1-All Sites Pooled

-Except 4 sites (S1,S4, L1,I1) with missing data for early sampling. The following 3 plots are just different flavours of the same graph:

From here on, plots show a smooth conditional mean with standard error:

2-Small Fragments

(except S1 &S4)

3- Large Fragments

(except L1)

4- Continuous Sites

(except I1. Same legend as above graphs)

5-Size effect

(both leaflitter types. From here on, removed Std.error to see slopes better)

6- Size effect by leaflitter type

##…Next steps, model exploration: Note: Very general model using Time as continuous variable and including Day0=>Final.mass=2g

M1 <- lme(log(Final.mass) ~ Days*Size.new*Leaves, data=deczero, random = ~1|Site,method="ML")
plot(M1)

anova(M1,test="Chisq")
##                      numDF denDF  F-value p-value
## (Intercept)              1   594 8418.048  <.0001
## Days                     1   594 1618.706  <.0001
## Size.new                 2    16    1.163  0.3376
## Leaves                   1   594   53.715  <.0001
## Days:Size.new            2   594    1.209  0.2993
## Days:Leaves              1   594   53.657  <.0001
## Size.new:Leaves          2   594    0.109  0.8964
## Days:Size.new:Leaves     2   594    0.735  0.4798