eLearning

used in this study make it difficult to unravel some
of the pieces of the N puzzle.
The N content of the aboveground portions of the
trees at 5 years was approximately 380 kg N/ha for
pitch pine and 420 kg N/ha for red pine. Only four
trees were sampled in each plot to provide these
estimates, so the 95% confidence intervals around
the estimates were very large (6175 kg N/ha for
pitch pine, 6255 kg N/ha for red pine). At year 1,
the 0–20-cm-depth soil was sampled in these plots,
but not the deeper soil. At year 5, the authors
resampled the upper soil and sampled the lower soil
for the first time. The 0–20-cm depth lost approxi-
mately 75 kg N/ha under pitch pine and 65 kg N/ha
under red pine. They estimated the 20–135-cm-
depth soil gained 365 kg N ha-1
(pitch pine) to 415
kgNha-1
(red pine), because the soil at this depth
contained more N than the same depth soil that, in
fact, had been sampled in a different, unvegetated
sandbox at the beginning of the study. The uncer-
tainty about the estimate of change in N content
was large: estimated N accretion of 290 under pitch
pine had a 95% confidence interval that was larger
than the mean (390 kg N/ha) because only four
trees were sampled in each sandbox. Similarly, the
accretion estimate of 350 kg N/ha under red pine
had a 95% confidence interval of 470 kg N/ha.
Despite the assumption that the 20–135-cm-depth
soil was originally similar among sandboxes, the net
change in soil N was not significantly different from
0 under either pine species. The N content of the
trees was significantly greater than 0, but the esti-
mate has low precision. The overall net accretion
estimate for the red pine sandbox was 390 kg N/ha
in the soil (nonsignificant) plus 380 kg N/ha in
aboveground pine biomass, for 770 kg N/ha, or
approximately 150 kg N ha-1
y-1
of occult N accre-
tion. The pitch pine stand had a net accretion of 680
kg N/ha, or 130 kg N ha-1
y-1
. The authors did not
have high confidence in the change in soil N from
20–135–cm depth and restricted their estimate of
occult N input to the vegetation plus 0–20-cm soil
depth. Their final estimate was approximately 54 kg
Nha-1
y-1
of occult N accumulation, with perhaps
5–10 kg N ha-1
y-1
expected from atmospheric N
deposition.We also note that the unvegetated sand-
box appeared to lose 290 kg N/ha from the
0–20-cm depth and to gain 150 kg N/ha in the
20–135-cm depth, but there was high variance in
the change estimate for the lower soil (95% confi-
dence interval was three times the mean). The au-
thors conclude they had an occult loss of N on the
order of 90 kg N ha-1
y-1
from the 0–20-cm depth of
the unvegetated sandbox, but much of this simply
may have resided in the uncertainty in the estimate
of N content of the deeper soil.
The sandbox forests were assayed for N fixation
by bacteria in the rhizosphere of the pine trees at
the end of the experiment. They did not report the
number of acetylene reduction assays, but the high-
est rate measured was 1.8 nmol of ethylene pro-
duction per gram of root per hour; they concluded
that rapid N fixation had occurred in the pine sand-
boxes. However, even if this maximum observed
rate represented a true average for the ecosystems,
total annual N fixation would be only on the order
of 0.02 kg N ha-1
y-1
(24 hours/day, fine root bio-
mass of 3000 kg ha-1
, growing season of 180 days,
molar conversion ratio of 1:3 of a mole of N2 fixed
per mole C2H2 reduced). This extrapolation of very
low N fixation in the rhizosphere of pine roots also
was supported by Barkmann and Schwintzer
(1998), who measured acetylene reduction activity
in cores of the forest floor and 15-cm-depth mineral
soil beneath 18 stands of pitch pine, red pine and
white pine stands in Maine. They estimated that
less than 0.06 kg N ha-1
y-1
was fixed in these soil
horizons.
Overall, we conclude that the sandbox approach
was a good idea, but that low precision (in the esti-
mates of the sandbox components, particularly the
vegetation and the 20–135-cm soil) precludes high
confidence in the authors’ estimate of occult N inputs.
Hubbard Brook Watershed 5
Johnson (1995) and coworkers undertook the
daunting task of attempting to measure changes in