Observations of nitrate (NO3−) uptake across the photic zone of the tropical and subtropical Atlantic Ocean are described. High NO3− uptake rates were commonly found at depth (>75 m), coincident with the deep chlorophyll maximum and the nitracline, whereas highest rates of carbon fixation were found in near-surface waters (0−50 m). Thus NO3− based estimates of in situ new production (NO3− uptake × 6.6) at depth are in excess of in situ measurements of carbon fixation implying NO3− uptake exceeds production requirements. Such surplus NO3− uptake may ultimately result in the production of dissolved organic nitrogen with its measurement within the particulate phase representing a transient enrichment. When only upper euphotic zone (to 14% irradiance depth) rates of NO3− uptake and carbon fixation are considered, new production is less than total carbon fixation, suggesting, as expected, that additional nitrogen sources are required to support the observed production levels. Comparison of the deep NO3− uptake rates with literature estimates of seasonal NO3− drawdown within the nitracline suggests that the drawdown observed represents only 1% of the actual NO3− uptake occurring in this region of the water column, and hence most NO3− uptake must be recycled locally rather than exported.
 Nitrate (NO3−) is an important macronutrient for phytoplankton growth and one which above all others has the most relevance to attempts at assessing the magnitude of the biological pump [Dugdale and Goering, 1967; Eppley and Peterson, 1979]. Until relatively recently it was believed that measuring the rate of NO3− uptake in the euphotic zone gave, under a steady state assumption, a reasonable estimation of the vertical flux of NO3− into the surface ocean, and allowed an estimation of new production and an approximation of the amount of nitrogen that was exported from the euphotic zone (i.e., export production). This now appears to be overly simplistic given the recognition of several additional nitrogen supply mechanisms [Lipschultz et al., 2002], including atmospheric nitrogen deposition [Baker et al., 2003] and N2 fixation [Capone et al., 2005] as important sources of new nitrogen to the surface ocean. An additional complication is also arising from a growing recognition of euphotic zone nitrification [Dore and Karl, 1996; Lipschultz, 2001; Fernández et al., 2005; Rees et al., 2006; Martin and Pondaven, 2006], which was classically considered to be insignificant in the surface ocean [Dugdale and Goering, 1967].
 The oligotrophic subtropical ocean presents strong challenges for phytoplankton as light and the major source of new nutrients (NO3− from the deep ocean) are vertically separated. With increasing depth irradiance intensities and photosynthetic rates decrease, whereas in parallel, NO3− concentrations increase owing largely to permanent stratification which prevents the homogenous distribution of NO3− but also owing to the biological uptake of NO3−. A strong gradient in NO3− concentrations (i.e., the nitracline) is observed between the well-lit nutrient-poor surface waters and the dimly lit but nutrient-rich deeper waters of the euphotic zone. Critically, the nitracline is commonly found below, and separate from, the base of the mixed layer, implying a biological explanation for the position of the nitracline and that it does not simply represent the boundary between the mixing of NO3− poor and NO3− rich waters. Furthermore, there is now evidence to suggest that the relative position of the nitracline migrates over seasonal timescales as a function of seasonally varying irradiance intensities at depth [Letelier et al., 2004] and that active uptake must occur beneath the mixed layer in a region where irradiance levels are low. The strong nutrient and light gradients have led some researchers to postulate the presence of two distinct layers within the euphotic zone [Jenkins and Goldman, 1985; Goldman, 1988]. These are: (1) a well-mixed upper layer where primary production is supported primarily by regenerated forms of nitrogen such as ammonium (NH4+) and dissolved organic nitrogen (DON) notably urea, and (2) a lower layer where production is predominately supported by new nutrients (i.e., NO3−) and which may therefore be important for carbon sequestration to the deep ocean.
 The aims of the present study were to examine NO3− uptake rates in the tropical and subtropical Atlantic Ocean in terms of (1) vertical patterns and (2) latitudinal variability. In doing so, we have identified an additional complication to the general practice of equating NO3− uptake with new particulate production that is particularly acute at the nitracline, and if left unaccounted for, may greatly overestimate export from the lower euphotic zone.
 Rates of NO3− uptake were determined between the Falkland Islands and the UK along a basin-scale Atlantic Meridional Transect (AMT) cruise during April−May 2004 (Figure 1). Samples were collected from six light depths (97, 55, 33, 14, 1 and 0.1% of surface incident irradiance) during predawn (0200–0400 local time (LT)) deployments of a SeaBird 911 + CTD equipped with 24 × 20 L Niskin bottles and a WetLabs fluorometer. Light depths were determined from the previous days noon (1100−1300 LT) optics cast as well as from assuming that the fluorescence maximum approximated the base of the euphotic zone (∼1% irradiance [after Agusti and Duarte, 1999]). In parallel to measurements of NO3− uptake, rates of carbon fixation via 14C uptake were also simultaneously determined [Poulton et al., 2006]. Samples for the determination of NO3− and chlorophyll a concentrations were collected from 10 to 12 depths (including the six light depths). The methodology for chlorophyll a determination followed Welschmeyer  and further details are given by Poulton et al. .
2.2. Nitrate Concentrations
 Total oxidized nitrogen (NO3− + NO2−) and nitrite (NO2−) at micromolar concentrations were measured colorimetrically using a five-channel Bran and Luebbe AAIII segmented flow autoanalyzer. The analytical methods used were based on work by Brewer and Riley  for NO3− and Grasshoff  for NO2−, and are summarized by Krom et al. . Where ambient concentrations were below the detection limit of the autoanalyzer (<0.1 μmol L−1) nutrient concentrations were measured using a colorimetric segmented flow analytical system with a 2-m-long Liquid Waveguide Capillary cell (World Precision Instruments) and photo-diode detectors providing nanomolar levels of detection; detection limit ∼2 nmol L−1 [Woodward, 2002; Krom et al., 2005]. Sampling was carried out using clean techniques and all samples were analyzed within 2 hours of sampling, with samples held in the dark in a refrigerator until analysis. No samples were frozen, preserved or stored.
2.3. Nitrate Uptake and New Production
 NO3− uptake was measured with the stable isotope tracer technique of Dugdale and Goering , where a 15N labeled NO3− salt is added to volumetrically measured seawater samples at ∼10% or less [Glibert and Capone, 1993] of the ambient NO3− concentration. Water samples from the six light depths were carefully measured into 2-L Nalgene polycarbonate bottles, spiked with 15N labeled KNO3− (99.5% purity, obtained from CIL laboratories) and placed into simulated in situ on-deck incubators within 1 hour of collection. In situ light levels were replicated using light blue and grey neutral density filters (Lee Filters, Hampshire, UK) and incubators were either cooled with sea surface water (upper light depths) or chilled freshwater (1 and 0.1%).
15NO3− additions made during the cruise ranged from 0.1 to 70 nmol L−1 varying with latitude and with depth. Highest additions were made in the temperate North and South Atlantic where surface concentrations of NO3− were in the micromolar range. Oligotrophic subtropical surface waters had maximum additions of 1.2 nmol L−1 while the deeper nitracline waters had additions typically up to 25 nmol L−1 but in rare instances, and exclusive to the deepest samples (0.1% irradiance), additions of 50 nmol L−1 were made when ambient concentrations dictated this was appropriate. Post-cruise analysis has revealed that virtually all spike additions were <5% (mean 2.3%) of the ambient NO3− concentration. The exceptions are those additions made to surface water samples collected at 38°S, which were up to 15% of the ambient NO3− pool and resulted from a sharp reduction in NO3− concentration as we passed through the subtropical convergence. Spike additions are therefore well within suggested guidelines for this type of measurement and perturbation effects are considered minimal.
 All samples were incubated from local dawn to local dusk (10−12 hours) before removal from the incubators and filtration onto preashed (450°C for >4 hours) Whatman 25-mm GFF filters (nominal pore size 0.7 μm). Filtration was performed under dim lighting and gentle vacuum pressure (<200 mbar). Filters were stored frozen (−20°C) until analysis at the Stable Isotope Ratio Mass Spectrometry (SIRMS) laboratory at the National Oceanography Centre, Southampton, using a Eurovector elemental analyzer coupled to a GV Isoprime mass spectrometer (GV Instruments, Manchester, UK). Measurements of PON concentration and 15N atom% enrichment were obtained simultaneously via the mass spectrometer analyses and uptake rates (nmol L−1 h−1) were obtained following the equations described by Dugdale and Goering . Rates of NO3− uptake were subsequently converted to units of carbon by multiplying by a C:N ratio of 6.625:1 [Redfield et al., 1963; Eppley, 1989; Dugdale et al., 1992].
3.1. Hydrography and Chlorophyll a
 The density structure along the transect (Figure 2a) reveals a general tendency for a south to north shoaling of individual isopycnals, while a pronounced and shallower pycnocline is evident at stations in the equatorial region. NO3− profiles reveal a very nutrient depleted surface ocean with concentrations commonly <15 nmol L−1 but occasionally near detection limits (Figure 2b). The nitracline (defined herein by the 1 μmol L−1 NO3− contour in Figure 2b) outcropped at both the southern and northern subtropical fronts (∼40°N/S) and serves as a useful indicator of the spatial extent of the oligotrophic surface waters of the subtropical ocean. Maximum nitracline depths were approximately 160 m in the interior of both subtropical gyres but shoaled to around 40 m at the equator owing to equatorial upwelling. Chlorophyll concentrations along the transect reveal the presence of a deep chlorophyll maximum as a continuous feature of the tropical and subtropical ocean at depths of 100−150 m (Figure 2c). Within the subtropical gyres surface chlorophyll concentrations were typically <0.1 μg L−1, rising to >0.20−0.25 μg L−1 at the deep chlorophyll maximum. Within the equatorial region, the chlorophyll maximum shoaled to depths of 25−75 m and maximum concentrations were >0.4 μg L−1. In temperate latitudes (>40°N/S), maximum chlorophyll concentrations were typically found at the surface as the deep chlorophyll maximum shoals to the sea surface when crossing the subtropical convergences and in parallel with increases in NO3− concentrations in surface waters.
3.2. Nitrate Uptake
Figure 2d presents a contoured section of NO3− uptake. Surface (<50 m) uptake rates are highest at the southern and northern ends of the transect (>5 nmol L−1 h−1), coincident with the shoaling of the nitracline and higher surface NO3− concentrations (Figure 2b). Low surface rates of NO3− uptake (<1 nmol L−1 h−1) were observed in both subtropical gyres, with elevated rates in equatorial surface waters (>2 nmol L−1 h−1). High rates of surface NO3− uptake (up to 5 nmol L−1 h−1) were observed at several stations in the South Atlantic (20°S–25°S), which may reflect a degree of patchiness within the ocean, although neither NO3− concentrations (Figure 2b) nor chlorophyll concentrations (Figure 2c) are similarly elevated to fully support this. With depth rates of NO3− uptake increase from ∼1−2 nmol L−1 h−1 at intermediate depths to ∼5−10 nmol L−1 h−1 at deeper levels, or even higher if in association with the deep chlorophyll maximum (Figure 2c) and with the nitracline (Figure 2b). Deep uptake maxima, or steady increases in NO3− uptake with depth, are a common observation in subtropical waters [e.g., Eppley and Koeve, 1990; Eppley and Renger, 1992; Allen et al., 1996; Varela et al., 2005]. However, as light-dependent photosynthetic rates are typically highest in surface waters [Maranon et al., 2000; Poulton et al., 2006], there is a vertical separation between the region of the water column primarily involved in NO3− uptake and that involved in carbon fixation.
 To assess the relevance of this deep NO3− uptake to export in a classical sense, we consider integrated NO3− uptake over three portions of the water column, (1) the photic zone (surface to a depth of 0.1% surface irradiance); (2) the traditional euphotic zone (surface to a depth of 1% surface irradiance); and (3) the upper euphotic zone (surface to a depth of 14% surface irradiance; see Table 1). When integrating to photic zone depths we obtain a cruise mean NO3− uptake of 12.9 mmol N m−2 d−1 but there is considerable variability with integrated uptake rates ranging from 0.8 to 43.6 mmol N m−2 d−1 (Table 1). When integrations to euphotic zone depths are considered the cruise mean NO3− uptake reduces by 50% to 6.5 mmol N m−2 d−1 (range 0.2–23.1 mmol N m−2 d−1; Table 1). Clearly a considerable amount of NO3− is taken into the particulate phase between the 1 and 0.1% irradiance horizons. Our third integration is restricted to measurements down to the 14% irradiance horizon. We calculate a cruise mean of 1.5 mmol N m−2 d−1 (range 0.1–6.1 mmol N m−2 d−1; Table 1) which represents ∼12% of the uptake calculated across the photic zone and ∼23% across the euphotic zone. The surface ocean (to 14% irradiance levels) therefore accounts for only a small fraction of total NO3− uptake.
Table 1. NO3− Uptake and Carbon Fixation Integrated Over Photic Zone, Euphotic Zone, and Upper Euphotic Zone Depthsa
See text for definitions. New production has been calculated by multiplying NO3− uptake by a C:N ratio of 6.625:1 and is expressed both in terms of an equivalent carbon flux and as a percentage of measured carbon fixation.
Values in italics are not strictly comparable to other values in these columns as they relate to eutrophic conditions whereas the remaining values are from oligotrophic waters.
These are mean values for all stations.
These are mean values for stations between 40°S and 40°N.
 We repeat the above comparisons but restrict ourselves to measurements made within subtropical/tropical latitudes. We exclude stations from outside the subtropical latitudes (>40°S/N) owing to differences in ambient NO3− concentrations (eutrophic versus oligotrophic; see Figure 2b). Mean NO3− uptake in the upper euphotic zone was 0.9 mmol m−2 d−1 (range 0.1–1.3 mmol m−2 d−1). This represents ∼15% of the NO3− uptake calculated across the euphotic zone (euphotic zone mean 6.0 mmol m−2 d−1) and ∼7% of the NO3− uptake calculated across the photic zone (photic zone mean 13.2 mmol m−2 d−1) for this subset of stations (Table 1). From our integrations it is therefore apparent that the majority (∼90%) of NO3− removed from solution across the photic zone of the subtropical ocean is done so in the lower 14% of the photic zone. Of this approximately 60% is removed from the water column between 0.1 and 1% irradiance depths and 40% between 1 and 14% irradiance depths.
4.1. Deep Nitrate Uptake
 Maximum rates of NO3− uptake were typically measured within the deep chlorophyll maximum (Figure 3), a biological feature previously suggested to act as a major trap to the upward diffusion of NO3− [Banse, 1987]. On occasion however, maximum NO3− uptake occurred beneath the deep chlorophyll maximum (e.g., 14.76°N; Figure 3). The position of the deep chlorophyll maximum usually approximated the 1% irradiance depth, but elevated chlorophyll concentrations (relative to surface values) were measurable below this. The consistent measurement of deep NO3− uptake in excess of parallel carbon fixation rates at stations along the transect argues against this being an anomalous or transient result, as suggested by Eppley and Renger , who observed a similar discrepancy at the base of the euphotic zone in southern Californian waters.
 Heterotrophic bacterial uptake of NO3− is thought to be of minor importance representing <10% of total uptake [Kirchman, 1994, 2000; Kirchman et al., 1994] and therefore cannot explain these observations. Variations in the algal C:N uptake ratio and particulate C:N ratio from Redfield stoichiometry are likely to occur in oligotrophic waters [Banse, 1994; Anderson and Pondaven, 2003; Falkowski and Davis, 2004; Christian, 2005]. However, our use of Redfield stoichiometry is broadly in agreement with C:N ratios in particulate material collected along the transect (Figure 4). Our POC/PON measurements do not reveal large distortions from expected stoichiometric ratios (range 5.5–9.3), although the data do reveal a latitudinal gradient with higher C:N ratios (more C) in the South Atlantic compared to the North Atlantic. Consequently, while knowledge of the variability in the algal C:N uptake ratio would allow better estimates of potential carbon flux to be derived from NO3− uptake measurements, it is difficult to reconcile high NO3− uptake and low carbon fixation at depth (i.e., C:NO3− uptake ratios < 1) with known processes resulting in particulate matter with C:N ratios of ∼6. We believe therefore that the observation of high NO3− uptake at depth is unrelated to permanent or transient deviations in the algal C:N uptake ratio to very low levels if the sole explanation is particulate formation. Thus the decoupled uptake ratios must be indicative of alternative processes representing considerable metabolic activity within the NO3− pool that is separated from carbon uptake.
 Luxury uptake of NO3− by vertically migrating phytoplankton capable of creating concentrated internal pools of NO3− [Villareal et al., 1993; Villareal and Lipschultz, 1995; Villareal et al., 1996] is unlikely to be important as the taxa involved (e.g., Rhizosolenia spp.) are relatively rare and poorly sampled by Niskin bottles (i.e., we did not observe the presence of a single Rhizosolenia mat). In any case the net effect of deep NO3− uptake and later surface photosynthesis should produce conceptually reasonable estimates of Redfield balanced new production and carbon fixation. Using a similar argument, vertical mixing is also unlikely to reduce the offsets between integrated new production and carbon fixation rates owing to strong density barriers preventing this from occurring.
 Over the past decade, the production of DON has increasingly been recognized as an important sink for inorganic nitrogen [Bronk et al., 1994; Bronk and Ward, 1999] and therefore traditional 15N (net) uptake rates are considered to be underestimates of true (gross) uptake rates, although the magnitude of the release to the DON pool is debatable [Bronk et al., 1994; Slawyk et al., 1998; Bronk and Ward, 2000; Slawyk et al., 2000; Bronk and Ward, 2005]. Although we have no direct measurement of DON production and cannot comment directly on the magnitude of this process with respect to our results, the recent findings of Bronk and Ward  strongly suggest that in the nitracline region most of the NO3− removed from solution ultimately ends up in the dissolved organic pool rather than in the particulate (and therefore potentially exportable) pool. At the present time we suspect that this may offer the best explanation for the high NO3− uptake rates we observe at depth.
 Assuming that all measured carbon fixation is supported by NO3− uptake and in line with Redfield stoichiometry, we calculate a lower estimate of mean surplus NO3− uptake of 57 ± 26% across the photic zone. As the integration depth is reduced first to the euphotic zone and then to the upper euphotic zone the mean surplus NO3− uptake reduces (12 ± 74% and −92 ± 120%, respectively) becoming a deficit in the surface ocean requiring, as expected, the contribution of additional N sources (e.g., NH4+) to support the observed rates of carbon fixation. Although these calculations are based on column integrated estimates of net NO3− uptake rates and include a large underlying stoichiometric assumption they do fit with the findings of Bronk and Ward  who found that ∼40% and ∼90% of gross NO3− uptake led to DON production in surface and nitracline waters respectively in a study from the Southern California Bight and therefore did not support particulate production.
4.2. New Production, Carbon Fixation, and Export Production
 Owing to the differences in the depth profiles of NO3− uptake (low at surface, high at depth) and carbon fixation (high at surface, low at depth) (Figure 3) and the excessive uptake of NO3− at depth, estimates of new production (integrated NO3− uptake × Redfield C:N) are in excess of carbon fixation (see Table 1). When we consider the differences between new production and carbon fixation over different portions of the water column cruise mean new production estimated over the photic zone represents ∼400% of total carbon fixation (range 41–1463%); mean new production estimated over the euphotic zone represents ∼210% of total carbon fixation (range 12–819%); and mean new production estimated over the upper euphotic zone represents ∼70% of total carbon fixation (range 7–214%). Clearly, when the deep NO3− uptake, in association with the nitracline and deep chlorophyll maximum, is ignored from estimates of new production, much lower and sustainable estimates of new production are provided. Repeating the calculations for the subtropical stations only, we calculate mean new production contributions to total carbon fixation of ∼450, 225 and 50% for the photic, euphotic and upper euphotic depths respectively.
 Several studies from the nitracline and deep chlorophyll maximum regions suggest that NO3− is rapidly taken into the particulate phase with the ultimate fate of this NO3− possibly being DON production rather than exportable PON production [Banse, 1987; Ward et al., 1989; Bronk and Ward, 1999; Raimbault et al., 1999; Letelier et al., 2004; Bronk and Ward, 2005]. Our results confirm that high levels of NO3− uptake can occur at depth (at the nitracline) but probably cannot be linked to an exportable flux as the C:N ratio of the resulting material would be very low, which is not the case (Figure 4).
Letelier et al.  show a seasonal drawdown of NO3− at the nitracline in the North Pacific subtropical gyre and calculated that the summertime removal (equivalent to 36 mmol m−2 yr−1) represented ∼34% of the mean annual PON export reported by Karl et al.  as being 106 mmol N m−2 yr−1. This was taken to indicate the presence of bloom-like conditions at depth due to seasonally varying irradiance levels leading to the rapid sedimentation of material, since local in situ biomass increases were insufficient to match the observed drawdown.
 Separately, Bronk and Ward  demonstrated that 90% of gross NO3− removal at the nitracline in the Southern California Bight region was partitioned into DON. Similarly, in the equatorial Pacific, Raimbault et al.  concluded from the meridional NO3− gradient that 68% of the ambient NO3− may be partitioned into DON. On the basis of a similar rate of partitioning NO3− uptake into the DON pool we have estimated what fraction of the NO3− uptake we measure is recycled as DON relative to that which is sedimented. We have done this by comparing our daily NO3− uptake rates in the nitracline relative to the NO3− drawdown in the nitracline calculated by Letelier et al.  at Station ALOHA. Although comparative studies between the Pacific (HOTS) and the Atlantic (BATS) time series sites [e.g., Brix et al., 2006] demonstrate some important differences in community production we think that the NO3− drawdown at the nitracline of the Pacific is probably of the correct order of magnitude for subtropical oligotrophic environments and as such is applicable to our study region. Furthermore the nitracline migration distance (∼30 m) observed by Letelier et al.  is similar to the (spring−autumn) displacement observed within the AMT database [Painter, 2006]. Integrating our typical nitracline NO3− uptake rate of 10 nmol L−1 h−1 over 30 m and 360 days leads to an estimate of PON production of 2.6 mol N m−2 yr−1. This is about 2 orders of magnitude higher than the drawdown observed at Station ALOHA which led to the 30 m migration of the nitracline and suggests that nearly all of the NO3− uptake we measure is recycled locally with only a small fraction (∼1%) being sedimented and thus leading to the vertical migration of the nitracline. Raimbault et al.  reached a similar conclusion in oligotrophic waters south of the equatorial Pacific (10°S–16°S) finding that only 1–3% of assimilated nitrogen ultimately escaped the surface ocean as particulate material.
 If substantiated by further investigations this would suggest that the two layered ocean hypothesis is far more complex than Jenkins and Goldman  and Goldman  suggest. Although production in the lower layer may be supported by NO3− the fraction of this production that escapes as particulate material is much smaller than NO3− uptake rates imply. It may be prudent therefore to conceptualize the marine nitrogen cycle of the subtropical ocean photic zone as a combination of a surface regenerative (microbial) loop based on NH4+ (and in some areas N2 fixation) and a deeper regenerative loop based on NO3− and nitrification.
 While these simple calculations have considerable uncertainties associated with them, it is clear that a large fraction of the NO3− uptake we measure must be recycled on short timescales, since the seasonal NO3− draw down is much less than uptake rates otherwise suggest. Arguably, therefore, deep NO3− uptake is unimportant for export but nitrification must be important for maintaining the ambient NO3− pool. However, there are two consequences of even these low rates of NO3− export. First, the slow loss of NO3− from the nitracline leads to a downward migration of the nitracline over seasonal timescales that has the effect of moving the nitracline farther from the ocean surface and therefore retarding light-driven new production in the upper water column where most carbon fixation occurs. Its second consequence is that some of the energy required for the ammonification of the DON produced from NO3− uptake must be derived from organic carbon sedimenting from the water column above. An interesting and open question arises: To what extent therefore does the rapid cycling of NO3− to PON to DON and back act as a trap for exported carbon from the surface ocean?
 In this study we report consistent observations of high NO3− uptake concurrent with low levels of carbon fixation around the deep chlorophyll maximum. Clearly these high uptake rates cannot lead to export in any significant way since if they did the nitracline would migrate far more than it does. We hypothesize instead that most is released as DON which is then ammonified and nitrified with only a minor fraction sinking. The C:N ratio of this material is unknown, presumably it is nitrogen rich and as such might be expected to result in relatively low DOC:DON ratios around the thermocline. Careful examination of the DOC:DON ratio around the thermocline from AMT data (X. Pan, unpublished data, 2006) suggests that this is not the case, either because the signal is masked by the extremely long-lived refractory DOC and DON species which dominate the bulk pools [Hopkinson and Vallino, 2005], or because our hypothesis is incorrect. Simple calculations (not shown) demonstrate that our hypothesized signal would represent <0.5% of the ambient DON pool and as such would certainly be masked by the larger refractory pool given current analytical detection limits. A search of culture experimentation literature looking for corroborating evidence indicating organisms from the deep chlorophyll maximum/nitracline release large quantities of N rich DON under near darkness conditions has not revealed any such results presumably because this type of investigation has yet to be performed. However, Bronk and Ward  reported a gradual increase in DON release with depth in a study from Monterey Bay and observations reported byVarela et al.  do support the general idea that NO3− uptake leads to proportionately higher DON production than other nitrogen sources. An alternative pathway is that our hypothesized return mechanism is short-circuited by the release of nitrite, potentially via the processes that form the primary nitrite maximum [Lomas and Lipschultz, 2006] which occurs at a similar depth.
 Subtropical environments are physically rather stable so variations in the seasonal irradiance cycle must be considered a mechanism regulating the seasonal migration of the nitracline. This hypothesis, suggested by Letelier et al. , argues that the nitracline migrates up when seasonal irradiance levels are low and light-dependent uptake at depth declines thereby allowing isopycnal mixing processes to gradually transfer NO3− upward. This implies some degree of light dependency for deep NO3− uptake, which suggests that algal processes (as typified by the deep chlorophyll maximum), rather than bacterial processes, must be of key importance in this region of the water column.
 High rates of NO3− uptake occur in the lower portion of the photic zone at the surface of, and within, the nitracline. The deep chlorophyll maximum appears to be colocated with this uptake although high uptake can occur below it. NO3− uptake rates exceed requirements needed to support local carbon fixation rates hence this NO3− uptake is decoupled from carbon fixation in a Redfield sense. We speculate that a tightly coupled regenerative loop centered upon NO3− and analogous to the NH4+ loop of the upper ocean may be present in the lower water column from which only a very small percentage (∼1%) of particulate material escapes to the deep ocean. Furthermore the presence of this lower regenerative loop may act as a trap for exported carbon from the surface ocean.
 We would like to thank the officers and crew of the RRS James Clark Ross, the technical support of the UKORS technicians and Mike Bolshaw for masterminding the mass spectrometry analyses. This study was supported by the UK Natural Environmental Research Council through the Atlantic Meridional Transect consortium (NER/O/S/2001/00680). This is contribution 141 of the AMT program.