Blackwell Publishing LtdOxford, UKZOJZoological Journal of the Linnean Society0024-4082The Lin-

nean Society of London, 2006? 2006

1474
473488
Original Article
ORIGIN OF ENDOTHERMYT. S. KEMP

Zoological Journal of the Linnean Society, 2006, 147, 473–488. With 4 figures

The origin of mammalian endothermy: a paradigm for the

evolution of complex biological structure
T. S. KEMP*

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

Oxford University Museum of Natural History and St John’s College, Oxford OX1 3JP, UK

Received April 2005; accepted for publication October 2005

Several mutually incompatible theories exist about how and why endothermy evolved in mammals and birds. Some
take the primary function to have been thermoregulation, selected for one adaptive purpose or another. Others take
the high aerobic metabolic rate to have been primary. None of these theories is incontrovertibly supported by evi-
dence, either from the fossil record of the synapsid amniotes or from observations and experiments on modern organ-
isms. Furthermore, all are underpinned by the tacit assumption that endothermy must have evolved in a stepwise
pattern, with an initial adaptive function followed only later by the addition of further functions. It is argued that
this assumption is unrealistic and that the evolution of endothermy can be explained by the correlated progression
model. Each structure and function associated with endothermy evolved a small increment at a time, in loose linkage
with all the others evolving similarly. The result is that the sequence of organisms maintained functional integration
throughout, and no one of the functions of endothermy was ever paramount over the others. The correlated pro-
gression model is tested by the nature of the integration between the parts as seen in living mammals, by computer
simulations of the evolution of complex, multifunctional, multifactorial biological systems, and by reference to the
synapsid fossil record, which is fully compatible with the model. There are several potentially important implications
to be drawn from this example concerning the study of the evolution of complex structure and the new higher
taxa that manifest it. © 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147,
473–488.

ADDITIONAL KEYWORDS: correlated progression – megaevolution – origin of mammals – Synapsida –

Therapsida.

INTRODUCTION the possession of feather impressions in certain dino-

saurs, which is taken to indicate that such forms were
Nothing is more fundamental for understanding the
already endothermic (Seebacher, 2003; Xu et al.,
biological nature of birds and mammals than their
2004). However, in the case of the evolution of the
endothermic temperature physiology: for organisms to
mammalian version of endothermy there exists the
pay the price of some ten-fold increase in their daily
celebrated fossil record of basal Synapsida, or ‘mam-
food requirements it must be balanced by a very sub-
mal-like reptiles’, which includes a series of grades of
stantial benefit indeed to their ultimate reproductive
organisms ranging from the indisputably ectothermic
potential. It is little wonder therefore that a great deal
pelycosaurs through intermediate levels to the equally
of attention has been paid to the problem of why, how
indisputably endothermic early mammals (Bennett &
and when endothermy arose. What is rather more sur-
Ruben, 1986; Kemp, 2005). The preserved anatomy of
prising is how little agreement there is. At least six
these fossils has been extensively scrutinized for sup-
different hypotheses currently exist (Hayes & Gar-
port by authors of various of the respective interpre-
land, 1995; Kemp, 2005). There is relatively little fos-
tations of the origin of endothermy, although again
sil evidence available that bears helpfully on the
with little mutual agreement on what structures, if
details of the origin of the avian version of endothermy
any, are reliable indicators of the nature of the ani-
(Ruben, 1995; Schweitzer & Marshall, 2001), beyond
mal’s temperature physiology.
The number of contradictory explanations for the
*E-mail: [email protected] origin of endothermy in mammals may in large part be

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488 473
474 T. S. KEMP

due to this scarcity of direct evidence. More fundamen- aethiopicus has a BMR only half, and the semi-arid
tally, it also points to the possibility that these adapted Hemiechinus auritus only three-quarters that
considerations are hampered by a mistaken implicit of the temperate-adapted European hedgehog Erina-
assumption concerning the evolution of complex struc- ceus europeaus, yet all maintain the same body temper-
tures and associated functions in organisms. As will be ature of 34 °C, and have similar activity levels. The
discussed later, a more fruitful approach is to pay elevated body temperature also cannot be regarded as
more than the usual lip service to the presumption an adaptation per se, because it too is very variable
that an evolving lineage of organisms must necessar- among otherwise comparable mammal species, and
ily maintain functional integration throughout its his- because many ectothermic amniotes operate with body
tory. The concept of ‘correlated progression’ (see Kemp, temperatures as high as, or higher than, those of mam-

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

1999) seems to be widely accepted in principle (e.g. Lee mals. This leaves two basic direct functions of the
& Doughty, 1997; Budd, 1998), but largely ignored in endothermic physiological system, thermoregulation
practice when discussing major evolutionary transi- and elevated aerobic activity levels, respectively.
tions to new higher taxa. Before taking up this issue,
the biological nature of endothermy must be clarified,
and the menu of current theories purporting to THERMOREGULATION
explain its origin in mammals briefly reviewed. Maintaining a constant body temperature is an eco-
logical adaptation for remaining active over a wider
range of ambient temperature. Although in principle
THE PHYSIOLOGY OF ENDOTHERMY IN
this includes high ambient temperatures, the effect is
LIVING MAMMALS
perhaps more striking at the low ambient tempera-
Typical endothermic temperature physiology of mam- ture of night-time. Thermoregulation also serves
mals is characterized by four measurable aspects: another, distinctly physiological function. Mainte-
nance of a precisely constant internal temperature is
1. The basal or resting metabolic rate (BMR) is high.
an essential prerequisite for the higher degree of orga-
It varies typically from five to as much as ten times
nizational complexity of endotherms compared with
that of an ectotherm of similar body size (e.g. Hayes
ectotherms. The rates of enzyme-controlled metabolic
& Garland, 1995; Hulbert & Else, 2000).
pathways, diffusion rates such as that of transmitter
2. The body temperature ( Tb) is higher than the ani-
molecules across synapses, and the viscosity and
mal’s normal ambient temperature, and lies
therefore speed of contraction of muscle fibres such as
between about 28 °C and about 40 °C depending on
the heart muscle are all temperature dependent, and
species (e.g. Crompton, Taylor & Jagger, 1978;
therefore can only occur at the reliably predictable
Eisenberg, 1981).
rates necessary for sustaining the integrity of a com-
3. The core body temperature is maintained at a
plex system if the body temperature is maintained
remarkably constant value, not normally varying
constant. Given the degree of sensitivity of the greatly
by more than 1–2 °C over the 24-h diurnal cycle
enlarged and complex mammalian brain to induced
(e.g. Eisenberg, 1981).
temperature changes, it is evident that the central
4. The maximum aerobic metabolic rate (MAMR) that
nervous system more than anything depends critically
the organism is capable of sustaining is greatly ele-
for its proper functioning on maintenance of the cor-
vated over that of ectotherms. There is a very
rect temperature.
approximately constant ratio of 10–15 between the
As recently reviewed in the context of the origin of
basal rate and the maximum aerobic rate in
avian endothermy (Schweitzer & Marshall, 2001), the
amniotes, and therefore, like the BMR, the MAMR
biochemical basis for the elevated BMR lies in an
of an endotherm is typically somewhere between
increased number of mitochondria in the cells, mostly
five and ten times that of an ectotherm of similar
those of the visceral organs. They metabolize aerobi-
body size (e.g. Taylor et al., 1987; Hinds et al., 1992;
cally and inefficiently, with a consequent increase in
Hayes & Garland, 1995).
heat production (Hulbert & Else, 1989, 1990, 2000). For
Disentangling which of these aspects are the functions this to be linked efficiently to a thermoregulatory func-
and which are the mechanisms responsible for the tion requires first that the body is sufficiently well insu-
functions seems clear now, although for many decades lated for the body temperature to rise high enough to
there was considerable confusion. The elevated BMR create a heat gradient between the body and the out-
cannot be considered to be of direct functional impor- side world: the surface conductance has to be low
tance because it is very variable among otherwise sim- enough. But secondly, there must be a means of varying
ilar and closely related species. To name but one of the conductance of the surface with great speed and
many examples, the study of Shkolnik (1980) showed precision so that the rate of heat loss can be rapidly
that the desert-adapted African hedgehog Paraechinus adjusted to changes in the net rate of heat input, which

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 475

is the sum of the BMR and any heat of activity being constant fraction of the MAMR in order to maintain a
generated by muscular exercise. Variable insulation of rapidly increased oxygen supply via the circulatory
the skin by variable piloerection, variable blood flow system during high activity levels (Krosniunas & Ger-
through the cutaneous capillaries and variable posture stner, 2003). This problem will be returned to.
are the well-known mechanisms for achieving this rap-
idly adjustable rate of conductance of heat from the
CURRENT THEORIES OF THE ORIGIN OF
body. Many other features such as a more effective cir-
ENDOTHERMY
culatory system for heat distribution and a higher oxy-
gen-carrying capacity of the haemoglobin are also The lack of an explicit reason for a relationship
necessary, as will be discussed later. between the physiology of thermoregulation on the

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

one hand and of aerobic capacity on the other has led
to a taxonomy of theories that divides them into two
HIGH AEROBIC ACTIVITY categories, the ‘thermoregulation first’ and the ‘aerobic
Within the thermo-neutral zone of ambient tempera- capacity first’ views. Within each of these there are dif-
ture, much the greater proportion of the additional ferences in what is hypothesized to have been the ini-
metabolism above BMR is devoted to muscular activ- tial selective pressure promoting the evolutionary
ity, and therefore the five-fold or more increase in the progress towards endothermy.
maximum sustainable level of aerobic metabolism in
endotherms potentially impacts on all the activities of
the animal. For example, Bennett & Ruben (1979) THERMOREGULATION FIRST – PHYSIOLOGICAL VERSION
quoted a maximum sustainable speed of a 1-kg Iguana The most detailed version of several claims that endot-
as 0.5 km h−1 and of a mammal of similar body weight hermy arose initially for the benefit that a constant
as 4.1 km h−1; Farmer (2000) quoted the huge daily body temperature endows on the physiological inte-
locomotory investment in birds foraging for their nest- gration of the organism is McNab’s (1978) theory of
lings. The mechanism behind the raised MAMR is miniaturization. He envisaged that the synapsids
quite simple. A larger number of mitochondria with a originated from small-bodied ectotherms. They then
larger net membrane area in the muscle tissue, cou- evolved increased body size, up to 30–100 kg, at which
pled with adequate oxygen delivery by the vascular point they became inertial homeotherms, the condi-
system, permits a greater rate of ATP synthesis and tion where the animal is buffered by its low surface
its conversion to mechanical energy. area to volume ratio against short-term fluctuations in
The relationship between the increased level of the ambient temperature. Insulation in the form of a
maximum aerobic activity and the thermoregulatory pelt may have been acquired at this stage to enhance
function is poorly understood: the main site of metab- the effect. Therefore, the body temperature remained
olism for thermoregulation is the viscera, that for close to constant, with the consequent benefits of con-
aerobic activity is the musculature, but there is no stant rates of enzyme activity, nerve conduction rates,
obvious mechanical or functional reason why they etc. The second step was a process of reduction in body
should be linked to one another (Bennett & Ruben, size, but in order to retain the constancy of the body
1979). Nevertheless, a considerable body of empirical temperature there had to be a relative increase in
evidence in living organisms demonstrates the metabolic heat production, and in insulation. This evo-
roughly constant ratio between the two values in both lutionary sequence was represented by successively
ectotherms and endotherms (e.g. Bennett & Ruben, smaller non-mammalian cynodonts, and culminated
1979; Hayes & Garland, 1995; Ruben, 1995; Krosniu- in the very small body mass of the first mammals,
nas & Gerstner, 2003). Suggestions of possible reasons which had a body weight of the order of 10–20 g.
for the correlation include: a relationship between an McNab (1978) claimed that the actual body sizes
increase in the metabolic activity of the muscle tissue seen in the fossil record support his theory, with the
and a corresponding increase in the metabolic func- small ectotherms represented by basal amniotes (‘cot-
tions of the viscera necessary to maintain the muscle ylosaurs’), and the inertial homeotherm stage repre-
tissue (Ruben, 1995); a coincidental correlation sented by large-bodied Late Carboniferous and Early
between a raised BMR for maintaining a higher incu- Permian pelycosaurs and the Late Permian basal
bation temperature for the juvenile and an increased therapsid taxon Pristerognathidae. This is followed by
need for higher muscle activity levels associated with reduction in size in the sequence of Triassic cynodonts.
food collection for provisioning the young (Farmer, As well as becoming smaller in body size, the cynodont
2000); a correlation between increased locomotory sequence also exhibits the evolution of a secondary
activity for food collecting and increased visceral palate, indicating an increase in ventilation rate and
metabolism for assimilation of the extra food (Koteja, by implication of metabolic rate. Unfortunately, the
2000); and a requirement that the BMR remains a phylogeny of therapsids adopted by McNab was

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
476 T. S. KEMP

pre-cladistic and is now grossly out of date. The widely However, it may be the case that the evolution of
accepted modern version (Hopson, 1991; Kemp, 2005) thermoregulatory ability was itself necessary for the
offers no support at all to his argument. In the first evolution of the large, complex brain of endotherms,
place, there are medium- and small-bodied pelyco- typically ten times the size of that of ectotherms. Cer-
saurs, as well as the large-bodied ones, throughout tainly the functioning of the mammalian (and avian)
their history (Reisz, 1986); in the second place, the brain today depends on precise control of the internal
pristerognathids are now regarded as part of a mono- environment of the body including temperature (Jeri-
phyletic Therocephalia, which is the sister group of son, 1971; Allman, 2000). Evidence from the fossil
cynodonts plus mammals, and which includes a con- record concerning synapsid brain size is ambiguous
siderable size range of members with skull lengths because it was not until the mammals themselves that

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

ranging from 30 cm to less than 10 cm. Cladistic anal- the brain actually filled the cranial cavity, permitting
ysis of the fossils implies that the common ancestor of accurate estimates of its size. Indeed, the basal cyn-
cynodonts and therocephalians was in fact a relatively odont brain has been reconstructed on the one hand as
small-bodied form. In the third place, the cynodonts significantly increased in size (Kemp, 1979), and on
also embrace a wide size range. There are small- the other as scarcely enlarged beyond the range seen
bodied forms dating from the Early Triassic such as in living reptiles (Hopson, 1979; Rowe, 1996). If the
Thrinaxodon and Galesaurus, and from the Middle former estimate is accepted, then enlargement of the
Triassic such as Probainognathus, skull lengths of brain coincided with the evolution of the cynodont fea-
which are approximately 7 cm. But there are also tures frequently associated with an increased meta-
large bodied Early, Middle and Late Triassic forms, bolic rate and the beginning of endothermy.
with skull lengths ranging around 30 cm, for instance
Cynognathus and Exaeretodon. Yet, small-, medium-
and large-bodied alike, these cynodont-grade ther- THERMOREGULATION FIRST – ECOLOGICAL VERSION
apsids all possess the same characteristics such as a The theory that the initial function of endothermy was
fully differentiated dentition, enlarged and reorga- maintenance of a constant body temperature so that
nized jaw musculature, a secondary palate, reduced the organism could be active at night has been pro-
lumbar ribs and a mammal-like hindlimb. In so far as mulgated most cogently by Crompton et al. (1978) and
features such as these are presumed to be correlated Taylor (1980). By their calculations, a small animal
with metabolic rate, the similarity implies that they that is well enough insulated by fur can maintain a
all possessed a comparable temperature physiology constant body temperature with little or no increase in
irrespective of body size. Only the final stage in the BMR, provided that the ambient temperature is low
transition from cynodont-grade to Mammalia, which enough. They argued that this was indeed the case in
occurred in a single Late Triassic lineage, corresponds the ancestral mammals, where a virtually reptilian-
to the expectations of McNab’s theory: all the earliest level metabolic rate would have been adequate to
mammals were indeed small. maintain a constant body temperature of between 25
Bennett, Hicks & Cullum (2000) attempted to refute and 30 °C at the ambient temperatures prevailing at
experimentally this, or any other theory which pro- night. The ability of most modern mammals to remain
poses that a rise in metabolic rate immediately active during the day requires a higher body temper-
endows a degree of thermoregulation. They gave liz- ature in order to have the ability to lose heat fast
ards a very large meal, which has the effect of increas- enough; this in turn requires a higher BMR to gener-
ing the metabolic rate of the viscera by 3–4 times, ate it. Therefore, the shift to diurnality was presumed
presumably to increase the rate of digestion. The to be a secondary stage, not found until the post dino-
authors then tested for an enhanced thermoregulatory saur radiation of mammals in the Cenozoic.
ability, but found no significant increase in the body Evidence for the theory is two-fold. There are sev-
temperature or decrease in rate of cooling, compared eral groups of living nocturnal mammals, notably ten-
with the control lizards that had not been force fed. recs, hedgehogs and small marsupials, with relatively
low body temperatures and metabolic rates. Crompton
et al. (1978) claimed that these had retained the
THERMOREGULATION FIRST – BRAIN SIZE VERSION ancestral mammalian condition. In fact, the distance
Hulbert (1980) pointed out that 5–10% of the BMR of of the relationships between this very disparate sam-
a mammal is due to the metabolic activity of the brain, ple of mammals makes such an inference exceedingly
and that therefore selection for an enlarged brain of unparsimonious, and all are surely secondarily modi-
itself may in part have led to endothermy. In fact, the fied independently of one another. Nevertheless, the
relationship between brain size and BMR across spe- very existence of mammals with such physiological
cies does not support such a simple interpretation characteristics, even if specialized among today’s
(Harvey & Krebs, 1990; Harvey & Pagel, 1991). mammals, indicates that the theory is feasible. The

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 477

second line of evidence is that the earliest mammals the sensitivity of vertebrate embryos generally to
show several signs of having indeed been nocturnal. their ambient temperature, from the point of view of
The small body sizes, agile skeletons and insectivore- both avoiding developmental abnormalities and
adapted shearing molar teeth all point to a close anal- increasing the rate of development, a generality dis-
ogy in life style with modern insectivorous mammals, puted by Angilletta & Sears (2003). Furthermore, the
which are primarily nocturnal in habit. The evidence dual role of thyroid hormones in reproduction and in
for increased acuity of hearing as indicated by the evo- the control of metabolic rate suggested to Farmer a
lution of the ear ossicles, and even more so of olfaction possible mechanism for the initial evolution of the
as indicated by the development of the neocortex in enhanced BMR. In this particular theory, the
the primitively olfactory telencephalon region of the increase in maximum aerobic activity in endotherms

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

brain, both imply nocturnal rather than diurnal hunt- may have evolved subsequently as a further manifes-
ing, in which vision would be expected to dominate. tation of parental care, namely to increase the rate of
However, there are arguments against this noctur- food collecting to nourish the young. No fossil evi-
nalization theory. The most important is that it dence is quoted.
implies that there was no increase in metabolic rate in
cynodonts and early mammals, despite all the mor-
phological features found in these fossil forms that AEROBIC CAPACITY FIRST – SUSTAINED ACTIVITY
apparently indicate an enhanced metabolic rate. VERSION
These include the presence of a secondary palate, and The theory that selection acted initially to increase the
the remodelling of the rib cage to the morphology asso- MAMR was first explicated in detail by Bennett &
ciated in living mammals with a functioning dia- Ruben (1979, 1986) and Ruben (1995) in their ‘aerobic
phragm, both of which are indicative of a higher capacity’ model, and is one of the most influential
ventilation capacity. The modification of the dentition, explanations for the origin of endothermy. Asserting
mandibular anatomy and jaw musculature, which that a small increment in BMR in the absence of asso-
together permit forceful but precise tooth occlusion, is ciated insulation would have a negligible effect on
hard to interpret in any way other than as an adap- thermoregulatory ability, they proposed that the ini-
tation for increased rate of food assimilation. The tial selection for endothermy was to increase aerobic
virtually mammalian organization of the postcranial scope: even a small incremental increase in the max-
skeleton points to the existence of high levels of loco- imum sustainable aerobic level of activity would be
motory activity, at least facultatively. expected to have had an immediate impact on the life
Experimental testing of the general idea that of the animal.
insulation alone can endow an ectotherm with endo- Actual evidence for the aerobic capacity theory
thermic abilities was provided by Cowles (1958), by comes from the observation that the basal and maxi-
covering a lizard with a mink fur coat and assessing mum aerobic metabolic rates are decoupled from one
its thermoregulatory ability. He reported that, as another to the extent that the former is due mainly to
expected under the laws of simple physics, there was a the metabolism of the tissues of the visceral organs
decrease in the rate of warming of the specimen when while the latter is due mainly to that of the muscula-
it was removed from cold to hot conditions, and of cool- ture. Therefore, it is inferred that the two functions
ing when placed back into cool conditions. However, as evolved independently, and because selection could
Cowles himself cautions, it is not clear how reliable an not have been for thermoregulation initially, it must
indicator of an actual evolutionary transition such an have been for increased sustainable aerobic activity
observation is. A modern ectotherm’s skin and its level of the muscles. However, due to the correlation
associated vascularization is primarily designed as a that exists between BMR and MAMR, the BMR also
heat-absorbing surface, whereas the external surface increased and this paved the way for the appearance
of an endotherm is adapted for differential heat loss, of thermoregulation at a later stage in the evolution of
under the control of neuronal and endocrine mecha- fully developed endothermy. A number of authors have
nisms absent from reptiles. proposed that a demonstration of a significant corre-
lation between BMR and MAMR in living animals
constitutes a test of the aerobic capacity theory (e.g.
THERMOREGULATION FIRST – GROWTH OF OFFSPRING Hayes & Garland, 1995; Dohm, Hayes & Garland,
VERSION 2001; Boily, 2002; Gomes et al., 2004). Interesting and
Farmer (2000, 2003) speculated that the initial selec- important for understanding the origin of endothermy
tion pressure for the evolution of endothermy was for as this issue is, strictly speaking the aerobic capacity
increased parental ability to maintain a higher incu- theory does not require that there has to be a correla-
bation temperature for the developing young. The tion between the two rates, but only that increased
argument is based mainly on observations regarding activity level could evolve prior to, and independently

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
478 T. S. KEMP

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

A

nt? mxt? etht?

B
Figure 1. A, reconstruction of the skeleton of the therocephalian therapsid Regisaurus in lateral and dorsal views (from
Kemp, 1986). B, internal view of the nasal cavity of the therocephalian Glanosuchus (from Hillenius, 1994). Abbreviations:
etht?, possible ethmo-turbinal ridge; mxt?, possible maxillo-turbinal ridge; nt?, possible naso-turbinal ridge.

of, thermoregulation, whatever the nature of the occurrence of levels of increased aerobic activity in
mechanisms underlying the two respective functions. therapsids that are otherwise very primitive com-
Evidence from the fossil record has also been pared with mammals (Bennett & Ruben, 1986). The
claimed to support this theory, by demonstrating the modification of the locomotory system in basal ther-

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 479

apsids (Fig. 1A) was interpreted by Carrier (1987) as a subsequently in the case of mammals as lactation.
means of increasing the ventilation capacity. Getting Increased daily energy expenditure was required for
rid of lateral undulation of the vertebral column, rais- the enhanced locomotory efforts devoted to food collec-
ing the body permanently off the ground by reposition- tion. An increase in the metabolic activity of the vis-
ing the limbs so that the feet lie closer to the mid-line, ceral organs was also required, to assimilate the
and reducing the lumbar ribs all suggest that a dia- increased food intake by the parent. Thermoregula-
phragm had evolved, to complement costal breathing tion was not initially a part of the selective regime, but
during times of high activity, when increased oxygen evolved later as a side-effect of the leakiness of the
was required if aerobic activity was to be sustained. plasma membranes that inevitably accompanied the
Reasonable as this interpretation of the evolution of increased metabolic rate. By his own admission, fea-

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

the postcranial skeleton is, it may be incomplete in so sible as it might be, Koteja’s theory is virtually untest-
far as there are additional possible explanations for able, either from the fossil record or from comparative
the change in limb posture in therapsids, notably studies of living endotherms.
increased acceleration and agility (Kemp, 1982).
These attributes of locomotion do not necessarily cor-
AN INTEGRATED VIEW OF THE ORIGIN OF
relate with high levels of aerobic activity, as may be
ENDOTHERMY
true in the case of the facultatively erect gait of croc-
odiles. In any case, demonstration of elevated levels of Why has the problem of the origin of endothermy
aerobic locomotory activity in therapsids would not failed to be solved to general satisfaction? In the first
necessarily exclude the possibility that thermoregula- place, it is because the available empirical evidence is
tion was also present. very limited and invariably ambiguous. Morphological
The second fossil evidence cited is the interpretation features of fossils claimed to be associated with a par-
of certain fine ridges on the inner surface of the nasal ticular mode of thermal physiology are always open to
cavity found in the therocephalian Glanosuchus alternative interpretations, such as the possible func-
(Fig. 1B) as the site of attachment of maxillo-turbinals tion of a secondary palate in mastication, complex
(Hillenius, 1992, 1994; Ruben, 1995), which serve to dentition for a specialist rather than simply abundant
humidify and warm the inspired air in mammals. The diet, and the reorganization of the postcranial skele-
absence of the actual turbinal bones is presumed to be ton for agility. There does appear to be a relationship
due to either their cartilaginous or their very delicate between bone histology and growth rates, although it
osseous nature. However, the supposed maxillo-turbi- is far from clear how this in turn might be related to
nal ridges are very small, and could equally well be metabolic rates (e.g. Horner, de Ricqles & Padian,
the sites of turbinals associated with olfactory epithe- 1999; Botha & Chinsamy, 2000; Ray, Botha & Chin-
lium. Furthermore, their presence is highly inconsis- samy, 2004). Predator–prey ratios, although some-
tent in therapsids. There are various ridges in the times proposed as an indicator of the high levels of
nasal cavity of gorgonopsians (Kemp, 1969) but all lie food requirements of endotherms, have been dis-
above or behind the path of the inspired air. Similarly missed as too unreliable (Bennett & Ruben, 1986).
the dicynodont Oudenodon baini (my pers. observ.) Living amniotes are no more helpful than fossils. The
lacks evidence for a suitably positioned maxillo- various ‘subendothermic’ modern mammals can
turbinal. Sigurdsen (unpubl. data) recently sectioned readily be shown by cladistic analysis to be second-
a snout of a therocephalian and did not find the spe- arily specialized rather than relicts of earlier evolu-
cific ridge described for Glanosuchus. The basal cyn- tionary stages: neither daily torpor in bats, nor low
odont Procynosuchus has turbinal ridges on the dorsal metabolic rates of desert hedgehogs, nor the virtually
surface of the nasal cavity, but no trace on the lateral ectothermic nature of the naked mole rat are reliable
surface, where a maxillo-turbinal would be expected indicators of actual ancestral grades. Attempts to
(Kemp, 1979). Again it should be pointed out that if illustrate steps in the evolution of endothermy by sim-
indeed present, maxillo-turbinals would be just as ple experimental manipulations of modern animals
consistent with other views on the origin of endot- inevitably ignore the very complexity of the system
hermy as with the aerobic capacity first theory. that is being addressed: the reaction to artificial fur
coats or large, force-fed meals by ectotherms that oth-
erwise possess none of the subtle complex of regula-
AEROBIC CAPACITY FIRST – JUVENILE PROVISION tory devices of endotherms are unlikely to illuminate
VERSION realistic intermediate stages.
Koteja (2000) proposed what he termed the ‘assimila- However, there is a second and more fundamental
tion capacity model’. The primary selective force reason for the continuing dispute about the origin of
involved was intensified parental provisioning of the endothermy, which is that the majority of authors of,
juveniles, presumably in the form of food initially, but and commentators on, the respective current theories

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
480 T. S. KEMP

subscribe to the assumption that all the physiological 2. Natural selection tests the fitness of an organism
mechanisms and ecological consequences of modern as a whole, not of any of its individual characteristics.
endothermy could not possibly have evolved at the 3. Therefore, over the course of the evolution of new
same time. Therefore, the implicit argument continues: complex structure, including the broad biological reor-
there must have been a temporal sequence starting ganization associated with the origin of a new higher
with selection for a single, primary function, to be fol- taxon, all the structures and associated functions
lowed only later by selection for secondary functions must evolve by respective sequences of small steps in
and the further structures associated with them. What loose correlation with each other in order to maintain
the respective authors seek is the most feasible recon- continuous functional integration.
struction of this presumed sequence, supported by

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

arguments of such dubious merit as ‘It seems unlikely Under this model, explaining the course of evolution of
that small endotherms could evolve directly from small any particular case of megaevolution, such as the ori-
ectotherms because the low rates of metabolism and gin of endotherms, becomes a matter of understanding
high conductances typical of ectotherms must be con- the nature of the integration between all the struc-
verted to the very high rates and low conductances of tures and functions involved in the evolutionary
small endotherms.’ (McNab, 1978: 4), or ‘It is difficult transition, rather than identifying one particular
to conceive of these metabolic increments occurring for structure or function as paramount over the rest as far
strictly thermoregulatory purposes, particularly at the as selection is concerned. Clearly from this perspec-
initial stages of the evolution of endothermy when they tive, it is unrealistic to consider even the origin of
would have been ineffective in establishing thermosta- endothermy alone, because it too cannot be under-
bility.’ (Bennett & Ruben, 1979: 650) [emphasis added]. stood in isolation of other structures and functions of
A few authors have considered the origin of endo- the organisms that are not immediately associated
thermy from a more integrated point of view. Kemp with temperature physiology, a point taken up later.
(1982, 2005) presented a flow-diagram indicating how To illustrate in broad terms the concept as applied
the various structures and functions of mammals, to the origin of endothermy, imagine, for example, a
including those associated with endothermy, are mutation that resulted in an increased number of
related to one another in the integrated organism, and mitochondria in all the cells, say by 10%, and consider
how this constrains theories about the origin of mam- the effects this might be expected to have (Fig. 2).
malian biology in general. Koteja (2000) presented a Assuming that the animal’s pulmonary and vascular
much more limited but logically similar scenario, systems were capable of delivering the extra oxygen
focused on the effect different aspects of endothermic required, there would be an instant increase of around
physiology might have on postnatal care. In a partic- 10% in the maximum possible rate of aerobic activity
ularly illuminating review of the origin of avian endot- of the muscles, with the effect of increasing the ani-
hermy, Schweitzer & Marshall (2001) recognized that mal’s maximum sustainable running speed by a few
a large number of correlated characteristics had to per cent. An increase of comparable magnitude in the
change to achieve full endothermy. Their particular resting heat production by all the tissues would also
model begins with a molecular mutation increasing be expected, and provided the rate of cutaneous blood
the oxygen affinity of haemoglobin and includes sub- flow and therefore of conductive heat exchange
sequent changes in mitochondria, ventilatory and through the skin was already capable of some control,
musculo-skeletal systems, insulation and reproduc- as it is in living ectotherms, there would immediately
tive strategy. exist the capacity for the organism to remain at an
There are good grounds for starting with different active body temperature for a few extra minutes at
fundamental assumptions about how complex biolog- either end of its working day. Potentially this same
ical systems and the new higher taxa that express increment in internal heat production would also
them evolve. Together these constitute the long-intro- result in a reduction in the amplitude of the daily fluc-
duced but in practice largely ignored correlated pro- tuation of body temperature, thereby reducing by a
gression model of the origin of new higher taxa, or few per cent the variance in rates of enzyme activity,
megaevolution (see Kemp, 1985, 1999). The assump- and physiological processes such as neurotransmis-
tions of the correlated progression model may be sion in brain function. In so far as the organism
stated thus: already had a modicum of parental care, the enhanced
activity level and foraging time, and the more stable,
1. The characteristics of an organism are highly inte- elevated body temperature would all contribute to an
grated with one another, so that no single one of them incremental increase in the level of that care. In short,
can evolve by more than a small increment at any one the hypothesized 10% increase in mitochondrial num-
time without losing its functional integration within ber would have had an instant incremental effect on
the organism as a whole. all of the functions of endothermy at the same time.

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 481

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

Figure 2. The proposed effect of a small increase in the number of mitochondria per cell on several functions of
endothermy.

The consequence for such a hypothetical organism processing systems to acquire the necessary incremen-
would be a minor shift in several of the dimensions of tal increase in food requirements; the skin to regulate
its niche simultaneously; given the right circum- heat flow; and so on. What would happen if one of the
stances, such may well be of selective advantage. functions did not have what might be termed this
As will be appreciated, the correlated progression ‘spare functional capacity’? This particular function
hypothesis requires that the nature of the functional would now be limiting, and the correlated progression
interrelationships between the structures and pro- model predicts that a mutation causing it to be suit-
cesses involved are such that a small mutational mod- ably modified is necessary before further progress is
ification in one can be accommodated immediately, possible. For example, suppose the limit was imposed
by appropriate non-genetic adjustments, in others. by the inability of the ventilation system to acquire
Indeed as will be addressed later, this requirement enough gaseous oxygen for any higher metabolic rate.
underlies the main test of the hypothesis. For Further progress towards endothermy would have to
instance, the supposed heritable increase in mitochon- await a mutation that resulted in, say, a larger lung
drial number could only affect metabolic rates if there capacity or a higher ventilation rate. Or suppose that
was an existing ability of the lungs, heart and arterial the conductivity of the skin proved to be too high for
system to deliver increased oxygen to the tissues. For any higher level of metabolic heat to be effectively
the potential functions of the elevated metabolic rate retained. A mutation causing a decreased conductivity
to be manifested, the existing central and peripheral would be awaited, at which point a higher level of heat
nervous systems must also be able to control the production would now be viable, thereby enhancing
enhanced locomotory activity; the food-collecting and the incipient thermoregulation.

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
482 T. S. KEMP

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

Figure 3. The interrelationships of the structures and functions responsible for or affected by endothermic temperature
physiology of a mammal.

Figure 3 illustrates the principle of the evolution of when one particular feature is critical in the sense
endothermy by correlated progression. All the original that it is limiting the changes in others, but the limit
proposed functions (in boxes) are now seen to be will be duly removed when that feature is itself subject
deeply and inextricably embedded within the complex to an appropriate mutational modification.
functioning of the whole. Other structures and func- The theory of the origin of endothermy presented
tions could readily be added, and of those shown, most here is that the separately recognizable functions of
represent complex attributes in their own right: as endothermy in mammals evolved as a single, inte-
well as fitting congruently into the overall scheme, a grated complex, and that it is meaningless to suppose
character such as ‘gait’ or ‘brain’ itself consists of inte- that any one function was ever primary. Over the time
grated parts. Solely for illustrative purposes, this sim- it took, small incremental changes in individual fea-
ple account of how the correlated progression model tures occurred in parallel, and this produced small,
might work for the origin of endothermy started with integrated enhancements of all the functions in corre-
a hypothesized increase in the number of mitochon- lation with one another.
dria per cell. In fact any other structure or process
could equally well have been chosen to begin with, and
TESTING THE HYPOTHESIS
the characters could have changed, increment by
increment, in virtually any order. The essential point It is entirely plausible that this interpretation of the
is that only a small change in any one characteristic at process by which the complex, multifunctional charac-
any moment in evolutionary time is possible, but the teristic called endothermy arose is correct, but how to
characteristic in question can be one of many. It is nei- test it as a scientific hypothesis may seem problem-
ther important nor predictable which one it will be. atic. The correlated progression hypothesis predicts
There will be moments in the evolutionary sequence that the exact sequence of very small, successive

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 483

changes associated with the evolution of fully oxygen at each stage is dependent on that of the pre-
expressed mammalian endothermy will be far below vious stage, suggesting that an increase in transport
any possibility of the fossil record to resolve. Nor can rate at one point in the system could automatically be
evidence from living amniotes reveal the sequence, accommodated by increases at others.
because the ectotherms show few relevant signs of There is direct evidence, notably in aquatic chelo-
incipient endothermy, and the mammals all evolved nians (e.g. Jackson & Prange, 1979), concerning the
from a common ancestor already possessing full endot- ability of the incompletely divided reptilian heart to
hermy. What can be done is to apply tests at a lower adjust its output between virtually complete mixing
resolution, in other words to assess whether the cor- and virtually complete separation of the systemic and
related progression model is the most realistic general pulmonary flows. This implies that cardiac output

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

model for accounting for the origin of endothermy. If could cope immediately with an evolutionarily
this proves to be so, then it has to be accepted that imposed increased oxygen requirement.
seeking a single, specified initial selection force or A second relevant category of study of living organ-
function for endothermy is in vain. Nor will it be pos- isms is comparative. Considering the relationship
sible to discover a particular sequence of evolutionary between metabolic rate and thermoregulatory struc-
steps by which full mammalian endothermy evolved. tures, Shkolnik’s (1980) study of hedgehogs from three
There are three kinds of argument or test that do, different habitats referred to earlier showed that
indeed, support the general correlated progression although the same body temperature of 34 °C was
model: (i) inference from the nature of the integration found in all three, the BMR varied. Furthermore, the
between the structures and processes associated with conductance value of the skin varied, but in contrast
endothermy in living amniotes; (ii) inference from the rate of evaporative loss under hot ambient condi-
computer modelling of how complex systems in gen- tions was approximately the same. In all three species,
eral evolve under a selection process; and (iii) demon- the values of the parameters coincided with expecta-
stration from the fossil record that what relatively tions for, respectively, a temperate, a semi-arid and a
little is known about the pattern of acquisition of char- desert-adapted species of otherwise very similar,
acters associated with endothermy in mammals is closely related mammals. The implication from this
nevertheless more closely compatible with correlated and numerous similar cases of small differences in
progression than with alternative models. various parameters found in the particular versions of
endothermy among various mammals is that different
features associated with endothermy can evolve to
(I) INTEGRATION BETWEEN STRUCTURES AND some extent independently of one another, and yet
PROCESSES IN LIVING MAMMALS remain integrated with other features to the extent
The model predicts that the quantitative variability in that the overall temperature physiology of each spe-
function of virtually all the integrated individual pro- cies is adapted for the particular ambient conditions
cesses is sufficient to allow instant adjustment to the that its members normally meet.
effect of a genetic modification of any one of them. A Another, particularly illuminating illustration of
large, though disconnected, volume of observational the nature of the correlation between different fea-
and experimental work indicates that the range and tures is the relationship between the high metabolic
pattern of variability in the parameters of tempera- rate of the visceral organs that is associated mainly
ture physiology are indeed compatible with the with the high BMR, and that of the muscle tissues
required degree of loose functional correlation. There associated mainly with the high MAMR. None of the
are several categories of such evidence, a few exam- various attempts mentioned earlier to discover a func-
ples of which will illustrate the argument. tional connection between them that accounts both for
Some studies show that appropriate adjustments in the general correlation and for the variation in the
existing functions in response to a change in another ratio among different species is convincing. The corre-
occur. For example, Hicks & Farmer (1999) used com- lated progression model not only explains the connec-
parative physiological literature and a theoretical tion, but predicts it. The two metabolic rates, BMR
analysis to show that the basic reptilian lung struc- and MAMR, evolved as separate features in so far as
ture is capable in principle of supporting the increased they are directly associated with the separate respec-
oxygen requirements of endothermic physiology; tive functions of thermoregulation and increased
indeed, the oxygen consumption of some modern liz- activity. But, as illustrated in Figure 3, these functions
ards exceeds that of some living endotherms of the are themselves correlated with one another indirectly
same body weight. Furthermore, the oxygen transport because they are related through many other struc-
system from lungs to tissues consists of a series of tures and functions, and so they both contribute to the
steps, diffusive and convective, that are closely inte- multiple, inextricably intertwined functions of endo-
grated with one another. The partial pressure of thermy in the life of the mammal in its habitat. At the

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
484 T. S. KEMP

same time, however, the correlation is sufficiently runs was a considerable range of morphologies (Fig. 4)
loose that species in different niches with different which all represent compromise structures that are
particular requirements are able to evolve different more or less equally well adapted. Interesting too, the
quantitative values for these parameters. various outcomes were frequently similar in general
A third kind of evidence is derived from studies of form to known fossil primitive plants.
responses to artificially imposed changes. For exam- The inferences to be drawn from this study are
ple, the mammalian lung has a marked ability to three-fold, and by analogy they point to the correlated
undergo compensatory growth in response to damage, progression pattern of evolution. The first is that nat-
as reviewed recently by Hsia (2004). This indicates ural selection is concerned with the total fitness of the
how readily a small, phenotypic adjustment to organism, as contributed to by all its structures and

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

increased oxygen requirements could occur in this their functions simultaneously. Perhaps in some sin-
particular part of the oxygen transport system. In fact, gle generation one character alone might be critical
the entire concept of developmental feedback mecha- and therefore variation in that character alone will
nisms that induce compensatory changes in one sys- determine the direction of evolution. But this will be
tem, be it neural, muscular, osteological, enzymatic, at most a transient stage and normally variation in
etc., in response to a genetically or environmentally many characters will contribute to the overall fitness
induced modification in another – one of the funda- level. The best compromise among the conflicting
mental properties of life – speaks for the correlated optima for several discrete functions will be the vari-
progression model. ant that is selected. The second inference is a corollary
of the first. There will be many possible sequences of
increasingly well-adapted compromises. The sequence
(II) INFERENCE FROM COMPUTER MODELLING OF HOW actually followed in a particular case will depend on a
COMPLEX BIOLOGICAL SYSTEMS EVOLVE variety of contingent circumstances, such as the exact
Niklas (1995, 1997, 2000) used a computer simulation variation available for selection to act on, and the fine
program for evolution by natural selection of hypo- details of the environment of the moment (see Mani &
thetical simple terrestrial plants. He considered the Clarke, 1990, for a computer model that makes a vary
optimal requirements for each of up to four functions similar point but at the level of intrapopulation genet-
in such plants, namely mechanical stability (low or ics). The third inference is that at any given time,
closely spaced branches), light interception (widely change in many of the characters will not be due to
spreading branches), water conservation (low plant natural selection acting directly on those characters.
with few branches) and spore dispersal (multiple high Rather, these changes will be due to a form of drift as
branches). The program was arranged to select at each they hitchhike as part of those variant organisms that
step the randomly generated variant with the best are selected because they momentarily possess a more
compromise structure between the conflicting require- favourable overall combination of characters. Indeed,
ments for these functions. The result of the series of a variant of a character could well survive even

Figure 4. On the left, computer-generated walks through a multi-task landscape requiring adaptation simultaneously for
light interception, mechanical stability and reproductive success. On the right, some of the optimal compromise morphol-
ogies generated by different walks (from Niklas, 1995).

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 485

though it is mildly disadvantageous as far as its par- characters associated with several different functions
ticular function is concerned. were found to have changed. On virtually every occa-
Lenski et al. (2003) used a logically comparable but sion these include characters associated, respectively,
much more sophisticated computer program to inves- with locomotion and feeding, but usually others as
tigate the evolution of complex characters. It starts well, such as those associated with ventilation, sense
with simple, replicating digital ‘organisms’ but which organs and brain size. For example, comparing the
have the possibility of acquiring certain random reconstructed hypothetical common ancestor of sphen-
‘mutations’ (logical functions, in computer terms). acodontid pelycosaurs and therapsids with that of the
Several particular ‘mutations’ increase the units of therapsids indicates that, among others, the following
‘energy’ (computational merit) available to the organ- evolutionary changes had occurred:

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

ism. Complex structure is represented by particular
1. enlargement of the canines;
combinations of several such favourable ‘mutations’,
2. enlargement of the temporal fenestra and reflected
which have ever-greater units of ‘energy’ available,
lamina of the angular;
and which therefore represent ever-greater ‘reproduc-
3. gracile limb bones and a facultative parasagittal
tive fitness’. The maximum ‘energy’ is available to a
gait;
stage referred to as EQU, which in principle could be
4. less constrained shoulder joint;
achieved by as little as 16 ‘mutations’, as long as they
5. reduced digit length;
were the right ones appearing in the right order. The
6. lighter stapes indicating more acute hearing.
actual outcome of the simulations was that of 50 runs,
EQU was achieved 26 times, but that the number of The transition from the common ancestor of the thero-
‘mutations’ occurring in these respective lineages to cephalians plus cynodonts to the common ancestor of
the point at which EQU arose varied from 51 to 721 the cynodonts can likewise be inferred to have
steps. The precise sequence of ‘mutations’ by which involved change in a variety of its functional systems,
EQU was achieved also varied, illustrated starkly by including:
the fact that the final ‘mutation’ leading to EQU dif-
1. mammalian style of tooth differentiation, including
fered from lineage to lineage. In the course of evolu-
multicusped postcanines;
tion of the lineages, no particular ‘mutation’ was
2. further enlargement of temporal fenestra, and coro-
essential for the subsequent appearance of EQU, and
noid process of dentary;
there were cases of ‘mutations’ that were deleterious
3. reduction of postdentary bones and quadrate, asso-
when they first arose, or that were involved in a
ciated with enhanced high-frequency hearing;
tradeoff with more favourable ‘mutations’.
4. nasal cavity with a more complex internal struc-
Like the simpler, morphological analogy program of
ture, associated with olfaction.
Niklas, here too the pattern of acquisition of charac-
5. secondary palate;
ters leading to the evolution of a new complex system
6. enlarged brain;
that is composed of functionally interrelated parts cor-
7. full differentiation of vertebral column, including a
responds to a correlated progression. Selection is con-
basically mammalian atlas-axis;
cerned with the overall operation of the system rather
8. enlarged ilium and reduced pubis.
than its individuated parts, the parts are functionally
correlated, and there are so many routes to achieving These two transitions are the largest morphological
the same overall level of complexity of function that it shifts within the fossil record of non-mammalian syn-
is not possible to predict which precise route will be apsids. Whether this is because there really was a
followed. In so far as the evolution of complex biolog- rapid evolutionary transition or whether it is an arte-
ical structure under the influence of natural selection fact of missing intermediate-grade fossils cannot of
can be expected to follow the same logical rules as course be ascertained. However, Sidor & Hopson
these computer simulations, then the correlated pro- (1998) showed a surprisingly good correlation
gression model is corroborated. between the number of new characters at each of the
nodes of their synapsid cladogram and the estimated
time between the nodes, as inferred from the dates of
(III) INFERENCE FROM THE SEQUENCE OF ACQUISITION appearance in the record of the relevant fossils. At
OF CHARACTERS INFERRED FROM THE FOSSIL RECORD the admittedly limited taxonomic resolution that this
Kemp (1982) explicitly inferred the pattern of acqui- represents, it is consistent with a relatively constant
sition of mammalian characters in the synapsids from rate of acquisition of mammalian characters, which
the cladogram of the known fossil members, and too is what the correlated progression model would
showed that it is consistent with a correlated progres- predict.
sion. At each hypothetical stage that can be recon- The existing synapsid fossil record is inadequate to
structed from the morphology of a known fossil form, illustrate anywhere near the full sequence of acquisi-

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
486 T. S. KEMP

tion of mammalian characters, including those associ- descendant morphology, from original to new habitat
ated directly with endothermy, and there are many or way of life, was possible.
physiological characters whose evolutionary history is 3. Given the assumed loose correlation between the
unlikely ever to be revealed. On the other hand, at the parts, it is easy to explain how relatively minor vari-
level of the taxonomic and organismic resolution that ations, of the kind associated with radiation at the
it does have, the fossil record is fully compatible with species and genus level, can occur at every stage along
correlated progression. There is no anatomical evi- the main evolving lineage. The fossil record of the
dence indicating unambiguously that any one partic- synapsids illustrates this well, where radiations of a
ular function was ever the sole focus of selection at any variety of differently adapted kinds of taxa occur
particular time. at several levels, pelycosaur, basal therapsid and

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

cynodont.
4. There is a degree of randomness about the evolu-
CONCLUSIONS tion of individual characters. In some cases a variant
The integrated theory of the origin of endothermy pre- of a character at any particular time may not be
sented here is that the several individually definable selected even though it would improve the efficiency of
functions of endothermy, and the structures and pro- the function it directly serves. Conversely, a variant of
cesses responsible for it, all evolved in a loosely corre- a character may be fixed because it happens by chance
lated progression, a small step in this structure here, to be present in what is the overall fittest phenotype,
a small step in that process there, and so on. As so while itself being selectively neutral or even to a
expressed, even this is a limited view, because endo- degree deleterious. Thus, the evolutionary pathway
thermy itself cannot be disentangled as a discrete followed by the lineage perhaps wanders rather more
function from its inextricable integration with all the aimlessly through morphospace than often supposed.
features of the biology of the organism, as illustrated
in Figure 3. Virtually everything in the biology and
life of a mammal is either contributory to, or affected REFERENCES
directly or indirectly by, the endothermic temperature Allman J. 2000. Evolving brains. New York: Scientific Amer-
strategy. ican Library.
The theory reconciles the various previous theories, Angilletta MJ, Sears MW. 2003. Is parental care the key to
in the sense that each of them is regarded as partly understanding endothermy in birds and mammals? The
correct. It also appears to correspond with realistic American Naturalist 162: 821–825.
assumptions about how complex biological structure, Bennett AF, Hicks JW, Cullum AJ. 2000. An experimental
and therefore new higher taxa that manifest it, can be test of the thermoregulatory hypothesis for the evolution of
expected to evolve. From this broader perspective, endothermy. Evolution 54: 1768–1773.
there are several corollaries. Bennett AF, Ruben JA. 1979. Endothermy and activity in
vertebrates. Science 206: 649–654.
1. No single, identifiable selection pressure for any Bennett AF, Ruben JA. 1986. The metabolic and thermoreg-
one characteristic of an organism persists beyond the ulatory status of therapsids. In: Hotton I, Nicholas MacLean
possible brief periods when a particular structure or PD, Roth JJ, Roth EC, eds. The ecology and biology of
function may be limiting further evolutionary change mammal-like reptiles. Washington, DC: Smithsonian
of the whole organism. Because the proposed change Institution, 207–218.
Boily P. 2002. Individual variation in metabolic traits of wild
possible in any one characteristic at any one time is
nine-banded armadillos (Dasypus novemcinctus) and the
presumed always to be small, this will never be more
aerobic capacity model for the evolution of endothermy.
than a transient stage, perhaps represented by a sin-
Journal of Experimental Biology 205: 3207–3214.
gle or a few speciation events.
Botha J, Chinsamy A. 2000. Growth patterns deduced from
2. The exact pattern of small changes from the initial
the histology of the cynodonts. Journal of Vertebrate Paleon-
ancestral condition to the final derived condition will tology 20: 705–711.
be below the resolution of any available evidence to Budd GE. 1998. Arthropod body-plan evolution in the Cam-
illustrate, be it palaeontological or neontological. brian with an example from anomalocaridid muscle. Lethaia
Therefore, the question to ask when seeking to explain 31: 197–210.
a major evolutionary transition such as the origin of Carrier DR. 1987. The evolution of locomotor stamina in tet-
mammals is not, as traditionally, what particular rapods: circumventing a mechanical constraint. Paleobiology
aspect of the environment provided the selection pres- 13: 326–341.
sure. Rather, the research programme should be Cowles RB. 1958. Possible origin of dermal temperature reg-
focused on how the attributes of the organisms in ulation. Evolution 12: 347–357.
question were functionally interrelated to one another Crompton AW, Taylor CR, Jagger JA. 1978. Evolution of
such that a correlated progression from ancestral to homeothermy in mammals. Nature 272: 333–336.

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
 ORIGIN OF ENDOTHERMY 487

Dohm MR, Hayes JP, Garland T. 2001. The quantitative Hulbert AJ, Else PL. 2000. Mechanisms underlying the cost
genetics of maximal and basal rates of oxygen consumption of living in animals. Annual Review of Physiology 62: 207–
in mice. Genetics 159: 267–277. 235.
Eisenberg JE. 1981. The mammalian radiations: an analysis Jackson D, Prange HD. 1979. Ventilation and gas exchange
of trends in evolution, adaptation, and behaviour. London: during rest and exercise in adult green sea turtles. Journal
The Athlone Press. of Comparative Physiology 134: 315–319.
Farmer CG. 2000. Parental care: the key to understanding Jerison HJ. 1971. More on why birds and mammals have big
endothermy and other convergent features in birds and brains. American Naturalist 105: 185–189.
mammals. The American Naturalist 155: 326–334. Kemp TS. 1969. On the functional morphology of the gor-
Farmer CG. 2003. Reproduction: the adaptive significance of gonopsid skull. Philosophical Transactions of the Royal

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

endothermy. The American Naturalist 162: 826–840. Society B 256: 1–83.
Gomes FR, Chaui-Berlinck J-G, Bicudo J-EPW, Navas Kemp TS. 1979. The primitive cynodont Procynosuchus: func-
CA. 2004. Intraspecific relationships between resting and tional anatomy of the skull and relationships. Philosophical
activity metabolism in anuran amphibians: influence of ecol- Transactions of the Royal Society B 285: 73–122.
ogy and behavior. Physiological and Biochemical Zoology 77: Kemp TS. 1982. Mammal-like reptiles and the origin of mam-
197–208. mals. London: Academic Press.
Harvey PH, Krebs JR. 1990. Comparing brains. Science 249: Kemp TS. 1985. Synapsid reptiles and the origin of higher
140–146. taxa. Special Papers in Palaeontology 33: 175–184.
Harvey PH, Pagel MD. 1991. The comparative method in evo- Kemp TS. 1986. The skeleton of a baurioid therocephalian
lutionary biology. Oxford: Oxford University Press. therapsid from the Lower Triassic (Lystrosaurus zone) of
Hayes JP, Garland T. 1995. The evolution of endot- South Africa. Journal of Vertebrate Paleontology 6: 215–
hermy: testing the aerobic capacity model. Evolution 49: 232.
836–847. Kemp TS. 1999. Fossils and evolution. Oxford: Oxford Univer-
Hicks JW, Farmer CG. 1999. Gas exchange potential in rep- sity Press.
tilian lungs: implications for the dinosaur–avian connection. Kemp TS. 2005. The origin and evolution of mammals.
Respiration Physiology 117: 73–83. Oxford: Oxford University Press.
Hillenius WJ. 1992. The evolution of nasoturbinates and Koteja P. 2000. Energy assimilation, parental care and the
mammalian endothermy. Paleobiology 18: 17–29. evolution of endothermy. Proceedings of the Royal Society of
Hillenius WJ. 1994. Turbinates in therapsids: evidence for London B 267: 479–484.
Late Permian origins of mammalian endothermy. Evolution Krosniunas EH, Gerstner GE. 2003. A model of vertebrate
48: 207–229. resting metabolic rate: balancing energetics and O transport
Hinds DSR, Baudinette RE, MacMillen RE, Halpern EA. in system design. Respiratory Physiology and Neurobiology
1992. Maximum metabolism and the aerobic factorial scope 134: 93–113.
of endotherms. Journal of Experimental Biology 182: 41–56. Lee MSY, Doughty P. 1997. The relationship between evolu-
Hopson JA. 1979. Paleoneurology. In: Gans C, Northcutt RG, tionary theory and phylogenetic analysis. Biological Reviews
Ulinski P, eds. Biology of the reptiles. London: Academic 72: 471–495.
Press, 39–146. Lenski RE, Ofria C, Pennock RT, Adami C. 2003. The evo-
Hopson JA. 1991. Systematics of nonmammalian Synapsida lutionary origin of complex features. Nature 423: 139–144.
and implications for patterns of evolution in synapsids. In: Mani GS, Clarke BC. 1990. Mutational order: a major sto-
Schultze H-P, Trueb L, eds. The origin of higher groups of tet- chastic process in evolution. Proceedings of the Royal Society
rapods: controversy and consensus. Ithaca, NY: Cornell Uni- of London B 240: 29–37.
versity Press, 635–693. McNab BK. 1978. The evolution of endothermy in the phylog-
Horner JR, de Ricqles A, Padian K. 1999. Variation in dino- eny of mammals. The American Naturalist 112: 1–21.
saur skeletochronology indicators: implications for age Niklas KJ. 1995. Morphological evolution through complex
assessment and physiology. Paleobiology 25: 295–304. domains of fitness. In: Fitch WM, Ayala FJ, eds. Tempo and
Hsia CCW. 2004. Signals and mechanisms of compensa- mode in evolution. Washington, DC: National Academy
tory lung growth. Journal of Applied Physiology 97: Press, 145–165.
1992–1998. Niklas KJ. 1997. The evolutionary biology of plants. Chicago:
Hulbert AJ. 1980. The evolution of energy metabolism in Chicago University Press.
mammals. In: Schmidt-Nielsen K, Bolis L, Taylor CR, eds. Niklas KJ. 2000. Modelling fossil plant form–function rela-
Comparative physiology: primitive mammals. Cambridge: tionships: a critique. In: Erwin DH, Wing SL, eds. Deep time:
Cambridge University Press, 130–139. paleobiology’s perspective. Washington, DC: The Paleontolog-
Hulbert AJ, Else PL. 1989. Evolution of mammalian endo- ical Society, 289–304.
thermic metabolism. American Journal of Physiology 256: Ray S, Botha J, Chinsamy A. 2004. Bone histology and
R63–R69. growth patterns of some nonmammalian therapsids. Jour-
Hulbert AJ, Else PL. 1990. The cellular basis of endothermic nal of Vertebrate Paleontology 24: 634–648.
metabolism: a role for ‘leaky’ membranes? News in Physio- Reisz RR. 1986. Pelycosauria. Gustav Fischer Verlag,
logical Sciences 5: 25–28. Stuttgart.

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488
488 T. S. KEMP

Rowe T. 1996. Coevolution of the mammalian middle ear and Sidor CA, Hopson JA. 1998. Ghost lineages and ‘mammal-
neocortex. Science 273: 651–654. ness’: assessing the temporal pattern of character acquisi-
Ruben J. 1995. The evolution of endothermy in mammals: tion in the Synapsida. Paleobiology 24: 254–273.
from physiology to fossils. Annual Review of Physiology 57: Taylor CR. 1980. Evolution of mammalian homeothermy: a
69–95. two-step process?. In: Schmidt-Nielsen K, Bolis L, Taylor
Schweitzer MH, Marshall CL. 2001. A molecular model for CR, eds. Comparative physiology: primitive mammals. Cam-
the evolution of endothermy in the theropod-bird lineage. bridge: Cambridge University Press, 100–111.
Journal of Experimental Zoology (Molecular Development Taylor CR, Karas RH, Weibel ER, Hoppeler H. 1987. Adap-
and Evolution) 291: 317–338. tive variation in the mammalian respiratory system in rela-
Seebacher F. 2003. Dinosaur body temperature: the occur- tion to energetic demand. Respiration Physiology 69: 7–26.

Downloaded from https://academic.oup.com/zoolinnean/article-abstract/147/4/473/2629149 by guest on 19 August 2019

rence of endothermy and ectothermy. Paleobiology 29: 105– Xu X, Norell MA, Kuang X, Wang X, Zhao Q, Jia C. 2004.
122. Basal tyrannosaurids from China and evidence for
Shkolnik A. 1980. Energy metabolism in hedgehogs: ‘primi- protofeathers in tyrannosaurids. Nature 431: 680–684.
tive’ strategies?. In: Schmidt-Nielsen K, Bolis L, Taylor CR,
eds. Comparative physiology: primitive mammals. Cam-
bridge: Cambridge University Press, 148–154.

© 2006 The Linnean Society of London, Zoological Journal of the Linnean Society, 2006, 147, 473–488