Determinants of inter-specific variation in basal metabolic rate

Abstract

Basal metabolic rate (BMR) is the rate of metabolism of a resting, postabsorptive, non-reproductive, adult bird or mammal, measured during the inactive circadian phase at a thermoneutral temperature. BMR is one of the most widely measured physiological traits, and data are available for over 1,200 species. With data available for such a wide range of species, BMR is a benchmark measurement in ecological and evolutionary physiology, and is often used as a reference against which other levels of metabolism are compared. Implicit in such comparisons is the assumption that BMR is invariant for a given species and that it therefore represents a stable point of comparison. However, BMR shows substantial variation between individuals, populations and species. Investigation of the ultimate (evolutionary) explanations for these differences remains an active area of inquiry, and explanation of size-related trends remains a contentious area. Whereas explanations for the scaling of BMR are generally mechanistic and claim ties to the first principles of chemistry and physics, investigations of mass-independent variation typically take an evolutionary perspective and have demonstrated that BMR is ultimately linked with a range of extrinsic variables including diet, habitat temperature, and net primary productivity. Here we review explanations for size-related and mass-independent variation in the BMR of animals, and suggest ways that the various explanations can be evaluated and integrated.

Appendices

Appendix 1: Phylogenetic methods for the analysis of the effect of diet on avian basal metabolic rate

The relationship between log transformed basal metabolic rate (BMR), log transformed body mass (M) and dietary categories was analysed using phylogenetic generalised least squares (PGLS) (Grafen 1989; Martins and Hansen 1997; Garland and Ives 2000) in the analysis of phylogenetics and evolution (APE) package (Paradis et al. 2004) within R (Ihaka and Gentleman 1996) according to established procedures (Halsey et al. 2006; Duncan et al. 2007; White et al. 2009). Data for avian BMR matched to a phylogenetic hypothesis were obtained from a published analysis of the scaling of BMR (Kabat et al. 2008), and were matched to dietary categories provided by McNab (2009). Matched BMR and diet data were available for a total of 287 species. Since the true branch lengths in the phylogeny are unknown, two branch length assumptions were compared: all branches set equal to 1, and an alternative assumption that branch lengths were proportional in length to the number of taxa descended from the node to which the branch leads (Grafen 1989). A measure of phylogenetic correlation, λ (Pagel 1999; Freckleton et al. 2002), was estimated by fitting PGLS models with different values of λ and finding the value that maximises the log likelihood. The degree to which trait evolution deviates from Brownian motion (λ = 1) was accommodated by modifying the covariance matrix using the maximum likelihood value of λ, which is a multiplier of the off-diagonal elements of the covariance matrix (i.e., those quantifying the degree of relatedness between species). All models were compared on the basis of Akaike’s information criterion (AIC) as a measure of model fit (Burnham and Anderson 2001, 2002). The relative support of alternative models was compared on the basis of Δ _i (=AIC − minimum AIC); models having Δ _i  ≤ 2 have substantial support, those where 4 ≤ Δ _i  ≤ 7 have considerably less support, while models having Δ _i  > 10 have essentially no support (Burnham and Anderson 2001).

Appendix 2: Methods for the generation of an allometric association between metabolic rate and body mass

The model for allometric scaling is based on Monte Carlo simulations developed to understand the causes of the observed right-skewed lognormal distribution of mammalian body masses (Maurer et al. 1992; Blackburn and Gaston 1994, 1998, 1999). Initially, 400 ‘species’ with a mass (M) of 1 and a metabolic rate (MR) of 1 were generated. For each species, a random change in M was then generated by multiplying M by a normal deviate with a mean of 0 and standard deviation of 0.02 and then adding M. This was then repeated a total of 5,000 times for each ‘species’. Thus, for each of the 5,000 time steps, mass varied randomly with a standard deviation of 2 % of the value of M at the previous time step. Because the genetic correlation between MR and M is positive and often <1 (Table 7), factorial changes in MR at each time step were randomly smaller than the changes in MR (see e.g. Fig. 4). This procedure generates lognormal distributions of M and MR, consistent with the idea that body size evolves multiplicatively, and could be made more realistic by the introduction of size-biased selection and extinction, and anagenetic size change within species between speciation and extinction events (e.g. Stanley 1973; Maurer et al. 1992; Kingsolver and Pfennig 2004; Clauset and Erwin 2008; Mattila and Bokma 2008; Clauset et al. 2009). The consequences of variation in MR for allometric scaling could be examined by including selections against low (e.g. Jackson et al. 2001) or high (e.g. Artacho and Nespolo 2009) MR.