EPSL

ELSEVIER Earth and Planetary Science Letters 148 (1997) 5 17-525

History of water-column anoxia in the Black Sea indicated bY

pyrite framboid size distributions
Richard T. Wilkin a**, Michael A. Arthur a, Walter E. Dean b
’ Department ofGeosciences, The Pennsylrania State Unir~ersity lJnic,ersic Park, PA 16802. USA
b U.S. Geological Suwe,v. MS Y3Y Federul Center. Denr,er. CO 80225, USA

Received 28 October 1996; revised 25 February 1997: accepted 3 March 1997

Abstract

A detailed study of size distributions of framboidal pyrite in Holocene Black Sea sediments establishes the timing of a
change from deposition under an oxic water column to deposition under an anoxic and sulfidic water column. In the most
recent carbonate-rich sediments (Unit I) and in the organic carbon-rich sapropel (Unit II). framboid size distributions are
remarkably uniform (mean diameter = 5 p,m); over 95% of the framboids in Unit I and Unit II are < 7 pm in diameter.
These properties of framboidal pyrite are consistent with framboid nucleation and growth within an anoxic and suh‘idic water
column, followed by transport to the sediment-water interface, cessation of pyrite growth due to the exhaustion of reactive
iron, and subsequent burial. In contrast, the organic carbon-poor sediments of lacustrine Unit III contain pyrite framboids
that are generally much larger in size (mean diameter = IO pm). In Unit III, over 95% of the framboids are < 25 I.i,rn in
diameter, 40% of framboids are between 7 pm and 25 p,m, and framboids up to 50 pm in diameter are present. This
distribution of sizes suggests framboid nucleation and growth within anoxic sediment porewaters. These new data on size
distributions of framboidal pyrite confirm that the development of water-column anoxia in the Black Sea coincided with the
initiation of deposition of laminated Unit II sapropels.

Krword.~: Black Sea: anaerobic environment; framboidal texture: pyrite: black shale

1. Introduction carbon-rich such as ‘black shales’ (e.g.

sediments
[2]) and has been implicated as a causal
recently
Periods of nearly global oceanic anoxia appar-
factor in specific mass extinctions of benthic biota
ently have occurred intermittently during most of the
(e.g. [3.4]). Sedimentological and geochemical crite-
earth’s history (e.g. [l]). Few reliable techniques,
ria used by earth scientists to deduce whether or not
however, are available to infer paleo-depositional
particular sediments were deposited under oxic or
redox conditions, although the condition of water-
anoxic conditions include: the presence or absence of
column anoxia is frequently invoked as being essen-
benthic fauna; non-laminated versus laminated sedi-
tial for the accumulation and preservation of organic
mentary structures; iron-sulfur-carbon systematics
[5,6]; trace-metal abundances and ratios 171; and
* Corresponding author. E-mail: [email protected] organic biomarker indicators [8]. These techniques

0012-821X/97/$17.00 Copyright 0 1997 Elsevier Science B.V. All rights reserved

P/I s0012-821x(97)00053-8
518 R.T. Wilkin et &./Earth and Planetap Science Letters 148 (1997) 517-525

are most useful when used in combination because, 20% organic carbon. Homogeneous, fine-grained
in some cases, secondary diagenetic overprints may muds (turbidites) of variable thickness are commonly
lead to mistaken interpretations regarding paleo-de- found interbedded with Units I and II in the deep
positional environments. portion of the basin (e.g. [ 171). A range of dates have
Recently, Wilkin et al. [9] have shown that the been reported for the Unit I-Unit II contact: from
size distribution of pyrite framboids contained in 1.63 ka, based on varve counts [ 14,151, to 2.7 ka,
modern sediments corresponds to bottom-water re- based on radiocarbon analyses of samples collected
dox conditions. Framboids in sediments of modem from the deep basin [18]. The oldest Pleistocene-Ho-
euxinic basins, including the Black Sea, are charac- locene sediments, Unit III, are composed of massive
teristically smaller and less variable in size than to macrobanded, organic carbon-poor clays ( < 1%
those contained in sediments underlying oxic and organic carbon). Taken together, these units record
dysoxic water columns. Kinetic factors, such as the transition of the Black Sea from a largely oxic
growth rate and duration, are the principal controls Pleistocene freshwater lake (Unit III) to a brackish,
on framboid size. The smaller sizes of framboids in anoxic marginal marine basin formed during the last
sediments from euxinic basins apparently reflect post-glacial sea-level transgression (Unit II) and, fi-
shorter average growth periods relative to framboids nally, to the present anoxic marine basin (Unit I).
formed in ‘normal’-marine sediments; that is, those When did the Black Sea water column become
underlying oxygenated water columns. Shorter anoxic and sulfidic? Deuser [ 193 proposed that oxy-
growth periods in euxinic environments, in turn, are gen depletion was initiated when Mediterranean wa-
the result of the hydrodynamic instability of fram- ters began spilling into the basin through Bosporus
boids that form within the water column subjacent to Strait ca. 9 ka. The influx of saline waters led to a
the oxic-anoxic interface (see [9-l 11). density stratification in the basin, progressive deple-
The Black Sea is the largest modem permanently tion of deep-water oxygen, and the rise of the O,-
anoxic basin and, therefore, is often cited as an H,S interface into the deepest part of the basin at ca.
analogue to the type of depositional environment 7.3 ka, followed by the continued rise of the redox
present during the formation of organic carbon-rich transition to the present depth of about 100 m below
shales (e.g. [12]). Despite the considerable interest in the surface. It also is possible that the position of the
the geochemical history of the Black Sea, contro- oxic-anoxic boundary has been largely stationary
versy remains about the evolution of anoxia in the over the past 7.3 ka with the exceptions of possible
Black Sea water column. In this paper we report the short-term advances and retreats (e.g., [ 16,20,2 I]). In
size distributions of pyrite framboids preserved the sediments, the transition to anoxia is apparently
through time in the Black Sea sediments and discuss marked by the Unit III-Unit II transition or at the
implications regarding the history of water-column point where the sediments become laminated and
anoxia in the Black Sea. enriched in organic carbon [16]. However, Pedersen
and Calvert [22], Calvert [23], and Calvert et al. [24]
argued that Unit II sapropels are a result of increased
2. Black Sea sedimentology primary productivity, and were, in fact, deposited
under an oxic water column.
The principal features of the Black Sea sediments
are described in Ross and Degens [13], Hay et al.
[14], Arthur et al. [151, and Arthur and Dean [16]. In 3. Methods and materials
the uppermost meter of the deep-basin sediments,
four distinct units are distinguished. The most recent Gravity core GC59 was recovered from a depth of
sediments (Unit I> consist of up to 45 cm of finely 600 m in the southeastern region of the Black Sea
laminated, coccolith marl that contains l-5% or- (station 32, 41”58.5’N, 36”33.3’E) during Leg 1 of
ganic carbon and lo-75% CaCO, [15]. Below Unit I the 1988 R/V Knorr expedition. This core pene-
is an approximately 50 cm thick, finely laminated, trated well into Unit III and is therefore ideal for
carbonate-poor sapropel (Unit II) that contains I- studying the Unit III-Unit II transition. It also was
 R.T. Wilkin et al. /Earth and Planetap Science Letters 148 f 1997) 517-525 519

located on a ‘basin center’ high, away from major the measured framboid size distributions are slightly
lateral sediment inputs, and provides an undisturbed skewed to smaller sizes than found in the true size
record of Unit II and Unit III deposition, with no distributions (see [9]). Several additional sediment
apparent influence from mud flows or turbidites. samples from GC59 and BC53 were mounted,
Samples were also obtained from box core BC53 ground, and polished in 3 cm diameter molds. In
collected during the 1988 cruise at a depth of 2154 these samples, framboid size distributions were mea-
m (station 103, 42”39.7’N, 37”37.0’E). Organic and sured in reflected-light mode using a calibrated ocu-
carbonate carbon were determined using a Coulo- lar graticule.
metrics Coulometer System 140 on powdered sam-
ples at selected depth intervals [ 151.
Total iron and aluminum in the sediments were 4. Results
determined using inductively coupled plasma emis-
sion spectroscopy (ICP-ES) on acid-digested sam- The concentrations of organic carbon, carbonate
ples. The amount of chromium-reducible sulfur (as CaCO,), total iron, and pyritic iron from core
(pyrite S + acid - volatile S + elemental sulfur) was GC59 are shown in Fig. la,b. The profile in Fig. la
determined directly with a sulfur coulometer from shows the markedly lower organic carbon concentra-
the quantity of hydrogen sulfide released from pow- tions in Unit III compared to both Unit II and Unit I.
dered samples reacted with a boiling 1 M CrCl ,-HCl Pyrite content in Unit II (mean 0.90 rt 0.13 1 s.d.,
solution [25]. Because elemental sulfur is typically n = 22) is similar to that in Unit III above 380 cm
negligible compared to pyrite sulfur in recent sedi- (mean 0.87 f 0.17 1 sd., II = 12). These values are
ments and because no precautions were taken to slightly lower but comparable to average pyrite con-
prevent iron monosulfide oxidation during sample centrations reported for modern Unit I sediments
storage, the chromium-reducible sulfur is assumed (- 0.6-1.5 wt.% Fe; e.g. [9,17,29,30]). Below 380
here to be entirely composed of pyrite. The extent of cm, pyrite content abruptly drops to < 0.2 wt%.
sulfidization of iron minerals was assessed by com- This transition from lower to higher pyrite concentra-
paring the amount of pyritic iron with the amount of tions in Unit III might record the up-section change
iron released from about 50 mg of sediment after from freshwater to marine deposition (see [29,31])
24 h of reaction with 1 M HCl at room temperature and, potentially, could be used as a stratigraphic
(see [26,27]). marker across the basin. However, this transition
Polished thin sections were prepared by impreg- may also, in part, represent a downward diffusion
nating GC59 sediments with a low-viscosity epoxy front of either ZH2S (HZScaq) + HS) or SO:-.
[28]. The epoxy effectively replaced pore spaces Total iron in core GC59 shows relatively small
filled with seawater and prevented any shrinkage or variations with depth (mean 4.30 + 0.41 1 s.d.,
expansion of the core. After the epoxy cured, the n = 140), most of which can be accounted for by
core sections were cut, ground, and polished with CaCO, dilution. For example, the increase of CaCO,
diamond pastes. Framboid size distributions were in upper Unit III sediments between 190 cm and 260
measured using a optical microscope by photograph- cm corresponds to a low in total iron. In addition, the
ing a series of fields for 1 cm intervals of the cores. weight ratio of Fe to Al does not show much vari-
Framboid diameters were then measured using a ability and suggests a time-independent input of iron
calibrated scale prepared from a photograph of an minerals (Fig. 1~). Above 380 cm, about 20% of the
engraved micrometer. All photographs were taken in total iron is present as pyrite. An additional 15% of
transmitted light through thick sections (> 50 pm). the total iron is extractable with cold 1 M HCI and
For the majority of the particles entirely contained potentially available for pyrite formation [26]. The
within the thick sections, maximum or ‘true’ fram- remaining non-extractable iron is presumably present
boid diameters were measured. However, for those in silicates and unavailable for reaction with aqueous
particles contained on the edge of the sections, mea- sulfides (e.g. [32]).
sured particle diameters are skewed to smaller sizes Descriptive statistics of the framboid size distribu-
relative to the true diameters. The net effect is that tions are summarized in box-and-whisker plots in
520 R.T. Wilkin et al./ Earth and Planetary Science Letters 148 f 1997) 517-525

Organic C, C&O,/1 0, wt.% Fe, wt. % DOP, Fe/Al

0 2 4 6 8 10 0123456 0.0 0.4 0.6 1.2 1.6

Unit II
Unit Ill

400

c 0a 500

Fig. 1. Depth distributions in Black Sea core GC59 of: (a) organic carbon and carbonate (+ IO); (b) total iron and pyritic iron; and (c)
degree of pyritization (DOP) and Fe/AI (weight ratio). The DOP values 1261 are expressed as: DOP = [Pyrite Fe/(Pyrite Fe + HCl - soluble
Fe)]. Unit I geochemical data are not available for GC59.
Fig. 2. The boxes in Fig. 2 extend from quartile
Q = 0.25 to quartile Q = 0.75, thus including 50%
of the data. The median value is indicated as a bold
line. The lines extending to the left and right of the
box delineate minimum and maximum values, re-
m 2 cm, n=228
spectively. In general, framboid size distributions are
m S cm, n=227
unimodal and positively skewed. Framboid sizes in
m
Unit I and Unit II are quite similar but contrast
22 cm. n=358
- significantly with those in Unit III above 380 cm, as
UNIT I 55 cm, n=423 indicated on a cumulative size distribution plot (Fig.
UNITII 56 cm, n=348 3a). Note that, in both Unit I and Unit II, about 95%
of all framboids are < 7 km; whereas in Unit III
about 95% of all framboids are < 25 km and 40%
of framboids are between 7 pm and 25 km. In both
Unit I and Unit II, framboids larger than 18 km
were not observed and the average mean framboid
size is 4.7 + 0.5 Frn. In contrast, framboids in Unit
III average 10 km in diameter and some 50 km
framboids are present. Below 380 cm, where pyrite
174 cm, n=327
concentrations are low, framboids are noticeably ab-
UNIT II sent; pyrite is present as euhedral grains and irregu-
UNIT II lar masses that range in size from < 1 pm to
> 20 km.
Framboids are generally not found in clusters or
as polyframboids in either Unit I or Unit II; instead,
they tend to be dispersed and more concentrated in
10 20 30 40 50 the dark, organic carbon-rich layers (Fig. 4). How-
Framboid diameter, urn ever, in the uppermost sediments of Unit III, clusters
Fig. 2. Box-and-whisker plots showing framboid size distributions of lo-50 framboids are relatively common (Fig. 4).
from Black Sea cores GC59 and BC53 (see text). These clusters contain a range of different framboid
 R.T. Wilkin et al. /Earth and Planetary Science Letters 14X (1997) 517-525

0 60
b
a
.g 60
"I
2 40

3 20

0 0
0 50 0 50

Fig. 3. Cumulative framboid size distribution plots. (a) + = Unit III framboids: 0 = range of Unit II framboids: 0 = mnge of Unit I
framboids. (h) A = Unit III framboids: A = Unit III framboids in clusters.

sizes, as shown in Fig. 3b, but overall the size in fact, the core studied by Calvert [23] and Pedersen
distribution of the matrix framboids in Unit III is and Calvert [22] was anomalous in its geochemical
similar to that found in the clusters. characteristics.
More recently, however. Calvert et al. [29] studied
the sulfur isotope systematics of pyrite contained in
5. Discussion and conclusions Unit I, II, and III sediments collected from the
central region of the Black Sea at a depth of 22 18 m.
A distinct transition in pyrite crystallization dy- Their results show that pyrite in Unit 1 and Unit II is:
namics is recorded, based on framboid size distribu- (1) depleted in 34S compared to pyrite in both a
tions, at the Unit III to Unit II boundary. Wilkin et turbidite and in Unit III; (2) similar in composition
al. [9] proposed that the decreased variability of to dissolved sulfide at the chemocline; and (3) slightly
framboid size in sediments from modem euxinic enriched in 34S compared to dissolved sulfide in
basins is a consequence of pyrite framboid formation waters of the deep basin. Following the model pre-
above the sediment-water interface and subjacent to sented by Muramoto et al. [ 111, Calvert et al. [29]
the oxic-anoxic boundary. The comparatively larger conclude, based on their S-isotope study, that pyrite
mean framboid size and increased size variability in in Unit I and Unit II formed mainly in an anoxic and
Unit III sediments are properties shared with several sulfidic water column, consistent with our conclu-
modem salt marsh environments (see [9]). The data sions using framboid size distributions. However, we
in Figs. 2 and 3 suggest that both Unit I and Unit II note that, although our data would support this inter-
were deposited under continuously anoxic and sul- pretation, caution must be applied when evaluating
fidic waters, and that the Unit III-Unit II contact paleo-depositional environments of ancient sedi-
represents an abrupt transition to euxinic conditions. ments based on sulfur isotopic compositions of pyrite
Pedersen and Calvert [22] have argued that Unit II because isotopically depleted pyrite is expected to be
was deposited under an oxygenated water column, formed and preserved in any marine system. oxic or
based on the intervals of anomalous Mn concentra- anoxic, where iron availability limits pyritization
tions and increased ratios of halogens (i.e., I and Br) (e.g. [331X
to organic carbon in Unit II. They suggested that the Framboid size distribution as an indicator of bot-
enrichment of organic carbon was a consequence of tom-water redox conditions is most useful when
high primary production in the surface waters and an secondary pyrite growth is minimal; that is, when
increased flux of organic carbon to the sediment- primary framboidal texture is preserved without in-
water interface. Arthur and Dean [16], however, filling or overgrowths of pyrite. In modem, marine
cited a variety of textural and geochemical evidence euxinic basins, pyrite formation appears to be limited
that argued against the presence of bottom-water by the input of reactive iron rather than by the
oxygen during Unit II deposition. They suggest that, supply or quality of organic carbon (e.g. [ 17.34-361).
522 R.T. Wilkin et al/Earth and Planetary Science Letters 148 (1997) 5I 7-525
 R.T. Wilkin et al. / Earth and Planetary Science Letters 148 (1997) 517-525 523

Indeed, the data plotted in Fig. 1 show no correlation within Unit III during the deposition of Unit II. Also
between the concentrations of organic carbon and note that, immediately below the Unit II-Unit III
pyrite in Units II and III. Lyons and Bemer [ 171 boundary there is a pronounced spike in the concen-
found down-core trends of pyritic sulfur and degree trations of total iron, pyritic iron, and Fe/AI at 184
of pyritization to be approximately constant in deep- cm (Fig. 1). These geochemical features may repre-
water Unit I sediments. Furthermore, porewater dis- sent a paleo-pyritization front, produced by upward-
solved sulfide concentrations do not show evidence diffusing ferrous species precipitation with 2 HZ S
of removal, as might be expected if iron sulfides diffusing downward from a recently emplaced sul-
were precipitating within the sediments. Based on fidic water mass. The continued downward diffusion
these trends, Lyons and Bemer [17] concluded that of sulfide and subsequent pyrite precipitation in Unit
the majority of pyrite in deep-water Black Sea sedi- III are generally consistent with observed increases
ments formed at or above the sediment-water inter- in the degree of framboid infilling and proportion of
face. Therefore, in the modem regime, reactive iron large euhedral pyrite grains ( > 10 pm). On the other
is apparently efficiently cycled in the Black Sea hand, it is plausible that pyritization was rapid enough
chemocline from Fe(II1) oxides and oxyhydroxides to exhaust the supply of reactive iron during Unit III
to Fe(I1) monosulfides and disulfides; consequently, deposition and that pyrite formation was limited by
pyrite formation is effectively cut off in the sedi- the availability of sulfate in Unit III below 380 cm.
ments, consistent with the lack of infilled framboids This scenario, however, would require a suitable
and euhedral grains in Unit I (i.e., pyrite textures supply and quality of organic carbon in Unit III that
indicating secondary growth) [9]. is not readily apparent by considering the data in Fig.
In GC.59, depth-independent pyrite concentrations la.
and DOP values continue downward into Unit II and, Water-column anoxia during the sedimentation of
interestingly, into Unit III above 380 cm (Fig. lb,c). Unit I and Unit II also is supported by several
Thus, a similar argument for iron-limited, water organic biomarker studies (e.g. [8,37]). Repeta [37]
column pyrite formation can be made for the Unit II reported high-resolution, down-core distributions of
sapropels. However, although pyrite concentrations the pigment isorenieratene, a compound indicative of
and the degree of pyritization do not show immedi- the activity of phototrophic Chlorobium bacteria.
ate transitions across the Unit II-Unit III boundary These organisms oxidize H,S to S” and SO:- by a
(Fig. lb,c), the framboid size distributions change photosynthetic pathway and are active in water
significantly and indicate a relatively abrupt switch columns in which dissolved sulfide is present above
in bottom-water redox conditions (Fig. 2). The con- the photic zone. Isorenieratene is present at varying
stancy of DOP values across the Unit II-Unit III concentrations throughout Unit I and the bottom
boundary suggests that iron reactive towards ZH,S portion of Unit II; however, it is conspicuously
was nearly completely consumed to produce iron absent in the upper half of Unit II. This absence of
sulfides in both the anoxic water columns present isorenieratene does not necessarily reflect the ab-
during the deposition of Unit I and Unit II and, also, sence of dissolved sulfide from the water column but
the anoxic sediments of Unit III, that apparently may, in fact, be due to higher freshwater inflow to
were overlain by an oxygenated water column. Thus, the basin and a depressed or eroded chemocline
a change in bottom-water redox conditions across the [16,37].
Unit III-Unit II boundary would not have been In conclusion. framboid size distributions in the
interpreted on the basis of the DOP indicator [6]. Holocene Black Sea sediments are consistent with
This lack of change in the DOP might be due to the the hypothesis that the laminated sapropels of Unit II
downward diffusion of ,CH,S, concomitant pyritiza- were deposited under an anoxic and sulfidic water
tion, and consequent secondary increases in DOP column. The framboid size distribution criterion is a

Fig. 4. Photomicrographs of pyrite from Black Sea sediments. (a) Dispersed framboids within Unit 11 (transmitted light). (b) Matrix
framboids within Unit III (transmitted light). (c) Cluster of framboids in Unit III (reflected light).
524 R.T. Wilkin et al. /Earth and Planetary Science Letters 148 (1997) 517-525

promising tool for evaluating bottom-water redox tion of framboidal pyrite in modem sediments: An indicator
conditions in ancient sedimentary rocks. The method of redox conditions, Geochim. Cosmochim. Acta 60 (1996)
3897-3912.
may be particularly useful when used in combination
I101 J.M. Skei, Formation of framboidal iron sulfide in the water
with organic biomarker analyses; for example, the of a permanently anoxic fjord-Framvaren, south Norway.
presence or absence of isorenieratene, which indi- Mar. Chem. 23 (1988) 345-352.
cates the positioning of the redox transition relative llll J.A. Muramoto, S. Honjo, B. Fry, R.W. Howarth. J.L. Cisne,
to the photic zone. Sulfur, iron and organic carbon fluxes in the Black Sea:
Sulfur isotopic evidence for origin of sulfur fluxes, Deep-Sea
Res. 38 (suppl. 2) (1991) S1151-S1187.
I121 E.T. Degens, P. Stoffers, Stratified waters as a key to the
past, Nature 263 (1976) 22-26.
Acknowledgements
[I31 D.A. Ross, E.T. Degens, Recent sediments of the Black Sea,
in: E.T. Degens, D.A. Ross (Eds.). Black Sea-Geology,
This research was supported by the National Sci- Chemistry, and Biology, Am. Assoc. Pet. Geol. Mem. 20
ence Foundation (grant EAR-9628344). We thank E. (1974) 183-199.
Neff (Graduate School of Oceanography, Univ. [141 B.J. Hay, M.A. Arthur, W.E. Dean. E.D. Neff, S. Honjo.
Sediment deposition in the Late Holocene abyssal Black Sea
Rhode Island) for careful preparation of the thin
with climatic and chronological implications, Deep-Sea Res.
sections of GC59. R. Bemer and D. Canfield pro- 38 (1991) 1211-1235.
vided thoughtful reviews of the manuscript. We also [I51 M.A. Arthur, W.E. Dean, E.D. Neff, B.J. Hay. J. King, G.
acknowledge the use of the laboratory facilities of Jones, Varve calibrated records of carbonate over the last
2000 years in the Black Sea, Global Biogeochem. Cycles 8
the Ore Deposits Research Section of The Pennsyl-
(1994) 195-217.
vania State University (Dir. H. Barnes). 04Kl
ll61 M.A. Arthur, W.E. Dean, Organic-matter production and
preservation, and evolution of anoxia in the Holocene Black
Sea, Paleoceanography (1996) in press.
1171 T.W. Lyons, R.A. Berner, Carbon-sulfur-iron systematics
References of the uppermost deep-water sediments of the Black Sea,
Chem. Geol. 99 (1992) l-27.
Ill M.A. Arthur, B.B. Sageman, Marine black shales: Deposi- [I81 G.A. Jones, A.R. Gagnon. Radiocarbon chronology of Black
tional mechanisms and environments of ancient deposits, Sea sediments, Deep-Sea Res. 41 (1994) 531-557.
Ann. Rev. Earth Planet. Sci. 22 (1994) 499-551. 1191 W.G. Deuser, Evolution of anoxic conditions in Black Sea
I21 G.J. Demaison, G.T. Moore, Anoxic environments and oil during Holocene, in: E.T. Degens, D.A. Ross (Eds.), Black
source bed genesis, Am. Assoc. Pet. Geol. Bull. 64 (1980) Sea-Geology. Chemistry. and Biology, Am. Assoc. Pet.
1179-1209. Geol. Mem. 20 (1974) 133-136.
[31 M.M. Joachimski, W. Buggisch, Anoxic events in the late DOI C.R. Glenn, M.A. Arthur, Sedimentary and geochemical
Frasnian-causes of the Frasnian-Famennian fauna1 crisis?, indicators of productivity and oxygen contents in modem and
Geology 21 (1993) 675-678. ancient basins: The Holocene Black Sea as the “type”
[41 P.B. Wignall, R.J. Twitchett. Oceanic anoxia and the end anoxic basin. Chem. Geol. 48 (1985) 325-354.
Permian mass extinction. Science 272 (1996) 1155-I 158. PII T.W. Lyons, R.A. Bemer, R.F. Anderson, Evidence for large
[51 J.S. Leventhal, An interpretation of carbon and sulfur rela- pre-industrial perturbations of the Black Sea chemocline,
tionships in Black Sea sediments as indicators of environ- Nature 365 (1993)538-540.
ments of deposition, Geochim. Cosmochim. Acta 47 (1983) WI T.F. Pedersen, SE. Calve& Anoxia vs. Productivity: What
133-138. controls the formation of organic-carbon-rich sediments and
I61 R. Raiswell. F. Buckley, R.A. Bemer, T.F. Anderson, Degree sedimentary rocks. Am. Assoc. Pet. Geol. Bull. 74 (1990)
of pyritization of iron as a paleoenvironmental indicator of 454-466.
bottom-water oxygenation. J. Sediment. Petrol. 58 (1988) [23] S.E. Calvert, Geochemistry and origin of the Holocene sapro-
812-819. pel in the Black Sea, in: V. Ittekot, S. Kempe, W. Michaelis.
[71 B. Jones, D.A.C. Manning, Comparison of geochemical in- A. Spitzy (Eds.), Facets of Modem Biogeochemistry,
dices used for the interpretation of paleoredox conditions in Springer, 1990, pp. 326-352.
ancient mudstones, Chem. Geol. 111 (1994) 11 l-129. [24] S.E. Calve& R.E. Karlin, L.J. Toolin. D.J. Donahue. J.R.
I81 J.S. Sinninghe Damstt, S.G. Wakeham, M.E.L. Kohnen, Southon, J.S. Vogel, Low organic carbon accumulation rates
J.M. Hayes, J.W. de Leeuw, A 6,000-year sedimentary in Black Sea sediments, Nature 350 (199 1) 692-695.
molecular record of chemocline excursions in the Black Sea, [25] N.N. Zhabina, 1.1. Volkov, A method of determination of
Nature 362 (1993) 827-829. various sulfur compounds in sea sediments and rocks, in:
[9] R.T. Wilkin, H.L. Barnes, S.L. Brantley, The size distribu- W.E. Krumbein (Ed.), Environmental Biogeochemistry and
 R.T. Wilkin et al. / Earth and Planetun Science Letters 138 f 1997) 517-52.5 525

Geomicrobiology. Vol. 3. Ann Arbor Science, 1978. pp. istry. and Biology, Am. Assoc. Pet. Geol. Mem. 20 (1974)
735-146. 524-53 I.
[26] R.A. Berner, Sedimentary pyrite formation. Am. J. Sci. 268 [32] D.E. Canfield, R. Raiswell, S. Bottrell, The reactivity ol
(1970) l-23. sedimentary iron minerals toward sulfide. Am. J. Sci. 293
[27] J.S. Leventhal. C. Taylor. Comparison of methods to deter- (1992) 659-683.
mine degree of pyritization, Geochim. Cosmochim. Acta 54 [33] D.L. Gautier. Isotopic composition of pyrite: Relationship to
( 1990) 262 I-2635. organic matter type and iron availability in some North
[28] J.F. Swartz. N. Lindsley-Griffin, An improved technique for American Cretaceous shales. Chem. Geol. 65 (1987) 293-
studying structure of unlithified cohesive sediments. Proc. 303.
ODP Sci. Results I I2 ( 1990) 87-91. [34] C. Boesen. D. Postma. Pyrite formation in anoxic envtron-
[29] S.E. Calvert. H.G. Thode. D. Yeung, R.E. Karlin. A stable men& of the Baltic Sea, Am. J. Sci. 288 ( 1988) 575-603.
isotope study of pyrite formation in the Late Pleistocene and [35] S.E. Calvert. R.E. Karlin. Relationships between sulphur,
Holocene sediments of the Black Sea. Geochim. Cosmochim. organic carbon, and iron in the modern sediments of the
Acta 60 (1996) 1261-1270. Black Sea, Geochim. Cosmochim. Acta 55 ( 1991 t 21833
[30] A.G. Rozonov. 1.1. Volkov, T.A. Yagodinskaya, Forms of 2490.
iron in surface layer of Black Sea sediments, in: E.T. De- [36] J.J. Middelburg, Organic carbon. sulphur. and iron in recent
gcns. D.A. Ross (Eds.). Black Sea-Geology, Chemistry, semi-euxinic sediments of Kau Bay. Idonesia, Geochim.
and Biology, Am. Assoc. Pet. Geol. Mem. 20 (1974) S32- Cosmochim. Acta 55 ( 199 I ) 8 15-828.
541. [37] D.J. Repeta. A high resolution historical record of Holocene
[31] R.A. Berner, Iron sulfides in Pleistocene deep Black Sea anoxygenic primary production in the Black Sea, Geochim.
sediments and their paleo-oceanographic significance. in: Cosmochim. Acta 57 ( 1993) 4337-1332.
E.T. Degens. D.A. Ross (Eds.). Black Sea-Geology. Chem-