Ecological Archives E096150A1
Thomas L. Anderson and Howard H. Whiteman. 2015. Asymmetric effects of intra and interspecific competition on a pondbreeding salamander. Ecology 96:1681–1690. http://dx.doi.org/10.1890/140479.1
Appendix A. Additional figures and tables of experimental design, model selection results, and parameter estimates.
Fig. A1. Density combinations of the response surface design. All treatments had five replicates, except for the 12:0 and 6:0 (A. talpoideum: A. maculatum) treatments which had four replicates due to erroneous mixing of animals resulting in their exclusion from analyses.
Multinomial Regression Results
Table A1. Multinomial regression models that compared the production (# of individuals) of each phenotype in the mole salamander (Ambystoma talpoideum). "Spotted" refers to the spotted salamander
(A. maculatum). The metamorph phenotype was set as the reference level. ω is the AICc weight.
Model 
Covariates 
ΔAICc 
ω 
Additive 
Mole Density+Spotted Density 
0.0 
0.76 
Interactive 
Mole+Spotted+Mole:Spotted 
2.3 
0.24 
Overall 
Overall Density 
12.7 
<0.001 
Conspecific 
Mole Density 
38.4 
<0.001 
Spotted Date 
Mean Date of Spotted Metamorphosis 
132.3 
<0.001 
Spotted Survival 
Spotted Survival 
136.2 
<0.001 
Null 
Null (intercept only) 
136.7 
<0.001 
Heterospecific 
Spotted Density 
139.9 
<0.001 
Table A2. Parameter estimates, standard error and 95% confidence intervals (lower and upper) from the top model of the multinomial regression (Model 1 in Table A1), which included terms of the initial abundance of mole (Ambystoma talpoideum; "Mole") and spotted salamanders (A. maculatum; "Spotted"). The metamorph phenotype was set as a reference level.
Covariate 
Estimate 
SE 
lower 
upper 
larva(intercept) 
3.68 
0.54 
4.74 
2.62 
paedomorph (intercept) 
1.10 
0.46 
0.2 
2.01 
larva:Mole 
0.19 
0.03 
0.14 
0.24 
paedomorph:Mole 
0.19 
0.04 
0.27 
0.11 
larva:Spotted 
0.15 
0.03 
0.09 
0.21 
paedomorph:Spotted 
0.05 
0.03 
0.11 
0.01 
Table A3. Model selection results comparing the Linear, Ricker, and Exponential competition models for body size of the three phenotypes of mole salamanders (Ambystoma talpoideum). *Ricker would not converge. ω is the AICc weight.
Phenotype 
Model 
ΔAICc 
df 
ω 
Metamorph 
Linear 
0 
10 
1 

Exponential 
28.7 
10 
<0.001 

Ricker 
188.2 
10 
<0.001 
OW Larvae 
Linear 
0 
10 
1 

Exponential 
129.7 
10 
<0.001 

Ricker 
320.2 
10 
<0.001 
Paedomorph* 
Linear 
0 
10 
0.9005 

Exponential 
4.4 
10 
0.0995 
Inividual Growth Rate Results
Table A4. Model selection results for growth rates of individually marked larvae of mole salamanders (Ambystoma talpoideum).
Model 
ΔAICc 
ω 
MichaelisMenten 
0.0 
1.0 
Power Law 
237.1 
<0.001 
Gompertz 
475.2 
<0.001 
Table A5. Parameter estimates for individual growth rates of each phenotype (metamorph, paedomorph, and overwintered (OW) larvae) of mole salamanders (Ambystoma talpoideum). Abundance is the initial number of A. talpoideum. a is the asymptote, b is the halfmaximum scale parameter, and b*a1 is the initial slope. The lower and upper CI are 95% confidence intervals. "n" is the total sample size (number of individuals from all tanks with that density).
a 
b 
b*a^{1} 

Phenotype 
Initial tank 
n 
lower CI 
estimate 
upper CI 
lower CI 
estimate 
upper CI 
estimate 
Metamorph 
3 
5 
52.61 
57.46 
62.30 
35.94 
45.51 
55.07 
0.79 

6 
19 
41.19 
51.44 
61.69 
12.88 
33.07 
53.27 
0.64 

9 
10 
33.45 
43.79 
54.14 
4.29 
24.76 
45.22 
0.57 

12 
15 
41.11 
51.42 
61.73 
18.25 
38.71 
59.18 
0.75 

18 
6 
36.52 
47.55 
58.58 
15.44 
37.98 
60.51 
0.80 

24 
6 
34.12 
45.36 
56.61 
12.41 
35.30 
58.19 
0.78 
OW Larva 
6 
5 
48.64 
53.11 
57.58 
31.24 
39.11 
46.98 
0.74 

12 
6 
32.26 
42.75 
53.24 
7.12 
25.52 
43.91 
0.60 

18 
3 
31.71 
42.85 
54.00 
9.78 
29.48 
49.17 
0.69 

24 
7 
26.97 
37.26 
47.55 
1.55 
19.68 
37.81 
0.53 
Paedomorph 
3 
5 
62.09 
69.13 
76.17 
45.81 
56.79 
67.76 
0.82 

6 
5 
37.72 
54.61 
71.50 
10.04 
36.05 
62.06 
0.66 

12 
4 
39.09 
56.58 
74.06 
15.85 
43.20 
70.54 
0.76 