Geochimica et Cosmochimica Acta, Vol.60, No. 18, pp.

3457-3470, 1996
Copyright 0 1996 ElsevierScienceLtd
Pergamon Printedin the USA.All rightsresened
0016-7037/96 $15.00+ .oO
PI1 SOO16-7037( 96)00171-8

Comparative determinations of trace and minor elements in coral aragonite by ion

microprobe analysis, with preliminary results from Phuket, southern Thailand
N. ALLISON*
Department of Geology and Geophysics, Grant Institute, Kings Buildings, West Mains Road, Edinburgh EH9 3JW, UK

(Received July 7, 1995; accepted in revised form May 19, 1996)

Abstract-The ion microprobe technique was developed for semiquantitative analysis of trace and
minor elements in coralline aragonite. The size of the microprobe beam (ca. 30 pm diameter) allows
an examination of the spatial geochemical heterogeneity of coral skeletons on a much finer scale than
has previously been possible. Of the elements investigated magnesium, calcium, rubidium, strontium,
barium, and uranium appeared suitable for analysis under the operating conditions used although it was
not possible to unequivocally confirm that “Rb and 238Uwere not subject to molecular interferences. A
homogenous carbonate standard was only available for magnesium so some relative ion yields were
calculated using a standard silicate glass. Magnesium, rubidium, strontium, and barium concentrations
were in reasonable agreement with previously reported values. Uranium concentrations were considerably
lower, probably reflecting a miscalibration of the ion microprobe. Reproducibility of analysis was investi-
gated by comparing immediately adjacent pairs of analyses on coral material. Although analytical
accuracy was questionable, precision was good and concentration differences between samples reflect
true variations in geochemistry. High resolution time series analyses of magnesium, rubidium, strontium,
barium, and uranium in Porites lutea corals indicated that the aragonite was heterogeneous on a small
spatial scale (over distances attributed to less than one month’s skeletal growth). This heterogeneity
does not appear to reflect expected variations in seawater temperature or chemistry. It may reflect the
metal binding capacity of a heterogeneous organic matrix unevenly distributed throughout the coral
skeleton. Individual coral colonies deposited substantially different ranges of Mg and Ba, within and
between reef sites, which could lead to significantly different seawater:coral distribution coefficients
being calculated between individuals of the same coral species.

1. INTRODUCTION examination of the spatial heterogeneity of coral skeletons on

The incorporation of trace and minor elements in coral skele- a much finer scale than has previously been possible. An
tons has been examined in a variety of contexts: in the eluci- understanding of the extent of heterogeneity in coral skeletons
dation of calcification processes, in relation to polluting ac- could provide valuable insights into the mode of incorporation
of trace and minor elements and their likely reliability as
tivities and as a tracer of the chemistry and temperature of
tracers of past environmental conditions. In addition, if the
local seawater. For example, variations in skeletal strontium
microprobe beam is positioned accurately on thin sections,
content have been correlated with changes in seawater tem-
then only those elements intimately associated with the skele-
perature (Schneider and Smith 1982; Beck et al., 1992).
tal carbonate will be analysed. Particulate materials in skeletal
Coralline barium has been related to variations in seawater
pores and ions adsorbed onto the skeletal walls are avoided.
temperature, to variations in the Ba/Ca seawater ratio de-
Previous work (Allison and Tudhope, 1992) has shown that
pendant on oceanic upwelling (Lea et al., 1989; Shen et al.,
centres of calcification and microborings in the skeleton can
1992), and to terrestrial inputs from freshwater runoff or
contain elevated concentrations of certain elements, e.g., bar-
from river discharge (Shen and Sanford, 1990). Several
ium and magnesium, which do not accurately reflect the ele-
pathways exist for the incorporation of minor elements in
ment concentration in the surrounding aragonite.
the skeleton. Single atoms or whole metal complexes, e.g.,
Although the ion microprobe technique has been estab-
uranyl carbonates (Swart and Hubbard, 1982) may substitute
lished for some time, there is little published literature on
in place of calcium atoms or CaC03 groups in the aragonite
its application to carbonate minerals (Swart, 1990; Mason,
lattice. Metals may precipitate directly onto exposed skeleton
1987; Veizer et al., 1987). As a consequence, a considerable
if the overlying tissue becomes damaged or is retracted under
amount of time in this study was spent developing sample
stressful conditions (Brown et al., 1991). Other elements
preparation techniques and determining the reproducibility
may be incorporated in an organic phase, derived from the
of analyses.
matrix produced by the coral during calcification, which may
have a substantial metal binding capacity (Mitterer, 1978).
2. MATERLALS AND METHODS
During the present study, polished thin sections of coral
material were analysed using an ion microprobe. The size 2.1. Sample Preparation
of the microprobe beam (ca. 30 ,um diameter) allows an
Thin sections were prepared from several specimens of
the massive coral Porites lutea. Corals were collected in July
* Present address: Department of Pharmacy, University of Brigh- 1991 from three reefs around the Phuket Marine Biological
ton, Cockcroft Building, Moulsecoomb, Brighton BN2 4GJ, UK. Centre, southern Thailand. Two of the reefs, Porites Bay
3457
3458 N. Allison

(PB) and Aquarium Reef (AQ), occur on the southeastern water vapour in the system. Interferences caused by molecular ions
were removed or reduced by the energy filtering method as detailed
peninsula of Ko Phuket, while the third reef, Ko Racha
in Shimizu et al. (1978) and Zinner and Crozaz (1986). In brief,
(KR) , fringes one of the offshore islands. A detailed descrip- the kinetic energy spectra for molecular ion species are usually
tion of the field area and reefs is not relevant to this study narrower and drop off in counts more rapidly than for atomic species.
but can be found in Brown et al. (1986). By using an altered accelerating voltage (a voltage offset), data can
Coral heads were soaked in a diluted solution of sodium be collected from a segment of the energy spectrum which contains
proportionally less of these interfering species. Throughout this study
hypochlorite (ca. 3-4%) for 48 h after collection to remove a voltage offset of -75 V and an energy window of 40 V was used.
coral tissue and were rinsed in free-flowing water for a fur- This is discussed in more detail below.
ther 48 h. Coral skeletal structures were examined by binocu- Secondary ion counts were measured using an electron multiplier
lar microscope to aid species identification. All corals in- by sequential stepping of the magnetic field through a cycle of
relative atomic mass. Counting times for each isotope varied ac-
cluded in this study exhibited morphological characteristics
cording to its abundance but did not exceed 30 s. Data were collected
of Porites lutea (using the keys of Veron, 1986), although for ten consecutive cycles in each analysis. Series of consecutive
considerable variations in corallite structure were observed cycles were averaged, the standard error calculated and the result
between colonies of apparently the same species. corrected for background counts (ranging from 0.00 to 0.06) and
Corals were sectioned along the axis of vertical growth deadtime. Data were collected during four sessions from December
1991-January 1993. Initially the following isotopes were analysed:
into 1 cm thick slices using a rock saw and were photo- 26Mg,27Al,?Si, “Ca, s4Fe, 55Mn,“Rb, 88Sr,“?Zd, “?Sn, ‘38Ba,13ta.
graphed under UV radiation to record fluorescent banding l4OCe,*08Pb, *?h, and 238U As discussed below, some of these
patterns. Bright-dull fluorescent couplets represent approxi- elements proved unsuitable for analysis due to possible contamina-
mately annual growth in the corals in this area, with bright tion problems, molecular interferences, or poor counting statistics.
band accretion beginning in approximately November (Scof-
fin et al., 1992). Cross-sections were cut through each slice, 2.2.1. Poor counting statistics
along the axis of maximum growth and were made into thin Cadmium, tin, lanthanum, cerium, and thorium concentrations in
sections. Two cross-sections were prepared from coral AQ8, coral are low and the relevant isotopes for these elements were
called AQ8i and AQSii. All thin sections were produced frequently not detectable by this method.
from carbonate deposited during the period approximately
November 1988-July 1991. Samples were fixed in Buehler 2.2.2. Confaminafion problems
epo-thin low viscosity epoxy resin onto 25 mm diameter “Al and 28Sicounts exhibited an exponential decrease, tailing off
glass discs and ground down to approximately 30 pm thick- with time over the ten cycle analytical period, suggesting that these
ness. Samples were finely polished using a variety of regimes isotopes or their isobars were present as contaminants on the surface
including mechanical polishing using Kemet industrial dia- of the thin section. This contamination may be introduced from the
polishing compounds or may result from the smearing of the epoxy
mond paste or alumina paste and polishing by hand using
resin fixative over the coral surface. ‘s’Ba and 138Bacounts also
diamond grit suspended in petroleum jelly. Analyses from decreased with time in the alumina polished samples but were stable
sections prepared by the different methods were compared in those mechanically and hand polished with diamond paste. All
to assess the potential for sample contamination during prep- barium concentrations reported in this study were obtained from thin
sections polished with diamond paste.
aration and polishing. Detailed photographs were produced
A decrease of 2-4% was also observed in the counts of *6Mg
of each thin section in transmitted light to record the place- over a typical analysis of 20 min duration. However, this decrease
ment of microborings and centres of calcification. These was approximately linear and showed no sign of tailing off even in
areas may be associated with geochemical anomalies and analyses which exceeded 1 h. This decrease may reelect the declining
were avoided during analysis. Polished sections were sput- signal of an interference which is either associated with the surface
of the thin section or migrates into the sputtering site during analysis.
ter-coated with a thin even-thickness gold coat (ca. 20 nm This decrease may have been avoided by using a lengthy burn-in
thick) to inhibit build up of charge on the sample during time or by using a higher primary beam current. As the magnesium
analysis. decrease is of the same order of magnitude in all analyses, compari-
sons of magnesium concentration can be made between analyses
which are performed for the same time period.
2.2. Ion Microprobe Analysis
No information on the trace element composition of the epoxy
Analyses were performed using a Cameca ims-4f ion microprobe resin was available from the manufacturer. In an attempt to gauge
with Charles Evans and Associates interface and computer control. the relative composition of the resin, two analyses were performed
Analysis was by Secondary Ion Mass Spectrometry (SIMS) in which approximately 50 pm apart on a thin section, one on coral material
the specimen is bombarded with a focused primary ion beam and and the other on epoxy resin (Table 1) . This table does not necessar-
the sputtered secondary ions produced are accelerated into a mass ily reflect the trace metal composition of the two media, as isotope
spectrometer. As the elements studied here most readily form posi- counts may include any signal-produced by molecular interferences.
tive ions, a primary beam of 160- ions was used to enhance second- Counts of Z6Mp.“Ca, “Rb. *‘Sr, “*Ba, and 23*U(or their isobars)
ary ion formation. This primary beam was mass filtered to leave in the epoxy a&lysis were a fraction of the values observed in the
an isotopically pure l6O- beam and avoid implantation of other coral analysis and it is unlikely that epoxy resin contamination will
contaminants in the sample. The primary beam was accelerated by increase the concentrations of these isotopes. Count rates of 27A1,
a primary voltage of 10 kV and a voltage of +4.5 kV was applied z8Si154Fe155MnI and *“‘Pb in the epoxy resin are similar to or greater
to the sample giving a total primary accelerating voltage of 14.5 than the counts observed in the coral and it is possible that contami-
kV. The primary beam current of ca. 5.0 nA was focused using a nation of coral analyses will significantly alter the concentrations of
series of electrostatic lenses to a diameter of about 30 pm on the these isotopes. These isotopes were not counted in subsequent analy-
sample surface. Sputtered secondary ions consist of a complex mix- ses and no attempt was made to assess their potential interferences.
ture of ionic and molecular species derived from the sample matrix,
the primary beam, and any gases present in the background of the 2.2.3. Interfering species
mass spectrometer, e.g., M’ (the desired signal), MO+, MOH’ ,
M2”, etc. Samples were maintained at high vacuum (between 2.10-a Interfering species were minimised using the energy filtering tech-
and 5.10-’ Torr) to reduce any interferences caused by residual nique @inner and Crozaz, 1986) at a voltage offset of -75 V and
 Comparative determinations of trace and minor elements in coral aragonite 3459

Table 1. Isotope counts per second on where

adjacent coral and epoxy resin analyses. X, = concentration of element i in the carbonate
Xc, = concentration of calcium in the carbonate
CPS, = counts per second for element i (corrected for deadtime,
Isotope Coral Epoxy background, interference and isotopic abundance)
Analysis Resin CPSc. = counts per second for calcium (corrected for deadtime,
background, interference and isotopic abundance)
“‘Mg 461 22 RIY, = ion yield of element i relative to that of calcium in standard.
2’Ai 373 648
The ion probe was calibrated for magnesium and strontium using
2*Si 110 128 a natural carbonate standard, Norman Cross Calcite (NCC), supplied
44Ca 77300 11 by A.E. Dickson, University of Cambridge. This standard has been
analysed by a variety of methods and workers and averages of several
“*Fe 1.05 0.76 of these values were used to calibrate the ion probe (2749 ppm Mg;
““Mn 1.48 1.14 40.08% Ca; 845 ppm Sr). The standard was used to determine
85Rb 0.42 0.07 relative ion yields (RIY, i.e., the ion yield of the element relative
to that of calcium in the standard) for these elements at the start of
@%r 26700 2 each of the four probe sessions (Table 3). The concentration of
‘38Ba 5.63 0.36 strontium in the standard is an order of magnitude less than that
observed in the samples; however, as RIY is independent of concen-
*“Pb 0.14 0.16
238 tration, extrapolating between the sample and standard should not
U 0.39 0.00 affect analytical accuracy. Multiple analyses were performed on the
standard to test its homogeneity for magnesium and strontium (Ap-
pendix A). The standard deviation between replicate analyses for
magnesium (6.0%) is in reasonable agreement with the analytical
anenergy window of 40 V. Under these conditions, trimers (molecu- error of each analysis calculated from counting statistics (5.4-
lar ions comprised of three atoms) are negligible (Schauer and Wil- 6.8%), suggesting that the standard is homogenous within analytical
liams, 1990). The major atomic and dimer interferences for the error for magnesium. However, the standard was heterogeneous with
isotopes studied are shown in Table 2. 26Mgis subject to few interfer- respect to strontium (the standard deviation of multiple analyses is
ences and Mason ( 1987) concluded that these amounted to an appar- 10.3% compared to an analytical error of 1.O- 1.6%). This heteroge-
ent signal equivalent to about 1 ppm magnesium at little or even no neity makes it impossible to ensure an accurate calibration of the
energy filtering. The conditions used in this study are sufficient to ion probe in any one session or to obtain comparable calibrations of
eliminate interferences for ‘*Sr (Swart, 1990). the instrument between sessions. To assess the extent of calibration
To assess possible interferences on barium, two isotopes ( 13’Ba differences between probe sessions, immediately adjacent strontium
and 13’Ba) were counted for several analyses. Assuming negligible analyses were performed on the same coral samples in different
molecular interferences, count rates for each isotope should occur probe sessions: i.e., some of the coral samples were themselves
in their relative isotopic proportions. Count rates did occur in the used as standards. Calculated element concentrations were averaged
expected proportions, within counting errors (2-7%), indicating between the analyses for each probe session and are shown in a
no significant interferences for these isotopes. The more abundant footnote to Appendix C. Variations in the calibration between ses-
isotope, 13’Ba,was counted for the remaining analyses. sions introduced an error of up to 0.9 mm01 Sr mol-’ Ca (ca. 10%).
It was not possible to check for molecular interferences on rubid- Correction factors were calculated which could be multiplied to each
ium or uranium. Although two isotopes of Rb exist, only “Rb was session of probe data to make results comparable between sessions.
counted as 87Sr produces a significant interference with *‘Rb. The All data were corrected relative to the March 1992 dataset.
largest expected dimer interference on 85Rb is ?a”Ca. Under the Carbonate standards were unavailable for calibration of the other
current operating conditions this species produces approximately elements due to the generally low concentrations of trace elements
0.008 counts s-’ , calculated from @Ca count rates and calcium isoto- in these minerals. Instead the ion probe was calibrated using the
pic abundances. This will not significantly interfere with “Rb counts standard silicate glass, NBS610 (nominal concentrations: 500 ppm
which are ca. 0.42 counts s-‘. Only one isotope of U (*“U) was Mg; 8.36% Ca; 512 ppm Sr; 426 ppm Rb; 500 ppm Ba; 461 ppm U).
sufficiently abundant for detection by the ion microprobe. No major RIYs were calculated from this standard several times throughout the
molecular interferences were expected for this isotope. fourteen month study period but essentially remained unchanged
In summary, “?d “*Sn ‘39La ‘@‘Ce,and 232Thin the corals were (Table 3). RIY s may differ between carbonates and silicates due to
frequently below the’limit of detection and “Al, “Si “Fe, s5Mn, and different matrix effects between the two media. However, barium
‘OsPbwere rejected from further analysis due to possible contamina- values from fragments of two corals analysed by atomic absorption
tion problems. The following isotopes appeared suitable for analysis, (3 and 10 ppm, A. Mennim, pers. commun., 1993) were in reason-
under the current operating conditions: 26Mg,?a, 85Rb,88Sr, “*Ba,
and 23sU.As it was not possible to unequivocally confirm that “Rb
and z38Uwere not subject to molecular interferences, data for these
isotopes should be viewed with some caution, TABLE 2. Isotopic abundances of analysed elements and their
major element interferences (all oxygen isotopes assumed to be
‘60).
2.2.4. Calibration
Isotope Abund.(%) Possible interferences
The sputtering and ionisation processes during analysis are poorly
understood and no physical model exists to adequately predict ion LbMg 11.29 ‘Y’%+
yields. To compensate for the effects of variations in analytical 44Ca 2.13 88Sr2’
conditions, e.g., beam density, beam current, and spot size between s5Rb 72.15 42Ca43Ca+'84,&'H+
analyses, trace element count rates are normalised to a reference
species, in this case ?Za. Concentrations are then calculated relative “Rb 27.85 *‘Sr+, 43Ca44Ca+
to a standard of known composition. The normal&d ratio, [“ill **Sr 82.56 44Ca44Ca+
[*Cal, is converted into a concentration using the following for-
mula: 13’Ba 11.32
13BLa+. 13SCe+
13’Ba 71.66
2X - CPS, x 1 238
U 99.28
(1)
XC, CPSC, RIY, ’
3460 N. Allison

Table 3. Relative ion yields (relative to Ca) for Mg and Sr calculated from the
carbonate standard, NCC, for each ion microprobe session and for all elements
calculated from the silicate standard NBS 610.

NCC 1 NBS 610

Element Dee 1991 Mar 1992 May 1992 Jan 1993

Mg 0.550 0.607 0.599 0.518 0.951

Ca 1.000 1.000 1.000 1.000 1.000
Sr 0.513 0.443 0.487 0.498 0.407
Rb _ 0.141
Ba 0.183
U _ 0.0692

able agreement with values obtained from single ion microprobe that local seawater chemistry or temperature will significantly alter
analysis of adjacent coral material (6 and 11 ppm, respectively, over this time. The calculated element concentrations for each pair
Allison, 1994). Although it was not possible to accurately calibrate of analyses are shown in Appendix B. The reproducibility of each
the ion probe using the silicate standard, detectable variations be- pair was calculated as a concentration (half the concentration differ-
tween samples proportionally mirror trace element variations and the ence between the two values) and as a percentage (the concentration
instrument can still be used to investigate geochemical heterogeneity. reproducibility as a percentage of the lowest concentration of the
pair). The reproducibility ranges and averages for each element are
2.2.5. Analytical reproducibility shown in Table 4. The reproducibility error obtained for magnesium
(5%) is in good agreement with the standard deviation observed
The ion probe unit’s computer system calculates an analytical during multiple analyses on the carbonate standard, NCC, which
error based purely on counting statistics which is calculated as fol- appeared homogeneous with respect to magnesium. Reproducibility
lows: errors for the remaining elements are usually well in excess of count-
ing statistic errors. It is not possible to conclude whether this reflects
% error (1~7) = 7 X 100, real small scale geochemical heterogeneity or instrumental varia-
(2)
tions. As such, the reproducibility errors calculated between analyti-
cal pairs are probably an overestimate of analytical errors.
where N = total number of counts for the isotope over the analytical
period. The ranges of counting statistic errors observed for each
element are shown in Table 4. As a function of the total number of 2.3. Trace and Minor Element Concentrations in Corals
counts, these errors take no account of differences in beam focusing
or fluctuations in beam current between analyses and essentially Analyses were performed vertically (with time) across each thin
represent the smallest possible errors which can be expected. To section to produce a chronological record of trace and minor element
test analytical reproducibility under experimental conditions it is variations. Analyses were performed at spacings of approximately
necessary to perform multiple analyses on a homogeneous carbonate 500-1000 pm. The relative position of each sample in the annual
material. The carbonate standard is homogeneous with respect to bright-dull fluorescent couplet was used to estimate the approximate
magnesium but heterogeneous for strontium and does not contain deposition time of the carbonate sampled in each analysis. Corals
rubidium, barium or uranium in sufficiently high concentrations for were analysed over the three most recently deposited fluorescent
analysis. The next best way to assess reproducibility was to perform couplets, i.e., approximately November 1988-July 1991.
immediately adjacent analyses on the coral samples and to assume
that the coral was homogeneous over this distance. The corals in
3. RESULTS
the Phuket area have typical linear extension rates of 15-20 mm
y -’ corresponding to an average growth of 40-55 pm day-‘. Al-
Although considerable time was dedicated to positioning
though corals do not extend linearly (they produce radiating fans of
aragonitic needles), the distance between adjacent analyses probably the primary beam on the sample, many of the analyses (about
accounts for no more than two or three days growth. It is unlikely 10%) were subsequently excluded from the dataset as they

Table 4. The range of errors for analysis of each isotopef4Ca ratio calculated
from counting statistics (single analyses) and by comparing analyses between
adjacent pairs (n = number of pairs).

Isotope Range of Reproducibility range Mean reproducibility n

Counting Stat. between pairs of between n pairs (%)
Error (%) analyses (“A)
26Mg/Ca 0.3 - 1.2 o-14 5 27
s”Rb/Ca 11-22 o-37 13 15
88SrlCa 0.1 - 0.2 o-4 1 27
’j'Ba/Ca 2-7 o-14 5 27
238UICa 6 - 16 1 - 12 4 5
 Comparative determinations of trace and minor elements in coral aragonite 3461

Table 5. Ranges of trace and minor element concentrations determined by ion

microprobe analysis (mmol mof’ Ca for Sr and Mg; pmol mof’ Ca for all other
elements). Approximate concentrations of each element in ppm (assuming Ca =
38 wt%) are also shown.

Observed Approximate Concentrations

range of range reported in the
concentrations (ppm) literature (ppm)
3.25 - 6.23 748 - 1435 1000’ - 21002
0.06 - 0.81 0.05 - 0.66 o.873
7.42 - 9.07 6160 - 7529 69004 - 78005
1.51 - 8.61 2.0 - 11.2 56 - 15’
0.00 - 0.40 0.0 - 0.90 1.5 - 3.08

’Howard and Brown 1987, Pocillopora spp.

2 Goreau 1977, Montastrea spp.
3 Sreekumaran and Gogate 1972, Porites spp.
’ Weber 1973, DipIoria spp.
’ Amiel et al. 1973, Diploria spp.
6 Shen et al. 1992, Pavona claws
’ Pingitore et al. 1989, Acropora palmata
* reviewed in Swart and Hubbard 1982

were contaminated by microborings and centres of calcifica- interval, i.e., from approximately November 1990 to July
tion. These areas are best visualised in transmitted light and 1991 (Fig. 4).
were difficult to distinguish using the reflected light micro- A comparison of the ranges of element concentrations
scope attached to the ion microprobe. Microborings and cen- deposited by different coral colonies was not the aim of this
tres of calcification can be associated with elevated concen- study. However, it is apparent that corals from the same reef
trations of barium and magnesium (Allison and Tudhope, site deposit substantially different concentrations of some
1992). All uncontaminated analyses obtained over the four elements. For example, coral AQ2 deposits a much higher
ion probe sessions are shown in Appendix C. and wider range of barium concentrations (3.14-8.61 pmol
The ranges of concentrations of each element are shown mol -’ Ca) than observed in coral AQSi ( 1Sl -3.16 pmol
in Table 5. The approximate concentrations of all elements mol-’ Ca) or AQ8ii (1.87-4.27 pmol mol-’ Ca). Coral
in ppm have been calculated to allow easier comparisons KR6 deposits a higher range of magnesium concentrations
with data produced by previous authors (also shown). Con- (4.31-6.32 mm01 mol-’ Ca) than coral KR4 (3.25-5.15
centrations in ppm have been calculated assuming a nominal pm01 mol-’ Ca).
calcium concentration of 38%. In pure CaC03, calcium com-
prises 40.08% of the compound mass. Coral aragonite con- 4. DISCUSSION
tains less calcium due to the inclusion of several contaminant
ions, primarily sodium (ca. 10000 ppm), strontium (ca. 7000 Element concentrations for magnesium, rubidium, stron-
ppm), and magnesium (ca. 1000 ppm) . tium, and barium are in reasonable agreement with values
Seasonal variations of magnesium, rubidium, strontium, previously reported in the literature (see Table 5). The mag-
barium, and uranium in each coral are illustrated graphically nesium concentrations determined here cover a slightly
(Figs. l-5). Element concentrations are plotted against the lower range than has previously been reported. However, as
relative position of each analysis in the coral, i.e., the approx- this study avoided the analysis of adsorbed ions, which may
imate deposition date. In general, seasonal variations in ele- account for up to 20% of skeletal magnesium (Amiel et al.,
ment concentrations were not apparent and for most corals 1973), a lower range of concentrations was expected.
the data points appear randomly scattered across each graph. The uranium values obtained here are considerably lower
However, variations in concentration between points far ex- than those previously reported. It is most likely that this
ceed analytical errors and reflect real geochemical variations. represents a m&calibration of the ion microprobe due to
Large variations in element concentration, which approach matrix differences affecting RIYs between the silicate stan-
the total range of values recorded, are frequently observed dard and the aragonite samples. Although it is not possible
over relatively small distances corresponding to less than to estimate uranium concentrations using the ion microprobe,
one month’s growth. In addition, almost all the corals exhibit differences between samples reflect true variations in geo-
an increase in barium concentration in the outermost growth chemistry.
3462 N. Allison

Probe Session: A Dee 1991 March 1992 * May 1992 ?? January 1993

Analytical error = +/- 5%

T 7

5.4 AQ8a
4.8
4.2
3.6 A

5.1 AQ8b
4.5
3.9
3.3
3 5.1 AQ2 A
U
4.5 * A

3.9
3.3
6.0 PBl
5.4
4.8
6.4 PB9
.
5.8
5.2
.
4.6 .

4.0
5.1 KR4
4.5
3.9
3.3
6.8
KR6 . .
6.2
5.6
5.0
4.4

Nov 1988 Nov 1989 Nov 1990

Approximate deposition dates determined from the dull-bright fluorescent banding

FIG. I. Magnesium chronologies in corals from Porites Bay (PB),Aquarium Reef ( AQ) , and Ko Racha (KR)

4.1. Spatial Variations in Trace and Minor Element temperature range (2-4°C) recorded in the skeletal stron-
Concentrations tium concentration. According to Beck et al. (1992) this
temperature variation should correlate with an annual skele-
Significant variations in element concentration were ob- tal strontium range of ca. 0.2-0.3 mmol mol-’ Ca (186-
served between the chronological analyses of all the ele- 248 ppm) . Strontium concentrations in this study vary by as
ments studied here, i.e., magnesium, rubidium, strontium, much as 0.6 mm01 mall’ Ca (approximately 480 ppm) over
barium, and uranium. Large variations in element concentra- relatively small skeletal distances. It seems unlikely that this
tion, which approach the total range of values recorded, are small scale spatial geochemical heterogeneity of strontium
frequently observed over relatively small distances corre- and other elements reflects short-term variations in seawater
sponding to less than one month’s growth. This small scale chemistry and/or temperature. While it is possible that
geochemical heterogeneity has not been recorded previously. warming and evaporation of the shallow water over the reef
No seasonal variations in geochemistry were detected. In flat may lead to some small variations in local water chemis-
particular, there is no evidence of the local annual seawater try and temperature on a daily or weekly basis, I detected
 Comparative determinations of trace and minor elements in coral aragonite 3463

Probe Session: A Dee 1991 ?? March 1992 * May 1992 ?? January 1993

Analytical error = +/- 13%

0.3 AQ8a

0.2

0.6 AQ8b
0.5 ..
.
0.4 00
.
0.3 .
. .
0.2

0.1

AQ2

PBl

0.8 KR6
.
0.7

0.6
.
0.5 .
0.4 .

. .
0.3

0.2 . A A

Nov 1988 Nov 1989 Nov 1990

Approximate deposition dates determined from the dull-bright fluorescent bandin

FIG. 2. Rubidium chronologies in corals from Porites Bay (PB), Aquarium Reef (AQ), and Ko Racha (KR)

no significant variations in salinity or temperature along the ever, several possible explanations can be discussed. The
reef front at Porites Bay during the day (N. Allison, unpubl. incorporation of some elements in the aragonite may be
data) and it seems unlikely that any significant variations in affected by calcification rate (de Villiers et al., 1994) or
seawater composition occurred. In addition, daily variations the trace element composition of the dissolved inorganic
in seawater chemistry or temperature would have been re- carbonate pool available for calcification may be modified
corded in the skeleton over very small distances and should by the coral symbiotic zooxanthellae (Buddemeier et al.,
have been included in the estimates of analytical error, ob- 1981). However, it is likely that the largest variations in
tained by comparing pairs of analyses up to 50 pm apart. If calcification rate occur on a daily basis (day vs. night) and
the strontium vs. temperature relationship does occur in the geochemical heterogeneity attributed to this would have been
coral skeletons of this area, then it is effectively swamped included in the estimates of analytical error. Weekly or
by this smaller scale geochemical heterogeneity. The cause monthly variations in calcification rate or in algal efficiency
of this heterogeneity was not discovered in this study; how- may cause some geochemical heterogeneity.
3464 N. Allison

Probe Session: A Dee 1991 March 1992 * May 1992 ?? January 199

Analytical error = +/-I %

8.8
AQ8a
8.4

8.0

AQ8b ??

8.2

8.2

8.2

7.8

PB9

8.8
c
8.4

8.0

8.6 KR6

8.2
.
7.8

7.4

Nov 1988 Nov 1989 Nov 1990

Approximate deposition dates determined from the dull-bright fluorescent bandin

FIG. 3. Strontium chronologies in corals from Porites Bay (PB), Aquarium Reef (AQ), and Ko Racha (KR)

Barnes and Lough (1993) suggested that while calcifica- The geochemical heterogeneity may relate to the metal
tion is concentrated at the surface of the coral colony, thick- binding capacity of the organic matrix produced by the coral
ening of the skeletal elements occurs throughout the tissue tissue during calcification and disseminated throughout the
layer, which occupies skeleton deposited over a period of coral skeleton. The coral tissue may contain substantially
4- 13 months. This lateral thickening may account for up to higher concentrations of certain metals (e.g., barium and
50% of the total material deposited. Material deposited in uranium, Flor and Moore, 1977) than the associated coral
this fashion may have a substantially different geochemical skeleton and contamination of the skeleton with organic ma-
signature to the originally deposited material either because terial may result in substantial geochemical heterogeneity.
of changes in seawater chemistry or temperature between Mitterer ( 1978) suggested that the organic matrix itself is
the two deposition dates or due to the reduction in calcifica- heterogeneous, containing areas of varying metal binding
tion rate. However, once again, I expected any small scale capacity.
heterogeneity relating to this phenomena to be included in The small scale skeletal geochemical heterogeneity ob-
estimates of analytical error. served here does not appear to reflect variations in seawater
 Comparative determinations of trace and minor elements in coral aragonite 3465

Probe Session: A Dee 1991 ?? March 1992 * May 1992 ?? January 1993

Average analytical error = +/- 5%

3 AQ8a AA
A AA
1 A~

A
*
1
A
AL
A

6-1
I
m
.
i
??
.
??
?? II

8 KR6
A
i A
6 . .
. ?? .
.
. A
. . . A

0 8 A

Nov 1988 Nov 1989 Nov 1990

Approximate deposition dates determined from the dull-bright fluorescent bandin

FIG. 4. Barium chronologies in corals from Porites Bay (PB), Aquarium Reef ( AQ), and Ko Racha (KR)

temperature or chemistry. It seems most probable that the be the case in this study. It is important to evaluate the extent
heterogeneity reflects the binding of metals by a heteroge- to which this small scale spatial heterogeneity affects corals
neous organic matrix which is itself distributed unevenly from other sites and to determine how it may influence the
throughout the skeleton. geochemical signature of coral skeletons.
Other authors have successfully reported excellent correla-
tions between coral geochemistry and seawater characteris- 4.2. The Barium Increase in the Outermost Part of the
tics. Either this heterogeneity did not occur in the corals they Skeletons
analysed or the geochemical variations were averaged out
in the analysis of the larger sample sizes they used. Research- Almost all the corals exhibit an increase in barium concen-
ers in the field of coral palaeoclimatology generally assume tration in the outermost growth interval although no other
that the geochemical signature in the skeleton is directly seasonal variations are apparent. This may be a reflection of
related to seawater characteristics. This does not appear to the coral bleaching event which occurred in the Phuket area
3466 N. Allison

Probe Session: . January 1993

Analytical error = +I- 4%

0.4

0.3

0.2

0.1

0.0 .

00
PB9
0.2 .
.
.
.
0.1

0.2

0.1
t
Nov 1988 Nov 1989 Nov 1990

Approximate deposition date determined from the dull-bright fluorescent banding.

FIG. 5. Uranium chronologies in corals from Porites Bay (PB), Aquarium Reef (AQ), and Ko Racha (KR).

from early June 1991 onwards during a period of anoma- that these differences are due to variations in seawater chem-
lously high seawater temperatures. However, none of the istry or temperature. The geochemistry of the skeletons may
corals used in this study appeared visibly affected by the be influenced by the factors discussed above, i.e., variations
bleaching event and there is no evidence of a significant in calcification rate between colonies, in concentrations of
reduction in skeletogenesis over the outermost growth inter- zooxanthellae or in quantity and composit?on of organics in
val (Allison, 1994). There is no evidence that local environ- the skeleton. Alternatively the different element ranges may
mental factors which may increase seawater barium concen- reflect genetic differences between coral colonies which af-
trations, eg., freshwater runoff or coastal upwelling, were fect trace and minor element distribution coefficients. All
more pronounced than compared to previous years (Allison, corals exhibited skeletal characteristics of Porites lutea
1994; T. K&be, pers. commun., 1991; respectively). (Veron, 1986) ; however, considerable variations in corallite
Similar increases in skeletal barium have been observed morphology were apparent between colonies.
by S. Tudhope (pers. commun. 1993, coral from Oman) and The cause and extent of this geochemical anomaly be-
Hart and Cohen (1996, coral from South Africa) in corals tween colonies should be further investigated. Distribution
from areas unaffected by coral bleaching events. It seems coefficients, reflecting the extent to which an element in
most probable that this barium is not skeletally bound but seawater is incorporated into the skeleton, have previously
is associated with some organic material which may age and been calculated for several coral species and for several
be removed from the skeleton in time. elements. In this study, calculations of barium distribution
coefficients will vary by as much as 2-3 between corals
4.3. Differences in Element Concentration between AQ2 and AQ8 although the corals appeared to be the same
Coral Colonies species. Other researchers estimates of distribution coeffi-
cients may have incorporated this degree of error.
It became apparent during the study Chatindividual corals,
of apparently the same species, deposit substantially differ- 5. SUMMARY
ent ranges of magnesium and barium concentrations both
within and between reef sites. Variations within each site 1) The ion microprobe technique has been developed to
are as large as those observed between sites and it is unlikely produce comparative determinations of certain elements
 Comparative determinations of trace and minor elements in coral aragonite 3461

in coral aragonite. Magnesium, rubidium, strontium, and T. P. ( 1991) A novel mechanism for iron incorporation into coral
skeletons. Coral Reefs 10, 21 l-21 5.
barium concentrations were in reasonable agreement with
Buddemeier R. W., Schneider R. C., and Smith S. V. (1981) The
values previously reported in the literature. Uranium con- alkaline earth chemistry of corals. Proc. 4th Int. Coral Reef Symp.
centrations were considerably lower than previous esti- 2, 81-85.
mates probably reflecting a miscalibration of the ion mi- de Villiers S., Shen G. T., and Nelson B. K. (1994) The Sr/Ca
croprobe. temperature relationship in coralline aragonite: Influence of vari-
ability in SrlCa (seawater) and skeletal growth parameters. Geo-
2) The coral skeletons examined here are heterogeneous chim. Cosmochim. Acta 58, 197-288.
with respect to magnesium, rubidium, strontium, barium, Flor T. H. and Moore W. S. ( 1977) Radium/calcium and uranium/
and uranium on a very small scale. The whole concentra- calcium determinations for western Atlantic reef corals. Proc. 3rd
tion range of an element is frequently observed over Intl. Coral Reef Symp. 2, 555-561.
Goreau T. J. (1977) Coral skeletal chemistry: physiological and
relatively short skeletal distances corresponding to less environmental regulation of stable isotopes and trace metals in
than one months growth. Perhaps the most plausible ex- Montastrea annularis. Proc. Roy. Sot. London B 196, 291-315.
planation for this heterogeneity is a heterogeneous metal- Hart S. R. and Cohen A. L. (1996) SrlCa in corals: An ionprobe
binding organic matrix which is unevenly dispersed study of annual cycles and microscale coherence with other trace
elements. Geochim. Cosmochim. Acta 60, 3075-3084.
throughout the skeleton. Howard L. S. and Brown B. E. (1987) Metals in Pocillopora dami-
3) Seasonal variations in element concentrations were not cornis exposed to tin smelter effluent. Mar. Pollut. Bull. l&451-
apparent in any of the corals except for a general increase 454.
in barium in the outermost growth interval of the skele- Lea D. W., Shen G. T., and Boyle E. A. ( 1989) Coralline barium
records temporal variability in equatorial Pacific upwelling. Na-
ton. This may reflect an increase in skeletal organic mate-
ture 340, 373-376.
rial or may be an anomaly caused by the coral bleaching Mason R. A. ( 1987) Ion microprobe analysis of trace elements in
event in progress at the time. calcite with an application to the cathodluminescence zonation of
4) Individual colonies deposited substantially different limestone cements from the lower carboniferous of South Wales,
U.K. Chem. Geol. 64, 209-244.
ranges of magnesium and barium both within and be-
Mitterer R. M. (1978) Amino acid composition and metal binding
tween sites. The variation in these ranges could result capability of the skeletal protein of corals. Bull. Mar. Sci. 28,
in significantly different distribution coefficients being 173-180.
calculated between colonies. Pingitore N. E., Range1 Y., and Kwarteng A. (1989) Barium varia-
tion in Acropora palmata and Montastrea annularis. Coral Reefs
8, 31-36.
Acknowledgments-I wish to thank John Craven for technical sup- Schauer S. N. and Williams P. ( 1990) Elimination of cluster interfer-
port with the ion microprobe, Sandy T&hope for assistance with ences in secondary ion mass spectrometry using extreme energy
field work and Jane Foster and Mike Hall for preparation of thin filtering. Intl. .I. Mass Spec. Ion Processes 103, 21-29.
sections. I thank the director and staff at the Phuket Marine Biologi- Schneider R. C. and Smith S. V. (1982) Skeletal strontium content
cal Center for the use of their facilities. Both Sandy Tudhope and and density in Porites spp. in relation to environmental factors.
John Craven provided valuable comments on this manuscript. Ed Mar. Biol. 66, 121-131.
Boyle, Graham Layne, Stan Hart, and Rick Hervig reviewed the Scoffin T. P., Tudhope A. W., Brown B. E., Chansang H., and
manuscript and are thanked for their constructive criticisms. This Cheeney R. F. ( 1992) Environmental controls on the skeletogen-
work was undertaken when the author was in receipt of a NERC esis of Porites lutea, South Thailand. Coral Reefs 11, l- 11.
studentship (GT4/90/G3/47). Shen G. T. and Sanford C. L. ( 1990) Trace element indicators of
climate variability in reef-building corals. In Global Ecological
Consequences of the 1982-83 El Nitio (ed. P. W. Glynn), pp
Editorial handling: M. A. McKibben 255-283. Elsevier.
Shen G. T., Cole J. E., Lea D. W., Linn L. J., Mcconnaughey
T. A., and Fairbanks R. G. (1992) Surface ocean variability at
REFERENCES Galapagos from 1936- 1982: Calibration of geochemical tracers
in corals. Paleoceanography 7, 563-588.
Allison N. ( 1994) The environmental geochemistry of Porites lutea Shimizu N., Semet M. P., and Allegre C. J. (1978) Geochemical
coral skeletons from Phuket, South Thailand. Ph.D. dissertation, applications of quantitative ion microprobe analysis. Geochim.
Edinburgh Univ. Cosmochim. Actu, 42, 1321-1334.
Allison N. and Tudhope A. W. (1992) Nature and significance of Sreekumaran C. and Gogate S. S. (1972) Mineral constituents of
geochemical variations in coral skeletons as determined by ion some species of corals. Curr. Sci. 41, 241-244.
microprobe analysis. Proc. 7th Intl. Coral Reef Symp. 1, 173- Swart P. K. (1990) Calibration of the ion microprobe for the quanti-
178. tative determination of strontium, iron, manganese and magnesium
Amiel A. J., Friedman G. M., and Miller D. S. (1973) Distribution in carbonate minerals. Anal. Chem. 62, 722-728.
and nature of incorporation of trace elements in modem aragonitic Swart P. K. and Hubbard J. A. E. B. ( 1982) Uranium in scleractinian
corals. Sedimentology 20, 47-64. coral skeletons. Coral Reefs 1, 13- 19.
Barnes D. J. and Lough J. M. (1993) On the nature and causes of Veizer J., Hinton R. W., Clayton R. N., and Lerman A. (1987)
density banding in massive coral skeletons. J. Exp. Mar. Biol. Chemical diagenesis of carbonates in thin sections: ion microprobe
Ecol. 167, 91-108. as a trace element tool. Chem. Geol. 64, 225-237.
Beck J. W., Edwards R. L., Ito E., Taylor F. W., Recy J., Rougerie Veron J. E. N. (1986) Corals of Australia and the Indo-Pacific.
F., Joannot P., and Henin C. ( 1992) Sea-surface temperature from Univ. Hawaii.
coral skeletal strontium/calcium ratios. Science 257, 644-647. Weber J. N. (1973) Incorportaion of Sr into reef coral skeletal
Brown B. E., le Tissier M. D. A., Howard L. S., Charuchinda M., carbonate. Geochim. Cosmochim. Acta 37, 2173-2190.
and Jackson J. A. ( 1986) Asynchronous deposition of dense skele- Zinner E. and Crozaz G. (1986) A method for the quantitative
tal bands in Porites lutea. Mar. Biol. 93, 83-90. analysis of rare earth elements in the ion microprobe. Intl.
J.Mass
Brown B. E., Tudhope A. W., le Tissier M. D. A., and Scoffin Spec. Ion Processes 69, 17-38.
3468 N. Allison

Appendix A. Multiple analyses of Sr and Mg on

the carbonate standard, NCC. % errors calculated
from counting statistics are included. Averages and
standard deviations of the 19 analyses are shown.

Analysis Mg Sr
No. ppm % ppm %
error error
1 2580 6.0 830 1.3
2 2850 6.6 820 1.3
3 2820 6.5 890 1.4
4 2870 6.6 640 1.0
5 2410 5.6 1010 1.6
6 2820 6.5 790 1.2
7 2750 6.4 860 1.3
8 2930 6.8 670 1.0
9 2850 6.6 700 1.1
10 2870 6.6 800 1.2
11 2700 6.2 840 1.3
12 2320 5.4 850 1.3
13 2640 6.1 890 1.4
14 2670 6.2 840 1.3
15 2520 5.8 750 1.1
16 2600 6.0 870 1.3
17 2790 6.4 800 1.2
18 2820 6.5 880 1.4
19 2790 6.4 830 1.3
Average 2716 819
S.D. (la) 162 83
S.D. (%) 6.0 1.0

Appendix B. Trace and minor element concentrations on adjacent pairs of analyses up to 50 pm apart.

Mg (mmol Rb (pmol mol-’ Ca: Sr (mm01 mokzoar) Ba (pmol mol-’ Ca) U (pm01 mol-’ Ca)
Analyses Analyses Error Aualyses Analyses Error Analyses Error

4.77 5.44 7 0.25 0.27 5 8.10 8.21 4.00 4.25 3 0.08 0.10 12
5.40 4.77 6 0.20 0.20 0 8.04 8.14 4.13 4.38 3 0.16 0.15 1
4.77 5.07 3 0.20 0.27 15 8.04 8.21 4.00 4.38 5 0.20 0.19 1
5.07 5.44 4 0.20 0.25 11 8.04 8.10 4.25 4.38 1 0.16 0.18 5
4.98 5.23 2 0.20 0.25 11 8.10 8.14 4.13 4.25 2 0.21 0.21 1
6.28 5.57 6 0.30 0.33 6 8.16 8.21 7.27 7.38 1
5.15 6.03 8 0.30 0.35 7 7.96 8.26 5.87 5.92 0
5.15 4.94 2 0.10 0.15 24 8.25 8.26 6.14 6.46 3
4.90 5.11 2 0.07 0.15 37 8.05 8.25 6.34 6.14 2
5.15 4.56 6 0.29 0.33 8 8.27 8.39 1.68 1.82 4
5.28 4.94 3 0.31 0.37 9 8.49 8.56 2.16 2.31 3
5.78 4.65 11 0.27 0.35 12 8.54 8.73 1.27 1.34 3
5.65 4.77 8 0.11 0.19 28 8.50 8.63 0.94 1.21 12
4.90 5.86 9 0.17 0.20 10 8.28 8.39 1.25 1.26 0
4.77 4.61 2 0.24 0.27 7 8.34 8.35 1.78 1.92 4
3.73 3.42 4 8.23 8.39 2.21 2.36 3
4.08 4.23 2 8.52 8.52 5.71 6.30 5
5.32 5.44 1 8.35 8.63 6.63 8.24 11
4.82 4.27 6 8.60 8.68 2.42 3.02 11
3.87 3.52 5 8.69 8.75 3.31 4.39 14
3.72 4.07 4 8.55 8.71 3.42 3.90 6
4.16 4.19 0 8.32 8.54 4.31 4.90 6
4.94 4.61 4 8.23 8.34 2.45 2.95 9
6.32 4.73 14 7.23 7.45 2.59 2.83 5
4.52 5.07 6 6.87 7.46 2.56 2.62 1
3.61 3.98 5 7.74 7.77 2.56 2.56 0
7.38 7.75 5.61 5.87 2

The reproducibility between each pair has been calculated as a % error (the concentration difference between the two analyses as a
percentage of the average concentration). Comparisons are only made between analyses from the same microprobe sessions. No
corrections to Sr data have been made.
 Comparative determinations of trace and minor elements in coral aragonite 3469

Appendix C. Seasonal trace and minor element variations expressed as mm01 mol-’ Ca (Sr and Mg) and pmol mol.-’ Ca (Rb, Ba, U).
Corr. Corr.
Sample RYP Mg Rb Sr Sr Ba U Sample RYP Mg Rb Sr Sr Ba U
December1991I March 1992
AQ8iA 3.63 5.07 0.30 7.42 8.31 2.94 PBlA 3.37 - 0.20 8.14 - 4.14
3.58 3.58 0.21 7.73 8.65 3.01 PBlB 3.59 4.86 0.43 8.19 - 5.99
3.42 1.23 0.18 7.41 8.30 2.12 3.50 4.94 0.28 8.05 - 5.61
3.11 1.44 7.41 8.30 2.29 3.50 5.74 0.17 7.98 - 5.06
2.90 4.48 7.21 8.08 2.40 3.53 4.98 0.16 8.16 - 7.39
3.20 4.90 7.25 8.12 1.71 3.58 4.82 0.23 7.96 - 5.88
AQ2A 3.73 3.61 0.25 7.75 8.68 5.88 3.58 5.57 - 8.26 - 5.93
3.66 4.82 0.16 7.53 8.43 8.09 3.53 5.23 - 8.21 - 7.28
3.12 3.98 7.38 8.26 5.62 3.51 5.15 - 8.25 - 6.15
3.68 3.25 7.63 8.55 8.61 3.52 6.03 - 8.26 - 6.47
3.52 4.73 7.64 8.56 6.82 KR4B 2.65 4.04 - 8.39 - 2.53
3.55 1.65 7.49 8.39 6.83 2.29 4.16 - 8.34 - 2.96
KR4A 3.51 4.01 7.54 8.44 2.97 2.31 4.19 - 8.23 - 2.45
3.61 4.69 7.45 8.34 7.72 2.26 3.72 - 8.54 - 4.31
3.59 3.82 7.34 8.22 6.23 2.25 4.07 - 8.32 - 4.91
3.56 4.15 7.48 8.38 4.47 2.38 3.87 - 8.55 - 3.90
KR6A 3.58 4.98 7.46 8.35 6.32 2.37 3.52 - 8.71 - 3.43
2.32 5.78 7.60 8.51 7.16 KR4B 2.63 4.73 - 8.42 - 3.47
2.18 4.56 7.48 8.38 4.40 KR4A 3.60 5.15 - 8.27 - 7.28
2.93 4.44 7.41 8.30 2.40 3.31 3.95 - 8.54 - 4.59
3.10 4.40 1.54 8.44 3.83 3.13 4.69 - 8.35 - 3.82
KR4A 2.33 3.82 1.42 8.31 3.97 PBlC 3.17 5.07 - 8.16 - 6.26
2.11 4.3s 7.60 8.51 3.70 3.29 5.11 - 8.63 - 6.26
AQ2A 3.44 4.40 7.54 8.44 6.67 3.30 5.69 - 8.31 - 5.16
3.30 4.27 7.35 8.24 5.58 May 1992
3.07 3.90 7.53 8.43 3.80 AQ8iiB 2.62 3.50 0.17 8.93 - 2.59
AQ2A 3.12 4.07 7.61 8.52 4.30 2.64 4.56 0.24 8.47 - 2.90
2.94 4.40 7.31 8.18 3.14 2.54 4.77 0.28 8.68 - 3.53
2.60 4.86 7.68 8.60 4.28 2.49 3.80 0.42 8.99 - 3.70
2.79 4.44 7.40 8.29 3.63 2.38 4.98 0.17 8.79 - 3.19
2.77 4.48 7.45 8.34 3.62 2.37 3.80 0.26 9.07 - 3.12
AQ8iB 2.66 4.61 7.57 8.48 2.90 2.34 4.82 0.07 8.18 - 2.79
2.59 4.31 1.47 8.31 2.72 2.23 4.40 0.06 8.99 - 2.78
2.55 4.94 7.31 8.25 3.16 2.17 4.07 0.10 8.76 - 3.22
2.43 4.04 7.35 8.24 2.66 AQSiiB 2.06 4.31 0.16 8.74 - 4.27
2.39 5.23 7.25 8.12 2.60 AQ8iiA 3.64 4.61 0.42 7.80 8.42 3.35
2.36 5.15 7.55 8.46 2.91 3.47 3.48 0.26 7.76 8.38 3.10
AQ8iB 2.19 5.23 7.39 8.28 2.76 3.69 3.84 0.58 7.76 8.38 2.37
2.18 4.44 7.47 8.37 2.52 3.67 4.44 0.27 7.46 8.06 2.42
AQ8iA 2.80 4.90 7.39 8.28 2.46
January 1993
3.30 4.23 7.30 8.17 1.51 3.79 0.45 7.76 8.38 2.04 0.30
AQ8iiA 3.43
3.36 4.82 7.40 8.29 2.09
3.64 4.31 0.57 7.99 8.63 3.77 0.18
KR4A 3.02 3.69 7.48 8.38 3.03
3.58 4.40 0.39 7.81 8.43 2.54 0.30
3.53 4.94 7.16 8.02 3.12
3.19 4.19 0.07 7.73 8.34 2.42 0.00
3.44 4.40 7.46 8.35 4.93
2.99 5.07 0.27 7.77 8.39 2.57 0.10
KR4B 2.99 3.25 7.41 8.30 3.15
3.27 - 0.62 7.91 8.54 2.33 0.18
March 1992
3.49 4.52 0.48 7.75 8.37 2.69 0.20
PBlA 3.69 5.65 0.35 1.95 - 3.65
3.34 3.45 0.38 7.97 8.61 2.78 0.22
3.52 5.53 0.26 8.10 - 5.93
3.26 4.52 0.23 7.92 8.56 2.48 0.19
3.50 5.19 0.29 7.98 - 5.61
2.99 4.52 0.24 7.74 8.36 2.56 0.08
3.37 5.07 0.24 8.10 - 4.26
AQSiiC 1.99 4.69 0.32 7.96 8.59 1.87 0.15
3.37 5.40 0.27 8.21 - 4.01
1.91 3.88 0.25 7.82 8.44 2.11 0.20
3.37 4.77 0.20 8.04 - 4.38
8.14 - 6.31 1.91 4.18 0.49 7.66 8.27 2.47 0.19
3.44 6.20 0.25
3.62 6.16 0.30 7.97 - 4.10 1.86 4.23 0.41 7.98 8.62 2.58 0.18
3.70 5.65 0.32 8.12 - 4.63 1.83 4.48 0.25 7.85 8.48 2.55 0.21
PBlC 3.56 6.45 0.42 8.23 - 7.17 AQSiiC 1.87 4.73 0.49 7.76 8.38 2.20 0.16
3.54 6.03 0.27 8.05 - 5.55 1.82 4.82 0.41 7.74 8.36 2.18 0.21
3470 N. Allison

Appendix C (continued)
-I-
Corr. Corr.
Sample RYP Mg Rb Sr Sr BP U Sample RYP, Mg Rb Sr Sr Ba IJ
AQSiiC l.73 4.23 0.45 7.82 8.44 3.03 0.30 KRQA 3.19 5.28 0.21 7.69 8.31 6.11 0.14
1.66 - 0.36 8.33 9.00 2.63 0.22 2.80 5.15 0.36 7.62 8.23 3.40 0.21
PB9A 3.51 3.95 0.53 7.68 8.29 2.16 0.23
1.61 4.73 0.31 7.53 8.13 2.49 0.12
3.43 4.19 0.31 7.91 8.54 2.02 0.19
2.00 4.27 0.34 7.59 8.19 2.27 0.16
5.19 0.48 7.68 8.29 3.72 0.21 3.29 4.86 0.43 7.68 8.29 1.80 0.15
KR6C 1.76
4.82 0.31 7.39 7.98 2.52 0.23 3.08 4.61 0.34 7.69 8.31 1.81 0.12
1.51
4.61 0.41 7.31 7.89 5.40 0.12 2.89 6.03 0.38 7.75 8.37 1.02 0.17
1.93
KR6B 2.58 4.94 0.56 7.23 7.80 2.83 0.12 3.59 4.35 0.39 7.78 8.41 2.68 0.23
2.48 4.31 0.36 7.34 7.93 3.49 0.11 *g;; ;;; 3.81 0.45 7.67 8.28 3.15 0.20
2.15 6.32 0.81 6.87 7.42 2.62 0.38 4.61 - 7.45 8.04 2.59 -
KR6A 3.63 5.36 0.73 7.12 7.69 5.13 0.11 2.15 4.73 - 7.46 8.06 2.56 -
3.44 5.07 0.52 7.26 7.84 5.22 0.18 2.72 6.45 - 7.57 8.18 4.18 -
3.27 4.98 0.32 7.59 8.19 6.07 0.21

Relative yearly positions (RYP) are calculated by dividing the distance of the
sample along the fluorescent couplet by the width of the couplet. Samples prefixed
by 1 were deposited in tbe couplet approx. Nov ‘88 - Nov ‘89; by 2 in the couplet
approx. Nov ‘89 - Nov ‘90 and by 3 in the couplet approx. Nov ‘90 -Nov ‘91.

Average strontium concentrations (mmol rnol-’ Ca) on adjacent pairs analysed in

different ion microprobe sessions were used to calculate correction factors
relative to the March 1992 data set.