Approaches and constraints to the reconstruction of palaeoproductivity from Cape Basin abyssal benthic foraminifera (South Atlantic)

The characteristics of benthic foraminiferal assemblages from well-oxygenated deep-sea settings have been used to reconstruct past productivity conditions. None of the different approaches that have been developed is without complications or applies in all settings. In this study we assess the use of benthic foraminifera (accumulation rates and assemblages composition) as proxies for palaeoproductivity changes during the last glacial period (25 – 95 ka) in an abyssal core located in the south of Cape Basin (41.1 °S, 7.8 °E, 4981 m water depth). Assemblage characteristics indicate a generally food-limited environment receiving episodic inputs of labile organic carbon of variable strength. High seasonality in the delivery of organic material to the seafloor in the form of phytodetritus influences the assemblage characteristics because the corresponding response does not involve the whole community. Benefiting from this occasionally high organic input is the opportunistic species Epistominella exigua (Brady) that reproduces rapidly to build up large populations. In general, the rest of the species (i.e. less opportunistic compared to E. exigua) show only subtle variations in their population densities and fauna composition. Under those circumstances benthic foraminiferal accumulation rates seem to be independent of the amount of organic flux arriving at the sediment surface and respond instead to the strength of phytoplankton blooms.


INTRODUCTION
Diz The fossil record of benthic foraminifera has been widely used as a proxy for past environmental conditions (for reviews see Gooday 2003;Gooday & Jorissen 2012;Jorissen et al. 2007;Murray 2006).Despite the complexity of factors playing a role in foraminiferal distribution and the effect of post-depositional processes (Jorissen et al. 2007;Murray 2001), there is a general agreement that benthic foraminifera from deep sea environments largely respond to a few parameters: the organic carbon flux, bottom-water hydrodynamics (and related grain size), bottom and pore-water oxygenation and carbonate saturation (see reviews in Mackensen et al., 1995;Jorissen et al. 2007).Within these parameters, the estimation of past changes in the primary production (and organic carbon flux to the sea floor) from benthic foraminifera has received major attention in palaeoceanographic studies because of the implications of changes in the biological pump on past and future climate.Based on knowledge of the ecology of benthic foraminifera, a number of studies have suggested the abundance of benthic foraminifera (benthic foraminiferal accumulation rates, Herguera & Berger 1991), the contribution of different foraminiferal morphotypes (Corliss & Chen 1988) and species composition (e.g., Altenbach et al. 1999) as quantitative and qualitative proxies for past organic carbon fluxes arriving at the seafloor.Faunal composition might also provide information about the quality (refractory versus labile) of the organic carbon (e.g., Fontanier et al. 2005).The input of labile organic carbon to the sea floor derived from seasonal or episodic phytoplankton blooms at the surface also influences assemblage composition (e.g., Smart et al. 1994;Sun et al. 2006) and diversity (e.g., Corliss et al. 2009;Enge et al. 2011;Gooday et al. 2010Gooday et al. , 2012) ) so providing information about the periodicity of the organic carbon flux.
Evaluating the extent to which benthic foraminifera can be used to infer a particular set of past environmental conditions and understanding the complicating factors and limitations of such reconstructions is fundamental to advancing the use of proxy methods based on these  In this study we discuss palaeoenvironmental inferences based on benthic foraminiferal assemblage characteristics from an abyssal core located in the south of Cape Basin (South Atlantic).We examine the consistency of the most typical palaeoproductivity proxies by comparing information provided by our assemblages with environmental parameters from independent proxies obtained in the same core and other proxy archives.

MATERIAL AND METHODS
The marine sediment core discussed in this study (TNO57-21) was recovered in the south Cape Basin, slightly north of the Agulhas ridge in the southeast Atlantic Ocean (41.1ºS,7.8ºE,4981 m of water depth, 13.8 m long, Fig. 1) .It is presently bathed by the poorly ventilated, cold and fresh Antarctic Bottom Water (AABW).The location lies below at the Subtropical Convergence (STC) front that separates the cold low-salinity Subantarctic waters to the south and warm saline Subtropical waters to the north.The complex interaction of physicochemical factors associated with cross-frontal mixing of subantarctic and subtropical waters and the intense turbulence and eddy activity (Baker-Yeboah et al. 2010;Llido et al. 2004) results in relatively elevated pulses or episodes of primary production in the surface waters of the study area (Froneman et al. 1997;Machu et al. 2005).
Several previous micropalaeontological studies have been carried out on core TN057-21.The palaeoceanographic significance of planktonic foraminifera assemblages are discussed in Barker et al. (2009Barker et al. ( , 2010) ) and (Barker & Diz 2014).The linkages between the record of the benthic foraminifera Epistominella exigua and millennial-scale climate variability during the last 95 kyr are considered in Diz & Barker (2015).However, detailed information about benthic assemblage composition from ~25 to 95 kyr, and relevant additional palaeoenvironmental information, were beyond the scope of these earlier papers and are therefore addressed in the present study.The age model used in this study is described in Barker & Diz (2014) and the methods for the study of benthic fauna are indicated in Diz & Barker (2015).Briefly, foraminifera are picked from the > 125 µm fraction, mounted on faunal slides and identified following the generic assignations of (Loeblich & Tappan 1987) and published illustrations (see Table 1).The study of benthic foraminifera in the >125 µm size fraction potentially underestimate the abundance of small (63-125 µm) opportunistic taxa such as Alabaminella weddellensis which are typically abundant in areas characterized by episodic phytoplankton blooms (e.g., Diz et al. 2007;Smart et al. 2010;Thomas et al. 1995).
On the other hand, it allows for studying a feasible number of samples still showing major faunal patterns (e.g., Poli et al. 2012, Schönfeld et al. 2012).Raw counts can be found in Diz & Barker (2015) and the most characteristic species of benthic foraminifera are illustrated in this study (Plate I).The assemblages are largely dominated by Epistominella exigua (>40% on average), a phytodetritus related species.In order to assess the ecological significance of the other relevant species we discuss the assemblages after removing the contribution of E. exigua (i.e., the "phytodetritus free assemblage", see also Thomas et al. 1995).Additionally, we examined the record of benthic foraminiferal accumulation rate (BFAR, total number of foraminifera x cm -2 x kyr -1 ) and phytodetritus free benthic foraminiferal accumulation rate (BFAR phfree ).BFARs are calculated as the number of individuals per gram of total dry bulk sediment (#bf x g -1 , Diz & Barker 2015) x linear sedimentation rates (cm x kyr -1 ) x density (g x cm -3 ).Sedimentation rates are taken from the age model developed by Barker & Diz (2014) and vary from 8 to 16 cm kyr -1 .It is considered that the age model is sufficiently robust as to assume that differences in BFAR are not mainly caused by inaccuracies in the calculation of sedimentation rates.In the absence of direct measurements of dry or wet bulk density, we used the algorithm developed by Sachs & Anderson (2003) for the calculation of, which is based on the percentage of calcium carbonate in order to calculate densities in core TNO57-21.To do this, the high resolution % CaCO 3 record of core TNO57-21 (Sachs & Anderson Diz & Barker -6 -2003) was re-sampled and values interpolated to the lower resolution record of benthic foraminiferal abundance using the program Analyseries (Paillard et al. 1996).Then, density values were calculated using Sachs and Anderson´ formula.Calculated values vary between 0.5 to 0.7 g x cm -3 .
The core TNO57-21 was recovered from deep waters that are currently undersaturated with respect to calcium carbonate (4891 m water depth, see Supplementary Information by Barker et al. 2010).Several episodes of increased carbonate dissolution have been identified in planktonic foraminiferal (Barker & Diz 2014;Barker et al. 2009Barker et al. , 2010) ) as well as benthic foraminiferal (Diz & Barker 2015) assemblages By considering the information provided by multiple proxies (e.g., % of coarse fraction, % CaCO 3 , foraminiferal abundances, composition of assemblages) it was indicated that carbonate dissolution does not seem to exert an overall control on benthic abundance or species distribution with the exception of the Holocene (not considered in this study) and Greenland Stadial (GS) events GS19 and 21.Those poorly preserved periods (indicated in figure 2 by bars) are characterized by low percent of calcium carbonate, very low foraminiferal abundances and low number of benthic taxa.They are not used to infer palaeoenvironmental conditions (see detailed discussion in Diz & Barker 2015).
The comparison of the composition of the "dissolved assemblages" with modern samples (Mackensen et al. 1993) indicates that Nutallides umboniferus is very rare along core TNO57-21 (see Plate).N. umboniferus is generally related to corrosive bottom waters in the South Atlantic (Mackensen et al. 1993;Harloff & Mackensen, 1997;Schmield et al. 1997).It might suggest that other factors more than carbonate undersaturated waters are playing a role in this species distribution (i.e, oligotrophic conditions and absence of phytodetritus deposition, Smart & Gooday, 1997;Kurbejeweit et al. 2000;Carman & Keigwin, 2004).
The BFAR values in core TNO57-21 show large fluctuations ranging from 30 to 1100 foraminifera (> 125 µm fraction) cm -2 kyr -1 .On average, peak BFAR values are higher during northern stadial events occurring during MIS 3 than over MIS 5 (Fig. 3a).The lowest BFAR values occur during GS19 and GS21 (also the Holocene, not shown) when benthic foraminiferal dissolution was inferred (Diz & Barker 2015) and thus they are excluded from interpretation (Fig. 3a).In general, episodes of peak BFAR values are largely driven by the phytodetritus-related species Epistominella exigua (Fig. 3a), and show a tight temporal correspondence with abrupt climate changes occurring in the North Atlantic Ocean and concomitant antiphase response in the South Atlantic (see Fig. 3 and Diz & Barker 2015).

DISCUSSION
Reconstructing the flux of organic carbon to the seafloor from benthic foraminifera Benthic foraminiferal accumulation rates.The benthic foraminiferal accumulation rate (BFAR, number of foraminifera x cm -2 x kyr -1 ) has been suggested as a palaeoproductivity proxy (Herguera 1992(Herguera , 2000;;Herguera & Berger 1991).However, several studies indicated that obtaining a quantitative estimation of primary production or export production from BFAR values might be limited by several factors, with the difficulty of obtaining reliable calibrations, carbonate dissolution and taphonomical processes being the most important (see review in Jorissen et al. 2007).In this study no attempts have been made to estimate absolute palaeoproductivity or palaeo-carbon fluxes, but we evaluate the potential use of the calculated BFAR as a semi-quantitative proxy for food supply to the ocean floor (e.g., Thomas et al. 1995;Alegret & Thomas 2009) in core TNO57-21.
The BFAR values during the Holocene and late deglaciation (not plotted but calculated on the basis of Diz & Barker (2015) data; <100 foraminifera cm-2 kyr-1) are within the range of values for abyssal core tops from the Pacific open ocean at water depths > 4000 m likely affected by carbonate dissolution (Herguera 1992;Herguera & Berger 1991).However, peak BFAR values from MIS 3 to MIS 5 are four to eight times Holocene values (Figure 3).These accumulation rates are substantially higher than maxima values recorded in core-top samples from widely distributed open-ocean areas at various depths in the Pacific.Following the use of BFAR as a palaeoproductivity proxy, these results might suggest past increases in the organic carbon flux to the sea floor up to several times the present conditions.According to modern primary production distributions (see Machu et al. 2005, their Fig. 5) this might be equivalent to a change from oligotrophic subtropical open ocean areas to high productivity upwelling centres (e.g., Benguela upwelling).In fact, peaks of BFAR reach values found in continental shelf areas affected by seasonal upwelling such as Cape Blanc upwelling (Guichard et al. 1999;Zarriess & Mackensen 2010) or the Southwest African upwelling (Schmiedl & Mackensen 1997).Such a change in the palaeoenvironmental conditions, if ithad occurred, would have involved a large change in assemblage composition, i.e. from low productivity faunas (low abundances) to a high productivity assemblage (high BFAR).This seems to be unrealistic considering that assemblage changes are subtle in comparison to BFAR fluctuations (see Figure2-3) and they do not involve the appearance of species related to eutrophic conditions (see next section).In fact, variations in BFAR are mainly driven by the phytodetritus-related species Epistominella exigua (Table 1, Figure 3a).In the absence of this species fluctuations in BFAR phfree (BFAR phytodetritus free; "phfree") are of much lower magnitude and suggest only subtle changes in organic carbon fluxes.This could be explained by the fact of that when large quantities of fresh marine organic matter (phytodetritus) are available, even for short periods of time, opportunistic species, such as Epistominella exigua, reproduce rapidly and produce a large number of tests (see Thiel et al. 1989).The accumulation of foraminifera is not related in a predictable manner to organic flux and, in these circumstances, BFAR should not be used as a quantitative (or semiquantitative) proxy (Thomas et al., 1995;Diz et al., 2007;Smart, 2008;Zarriess & Mackensen, 2010) for the organic carbon flux reaching the sea floor.BFAR is rather related to the strength of the phytoplankton blooms.The episodic and labile nature of the Diz & Barker -10 -phytodetritus aggregates reaching the sea floor (Beaulieu 2002;Smith et al., 2008) could explain the weak correlation (r= -0.24, p= 0.000, n=467) between BFAR (also BFAR phfree ) and the alkenones flux, a proxy for averaged organic carbon flux to the seafloor (Sachs & Anderson 2003) obtained in the same core (Fig. 3c).
Benthic foraminiferal assemblage composition.The use of benthic foraminiferal assemblages as palaeo-productivity proxies is based on the relation of faunal composition to organic flux rates (e.g., Morigi et al. 2001).Again, the quantification of flux regimes from the relative abundance of major species is complicated by ecological factors, broad adaptability of some species to organic carbon fluxes and uncertainties related to estimations of carbon flux (Altenbach et al. 1999;Gooday 2003;Jorissen et al. 2007).Thus, the information provided by the faunal composition of core TNO57-21 is used here as qualitative proxy for the organic carbon flux or its quality/periodicity.
The relative abundance of particular species or group of species is related qualitatively to the predominance of a particular flux regime (i.e., high, intermediate, low-flux species), quality (i.e., labile, refractory) or seasonality of the flux (i.e., seasonal, sustained).In this study, groups are defined based on representative species with well-known ecology (see the ecological attributions of characteristic species in Table 1).Because organic flux dependent patterns should be best analysed within the environmental optimum of species (Altenbach et al. 1999) those species with percentages lower than 10% for most of the record (e.g., Eggerella bradyi, Quinqueloculina cf.seminula, Rutherfordoides tenuis, see Plate) and/or those species for whom the ecology is not well-constrained are not included in any of the groups.An example of the later is the high contribution to the assemblage of individuals belonging to the "Lagenina group" (>10%, Fig. 2c, see Plate).Unfortunately their ecology is poorly known (see Table 1) and consequently the contribution to the paleoenvironmental understanding of the assemblage changes is limited.
The most abundant species throughout the core is Epistominella exigua, an epifaunal/shallow infaunal abyssal deep-sea opportunistic foraminifera that flourishes and rapidly reproduces in the presence of phytodetritus aggregates (see ecological attributions in Table 1).Thus, high relative abundance of E. exigua (Figure 2b) along core TNO57-21 indicates overall low organic carbon fluxes punctuated by episodic phytoplankton blooms that resulted in the deposition of phytodetrital inputs (i.e., labile organic carbon).This interpretation is supported by the "phytodetritus free" assemblage which is composed by species typical of oligotrophic ("low flux") to moderately mesotrophic ("intermediate") environmental conditions.
The "low-flux species" are the epifaunal Fontbotia wuellerstorfi, Pyrgo murrinha and the shallow infaunal Oridorsalis umbonatus and Globocassidulina subglobosa (see Table 1).All together these represent > 30% of the phytodetritus free assemblage (Fig. 3f) reaching up to 60% during early MIS3.For some of these species (i.e.G. subglobosa, F. wuellerstorfi, P. murrinha) a relationship to phytodetritus input (See Table 1) has also been suspected.However, their downcore relative abundance does not covary with BFAR or percentages of phytodetritus-related species E. exigua (Figure 3).Instead, they must respond to a different type of phytodetritus or, more likely, they prefer sustained (and low) organic carbon flux to the sea floor rather than pulsed (see also Smart, 2008).
The group of species considered indicative of an "intermediate -flux" regime are mainly represented by the relatively common shallow infaunal Pullenia osloensis together with Melonis spp.and Siphotextularia rolshauseni (Figure 3e, Table 1).Those species show a statistically significant negative correlation (p<0.000) to the relative contribution of the species belonging to low -flux group (r O. umbon-S.rolshaseni =-0.38; r P. osloensis-P.murrinha = -0.39;r P.  3e) occurring in coincidence with high relative contribution of Epistominella exigua (Fig. 2-3) and the substantial increase in the alkenone flux measured during glacial MIS 4 (Fig. 3c).These two species show affinity for degraded organic carbon (Table 1) suggesting that during MIS 4 part of the sedimentary organic compounds are of low nutritional value.This might be explained by a different type and/or fate of phytodetrital material arriving to the seafloor which could promote the response of species that benefit from the bloom indirectly (i.e., bacteria colonizing aggregates, Koho et al., 2008).In fact, it has been inferred (Diz & Barker 2015) that a slightly different physical conditions promoting phytoplankton blooms in the surface waters occurred during MIS 4/MIS 5 transition and early MIS 4 (seasonality and windiness) than during MIS 3 and late MIS 5 (mainly eddy activity).
Notably, that assemblage change does not trigger a parallel increase in the BFAR phfree indicating that a change in the food quality does not force a change in the absolute numbers of benthic foraminiferal accumulation rate but only in the assemblage composition.

CONSTRAINTS AND APPROACHES: OVERVIEW
Benthic foraminifera (BFAR and assemblage composition) were analysed in abyssal core TN057-21 .Episodic phytoplankton blooms represents the main source of "disturbance" to the benthic environment and have important consequences for the structure and taxonomic composition of the benthic foraminifera record.The inferred phytodetritus deposition does not Diz & Barker -13 -seem to trigger a response by the whole community but it is limited to a single species, leading to a decrease in the diversity of the assemblages.The species benefiting from phytodetritus input is the opportunistic (r-strategist) Epistominella exigua which reacts quickly, producing high numbers of individuals and consequently high benthic foraminiferal accumulation rates.The response of the remaining fauna (the so called "phytodetritus free assemblage") to phytodetritus input is mainly constrained to a concomitant moderate increase in the foraminiferal abundance (BFAR phfree ) with overall little variations in the assemblage composition.This could be explained by the different ecological preferences of species with lower reproductive potential (k-strategists) and likely benefiting from other type organic carbon (less labile).All these data together suggest BFAR could not be a reliable proxy for palaeoproductivity in abyssal (food limited) environments when benthic foraminifera are highly dependent on the rate and nature of the input of labile organic material generated in the euphotic zone.Under these circumstances fluctuations in BFAR do not reflect the flux of organic carbon to the seafloor but rather the strength of phytoplankton blooms in the surface waters.In this case, the assemblage composition is relevant for interpreting palaeoenvironmental conditions.

Figure 3: Paleoproductivity from benthic foraminiferal accumulation rates and assemblage composition
The Benthic Foraminiferal Accumulation Rate (BFAR, total number of foraminifera >125 m cm -2 kyr -1 ) is indicated as a shaded black areas and Benthic Foraminiferal Accumulation Rate calculated on phytodetritus free basis (i.e., discounting the contribution of the phytodetritus-
Consequently, the difference between the two shaded areas is the BFAR of E. exigua (BFAR E.exigua ). Figure c displays the flux of alkenones (ng cm -2 kyr -1 ) in core TNO57-21, a proxy for organic carbon flux to the seafloor (Sachs & Anderson, 2005) plotted on Barker & Diz (2014) (Barker et al., 2011).Black crosses on the top indicate tie points and sedimentation rates respectively used in the age model construction and BFAR calculations.Grey line on the top plot shows density values calculated using model proposed by Sachs & Anderson (2003).

TABLE CAPTIONS:
Table 1: Ecological attributions of the most characteristic benthic foraminiferal species in core TNO57-21..

Epifaunal/ Epibenthic
F. wuellerstorfi is an epibenthic species (Linke & Lutze, 1993) that lives in well oxygenated and ventilated bottom waters away from high productivity coastal environments.Global distributional patterns indicate that it prefers organic fluxes below 2 gCm -2 yr -1 (Altenbach et al .1999).In the South Atlantic, the distribution of this species coincides with relatively sustained food fluxes in areas were carbonate dissolution is not too severe (Mackensen et al ., 1995).In deep sea sediments of the Indian Ocean F. wuellerstorfi co-occurs with the phytodetritus species E. exigua .On ocean wide scale it co-occurs with Uvigerina peregrina (a typical high-flux species) on the lower range of organic carbon fluxes of the later (Altenbach et al., 1999).This information suggest that F. wuellerstorfi is adapted to a wide range of environmental conditions, including strongly pulsed supply of organic carbon in the Indian Ocean (Gupta, 1997;De & Gupta, 2010).
Most of the individuals found in core TN057-21 are juvenile forms.

Shallow infaunal
On an ocean wide scale, O. umbonatus seems to be distributed in areas receiving low organic carbon fluxes (Altenbach et al . 1999).Mackensen et al . (1995) suggest that the distribution of O. umbonatus in the South Atlantic is related to relatively well-oxygenated pore waters receiving a relatively low but sustained food supply.
It co-occurs with E. exigua in the carbonate supersaturated and low organic carbon areas of the Weddell sea continental Shelf (Mackensen et al . 1990).In the eastern Indian Ocean, Murgese & DeDecker (2005)

Epifaunal
Pyrgo murrinha is an epifaunal species (Corliss & Chen, 1988) found in areas of low organic carbon fluxes (Altenbach et al., 1999) and high oxygenation such as the South Central Indian Ocean (De & Gupta, 2010) or eastern Indian Ocean (Murgese & De Deckker, 2005).Conversely, Gupta & Thomas (2003) interpret the abundance of P. murrinha in Indian cores as related to pulsed food and good carbonate preservation.

Epifaunal/ Infaunal
In the North Atlantic Sun et al .( 2006) found an inverse correlation of the G. subglobosa-N.umbonifera assemblage with seasonality of the organic matter.In the Southeast Atlantic, this species characterize sandy sediments of the Walvis Ridge (South Atlantic) with low organic carbon and high calcium carbonate content and enhanced bottom current velocities (Schmield et al . 1997).However, G. subglobosa has been associated with deposition of phytodetritus in the Northeast Atlantic Abyssal Plain Gooday (1988Gooday ( , 1993)), North West Africa (Eberwein & Mackensen (2006) and the Antarctic shelf (Suhr et al . 2003).

Epifaunal/ Shallow infaunal
This is an opportunistic species that rapidly colonizes and feeds on aggregates of phytodetritus produced by episodic phytoplankton blooms (e.g., Gooday, 1988, 1993, Cornelius & Gooday, 2004;Sun et al . 2006;Gooday et al . 2010).Epistominella exigua is well adapted to oligotrophic conditions well away from areas where the productivity is high and the flux of organic matter is continuous (Mackensen et al . 1995).In the eastern South Atlantic E. exigua isf o undinas s o c ia tio nw ithF.wuellerstorfi in low organic carbon areas on the flanks of the Walvis Ridge (Schmield et al. 1997).It is also abundant in the deep western South Atlantic (Harloff & Mackensen, 1997) the deep Weddell Sea continental slope (Anderson, 1975;Cornelius & Gooday, 2004) and South West Indian Ocean (Peterson, 1983;Corliss, 1983).

Not very well constrained Shallow to deep infaunal
The ecology of Pullenia osloensis is not well known.This species is described in deep and abyssal areas of the South Cape Basin (Boltovskoy & Boltovskoy, 1989), Southwest Indian Ocean (Corliss, 1979) and the Agulhas retroflection area (Diz et al . 2007).Most authors interpret the group of Pullenia spp.(mainly Pullenia bulloides ) as indicator of areas of high organic supply (Rasmussen et al . 2003) or high but varying surface productivity (Mackensen et al . 1995).In the Indian Ocean, the assemblage dominated by Epistominella exigua with Pullenia osloensis and Pullenia salisburyi as secondary species is considered indicator of low to intermediate organic flux and high seasonality (Gupta & Thomas, 2003).Pullenia osloensis is common along core TNO57-21 suggesting that it is adapted to a wide range of carbon flux regimes (from low to intermediate) and organic carbon qualities (i.e., labile and refractory).Pullenia salisburyi is also present in core TNO57-21 (Plate, Fig. 17) even though its relative contributions are lower than P. osloensis .

Not well constrained
Siphotextularia rolshauseni is an agglutinated species composed of calcareous fragments of various sizes and it is considered a junior synonym of Siphotextularia catenata Cushman (Corliss, 1979, Nees & Struck, 1994).There are a few references to this species in superficial sediments of the North Atlantic (see review Nees & Struck, 1994), Southwest Pacific (Kurihara & Kennett, 1986), South China Sea (Szarek et al ., 2006), the Indian Ocean (Corliss, 1979, Nees et al . 1997).Similarly, this species is reported as rare in South Atlantic superficial sediments by Mackensen et al . (1993, Siphotextularia sp.), Schmield et al . (1997, S. catenata )a sw e l la si n Pleistocene and Miocene sediments of the Walvis Ridge (Boltovskoy & Boltovskoy, 1989).Nees & Struck (1994) and Struck (1995) relate the occurrence of S. rolshauseni in the Greenland-Norwegian Seas during MIS 2 with a period of low trophic quality.This interpretation is also suggested by the downcore distribution of S. rolshauseni in core TNO57-21 which peaks together with Melonis spp.during MIS 4 (see text).

Flux Range?
The called "Lagenina Group" is composed by several species belonging to genus Amphycorina, Bifarilaminella, Cushmanina , Fissurina, Homaloedra ; Unknown The species belonging to the called " Lagenina group" are found as rare representatives of the living and recently dead assemblages in different marine settings.Thus, the ecological significance of this group is not well constrained.Some of the species belonging to Lagenina group are illustrated in Plate I: Fissurina staphyllearia (Plate, Fig. 3), osloensis-F.wuellerstorfi =-0.28; r P. osloensis-G.subglobosa=-0.23; r O. umbonatus-Melonis spp.=-0.35).The relative Diz & Barker -12 -contribution of the "intermediate-flux" group indicates slightly more eutrophic conditions during MIS 5 and MIS 4 than during MIS 3. Peaks of the intermediate flux fauna occurring over MIS 4 are mainly caused by noticeably increased contributions of Melonis spp.and S. rolshauseni (Figure 2d-f, Figure

Figure 2 :
Figure 2: Benthic foraminiferal assemblages in core TN057-21 age model.The group of "low flux" species (e) is composed by shallow infaunal O. umbonatus, G. subglobosa and epifaunal P. murrinha and F. wuellerstorfi.Intermediate flux species (light grey line, d) is constituted by individuals belonging to P. osloensis, S. rolhsauseni (medium grey line) and Melonis spp.(dark grey line).Percentages are calculated on the phytodetritus free assemblage basis.Records of Southern Hemisphere ice core temperature (f, Antarctica Epica Dome C δD, Jouzel et al., 2007, 3 points running average) and Northern Hemisphere ice core temperature record (b, Greenland NGRIP δ 18 O; EPICA, 2006) are showed as reference.Figures b and f are plotted on GICC05/NALPS timescale found the assemblage composed by O. umbonatus-E.exigua and P. murrhina indicator of low organic carbon.In Sulu area, Miao & Thunell (1993) indicated O. umbonatus as a species inhabiting shallow infaunal low organic carbon, well oxygenated sediments in agreement with Rathburn & Corliss (1994).Likewise, Burke et al .(1993), include O. umbonatus together with Eggerella bradyi within the group of low productivity species in the central Pacific.Conversely Gupta (1997) interpret the group of O. umbonatus , Eggerella bradyi ,a n dMelonis pompilioides as reflect intermediate flux of relatively degraded organic matter and Kaiho (1999) consider O. umbonatus as an indicator of low-oxygen conditions.