ecomorphological diversity of mesozoic mammals

Gareth Coleman gc12847 EASC30048 Palaeobiology Analytical Project Literature Review

1

The Ecomorphological diversity of Mesozoic mammals

Gareth Andrew Coleman

[email protected]

Abstract

There has been much recent progress in understanding the evolution and adaptive

radiations of early mammals and mammaliaformes, showing successive waves of

ecomorphological diversification and adaptive radiations. This is in contrast to many earlier

views on the early evolution and radiation of mammals, which characterised them

exclusively as small, insectivorous and generalised animals, with litter ecomorphological

diversity. Much of the progress has come from new fossil finds and the integration of many

techniques, such as dental microwear analysis, CT-scanning and finite element analysis.

These integrated techniques create a robust set analytical tools, and give us the potential

to increase our understanding of the early history of mammals.

Keywords: ecomorphology, diversity, mammals, Mesozoic

Main text word count: 3500

Total word count: 7635


2

Declaration

'I declare that this library project is entirely my own work, and does not contain any

plagiarised material'

Signature:

Name: Gareth Coleman

Date: 17/12/2014


3

Introducing the debate: past views on mammalian evolution

Mammals are one the most ubiquitous groups of animals present on Earth today

(Box 1). They exhibit much diversity and occupy a vast array of different ecological niches

on the land, in the oceans and in the air [1] [2]. There has been much research concerning

the diversification of mammals, and their adaptive radiations to fill the many niches they

presently hold, in an attempt to discern the key to their success and the timing of such

diversification. In particular, much of the recent research, coupled with many new fossils

being discovered at various localities, has shed more light and altered our views on

mammalian evolutionary history [3] [4] [5].

In the past, the traditional narrative associated with the early evolution of mammals

described them as small, living in the shadows of the much larger archosaurs which

dominated the ecosystem [6] [7]. They were therefore thought to have only exploited

niches where there was no direct competition with the archosaurs [6]. This limited

exploitable niches, and stopped specialisation, causing them to remain generalised [8].

This was thought to have forced early mammals to become largely insectivorous and

nocturnal. Becoming increasingly insectivorous would have further accelerated the

therapsid trend towards differentiated teeth with precise occlusion, in the need to capture

arthropods and crush their exoskeletons. It would also have restricted the size of the

animals, as they would not have been able to eat enough to sustain themselves [9]. The

decrease in body size, and becoming nocturnal, would have led to the need for insulation

in the form of fur, and thermoregulation [9] [10]. Nocturnal life was also thought to lead to

more acute senses, including acute sense of smell, and therefore a larger brain [11], and

loss of two of the four opsin cones present in amniotes, giving dichromatic vision, which

improved low-light vision [9].


4

Much of the evidence from the fossil record supported, and to a certain extent still

does support this model of mammal evolution. Many of the early mammal fossil are small

and insectivorous, such as Morgonucodon, Megazostrodon, Yanoconodon and

Zhangheotherium [3]. However, new techniques are now being used to uncover the real

diversity exhibited by these animals.

Looking at morphology

The most direct source of information on mammalian diet are the contents of its

stomach. However, these are rarely preserved in fossil mammals (with a few remarkable

exceptions), therefore morphology has historically been the best way to assess the diets of

extinct organisms. Ecomorphology is the relation between the morphology of an organism

and its feeding and ecological role. We can look at the morphologies of modern mammals

and how it corresponds with their diets, before applying this to fossils mammals. Many

modern mammal groups have evolved distinct morphologies in order to exploit particular

food sources and ecological niches.

The size of the animal is often important in giving information on their general diet.

Small mammals have high metabolic rates due to high ratio of heat-losing surface area to

heat-generating volume, and therefore need to feed much more often than larger

mammals. Subsequently, most small mammals are insectivorous, as they cannot tolerate

slower rates of food intake or digestion. Larger mammals can generate more heat, and

less of it is lost. They can therefore tolerate slower collection rate of food (if carnivorous) or

slower digestion (if herbivorous). Generally, larger mammals are not insectivorous, as they

would not be able to ingest enough food to sustain themselves, except for a few species

which feed on large insect colonies such as ants and termites. Therefore, the size of the

fossil mammal can potentially indicate its diet [1] [9].


5

Teeth and jaws are specialised and also convey information on the diet of the

animal. Mammals are heterodont, i.e. they have different kinds of teeth, with different

functions. Many insectivorous animals have very small, reduced teeth or lack dentition

altogether (such as anteaters). Herbivorous mammals generally have large molars for

grinding up vegetation, and some have blade-like incisors for cutting leaves. Specialised

herbivores, such as granivores (seed eaters) have larger, more robust molars for crushing

seeds, while gummivores (gum eaters), have tooth combs made up from the lower incisors

and canines. Carnivorous mammals have sharp teeth with large canines and carnassials

and strong jaws for a big bight force. Filter feeders, such as baleen whales, have baleen

plates for filter feeding. The teeth themselves often exhibit microwear patterns, which can

be used as a powerful tool to assess diet. Recently, analytical and numerical methods to

analyse microwear patterns have developed, as shall be discussed later [1].

There are also many other morphological adaptations to feeding. Long necks in

some mammals could have evolved so that it can take advantage of the leaves higher in

the trees, with trunks or probosces of some species being used for similar purposes, with

modern examples being giraffes and elephants. Many carnivorous mammals which

actively hunt have long legs for running and long tails for balance. These adaptation are

seen in many large cat species. Many insect eating mammals have hypertrophied limbs

built for scratch-digging and fossorial behaviours (e.g. aardvark, pangolin, armadillo and

echidnas). These may also been seen in exaptations for swimming, as in beavers, coupled

with webbed feet and tail. Mammals with these aquatic adaptation could be inferred to be

piscivorous. Scansorial (climbing) mammals often have elongated phalanges and limbs for

climbing, these being exaptations for gliding in volant mammals (e.g. in flying squirrels) [1].


6

Modern analytical methods to determine ecomorphology

As well as looking at morphology, a variety of different analytical methods can now

be applied to the remains of fossils mammals to determine their ecological niches and

diets. The most important among these are finite element analysis and microtextural

analysis of dental microwear.

Finite element analysis (FEA) is used to reconstruct stress, strain and deformation

in structures. [12] [13]. It can give insights into the mechanics of various different body

parts of fossil organisms. This can be used to determine function of skeletons and the

ecology of the mammals analysed, and why evolution has shaped the bones in a particular

fashion [14] [15] [16]. Analysis of the effects of forces on structures was hitherto

impossible, as the models formed by differential equations derived from first principles

were impossible to solve, unless they were the simplest geometric shapes with the

simplest boundaries. FEA can overcome these problems by breaking the structures down

into disjoint components of simple geometry, called finite elements, or elements. This

process is called discretisation, as it takes a continuous structure, and represents it as a

series of discrete problems, which are readily solvable by mathematics. The discrete

model is obtained by connecting all of the elements by nodes to form a finite element

mesh.

The first part of the process (Figure 1), the preprocessing stage, involves the

creation of a digital representation of the structure using computer-aided design (CAD).

The structure then undergoes discretisation. The elements in the mesh can then be

assigned specific material properties that represent the elasticity of the real structure.

Virtual loads are then added at the nodes. Constraints are also added, with mobility being

restricted to particular degrees of freedom. The loads and constraints applied at this stage

are collectively called the boundary conditions. When running the analysis, nodal


7

displacements are calculated in response to boundary conditions, taking the structure’s

predefined geometry and elasticity into account. These can be used to calculate structural

strain, stress and deformation, building up a picture of the mechanical behavior of a

structure under the predefined conditions. After the analysis, the final, postprocessing step

involves the representation and interpretation of the results, usually as a digital image. The

images usually have scaled colour plots and can be animated to represent structural

deformities. Results may be used to assess the accuracy of the mesh and boundary

conditions, and convergency tests can be run to assess how well the discrete model

represents the original structure. The analysis can rerun with successively smaller

elements to get more accurate results [17] [18].

While FEA is a powerful tool, it has limitations and problems, mostly the concern

that there are too many problems and assumptions associated with modelling biological

systems, particularly a fossil organism where less data is available. The main areas to

which this applies are the model creation and the application of boundary conditions. One

of the most fundamental issues faced is the level of detail which should be used, and how

this effects the accuracy of the analysis. Ultimately, this has to be a trade-off between how

accurate the model will be compared to the amount of time and computer power needed to

complete the analysis. The answer to this is complex and depends on the areas of the

skeleton that are being considered. It may be possible for certain areas of the skeleton to

be simplified more than others, while some parts have to be done in more detail to be

accurate. This also relates to whether one wishes to use 2-D or 3-D models, which will

depend on what questions are being asked. When creating the mesh, users must choose

between 2-D triangular or quadrilateral, or 3-D tetrahedral or cuboidal elements, and

whether linear of quadratic elements are required. Generally, the most accurate analyses

would be generated with cuboidal, quadratic elements, but these analyses require much


8

time and computing power, maybe more than is practical [19]. Material properties also

have to be added to the elements, which can be problematic when dealing with fossil

organism, where data on material properties is not necessarily available. The applications

of boundary conditions also present problems. Knowing which loads would have been

applied to the skull of the organism in life is impossible to know, and modern analogues

(using extant phylogenetic bracketing) are essential in predicting the behavior and

possible usages of the particular anatomy in question. Further, while constraints must be

applied for the analysis to work, applying too many can prevent the model from deforming

naturally and give inaccurate results [20].

The way we can validate the FEA, and therefore combat many of these problems, is

by doing FEA analyses on living animals and seeing what levels of detail and how

accurate material properties and boundaries need to be in order for the analysis to work.

So far, it seems that these have been done exclusively on vertebrates, with the macaque

monkey being the most used animal. FEA studies on the eating mechanics of the

macaque skull have shown that accurate models can be produced if the loads applied to

the muscles are accurate, and that muscle activation patterns are not an important

consideration. This means that palaeontologists only have to worry about reconstructing

the musculature of the animal. Biological materials are very complex and may necessitate

a number of simplifications. Further studies on macaque skulls have shown that, although

more accurate modelling of material properties improved the accuracy of the analysis, the

same broad patterns were repeated in more simplified models. Generally, it seems

possible to predict the nature of deformation, even when detailed information about the

material properties are not known. Material distribution may also be have an important

effect, especially in the different types of bone with different densities, which are not

readily identifiable from fossil remains. Some studies in macaques have shown that the


9

bone of the lower jaw can be represented by a solid mass of low Young’s modulus [20],

and fine anatomic detail is not necessary. However, this has not been widely applied to

other animals, with may have different arrangements of bones type in their jaws. Where

these have been done, the results vary considerably from the macaque and from each

other. Also, the problem of how teeth are set in the jaw (i.e. set directly or via ligaments)

and how this affects the models, continues to be debated [20].

FEA has had only limited use in palaeontology thus far [20]. However, important

papers, such as those on the feeding mechanics of theropod dinosaurs by Emily Rayfield

[21] [22], particularly a landmark paper in 2001 where FEA was applied to the skull of

Allosaurus [23], and in early mammaliaforms in Gill et al. 2014 [24], as well as several

important papers on various fossil mammals, have shown the potential impact that the use

of FEA can have on palaeontology [25] [26] [27] [28].

Microtextural analysis of dental microwear is another powerfull technique which can

be used to assess mammalian diet. Dental microwear are the many pits and scratches

found on the enamel of teeth, which vary with diet. Many studies have been done on fossil

mammals, particularly early hominids and other primates [29] [30] [31] [32] [33].

Conventional method of microwear analysis have been limited to two-dimensional imaging

studies using scanning electron microscopes to identify features on the teeth. This was

time-consuming, and could be very subjective and prone to error. New three-dimensional

microtextural analysis has given us a way to reliably quantify and analyse dental

microwear patterns, saving time, minimising the error and making the analysis more

objective. These measurements are taken using white-light confocal microscopy and

scale-sensitive fractal analysis, and charactierise microwear surface textures as 3-D

images [34]. The data can then undergo statistical analysis, where samples can be

compared with each other and with modern analogues. This data can be used in a


10

principle component analysis (PCA), using International Organization for Standardization

(ISO) roughness parameters from animals that are being analysed [35]. The data is plotted

onto a set of axes, where different the results for different animals can be compared.

Current views on ecomorphological diversity in Mesozoic mammals

All of the evidence currently being uncovered, from new fossil finds to more

powerful analytical techniques, has brought about a drastic change in the way we think of

mammals in the Mesozoic. Certainly, while many of Mesozoic mammalian clades are still

seen to contain mainly relatively small, generalised insectivors (with these making up

much of the mammalian fauna in ecosystems such as the Jehol Biota [36]), many lineages

and clades exhibit much higher levels diversity and specialisation. We also see is

successive radiations of mammals throughout the Mesozoic, coupled with much more

ecomorpological diversity than was previously thought [3] [24].

The discovery of many of these new fossils of basal mammaliaforms (Boxes 2 and

3), especially from China [36], have greatly increased our knowledge of the earliest part of

mammal evolution and shown us that these early mammals were more diverse than

expected [3] [37]. However, while much research has been done on the diversity of

mammaliaforms from the Mid Jurassic to the Cretaceous, until recently, little had been

done on the very earliest mammaliaforms of the Late Triassic and Early Jurassic. Now,

studies are showing diversity in these earliest mammaliaforms. Using a suite of

techniques, including finite element analysis and microtextural analysis of dental

microwear, Gill et al. (2014) found that early mammaliaforms Morganucodon and

Kuehneotherium had different feeding ecologies, with Morganucodon specialising in eating

insects with hard cuticles, such as beetles, and Kuehneotherium appearing to specialise in

eating soft bodied insects, such as moths [24]. This early specialisation shows “previously


11

hidden trophic diversity and niche partitioning at the base of the mammalian radiation,

supporting a hypothesis of coupled lineage splitting and ecomorphological adaptation of

the skull and jaws, even during the earliest stages of mammalian evolution” [24]. This

paper was important in its use of finite element analysis combined with microtextural

analysis of microwear, classical mechanics and imaging techniques to verify their findings

from independent lines of evidence which could be integrated together to form a more

complete picture.

Most other evidence for mammalian diet come mainly from morphological data.

Many mammaliaform lineages show many adaptations to various behaviors and ecological

specialisations. Adaptations for fossorial (digging) behavior are seen in the hypertrophied

burrowing limbs in Hadanodon, a doconodont (Box 3) from the Upper Jurassic of Portugal,

which shows many convergent characteristics with desmans and moles. These

adaptations also represent exaptations for swimming in Castorocauda, another

doconodont, from the Middle Jurassic. Castrocauda is larger than most mammaliaforms of

the period and showed further adaptations to swimming, such as a flattened tail and

possibly webbed feet, among other features showing convergence with modern

platypuses, otters and beavers. It was possibly piscivorous. It is further important in

demonstrating the earliest evidence of fur [38]. Adaptions for fossorial behavior were

originally known only from pre-mammaliaform cynodonts [39], but have now also been

demonstrated to be widespread in mammaliaform lineages.

The multituberculates (Box 4) were a very diverse group. They occupied a variety of

different niches, ranging from fossorial to scansorial and arboreal lifestyles [40]. The

ptilodonts of the Upper Cretaceous of North America, show convergences with modern

squirrels, with legs and longs tails which suggest a scansorial or aboreal lifestyle. The

ptilodonts also had enlarged and elongated last lower premolars, which may have been


12

used to crush seeds, suggesting granivory [41]. A European family of multituberculates,

the kogaionids, also developed these large, blade-like lower premolars, suggesting similar

diet had convergently evolved. Another family, the taeniolabids, were larger, similar in size

to beavers, and heavily built, suggesting that they were fully terrestrial. Some

multituberculates have adaptations for eating wood, particularly shown in the enlarged

front teeth [41] [42]. The euharamiyids are a group thought to be closely related to the

multuberculates and seem to have included some of the first herbivorous mammals. Some

species showed elongated limbs and tails, similar to modern squirrels, suggesting a

scansorial lifestyle [43].

The eutriconodonts were another diverse groups. Fossorial and scansorial

behaviors seem to have been common in the eutriconodonts, as demonstrated by long

legs and tails and the ability to abduct and adduct their big toes [44]. Scansorial traits may

have become exaptations for gliding in the eutriconodont Volaticotherium, which shows

convergence with modern sugar gliders and flying squirrels [45]. Scansorial adaptations

are also widespread in other linneages, including early therians Henkelotherium and

Eomaia, and the early metatherian Sinidelphis [3] [41] [42] [44] [46].

An interesting mammal from the Late Jurassic of North America of uncertain

affinities is Fruitafossor. It seem have adaptation for feeding on ant and termites

(myrmecophagy) and has many convergent characteristics with anteaters and aardvarks

[3] [47]. Basal mammaliamorph tritylodontids have balde like teeth with suggested that

they ate leaves [48] [49].

Predation and scavenging are also represented in Castorocauda [38], and basal

mammaliamorph Sinoconodon, shown in their large canines and strong jaws [50]. Another

very interesting find is a relatively large gobiconodont, Repenomamus, which was found

with a juivinile Psittacosaurus preserved in its stomach, showing that some Mesozoic


13

mammals were carnivorous and prayed on small vertebrates, such as young dinosaurs

[51] [52] [53].

What this ecomorphological diversity tells us about the patterns of evolution and

adaptive radiation in early mammals is of particular importance. Mammal evolution seems

best characterized not by long branches with little ecomorphological diversity stretching

deep into the Mesozoic, but lots of short lived branches coupled with ecological

specialisation, representing many successive episodes or waves of diversification and

adaptive radiation throughout the Mesozoic (Figure 2a) [3]. These short lived, dead-end

lineages iteratively evolved developmental homoplasies and convergent ecological

specialisations, parallel to those in modern mammal groups (Figure 2b) [3]. This means we

can see successive lineages adapting into similar niches and evolving similar

morphologies at different times during the Mesozoic, long before the analogous modern

mammals [3] [24]. It seems that “correlation of ecomorphological specialisations with

phylogenetic splitting is a basic feature of Mesozoic mammal evolution” [3].

Concluding remarks

While there is still some truth in the stereotype of Mesozoic mammals being small,

generalised insectivores, there is much more ecomorphological diversity and niche

specialization in these mammals than was previously known. A range of different diets,

feeding ecologies and adaptations for various niches are exhibited, including

insectivorous, herbivorous, myrmecophagous, carnivorous, fossorial and scansorial

lifestyles, and even adaptations for swimming and gliding. We also see lots of short-lived

lineages representing burst of iterative evolution through-out the Mesozoic, rather than

long branches with little ecomorphological diversity. There are now a range of techniques


14

which can be integrated to continue to shed light on this area of evolutionary history, as

well as many others.

Future research should concentrate on using these techniques and applying them

to a wider range of fossils to try and build a more complete picture of the ecomoprhological

diversity of Mesozoic mammals. Most research up to now has largely been looking at

comparative morphology, with many of these analytical techniques having been used only

in limited contexts. In particular, the use of FEA in vertebrate biomechanics, especially

applied to extinct organisms, in still in its infancy, and has yet to be applied to a wide range

of living animals, let alone fossil organisms. Further research would be needed to further

validate the approach. Also, few studies have combined this with other techniques (with

the notable execption of Gill et al. 2014), and it would therefore be important to use this

integrated approach on many different fossil samples in the future.

References

[1] Wilson, D. E. and Reeder, D. M., eds. (2005) Mammal Species of the World, Johns

Hopkins University Press

[2] Rowe, T. (1988) Definition, diagnosis, and origin of Mammalia. J. Vertebr. Paleontol.

8, 241-264

[3] Luo, Z.-X. (2007) Tranformation and diversification in early mammal evolution.

Nature 450, 1011-1019

[4] dos Reis, M. et al (2012) Phylogenomic datasets provide both precision and

accuracy in estimating the timescale of placental mammal phyologeny. Proc. of R.

Soc. B 279, 3491-3500


15

[5] Meredith, R. W. et al. (2011) Impacts of the Cretaceous terrestrial revolution and

Kpg extinction on mammal diversification. Science 334, 521-524

[6] Benton, M. J. (2004) Vertebrate Palaeontology, Oxford: Blackwell Science

[7] Ahlberg, P. E. and Milner, A. R. (1994) The Origin and Early Diversification of

Tetrapods. Nature 368, 507-514

[8] Kermack, D. M. and Kermack, K. A. eds. (1984) The evolution of mammalian

characters, Croom Helm

[9] Campbell, J. W. and Saunders, W. B. eds. (1979) Comparative Animal Physiology

(3rd ed.), 279-316.

[10] Ruben, J. A. and Jones, T. D. (2000) Selective Factors Associated with the Origin of

Fur and Feathers. Am. Zool. 40, 585-596

[11] Rowe, T. B. et al. (2011) Fossil evidence on origin of the mammalian brain. Science,

332, 955-957

[12] Strang, G. and Fix, G. eds. (1973) An analysis of The Finite Element Method,

Prentice Hall.

[13] Zienkiewicz, O. C. et al. eds. (2005) The Finite Element Method: Its Basis and

Fundamentals (Sixth ed.), Butterworth-Heinemann

[14] Goodship, A. E. et al. (1979) Functional adaptation of bone to increaed stress. J.

Bone Joint Surg. 61A, 539-546

[15] Lanyon, L. E. et al. (1982) Mechanically adaptive bone remodelling. J. Biomech. 15,

141-154

[16] Curry, J. D. (2002) A very accessible study of bone biology and biomechanics. In

Bones: Structure and Mechanics (Curry, J. D. eds.), p.436, Princeton University

Press


16

[17] Donald, B. J. M. eds. (2007) Practical Stress Analysis with Finite Elements,

Glasnevin

[18] Bright, J. A. and Rayfield, E. A. (2011b) The response of the cranial biomechanical

finite element models to variations in mesh density. Anat. Rec. 294, 610-620

[19] Davis, J. L. et al. (2010) Predicting bite force in mammals: two-dimensional versus

three-dimensional lever models. J. Exp. Biol. 213, 1844-1851

[20] Rayfield, E. J. (2007) Finite Element Analysis and Understanding the Biomechanics

and Evolution of Living and Fossil Organisms. Earth Planet. Sci. 35, 541-576

[21] Rayfield, E. J. (2004) Cranial mechanics and feeding in Tyrannosaurus rex. Proc. R.

Soc. B 409, 1451-1459

[22] Rayfield, E. J. (2005) Aspects of comparitive cranial mechanics in the theropod

dinosaurs Coelophysis, Allosaurus and Tyrannosaurus. Zool. J. Linn. Soc.-Lond.

144, 309-316

[23] Rayfield, E. J. et al. (2001) Cranial desing and function in a large theropod dinosaur.

Nature 409, 1033-1037

[24] Gill, P. G. et al.(2014) Dietry specializations and diversity in feeding ecology of the

earliest stem mammals. Nature 512, 303-305

[25] McHenry, C. R. et al.(2007) Supermodeled sabercat, predatory behaviour in

Smilodon fatalis revealed by high-resuolution 3D computer simulation. Proc. Natl.

Ac. Sci. USA 104, 16010-16015

[26] Christiansen, P. A. (2011) A dynamic model for the evolution of sabercat predatory

bit mechanics. Zool. J. Linn. Soc.-Lond. 162, 220-242


17

[27] Tseng, Z. J. and Binder, W. J. (2010) Mandibular biomechanics of Crocuta crocuta,

Canis lupus, and the late Miocene Dinocrocuta gigantea (Carnivora, Mammalia).

Zool. J. Linn. Soc. 158, 683-696

[28] Dumont, E. R. (2005) Finite Element Analysis of biting behaviour and bone stress in

the facial skeletons of bats. Anat. Rec. Part A 283, 319-330

[29] Scott, R. S. et al. (2005) Dental microwear textural analysis shows within-species

diet variability in fossil hominins. Nature 436, 693-695

[30] Ungar, P. (2007) Dental microwear textural analysis of varswater bovids and Early

Pliocene palaeoenvironments of Langebaanweg, Western Cape Province, South

Africa. J. Mamm. Evol. 14, 163-181

[31] Ungar, P. S. (2008) Dental microwear and diet of the PlioPleistocene hominin

Paranthropus boisei. PLoS ONE 3, e2044

[32] Calandra, I. et al. (2012) Teasing apart the contributions of hard dietry items on 3D

dental microtextures in primates. J. Hum. Evol. 63, 85-98

[33] Godfey, G. M. et al. (2004) Dental use wear in extinct lemurs: evidence of diet and

niche differentiation. J. Hum. Evol. 47, 145-169

[34] Boyde, A. and Fortelius, M. (1991) New confocal LM method for studying local

relative microrelief with special reference to wear studies. Scanning 13, 439-430

[35] International Organization for Standardization, ISO 25178-2:2012 Geometrical

Product Specifications (GPS) (2012) - Surface Texture: Areal - Part 2, Definitions

and Surface Texture Parameters

[36] Meng, J. et al. (2006) The mammal fauna in the Early Cretaceous Jehol Biota:

implications for diversity and biology of Mesozoic mammals. Geol. J. 41, 439-463


18

[37] Kielan-Jaworowska, Z. et al. eds. (2004) Mammals from the age of dinosaurs. pp.

113, Columbia University Press

[38] Ji. Q. et al. (2006) A Swimming Mammaliaform from the Middle Jurassic and

Eomorphological Diversification of Early Mammals. Science 311, 1123-1127

[39] Damiani, R. et al. (2003) Earliest evidence of cynodont burrowing Proc. R. Soc.

Lond. B. 270, 1747-1751

[40] Kielan-Jaworowska, Z. and Gambaryan, P. P. (1994) Post-cranial anatomy and

habits of Asian multituberculate mammals. Foss. Strat. 36, 1-92

[41] Wilson, G. P. et al. (2012) Adaptive radiation of multituberculate mammals before

the extinction of the dinosaurs. Nature 483, 457-460

[42] Butler, P. (2000) Review of early allotherian mammals. Acta Palaeontol. Pol. 45,

317-342

[43] Bi. S et al. (2014) Three new Jurassic euharamiyidan species reinforce early

divergence of mammals. Nature 514, 579-584

[44] Averianov, A. O. and Lopatin, A. V. (2011) Phylogrny of Triconodonts and

Symmetrodonts and the Origin of Extant Mammals. Dokl. Biol. Sci. 436, 32-35

[45] Meng. J. et al. (2006) A Mesozoic gliding mammal from northeastern China. Nature

444, 889-893

[46] Krause, D. W. and Jenkins, F. A. (1983) The postcranial skeleton of North American

multituberculates. Bull. Mus. Comp. Zool. Harv. 150, 199-246

[47] Luo, Z.-X. and Wible, J. R. (2005) A Late Jurassic Digging Mammal and Early

Mammal Diversification. Science 308, 103-107

[48] Ruta, M. et al. (2013) The radiation of cynodonts and the ground plan of mammalian

morphological diversity. Proc. R. Soc. B 208, 20131865


19

[49] Lucas, S. G. and Luo, Z. (1993) Adelobasileus from the upper Triassic of west

Texas: the oldest mammal. J. Vertebr. Paleontol. 13, 309-334

[50] Luo, Z.-X. et al. (2001) A New Mammaliaform from the Early Jurassic and Evolution

of Mammalian Characteristics. Science 292, 1535-1540

[51] Hu, Y. et al.(2005) Large Mesozoic mammals fed on youg dinosaurs. Nature 433,

149-152

[52] Montellano, M. et al. Late Early Jurassic Mammaliaforms from Huizachal Canyon,

Tamaulipas, México. J. Vertebr. Paleontol. 28, 1130-1143

[53] Li, J. et al. (2001) A new family pf primitive mammal from the Mesozoic of western

Liaoning, China. Chinese Sci. Bull. 46, 782-785

[54] Stewart, J. R. (1997) Morphology and evolution of the egg of oviparous amniotes. in

Amniote Origins - Completing the Transition to Land (1) (Stuart, S. and Martin, K. L.

M. eds.), pp. 291-326, Academic Press

[55] Gauthier, J. et al. (1988) "The early evolution of the Amniota," in The Phylogeny and

classification of the tetrapods, Volume 1: amphibians, reptiles, birds (Benton, M.

eds.) pp. 103-155, Clarendon Press

[56] Benton, M. J. and Donoghue, P. C. J. (2006) Palaeontological evidence to date the

tree of life. Molecular Biol. Evol. 24, 26-53

[57] Laurin, M. and Reisz, R. R. (1995) A reevaluation of early amniote phylogeny. Zool.

J. Linn. Soc. 113, 165-223

[58] Reisz, R. R. (1997) The origin and early evolutionary history of amniotes. Trends in

Ecology and Evolution 12, 216-222

[59] Romer, A. S. and Parsons T. S. eds. (1985) The Vertebrate Body (6th ed.),

Philidelphia: Saunders


20

[60] Hildebran, M. and Goslow, G. eds. (2001) Analysis of Vertebrate Structure. John

Wiley & Sons Inc.

[61] Benton, M. J. eds. (2006) When Life Nearly Died. The Greatest Mass Extinction of

All Time,Thames & Hudson

[62] Charig, A. J. (1984) Competition between therapsids and archosaurs during the

Triassic period: a review and synthesis of current theories. in Symposia of the

Zoological Society of London, London

[63] Liu. J and Olsen, P. (2010) The Phylogenetic Relationships of Eucynodontia

(Amniota: Synapsida). J. Mamm. Evol. 17, 151-176

[64] Kemp, T. S. (2005) The Origin and Evolution of Mammals, pp. 3, Oxford University

Press

[65] de Quieroz, K. (1994) Replacement of an essentialistic perspective on taxonomic

definitions as exemplified by the definition "Mammalia". Syst. Biol. 43, 497-510

[66] Lucas, S. G. (1992) Extinction and the Definition of the Class Mammalia. Syst. Biol.

41, 370-371

[67] Liu, J. and Olsen, P. (2010) The Phylogenetic Relatioships of Eucynodontia

(Amniota: Synapsida). J. Mamm. Evol. 17, 151-176

[68] Lucas, S. G. and Hunt, A. P. (1990) The oldest mammal. New Mex. J. Sci. 30, 41-49

[69] Luo, Z.-X. et al. (2002) In quest for a phylogeny of Mesozoic mammals. Acta

Palaeontol. Pol. 47, 2002.

[70] Kermack, K. A. et al. (1981) The skull of Morganucodon. Zool. J. Linn. Soc.71, 1-158

[71] Kermack, K. A. et al. (1973) The lower jaw of Morganucodon. Zool. J. Linn. Soci. 53,

87-175


21

[72] Prasad, G. V. and Manhas, B. K. (2001) First docodont mammals of Laurasian

affinity from India. Curr. Sci. India 81, 1235-1238

[73] Jacobs, L. L et al. (1989) Modern mammals origins: evolutionary grades in the Early

Cretaceous of North America. PNAS 86, 4992-4995

[74] Luo, Z.-X. et al. (2000) Dual origin of tribosphenic mammals. Nature 409, 53-57

[75] Rauhut, O. W. M. et al. (2002) A Jurassic mammal from South America. Nature 416,

165-168

[76] Krause, D. W. et al. (1997) Cosmopolitanism among Godwanan Late Cretaceous

mammals. Nature 390, 504-507

[77] Yuan, C.-X. et al. (2013) Earliest Evolution of Multituberculate Mammals Revealed

by a New Jurassic Fossil. Science 341, 779-783

[78] Chun-Ling, G. et al. (2010) A new mammal skull from the Lower Cretaceous of

China with implications for the evolution of obtuse-angled molars and 'amphilestid'

eutriconodonts. Proc. R. Soc. B. 277, 237-246

[79] Meng, J. et al. (2011) Transitional mammalian middle ear from a new Cretaceous

Jehol eutriconodont. Nature 472, 181-185

[80] O'Leary, M. A. et al. (2013) The Placental Mammal Ancestor and the Post-K-Pg

Radition of the Placentals. Science 339, 662-667


22

Box 1. Origin of Mammals

The first fully terrestrial vertebrates were the amniotes. Unlike earlier tetrapods, they had eggs with

internal membranes, allowing the embryo to breath while still retaining water [54] [55], and the eggs laid

on dry. They arose in the Late Carboniferous and split into two lineages, Sauropsida (reptiles and birds),

and Synapsida [7] [56] [57] [58]. Mammals fall within the Clade Synapsida [6], which also includes many

other extinct taxa traditionally called ‘mammal-like reptiles’. Synapsids are characterised by having a

single temporal fenestra behind each eye orbit, and differentiated teeth [6] [59] [60]. They came to

dominate the terrestrial fauna in the Permian, with the ‘pelycosaurs’ dominating the Early Permian, and

the more advanced therapsids dominating the Late Permian [6].

The therapsids differed from earlier pelycosaurs in several features of the skull, including larger temporal

fenestrae and incisors of equal size. The therapsid lineage leading to mammals gradually changed from

pelycosaurs-like animals, to mammal-like animals. Acquisition of mammalian traits included the gradual

development of a bony secondary palate, which may have been involved in the development of a faster

metabolism; the dentary becoming the main bone of the lower jaw, with the reduction of other lower jaw

bones (which would eventually form the bones of the middle-ear); and the evolution of erect limb posture

(although this process was erratic; indeed, modern monotremes still have a semi-sprawling gait) [8].

The Permian-Triassic extinction wiped out the majority of synapsids, with only three therapsid clades

surviving, the dicynodonts, therocephalians, and cynodonts [61]. However, the remaining synapsids were

soon overtaken as the dominant land vertebrates by the archosauran sauropsids during the ‘Triassic

takeover’ [62]. The archosaurs may have been able to diversify and radiate more quickly than synapsids

due to glandless skin and ability to eliminate nitrogenous waste as a solid uric acid paste with very little

water. This would have been an advantage over the synapsids, which excreted urine with lots of water, in

the increasingly dry climate [9].

The therocephalians only lasted into the early Triassic, and the dicynodont went extinct by the end of the Triassic, leaving the cynodonts as the only living synapsid lineage. The clade Eucynodontia, was divided into two clades, Cynognathia (such as Cynognathus) and the Probainognathia [63]. During the late Triassic, most of the large cynodonts disappeared, and the remaining became smaller and increasingly mammal-like. The probainognathans gave rise to the mammaliaforms at the Late Triassic [48] [64].


23

Box 2. What is a mammal?

The exact definition of Mammalia is not entirely uniform [65]. Some authors restrict the term Mammalia to

grown group mammals only (the group comprising of the common ancestor of all modern mammals and

of all its decedents – sometimes referred to as ‘true mammals’) [2] [66]. Basal families, such as the

morganucodonts, docodonts and kuehneotherians, are not included in this definition. To accommodate

those taxa falling outside this group, but which are more closely related to crown group mammals than to

any other taxa, the group Mammaliaformes was defined as comprising “the last common ancestor of

Morganucodontidae and Mammalia and all its descendants”. Some tradition, trait-based mammalian

taxa, such as Abelobaselius and Sinoconodon still fall outside of this definition. They are therefore

included with the grouped Mammaliamorpha, defined as the clade originating with the last common

ancestor of Tritylodontidae and the crow group mammals (Figure i) [2] [66].

Figure i. Mammaliamorph cladogram

However, many authors continue to use the traditional, morphological definition, which includes all of the

non-mammalian mammaliaforms as mammals [6] [64]. Other authors do not define mammals by the

crown group. For example, Kielan-Jawrowska et al. (2004) define Mammalia as the group originating

with the last common ancestor of Sinoconodon and living mammals, which would therefore include most

of the mammaliaforms within Mammalia [37].

The basal mammaliamorphs include the tritylodontids, Adelobasileus and Sinoconodon. The tritylodonts

were small, highly mammal-like, and are the most basal mammaliamorphs, although they retained the

reptilian quadrate-articular joint. They arose at the end of the Late Triassic and disappeared in the Late

Cretaceous [48] [67]. Adelobasileus is from the Late Triassic of Texas, and only known from a partial

skull. Its distinct cranial feature indicate it being close to mammals, possibly close to the common

ancestor of mammals, with some authors originally describing it as the oldest known mammal [49] [68].

Sinocnbodon was from the Early Jurassic, and had a dentary-squamosal jaw joint. This would put it

closer to mammals than the morganucodnts, however it was polyphylodont, like reptiles, and may thus

be more basal. It is possible that the jaw joint evolved independently [37] [69].


24

Box 3. Mesozoic Mammaliaforms

The morgnucodonts, named after the type species Morganucodon watsoni, were some of the first

mammaliaforms, appearing at the end of the Triassic, and are considered the most basal of the

mammaliaforms. They had an unusual ‘double-joint’ jaw structure, with the jaw articulation being made

up of a dentary-squamosal joint as well as a quadrate-articular joint [70]. The articular and the quadrate

would become the melleus and the incus in modern mammals. The double joint clearly shows the

transition from a ‘reptile-like’ jaw to a ‘mammalian-like’ jaw. Morganucodonts were also diphyodont

(having only two sets of teeth, like modern mammals), and the postcanine teeth were replaced with

molars and premolars, as in modern mammals [71].

The docodonts arose during the Middle Jurassic, and were among the most common mammaliaforms

until the Early Cretaceous, and exhibited a fair degree of diversity. They are distinguished by their

relatively sophisticated set of molars [72]. They are generally insectivorous or herbivorous.

Hadrocodium from the Early Jurassic may be very close the origin of mammals. Its jaw consists of only

the squamosal and dentary bones, with the articular and quadrate bones forming a nearly complete

middle-ear structure. It also had a relatively large brain case [50].

Comprising of the traditional clade ‘Symmetrodonta’ along with Kuehneotherium and Woutersia, the

kuehneotherians form the Late Triassic to Early Jurassic are known only from teeth and a few jaw

fragments. They are also potentially very close to the origin of mammals, with evidence from teeth

putting them closer to mammals than Hadrocodium [24] [69], although Kuehneotherium seems to retain

the plesiomorphic articular-quadrate jaw join, which would suggest it be more basal [37].

Kuehneotherium also appeared to specialise in eating soft bodied insects, such as moths [24].


25

Box 4. Crown-group Mammals

Auatralosphenida contains the monotremes (such as modern echidnas and platypus) and the enigmatic

clade Ausktribosphenidae [6]. The ausktribosphenids are strange in having tribosphenic molars,

otherwise only found in therians [73]. However, they come from the Early to Middle Cretaceous of

Australia, whereas the therians were confined in the northern hemisphere until much later. Therefore

their position within mammals is uncertain [69] [74]. The australosphenids represent the most basal

crown-group mammals, originating in the Early to Middle Jurassic, with Asfaltomylos being a potential

basal member, and were once widespread in the southern hemisphere [75]. Teinolophos and

Steropodon of the Early Cretaceous are the earliest monotremes.

Figure ii. Grown-group mammal cladogram

The eutriconodonts and the multitubercilates were the most common and diverse mammals of the

Mesozoic, with cosmopolitan distributions [76]. The eutriconodonts were named for their molars with

three main cusps in a row, and were a diverse group which occupied a variety of niches. The

multituberculates are so named for the multiple tubercles on their molars, and share many convergent

characteristics with rodents, often being called the ‘rodents of the Mesozoic’ [6] [37] [41]. They were very

diverse and existed for around 120 million years, from the late Jurassic [77] to the early Oligocene, when

they were outcompeted by rodents. There is also some evidence, based on the shape of the pelvic

bones, that the multituberculates gave birth to tiny, undeveloped young, similar to modern marsupials

[37] [41].

Most analyses place the multituberculates as being closer to therians than eutriconodonts, though some

authors have found that the multituberculates fall outside of crown-group mammals, perhaps even further

than the morganucodonts [42]. Others have put the eutriconodonts as closer to therians, within the clade

Holotheria [44] [78] [79]. The multituberculates, along with the gondwanatherians and haramiyids, form

the large and diverse group Allotheria, forming a sister group to Holotheria (which includes therians and

relatives) (Figure ii) [42] [69].

Theria contains eutherians and metatherians. They are characterised as having no clavicle or coracoid

bones, and having tribospenic molars and a crurotarsal-like ankle joint. The metatherians include

marsupials, as well as related extinct groups. The oldest know metatherian is Sinodelphis, from the Early

Cretaceous of China. The eutherians include placental mammals, as well as various extinct groups. The

oldest known eutherians are Juramaia from the Late Jurassic and Eomaia from the Early Cretaceous [6]

[37]. The oldest confirmed placental is Protungulatum, dating to the K-Pg boundary [80].


26

Figure 1. Steps in finite element analysis. From Rayfield (2007).


27

Figure 2. Diversity of Mezoic mammals. a) Phylogenies showing successive radiations of mammaliforms

through the Mesozoic. b) Ecomorpholotypes of Mesozoic mammals in comparison to modern analogues,

showing small, insectivorous stereotype and many newly discovered ecomorphologypes. Also shows

iterative evolution. From Luo (2007).

A

B


28

Glossary

Adaptive radiation: process whereby organisms diversify rapidly, filling many new

ecological niches. Cf. radiation.

Amniote: tetrapods with eggs that have an amnion, allowing eggs to be laid on land.

Includes modern day reptiles, birds and mammals.

Apomorphy: or derived state, an innovation which can be used to diagnose a clade.

Apomorphies derived in individual taxa are autapomorphies, and do not express anything

about relationships between groups. Synapomorphies are apomorphies which are shared

by two or more taxa and inferred to have been present in the most recent common

ancestor, but whose own ancestor lacked them. Synapomorphies can show relationships

between taxa.

Arborealism: living in trees

Archosaurs: group of diapsid reptiles, including dinosaurs (and modern day birds) and

crocodilians among others.

Canines: long, pointed teeth used to hold food to tear it apart.

Carnassials: large teeth in many carnivorous animals used for shearing flesh. Modified

fourth upper premolar and first lower molar.

Carnivory: deriving nutrition primarily from animal tissue.


29

Cladistics: biological classification where organisms are grouped together based on

shared characteristics that come from the groups last common ancestors. Cf.

phylogenetics.

Cladogram: tree diagram used in cladistics to show relations among organisms.

Convergent evolution: independent evolution of similar characters in different lineages.

Known as homoplasy in cladistics.

Crown group: in cladistics and phylogenetics, a group which contains modern

representatives, there most recent common ancestor, and all of its descendants.

Cynodonts: a group of mammals-like therapsids, first appearing in the Late Permian.

Includes mammals.

Dentary-squamosal joint: joint between the dentary bone in the lower jaw, and the

squamosal bone in the back of the skull. Jaw joint typical of mammals.

Diapsids: amniotes with two temporal fenestrae in each side of the skull. Includes reptiles

and birds.

Discretisation: process of breaking down a continuous structure into discrete parts.


30

Ecomorphology: the relationship between the ecological role of an organism and its

morphology or morphological adaptations.

Exaptation: shift in function of a trait during evolution.

Fossorial: organisms adapted to digging and living underground.

Generalised: uspecialised, exploits a variety of different ecological niches.

Herbivory: deriving nutrition primarily from plant matter.

Heterodont: having differentiated teeth/ different kinds of teeth.

Homoplasy: see convergent evolution.

Incus: bone in the mammalian middle ear, developed from the quadrate bone. Cf.

quadrate- articular joint.

Insectivory: deriving nutrition primarily from insects.

Malleus: bone in the mammalian middle ear, developed from the articular bone. Cf.

quadrate- articular joint.

Mammaliaforms: group of cynodonts containing mammals. See Box 2.


31

Mammaliamorphs: group of cynodonts containing mammaliforms. See Box 2.

Mesozoic: geological era lasting from 252 to 66 mya. Often called the age of reptiles or

the age of the dinosaurs. Dinosaurs (including birds), mammals, crocodilians, pterosaurs

and marine reptiles evolved during this time. Many of these groups went extinct at the end

of the Mesozoic, with mammals, birds and crocodilians surviving.

Monophyly: in pylogenetics and cladistics, a clade which includes and ancestral species

and all of its descendants. Mammals (Class Mammalia) are monophyletic.

Morphology: physical form and structure of an organism.

Myrmecophagy: feeding primarily on ants or termites.

Opsin: group of light sensitive proteins in the photoreceptor cells of the retina.

Paraphyly: a clade or group consisting of a common ancestor and some of its

descendants, but not others. The traditional grouping of reptiles (Class Reptilia) is

paraphyletic as it doesn’t include birds or mammals, but does include their ancestors, and

the most recent common ancestor between them and modern reptiles. Paraphyletic

groups are discouraged in cladistics and phylogenetics, although they are sometimes

useful, especially in relation to stem lineages and evolutionary grades.

Pelycosaurs: paraphyletic group of ‘primitive’ synapsids, contrasted with the more

‘advanced’ therapsids.


32

Plesiomorphy: or ancestral state, character state that a taxon has retained from its

ancestors. When two or more taxa that are not nested within each other share this a

plesiomorphy, it is called a symplesiomorphy. Cf. apomorphy.

Piscivorous: primarily feeding on fish.

Quadrate-articular joint: joint between the quadrate bone in the lower jaw, and the

articular bone in the back of the skull. Jaw joint typical of reptiles. Quadrate and articular

bones become the incus and malleus bones of the mammalian middle ear respectively.

Radiation: see adaptive radiation

Sauropsids: group of amniotes containing diapsids and relatives, and possibly anapsids

such as turtles.

Scansorial: lifestyle where the animal is adapted or specialised for climbing.

Secondary palate: anatomical structure which divides the nasal and oral cavities.

Specialised: adaptived to exploit specific niche.

Synapsids: amniotes with one temporal fenestra in each side of the skull. Includes

mammals and their extinct relatives (often referred to as ‘mammal-like reptiles’)


33

Taxon: groups of organisms which form a unit, e.g. species, family, class etc.

Temporal fenestrae: openings in the temporal bone in vertebrate skulls.

Therapsids: group of synapsids. More ‘advanced’ (or derived) in contrast to ‘pelycosaurs’.

Volant: capable of flying of gliding.

ecomorphological diversity of mesozoic mammals

Documents