Enzyme activities are averaged if there is more than one enzyme in the group

(Sinsabaugh et al., 1997).

 Potential decomposition models were constructed from enzymic data using

normalized enzyme activity across both the benthic and soil layers. This was done in

order to compare different enzymes at the same scale so that the more active enzymes do

not heavily weight the calculations of model components. The Enzymic Index of Carbon

Quality (EICQ) compares the activities of oxidative to hydrolytic enzymes (Sinsabaugh

and Findlay, 1995). In addition, several other indices were formulated from the data.

Ecell/En reflects apparent N control over C mineralization, Ecell/Ep is a relative measure

indicating P control over C mineralization, and Ecell/Eox reflects apparent lignin control

over C mineralization.

 To improve normality and heteroscedescity, data were log transformed before

statistical analysis using SAS version 8 statistical software (SAS, 1999). A split-plot

model for the data was adopted for subsequent ANOVA analysis using PROC GLM in

SAS. The whole plot corresponded to the sites in reference to the nutrient gradient while

the subplots referred to the specific vegetative communities within the whole plots.

Multiple comparisons were made between habitats, sites, and habitat-site combinations.

Other analyses were performed using the Statistical Package for the Social Sciences (TM

SPSS) version 11.0.0. All regressions and significant differences are significant at the

p<.05 level unless otherwise noted.

 Results

Nutrient Composition

 Average nutrient concentrations for the benthic and soil layers with standard errors

are presented in Table 3-1. Benthic TP values were significantly different among sites