle 1998) or thattheir cells contain relatively less plasma than iron-sufficientAntarctic Science 16 (4): 541–558 (2004) © Antarctic Science Ltd Printed in the UK DOI: 10.1017/S0954102004002317541The role of grazing in structuring Southern Ocean pelagicecosystems and biogeochemical cyclesVICTOR SMETACEK*, PHILIPP ASSMY and JOACHIM HENJESAlfred Wegener Institute for Polar and Marine Research, D-27570 Bremerhaven, Germany*corresponding author: [email protected]: This review examines the links between pelagic ecology and ocean biogeochemistry with anemphasis on the role of the Southern Ocean in global cycling of carbon and silica. The structure andfunctioning of pelagic ecosystems is determined by the relationship between growth and mortality of itsspecies populations. Whereas the key role of iron supply in conditioning the growth environment of land-remote oceans is now emerging, the factors shaping the mortality environment are still poorly understood.This paper addresses the role of grazing as a selective force operating on the structure and functioning ofpelagic ecosystems within the larger conceptual framework of evolutionary ecology. That mortality due tograzing decreases with increasing cell size is widely taken for granted. We examine the impact of thisprinciple across the range of size classes occupied by Southern Ocean plankton and show that relatively fewspecies play crucial roles in the trophic structure and biogeochemical cycles of the Southern Ocean. Underiron-sufficient conditions, high growth rates of weakly silicified diatoms and Phaeocystis result in build-upof blooms that fuel “the food chain of the giants” (diatoms-krill-whales) and drive the carbon pump. Incontrast, high grazing pressure of small copepods and salps on the regenerating microbial communitiescharacteristic of the iron-limited Southern Ocean results in accumulation of large, heavily silicified diatomsthat drive the silicon pump. The hypotheses we derive from field observations can be tested with in situ ironfertilization experiments.Received 19 January 2004, accepted 31 August 2004Key words: evolutionary ecology, iron fertilization, iron-limited systems, iron-replete systems, organic carbon pump, silicon pumpdiatoms (Franck et al. 2003), or both. Such a physiologicalmechanism could conveniently account for the decouplingbetween N and Si in the iron-limited ACC (Sarmiento et al.2004) but it is a proximate explanation that needs to beexamined in an ultimate or evolutionary framework. Thus,frustule thickness, manifested in the Si:N ratio, is anintrinsic morphological property of diatom species(Brzezinski 1985) that can vary fivefold, even betweenspecies of the same genus. So frustule thickness will have tohave some ecological significance. Interestingly, diatom biomass in the iron-depleted, openACC, although low, tends to be dominated by species thatare larger and have thicker frustules or stouter spines thanthe species that dominate the much higher diatom biomassin productive coastal waters (Hasle & Syvertsen 1996). Thisdifference in diatom community composition is reflected inthe underlying sediments. Whereas biogenic silica (BSi),dominated by the thick frustules of a few species,contributes about 50–70% of the sediments underlying thelow productive areas of the ACC, the sediments under theproductive regions have lower BSi (<10%) (Tréguer 2002)but twofold higher organic carbon percentage of bulksediment (Berger & Herguera 1992). Most of the diatomfrustules in coastal sediments are resting spores of thecosmopolitan diatom genus Chaetoceros (Crosta et al.1997). Clearly, the differences in sediment geochemistryand frustule composition between the high and lowproductive regions will be due to