Hydrobiologia 517: 1–13, 2004.© 2004 Kluwer Academic Publishers. Printed in the Netherlands.

1

On the celestite-secreting Acantharia and their effect on seawaterstrontium to calcium ratios

Patrick De DeckkerDepartment of Earth and Marine Sciences, The Australian National University, Canberra ACT 0200, AustraliaE-mail: patrick.dedeckker@anu.edu.au

Received 27 April 2003; in revised form 16 August 2003; accepted 10 September 2003

Key words: Acantharia, sarcodine protozoan, strontium sulfate, Indian Ocean, coral palaeothermometry, seawaterchemistry

Abstract

Significant variations in the Sr/Ca in waters from the eastern Indian Ocean in the vicinity of Australia, both from thesurface and from shallow depth profiles, are documented. The strontium sulfate-secreting protozoans Acantharia,which are common in the upper 400 m of the oceans, especially at low latitudes, contribute substantially to changesin the Sr/Ca of oceanic waters by extracting Sr for the formation of their skeletons. Below such depths, theseorganisms dissolve and the Sr/Ca of seawater regains its ‘conservative’ nature. This mechanism accounts for someof the variability of the Sr/Ca near the surface of the ocean, but this may still be found at depth as surface watersbecome entrained at greater depths. It is argued here that the noted Sr/Ca variations may explain discrepanciesbetween coral data sets from different parts of the oceans, and calls for caution when reconstructing sea-surfacetemperatures from the Sr/Ca of corals.

Introduction

Acantharians are common inhabitants in all oceans.This unique group of sarcodine protozoans secretesa skeleton made of the mineral celestite, SrSO4. Be-cause seawater is grossly undersaturated with respectto this mineral, the remains of these planktonic organ-isms readily dissolve and are very rarely encounteredas fossils. Hence, the study of acantharians has beenneglected by geologists and palaeoceanographers, al-though these organisms are among the most abundantmembers of the plankton in the oceans (Beers et al.,1975; Bishop et al., 1977; Michaels, 1988), fre-quently surpassing that of the calcareous test-formingforaminiferans.

Acantharian numbers in the equatorial AtlanticOcean can reach up to 744 specimens m−3, comparedto a value of 105 m−3 at high latitudes (Bottazzi, 1978;Bottazzi & Andreoli, 1982). In the North Pacific Cent-ral Gyre, Michaels (1988) counted 16 specimens perlitre in the upper 20 m of the water column, but num-bers dropped down to 10 between 40 m to 120 m.

In comparison, foraminifers amounted to less than 1specimen per litre in the upper 80 m of these samples(Michaels, 1988). Commonly, the maximum dens-ity of acantharians in the oceans is found between50 and 200 m, but figures vary between studies, andthe type of mesh size for plankton sampling devicescan substantially affect results; even the nature of thepreservative used is important as some selectively dis-solve acantharian spicules (Beers & Stewart, 1970).Also, to undergo gametogenesis, some acanthariansmigrate to depths between 300 and 400 m (Bottazzi,1978). With respect to seasonal abundance, the pi-oneering taxonomic and ecological observations ofSchewiakoff (1926) for the Bay of Naples showedthat these organisms are especially abundant from latespring to the beginning of autumn. In addition, studiesfrom the seas around Italy identified that salinity isalso a controlling factor in the abundance of acanthari-ans (Bottazzi, 1978), with high specimen counts evenat salinities up to 38.3% (Bottazzi & Andreoli, 1978).

During the two Australian cruises in the easternIndian Ocean, on board the RV Franklin, numer-

2

ous acantharian specimens were recovered from sur-face plankton tows at the sites where water samplesand CTD profiles were obtained. The abundanceof the acantharians collected was on average 100times higher than the live foraminifers and radiolari-ans; acantharian densities were consistent with otheroceans (Beers et al., 1975; Bishop et al., 1977;Bottazzi & Andreoli, 1982; Michaels et al., 1995).Acantharians were most abundant between latitudes14◦ and 25◦ S.

The aim of this paper is to present chemical ana-lyses of water samples collected during the 2 cruisesand also document the presence of acantharians col-lected at the same stations.

Material and methods

During two cruises with the RV Franklin in the easternIndian Ocean in late 1995 [Fr10/95] and early 1996[Fr2/96], numerous surface plankton samples werecollected. The plankton samples were collected bytowing a circular plankton net, 40 cm in diameter andwith a 150 µm mesh, through the surface of the oceanfor approximately 1 km. About 4 m3 of water passedthrough the plankton nets. No quantitative effort wasmade to estimate the number of plankton species pervolume of water sampled, although it is estimated thatapproximately 125 m3 of surface water were sampled.Once the towing was completed, the plankton samplewas immediately filtered through a small 10 cm dia-meter net (mesh 60 µm) and rinsed thoroughly withdistilled water to get rid of any remaining sea water(to prevent a flocculant from forming once ethanol isadded to the sample, and also to prevent dissolution ofthe acantharians as seawater is undersaturated with re-spect to celestite – see Discussion later). The sampleswere then preserved in 100% ethanol.

Water samples were collected from both the oceansurface and depth profiles. The water samples usedfor chemical analyses were filtered through a 0.45 µmMillipore� membrane using a peristaltic pump, acid-ified with a few drops of ultrapure HNO3. The sea-water samples were all analysed on a Perkin ElmerElan 6000 ICP-MS, with the use of a Perkin ElmerFIAS 2000 (used as an online auto-diluter) and a Per-kin Elmer AS-90 auto-sampler. Two sets of samplesand standards were prepared. Both sets of samplesand standards were analysed for Ca42,43,44, Sr86,87,88,Ba134,135, Sc45 and Y89. The first set comprised a1 in 10 dilution of the original samples and stand-

ards, the second set was prepared as the first, butincluded spikes of internal standards of Sc45 and Y89,resulting in end concentrations of 9 and 0.9 mg l−1,respectively, in every sample and standard. A further1 to 10 dilution was carried out for all samples andstandards using a Perkin Elmer FIAS 2000 running asan online auto-diluter, feeding directly into a RytonCross Flow Nebulizer/Spray Chamber attached to aPerkin Elmer Elan 6000 ICP-MS. Artificial seawaterwas used to make up the analytical standards; thesewere also crossed-checked against natural seawatersspiked with Ca, Sr and Ba standards to determineanalytical performance of the technique.

In the first set of samples, Sc and Y concentrationsproved to be low [less than 20 µg/l], but raw countswere still used as minor corrections for Sc45 and Y89

concentrations in the results from the analysis of thesecond set of samples. The results of the second setof samples were used only, as the corrected internalstandards, Sc45 for Ca44 and Y89 for Sr88, could beused to correct for matrix interference, instrument driftand short-term noise in the second analysis on theanalyses of interest. A further 8 sub-samples were ana-lysed as duplicates, and 6 sub-samples were analysedusing Methods of Standard addition on both a PerkinElmer Elan 6000 ICP-MS and a Perkin Elmer 3100Atomic Absorption Spectrometer, as cross-checks onthe original analysis. Considering ICP-MS analysis isnot the ‘favoured’ technique for the analysis of sea-water, this technique showed considerable promise.However, errors associated with large dilutions (thiswas partially corrected for with the addition of internalstandards at the beginning of the dilution stages), andthe production of a suitable standard solution, matrix-matched to the samples, continued to cause smallproblems. All samples with standards greater than 5%were re-analysed.

Results

A total of 65 surface water samples was analysed fromthe region between 108◦ E and 13◦ E and 12◦ S and32◦ S and their location are shown in Fig. 1a. Notethat none of the surface samples were taken in lessthan 50 km from the Australian coastline to preventpossible ‘contamination’ from likely fluvial sources.A broad range of salinities was encountered spanningfrom S = 33.69 to S = 35.91 and those values areschematically contoured in Fig. 1b. Variations werealso noted for Ca and Sr measurements, and again the

3

Figure 1. Maps of the eastern Indian Ocean adjacent to Western Australia showing the location of the sea-surface samples (dots) taken duringthe two Franklin cruises in December 1995 (= Fr10/95) and February–March 1996 (= Fr2/96) (1a) and the location of the rivers whichcontribute to continental outflow, especially during the monsoonal season. 1(b) shows the salinity contouring based on the data obtained duringthe 2 cruises; 1(c) displays the contouring for Sr/Cl (as µg/g); 1(d) the measured molar Sr/Ca values in mmol/mol; 1(e) and 1(f) display thecontouring for Ca and Sr in ppm respectively. Chlorinity values used in 1c were obtained using the relationship of S(�) = 1.80655 Cl(�)(Chester, 1990). Salinity values were obtained by conductivity measurements while at sea, and recalibrated later by CSIRO personnel. Thelocation of the 3 sites with CTD profiles displayed in Fig. 2 is shown by black crosses in the above map.

values are schematically contoured in Figs 1e, f. Ra-tios of Sr/Cl as well as Sr/Ca were also computed andare displayed in the same fashion in Figs 1c, d. Watersamples from three CTD [conductivity-temperature-depth] profiles were selected for chemical analysesand the Sr/Ca results are presented against salinityin Figs 2a–c. All data used for the present paper areavailable on the Springer server.

Discussion

Surface waters

Surface water data from these Franklin cruises showa typical patchiness with respect to all measured ele-ments, either tabled singly or displayed as ratios suchas Sr/Ca or Cl/Sr (Figs 1a–f). The latter ratio was usedto determine whether a relationship could be foundbetween salinity and the concentration of Sr. Thiscould not be found. Patchiness of Sr values in upperwaters in the oceans had already been documented inRenard’s (1985) review, and more recently for several

4

Figure 1(b). (Continued.)

sites adjacent to the Hawaiian islands where coralsgrow (de Villiers et al., 1995).

Close examination of the location of the majorrivers bordering the Western Australian coast (Fig. 1a)shows no obvious links between chemical and salin-ity variations found offshore (Figs 1b–f), despite thefact that half of the water samples were collectedin December 1995 when three cyclonic depressionsoccurred in the region with substantial rainfall, andriver discharge ensuing. Equally, no link was foundin February–March 1996 when the rest of the watersamples were taken.

Overall, salinity displays lower values at the low-est latitudes and highest salinities are recorded near28◦–30◦ S where evaporation is highest in the region.Nevertheless, salinity values ≥ 34.6 are noticeable on

the northwest shelf and these correspond also to theoverall high Sr/Ca values.

A possible explanation being that very likely fineneedles of CaCO3 may precipitate due to the highsalinities and temperatures at the end of the ‘dry’[=austral winter] season caused by some supersatur-ation in CaCO3. However, this hypothesis requirestesting as a comparison between the plots for salin-ity (Fig. 1b) and Cl/Sr (Fig. 1c) show no obvioussimilarity with Sr/Ca (Fig. 1d).

Depth profiles

Characteristically, in some profiles obtained us-ing conductivity-temperature-depth [CTD] equip-ment, Sr/Ca ratios also show a systematic depletionof Sr in the upper portion of the water column (see

5

Figure 1(c). (Continued.)

Figs 2a–c). This pattern is reminiscent of profilesfrom the Atlantic (McKenzie, 1964; Brass & Turekian,1972, 1974; Bernstein et al., 1987) and Pacific (Brassand Turekian, 1973, 1974; Brass, 1980) Oceans, andthe same was recently found also by Müller & DeDeckker (2003) from a region near Tasmania and NewZealand at the confluence of the Indian, Pacific andSouthern Oceans. More importantly, de Villiers (1999)concluded from her work in the Atlantic and PacificOceans that ‘seawater Sr and Sr/Ca exhibit spacialgradients of 2–3% globally with the deep ocean en-riched relative to the surface’. This is investigatedfurther below by attempting to determine the cause ofseawater Sr/Ca variability.

However, despite the lowest values recognizednear the surface in all three CTD profiles (for ex-planation see Discussion below), another phenomenon

showing a variation in the Sr/Ca with depth is notice-able. Examination of the salinity–temperature profilesfor the 3 CTD sites (Station Fr2/95-36 at 14◦ 48.68′ S114◦ 16.37′ E; Station Fr10/95-39 at 28◦ 44.35′ S112◦ 47.89′ E; station Fr10/95-28 at 20◦ 01.93′ S112◦ 43.37′ E) help identify the various water massesencountered in the region. Importantly are (1) the In-donesian Throughflow water (= ITF in Fig. 3) which ischaracteristically of low salinity, (2) the Indian CentralWater (= ICW in Fig. 3) which is at the other extremewith its highest salinity, and (3) the other two watermasses the Subtropical South Indian water (= SSIW inFig. 3) and the South Indian Central Water (= SICW inFig. 3). The latter two water masses are intermediarybetween the ITF and the ICW. It is noteworthy thatthe ICW, recognized in Fig. 2c at Station Fr10/95-28between ∼200 and 300 m water depth, also has a high

6

Figure 1(d). (Continued.)

Sr/Ca as for the other 2 profiles below the surficialSr/Ca-‘depleted’ layer. The surprising feature, how-ever, is that below the ICW lens at Station Fr10/95-28,Sr/Ca return to low values; these waters belong towatermasses (SSIW and SICW) that originated at thesurface at higher latitudes, and at the time of theirformation, may also have been ‘depleted’ in Sr/Ca andremained as such.

The acantharian skeleton and seawater Sr/Ca

The acantharian skeleton is made of fine SrSO4needles/spicules and rapidly becomes dismantledupon death once the protoplasm decomposes, fallsthrough the water column [probably fairly rapidly dueto its high (3.96) specific density] and consequentlydissolves readily (Beers et al., 1975). Acantharians

are rare or non existent below 1500 m as recognisedfrom trap studies (Beers & Stewart, 1970). This ex-plains the change in the Sr composition of seawateras acantharians utilise a substantial amount of Sr inthe upper parts of the water column where they liveand secrete their test, but this element is returned intosolution once spicules dissolve while they sink. Sev-eral studies have alluded to the substantial uptake of Srby acantharians from their ambient waters. Estimatesfor one acantharian vary between 20.6 µg cm−2 yr−1

(Bernstein et al., 1987) and 22 to 40 µg cm−2 yr−1

(Brass, 1980). Again, estimates vary between theamount of Sr acantharian skeletons contain; 0.2 nmol(Bishop et al., 1977) and between 5–49 nmol (Bern-stein et al., 1987) based on mean weights. Assuming amean value of 20 nmol of Sr, an acantharian specimenwould require 0.76 µg of Sr to fabricate its acicu-

7

Figure 1(e). (Continued.)

lar skeleton; for an average number of 10 specimensper litre at anyone time, up to 7.6 µg of Sr wouldbe taken from the ambient seawater. As an example,with 2 acantharian generations per year, and an aver-age residence time of 25 years for the upper watersin the ocean, an acantharian population would haveextracted 380 µg of Sr from 1 l of seawater over thoseyears. By this process, the Sr can be extracted fromthe original water sample, provided there is no mixingwith underlying waters, and explains the 5% areal andvertical variations of Sr/Ca in the upper 400 m of thewater column.

de Villiers (1999) discussed the characteristic de-pletion of Sr variations in the upper few hundredmetres in her samples form the Atlantic and PacificOceans with respect to deeper horizons below. Sheconcluded that the dissolution of acantharians contrib-

ute to a significant increase in the upper part of theocean [below the horizon near the surface which isdepleted in Sr as a result of Sr uptake from acanthari-ans], but that also a secondary influence is causedby calcium carbonate cycling. It is suggested herethat different upper-ocean water masses may, in fact,also have Sr/Ca values as illustrated in Fig. 2c for theprofile which displays an obvious correlation betweensalinity and Sr/Ca. It may be that partially-evaporatedseawater may have chemically evolved differently,such that aragonite needles may have precipitated athigh salinities [and high temperatures], thus forcing a‘partial’ depletion in Sr values. Such a phenomenoncould be expected on the northwest shelf of Australiawhere intense evaporation does frequently occur andhigh salinities are generated.

8

Plate 1. SEM photos of acantharians collected in plankton tows during the RV Franklin cruises in the eastern Indian Ocean. Collection stationswith coordinates are listed after the identification. (A) Acanthocolla cruciata Haeckel with spicules partly broken (Fr2/96–19; 15◦ 00.12′ S110◦ 29.80′ E); B, F: Acanthostaurus purpurascens Haeckel (Fr2/96-26; 21◦ 59.76′ S 107◦ 42.71′ E); C: Phyllostaurus siculus Haeckel withshrunk protoplasm still visible (Fr2/96-18); D: Amphibelone hydrotomica Haeckel with juvenile planktonic foraminifers stuck in between thespicules (Fr2/96-18); E: Coleapsis vaginata Haeckel (Fr2/96-26). Note scale bars for all specimens represent 50 µm and the background whichconsisted of carbon tape has been artificially erased.

9

Plate 2. SEM photos of acantharians collected in plankton tows during the RV Franklin cruises in the eastern Indian Ocean. Same remarks aboutcollection stations as for Plate 1. (A), (B) Diploconus aff. fasces Haeckel (note juvenile planktonic foraminifers for relative scale) (Fr2/96-19);(C) Amphibelone anomala Haeckel (Fr2/96-19; 26◦ 59.68′ S 109◦ 29.68′ E); D: Acanthospira spiralis Haeckel with possible radiolarian stuckamong the spicules and with some shrunk protoplasm still visible (Fr2/96-36; 14◦ 48.68′ S 114◦ 16.37′ E); (E) Phyllostaurus siculus Haeckelwith shrunk protoplasm still visible and a juvenile planktonic foraminifer adhering to some of the spicules (Fr2/96-36); (F) possible earlystage of Pleuraspis costata Müller with some protoplasm also visible (Fr2/96-19). Note scale bars for all specimens represent 50 µm and thebackground which consisted of carbon tape has been artificially erased.

10

Figure 2. Continuous salinity profiles measured by CTD (bold line)and Sr/Ca (in mmol mol−1) (thin line with dots showing the depthsof sampling using Niskin bottles) taken for 3 selected sites duringthe two Franklin cruises (see text for more details) to show thecharacteristic depletion in Sr in at least the upper 200 m of theocean, and the increase at various levels down to 1000 m whereacantharians would completely dissolve. Note the broad range ofsalinities resulting from the sampling of different water masses, andthe different depths displayed for those profiles. In particular, notethe relationship in (2c) between salinity and the Sr/Ca measured atstation Fr10/95-28.

Figure 3. Temperature-salinity profiles obtained with CTD equip-ment for the 3 stations discussed in the text (36 = Fr2/96-36; 39 =Fr10/95-39; 28 = Fr10/95-28]. The various watermasses of relev-ance to the discussion in this paper are labeled on these profiles.These are: ITF for Indonesian Throughflow; SSIW for SubtropicalSouth Indian Water; SICW for South Indian Central Water; ICW forIndian Central Water (nomenclature after Fieux et al., 1996).

Implications of seawater Sr/Ca

The residence times for calcium and strontium intoday’s oceans have been estimated to be of the orderof 1.1 × 106 and 4.9 × 106 years, respectively (Martin& Whitfield, 1983; Chester, 1990) and, consequently,the ratio of these two elements has been considered bymany to be constant for the entirety of the Quaternary.This assumption led to the use of Sr/Ca fluctuation de-tected in the skeleton of modem and fossil corals as aproxy for sea-surface temperature changes in tropicaloceans (Smith et al., 1979; Beck et al. 1992; McCul-loch et al., 1994; Min et al., 1995; Beck et al., 1997;Alibert & McCulloch, 1997; McCulloch et al., 1999;Schrag, 1999). Nevertheless, there have been amplepublished accounts documenting significant variationsin either Sr, or in the Sr/Ca and Sr/Cl ratios of seawaterin the open oceans (Andersen et al., 1970; Brass &Turekian, 1972, 1973, 1974; Bernstein et al. 1987;Bernstein et al., 1992; de Villiers et al., 1994; deVilliers, 1999). These findings have been ignored bymost of those using the coral Sr/Ca as a palaeother-mometer. These contradicting views need resolving,especially since there are now several claims for sub-stantial sea-surface temperature differences betweenthe present and previous episodes of the Late Quatern-ary (e.g., for the Last Glacial Maximum) (Guildersonet al., 1994), or younger episodes (Beck et al., 1992,

11

Plate 3. SEM photos of acantharians collected in plankton towsduring the RV Franklin cruises in the eastern Indian Ocean. Sameremarks about collection stations as for Plate 1. (A) Acanthostauruspurpurascens Haeckel with some shrunk protoplasm visible and asmaller acantharian adhering to some of its spicules (Fr2/96-19); (B)Amphibelone hydrotomica Haeckel (Fr2/96-15; 23◦ 44.23′ S 108◦31.91′ E); C: Amphilithium clavarium var. lanceolatum Schewiakoffwith several planktonic foraminifers and one radiolarian adheringto numerous other ancantharian spicule debris; one dinoflagel-late is also visible at the opposite end of the debris amalgamate(Fr2/96-36). Note scale bars for all specimens represent 50 µm andthe background which consisted of carbon tape has been artificiallyerased.

1997) that are in conflict with sea-surface temperat-ure reconstructions using other proxy methods suchas foraminiferal faunal assemblages (CLIMAP, 1981;Anderson et al., 1989; Martinez et al., 1997) or Uk37

analyses from the remains of calcareous nannoplank-ton (Bard et al., 1997). Here, additional chemicalanalyses of seawater samples from the eastern IndianOcean are presented and these identify that the Sr/Caof waters at the surface and through profiles is againnot constant throughout in an oceanic region wherecorals live.

Seawater Sr/Ca and coral palaeothermomet

The analytical data presented here seriously questionthe validity of using the Sr/Ca of fossil corals in recon-structions of past sea-surface temperatures, becausecalculated temperature shifts (Beck et al., 1992; Guild-erson et al., 1994; Min et al., 1995; Beck et al., 1997;Schrag, 1999; McCulloch et al., 1999) all assume aconstant or near-constant Sr/Ca for the water in whichthe corals grew. The present study, in addition to othersmentioned above, shows that the Sr/Ca of seawater,especially in the realm where corals grow, cannot beaccepted to be constant through time and space.

It is noteworthy that the several sites and periodsfor which fossil corals have been analysed representlocations – and periods of time – during which salinitymay have changed as a result of reduced precipitation.For example, the ‘postulated’ substantial temperaturedrop (6.5 ◦C) at Vanuatu (Beck et al., 1992, 1997)appears to coincide with the occurrence of a signific-ant period of aridity registered in northern Australiaat a similar latitude (De Deckker et al., 1991), im-plying a possible reduction of runoff from the islandthat could cause a change in the Sr/Ca but not ne-cessarily a change of sea-surface temperature (SST).Foraminifer assemblages from nearby ODP core 828(Martinez et al., 1997), detect a shift of summer SSTof no more that 1 ◦C over the last 20 Ka, challengingthe coral Sr/Ca evidence (Beck et al., 1997).

Similarly, for the Barbados coral (Guilderson et al.,1994) which grew during the Last Glacial Maximum(= LGM), a reconstructed SST (e.g., a drop of 5 ◦Ccompared to today) may be misleading because ofmuch more arid conditions in the region. So far, nolacustrine cores have returned LGM sediments fromthe island, very likely due to the extreme dry condi-tions there during the LGM. The postulated decreasein precipitation is confirmed for adjacent areas in theCaribbean (Bradbury et al. 1981) and could have

12

caused the increase in the salinity and δ18O of the am-bient waters that was originally interpreted as resultingfrom an SST drop (Guilderson et al., 1994). In addi-tion, at the LGM 1.2� must be added to the salinity toadjust for the fall in global sea level. Similarly, an in-crease of the Sr/Ca of seawater by 3% (due to a changein the supply of local Sr-depleted continental waters)near the coast where corals usually grow, or a lackof local acantharian productivity, for example, wouldhave registered as a 5 ◦C temperature drop by the coralSr/Ca palaeothermometer. Modelling studies (Stoll &Schrag, 1998) on the effect of sea level changes dur-ing the Quaternary also postulate that Sr in seawaterwould have changed during periods of low sea levels.No quantitative measurements have so far been madeto verify these models but they support the argumentthat the basis of coral Sr/Ca thermometry for parts ofthe Quaternary ought to be looked at more closely.

Conclusion

This paper argues that acantharians, despite their smallsize, play an important role in the Sr depletion inthe upper few hundred metres of the ocean’s watercolumn. It is argued also that if such ‘Sr-depleted’ wa-ters eventually ‘sink’ or become covered by the watermasses, the Sr depletion may remain at greater depths.Such a phenomenon may perhaps be used to actuallytrace the origin of a water mass.

Despite the lack of uniformity in the Sr/Ca of seawater near the surface as shown in this paper, thecoral Sr/Ca thermometer remains a very powerful toolfor demonstrating secular variations and for recenttimes’ reconstructions (e.g., the last millennium) asoceanic conditions for that time frame would havechanged little. Nevertheless, several studies have iden-tified that the temperature equation using the Sr/Ca inthe same species of coral varies from place to place(de Villiers et al., 1994; Beck et al., 1997; Alibert &McCulloch, 1997), showing that oceanic Sr/Ca var-ies between sites. However, in order to extend pastSST reconstructions for periods of the record whichregistered significantly different conditions – the ex-treme being during the LGM – such as salinity changes(which could affect acantharian productivity levels (ef-fectively changing the Sr/Ca of seawater locally), lightintensity, depth of the nutricline, wind-induced turbu-lence and mixing of the surface layer of the ocean)caution is definitely warranted. A number of prox-ies, used together at the same site – such as corals

and other methods/organisms – are essential beforeinferring substantial sea-surface temperature changesat low latitudes during the LGM.

Acknowledgements

I am thankful to the captains and crew of the Aus-tralian National Facility RV Franklin for facilitatingthe collection of all samples in the eastern IndianOcean, and the numerous colleagues who helped whileat sea, and for several ARC grants to support this work.I benefited from the insight of Dr D. Nürnberg (Geo-mar, Kiel) who suggested that acantharians play a rolein the uptake of Sr in the oceans. He also providedkey references and commented on an earlier draft ofthe manuscript. Mr F. Krikowa of the University ofCanberra performed the analyses presented here. DrT. Corrège provided some cross-checking analysesby ICP-MS. Very profitable discussions with MichaelShelley during the early stages of writing of this paperhelped improve the clarity of the arguments and text.Mr C. Hilliker drew the figures and Mr A. Sadekovhelped with the photographs. Dr R. Barwick was alsoprovided much valuable advice with the illustrations.

I am also grateful to Dr E. Gliozzi who man-aged to obtain an original of the publication andplates of the important and outstanding work of W.Schewiakoff from the Zoological Station in Naplesthrough negotiations with the librarian there.

References

Alibert, C. & M. T. McCulloch, 1997. Strontium/calcium ratios inmodern Porites corals from the Great Barrier Reef as a proxyfor sea-surface temperature: calibration of the thermometer andmonitoring ENSO. Paleoceanography 12: 345–363.

Andersen, N. R., J. D. Gassaway & W. E. Maloney, 1970. The re-lationship of the strontium: chlorinity ratio of water masses inthe tropical Atlantic Ocean and Carribean Sea. Limnology andOceanography 15: 467–472.

Anderson, D. M., W. L. Prell & N. J. Barratt, 1989. Estimates of seasurface temperature in the Coral Sea at the last glacial maximum.Paleoceanography 4: 615–627.

Bard E., F. Rostek & C. Sonzogni, 1997. Interhemispheric syn-chrony of the last deglaciation inferred from alkenone palaeo-thermometry. Nature 385: 707–710.

Beck, J. W., R. L. Edwards, E. Ito, F. W. Taylor, J. Récy, F. Rougerie,P. Joannot & C. Henin, 1992. Sea-surface temperature from coralskeletal strontium/calcium ratios. Science 257: 644–647.

Beck, W. T. , J. Récy, F. Taylor, R. L. Edwards & G. Cabioch,G., 1997. Abrupt changes in early Holocene tropical sea surfacetemperature derived from coral records. Nature 385: 705–707.

13

Beers, J. R., F. M. H. Reid & G. L. Stewart, 1975. Microplankton ofthe North Pacific Central Gyre. Population structure and abund-ance, June 1973. Internationale Revue Gesamtes Hydrobiologia60: 607–638.

Beers, J. R. & G. L. Stewart, 1970. The preservation of acanthari-ans in fixed plankton samples. Limnology and Oceanography 15:825–827.

Bernstein, R. E., P. R. Betzer, R. A. Feely, R. H. Byrne, M. F.Lamb & A. F. Michaels, 1987. Acantharian fluxes and stron-tium to chlorinity ratios in the North Pacific Ocean. Science 237:1490–1494.

Bernstein, R. E., R. H. Byrne, P. R. Betzer & A. M. Greco,1992. Morphologies and transformations of celestite in seawater:the role of acantharians in strontium and barium geochemistry.Geochimica et Cosmochimica Acta 56: 3272–3279.

Bishop, J. K. B., J. M. Edmond, D. R. Ketten, M. P. Bacon & W. B.Silker, 1977. The chemistry, biology, and vertical flux of par-ticulate matter from the upper 400 m of the equatorial AtlanticOcean. Deep-Sea Research 24: 511–548.

Bottazzi, E. M., 1978. Systematic-ecological aspects of Radiolariawith special reference to Acantharia. Bolletino Zoologica 45:133–144.

Bottazzi, E. M. & M. G. Andreoli, 1978. Distribuzione stagionaledegli Acantari e dei Radiolari (Protozoa, Sarcodina) in di-verse zone costiere dei mari italiani. L’ Ateneo Parmense ActaNaturalia 14: 477–500.

Bottazzi, E. M. & M. G. Andreoli, 1982. Distribution of adult andjuvenile Acantharia (Protozoa, Sarcodina) in the Atlantic Ocean.Journal of Plankton Research 4: 757–777.

Bradbury, J. P., B. Leyden, M. Salgado-Laboriau, C. Lewis, C.Schubert, M. Binford, D. Frey, D. Whitehead & F. Weibezahn,1981. Late Quaternary environmental history of Lake Valencia.Science 214: 1299–1305.

Brass, G. W., 1980. Trace elements in acantharian skeletons.Limnology and Oceanography 25: 146–149.

Brass, G. W. & K. K. Turekian, 1972. Strontium distributions in seawater profiles from the GEOSECS I (Pacific) and GEOSECS II(Atlantic) test stations. Earth and Planetary Science Letters 16:117–121.

Brass, G. W. & K. K. Turekian, 1973. Stontium and alkalinity vari-ations in the South Pacific. In Fraser, R. (ed.), Oceanography ofthe South Pacific 1972, New Zealand National Commission forUNESCO, Wellington: 15–18.

Brass, G. W. & K. K. Turekian, 1974. Strontium distribution inGEOSECS oceanic profiles. Earth and Planetary Science Letters23: 141–148.

Chester, R., 1990. Marine Geochemistry. Unwin Hyman, Boston.CLIMAP Project Members, 1981. Seasonal reconstruction of the

earth’s surface during the last glacial maximum. Map and ChartSeries, No. 36. Geological Society of America, Boulder.

De Deckker, P., T. Corrège & J. Head, 1991. Late Pleistocene recordof eolian activity from tropical northeastern Australia suggestingthe Younger Dryas is not an unusual climatic event. Geology 19:602–605.

de Villiers, S., 1999. Seawater strontium and Sr/Ca variability in theAtlantic and pacific oceans. Earth and Planetary Science Letters171: 623–634.

de Villiers, S., G. T. Shen & B. K. Nelson, 1994. The Sr/Ca temper-ature relationship in coralline aragonite: influence of variabilityin (Sr/Ca) seawater and skeletal growth parameters. Geochimicaet Cosmochimica Acta 58: 197–208.

de Villiers, S., B. K. Nelson & A. R. Chivas, 1995. Biologicalcontrols on coral Sr/Ca and δ18O reconstructions of sea surfacetemperatures. Science 269: 1247–1249.

Fieux, M., C. Andrié, E. Charriaud, A. G. Ilahude, N. Metzl, R.Molcard & J. C. Swallow, 1996. Hydrological and chlorofluoro-methane measurements of the Indonesian Throughflow enteringthe Indian Ocean. Journal of Geophysical Research 101 (C5):12,433–12,454.

Guilderson, T. P., R. G. Fairbanks & J. L. Rubenstone, 1994. Trop-ical temperature variations since 20,000 years ago: modulatinginterhemispheric climate change. Science 263: 663–665.

Martin, J. H. & M. Whitfield, 1983. The significance of river inputsof chemical elements to the ocean system. In Wong, W. S., E.Boyle, K. W. Bruland, J. D. Burton & E. D. Godberg (eds), TraceMetals in Sea Water. Plenum, New York: 265–296.

McCulloch, M. T., M. K. Gagan, G. Mortimer, A. R. Chivas & P. J.Isdale, 1994. A high resolution Sr/Ca and δ18O coral record fromthe Great Barrier Reef, Australia, and the 1982–1983 El Niño.Geochimica et Cosmochimica Acta 58: 2747–2754.

McCulloch, M. T., A. W. Tudhope, T. M. Esat, G. E. Mor-timer, J. Chappell, B. Pillans, A. R. Chivas & A. Omura,1999. Coral record of equatorial sea-surface temperatures dur-ing the penultimate deglaciation at Huon Peninsula. Science 283:202–204.

McKenzie, F. T., 1964. Strontium content and variable strontiumchlorinity relationship of Sargasso sea water. Science 146: 517–518.

Martinez, J. I., P. De Deckker & A. R. Chivas, 1997. New estimatesfor salinity changes in the Western Pacific Warm Pool duringthe Last Glacial Maximum: oxygen-isotope evidence. MarineMicropaleontology 32: 311–340.

Michaels, A. F., 1988. Vertical distribution and abundance ofAcantharia and their symbionts. Marine Biology 97: 559–569.

Michaels, A. F., D. A. Caron, N. R. Swanberg, F. A. Howse &C. M. Michaels, 1995. Planktonic sarcodines (Acantharia, Ra-diolaria, Foraminifera) in surface waters near Bermuda: abund-ance, biomass and vertical flux. Journal of Plankton Research 17:131–163.

Min, G. R., R. L. Edwards, F. W. Taylor, J. Recy, C. D. Gallup& J. W. Beck, 1995. Annual cycles of U/Ca in coral skeletonsand U/Ca thermometry. Geochimica et Cosmochimica Acta 59:2025–2042.

Müller, A. & P. De Deckker, 2003. Magnesium, calcium and stron-tium in waters of the southern Tasman Sea at the confluenceof the Indian, Pacific and Southern Oceans. Marine FreshwaterResearch 53 (7): 1115–1128.

Renard, M., 1985. Géochimie des carbonates pélagiques. Docu-ments du BRGM 85: 1–650.

Schewiakoff, W. 1926. Die Acantharia. Fauna e Flora del Golfo diNapoli 37: 1–755.

Schrag, D. P., 1999. Rapid analysis of high-precision Sr/Ca ratios incorals and other marine carbonates. Paleoceanography 14: 97–102.

Smith, S. V., Buddemeir, R. W., Redalje, R. C. & Houck, J. E., 1979.Strontium-Calcium thermometry in coral skeletons. Science 204:404–407.

Stoll, H. M. & D. P. Schrag, 1998. Effects of Quaternary sea levelcycles on strontium in seawater. Geochimica et CosmochimicaActa 62: 1107–1118.