Benton M.J. Introduction to paleobiology and the fossil record

Подождите немного. Документ загружается.

128 INTRODUCTION TO PALEOBIOLOGY AND THE FOSSIL RECORD

where their ecological tolerances determine

their evolutionary rates, and produce a super-

ﬁ cial appearance of species selection.

Is evolution hierarchical? And, if so, was

Darwin wrong? The case has been overstated

by critics: evolution occurs by natural selec-

tion, as Darwin said in 1859. Many proposed

examples of species selection can be explained

by natural selection, coupled with rapid asym-

metric geographic speciation and the effect

hypothesis. Nonetheless, species selection is a

possibility, and convincing examples may be

found in the future.

THE TREE OF LIFE

Tree thinking

As noted at the start of this chapter, Darwin

was the ﬁ rst to picture evolution as a great

branching tree, and to point out that all

species had evolved from a single ancestor at

the base of the tree. The idea of the tree of

life has come to the fore recently, with a

massive effort by biologists and paleontolo-

gists to discover the whole tree. From small-

scale questions like “Which species of ape is

closest to humans – the gorilla or chimp?”,

large teams of researchers are hastening to put

together complete trees of all species of

mammals, angiosperms, amphibians, spiders

and many other groups. As each complete

tree, that is, a tree containing all species, is

published, systematists are getting closer to

Darwin’s ideal of understanding the shape of

the whole tree of life.

It is important to distinguish trees from

ladders. Many people think that all plants and

animals are arranged in a series from simple

to complex, or “lower” to higher”. The

pattern of evolution is then like a ladder, a

single long line of progression from one species

to the next, an idea that was popular 200

years ago and termed the Scala naturae (see

p. 13). But all the evidence shows that evolu-

tion is a process of splitting and so the tree is

the correct analogy, not the ladder.

Fossils offer fundamental information on

the history of life and on large-scale patterns

of evolution. There has been a revolution in

the ways in which paleontologists interpret

evolutionary aspects of the fossil record, and

Impalas

Gnus and hartebeests

Time (Ma)

Figure 5.7 Reconstructed phylogeny of African antelopes. Two lineages diverged 6–7 Ma, the slowly

evolving impalas and the rapidly speciating gnus and hartebeests. The second group could be said to be

evolutionarily more successful than the ﬁ rst, and this might be interpreted as a result of species

selection of species-level characters – the rate of speciation. However, the gnus and hartebeests have

more specialized ecological preferences than do the species of impalas: perhaps selection has occurred

at the individual level (natural selection), and this has had an effect at the species level. Species numbers

14 and 26 are omitted in this study. (Based on Vrba 1984.)

MACROEVOLUTION AND THE TREE OF LIFE 129

this is true of biologists as well. Nearly all

studies of ecology, behavior and evolution are

tied to a phylogenetic tree of the organisms

involved. Since 1990 phylogenetic trees have

been springing up everywhere, both because

of new techniques for discovering trees and a

realization that nothing in biology means any-

thing without a tree.

Cladistics: reconstructing life’s hierarchy

For centuries, biologists have struggled with

the search for the true pattern of relationships

among organisms: how is the tree of life to be

discovered? Debates about whether birds

originated from dinosaurs, whether annelids

and arthropods are close relatives or not, or

whether the gorilla or chimp is the closest

relative of humans, all hinge on the need to

identify patterns of relationships correctly.

If you had the task of sorting out the rela-

tionships among 100 species of parrots, where

would you begin? You might note the color

of their feathers, and classify them into a blue

group, a red group and a green group. But

then you might notice that body size or beak

shape gives a different classiﬁ cation. If you

then looked at the internal anatomy of the

100 parrots, you might ﬁ nd an entirely differ-

ent classiﬁ cation based on the shape of the

skull, the bones of the wing, or the arrange-

ment of muscles or arteries. Up to 1960, sys-

tematists had a hard task in seeking to decide

which characters were “good” and which

were “bad”. Good characters are phyloge-

netically informative, that is, indicative of the

true phylogeny, but what about the bad, or

uninformative, characters?

Phylogenetically uninformative characters

fall into two main categories: convergences

and plesiomorphies. Convergence in evolu-

tion is when features, or organisms, evolve to

look the same perhaps because they live the

same way. The marsupial mole of Australia

looks just like the northern hemisphere mole,

with great paddle-like limbs, poor eyesight

and an excellent sense of smell, because they

both burrow and eat worms, and yet they are

not closely related. Two species of parrots

might have convergently evolved a red patch

of feathers on their wings as a signal. Plesio-

morphies are characters that are shared by the

organisms of interest, say parrots, but also by

other groups. So, all parrots have beaks, but

this is not a helpful character in sorting out

the phylogeny of parrots because all other

birds have beaks too. True parrots have blue

and green feathers that have a special irides-

cent quality not seen in the feathers of cocka-

toos. But such special light-reﬂ ecting feathers

are seen also in many other bird groups, and

so are plesiomorphic for parrots.

Phylogenetically informative characters

identify clades, or monophyletic groups. These

are groups that had a single origin and include

all the descendants of that common ancestor.

A good example of a clade is the Psittaci-

formes, the parrots, a group that has long

been identiﬁ ed as real and distinct from all

others by naturalists. Clades are distinguished

from two kinds of non-natural groups: (i)

paraphyletic groups, which had a single

common ancestor, but do not include all

descendants, such as the Reptilia, which

excludes birds and mammals; and (ii) poly-

phyletic groups, which are random assem-

blages of organisms that arose from more

than one ancestor, and so have no place in the

search for the tree of life.

Willi Hennig (1913–1976), an eminent

German entomologist, realized the difference

between phylogenetically informative and

uninformative characters, and between mono-

phyletic and paraphyletic/polyphyletic groups.

He stressed the need to develop a new, more

objective method in systematics, which has

come to be called cladistics. The fundamental

aim of cladistics is to identify clades, and

so to discover, or reconstruct, the tree of life.

Patterns of relationships are shown as branch-

ing diagrams, or cladograms (e.g. Fig. 5.8), in

shark

salmon

frog

shark

salmon

frog

sea squirt

amphioxus

VERTEBRATA

?bone

?skull

?tail

VERTEBRATA

bone

skull

(a) (b)

CHORDATA

tail

Figure 5.8 Reconstructing the phylogeny of

vertebrates by cladistic methods. (a) Are the

deﬁ ning features of vertebrates the possession of

bone, a skull and a tail? (b) The tail is found in

a wider group, termed the Chordata, but the

skull and bone deﬁ ne the Vertebrata.

130 INTRODUCTION TO PALEOBIOLOGY AND THE FOSSIL RECORD

which the most closely related species are

joined most closely. The branches in the clado-

gram join at branching points, or nodes, each

of which marks the base of a clade.

Hennig invented a rather complex termi-

nology for cladistics, but some terms are com-

monly used, and should be mentioned. He

called phylogenetically informative characters

apomorphies, or derived characters (and dis-

tinguished them from plesiomorphies, the

characters present in wider groups). Apomor-

phies shared by two or more species are

termed synapomorphies. Apomorphies are

features that arose once only in evolution, and

therefore diagnose all the descendants of the

ﬁ rst organism to possess that new character.

Synapomorphies of parrots, the bird Order

Psittaciformes include the deep, hooked beak

and the unusual foot in which two toes point

forward and two back.

The concept of an apomorphy actually cor-

responds to an older distinction between

homology and analogy in evolution. The fore-

limb of vertebrates is a good example of a

homology: even though the arm of a human

is very different from the wing of a bird or

the paddle of a whale, the detailed anatomy

of each is the same, and they clearly arose

from the same ancestral structure. On the

other hand, the swimming limbs of vertebrates

differ from group to group: in detail it can be

shown that the paddles of ichthyosaurs,

whales, plesiosaurs and seals (Fig. 5.9) are not

homologs; they are merely analogs.

Cladistics might seem relatively straightfor-

ward, but it is not in practice (Box 5.5). There

(a)

(e)

(f)

(g)

(b)

(c)

(d)

Figure 5.9 Swimming forepaddles of a variety of

reptiles (a–d) and mammals (e–g): (a) Archelon,

a Cretaceous marine turtle; (b) Mixosaurus, a

Triassic ichthyosaur; (c) Hydrothecrosaurus, a

Cretaceous plesiosaur; (d) Plotosaurus, a

Cretaceous mosasaur; (e) Dusisiren, a Miocene

sea-cow; (f) Allodesmus, a Miocene seal; and (g)

Globicephalus, a modern dolphin. The forelimbs

are all homologous with each other, and with

the wing of a bird and the arm of a human.

However, as paddles, these are all analogs: each

paddle shown here represents a separate

evolution of the forelimb into a swimming

structure.

Box 5.5 Cladistic analysis

There are three steps in drawing a cladogram: character analysis, outgroup comparison and tree

calculation. The character analysis is the process of listing characters that vary among the organisms

of interest, and identifying those that are possible apomorphies. Characters are often coded as pres-

ence/absence or 0/1. So, for example, in a character analysis of the apes, a possible apomorphy might

be “possession of an opposable thumb, used for gripping”. This would be coded as present (or 1)

in humans, and as absent (or 0) in chimpanzees and gorillas.

The outgroup comparison is the phase when characters and their codings are tested for validity.

If the group under study, the ingroup, consists of the apes, then the outgroup might be monkeys.

The outgroup could really be oak trees and worms, but they are so distantly related to the apes that

comparisons would be largely meaningless (oak trees and worms do not even have thumbs). Com-

MACROEVOLUTION AND THE TREE OF LIFE 131

parison with the outgroup of monkeys shows that monkeys, chimpanzees and gorillas have non-

opposable thumbs. So this is coded “0”, and the opposable thumb of humans appears to be unique,

an apomorphy, and so is coded “1”.

The ﬁ nal step is the tree calculation. All the characters and their codings are listed in a data

matrix, a tabulation of the data, listing all the species and all the characters and their codings. The

tree calculation is usually run by computer, and a search is carried out, using different methods, to

ﬁ nd the single tree, or group of trees, that best explain the data.

We give a simple worked example here, to determine the relationships of six vertebrates:

shark, salmon, frog, lizard, chicken and mouse. These are distant relatives, of course, representing

cartilaginous ﬁ shes, bony ﬁ shes, amphibians, reptiles, birds and mammals, respectively, but if we

can sort out their relationships, we have a broad outline of the phylogenetic tree of all

vertebrates.

Ten (out of many) characters are listed in the data matrix below, and their codings are shown

(0 = absent; 1 = present) for each of the six animals. Next, outgroup comparison will help to

sort out the phylogenetically informative characters from the plesiomorphies. We will choose two

examples: the shark and the salmon have ﬁ ns (character 1) while the others have legs (character 2),

and the chicken and mouse are warm-blooded and the others are cold-blooded (character 3). Out-

group comparison of these two sets of characters suggests that warm-bloodedness is probably an

apomorphy (because most members of the outgroup, such as clams, oak trees and bugs, are cold-

blooded), but it is harder to tell whether ﬁ ns or legs are apomorphies or not, so these two are

retained.

Scanning over the data in this table, it is clear that some groupings are indicated by several synapo-

morphies, but there are contradictions. For example, the diapsid skull (see p. 447) supports a pairing

of lizard and chicken, but warm-bloodedness suggests a pairing of chicken and mammal. Both pair-

ings are not possible, and one of these synapomorphies must be wrongly interpreted. The method

of testing at this point is to seek the most parsimonious pattern of relationships, that is, the one that

explains most of the data and implies least mismatch, or incongruence. The data may be run through

a computer program, such as PAUP (Swofford 2007) that ﬁ nds the most parsimonious cladogram

(Fig. 5.10a), and highlights the incongruent (i.e. probably misinterpreted) characters. The cladogram

is of course a best effort, and further study of the specimens, and the discovery of new characters,

can conﬁ rm or refute it.

The cladogram can be made into a phylogeny by the addition of a time scale (Fig. 5.10b). Here,

the fossil evidence for dates of origin of the various groups is used to give a picture of the true shape

of this part of the phylogeny of life.

Read more about cladistics in Forey et al. (1998), and about cladistics as applied to fossil organ-

isms in Smith (1994). Cladistic software includes PAUP (Swofford 2007), the most-used program,

and basic cladistic routines are available in PAST (Hammer & Harper 2005). Read more at http://

www.blackwellpublishing.com/paleobiology/.

Character Shark Salmon Frog Lizard Chicken Mouse

1 Fins 1 1 0 0 0 0

2 Legs 0 0 1 1 1 1

3 Warm-bloodedness 0 0 0 0 1 1

4 Bone 0 1 1 1 1 1

5 Diapsid skull 0 0 0 1 1 0

6 Loss of larval stage 0 0 0 1 1 1

7 Lung or swim bladder 0 1 1 1 1 1

8 Amniote egg 0 0 0 1 1 1

9 Elongated neck vertebrae 0 0 0 1 1 0

10 Marginal teeth 0 1 1 1 1 1

Continued

132 INTRODUCTION TO PALEOBIOLOGY AND THE FOSSIL RECORD

are many controversies as systematists try to

ﬁ nd agreement on disputed parts of the tree

of life. In some cases, the shape of the tree is

clear because each node is diagnosed by many

apomorphies, but in others the clades and

nodes are hard to pinpoint. Perhaps in those

cases, evolution happened so fast that apo-

morphies were not established, or perhaps

they have been overwritten in time.

The molecular revolution

The second approach in reconstructing the

tree of life is based on comparison of mole-

cules. With the birth of molecular biology in

the 1950s and 1960s, it became clear that

homologous proteins share similar structures

in different organisms. For example, many

animals share the molecule hemoglobin, a

shark

salmon

frog

lizard

chicken

mouse

shark

salmon

frog

lizard

chicken

mouse

(a)

(b)

1. fins

3. warm-bloodedness

6. loss of larval stage

2. legs

8. amniote egg

4. bone

7. lungs or swim bladder

10. marginal teeth

Time (Ma)

Cenozoic

Cretaceous

Jurassic

Triassic

Permian

Carboniferous

Devonian

Silurian

Diapsida

Amniota

Tetrapoda

Osteichthyes

Gnathostomata

100

200

300

400

5. diapsid skull

9. elongated neck vertebrae

Figure 5.10 The relationships of the major groups of vertebrates, tested using six familiar

animals. (a) Postulated relationships, based on the analysis of characters discussed in the text.

(b) Phylogenetic tree, showing the cladogram from (a) set against a time scale, and basing the

dating of branching points on the oldest known fossil representatives of each group.

MACROEVOLUTION AND THE TREE OF LIFE 133

protein that carries oxygen in the blood, and

that makes the blood red. Structurally, the

hemoglobin of all organisms that possess it is

very similar because it has to perform its

oxygen-carrying function – but there are

subtle differences. So, the hemoglobin of

humans and chimps is identical, but their

hemoglobin differs a little from that of a horse

or cow, and a great deal from the hemoglobin

of a shark or a salmon.

Comparisons of molecules allow analysts

to do two things: to draw up trees of relation-

ships and to estimate time. Trees of relation-

ships can be based on a simple comparison of

the amount of difference between protein

sequences, and a best-ﬁ tting dendrogram, or

branching diagram, is drawn. Identifying spe-

ciﬁ c amino acid changes, and treating them as

synapomorphies, allows the dendrogram to

be treated as a molecular cladogram.

Time estimation comes from the concept of

the molecular clock. The amount of difference

in the ﬁ ne structure of a protein between any

pair of species is proportional to the time

since they last shared a common ancestor.

Differences have been documented in the

primary structure of proteins, the sequence of

amino acids from end to end of the unfurled

protein backbone. There are some 20 amino

acids, and their sequence determines the shape

and function of a protein. Small changes in

the amino acid sequence of hemoglobin

occurred every few million years, somewhat

at random, and the rate of change allows a

time scale to be calibrated against the molecu-

lar tree.

Since 1990, attention has shifted almost

entirely from sequencing proteins to sequenc-

ing the nucleic acids such as DNA and RNA.

These are the molecules in the nucleus that

comprise the genetic code, and they may be

sequenced in a semiautomated manner using

a process called the polymerase chain reaction

(PCR). PCR is a means of cloning, or dupli-

cating, small samples of nucleic acid, and then

of determining the exact sequence of base

pairs, the four components of the nucleic acid

strand, adenine, cytosine, guanine and thymine

(or uracil), abbreviated as A, C, G and T (or

U). DNA and RNA may be sequenced from

the nucleus or the mitochondria of cells (see

p. 186), and molecular biologists generate

huge sequences of such information each year.

Indeed, the human genome project was one

of many examples of international programs

to determine the entire DNA sequence of all

the chromosomes of a single species. The PCR

method has also opened up the possibility of

sequencing the genetic material of extinct

organisms (Box 5.6).

Box 5.6 Fossil proteins: the real Jurassic Park?

Proteins were extracted from fossils in the 1960s and 1970s, but most of these were decay materials,

the proteins of bacteria that decomposed the original tissues. Even in cases of exceptional preserva-

tion where soft tissues are preserved (see p. 60), the proteins have usually long vanished. Until 1985,

the oldest DNA, recovered in tiny quantities, came from Egyptian mummies, 2400 years old.

Then came Jurassic Park! In the book by Michael Crichton (1990), and in the ﬁ lm by Steven

Spielberg (1993), a scenario was developed where molecular biologists extracted dinosaur DNA from

blood retained in the stomach of a mosquito preserved in amber. The fragments of dinosaur DNA

were cloned and inserted into the living cells of a modern frog (an odd choice when the nearest

living relatives of dinosaurs are birds), and the whole dinosaur genetic code was somehow recon-

structed and living dinosaurs recreated. Amazingly, science then followed the ﬁ ction for a time.

Michael Crichton was wise to choose amber as the means of preservation (see p. 63). Insects in

amber are trapped instantly, usually overwhelmed by the sticky resin, and no decay takes place; the

amber excludes oxygen and water so that no physical or chemical changes should occur during

subsequent millennia. A series of scientiﬁ c reports were published in high-proﬁ le journals through

the 1990s, announcing original DNA from a termite in Oligocene-Miocene amber, a weevil in Early

Cretaceous amber, Miocene leaves, and even supposed dinosaur DNA in 1994. These reports col-

Continued

134 INTRODUCTION TO PALEOBIOLOGY AND THE FOSSIL RECORD

lapsed like a pack of cards soon after. The “dinosaur” DNA turned out to be human: the PCR

technique is so sensitive that the tiniest fragment of DNA, in this case from sweat or sneezed mucus

of a lab assistant, can be ampliﬁ ed. Careful study showed that DNA is highly labile and breaks down

in even hundreds of years, and is pretty well all gone by 40,000 years, even in the most exceptional

preservation.

The most convincing studies of the DNA of fossil species come from ice age mammals such as

the cave bear, giant Irish deer, Neandertal man, woolly rhino and woolly mammoth. In a rush of

enthusiasm, three labs independently sequenced and published the complete mitochondrial genome

of the woolly mammoth in early 2006 (Krause et al. 2006; Poinar et al. 2006; Rogaev et al. 2006).

These studies gave conﬂ icting results: it is still not clear whether the closest living relative of

the mammoth is the Asian elephant Elephas maximus or the African elephant, Loxodonta africana

(Fig. 5.11). All studies though conﬁ rm that modern elephants and the mammoth are about as

closely related to each other as humans are to chimps, and that the species split apart 5–6 Ma.

Read more at http://www.blackwellpublishing.com/paleobiology/.

Figure 5.11 Relationships of the woolly mammoth based on mitochondrial DNA (mtDNA). This

analysis (Rogaev et al. 2006) places the mammoth Mammuthus primigenius closest to the Asiatic

elephant Elephas maximus, while other analyses of mammoth mtDNA place the mammoth closer to

the African elephant Loxodonta africana. Either way, the relationship to the modern elephants is

close, suggesting all three species diverged in the last 5–6 myr. Two samples of mtDNA for the two

modern elephants are included, and the outgroups are the sea cow Dugong dugon and the hyrax

Procavia capensis. The sets of digits at each branching point are various measures of robustness:

values range from 0 to 1 and 0 to 100, with 1.0 and 100% indicating maximum robustness of the

node. Scale bar is 0.1 base-pair substitutions per site. (Courtesy of Evgeny Rogaev.)

Mammuthus primigenus

Elephas maximus A

Loxodonta africana A

Loxodonta africana B

Elephas maximus B

Dugong dugon

Procavia capensis

1.0, 1.0,

100, 100

1.0, 0.88,

95, 82

1.0, 1.0,

100, 100

1.0, 1.0,

100, 100

0.1

MACROEVOLUTION AND THE TREE OF LIFE 135

The vast numbers of DNA and RNA

sequences that are generated daily from the

whole diversity of life are stored as letter

codes in open-access databases. Any investi-

gator may download the sequences of any

particular gene or chromosome for as many

species as are available. There are then two

key processes in extracting a phylogenetic

tree from such data. First, the nucleic acid

sequences must be aligned, that is matched,

so that the code of a particular gene in one

species is lined up with the same sequence in

another species. Alignment can be difﬁ cult

because species do not differ only in the place-

ment of particular base pairs, but sometimes

gaps in the sequence are introduced, or whole

sections may be repeated. Once the sequences

of a number of species are satisfactorily

aligned, the phylogenetic analysis is carried

out. The base-pair codes are treated like the

presence/absence (1/0) codes in a morphologi-

cal data matrix, and a variety of algorithms

are applied to extract the most likely tree that

explains the data.

The tree of life

Paleontologists and biologists are using mor-

phology and molecules to put together ever-

larger sectors of the tree of life. Desktop

computers are exploding in labs around the

world as analysts ask them to crank out ever-

larger trees. Bear in mind that the number of

possible trees is N = (2n − 3)!/(2

n−2

(n − 2)!).

So for three species, there are three possible

unique trees. For four species, there are 15

possible trees (Fig. 5.12), for eight species

168,210 possible trees, and for 50 species

about the same number as there are atoms in

the universe. You can do these calculations in

table 1.3.1 at http://www.talkorigins.org/faqs/

comdesc/section1.html.

And yet 50 species is not a demanding

number. Systematists want to know the com-

plete tree for all 240 species of primates, all

4500 species of mammals (now available:

Bininda-Emonds et al. 2007), and so on.

Mathematicians tinker with the tree-ﬁ nding

software so it ﬁ nds clever ways to ﬁ nd the

best-ﬁ tting tree quickly, even though many

millions or billions of potential trees are con-

sidered and rejected. Another approach is to

link existing trees for parts of the group of

interest to create a supertree that summarizes

the information in all existing trees. Research-

ers are currently using all methods and

approaches to draw major sectors of the tree

of life, and such huge trees will allow paleon-

tologists and biologists to carry out many

novel studies of macroevolution.

Review questions

1 Consider the four logical steps that sum-

marize natural selection (see p. 119), list

examples for each of numbers 1–3, and

consider how the mechanism could be

disproved.

2 Read books and web sites that present

evidence for intelligent design (ID). List

some speciﬁ c examples/case studies that

are cited by ID supporters, and present

these in the form of falsiﬁ able scientiﬁ c

hypotheses.

3 Find 10 paleontological case studies on

speciation, by searching the internet with

the key words “phyletic gradualism” and

(a)

DCB

D D

(b)

Figure 5.12 The number of unique trees for

three (a) and four (b) taxa. These cladograms

may be written more simply as (A(BC), (B(AC))

and ((AB)C) for the three-taxon cases, and

((AB)(CD)), ((AC)(BD)), ((AD)(BC)), etc. for the

four-taxon cases. Note that (A(BC)) and (A(CB))

are identical trees, and both versions count as

one.

136 INTRODUCTION TO PALEOBIOLOGY AND THE FOSSIL RECORD

“punctuated equilibrium”. Look at the

original papers, and consider how well

each one fulﬁ ls the four criteria for a good

study (abundant specimens, fossils with

living representatives, information on geo-

graphic variation and good stratigraphic

control; see p. 122). Which of the 10 studies

you have chosen is sufﬁ ciently well docu-

mented to give a meaningful conclusion?

4 How would you design a study to test

whether species selection might have

happened?

5 Construct a cladogram of the apes, and

identify apomorphies for each node. The

likely shape of the tree is (gibbon(orangu

tan(gorilla(chimp + human)))) (see p. 472)

– so read around and ﬁ nd out the cladistic

basis for the tree.