(Appendix) Vertebral number of topsmelt, Atherinops affinis, adults and offspring from different latitudes along the North-American Pacific coast

Organisms that are distributed across spatial climate gradients often exhibit adaptive local variations in morphological and physiological traits, but to what extent such gradients shape evolutionary responses is still unclear. Given the strong natural contrast in latitudinal temperature gradients between the North-American Pacific and Atlantic coast, we asked how increases in vertebral number (VN, known as Jordan's Rule) with latitude would differ between Pacific (Atherinops affinis) and Atlantic Silversides (Menidia menidia), two ecologically equivalent and taxonomically similar fishes with similar latitudinal distributions. VN was determined from radiographs of wild-caught adults (genetic + environmental differences) and its genetic basis confirmed by rearing offspring in common garden experiments. Compared to published data on VN variation in M. menidia (a mean increase of 7.0 vertebrae from 32 to 46°N, VN slope = 0.42/lat), the latitudinal VN increase in Pacific Silversides was approximately half as strong (a mean increase of 3.3 vertebrae from 28 to 43°N, VN slope = 0.23/lat). This mimicked the strong Atlantic (1.11°C/lat) versus weak Pacific latitudinal gradient (0.40°C/lat) in median annual sea surface temperature (SST). Importantly, the relationship of VN to SST was not significantly different between the two species (average slope = -0.39 vertebrae/°C), thus suggesting a common thermal dependency of VN in silverside fishes. Our findings provide novel support for the hypothesis that temperature gradients are the ultimate cause of Jordan's Rule, even though its exact adaptive significance remains speculative. A second investigated trait, the mode of sex determination in Atlantic versus Pacific Silversides, revealed patterns that were inconsistent with our expectation: M. menidia displays temperature-dependent sex determination (TSD) at low latitudes, where growing seasons are long or unconstrained, but also a gradual shift to genetic sex determination (GSD) with increasing latitude due to more and more curtailed growing seasons. Sex ratios in A. affinis, on the other hand, were independent of latitude and rearing temperature (indicating GSD), even though growing seasons are thermally unconstrained across most of the geographical distribution of A. affinis. This suggests that additional factors (e.g., longevity) play an important role in shaping the mode of sex determination in silverside fishes.

(Table 1) Tree age, post fire stand age, and soil organic matter thickness in the western Quebec black spruce boreal ecosystem

The observed long-term decrease in the regional fire activity of Eastern Canada results in excessive accumulation of organic layer on the forest floor of coniferous forests, which may affect climate-growth relationships in canopy trees. To test this hypothesis, we related tree-ring chronologies of black spruce (Picea mariana (Mill.) B.S.P.) to soil organic layer (SOL) depth at the stand scale in the lowland forests of Quebec's Clay Belt. Late-winter and early-spring temperatures and temperature at the end of the previous year's growing season were the major monthly level environmental controls of spruce growth. The effect of SOL on climate-growth relationships was moderate and reversed the association between tree growth and summer aridity from a negative to a positive relationship: trees growing on thin organic layers were thus negatively affected by drought, whereas it was the opposite for sites with deep (>20-30 cm) organic layers. This indicates the development of wetter conditions on sites with thicker SOL. Deep SOL were also associated with an increased frequency of negative growth anomalies (pointer years) in tree-ring chronologies. Our results emphasize the presence of nonlinear growth responses to SOL accumulation, suggesting 20-30 cm as a provisional threshold with respect to the effects of SOL on the climate-growth relationship. Given the current climatic conditions characterized by generally low-fire activity and a trend toward accumulation of SOL, the importance of SOL effects in the black spruce ecosystem is expected to increase in the future.

Age model for ODP Leg 175 sites

Upwelling along the western coast of Africa south of the equator may be partitioned into three major areas, each having its own dynamics and history: (1) the eastern equatorial region, comprising the Congo Fan and the area of Mid-Angola; (2) the Namibia upwelling system, extending from the Walvis Ridge to Lüderitz; and (3) the Cape Province region, where upwelling is subdued. The highest nutrient contents in thermocline waters are in the northern region, the lowest in the southern one. Wind effects are at a maximum near the southern end of the Namibia upwelling system, and maximum productivity occurs near Walvis Bay, where the product between upwelling rate and nutrient content of upwelled waters is at a maximum. In the Congo/Angola region, opal tends to follow organic carbon quite closely in the Quaternary record. However, organic carbon has a strong precessional component, while opal does not. Despite relatively low opal content, sediments off Angola show the same patterns as those off the Congo; thus, they are part of the same regime. The spectrum shows nonlinear interference patterns between high- and low-latitude forcing, presumably tied to thermocline fertility and wind. On Walvis Ridge, as in the Congo-Angola region, the organic matter record behaves normally; that is, supply is high during glacial periods. In contrast, interglacial periods are favorable for opal deposition. The pattern suggests reduction in silicate content of the thermocline during glacial periods. The reversed phase (opal abundant during interglacials) persists during the entire Pleistocene and can be demonstrated deep into the Pliocene, not just on Walvis Ridge but all the way to the Oranje River and off the Cape Province. From comparison with other regions, it appears that silicate is diminished in the global thermocline, on average, whenever winds become strong enough to substantially shorten the residence time of silicate in upper waters (Walvis Hypothesis, solving the Walvis Paradox of reversed phase in opal deposition). The central discovery during Leg 175 was the documentation of a late Pliocene opal maximum for the entire Namibia upwelling system (early Matuyama Diatom Maximum [MDM]). The maximum is centered on the period between the end of the Gauss Chron and the beginning of the Olduvai Chron. A rather sharp increase in both organic matter deposition and opal deposition occurs near 3 Ma in the middle of the Gauss Chron, in association with a series of major cooling steps. As concerns organic matter, high production persists at least to 1 Ma, when there are large changes in variability, heralding subsequent pulsed production periods. From 3 to 2 Ma, organic matter and opal deposition run more or less parallel, but after 2 Ma opal goes out of phase with organic matter. Apparently, this is the point when silicate becomes limiting to opal production. Thus, the MDM conundrum is solved by linking planetary cooling to increased mixing and upwelling (ramping up to the MDM) and a general removal of silicate from the upper ocean through excess precipitation over global supply (ramping down from the MDM). The hypothesis concerning the origin of the Namibia opal acme or MDM is fundamentally the same as the Walvis Hypothesis, stating that glacial conditions result in removal of silicate from the thermocline (and quite likely from the ocean as a whole, given enough time). The Namibia opal acme, and other opal maxima in the latest Neogene in other regions of the ocean, marks the interval when a cooling ocean selectively removes the abundant silicate inherited from a warm ocean. When the excess silicate is removed, the process ceases. According to the data gathered during Leg 175, major upwelling started in the late part of the late Miocene. Presumably, this process contributed to the drawing down of carbon dioxide from the atmosphere, helping to prepare the way for Northern Hemisphere glaciation.

Os isotope ratios of Late Eocene sediments from the tropical Pacific

High resolution records (ca. 100 kyr) of Os isotope composition (187Os/188Os) in bulk sediments from two tropical Pacific sites (ODP Sites 1218 and 1219) capture the complete Late Eocene 187Os/188Os excursion and confirm that the Late Eocene 187Os/ 188Os minimum, earlier reported by Ravizza and Peucker-Ehrenbrink (2003, doi:10.1016/S0012-821X(03)00137-7), is a global feature. Using the astronomically tuned age models available for these sites, it is suggested that the Late Eocene 187Os/188Os minimum can be placed at 34.5 +/- 0.1 Ma in the marine records. In addition, two other distinct features of the 187Os/188Os excursion that are correlatable among sections are proposed as chemostratigraphic markers which can serve as age control points with a precision of ca. +/-0.1 Myr. We propose a speculative hypothesis that higher cosmic dust flux in the Late Eocene may have contributed to global cooling and Early Oligocene glaciation (Oi-1) by supplying bio-essential trace elements to the oceans and thereby resulting in higher ocean productivity, enhanced burial of organic carbon and draw down of atmospheric CO2. To determine if the hypothesis that enhanced cosmic dust flux in the Late Eocene was a cause for the 187Os/188Os excursion can be tested by using the paired bulk sediment and leachate Os isotope composition; 187Os/188Os were also measured in sediment leachates. Results of analyses of leachates are inconsistent between the south Atlantic and the Pacific sites, and therefore do not yield a robust test of this hypothesis. Comparison of 187Os/188Os records with high resolution benthic foraminiferal delta18O records across the Eocene-Oligocene transition suggests that 187Os flux to the oceans decreased during cooling and ice growth leading to the Oi-1 glaciation, whereas subsequent decay of ice-sheets and deglacial weathering drove seawater 187Os/188Os to higher values. Although the precise timing and magnitude of these changes in weathering fluxes and their effects on the marine 187Os/188Os records are obscured by recovery from the Late Eocene 187Os/188Os excursion, evidence of the global influence of glaciation on supply of Os to the ocean is robust as it has now been documented in both Pacific and Atlantic records.

Middle Eocene bulk sediment, benthic and planktic foraminiferal stable isotope values, and relative weight percent CaCO3 content from ODP Site 207-1260

Major ice sheets were permanently established on Antarctica approximately 34 million years ago, close to the Eocene/ Oligocene boundary, at the same time as a permanent deepening of the calcite compensation depth in the world's oceans. Until recently, it was thought that Northern Hemisphere glaciation began much later, between 11 and 5million years ago. This view has been challenged, however, by records of ice rafting at high northern latitudes during the Eocene epoch and by estimates of global ice volume that exceed the storage capacity of Antarctica at the same time as a temporary deepening of the calcite compensation depth 41.6 million years ago. Here we test the hypothesis that large ice sheets were present in both hemispheres 41.6 million years ago using marine sediment records of oxygen and carbon isotope values and of calcium carbonate content from the equatorial Atlantic Ocean. These records allow, at most, an ice budget that can easily be accommodated on Antarctica, indicating that large ice sheets were not present in the Northern Hemisphere. The records also reveal a brief interval shortly before the temporary deepening of the calcite compensation depth during which the calcite compensation depth shoaled, ocean temperatures increased and carbon isotope values decreased in the equatorial Atlantic. The nature of these changes around 41.6 million years ago implies common links, in terms of carbon cycling, with events at the Eocene/Oligocene boundary and with the 'hyperthermals' of the Early Eocene climate optimum. Our findings help to resolve the apparent discrepancy between the geological records of Northern Hemisphere glaciation and model results that indicate that the threshold for continental glaciation was crossed earlier in the Southern Hemisphere than in the Northern Hemisphere.

Stable oxygen and carbon isotopes from ODP sites on Maud Rise and Kerguelen Plateau

Paleogene stable oxygen and carbon isotopes were measured in formainifera from ODP Sites 689 and 690 at Maud Rise in the Atlantic Ocean sector of the Southern Ocean, and from Sites 738, 744, 748 and 749 at the southern Kerguelen Plateau in the Indian Ocean sector. These data were compared with sedimentological data from the same sample set. Both benthic and planktic d18O values document a cooling trend beginning around 49.5 Ma at all sites. During the late middle Eocene planktic d18O values indicate a steepening latitudinal temperature gradient from 14°C at the northern sites towards 10°C at the southernmost sites. Terrigeneous sand grains of probably ice rafted origin and clay mineral assemblages point to the existence of a limited East Antarctic ice cap with some glaciers reaching sea level as early as middle Eocene time around 45.5 Ma. Between 45 and 40 Ma, average paleotemperatures were between 5° and 7°C in deep and intermediate water masses, while near-surface water masses ranged between 6° and 10°C. During the late Eocene, between 40 and 36 Ma, average temperatures further decreased to 4°-5°C in the deep and intermediate water masses and to 5°-8°C near the sea surface. Abruptly increasing d18O values at approximately 35.9 Ma exactly correlate with a sharp pulse in the deposition of ice-rafted material on the Kerguelen Plateau, a dramatic change in clay mineral composition, and an altered Southern Ocean circulation indicated by a differentiation of benthic d13C values between sites, increasing opal concentrations and decreasing carbonate contents. For planktic and benthic foraminifera this d18O increase ranges between 1.0 and 1.3 per mil, and between 0.9 and 1.4 per mil, respectively. We favour a hypothesis that explains most of the d18O shift at 35.9 Ma with a buildup of a continental East Antarctic ice sheet. Consequently, relatively warm Oligocene Antarctic surface water temperatures probably are explained by a temperate, wet-based nature of the ice sheet. This would also aid in the fast build-up of an ice sheet by enhancing the moisture transport on to the continent.

Gametogenic cycle of Mesodesma donacium

Large-scale environmental patterns in the Humboldt Current System (HCS) show major changes during strong El Niño episodes, leading to the mass mortality of dominant species in coastal ecosystems. Here we explore how these changes affect the life-history traits of the surf clam Mesodesma donacium. Growth and mortality rates under normal temperature and salinity were compared to those under anomalous (El Niño) higher temperature and reduced salinity. Moreover, the reproductive spatial-temporal patterns along the distribution range were studied, and their relationship to large-scale environmental variability was assessed. M. donacium is highly sensitive to temperature changes, supporting the hypothesis of temperature as the key factor leading to mass mortality events of this clam in northern populations. In contrast, this species, particularly juveniles, was remarkably tolerant to low salinity, which may be related to submarine groundwater discharge in Hornitos, northern Chile. The enhanced osmotic tolerance by juveniles may represent an adaptation of early life stages allowing settlement in vacant areas at outlets of estuarine areas. The strong seasonality in freshwater input and in upwelling strength seems to be linked to the spatial and temporal patterns in the reproductive cycle. Owing to its origin and thermal sensitivity, the expansion and dominance of M. donacium from the Pliocene/Pleistocene transition until the present seem closely linked to the establishment and development of the cold HCS. Therefore, the recurrence of warming events (particularly El Niño since at least the Holocene) has submitted this cold-water species to a continuous local extinction-recolonization process.

Age model and stable isotope record of DSDP holes 48-401 and 86-577

Early Paleogene warm climates may have been linked to different modes and sources of deepwater formation. Warm polar temperatures of the Paleocene and Eocene may have resulted from either increased atmospheric trace gases or increased heat transport through deep and intermediate waters. The possibility of increasing ocean heat transport through the production of warm saline deep waters (WSDW) in the Tethyan region has generated considerable interest. In addition, General Circulation Model results indicate that deepwater source regions may be highly sensitive to changing basin configurations. To decipher deepwater changes, we examined detailed benthic foraminiferal faunal and isotopic records of the late Paleocene through the early Eocene (~60 to 50 Ma) from two critical regions: the North Atlantic (Bay of Biscay Site 401) and the Pacific (Shatsky Rise Site 577). These records are compared with published data from the Southern Ocean (Maud Rise Site 690, Islas Orcadas Rise Site 702). During the late Paleocene, similar benthic foraminiferal delta18O values were recorded at all four sites. This indicates uniform deepwater temperatures, consistent with a single source of deep water. The highest delta13C values were recorded in the Southern Ocean and were 0.5 per mil more positive than those of the Pacific. We infer that the Southern Ocean was proximal to a source of nutrient-depleted deep water during the late Paleocene. Upper Paleocene Reflector Ab was cut on the western Bermuda Rise by cyclonically circulating bottom water, also suggesting a vigorous source of bottom water in the Southern Ocean. A dramatic negative excursion in both carbon and oxygen isotopes occurred in the latest Paleocene in the Southern Ocean. This is a short-term (<100 kyr), globally synchronous event which also is apparent in both the Atlantic and Pacific records as a carbon isotopic excursion of approximately 1 per mil. Faunal analyses from the North Atlantic and Pacific sites indicate that the largest benthic foraminiferal faunal turnover of the Cenozoic was synchronous with the isotopic excursion, lending support to the hypothesis that the extinctions were caused by a change in deepwater circulation. We speculate that the Southern Ocean deepwater source was reduced or eliminated at the time of the excursion. During the early Eocene, Southern Ocean delta13C values remained enriched relative to the North Atlantic and Pacific. However, the Southern Ocean was also enriched in delta18O relative to these basins. We interpret that these patterns indicate that although the Southern Ocean was proximal to a source of cool, nutrient-depleted water, the intermediate to upper deep water sites of the North Atlantic and Pacific were ventilated by a different source that probably originated in low latitudes, i.e., WSDW.

Sea surface temperature reconstruction from biomarker in sediments off Portugal

High resolution reconstructions of sea surface temperature (Uk'37-SST), coccolithophore associations and continental input (total organic carbon, higher plant n-alkanes, n-alkan-1-ols) in core D13882 from the shallow Tagus mud patch are compared to SST records from deep-sea core MD03-2699 and other western Iberian Margin cores. Results reveal millennial-scale climate variability over the last deglaciation, in particular during the LGIT. In the Iberian margin, Heinrich event 1 (H1) and the Younger Dryas (YD) represent two extreme episodes of cold sea surface condition separated by a marine warm phase that coincides with the Bølling-Allerød interval (B-A) on the neighboring continent. Following the YD event, an abrupt sea surface warming marks the beginning of the Holocene in this region. SSTs recorded in core D13882 changed, however, faster than those at deep-sea site MD03-2699 and at the other available palaeoclimate sequences from the region. While the SST values from most deep-sea cores reflect the latitudinal gradient detected in the Iberian Peninsula atmospheric temperature proxies during H1 and the B-A, the Tagus mud patch (core D13882) experienced colder SSTs during both events. This is most certainly related to a supplementary input of cold freshwater from the continent to the Tagus mud patch, a hypothesis supported by the high contents of terrigenous biomarkers and total organic carbon as well as by the dominance of tetra-unsaturated alkenone (C37:4) observed at this site. The comparison of all western Iberia SST records suggests that the SST increase that characterizes the B-A event in this region started 1000 yr before meltwater pulse 1A (mwp-1A) and reached its maximum values during or slightly after this episode of substantial sea-level rise. In contrast, during the YD/ Holocene transition, the sharp SST rise in the Tagus mud patch is synchronous with meltwater pulse IB. The decrease of continental input to the mud patch conflrms a sea level rise in the region. Thus, the synchronism between the maximum warming in the mid-latitudes off the western Iberian margin, the adjacent landmasses and Greenland indicates that mwp-lB and the associated sea-level rise probably initiated in the Northern Hemisphere rather than in the South.

Planktonic foraminifera distribution during the mid-Piacenzian warm period in the North Atlantic

The mid-Piacenzian (MP) warm period (3.264-3.025 Ma) has been identified as the most recent time in geologic history during which mean global surface temperatures were considerably warmer than today for a sustained period. This interval has therefore been proposed as a potential (albeit imperfect) analog for future climate change and as such, has received much scientific attention over the past two decades. Central to this research effort is the Pliocene Research, Interpretation, and Synoptic Mapping (PRISM) project, an iterative paleoenvironmental reconstruction of the MP focused on increasing our understanding of warm-period climate forcings, dynamics, and feedbacks by providing three-dimensional data sets for general circulation models. A mainstay of the PRISM project has been the development of a global sea surface temperature (SST) data set based primarily upon quantitative analyses of planktic foraminifer assemblages, supplemented with geochemical SST estimates wherever possible. In order to improve spatial coverage of the PRISM faunal data set in the low and mid-latitude North Atlantic, this study provides a description of the MP planktic foraminifer assemblage from five Ocean Drilling Program sites (951, 958, 1006, 1062, and 1063) in the subtropical gyre, a region critical to Atlantic Ocean circulation and tropical heat advection. Assemblages from each core provide evidence for a temperature- and circulation-driven 5-10° northward displacement of MP faunal provinces, as well as regional shifts in planktic foraminifer populations linked to species ecology and interactions. General biogeographic trends also indicate that, relative to modern conditions, gyre circulation was stronger (particularly the Gulf Stream, North Atlantic Current, and North Equatorial Current) and meridionally broader. A comparison of mid-Piacenzian and modern North Atlantic planktic foraminifer assemblages suggests that low latitude western boundary currents were less than 1 °C warmer while eastern boundary currents were ~1-2 °C warmer, supporting the hypothesis of enhanced northward heat advection along western boundary currents and warming of high latitude Northeast Atlantic source regions for the Canary Current. These findings are consistent with a model of reduced meridional SST gradients, with little-to-no low latitude warming, and more vigorous ocean circulation. Results therefore support the theory that enhanced meridional overturn circulation and associated northward heat advection made an important contribution, in conjunction with elevated atmospheric CO2 concentrations, to the 2-3 °C global surface temperature increase (relative to today) and strong polar amplification of SST warmth during the MP warm period.

Methane concentrations in sediments of ODP sites 186-1150 and 186-1151

Substantial amounts of adsorbed methane were detected in authigenic carbonate concretions recovered from sedimentary layers from depths between 245 and 1,108 m below seafloor during Ocean Drilling Program Leg 186 to ODP sites 1150 and 1151 on the deep-sea terrace of the Japan Trench. Methane contents were almost two orders of magnitude higher in the concretions (291-4,528 nmol/g wet wt) than in the surrounding bulk sediments (5-93 nmol/g wet wt), whereas methane/ethane ratios and stable carbon isotopic compositions were very similar. Carbonate content of surrounding bulk sediments (0.02-3.2 wet wt%) and methane content of the surrounding bulk sediments correlated positively. Extrapolation of the carbonate contents of bulk sediments suggests that 100 wt% carbonate would correspond to 1,886±732 nmol methane per g bulk sediment, which is similar to the average value observed in the carbonate concretions (1,321±1,067 nmol/g wet wt, n = 13). These data support the hypothesis that, in sediments, adsorbed hydrocarbon gases are strongly associated with authigenic carbonates.

The occurrence and significance of Pleistocene and Upper Pliocene sapropels

Numerous sapropels and sapropelic strata from Upper Pliocene and Pleistocene hemipelagic sediments of the Tyrrhenian Sea show that intermittent anoxia, possibly related to strongly increased biological productivity, was not restricted to the eastern Mediterranean basins and may be a basin-wide result of Late Pliocene-Pleistocene climatic variability. Even though the sapropel assemblage of the Tyrrhenian Sea clearly originates from multiple processes such as deposition under anoxic conditions or during spikes in surface water productivity and lateral transport of organic-rich suspensates, many "pelagic sapropels" have been recognized. Stratigraphic ages calculated for the organic-rich strata recovered during ODP Leg 107 indicate that the frequency of sapropel formation increased from the lowermost Pleistocene to the base of the Jaramillo magnetic event, coinciding with a period when stable isotope records of planktonic foraminifera indicate the onset of climatic cooling in the Mediterranean. A second, very pronounced peak in sapropel formation occurred in the Middle to Late Pleistocene (0.73-0.26 Ma). Formainifers studied in three high-resolution sample sets suggest that changes in surface-water temperature may have been responsible for establishing anoxic conditions, while salinity differences were not noted in the faunal assemblage. However, comparison of sapropel occurrence at Site 653 with the oxygen isotopic record of planktonic foraminifers established by Thunell et al. (1990, doi:10.2973/odp.proc.sr.107.155.1990) indicates that sapropel occurrences coincide with negative d18O excursions in planktonic foraminifers in thirteen of eighteen sapropels recognized in Hole 653A. A variant of the meltwater hypothesis accepted for sapropel formation in the Late Pleistocene eastern Mediterranean may thus be the cause of several "anoxic events" in the Tyrrhenian as well. Model calculations indicate that the amount of oxygen advection from Western Mediterranean Deep Water exerts the dominant control on the oxygen content in deep water of the Tyrrhenian Sea. Inhibition of deep-water formation in the northern Adriatic and the Balearic Basin by increased meltwater discharge and changing storm patterns during climatic amelioration may thus be responsible for sapropel formation in the Tyrrhenian Sea.

Age determination and sea surface temperature reconstruction from marine and lake records, Iberian Peninsula

Selected multi-proxy and accurately dated marine and terrestrial records covering the past 2000 years in the Iberian Peninsula (IP) facilitated a comprehensive regional paleoclimate reconstruction for the Medieval Climate Anomaly (MCA: 900-1300 AD). The sequences enabled an integrated approach to land-sea comparisons and, despite local differences and some minor chronological inconsistencies, presented clear evidence that the MCA was a dry period in the Mediterranean IP. It was a period characterized by decreased lake levels, more xerophytic and heliophytic vegetation, a low frequency of floods, major Saharan eolian fluxes, and less fluvial input to marine basins. In contrast, reconstruction based on sequences from the Atlantic Ocean side of the peninsula indicated increased humidity. The data highlight the unique characteristics of the MCA relative to earlier (the Dark Ages, DA: ca. 500-900 years AD) and subsequent (the Little Ice Age, LIA: 1300-1850 years AD) colder periods. The reconstruction supports the hypothesis of Trouet et al. (2009, doi:10.1126/science.1166349), that a persistent positive mode of the North Atlantic Oscillation (NAO) dominated the MCA.

Geochemistry of tropical Atlantic sediments of Marine Isotope Stage 4

Antarctic Intermediate Water (AAIW) and Subantarctic Mode Water (SAMW) are the main conduits for the supply of dissolved silicon (silicic acid) from the deep Southern Ocean (SO) to the low-latitude surface ocean and therefore have an important control on low-latitude diatom productivity. Enhanced supply of silicic acid by AAIW (and SAMW) during glacial periods may have enabled tropical diatoms to outcompete carbonate-producing phytoplankton, decreasing the relative export of inorganic to organic carbon to the deep ocean and lowering atmospheric pCO2. This mechanism is known as the "silicic acid leakage hypothesis" (SALH). Here we present records of neodymium and silicon isotopes from the western tropical Atlantic that provide the first direct evidence of increased silicic acid leakage from the Southern Ocean to the tropical Atlantic within AAIW during glacial Marine Isotope Stage 4 (~60-70 ka). This leakage was approximately coeval with enhanced diatom export in the NW Atlantic and across the eastern equatorial Atlantic and provides support for the SALH as a contributor to CO2 drawdown during full glacial development.