Limnol. Oceanogr..
41(8),
1996, 1629-1650
0 1996, by the American Society of Limnology and
Oceanography,
Inc.
Pathways of carbon oxidation in continental margin
sediments off central Chile
Bo Thamdrup and Donald E. Canfield’
Max Planck Institute for Marine Microbiology, Celsiusstr. 1, D-283 59 Bremen, Germany
Abstract
Rates and oxidative pathways of organic carbon mineralization were determined in sediments at six stations
on the shelf and slope off Conception Bay at 36.5”s. The depth distribution of C oxidation rates was determined
to 10 cm from accumulation of dissolved inorganic C in l-5-d incubations. Pathways of C oxidation were
inferred from the depth distributions of the potential oxidants (0,, N03-, and oxides of Mn and Fe) and
from directly determined rates of SOd2- reduction. The study area is characterized by intense seasonal
upwelling, and during sampling in late summer the bottom water over the shelf was rich in NO,- and depleted
of OZ. Sediments at the four shelf stations were covered by mats of filamentous bacteria of the genera
Thioploca and Beggiatoa. Carbon oxidation rates at these sites were extremely high near the sediment surface
(> 3 pmol cmm3 d-l) and decreased exponentially with depth. The process was entirely coupled to SOd2-
reduction. At the two slope stations where bottom-water O2 was > 100 PM, C oxidation rates were lo-fold
lower and varied less with depth; C oxidation coupled to the reduction of 02, N03-, and Mn oxides combined
to yield an estimated 15% of the total C oxidation between 0 and 10 cm. Carbon oxidation through Fe
. reduction contributed a further 12-29% of the depth-integrated rate, while the remainder of C oxidation was
through SOd2-
reduction. The depth distribution of Fe reduction agreed well with the distribution of poorly
crystalline Fe oxides, and as this pool decreased with depth, the importance of SOd2- reduction increased.
The results point to a general importance of Fe reduction in C oxidation in continental margin sediments.
At the shelf stations, Fe reduction was mainly coupled to oxidation of reduced S. These sediments were
generally H,S-free despite high SOd2-
reduction rates, and precipitation of Fe sulfides dominated H,S scav-
enging during the incubations. A large NO,- pool was associated with the Thioploca, and the shelf sediments
were thus enriched in N03- relative to the bottom water, with maximum concentrations of 3 pmol cm-3.
The NO,- was consumed during our sediment incubations, but no effects on either C or S cycles could be
discerned.
The mineralization of organic matter in the sea floor
proceeds through a complex web of fermentative and
respiratory microbial pathways where the oxidation of
organic carbon to CO2 is balanced overall by concomitant
reduction of the inorganic electron acceptors 02, N03-,
oxides of Mn and Fe, and S042-. To understand the
regulation of carbon oxidation, independent quantifica-
tion of each pathway would be ideal, but such quantifi-
cation is only possible for a few of the pathways. The
radiotracer technique for measuring S042- reduction rates
(Jorgensen 1978) is one of the most robust methods avail-
able. Most importantly, S042- reduction is an entirely
biological process that uses only organic substrates (and
H2) and has H2S as the only immediate product. All of
the other electron acceptors may be reduced in sediments
either abiotically or through bacterial catalysis by one or
more of the reduced inorganic species (H2S, Fe2+, Mn2+,
l Present address: Institute of Biology, University of Odense,
DK-5230 Odcnse M, Denmark.
Acknowledgments
We are grateful to Henrik Fossing, Victor Aricl Gallardo, and
Bo Barker Jorgensen for initiating, organizing, and leading the
Thioploca Expedition 1994 and to Kirsten Neumann and Swantje
Fleischer for skillful technical assistance. We also thank the
Captain and crew of RV Vidal Gormaz as well as all members
of the scientific party for good company and collaboration dur-
ing the expedition and for permission to cite unpublished results
from the cruise. The comments of the reviewers are appreciated.
or NH4+), and no technique is available for measuring
organotrophic respiration alone. Hence, the quantifica-
tion of these processes relies on multilateral approaches
wherein different types of measurements together con-
strain the rates.
Determinations of bacterial S042- reduction rates in
sediments are numerous and have shown that this path-
way accounts for lo-90% of the C oxidation in coastal
sediments, with 50% as a median value (Jorgensen 1982;
Henrichs and Reeburgh 1987; Canficld 1993). Most of
the sulfide produced from S042- reduction is typically
rcoxidized within the surface sediment rather than being
buried. Thus, although benthic O2 uptake provides a good
estimate of sediment metabolism in most continental
margin sediments, a significant part of this O2 consump-
tion is typically not directly coupled to C oxidation, but
instead is coupled to the reoxidation of inorganic com-
pounds such as H2S (Jorgensen 1982). No technique is
presently available for direct quantification of aerobic
respiration in sediments.
Several methods have been used for measuring sedi-
mentary N03- reduction (Sorensen 1978; Seitzinger et
al. 1984; Nielsen 1992), with results that generally dem-
onstrate a minor role (I 10%) for C oxidation in coastal
sediments (Sorensen et al. 1979; Canfield et al. 1993a).
The significance for C oxidation may be higher in areas
with low O2 and (or) elevated N03- concentrations in the
bottom water (cf. Canfield 1993).
The contributions of Mn and Fe reduction to C oxi-
dation are the most infrequently quantified of the respi-
1629
1630
Thamdrup and Canfield
Fig. 1. Map showing sampling locations.
ratory types. Estimates based on pore-water gradients of
Mn2+ and Fe2+ have found the processes to be of little
importance (e.g. Bender and Heggie 1984; Reimers et al.
1992). In the presence of sediment and pore-water trans-
port by bioturbation, however, such estimates underes-
timate actual turnover rates (Aller 1980; Canfield 1993),
and a further complication arises from the adsorption or
precipitation of reduced Mn and Fe (Canfield et al.
19933;
Thamdrup et al.
1994a).
Indeed, Aller (1990) found that
in Mn oxide-rich sediments of the Panama Basin, Mn
reduction dominated C oxidation and that the process
was supported by bioturbation, which mixed the oxides
and organic matter into anoxic sediment layers.
A major difficulty in quantifying dissimilatory Mn and
Fe reduction arises from the competing abiotic reactions
with H2S and, for Mn oxides, with Fe2+ (Pyzik and Som-
mer 1981; Postma 1985; Burdige and Nealson 1986).
Reaction with H2S may be so rapid that this compound
may be undetectable even in sediments with high rates
of SOd2- reduction (Goldhaber and Kaplan 1974; Can-
field et al. 1992). Reaction with H2S was found to dom-
inate Fe reduction in a coastal bay, with dissimilatory Fe
reduction contributing <5% to C oxidation (Thamdrup
et al.
1994a).
Similar conclusions have been reached for
Mn reduction in a number of relatively Mn-poor sedi-
ments (Canfield et al.
19933;
Aller 1994; Thamdrup et
al.
1994a).
By contrast, Sorensen (1982) showed that in oxidized
sediment, Fe reduction proceeded through biological ca-
talysis and was independent of SOd2- reduction. Fur-
thermore, from the depth distribution of 02, N03-, and
rates of SOd2-
reduction in some coastal sediments, a
zone dominated by Fe reduction has been suggested (So-
rensen and Jorgensen 1987; Hints et al. 199 1). In a study
of three sites in the Norwegian Trough, Canfield et al.
(19933)
compared total mineralization rates to SOb2- re-
duction rates in anoxic sediment incubations. Based on
excess total mineralization relative to the mineralization
due to SOd2- reduction, Fe reduction was found to con-
tribute 44 and 69% of anaerobic C oxidation at two sites,
whereas at the third site, essentially all mineralization
was coupled to Mn reduction. These large contributions
Table 1. Positions and bottom-water data for stations sam-
pled 12-23 March 1995 (nd-not detected).
Posit ion
Depth Temp
02”
N03-
(36’S+, ‘73”W+) (m)
(“Cl (CLW (PM)
C6
37.3’S, 00.5’W
34
11.4 nd 3
c7
36.5’S, 00.6’W 37
11.6 nd
21
Cl8 30.8’S, 07.7’W
87 11.7 nd
18
C26 25.9’S, :!3.4’W
122
11.2 nd
15
C40
20.1 ‘S,
43.7’W
1,015 6.0
105 43
c41
19.6’S, 49.3/W
2,000 3.9
220 35
* Data provided by J. K. Gundcrsen and R. N. Glud.
were associated with strong bioturbation and, in the last
case, with a very high Mn oxide content. From compar-
ison with benthic O2 uptake rates, it was concluded that
the anaerobic processes dominated C oxidation, whereas
aerobic respiration contributed only 4-l 7%; furthermore,
it was suggested that the role of aerobic respiration in C
oxidation has generally been overestimated in continental
margin sediments (Canfield ct al. 1993a).
Although the studies cited above have provided the
first evidence on the diversity of C oxidation pathways,
it is also clear that a general understanding of the relative
importance of these pathways and their regulation in sed-
iments has not been reached. We have applied an ap-
proach similar to that of Canlield et al. (1993b) to sedi-
ments from the upwelling region off central Chile. Stations
were chosen on a transect that ran from a highly pro-
ductive shelf area with O,-deficient bottom waters to half-
way down the continental slope. This approach gave us
an opportunity to study how the pathways of C oxidation
responded to variations in bottom-water O2 levels. This
study extended .:he small database on the significance of
dissimilatory Fe and Mn reduction and also served as a
further test of tie methodology.
Study area
As part of the,
“Thioploca
Cruise 1994” onboard RV
Vidal Gormaz
(Fessing ct al. 1995), sediments were col-
lectcd on the car tinental margin off Conception Bay (Fig.
1, Table 1). Stations were distributed along a transect
from the mouth of the Bay at 34-m depth, across the 40-
km-wide shelf, to 2,000-m depth on the slope. The region
is characterized
by
a very high upwelling-driven primary
production during summer (September-April), with up-
welling induced by south-southwest winds that typically
blow for period; of a week, followed by a few days of
calm or north winds (Ahumada et al. 1983; Arcos and
Wilson 1984; Peterson et al. 1988). The noncontinuous
upwelling pattern causes strong temporal variations in
primary productivity. Sediments were sampled during
late summer (12-23 March), and upwelling-favorable
winds blew throughout the first half of March.
A pronounced O2 minimum zone is associated with the
south-flowing equatorial subsurface water, which is found
at 100-400-m depth in nonupwelling periods (Brandhorst
1959; see Sobarzo 1994). During summer upwelling, this
C oxidation in Chile margin
1631
water mass may reach 20-m depth and may cover the
shelf and reach into the Conception Bay, thereby reducing
bottom-water O2 concentrations over the shelf (Ahumada
et al. 1983; Peterson et al. 1988). Thus, O2 concentrations
< 11 PM have been reported from the shelf bottom water
during this season (Roa et al. 1995). Inside the bay, bot-
tom-water anoxia, H2S release, and faunal mass mortality
are recurring phenomenona (Ahumada ct al. 1983). Dur-
ing the sampling period, an O,-deficient zone was found
at 100-300-m depth offshore, and it stretched up over
the shelf to 30-40-m depth, so that all shelf sediments
studied were fully anoxic (Table 1; R. N. Glud et al. in
prep.).
by driving the sediment core from below toward a sta-
tionary lid with a central l-mm-wide hole covered with
a paper filter. Water was collected in 2-ml aliquots, each
corresponding to -0.9-mm depth in the core, depending
on porosities.
Sedimentation rates from 210Pb distributions at stations
C4 1, C40, Cl 8 (Fig. l), and at locations close to C6 and
C7, all fall in the range l-2.2 mm yr-l (mean, 1.2; M.
Salamanca pers. comm.). Giant filamentous bacteria of
the genus
Thioploca
were first described from the shelf
sediments outside the bay, where they occur in matlike
masses in the O2 minimum zone (Gallardo 1977). On the
inner shelf, where O2 levels vary strongly over the year,
the mats proliferate in summer and may break up and
become exported with bottom currents during winter (V.
A. Gallardo pers. comm.). Single filaments are up to 40
pm wide and 7 cm long (Fossing et al. 1995) and arc
found in conspicuous (up to 1.5-mm-thick) sheathed bun-
dles. These filaments belong to the species
Thioploca chi-
lea
and
Thioploca araucae
(Maier and Gallardo 1984a;
Teske et al. 1996). Sulfide oxidation has been proposed
as their main method of metabolism (Maier and Gallardo
19843), but despite wide distribution of these species in
the Chilean and Peruvian upwelling areas (Gallardo 1977),
their ecology is poorly known and was a main subject of
investigation during the cruise (Fossing et al. 1995).
For the determination of carbon mineralization rates
and pathways, sediment from the upper 10 cm of six cores
(434 cm2 in total) was incubated as described by Canfield
et al. (1993b) in laminated ethyl-vinyl-alcohol plastic
Wiirgler bags (Ril-O-Ten; Hansen 1992; Kruse 1993) af-
ter it had been sectioned into eight depth intervals and
the parallel intervals from different cores were mixed.
Sediments were processed over - 2 h in an N,-filled glove
bag beginning within 4 h of retrieval. The incubation bags
were first closed with a big clip, which eliminated any
head space, and they were then heat-sealed behind the
clip, thus avoiding heating of the sediment. The bags were
incubated dry at bottom-water temperature (Table 1) and,
in addition to the initial sample, sampled four times at
regular intervals in the glove bag. Apart from the initial
exposure during loading, the bags were exposed to room
temperature (-- 18°C) for < 10 min during samplings to
avoid temperature artifacts on rates. The generally linear
accumulation of metabolitcs during our incubations ar-
gues against any significant transient stimulation of met-
abolic activity resulting from initial sample heating and
handling. During incubation, the bags were not placed in
a larger N,-filled bag as practiced by Canfield et al. (1993b).
The range of oxygen permeability reported for the plastic
(Hansen 1992; Kruse 1993) at a maximum bag surface :
volume ratio of 0.2 m2 liter-’ corresponds to an oxygen
input of l-4 nmol cm-3 d-l, which is insignificant for
the metabolically active sediments studied here. Gener-
ally, however, we recommend the use of an outer anoxic
bag for sediment incubations.
Methods
The sediments were sampled with a multiple corer
(Barnett et al. 1984) that retrieved up to eight cores in
polycarbonate liners of 9.6~cm i.d, although for stations
C7 and C40, a 30 x 30-cm box core was subsamplcd into
such liners on deck. Cores were only accepted when the
surface appeared neither affected by resuspension nor, in
presence of surface mats, disrupted or pushed down dur-
ing coring. Cores were immcdiatcly transferred to an in-
cubator set at bottom-water temperature (Table 1).
For stations C7-C26, the total incubation time was 30-
35 h; station C40 was incubated for 4 d and C4 1 for 5 d.
At station C6, the incubation was ended after the second
sampling (5-h incubation) because of incubator failure.
With stations C7 and Cl 8, we prolonged the incubations
to 6 and 10 d, respectively, to monitor long-term changes
in the pore water and solid phases. For each sampling,
sediment was loaded into centrifuge tubes and pore water
was retrieved under N2 as described above.
Sediment incubation and pore- water extraction
-For
the pore-water chemistry, one or two cores were pro-
cessed, generally within 1 h of retrieval. In an N,-filled
glove bag, the sediment was sectioned in appropriate in-
tcrvals and loaded into polypropylene centrifuge tubes.
The tubes were tightly capped and centrifuged for 5-10
min, and after reintroduction into the glove bag, the pore
water was sampled and filtered through Whatman GF/F
glass-fiber filters. Conservation and analyses of samples
are described below.
During the short-term incubations, rates of sulfate re-
duction were determined once at C6 and C7 and twice at
all other sites in splits of lo-ml sediment with 35S042-
tracer (Jorgensen 1978) in 3-h (shelf sites) to 10-h (C4 1)
incubations at the same temperature as the bags. For
termination, the sediment was fixed in 20% zinc acetate
and frozen. The reduced 35S was rccovcred by distillation
with boiling acidic Cr2+
solution (Zhabina and Volkov
1978; Canfield et al. 1986; Fossing and Jorgensen 1989).
At the end of the short-term incubations, pore-water
pH was detcrmincd with a glass electrode that was in-
serted into subsamples of the sediment and was calibrated
with NBS standards.
A finer spatial resolution of the N03- distribution near
Pore- water analyses
-Portions of pore water for X02
the sediment surface was attempted with a whole-core
(dissolved inorganic carbon) and NH4+ analyses were fil-
squeezer (Bender et al. 1987). Pore water was extruded
tered into 1.8-ml glass vials that wcrc capped with Teflon-
1632
Thamdrup and Canfield
coated butyl rubber septa, leaving no gas phase and main-
taining anoxia. The samples were stored at 4°C and-an-
alyzed right after termination of the short-term incuba-
tions or, for the pore-water cores, within 5 d by means
of flow-injection systems with gas-exchange and conduc-
tivity detection for both spccics (Hall and Aller 1992; SD
2% for both). Although we cannot exclude slight X02
production from DOC oxidation during the brief storage,
we expect this to be insignificant rclativc to the produc-
tion in the sediments given the storage conditions. Sam-
ples that contained H2S from C6 and samples from pro-
longed incubations at C7 and Cl 8 that contained H2S
were first analyzed for NH,+ and then treated with 5%
vol of a 10% H202 solution (R. C. Aller pcrs. comm.).
This treatment removed all H2S within 10 min and had
no effect on t;COz measurements, as determined from
similarly treated standards.
Pore water for determination of N03- +N02- was
stored frozen. For analysis, N03- was reduced to N02-
by shaking the sample with spongy cadmium, and total
N02- was subsequently determined (Jones 1984). Tests
showed that 3 h of reaction ensured complete reduction
to N02- without overreduction to NH4+.
Dissolved Fe2+ in the pore-water cores was determined
immediately after filtration by colorimctry with a Fer-
rozine solution without reducing agent (dct. limit 1 PM;
SD 2%; Stookey 1970), whereas pore-water Fe2+ in the
incubated sediment was determined in acidified samples
at the end of the incubation. Dissolved Mn2+ and Ca2+
were analyzed in acidified pore water by flame atomic
absorption spectroscopy (Mn2+ det. limit 0.5 PM; SD 2%
for both).
Sulfate was determined by nonsuppressed anion chro-
matography (SD 1%). Samples for H2S analysis (100-300
~1) were fixed immediately in 50 ~1 20% Zn acetate and
frozen for later analysis by the methylenc blue method
(dct. limit 1 PM; SD 5%; Cline 1969).
Solid-phase analysis -
Portions of sediment from the
beginning of the incubations and the scdimcnt remaining
in the bags at the end were sampled under N2, frozen,
and later analyzed for solid Fe, Mn, and S phases. Iron
was extracted with dithionite-citrate-acetic acid (DCA;
pH 4.8; Lord 1980; Canfield 1989) and acidic ammonium
oxalate (pH 3.0; Schwertmann 1964). DCA extracts all
free iron oxides (except some magnetite; Canfield 1988;
1989) together with Fe(II) phases such as FcS and FcCO,
(Thamdrup et al. 1994a; Kostka and Luther 1994) but
does not allow a separate determination of Fe(III) and
Fe(II). Oxalate has been used to extract poorly crystalline
Fe oxides from soils and scdimcnts (Schwertmann 1964;
Canheld 1988, 1989). Its selectivity is, however, dcpen-
dent on the absence of Fc2+, because the Fe2+-oxalate
complex efficiently catalyzes the dissolution of crystalline
Fe(III) phases (Fischer 1973; Sutcr et al. 1988). Because
ferrous phases such as FeS and FeC03 are also dissolved
by this extraction, unwanted dissolution of unreactive
Fe(III) phases from anoxic Fe(II)-bearing sediments re-
sults (Thamdrup et al.
1994a;
Kostka and Luther 1994).
To avoid this catalytic interference of Fe2+, we air-dried
the sediment befi1t-e extraction, thereby oxidizing FeS and
FeCO, to ferrihlrdrite (Canfield 1988, 1989; Raiswell et
al. 1994). Calibration experiments with pure Fe phases
mixed into dried sediment conhrmcd the selectivity to-
wards poorly crystalline oxides (D. E. Can field unpubl.).
In parallel with the (oxic) extraction of dried sediment,
Fe2+ was determined in an anoxic oxalatc extraction of
fresh (frozen) sediment (Phillips and Lovley 1987) in or-
der to quantify the contribution of Fe(II) phases to the
oxalate-extractable Fe. Oxalate-extractable Fe(III) was
defined as the difference between thcsc two measures.
We used IO-ml extractants with - 150 mg wet or 50
mg of dry sediment. Extraction times were 1 h for DCA
and 4 h for oxaiate, and the extraction conditions were
otherwise as described by Canfield et al. (1993b) and
T.hamdrup et al. (1994a). Total extractable Fe and Fe(II)
were detcrminec. with Ferrozine (above references). Man-
ganese was determined in the DCA extracts by flame
AAS.
Acid-volatile sulfide (AVS = FeS + H2S) and chro-
mium-reducible sulfur (CRS = FcS, + So) were deter-
mined after a two-step distillation with cold 2 N HCl and
boiling 0.5 M Cr 2+
solution (Fossing and Jorgensen 1989).
Elemental sulfur was extracted by shaking the samples
with methanol for 24 h and was then determined with
HPLC (T. G. Ferdelman et al. unpubl.). For this deter-
mination, sediment fixed with Zn acetate from the SOd2-
reduction measurements was used. Concentrations of FeS
and FeS, were calculated as AVS - H2S and (CRS - So)/
2, respectively.
Organic C and total N and S were determined in dried
sediment on an elemental analyzer (Carlo Erba; SD 2%);
inorganic C was removed by acidification (HCl) and dry-
ing of the sample boats prior to analysis. Inorganic C in
carbonates was determined by acidification of 0.1-0.5 g
of dried sediment with 10 ml of 4 M HCl in 120-ml serum
bottles and measurement of the evolved CO2 in the head-
space by gas ch.romatography with thermal conductivity
detection (SD 5%). These values were corrected for car-
bonate alkalinity.
Thioploca
analysis and N03- extractions-
For deter-
mination of the bulk composition of
Thioploca,
a IO-cm-
deep sediment core with a well-developed mat at the
surface was sieved (1 mm) and washed with surface sea-
water. Remainmg detritus and macrofauna were removed
with forceps, leaving a visually clean pellet of sheathed
Thioploca,
which was then frozen. About 1 g wet wt of
Thioploca
was ‘reeze-dried and ground for analysis. Or-
ganic C, N, and S were determined with an elemental
analyzer as above. The N03- content was analyzed as
described below. To estimate the contribution of salt to
the dry weight, Cl- was extracted in water and analyzed
by ion chromatography. The total salt content was cal-
culated from Cl- and the typical composition of seawater.
Because we discovered during the cruise that
Thioploca
concentrated N03- far beyond pore-water concentra-
tions, an N03- extraction of total sediment was tested
and applied. Drying and rewetting had proven to be an
efhcient way to release N03- from single filaments of
C oxidation in Chile margin
1633
Corg (%), C/N (mole/mole)
of the test extractions (see
below).
We interpret this stable
background concentration as an artifact for which we
have no certain explanation, but which could be the result
of nitrification during the extraction. This background
was subtracted in calculations of N03- pool sizes.
Results
Sediment description -
White frlamcntous bacteria
(Thioploca
spp. and
Beggiatoa
spp.) were observed in
masses in the anoxic shelf sediments and in mats formed
on the surface. The sediment at station C6 was black and
sulfidic and had a web of
Beggiatoa
filaments practically
floating on 2 cm of fluff that graded into silt. At C7 and
C18,
-2-cm-thick matlike structures containjng
Thio-
ploca,
the small suspension feeding polychaetc
Paraprion-
ospio pinnata
(- 1 cm), and juvcnilc squat lobsters,
Pleu-
roncodes monodon
(- 1 cm) were observed on the surface
of the otherwise silty, dark brown scdimcnts. Single strands
of
Thioploca
and a few burrows of larger polychaetes were
seen to IO-cm depth. Station C26 did not have a distinct
bacterial mat at the surface, but strands
of Thioploca
were
distributed within the upper 10 cm. This sediment con-
sisted of green silt to 10-l 5-cm depth and gray clay with
a purple tinge below this lcvcl. At the slope stations, the
sediment was brown and silty and infauna was indicated
by burrow structures.
Solid phase and pore- water chemistry-
To facilitate di-
rect comparison of rates and pool sizes, concentrations
of solids (except C,, and N) are given on a volume rather
than on a dry weight basis. To some extent, this choice
masks changes that occur within the solids, because a
component that makes up a constant fraction of the solids
increases with depth solely owing to compaction. This
masking effect is particularly significant at the inner sites,
where porosities decreased from > 0.9 5 in the surface mat
to -0.85 at IO-cm depth.
The distribution of the organic C and the Corg : N ratio
in the scdimcnts are shown in Fig. 2. Only the surface
layers showed pronounced changes through the transect,
with high Corg
concentrations of up to 8% dry wt at the
shallower stations decreasing to - 3O/o at the slope stations.
A concentration -3% was also approached in the deeper
parts of the shelf cores. At C26, Gory was -4% in the
upper 10 cm and dropped to 1% below this. This change
coincided with the abrupt color change from green to gray.
The station also stood out with respect to other solid
components (see
below).
Low C,, : N ratios of 6-7 were
associated with the high C,, in the mats; otherwise, values
were -9.
The distribution of inorganic carbon (ZC02) and NH4+
in the pore waters showed a uniform trend along the
transect, and the extremely steep concentration gradients
and strong convex curvature of the profiles at the near-
shore stations indicated high rates of C mineralization
(Fig. 3). Offshore, both the surface gradients and the cur-
vature of the concentration profiles decreased dramati-
cally. The ZC02 and NH4 + distributions were generally
parallel. At C7 and in one core at C18, however, a pref-
erential release of NH4+
was indicated near the surface.
0 2 4 6 810 0 2 4 6 810
l + .
0
0
C6
0
L
0
c7
0
:
C18:+
i-7
O0 %g
l + C/N
l
I
0
C26
l
10 0"
1
0
l
0
c40
l
I
0
l
I
0
c41
l
20 L
Fig. 2.
Depth distributions of organic carbon and C,, : N
ratios, Data both from scdimcnt used for incubations (dia-
monds) and from a core for port-water analysis (circles) are
shown.
Thioploca
(L. P. Niclscn pcrs. comm.). For the extraction,
frozen sediment was dried either at 80°C after brief thaw-
ing or in a freeze-drier that avoided thawing completely.
The sediment was ground by hand and 50-100 mg (5 mg
for
Thioploca
samples) was then extracted in 10 ml water
for 15 min. After centrifugation, N03- in the supernatant
was determined as in the pore-water samples. The ex-
traction procedure was tested with surface sediment from
the N03- rich C7. Large and highly reproducible amounts
of N03- were released in the first 15-min extraction, and
the yield decreased sharply in a second wash. The release
of N03- decreased little in three further washes and sta-
bilized at concentrations -0.2 pmol g-l, corresponding
to l-2 PM in the extract independent of the yield in the
first wash. Extraction in 1 M KC1 or sonication during
the extraction did not increase the yield (data not shown).
Also, extraction of the decpcst sediment sections and of
sediment from the prolonged incubations, where strongly
reducing conditions developed, consistently yielded non-
zero N03- concentrations around the “background” level