Page 13 - Volume 2, Issue 1, January 2009
Basic HTML Version
Troyanos
et al.
12
© Benaki Phytopathological Institute
od (4).
[NO
3
-
- N] measurements were carried
out with the standard method after extract-
ing each sub-sample with 250 ml de-ion-
ized water for 0.5 hour. The extracts were fil-
tered using a 125 mm filter paper which had
been rinsed with 25 ml of de-ionized water.
After filtration, 0.75 ml of a solution contain-
ing 1 M NaOH and 0.12 M Na
3
PO
4
.12H
2
O was
added per 25 ml of extract to remove the Ca
and Mg ions (5). Afterwards, a volume of 2
ml extract was taken from the clear super-
natant and used for colorimetric determina-
tion of [NO
3
-
-N] (5). If the results of the anal-
yses of the two sub-samples differed more
than 10%, a new sub-sample was taken and
analyzed for a third time and the means of
the three measurements were used. Mea-
surements with the “test strip” method were
carried out according to Hartz
et al.
after ex-
tracting the soil with 0.01 M CaCl
2
(4).
Regression analysis was performed be-
tween the [NO
3
-
-N] determined by the “test
strip” and the standard method (Figure 1).
The analysis showed that a linear curve was
significant (P<0.001) with a high coefficient
of determination (R
2
=0.85, n=67). The slope
of the line was estimated to be 0.953, it was
different from zero (P<0.001) and had a stan-
dard error (SE) of 0.053. The estimated con-
stant was 2.408, it was different from zero
(P<0.001) and had a SE 0.919 (P=0.011). Based
on this analysis, themean [NO
3
-
-N] measured
by the standard method was 14.22 ppm and
the mean [NO
3
-
-N] measured by the “test
strip” method was 15.96 ppm showing an
overestimation of approximately 12.23 % by
the “test strip” method. From Figure 1 it is
evident that the constant of the regression
equation causes the regression line to be
different from the line (y=x) which describes
the agreement between the two methods.
To evaluate the differences between
the two methods the procedure of Bland
and Altman (3) was used. According to this
method the plot of the differences between
the [NO
3
-
-N] determined with the two meth-
ods against their means are indicative of
their discrepancies. The mean of the differ-
ences ( ) (“test strip” – standard method)
was 1.74 ppm and the standard deviation
(SD) of the differences was 3.99 (Figure 2).
Therefore, 95% of the differences between
the methods are expected to lie between
±1.96 SD. From Figure 2 it is evident that the
upper limit of the differences between the
methods could be 9.56 ppm and the low
limit 6.08 ppm. Therefore, the maximum dif-
ferences that could be expected to occur
when the “test-strip” method is used in the
field are 10 ppm above or 6 ppm [NO
3
-
-N]
below the standrd method. A 10 ppm over-
estimation corresponds to approximately 30
kg of N per ha (0-30 cm soil depth) which is
acceptable and therefore, the “quick test”
method could be used reliably to monitor
the “in-season” soil N status of drip irrigated
processing tomatoes.
Figure 1.
Regression analysis between the “test strip” and
the standard method and the line of agreement (y=x) be-
tween the methods.
Figure 2.
[NO
3
-
- N] (ppm) differences between the methods
vs. average values of the methods for each composite soil sam-
ple. ( = mean of the differences and SD = standard devia-
tion).
