452
Holmesetal.
mean = 9.0%0; s.d. = 0.8%0) and the samples that
were rinsed, but not decalcified (n = 5; mean =
9.9%0, s.d. = 0.6%0) were not significantly different from the untreated samples (Anova: P < 0.05).
The higher standard deviations ofthe "North Sea
Standard" subjected to decalcification or rinsing
relative to "WST" may be due to the much lower
N-content of the former. Based on the tests done
in both laboratories, it appears that the removal of
carbonate with weak HCI did not affect the measured 8 15 N values ofthe sediments discussed in the
present work, whether the sediments are rinsed or
not. However, the high standard deviations noted
in the treated "North Sea Standards" may indicate
that 8 15 N values in very nitrogen-poor sediments
are more liable to be affected by acid treatment and!
or rinsing with water.
Calcium carbonate in the sediment samples discussed in the present work was removed in using
the same procedure as for the "Watt Standard".
Sediment was decalcified in I N HCI for I hour at
50°C, filtered, then dried in a 50°C drying oven
overnight. The decalcified sediment samples (25ISO mg) were reground with copper oxide (CuO;
50% of the sample weight) to facilitate combustion. The sediment!CuO mixture was placed into
tin boats and pressed into small cylinders (-7 x 4
mm) prior to isotopic measurement in the system
at the University of Bremen described above.
Sediment sample standard deviations based on repeated measurement of reference sediment standards (WST 1,2 and 3) were:O: 0.2%0. This is similar to the results of Fry et al. (1992), who reported
a standard deviation of 0.21 %0 based on analyses
of riverine sediments using an automated system
similar to the one used in our laboratory. Sediment
trap samples had higher standard deviations associated with them (:0: 0.4%0) due to their small sample size.
Several different standards were measured to
control the quality of the nitrogen isotopic measurements. At least one laboratory working standard (WST 1,2 or 3) was measured per five samples. Additionally, IAEA N-I and IAEA N-2 standards from the International Atomic Energy Agency
(Vienna, Austria) and peptone (Merck chemicals)
were measured frequently to check the reliability
of sample combustion and mass spectrometric
measurements.
Although it is possible to measure both8 13 C OCg
and 8 15 N on one sample, we found that with the
large sample sizes needed for 8 15 N measurement,
organic carbon was present at such high concentrations that the CO, produced during sample combustion could not be completely isolated from the
nitrogen gas using the cryogenic method of
the trapping box. This resulted in unreliable isotopic
measurements. To circumvent this problem,
an extra cold trap immersed in liquid nitrogen was
added upstream ofthe trapping box to retain CO 2 ,
but let N, pass through to the trapping box.
This meant that stable carbon isotopic measurements had to be carried out on separate samples.
Samples for 8 J3 C OCg measurements were placed in
silver foil crucibles, decalcified with 6 M HCI and
dried on a hot plate at approximately 80°C.8 13 C OCg
analyses were performed using a Finnigan MAT
delta E mass spectrometer coupled to a Heraeus
CHN elemental analyzer and automated trapping
box. The standard deviation was < 0.1 %0 based on
duplicates and repeated measurements of the reference sediment "WST I".
Study Area
The Benguela region, off the southwest coast of
Africa, is comprised of the Angola Basin in the
north and the Cape Basin in the south. These two
hydrographically and productively distinct areas are
separated by a topographic high (Walvis Ridge) and
a hydrographic feature known as the AngolaBenguela Front (ABF). The warm, low nutrient
South Equatorial Countercurrent (SECC) flows
southeastward toward the Angola Basin at subsurface depths until it reaches the surface at around
100S and continues southward as the Angola Current (AC) (Moroshkin et al. 1970; Gordon and
Bosley 1991). The Benguela Current (BC), the
eastern boundary current of the South Atlantic
subtropical gyre, brings cold surface water into the
Cape Basin. The BC splits into the Benguela Oceanic Current (BOC) and the Benguela Coastal
Current (BCC) at around 28°S (Stramma and
Peterson 1989). The BOC flows offshore toward
Holmesetal.
mean = 9.0%0; s.d. = 0.8%0) and the samples that
were rinsed, but not decalcified (n = 5; mean =
9.9%0, s.d. = 0.6%0) were not significantly different from the untreated samples (Anova: P < 0.05).
The higher standard deviations ofthe "North Sea
Standard" subjected to decalcification or rinsing
relative to "WST" may be due to the much lower
N-content of the former. Based on the tests done
in both laboratories, it appears that the removal of
carbonate with weak HCI did not affect the measured 8 15 N values ofthe sediments discussed in the
present work, whether the sediments are rinsed or
not. However, the high standard deviations noted
in the treated "North Sea Standards" may indicate
that 8 15 N values in very nitrogen-poor sediments
are more liable to be affected by acid treatment and!
or rinsing with water.
Calcium carbonate in the sediment samples discussed in the present work was removed in using
the same procedure as for the "Watt Standard".
Sediment was decalcified in I N HCI for I hour at
50°C, filtered, then dried in a 50°C drying oven
overnight. The decalcified sediment samples (25ISO mg) were reground with copper oxide (CuO;
50% of the sample weight) to facilitate combustion. The sediment!CuO mixture was placed into
tin boats and pressed into small cylinders (-7 x 4
mm) prior to isotopic measurement in the system
at the University of Bremen described above.
Sediment sample standard deviations based on repeated measurement of reference sediment standards (WST 1,2 and 3) were:O: 0.2%0. This is similar to the results of Fry et al. (1992), who reported
a standard deviation of 0.21 %0 based on analyses
of riverine sediments using an automated system
similar to the one used in our laboratory. Sediment
trap samples had higher standard deviations associated with them (:0: 0.4%0) due to their small sample size.
Several different standards were measured to
control the quality of the nitrogen isotopic measurements. At least one laboratory working standard (WST 1,2 or 3) was measured per five samples. Additionally, IAEA N-I and IAEA N-2 standards from the International Atomic Energy Agency
(Vienna, Austria) and peptone (Merck chemicals)
were measured frequently to check the reliability
of sample combustion and mass spectrometric
measurements.
Although it is possible to measure both8 13 C OCg
and 8 15 N on one sample, we found that with the
large sample sizes needed for 8 15 N measurement,
organic carbon was present at such high concentrations that the CO, produced during sample combustion could not be completely isolated from the
nitrogen gas using the cryogenic method of
the trapping box. This resulted in unreliable isotopic
measurements. To circumvent this problem,
an extra cold trap immersed in liquid nitrogen was
added upstream ofthe trapping box to retain CO 2 ,
but let N, pass through to the trapping box.
This meant that stable carbon isotopic measurements had to be carried out on separate samples.
Samples for 8 J3 C OCg measurements were placed in
silver foil crucibles, decalcified with 6 M HCI and
dried on a hot plate at approximately 80°C.8 13 C OCg
analyses were performed using a Finnigan MAT
delta E mass spectrometer coupled to a Heraeus
CHN elemental analyzer and automated trapping
box. The standard deviation was < 0.1 %0 based on
duplicates and repeated measurements of the reference sediment "WST I".
Study Area
The Benguela region, off the southwest coast of
Africa, is comprised of the Angola Basin in the
north and the Cape Basin in the south. These two
hydrographically and productively distinct areas are
separated by a topographic high (Walvis Ridge) and
a hydrographic feature known as the AngolaBenguela Front (ABF). The warm, low nutrient
South Equatorial Countercurrent (SECC) flows
southeastward toward the Angola Basin at subsurface depths until it reaches the surface at around
100S and continues southward as the Angola Current (AC) (Moroshkin et al. 1970; Gordon and
Bosley 1991). The Benguela Current (BC), the
eastern boundary current of the South Atlantic
subtropical gyre, brings cold surface water into the
Cape Basin. The BC splits into the Benguela Oceanic Current (BOC) and the Benguela Coastal
Current (BCC) at around 28°S (Stramma and
Peterson 1989). The BOC flows offshore toward