Agriculture Reference
In-Depth Information
Activities of nine soil enzymes indicative of C-cycling (α-galactosidase,
β-glucosidase), C/N-cycling (β-glucosaminidase), N-cycling (aspartase,
l-asparaginase, urease), P-cycling (acid phosphatase, alkaline phospho-
monoesterase) and S-cycling (arylsulfatase) were evaluated. These en-
zyme activities were assayed using 1 g of air-dried soil with their appropri-
ate substrate and incubated for 1 h (37 °C) at their optimal pH as described
by Tabatabai (1994) and Parham and Deng (2000). Enzyme activities were
assayed in duplicate with one control, to which substrate was added after
incubation and subtracted from the sample value.
13.2.4 STATISTICAL ANALYSES
Box plots were graphed using the default settings from the
ggplot2
package
in R (Wickham, 2009 and R Development Core Team, 2012). The horizontal
line is the mean, and upper and lower “hinges” are the first and third quartiles,
respectively. Upper and lower “whiskers” extend to the highest or lowest val-
ue, respectively, within 1.5 times the inter-quartile range (the distance between
the first and third quartiles). Data beyond this range are plotted as points.
Linear relationships between selected soil physicochemical character-
istics were tested with the
lm()
function in R using fi eld averages (i.e. n
= 13) in order to examine broad relationships across the fi elds. Field was
considered an explanatory factor (13 levels) in analysis of variance (ANO-
VA) models with separate analyses for each response variable, including
9 enzyme activities and 9 taxonomic groups compiled from indicator FA-
MEs. All models were statistically signifi cant at the p < 0.001 level. F-
statistics, i.e. the ratio of variance among fi elds to variance within fi elds,
derived from these analyses were used to compare the relative magnitude
of the fi eld effect for each variable. Hence, greater between fi eld vari-
ability for a given enzyme (or taxon) relative to other enzymes (or taxa) is
refl ected by a larger F-statistic.
Principal components analysis (PCA) was performed in the vegan
package in R ( Oksanen et al., 2012) using a correlation matrix. Compo-
nent scores for each of the six plots within a fi eld were used to generate
95% confi dence ellipses around each fi eld using the
ordiellipse()
function
within
vegan.
PCA of FAMEs used data expressed as mol percent.
