A recent article by Berlin et al. (2007) (also reviewed by Marais, 2007) reported that the low mitochondrial variability in birds (relative to mammals) is most easily explained by Hill–Robertson effects on the W chromosome. These authors suggest selection and linkage between the W chromosome (in heterogametic females) and mtDNA act to lower avian mtDNA diversity. This may well be correct; however, a well-known biological phenomenon, which appears unique to birds, was excluded from these analyses, and may in fact provide a much simpler and more plausible explanation.

By comparison to mammals, birds are remarkable in several physiological parameters such as athletic performance, capacity to regenerate neuronal damage and their high respiratory efficiencies. However, perhaps more remarkable are several metabolic avian features. Birds have metabolic rates that are 2- to 2.5-fold greater, and estimated lifetime energy expenditures 15 times that of mammals of equivalent body mass. Not only do birds maintain body temperatures 3 °C hotter than mammals, but many birds have blood glucose levels 2–4 times that of mammals, which in part defines them as diabetic (Holmes et al., 2001). This last feature occurs without the associated pathological complications seen in mammals and highlights considerable protection from oxidative damage.

With few exceptions, birds are also very long living relative to body mass, as many birds live three times longer, or more, than mammals of equivalent mass, and birds age slower at the cellular level (Holmes et al., 2001). Parrots may live over 100 years, and even the tiny 5 g broad-billed hummingbird (Selasphorus platycercus) can live for 14 years (Holmes and Austad, 1995), while the maximum recorded lifespan of a 20 g house mouse is only 4 years (Holmes et al., 2001). Zoo and wild-tagging data also can mostly eliminate confounding influences of reduced predation through flight (Ricklefs, 1998), and the ‘exception species’, which are generally domesticated species, still live relatively long (for example, chickens 20 years, Cortunix quail 6–7 years) (Holmes and Austad, 1995).

Just how birds achieve such exemplary resistance to age appears to be largely explained by different mitochondrial properties. In mammals, 2–4% of all consumed oxygen is released as reactive oxygen species (ROS, for example, superoxide O2−, hydroxyl radical OH−, hydrogen peroxide H2O2) from the electron transport system (ETS) complexes I and III. In health (and more so with numerous pathologies), mitochondria are generally the largest source of ROS (Turrens, 2003). Avian mitochondria produce considerably less ROS than mammalian mitochondria, with pigeon liver and heart mitochondria producing up to 10-fold less H2O2 than rats (this depends on respiration state, and ROS predominates from complexes I and III as O2−, which is converted to H2O2; Herrero and Barja, 1997; Barja, 2004). Furthermore, parrots and canaries show considerable resistance to lipid peroxidation relative to rodents, and isolated kidney epithelial cells from other long-lived bird species are much more resilient to pro-oxidant challenge by paraquat, H2O2 and 95% O2 with markedly less DNA damage than mouse cells (Ogburn et al., 1988). The differences between birds and mammals should not also be assumed to be adaptive. ROS release is not necessarily a by-product of less efficient ETS function, as ROS provides feedback to cells and mitochondria (Barja, 2004), which explains why antioxidants can often be of detriment (Lane, 2005). These data do, however, illustrate increased ROS protection and lower ROS production in numerous bird species, which results in less DNA damage (Barja, 2004).

These avian physiological features were overlooked by Berlin et al. (2007). This is surprising given that the mitochondrial ETS is juxtaposed to mtDNA, and that ETS-derived ROS makes the greatest contribution to mtDNA damage, and hence diversity (Barja, 2004). Admittedly, ROS production would be a difficult parameter to measure for many species, although ROS production correlates more tightly with longevity than body mass (Barja, 2004). Therefore, correlation of π with mammalian and avian longevity may provide greater insight (note that this assumes the metabolic theory of ageing). Potentially, the lower mitochondrial ROS output of bird mitochondria may provide another and potentially stronger physiological explanation for the low mtDNA diversity of birds.