Effect of Mg/Ca ratio in seawater on shell composition in shallow benthic foraminifera



[1] The ratio of magnesium to calcium (Mg/Ca) in CaCO3 shells of foraminifera is widely used to determine paleotemperatures. However, Mg/Ca is highly variable within and between species, suggesting a strong physiological influence on the incorporation of Mg2+ into the shells. While most field and laboratory calibrations have focused on the effect of temperature, we chose to study the effect of ambient Mg/Ca on the calcification process and the final shell composition. We cultured two species of symbiont-bearing benthic foraminifera, Amphistegina lobifera and Amphistegina lessonii, in seawater with different Mg/Ca ratios. Electron probe analysis of the shell Mg/Ca revealed a positive (but not entirely linear) correlation with Mg/Ca in the culturing media with slightly different curves for each species. Partition coefficients of Mg2+ (DMg) in the calcite shells showed a decrease by a factor of roughly 2 between the lowest and highest Mg/Ca in the ambient water. This was previously demonstrated in inorganic calcite precipitation experiments. However, the biogenic DMg was significantly lower than the inorganic one, suggesting a physiological mechanism that reduces Mg/Ca at the calcification site. Unlike inorganic experiments that display a dependence of DMg on the kinetics of precipitation, the biogenic DMg is not correlated with the rate of calcification. Both DMg and calcification rates in our experiment were sensitive to the Mg/Ca ratio rather than the concentration of either Ca2+ or Mg2+. The largest addition of CaCO3 was obtained at Mg/Ca of 1, and not at present-day seawater ratio (Mg/Ca = 5). This may reflect the Mg/Ca that prevailed during the Eocene (Mg/Ca ∼ 1.5), when this genus evolved.

1. Introduction

[2] Foraminifera are unicellular calcifying marine organisms that first appeared in the early Cambrian [Loeblich and Tappan, 1988]. Foraminiferal shells are excellent indicators of past oceanic conditions based on trace elements and stable isotopes in their CaCO3 shells. Recently, the ratio of magnesium to calcium (Mg/Ca) in foraminiferal shells has been used to determine paleotemperatures; and together with oxygen isotopes this allows a singular solution to the well-known paleotemperature equation [e.g., Lear et al., 2000]. The relationship between Mg/Ca and temperature can be predicted from thermodynamic considerations [Lea et al., 1999]. Mg2+ incorporation into calcite increases with temperatures in both biogenic and inorganic calcite. However, Mg/Ca is highly variable within and between species, suggesting a strong physiological influence on the incorporation of this metal into the shells. Furthermore, the actual Mg content in perforate foraminifera is much lower than that of inorganic calcite [e.g., Rosenthal et al., 1997]. On the other hand the Mg/Ca in the imperforate foraminifera (porcelaneous) is higher than that of inorganic calcite reaching up to 20 mole% MgCO3 in their shells [Milliman, 1974]. The temperature dependence of Mg in perforate foraminifera is also species dependent and different from that of inorganic calcite [e.g., Toyofuku et al., 2000]. This variability must be related to the biomineralization process, which may be influenced by various environmental and physiological factors [e.g., Erez, 2003]. One of these factors may be the Mg/Ca ratio in seawater that has changed significantly during the geological past due to tectonic processes and biological evolution [e.g., Horita et al., 2002; Stanley and Hardie, 1999]. Stanley et al. [2002] first conducted an experimental study on Mg fractionation by coralline algae. Recently, Ries [2004] has showed experimentally that various marine invertebrates (echinoids, crabs, shrimps and calcareous tube worms) record in their skeleton the Mg/Ca ratio of the seawater. The resulting partition coefficients (DMg) were in almost all cases lower than that of inorganic calcite. In addition, the DMg values were not entirely constant and showed significant changes with the ambient Mg/Ca. While the temperature effect on Mg incorporation into foraminifera has been intensively examined both in natural distribution studies (field calibrations) and in laboratory culture experiments [e.g., Lea, 2003], the effects of ambient Mg/Ca on shell composition was reported only in one study [Delaney et al., 1985], for planktonic foraminifera in a range of seawater Mg/Ca between 3 and 10. The relationship between Mg/Ca in the shells and that in the solutions was roughly linear but not conclusive, mainly due to large experimental errors. We carried out an experimental study on the effect of seawater Mg/Ca ratio on foraminiferal calcification rate and shell chemistry. Culturing experiments with two benthic species of foraminifera (Amphistegina lobifera and Amphistegina lessonii) were conducted in artificial seawater with different Mg/Ca ratios.

2. Experimental Methods

2.1. Culturing Conditions

[3] Live symbiont-bearing specimens of A. lobifera and A. lessonii were collected in the Gulf of Eilat, Israel. Uniform populations of ∼400 μm diameter were placed in sealed 60 mL Erlenmeyer bottles filled with various proportions of artificial and natural seawater that produced eight different Mg/Ca combinations (Table 1). The natural seawater was surface nutrient depleted water of the Gulf of Eilat (with salinity of 40.7‰ that was adjusted to 35‰ by adding DW). The experiment lasted 52 days and was carried out at salinity of 35‰ and constant temperature of 24 ± 0.2°C in a temperature controlled water bath. Water in each bottle was replaced twice a week with a fresh solution. The stock solution of each treatment was kept sealed in the dark. All the bottles received the same amount of natural light and were examined daily for development of algae and precipitation of inorganic deposits. Measurements of pH and alkalinity were conducted twice a week with every replacement of water as well as on the stock solutions to ensure a constant pH with each refill (see analytical methods below). Each experimental treatment was tested in duplicate bottles that contained ∼60 individuals, ∼30 individuals of each species. The foraminifera were not fed. Previous experiments in our laboratory have demonstrated that highest growth rates are obtained with replacement of fresh seawater and with no feeding. Each bottle was opened for several minutes only during water replacement.

Table 1. Addition of CaCO3 per Individual, Mg/Ca Ratios and Ion Concentrations in the Culture Mediaa
Mg/Ca Rounded RatiobMg/Ca Calculated RatioMg2+, mMCa2+, mMAddition of CaCO3 per Individual,c μgAddition of CaCO3 per Individual,d μg
  • a

    The initial weight/individual was 43.6 ± 3 μg.

  • b

    As in text.

  • c

    Calculated from alkalinity depletions.

  • d

    Calculated from weight gain.

  • e

    Normal seawater. The anomaly of high Mg/Ca is typical of the Gulf of Eilat.

0.50.5305.3110.033.4 ± 1030.7 ± 2.0
11.0610.610.050.0 ± 1246.5 ± 9.2
11.0530.530.152.1 ± 1140.2 ± 2.4
2.52.6326.510.147.0 ± 2.645.9 ± 2.0
2.52.6353.120.246.0 ± 2.946.0 ± 1.7
5e5.2053.110.239.9 ± 1.235.9 ± 1.9
7.57.6652.06.8032.1 ± 1149.7 ± 9.7
1010.151.55.1027.4 ± 4.725.5 ± 2.4

2.2. Preparation of Culture Media

[4] The artificial seawater was prepared according to Millero and Sohn [1992] by mixing salts (SIGMA) corresponding to the composition of seawater but without Mg2+ and Ca2+ that were added later (as chloride solutions) in different proportions to the culture media. Solutions of MgCl2·6H2O and CaCl2·2H2O were prepared using salts (SIGMA) and DDW to achieve a concentration of 1M in each solution. The exact concentrations of the solutions were determined in the Geochemical Laboratory of the Hebrew University in Jerusalem, using ICP-OES on samples of the MgCl2·6H2O and CaCl2·2H2O stock solutions following the analytical methodology described by Starinsky and Katz [2003]. Exact concentrations of Mg and Ca in the natural seawater were taken from Krumgalz and Erez [1984]. In two cases (Mg/Ca ratio of 1 and 2.5), the ratios were obtained with two different combinations of Mg2+ and Ca2+ concentrations (Table 1). The carbonate chemistry of the culture media was monitored throughout the experiment (twice a week with every water replacement). There were no significant changes in these parameters in the stock solution for the duration of the experiments.

2.3. Cleaning Method

[5] Upon termination of the experiment, the cultured foraminifera were rinsed with DDW and dried overnight at 50°C. Specimens were placed in dilute sodium hypochlorite solution (1:3) for about 7 hours followed by several DDW rinses and overnight drying at 50°C.

2.4. Analytical Methods

[6] Salinity was determined using an optical salinometer (American Optics) with precision of 0.1‰ PSU. pH was measured on the alkalinity samples using the electrodes of the titrator (pH electrode Radiometer #PHG201-7 and reference electrode Radiometer REF201). The electrodes were calibrated using Radiometer buffers (Radiometer analytical traceable to IUPAC and NIST, 4.005 and 7.000). Alkalinity measurement was conducted once a week on a combined solution from the two last water replacements. Alkalinity titrations were made using a Radiometer automatic titrator (ABU 91), on accurately weighed duplicate samples of ∼10 ml. Gran calculations were preformed according to Sass and Ben-Yaakov [1977] to determine the total alkalinity. The average precision was ±1.4 μeq kg−1, and the typical alkalinity depletions in comparison to the initial solution were at least several tens of μeq kg−1. Weight measurements were conducted using a CAHN analytical scale. Four control groups (additional to the experimental groups) were weighed after they were rinsed with DDW and oven-dried, their average initial weight was 43.6 ± 3 μg. The experimental groups were weighed by the same procedure upon termination of the experiment.

[7] Analysis of skeletal Mg/Ca molar ratio was carried out on the knob area of unpolished foraminifera with JEOL JXA 8600 SUPERPROBE electron microprobe (WDS, accelerating potential = 10 kV, beam diameter = 15 μm, beam current = 8–10 nA, counting time = 60sec). The analyses were performed on the center of the knob area in order to ensure analysis of newly formed CaCO3 only. This area is covered with a CaCO3 layer each time a new chamber is being formed and its thickness is several hundred μm. Parameters of the electron microprobe were adjusted to obtain signals only from the upper few μm representing the newest CaCO3 layer deposited by the organism. Two specimens of each species from every bottle (for each Mg/Ca treatment we had two duplicate bottles) were analyzed. These specimens were all larger than the final average weight/individual hence their weight was more than doubled during the experiment. Three analyses were preformed on each specimen yielding a total of 12 measurements per species per treatment. Average precision (1σ) of the Mg/Ca ratio for A. lobifera within a specimen was ±6% and between two specimens ±9%. The average precision (1σ) of the ratio for A. lessonii within a specimen was ±6% and between two specimens ±15%. Mg/Ca quotient error was calculated according to Topping [1972].

3. Results

3.1. Shell Mg/Ca

[8] Both A. lobifera and A. lessonii incorporate more Mg2+ to their shell as the Mg/Ca of the medium increases (Figure 1). The two species have incorporation curves that are significantly different from each other at a probability >0.06. Both curves display an almost linear positive correlation with ambient Mg/Ca ratios lower than 5, and a more moderate slope in higher ratios. Partition coefficients (DMg) calculated for A. lessonii showed a decrease from 1.8 * 10−2 for ambient Mg/Ca of 0.5, to 0.5 * 10−2 for Mg/Ca of 10 (Figure 2). Similarly for A. lobifera DMg decreased from 1.5 * 10−2 to 0.8 * 10−2 for ambient Mg/Ca ratios of 0.5 and 10, respectively (Figure 2). We can rule out changes in the composition of the solution since the amount of Mg and Ca removed bycalcification was insignificant compared to the dissolved Mg and Ca concentrations. For example, the average alkalinity depletion of the last 20 days, which most likely represents the CaCO3 measured by the Electron Probe, was 36 μeq kg−1. This depletion will cause an increase in the Mg/Ca ratio of the culture media and a corresponding decrease in DMg of 1.8% for Mg/Ca = 5, and 3.6% for Mg/Ca ratio of 10. These changes are negligible compared to the decrease in the DMg that was observed (200–300%). Skeletal Mg/Ca did not show correlation with Ca2+ or Mg2+ concentration in the culture media (Figure 3, Figure 4). At present-day seawater concentrations (Ca2+ ∼ 10 mM, Mg2+ ∼ 50 mM) variable values of skeletal Mg/Ca were produced, indicating a correlation with ratio and not concentration (Figure 3, Figure 4).

Figure 1.

Mg/Ca in foraminifera shells versus Mg/Ca in their culture media.

Figure 2.

Average partition coefficient (DMg) versus Mg/Ca in culture media for two species of shallow benthic foraminifera. Error bars represent the standard deviation (2σ).

Figure 3.

Shell Mg/Ca versus Mg2+ concentrations in the culture media. Each circled set of points represents a different Mg/Ca ratio in the culture media (indicated by numbers).

Figure 4.

Shell Mg/Ca versus Ca2+ concentrations in the culture media. Each circled set of points represents a different Mg/Ca ratio in the culture media (indicated by numbers).

3.2. Addition of CaCO3

[9] The highest addition of CaCO3 per individual occurred in Mg/Ca ambient ratio of 1 (Table 1, Figure 5). This result was observed both in calculations based on alkalinity depletions as well as calculations based on weight gain measurements. Both of the measuring methods (alkalinity and weight) exhibit similar trends. Higher ratios than 1 produced a gradual and consistent decrease in CaCO3 addition. The ratio of 0.5 produced in both methods a lower CaCO3 addition than the ratio of 1.

Figure 5.

Cumulative addition of CaCO3 per individual versus Mg/Ca in the culture media. Two independent sets of data are reported: (1) CaCO3 addition per individual based on weight measurements. (2) Cumulative CaCO3 addition per individual based on alkalinity depletion measurements. Error bars represent the variability between two duplicates. Note that at the ratio of 7.5 only one duplicate is presented for both parameters. The duplicates not presented yielded exceptionally high and low values for weight and alkalinity, respectively.

4. Discussion

[10] The DMg calculated in our experiment exhibits similar trends to that of inorganic calcite [Morse and Bender, 1990] and to those observed in similar experiments conducted on other marine invertebrates [Ries, 2004]. An important difference is the lower value of the biogenic DMg in comparison with inorganic precipitations. For example in present-day seawater with Mg/Ca = 5, Amphistegina calcite contains ∼6 mole% MgCO3 which yields a biogenic DMg of about 1 * 10−2 (Figure 6). DMg for inorganic calcite precipitated from seawater and containing ∼8 mole% MgCO3 is about 1.5 * 10−2 [Morse and Mackenzie, 1990; Morse and Bender, 1990]. This may suggest a physiological mechanism for Mg2+ removal from the calcification site [Bentov and Erez, 2006].

Figure 6.

DMg of A. lobifera (white diamonds, dashed trend line) and A. lessonii (black diamonds, black trend line) superimposed on a plot of DMg for inorganic calcite adapted from Morse and Bender [1990]. K is for the partition coefficient of Katz [1973].

[11] Previous field calibrations and laboratory experiments reported the relations between Mg/Ca ratios and temperature for various benthic and planktonic foraminifera. Using these relations we calculated the DMg values for 24°C and compared them with the results for A. lobifera and A. lessonii (Table 2). It is obvious that foraminifera display a large range of DMg values ranging from 0.028 in the neritic hyaline benthic species P. opercularis [Toyofuku et al., 2000], to 0.0006 in the planktonic G. sacculifer [Nürenberg et al., 1996]. Other planktonic species show similarly low values of 0.0016 and 0.0012 for O. universa and G. bulloides, respectively [Russell et al., 2004; Lea et al., 1999]. Deep benthic foraminifera (different Cibicidoides species) show values of 0.0030 [Rosenthal et al., 1997] and 0.0023 [Lear et al., 1999]. These values are higher than the values obtained for the planktonic species by a factor of up to 5. The average DMg values for the Amphistegina species are 0.012 and 0.010 for A. lobifera and A. lessonii, respectively. These values are intermediate between the deep benthic and the shallow benthic species. The planktonic species that bear symbiotic algae (i.e., G. sacculifer and O. universa) show values above and below that of G. bulloides, which does not have symbionts. Amphistegina species, which have symbionts, show DMg values that are higher by a factor of 3–4 than those of the deep benthic species (without symbionts) but are lower by a factor of 3 compared to P. opercularis that have no symbionts. It seems therefore that symbiotic algae do not affect the DMg in foraminifera. Obviously pressure cannot explain this variability and as Toyofuku et al. [2000] have shown, salinity does not affect the DMg of the foraminifera they studied. This suggests that the large variability is caused mainly by vital effects on the incorporation of Mg into foraminiferal calcite (see Bentov and Erez [2006] for detailed discussion).

Table 2. Comparison Between Different Species of Foraminifera and Inorganic Calcitea
SpeciesMg/Ca in 24°C, mmole/moleDMg in Present SeawaterSource
  • a

    The ratio of Mg/Ca in 24°C was calculated using the equations given in each cited reference in the table. In order to obtain the value of DMg the calculated ratio was divided by present-day seawater Mg/Ca of 5.14 [Broecker and Peng, 1982].

  • b

    The DMg of Katz [1973] was obtained at 25°C.

Inorganic calciteb 0.057Katz [1973]
P. opercularis1420.028Toyofuku et al. [2000]
G. sacculifer3.300.00064Nürenberg et al. [1996]
G. bulloides6.130.0012Lea et al. [1999]
C. floridanus15.50.0030Rosenthal et al. [1997]
Different Cibicidoides11.90.0023Lear et al. [2002]
O. universa8.510.0016Russell et al. [2004]
A. lobifera58.60.012this study
A. lessonii49.30.010this study

[12] An important observation is that DMg in our experiments was sensitive to the Mg/Ca ratio rather than the concentration of either Ca2+ or Mg2+. The same behavior was previously reported for MgCO3 incorporation in inorganic calcite overgrowth [Mucci and Morse, 1983]. A physiological mechanism sensitive to ratio, such as a competition between Mg2+ and Ca2+ on the same transport apparatus, remains to be explored. Finally, if Mg/Ca variations in the past oceans can be estimated independently, they should be taken into account while reconstructing paleotemperatures based on Mg/Ca ratios in foraminifera.

[13] In present-day seawater with Mg/Ca ratio of ∼5, the kinetically preferred mineral to be precipitated inorganically is aragonite because calcite growth is inhibited by Mg and possibly other ions [e.g., Davis et al., 2000; Morse and Bender, 1990]. Aragonite is also the mineral deposited by corals and many mollusks. However, except for one family (Robertinacea), all foraminifera precipitate a calcite shell. The large variability in Mg contents of co-existing species suggests that foraminifera have a strong control on the Mg content of their shell. It has been speculated that foraminifera reduce the Mg concentration in the solution from which they precipitate their calcite shells, a process that may require energy [Erez, 2003; Bentov and Erez, 2006]. Because foraminifera use seawater vacuoles as their main source of ions for biomineralization [Bentov and Erez, 2001; Erez, 2003; Bentov and Erez, 2006], a lower ambient Mg/Ca ratio may result in higher calcification rates. This may explain why the highest addition of CaCO3 was observed at Mg/Ca of 1 rather than at present-day ratio (5). Surprisingly, the Mg/Ca ratio of 0.5 produced lower calcification than the Mg/Ca ratio of 1, suggesting that the biomineralization process may require a minimum Mg2+ concentration.

[14] The geological record shows that when the genus Amphistegina first appeared in the Eocene [Loeblich and Tappan, 1988] the seawater Mg/Ca ratio was around 1.5 [Stanley and Hardie, 1999]. This is in agreement with the ratio that displayed the largest addition of CaCO3 in our experiment. Hence it may be suggested that the current Mg/Ca ratio in the oceans is not optimal for foraminiferal calcification. Therefore variations in Mg/Ca ratios of seawater should be taken into account when evaluating the role of foraminifera in the past oceanic carbon cycle.


[15] The authors would like to thank A. Katz for his assistance with solutions calibration. We thank Y. Kolodny, A. Starinsky, A. Katz, S. Bentov, and J. Silverman for critically reading the manuscript and providing helpful comments and discussions, and K. Schneider for assistance with statistics. We thank an anonymous reviewer, who has considerably improved the presentation of this paper. Funding for this project was provided by the US-Israel BSF (2000284) and GIF (G-720-145.8/01). This work is part of the M.Sc. thesis of E. Segev at the Hebrew University of Jerusalem.