Table 9. Statistics for discrete and stirred composite samples for the four subdivi-
sions.
White phosphorus conc. (g/g dry weight)
Discrete*
Stirred composite
samples (n = 12
Sediment
1:1 Sed:water
1:2 Sed:water
means of dups)
(n = 5)
(n = 5)
(n = 5)
Rows 1 to 3
Mean
0.0166
0.0271
0.0250
0.0226
2.37105
3.39106
8.02106
2.41104
Variance
Standard deviation
0.0155
0.00487
0.00184
0.00283
RSD (%)
93.0%
18.0%
7.35%
12.5%
Range
0.0042 to 0.0529
0.0242 to 0.0357
0.0229 to 0.0274
0.0191 to 0.0264
Median
0.0102
0.0250
0.0248
0.0228
Geometric mean
0.0116
0.0268
0.0250
0.0224
Rows 4 to 6
Mean
1.39
1.95
0.970
0.750
Variance
8.40
3.20
0.135
0.012
Standard deviation
2.90
1.79
0.367
0.109
RSD (%)
208%
91.6%
37.9%
14.5%
Range
0.0122 to 8.26
0.814 to 4.97
0.701 to 1.53
0.627 to 0.895
Median
0.101
0.892
0.741
0.737
Geometric mean
0.141
1.47
0.920
0.744
Rows 7 to 9
Mean
24.0
7.43
3.62
2.95
Variance
3555
46.4
3.61
5.19
Standard deviation
59.6
6.81
1.90
2.28
RSD (%)
248%
91.7%
52.4%
77.1%
Range
0.0240 to 206
2.18 to 19.2
1.96 to 6.41
1.84 to 7.02
Median
0.773
4.49
2.73
1.93
Geometric mean
1.01
5.63
3.26
2.5
Rows 10 to 12
Mean
58.9
31.0
31.5
15.6
Variance
34740
95.4
106
5.77
Standard deviation
186
9.77
10.3
2.40
RSD (%)
316%
31.5%
32.8%
15.4%
Range
0.0590 to 650
18.8 to 42.7
19.2 to 42.3
12.0 to 18.2
Median
0.913
30.8
36.3
16.5
Geometric mean
1.83
29.7
30.0
15.5
* The variances and the RSDs (%) are not strictly valid because of non-normality of the distribution,
but they are useful for comparison.
Sieved composite samples
each subdivision sampled. Both ANOVA and the
The ponded area we sampled was sparsely
Krusal-Wallis test indicated that the means were
vegetated. Despite the large amount of sediment
not significantly different within three of the four
subdivisions (p > 0.05) (Table 8). The significant
collected to form the composite samples (Table 3),
once they were sieved, the volume in all cases
difference was observed for the composite
was reduced to less than 100 mL. This volume of
formed from rows 10 to 12, where the white phos-
sediment can be analyzed without subsampling
phorus particles were most numerous and the
in the laboratory using headspace SPME. The
highest concentrations found.
SPME analysis detected a large amount of white
In terms of ranking the four subdivisions of the
phosphorus in each of the sieved composite sam-
area sampled by level of contamination, the dis-
ples collected.
crete sampling showed that the level of contami-
Originally, we planned to count white phos-
nation increased sequentially with the first subdi-
phorus particles by spreading the sieved compos-
vision (i.e., rows 1 to 3) having the least contami-
ite sample on a pan, heating, and counting
nation and the last subdivision (rows 10 to 12)
flames. This approach was used for the first sam-
having the most contamination. The mean com-
ple examined (2.33-m-square grid, rep. 1) (Table
posite concentrations ranked the subdivisions in
10); however, so many particles were present in
this same order with fewer analyses (Table 9).
13
Previous Page
