Ecological Archives E091-198-A3

Isabella Capellini, Chris Venditti, and Robert A. Barton. 2010. Phylogeny and metabolic rates in mammals. Ecology 20:2783–2793.

Appendix C. Additional figures of the residuals of the metabolic rates on body mass.

figC1a
 
figC1b
 

   FIG. C1. Residuals of basal (a) and field (b) metabolic rates on predicted values (empty circles = placental mammals; filled circles = marsupials; diamonds = monotremes) to check for curvilinearity in the PGLSλ regression model of metabolic rates on body mass in mammals (a random pattern should be observed). The figure shows considerable phylogenetic clustering in the residuals, confirming that relative metabolic rates exhibit a phylogenetic signal in mammals and therefore that also the analysis of residuals should be done with phylogenetically controlled statistical methods (see main text). The slight curvilinearity observed in the residuals is much attenuated once phylogeny is accounted for and, albeit significant, a quadratic model increases the R2 only marginally (0.003 for BMR and 0.022 for FMR) (see Methods and Results in main text). 


 

figC2a
 
figC2b
 

   FIG. C2. Phylogenetic independent contrast residuals of metabolic rates on body mass. There is no curvilinearity in this relationship, thus supporting the conclusion that lack of curvilinearity in OLS analysis is largely a phylogenetic artefact [(a) BMR in ‘true insectivores’ and (b) FMR in rodents]. PIC analysis was carried out only in those lineages that showed a λ = 1 for relative metabolic rates since only in this case PIC is equivalent to PGLSλ models (see Methods and Results in the main text).


[Back to E091-198]