Ecological Archives E096-220-A1
Michael P. Moore, Tobias Landberg, and Howard H. Whiteman. 2015. Maternal investment mediates offspring life history variation with context-dependent fitness consequences. Ecology 96:2499–2509. http://dx.doi.org/10.1890/14-1602.1
Appendix A. Detailed methods used for reassigning tanks to density treatments.
Re-binning Methods
Preliminary analyses of growth, using the maximum number of larvae captured from each tank as a continuous estimate of density (Table A1), indicated strong density effects (all χ² > 12.0, df = 1, P < 0.001; see Results). However, data visualization indicated clustering of responses across the range of densities, and thus densities were reassigned to one of three binned densities according to clustering. To ensure that these new densities better reflected the environmental conditions under which the larvae grew, analyses of all responses significantly affected by density were run separately using the initial, continuous, and binned densities, in addition to the other main effects and interactions, and the model fits were compared using AIC corrected for small sample sizes (Table A2; Burnham and Anderson 2002).
Table A1. Re-assignment of tank density treatments. The experiment was designed as a 3 × 3 factorial replicated in six spatial blocks. Substantial mortality resulted in no tanks exhibiting their initial densities, and densities were re-assigned to reflect more realistic larval growth environments. A tank’s maximum known density was the maximum number of individuals ever captured from a single tank during the larval growth period.
Block |
Embryonic Treatment |
Initial Density |
Maximum Known Density |
"Binned" Density |
A |
Reduced |
High |
15 |
High |
A |
Unmanipulated |
Low |
1 |
Low |
A |
Unmanipulated |
High |
3 |
Low |
A |
Surgical Control |
High |
8 |
Medium |
A |
Unmanipulated |
Medium |
2 |
Low |
A |
Reduced |
Low |
2 |
Low |
A |
Reduced |
Medium |
1 |
Low |
A |
Surgical Control |
Medium |
1 |
Low |
A |
Surgical Control |
Low |
4 |
Low |
B |
Surgical Control |
Medium |
6 |
Medium |
B |
Reduced |
Low |
3 |
Low |
B |
Reduced |
High |
0 |
NA |
B |
Unmanipulated |
Low |
2 |
Low |
B |
Reduced |
Medium |
0 |
NA |
B |
Surgical Control |
Low |
5 |
Medium |
B |
Surgical Control |
High |
16 |
High |
B |
Unmanipulated |
High |
9 |
High |
B |
Unmanipulated |
Medium |
6 |
Medium |
C |
Reduced |
High |
7 |
Medium |
C |
Surgical Control |
Low |
0 |
NA |
C |
Surgical Control |
Medium |
3 |
Low |
C |
Reduced |
Low |
3 |
Low |
C |
Surgical Control |
High |
14 |
High |
C |
Unmanipulated |
High |
0 |
NA |
C |
Unmanipulated |
Low |
1 |
Low |
C |
Unmanipulated |
Medium |
2 |
Low |
C |
Reduced |
Medium |
0 |
NA |
D |
Reduced |
Medium |
8 |
Medium |
D |
Reduced |
High |
16 |
High |
D |
Surgical Control |
High |
13 |
High |
D |
Surgical Control |
Medium |
2 |
Low |
D |
Unmanipulated |
Medium |
6 |
Medium |
D |
Reduced |
Low |
3 |
Low |
D |
Surgical Control |
Low |
4 |
Low |
D |
Unmanipulated |
Low |
1 |
Low |
D |
Unmanipulated |
High |
16 |
High |
E |
Reduced |
Low |
0 |
NA |
E |
Reduced |
Medium |
0 |
NA |
E |
Unmanipulated |
Low |
0 |
NA |
E |
Surgical Control |
Low |
4 |
Low |
E |
Surgical Control |
High |
13 |
High |
E |
Unmanipulated |
Medium |
1 |
Low |
E |
Unmanipulated |
High |
15 |
High |
E |
Reduced |
High |
13 |
High |
E |
Surgical Control |
Medium |
2 |
Low |
F |
Surgical Control |
Medium |
2 |
Low |
F |
Unmanipulated |
Medium |
9 |
High |
F |
Unmanipulated |
High |
0 |
NA |
F |
Surgical Control |
Low |
5 |
Medium |
F |
Unmanipulated |
Low |
2 |
Low |
F |
Reduced |
High |
6 |
Medium |
F |
Reduced |
Low |
4 |
Low |
F |
Surgical Control |
High |
12 |
High |
F |
Reduced |
Medium |
7 |
Medium |
|
||||
Table A2. AICc comparison of models fit with binned, initial, and continuous estimates of larval density. Density estimate receiving the best support, as indicated by the lowest AICc value, are shown in bold. Binned density models always fit the data better than initial or continuous densities, except in one circumstance where the binned density model fit equally well as the continuous density model.
Model Density |
AIC |
AICc |
ΔAICc |
|
Larval Growth |
||||
Binned |
318.50 |
320.22 |
0.00 |
|
Initial |
358.09 |
359.82 |
39.60 |
|
Continuous |
403.66 |
404.50 |
84.27 |
|
|
||||
Morph Expression |
||||
Binned |
381.78 |
386.93 |
0.48 |
|
Initial |
385.20 |
390.34 |
3.90 |
|
Continuous |
385.45 |
386.45 |
0.00 |
|
|
||||
Metamorph Size |
||||
Binned |
88.25 |
92.45 |
0.00 |
|
Initial |
97.94 |
102.14 |
9.70 |
|
Continuous |
94.30 |
96.12 |
3.67 |
|
|
||||
Reproductive Output |
||||
Binned |
65.22 |
67.55 |
0.00 |
|
Initial |
86.23 |
88.56 |
21.01 |
|
Continuous |
73.18 |
73.80 |
6.25 |
|
Literature Cited
Burnham, K. P., and D. R. Anderson. 2002. Model selection and inference: a practical information-theoretic approach. Springer-Verlag, New York, New York, USA.