Ecological Archives E096118A2
Qiang He, Andrew H. Altieri, and Baoshan Cui. 2015. Herbivory drives zonation of stresstolerant marsh plants. Ecology 96:1318–1328. http://dx.doi.org/10.1890/140937.1
Appendix B. Summaries of the results of the generalized linear models testing treatment effects in the pot, field herbivory and common garden experiments, respectively.
Table B1. Summary of the results of the generalized linear models (GLM) testing the effects of salinity (S), waterlogging (W), and competition (C) on the density and biomassof Salicornia and Suaeda in the pot experiment. The experiment had three levels of salinity (30, 60, 90 PSU), two levels of waterlogging (waterlogged and nonwaterlogged), and three levels of competition (Salicornia only, Suaeda only, and both species). The GLMs were conducted using quasiPoisson distribution with log link for density and Gamma distribution with inverse link for biomass. The effects were tested by comparing the resulting deviances to Wald χ2 test statistics using Type II sum of squares.


Salicornia 

Suaeda 

df 
Wald χ2 
P 
Wald χ2 
P 

Density 





C 
1, 60 
9.45 
0.0023** 
0.00 
0.98 
S 
2, 60 
3.54 
0.18 
220.52 
< 0.0001*** 
W 
1, 60 
3.35 
0.07 
0.10 
0.75 
C:S 
2, 60 
8.03 
0.018* 
7.65 
0.0213* 
C:W 
1, 60 
0.51 
0.46 
1.47 
0.22 
S:W 
2, 60 
13.01 
0.0016** 
23.85 
<0.0001*** 
C:S:W 
2, 60 
1.98 
0.40 
2.10 
0.32 
Biomass + 0.01 

C 
1, 60 
1.69 
0.19 
5.24 
0.0221* 
S 
2, 60 
64.33 
<0.0001*** 
46.44 
<0.0001*** 
W 
1, 60 
28.06 
<0.0001*** 
0.14 
0.71 
C:S 
2, 60 
1.12 
0.57 
5.89 
0.0526 
C:W 
1, 60 
1.64 
0.20 
0.70 
0.40 
S:W 
2, 60 
1.31 
0.52 
8.12 
0.0172* 
C:S:W 
2, 60 
1.14 
0.57 
6.44 
0.0399* 
***, P < 0.001; **, P < 0.01; and *, P < 0.05.
Table B2. Summary of the results of the GLMs testing the effects of crab exclusion, neighbor removal, and zone on Salicornia and Suaeda survival in the field herbivory experiment. The experiment had two levels of crab exclusion (control and exclusion), two levels of neighbor removal (removal and control), and two zones (Salicornia zone and Suaeda zone). The GLMs were conducted using quasibinomial distribution with logit link. The effects were tested by comparing the resulting deviances to Wald χ2 test statistics using Type II sum of squares.
Salicornia 
Suaeda 

Effect 
df 
Wald χ2 
P 
Wald χ2 
P 
Zone 
1, 40 
1.20 
0.27 
18.91 
<0.0001*** 
Crab 
1, 40 
31.26 
<0.0001*** 
11.22 
<0.0008*** 
Neighbor 
1, 40 
0.50 
0.48 
0.88 
0.35 
Zone:Crab 
1, 40 
0.00 
1.00 
5.02 
0.025* 
Zone:Neighbor 
1, 40 
4.89 
0.027* 
3.39 
0.07 
Crab:Neighbor 
1, 40 
0.33 
0.56 
0.03 
0.87 
Zone:Crab:Neighbor 
1, 40 
0.05 
0.82 
0.13 
0.71 
***, P < 0.001; **, P < 0.01; and *, P < 0.05.
Table B3. Summary of the results of the GLMs testing the effects of waterlogging and crab addition on Suaeda survival and biomass in the common garden experiment. The experiment had two levels of crab addition (addition and no addition) and two levels of waterlogging (waterlogged and nonwaterlogged). The GLMs were conducted using quasibinomial distribution with logit link for survival and Gamma distribution with inverse link for biomass. The effects were tested by comparing the resulting deviances to Wald χ2 test statistics using Type II sum of squares.
Survival 
Biomass + 0.01 

Effect 
df 
Wald χ2 
P 
Wald χ2 
P 
Waterlogging 
1, 20 
11.68 
0.0006*** 
2.37 
0.12 
Crab addition 
1, 20 
29.22 
<0.0001*** 
0.83 
0.36 
Waterlogging: Crab addition 
1, 20 
11.75 
0.0006*** 
17.49 
<0.0001*** 
***, P < 0.001; **, P < 0.01; and *, P < 0.05.