76
Many investigators use both light and dark bottles for a 14C incubation. However, in this case, dark
bottles are a combined measure of particulate contamination in the tracer (a blank control that should
be subtracted) and dark carbon fixation, a real process for incorporating carbon into plankton that is
temporally independent of light. The careful application of killed controls or deep-water samples can
partially clarify this distinction. In careful incubations, much of the dark bottle uptake is dark carbon
fixation. For an estimate of the total carbon incorporation into the ecosystem, this rate should be
added to the light bottle rate. For an estimate of the
light-dependent production rate, it should be subtracted. Fortunately, the dark carbon fixation rate is
rarely more than 5% of the total integrated production, so, at the ecosystem level, it is a minor issue
compared with some of the larger concerns with the
accuracy and interpretation of the method.
As with the light and dark bottle oxygen method,
there are many problems with this technique regarding the scaling from volume to surface area,
the scaling from the time period of the incubation
to a daily rate, the sampling of the water, and the
isolating of the water sample in a bottle during incubations. These problems are discussed in the section on sampling issues below.
There are some problems or concerns specific to
the 14C02 method. One of these is that 14C-labeled
organic matter formed during the incubation is
sometimes released to solution, either excreted by
the living phytoplankton or lysed from cells during
the filtration. This results in an underestimation of
primary production, if as is usually the case, the
soluble 14C material is not counted. An alternative
to filtering the samples at the end of the incubation
is to acidify the sample or filtrate, bubble off the
remaining 14C that has not been fixed in primary
production, then sample some of the degassed water and count the 14C (which represents both particulate and dissolved 14C-labeled organic matter if it
was not filtered) on a scintillation counter. Unfortunately, the rather small volume that can be
counted in a scintillation results in a less sensitive
assay than can be obtained by concentrating the
particles from a larger volume onto a filter. Also,
care must be taken to ensure that no 14C-labeled
DIC remains in solution (dark or killed controls can
confirm this). The converse of this correction is that
it allows the explicit separation of the productivity
Robert W. Howarth and Anthony F. Michaels
into particulate and soluble, a valuable distinction
in ecosystems dominated by the microbial food
web. Even here, the interpretation will have a few
caveats as the glass-fiber filters commonly used for
these experiments adsorb some ofthe dissolved organic matter, and some of the excreted dissolved
organic matter will be respired to 14C-labeled DIC
by bacteria during the course of the incubation.
A conceptual problem with the 14C technique is
that it measures neither GPP nor NPP, but rather,
in many ecosystems, some rate between these two
(Peterson 1980; Bender et al. 1987; Collos et al.
1993). Simply stated, the method would measure
GPP if incubations were instantaneously short. As
an incubation starts, none of the organic matter
within phytoplankton is labeled with 14C, and so
none of the organic matter respired by phytoplankton is labeled with 14c. Therefore, in some types of
ecosystems, a very short incubation measurement
can sometimes approach an estimate of GPP (Collos et al. 1993), particularly in systems with high
light irradiance (Sand-Jensen and Krause-Jensen
1997). However, as the incubation proceeds in
time, more and more of the phytoplankton biomass
is labeled with the radioisotope, and so an increasing percentage of the phytoplankton respiration includes respiration of the 14C-labeled tissues. That
is, as the incubation time increases, the net rate of
14C uptake decreases from the rate of GPP and
moves toward the rate of NPP. If phytoplankton
were not grazed and the extracellular products of
photosynthesis were not consumed by bacteria,
then eventually a steady state would be approached
where the phytoplankton biomass would be uniformly labeled and the amount of 14C02 respired
would be a constant percentage of the total CO 2
respired. At this point, the 14C method would be
measuring NPP (assuming that the incubation included the nighttime dark cycle). However, it is unclear how long a time is required for an incubation
to truly represent a measure of NPP, and the necessary length of time for this to occur probably varies among ecosystems, being faster in ecosystems
with higher rates of turnover of the phytoplankton
biomass. Zooplankton grazing on phytoplankton
and bacterial consumption of dissolved organic carbon aggravate the interpretation of the 14C method,
for with longer incubations, these processes increasingly include the ingestion and respiration of
labeled 14C. Thus, long incubations may tend to-
Many investigators use both light and dark bottles for a 14C incubation. However, in this case, dark
bottles are a combined measure of particulate contamination in the tracer (a blank control that should
be subtracted) and dark carbon fixation, a real process for incorporating carbon into plankton that is
temporally independent of light. The careful application of killed controls or deep-water samples can
partially clarify this distinction. In careful incubations, much of the dark bottle uptake is dark carbon
fixation. For an estimate of the total carbon incorporation into the ecosystem, this rate should be
added to the light bottle rate. For an estimate of the
light-dependent production rate, it should be subtracted. Fortunately, the dark carbon fixation rate is
rarely more than 5% of the total integrated production, so, at the ecosystem level, it is a minor issue
compared with some of the larger concerns with the
accuracy and interpretation of the method.
As with the light and dark bottle oxygen method,
there are many problems with this technique regarding the scaling from volume to surface area,
the scaling from the time period of the incubation
to a daily rate, the sampling of the water, and the
isolating of the water sample in a bottle during incubations. These problems are discussed in the section on sampling issues below.
There are some problems or concerns specific to
the 14C02 method. One of these is that 14C-labeled
organic matter formed during the incubation is
sometimes released to solution, either excreted by
the living phytoplankton or lysed from cells during
the filtration. This results in an underestimation of
primary production, if as is usually the case, the
soluble 14C material is not counted. An alternative
to filtering the samples at the end of the incubation
is to acidify the sample or filtrate, bubble off the
remaining 14C that has not been fixed in primary
production, then sample some of the degassed water and count the 14C (which represents both particulate and dissolved 14C-labeled organic matter if it
was not filtered) on a scintillation counter. Unfortunately, the rather small volume that can be
counted in a scintillation results in a less sensitive
assay than can be obtained by concentrating the
particles from a larger volume onto a filter. Also,
care must be taken to ensure that no 14C-labeled
DIC remains in solution (dark or killed controls can
confirm this). The converse of this correction is that
it allows the explicit separation of the productivity
Robert W. Howarth and Anthony F. Michaels
into particulate and soluble, a valuable distinction
in ecosystems dominated by the microbial food
web. Even here, the interpretation will have a few
caveats as the glass-fiber filters commonly used for
these experiments adsorb some ofthe dissolved organic matter, and some of the excreted dissolved
organic matter will be respired to 14C-labeled DIC
by bacteria during the course of the incubation.
A conceptual problem with the 14C technique is
that it measures neither GPP nor NPP, but rather,
in many ecosystems, some rate between these two
(Peterson 1980; Bender et al. 1987; Collos et al.
1993). Simply stated, the method would measure
GPP if incubations were instantaneously short. As
an incubation starts, none of the organic matter
within phytoplankton is labeled with 14C, and so
none of the organic matter respired by phytoplankton is labeled with 14c. Therefore, in some types of
ecosystems, a very short incubation measurement
can sometimes approach an estimate of GPP (Collos et al. 1993), particularly in systems with high
light irradiance (Sand-Jensen and Krause-Jensen
1997). However, as the incubation proceeds in
time, more and more of the phytoplankton biomass
is labeled with the radioisotope, and so an increasing percentage of the phytoplankton respiration includes respiration of the 14C-labeled tissues. That
is, as the incubation time increases, the net rate of
14C uptake decreases from the rate of GPP and
moves toward the rate of NPP. If phytoplankton
were not grazed and the extracellular products of
photosynthesis were not consumed by bacteria,
then eventually a steady state would be approached
where the phytoplankton biomass would be uniformly labeled and the amount of 14C02 respired
would be a constant percentage of the total CO 2
respired. At this point, the 14C method would be
measuring NPP (assuming that the incubation included the nighttime dark cycle). However, it is unclear how long a time is required for an incubation
to truly represent a measure of NPP, and the necessary length of time for this to occur probably varies among ecosystems, being faster in ecosystems
with higher rates of turnover of the phytoplankton
biomass. Zooplankton grazing on phytoplankton
and bacterial consumption of dissolved organic carbon aggravate the interpretation of the 14C method,
for with longer incubations, these processes increasingly include the ingestion and respiration of
labeled 14C. Thus, long incubations may tend to-