42
CORDON A. RILEY
in the water. Invariably the rate of production was smaller, and under
some experimental conditions the yield was negligible. Hence some
doubt began to arise as to the validity of the earlier work.
The matter became further confused by the fact that a variety of
techniques was used, with little knowledge as to the effects that they
might have on experimental results. Various kinds and pore sizes of
filters were employed. Plastic containers have been used because of
ease in fabrication, but they have been criticized on the grounds of
possible introduction of contaminants. Air supplies are always suspect.
Furthermore the quantity of carbon obtained is small, so that analytical
errors are troublesome. The above section on " Methods )' has outlined
precautions that are necessary for good analytical results ; this kind of
information has accumulated gradually over the years, and early
investigations did not include all of the precautions that are now regarded as necessary, nor did they always include enough replications
for good statistical tests of the results. However, the results of earlier
work should be summarized, and insofar as possible they will be reexamined in the light of more recent information about possible errors.
Riley et al. (1965, Table 5) described the results of bubbling experiments on surface and deep water from the Sargasso Sea. The water
was first filtered through 0.45 p Millipore filters in 15-20 litre lots into
polyethylene carboys, bubbled for 3-6 hours with filtered ambient air
supplied through an aquarium pump, and then re-filtered through
glass fibre filters for measurements of particulate carbon. The conclusion at that time was that all of these waters, surface and deep,
contained filter passing material which could be converted into particulate matter by the bubbling process.
The use of polyethylene and ambient air and the lack of unbubbled
controls have been criticized and raise some doubt as to absolute values ;
however, the whole series was done under essentially the same kind of
conditions, permitting internal comparison. There were 44 experiments in all, about equally divided between samples taken in the upper
200 m and those obtained at various depths from 500 m to 3 000 m.
The mean yields in these two series were 70 and 39 pg C/litre, respectively, and the difference was significant to the 1% level. Any errors
due to artifacts or lack of control should be about equal in the two
sets. Thus, while one cannot say with certainty that bubbling produced
particles in the deep water samples, the higher values obtained in the
surface layer were a definite indication that this process was significant,
even though the region under investigation is relatively poor in organic
productivity. The other question, whether organic matter in deep
water also can be converted into particles by the bubbling process,
CORDON A. RILEY
in the water. Invariably the rate of production was smaller, and under
some experimental conditions the yield was negligible. Hence some
doubt began to arise as to the validity of the earlier work.
The matter became further confused by the fact that a variety of
techniques was used, with little knowledge as to the effects that they
might have on experimental results. Various kinds and pore sizes of
filters were employed. Plastic containers have been used because of
ease in fabrication, but they have been criticized on the grounds of
possible introduction of contaminants. Air supplies are always suspect.
Furthermore the quantity of carbon obtained is small, so that analytical
errors are troublesome. The above section on " Methods )' has outlined
precautions that are necessary for good analytical results ; this kind of
information has accumulated gradually over the years, and early
investigations did not include all of the precautions that are now regarded as necessary, nor did they always include enough replications
for good statistical tests of the results. However, the results of earlier
work should be summarized, and insofar as possible they will be reexamined in the light of more recent information about possible errors.
Riley et al. (1965, Table 5) described the results of bubbling experiments on surface and deep water from the Sargasso Sea. The water
was first filtered through 0.45 p Millipore filters in 15-20 litre lots into
polyethylene carboys, bubbled for 3-6 hours with filtered ambient air
supplied through an aquarium pump, and then re-filtered through
glass fibre filters for measurements of particulate carbon. The conclusion at that time was that all of these waters, surface and deep,
contained filter passing material which could be converted into particulate matter by the bubbling process.
The use of polyethylene and ambient air and the lack of unbubbled
controls have been criticized and raise some doubt as to absolute values ;
however, the whole series was done under essentially the same kind of
conditions, permitting internal comparison. There were 44 experiments in all, about equally divided between samples taken in the upper
200 m and those obtained at various depths from 500 m to 3 000 m.
The mean yields in these two series were 70 and 39 pg C/litre, respectively, and the difference was significant to the 1% level. Any errors
due to artifacts or lack of control should be about equal in the two
sets. Thus, while one cannot say with certainty that bubbling produced
particles in the deep water samples, the higher values obtained in the
surface layer were a definite indication that this process was significant,
even though the region under investigation is relatively poor in organic
productivity. The other question, whether organic matter in deep
water also can be converted into particles by the bubbling process,