Table 4-6. Soil layer nutrient and enzyme correlation coefficients. Values in bold
 are significant at the p<0.0001 level, all other values are significant at the
 p<0.05 level. Lig=lignin, Cell=cellulose. TN=total nitrogen, TP=total
 phosphorus, TOC=total organic carbon, C:P=ratio of TOC to TP, Ca=calcium,
 Mg=magnesium, GLU=glucosidase, Ecell/En= apparent N control on carbon
 mineralization, Ecell/Ep=apparent P control on carbon mineralization LEU=
 leucine aminopeptidase, PHO=phosphatase, PHE=phenol oxidase,
 CRR=cotton rottenness rate.


 Ecell/ Ecell/
 TP TN TOC Lig Ca Mg En Ep PHO
TP
TN .63
TOC .63 .99
Lig .48 .76 .79
Cell
Ca -.59 -.62 -.78
Mg
C:P .48 .51 .41
EC/EN .37 -.49
EC/EP .53 .61 .37 .62
BGL -.44 .93 .57
LEU -.69 -.54 -.39
PHO -.49 -.61 -.62 -.51 .57 -.81
PHE -.37 -.40 -.39 -.48 .57
CRR


 A model was created using enzyme data to test the validity of prediction against

actual CRR values. Ecell/Ep and TP were found to be the most powerful predictors of

CRR, accounting for 62% of the variability in the benthic layer. The resulting benthic

model: Predicted CRR= (0.04(LnEc:Ep)+1.4) + (0.35(LnTP)+1), was strongly correlated in

sampling period 1 (r2 =0.92) and more weakly correlated in sampling period 3 (r2=0.46).

This benthic model predicted significantly greater cellulose decomposition at the

enriched sites in all four areas at all time periods (p<0.0001). A separate significant

model could not be created for the soil layer. However, the benthic model predicted

significantly greater decomposition in the enriched sites in 75% of the contrasts.