Recent discoveries and analyses have revolutionised our knowledge of Mesozoic Mammals, revealing novel aspects of their ecology, development, systematics, and macroevolution. However, details of physiology are more difficult to determine from fossils, and our knowledge of physiological evolution remains comparatively poor. Living Mammals are endotherms, possessing the ability to control and maintain metabolically produced heat and have a substantially higher capacity for sustained aerobic activity than ectothermic Animals. The origin of endothermy is an important event in Mammalian evolution, often noted as key to their success. There are a number of competing evolutionary hypotheses for the origin of endothermy: (a) selection for higher maximum metabolic rates enhanced sustained aerobic activity, (b) selection for higher basal metabolic rates enhanced thermoregulatory control, or (c) maximum metabolic rates and basal metabolic rates evolved in lockstep with each other.
Direct evidence from living Mammals to support these hypotheses is equivocal. Recent analyses find no long-term evolutionary trend in basal metabolic rates contradicting earlier suggestions of increasing basal metabolic rates throughout the Cainozoicand so implying that the Middle Jurassic (roughly 170 million years ago) most recent common ancestor of living Mammals possessed a basal metabolic rate within the range of present-day Mammals. Several indirect indicators of metabolic physiology in fossil Synapsids have been suggested but provide contradictory evidence for the timing of the origin of endothermy and its evolutionary tempo. These include: the presence of fibrolamellar long-bone histology, first seen in the Early Permian Synapsid Ophiacodon about 300 million years ago, the presence of an infraorbital canal and lack of parietal foramen, used to infer facial whiskers, fur, lactation and endothermy in Early Triassic (about 245 million years ago) Cynodonts; inferred maxillary nasal turbinates in the Late Permian (about 255 million years ago) Therapsid Glanosuchus, used to suggest that Mammalian levels of endothermy evolved by the Late Triassic (approximately 210 million years ago); a trend towards increased relative brain size initiated in Late Triassic non-Mammaliaform Cynodonts and the Mammaliaform (stem Mammal) Morganucodon; and acquisition of a parasagittal gait in the Early Cretaceous (roughly 125 million years ago) Therian Mammals Eomaia and Sinodelphys. Several recent studies provide more quantitative links to physiological parameters. Oxygen isotopes were used to infer elevated thermometabolism in Middle–Late Permian (roughly 270–255 million years ago) Eucynodonts, red blood cell size diminution in Late Permian (about 255 million years ago) Eutheriodontid Therapsids was linked via two proxies to increased maximum metabolic rates and osteocyte lacuna shape correlations suggested 'Mammalian' resting metabolic rates in Permo-Triassic (about 250 million years ago) Dicynodonts.
However, the inconsistency of these characters, in time and with respect to phylogeny, along with re-assessments of function in relation to endothermy, limit their use as conclusive indicators of modern mammalian levels of endothermy in fossil taxa. Such temporal and phylogenetic heterogeneity suggests that the evolution of Mammalian endothermy followed a complex, mosaic pattern with different physiological aspects likely evolving independently, and at separate rates, towards current Mammalian levels. Additionally, few of these physiological proxies are directly related to measurable aspects of metabolic rate.
In a paper published in the journal Nature Communications on 12 October 2020, a team of scientists led by Elis Newham of the School of Physiology, Pharmacology & Neuroscience at the University of Bristol, and the Bioengineering Science Research Group at the University of Southampton, present the results of a study which used two proxies to improve understanding of physiology at one of the most important nodes along this transition, in order to address these issues.
Newham et al. did so by estimating basal metabolic rates and growth rate, and calculating a known proxy for maximum metabolic rates, for two of the earliest known Mammaliaforms, Morganucodon and Kuehneotherium. Using cementochronology to estimate maximum lifespan by counting growth increments in synchrotron radiation-based micro-computed tomographic data of fossil dental cementum, Newham et al. estimated that both taxa had significantly longer lifespans than extant Mammals of comparable size. By regressing lifespan against basal metabolic rates and postnatal growth rate in extant Mammals and Reptiles, Newham et al. in turn estimated significantly lower values for both of these metrics for the earliest Mammaliaforms. However, when they compared the blood flow index (the ratio between femoral nutrient foramina area and femur length that serves as a proxy for maximum metabolic rate) of Morganucodon with those of extant taxa, Newham et al. found that Morganucodon had an intermediate value between living Mammals and Reptiles. These results suggest that basal Mammaliaforms occupied a metabolic grade similar to living Reptiles and had yet to achieve the endothermic physiology of living Mammals.
Newham et al. used maximum lifespan (i.e. the single longest known lifespan of a taxon) estimates for fossil Mammaliaform taxa as a proxy for both basal metabolic rate and postnatal growth rate. In extant Tetrapods, negative correlations exist between maximum lifespan and basal metabolic rate and between maximum lifespan and growth rate. In general, the longer a Mammal’s lifespan, the lower its size-adjusted basal metabolic rate and growth rate. Growth rates have been shown to correlate strongly with metabolic power in extant Vertebrates, with endotherms growing an order of magnitude faster than ectotherms. Maximum lifespan is an applicable value for fossil samples, as, unlike other metrics (e.g. 10% most long lived or mean lifespan of a cohort), it does not rely on cohort- or population-based statistics that fossil samples cannot fulfil. This value is also less susceptible to extrinsic population-level factors on lifespan, such as disease or predation, and relates most closely to the physiological limit of lifespan of an organism. An accurate assessment of maximum lifespan in fossil Mammals can therefore be used to estimate their metabolic potential.
To estimate Mammaliaform lifespans, Newham et al. used cementochronology. This well-established technique, which counts annual growth increments in tooth-root cementum, has been used to record lifespans in extant Mammals with over 70 species aged using this technique. Cementum is a mineralised dental tissue surrounding the tooth root, attaching it to the periodontal ligament and anchoring the tooth within the alveolus. Growth of cementum is continuous throughout life in extant Mammals and seasonally appositional in nature, forming a series of increments of differing thickness and opacity when viewed in histological thin sections under light microscopy. The correlation between increment count and chronological age is well documented, with one thick and one thin increment deposited every year. It has been shown that the thin, hyper-mineralised opaque increments record growth rate reduction in less favourable seasons.
Despite this potential, cementochronology has not previously been attempted for fossil Mammals older than the Pleistocene (2.6 million years old), because histological thin sections destroy fossils and provide only a restricted field of view. Newham et al. overcame these problems by using propagation phase-contrast X-ray synchrotron radiation microtomography to non-destructively image fossilised cementum increments. The sub-micrometre resolution, fast-throughput and three-dimensional nature of X-ray microtomography allows for large sample sizes and for increments to be imaged along their entire transverse and longitudinal trajectories in volumetric X-ray microtomography data sets. Cementum increments are known to occasionally split and coalesce, creating errors in counts based on single, or limited numbers of, two-dimensional thin sections created for each tooth. X-ray microtomography imaging and 3D segmentation of individual increments across extensive vertical distances within the cementum allowed us to confidently distinguish principal annual increments from any accessory increments created by lensing and coalescence.
Morganucodon and Kuehneotherium are Shrew-sized insectivores, which co-existed on a small landmass during the Early Jurassic marine transgression (Hettangian-early Sinemurian, about 200 million years ago) in what is now Glamorgan, South Wales, UK. Thousands of their bones and teeth were washed into karst fissures that have subsequently been revealed by quarrying. This provides a rare opportunity to analyse large samples of fossil material needed for confident estimation of maximum lifespan. Importantly, these are the earliest diphyodont taxa (taxa with two successive sets of teeth), with a single replacement of non-molar teeth and no molar tooth replacement, and so estimates of lifespan are accurate to the time of the measured tooth-root formation.
The fossil sample studied included both isolated teeth and mandibles with multiple teeth or roots in situ. Newham et al. applied X-ray microtomography to 87 Morganucodon specimens (52 isolated teeth, 35 dentaries, all Morganucodon watsoni) and 119 Kuehneotherium specimens (116 isolated teeth, 3 dentaries). From these, 34 Morganucodon and 27 Kuehneotherium specimens were sufficiently well preserved for three observers to independently estimate lifespan from cementum increments. These estimates were compared to validate their accuracy and precision. The remainder showed physical and/or diagenetic damage that prevented increment measurement.
The cementum of Morganucodon and Kuehneotherium is distinguished from dentine in Newham et ai.'s X-ray microtomography data by a distinct boundary layer separating the two tissues. This lies external to the granular layer of Tomes of the dentine and is interpreted as the hyaline layer of Hopewell–Smith. Synchrotron nanotomographic imaging (30 nm isotropic voxel size) highlights individual Sharpey’s fibre bundles (linking cementum to the periodontal ligament in extant Mammals) visible in several exceptionally preserved specimens, which can be traced radially through the cementum. Across the toothroot transverse axis, cementum is roughly 10–70 μm radial thickness and displays a series of contrasting light and dark circumferential increments representing different material densities. Higher-density increments (represented by greater greyscale values) are on average 2–3 μm radial thickness, and lower density increments are 1–3 μm radial thickness. Individual increments can be followed continuously both longitudinally and transversely through the entire scanned volume of a tooth root.
Newham et al. tested the accuracy of cementum increment counts for predicting lifespan in fossils by additional X-ray microtomography imaging and counting of increments in the cementum of several teeth along the tooth row in eight dentulous Morganucodon specimens with a range of teeth in situ and of growth increments (lines of arrested growth) in the periosteal region of the dentary bone in two of these. In both specimens where dentary lines of arrested growth are found, counts are identical with the cementum increments in the teeth (p3–m2). Also, comparisons between counts of cementum increments are identical across all four premolars (p1–p4) and the anterior molars (m1–m2), in all specimens where they occur together. This agreement between p1–m2 teeth and dentary increment counts indicates that growth in both teeth and jaws was following the same, circum-annual rhythm, as previously reported for multiple extant Mammal species. Newham et al. consider this to be strong support for the accuracy of lifespan estimates based on these increment counts.
The increment counts along Morganucodon dentary toothrows can also provide information on eruption sequence and timing. The first permanent premolar to the second molar all erupted within 1 year, with the first molar erupting prior to the third and fourth premolars (NHMUK PV M 27312). The ultimate incisor (i4), the canine and the third molar erupted in the following year. Newham et al. do not have information on eruption timing of more anterior incisors, or the fourth molar, and the fifth molar is only occasionally present. As they estimate that Morganucodon was long lived relative to comparatively sized extant Mammals, this pattern of most of the adult tooth row being in place during the first 2 years of life is also supportive of a relatively short (compared with its total lifespan) juvenile stage and determinate growth. The absolute length of these stages in Morganucodon is, however, considerably longer than extant Mammals of comparable body size. Unfortunately, dentulous specimens of Kuehneotherium are rare, and there are no tooth rows with cementum increment counts in Newham et al.'s sample.
Cementum increment counts provide a minimum estimate of maximum lifespan of 14 years for Morganucodon and 9 years for Kuehneotherium. These may underestimate true maximum lifespan, as any damage to outer cementum increments would reduce estimated maximum lifespan. One-way analysis of variance comparisons of mean intra-observer coefficient of variation between Newham et al's study and ten previous cementochronological studies of different extant Mammal species with similar age ranges suggest that values for X-ray microtomography data of Morganucodon and Kuehneotherium are significantly lower than previous thin section-based studies.
Newham et al. estimated body mass ranges of 10.7–25.0 g (mean 17.9 g) for Morganucodon and 14.9–32.7 g (mean 23.8 g) for Kuehneotherium (minimum mass estimates based on skull length and maximum mass estimates on dentary length). Maximum lifespan and mean body mass for the Mammaliaforms were compared with published data for large samples of terrestrial, non-volant wild extant Mammal (278) and non-Avian Reptile (256) species. Maximum wild lifespans of extant taxa were chosen for comparison with our fossil taxa, as these values are the closest analogue to our estimated lifespans, relative to captive lifespan values. To ensure robustness of their results, Newham et al. additionally analysed maximum captive lifespans of extant taxa below 100 g, which show an average increase above maximum wild lifespans of approximately 3.43 and 4.38 years per taxon for Mammals and Reptiles, respectively. Broad results of statistical tests, and the overall conclusions of our study, are unchanged regardless of whether wild or captive data are used for analysis and comparisons between our fossil lifespans and the lifespans of extant taxa.
Phylogenetic generalized least squares regression of log₁₀-transformed values shows that the fossil Mammaliaforms fall within the range of extant Reptiles and have longer maximum lifespans for their size and are further above the Mammal regression mean, than all extant Mammals under 4 kg (the long-lived and secondarily dwarfed Mouse Lemur, Microcebus murinus, is closest). Only the Short-beaked Echidna, Tachyglossus aculeatus, a Monotreme with long lifespan and low metabolic rate, exceeds the distance above the Mammalian mean for Kuehneotherium, but not for Morganucodon. One-way phylogenetic analysis of covariance comparisons show that regression slopes for extant Mammals and Reptiles are statistically similar, but their means are significantly separated, with Reptiles on average living 18.3 years longer than Mammals of the same body mass.
To estimate basal metabolic rates, Newham et al. used phylogenetic generalized least squares and recovered significant correlations between log₁₀-transformed values of maximum wild lifespan and mass-specific standard metabolic rate (measured in millilitres of oxygen per hour per gram of mass and analogous with basal metabolic rates in extant Mammals; standard metabolic rate was used as basal metabolic rates cannot be measured in Reptiles) from published data for 117 extant Mammals and 55 extant Reptiles. Using the correlation between maximum wild Reptile lifespan and mass-specific standard metabolic rate and plotting our mammaliaforms directly onto this regression line, Newham et al. estimated a reptile-derived mass-specific standard metabolic rate of 0.055 millilitres of oxygen per hour per gram of mass for Morganucodon and 0.08 millilitres of oxygen per hour per gram of mass for Kuehneotherium. Newham et al. additionally used the correlation between maximum wild Mammal lifespan and mass-specific standard metabolic rate and estimated a Mammal-derived mass-specific standard metabolic rate of 0.36 millilitres of oxygen per hour per gram of mass for Morganucodon and 0.46 millilitres of oxygen per hour per gram of mass for Kuehneotherium. When log₁₀ phylogenetic generalized least squares regression is used to regress these estimates against body mass, both Mammaliaforms fall outside the 95% predictor interval of the Mammalian data and within the Reptile range of mass-specific standard metabolic rate, regardless of whether Mammaliaform mass-specific standard metabolic rate is estimated from Reptilian or Mammalian data. This suggests that the Mammaliaforms had significantly lower mass-specific standard metabolic rate values when compared to extant Mammals of similar size. The comparably sized Mammal (under 100 g) of lowest mass-specific standard metabolic rate is the Marsupial Dasycercus cristicauda, with a maximum wild lifespan of 7 years and a mass-specific standard metabolic rate of 0.63 millilitres of oxygen per hour per gram of mass.
Newham et al. estimated growth rates using phylogenetic generalized least squares correlations between maximum wild lifespan and growth rate from published data from 115 extant Mammals and 30 extant Reptiles. From Mammal data, Newham et al. estimate growth rate constants of 1.085⁻² per day for Morganucodon, and 1.474⁻² per day for Kuehneotherium. From Reptile data, Newham et al. estimated growth rate constants of 4.91⁻⁴ per day for Morganucodon, and 6.65⁻⁴ per day for Kuehneotherium. Log₁₀-transformed phylogenetic generalized least squares regression against body mass again places both Mammaliaforms outside the Mammalian 95% predictor interval and within the Reptile growth rate range, whether estimated from Mammalian or Reptilian data. The lowest growth rate of any under 100 g extant mammal is 3.24⁻² per day for the Mongolian Gerbil, Meriones unguiculatus.
In summary, Newham et al.'s estimates of maximum lifespan provided by tomographic imaging of cementum increments in Morganucodon and Kuehneotherium are significantly longer than the maximum wild lifespan of any extant Mammal of comparable body mass. These lifespans provide estimates of standard metabolic rate/basal metabolic rates and growth rate that are significantly lower than comparably sized extant Mammals and instead correspond to those of extant Reptiles.
To compare their fossil Mammaliaform basal metabolic rates estimates with maximum metabolic rates, Newham et al. used a second proxy directly linked to maximum metabolic rates. The ratio between nutrient foramen area and femur length has been used as an index for relative blood flow through the femur during and after metabolically demanding exercise, previously shown to correlate well with maximum metabolic rates. From microtomography data of the six most complete Morganucodon femoral diaphyses available, Newham et al. segmented all nutrient foramina and estimated their area by measuring their minimal radii. Kuehneotherium could not be included as no suitable femoral specimens are known.
Newham et al. estimated an index for relative blood flow of 3.829⁻⁷ mm³ for Morganucodon and compared this with published and new data for extant Mammals (69 species) and Reptiles (30 species). The latter includes Varanids (8 species), which in the absence of Mammalian predators fill an active hunting niche and tend to have Mammalian maximum metabolic rates levels while retaining Reptilian basal metabolic rates levels. One-way phylogenetic analysis of covariance comparisons show that means of generalised least squares regression slopes for extant Mammals and non-Varanid Reptiles are significantly different, while the slopes are similar. Log₁₀ generalised least squares regression of body mass and index for relative blood flow shows that is further above (higher index for relative blood flow for its mass) the non-Varanid reptile mean than all non-Varanid Reptiles (phylogenetically informed statistical comparisons were not used here due to the non-significant lambda values showing no phylogenetic signal in the taxa used). However, is also slightly further from the Mammalian mean than the non-Varanid Reptile mean and considerably closer to small non-Varanid Reptile species data points than those of small Mammalian species. This intermediate index for relative blood flow and so inferred intermediate maximum metabilc rate, suggests that, while retaining typical Reptilian basal metabolic rates and growth rates, Morganucodon had above non-Varanid Reptiles but not as high as Mammals or actively foraging Varanid Reptiles.
Newham et al. have used two quantitative proxies to determine the metabolic status of early Mammaliaforms. Relatively long lifespans for both and result in standard metabolic rate/basal metabolic rate and growth rate estimates equivalent to modern Reptiles of comparable size and indeed at the higher lifespan/lower basal metabolic rate/slower growth end of the Reptile scale for Morganucodon. This is true whether Newham et al. compare their fossil estimates to wild lifespans of extant taxa or estimate fossil 'captive' lifespans and compare them to captive values for extant taxa. In contrast, femoral blood flow estimates suggest that the maximum metabolic rate of was intermediate between extant non-Varanid Reptiles and Mammals. Newham et al. therefore infer that in increased maximum metabolic rate (and so also absolute aerobic capacity which is considered to be equal to the maximum metabolic rate minus the basal metabolic rate) was initially selected for before basal metabolic rate and that the basal metabolic rate-first hypothesis is the best-supported model for the evolution of Mammalian endothermy. Newham et al. suggest that at least Morganucodon, if not also Kuehneotherium, occupied a metabolic grade approaching extant Varanids: able to undergo longer bouts of aerobically demanding activity than non-Varanid Reptiles but not capable of sustaining either Mammalian levels of aerobic activity or the elevated thermometabolism exhibited by living endotherms.
Comparison of Newham et al.'s results to those of other recent studies of physiology in fossil Synapsids supports the hypothesis of a complex, mosaic pattern for the evolution of endothermy, with different characters being selected for at different rates through time, and with respect to phylogeny. For example, the size diminution associated with the Cynodont–Mammaliaform transition may have reversed the evolutionary trajectory of some previous histological proxies for endothermy, contributing to the complex, contradictory patterns observed. Newham et al.'s study also suggests that more work is needed to compare fossil and extant ectothermic and endothermic taxa directly in order to better understand their relative metabolic properties. Many previous studies rely on simple binary divisions, such as the presence/absence of fibrolamellar bone and/or respiratory nasal turbinates. These proxies cannot represent accurately the complex series of physiological characteristics that range between 'ectothermy' and 'endothermy' and are frequently distributed homoplastically across the Synapsid phylogeny, individually and with respect to each other. Other studies provide relative data such as preserved apatite oxygen isotopes that allow comparisons with cohabiting ectothermic taxa but cannot be directly compared to extant data and so do not suggest where the studied fossil taxa fall in the metabolic spectrum of extant Vertebrates. However, Newham et al.'s results are compatible with recent work on living Mammals, suggesting that the basal metabolic rate of the Middle Jurassic (roughly 170 million years ago) Mammalian most recent common ancestor was comparable to present-day values. This indicates that evolution towards modern-day Mammalian endothermy occurred during the 25 million year-long Early Jurassic and suggests that the Mammalian mid-Jurassic adaptive radiation was driven by this or vice versa.
In conclusion, Newham et a\l.'s data offer a direct link to measurable aspects of endothermy, such as basal metabolic rate and maximum metabolic rate, at a key point in Mammalian evolution. Further work applying these methods to additional Mesozoic Mammaliaforms and Mammals, and comparison with evidence from other physiological characteristics, will allow the evolutionary tempo and mode of multiple aspects of Mammalian physiology to be determined. The early Mammaliaforms and possessed surprisingly low, Reptile-like metabolic rates plus a mixture of plesiomorphic and derived characters relating to life history and physiology. Ultimately, we can no longer assume that the endothermic metabolism of living Mammals had evolved in the earliest Mammaliaforms.
See also...
Online courses in Palaeontology.
Follow Sciency Thoughts on Facebook.
Follow Sciency Thoughts on Twitter.
