ÐÓ°ÉÔ­´´

The faces of our early ancestors have a powerful hold on our imagination.
Their distinctive anatomy attracts us because we know that our branch of
the family tree sprouted from theirs. But what were those earliest faces
really like? And how did they change during development, from infancy through
adolescence and into adulthood?

Traditional techniques give only a partial answer to these questions.
To go further it is necessary to uncover the mechanisms by which our ancestors’
faces grew and took shape. With this in mind, I began work on fossil faces
in 1980 with Becky Sigmon at the University of Toronto. The research has
shown that many of our earliest forebears were much more ape-like than many
anthropologists have come to believe.

Bones of the face grow and change shape through a combination of two
processes: deposition of new material and resorption of old (see Box 1 overleaf
for further details of these processes). The jaws, for instance, become
more prominent during growth if deposition of bone on the front surfaces
is combined with resorption at the back. Sites where these two processes
are occurring can be identified under the microscope. So, in theory, a survey
of the surfaces of the bones in a young hominid’s face should reveal a characteristic
pattern of deposition and resorption. From this, it should be possible to
identify areas that were advancing or retreating during development.

In practice it is trickier than that. The traditional method involves
cutting thin sections of bone and studying them under the microscope, a
procedure which was pioneered by Donald Enlow of Case Western Reserve University
in Cleveland, Ohio. It is one thing to cut sections of modern material,
but no curator of fossil ancestors is prepared to see their treasured objects
vanish under the knife. A nondestructive technique was needed if the analysis
was to be extended to our ancestors.

The cells responsible for deposition and resorption leave telltale traces
on the surfaces of bones, which I hoped might still be discernible in fossils.
In the 1970s, Alan Boyde and Sheila Jones of University College London pioneered
the use of the scanning electron microscope (SEM) to examine bone surfaces,
but they had studied fresh, undamaged material. Work with fossils required
a way of making a replica of the bone surface. At Toronto, Dennis Smith
and I developed a technique based on a liquid rubber-like material which
spreads like cake icing, coats the bone and then hardens – a replica which
preserved surface detail as small as 0.1 micrometre. This ‘negative’ is
sprayed with epoxy resin, which dries as a very thin hard shell, a perfect
replica of the original and perfectly suitable for studying with the SEM.

The next task was to assemble a collection of replicas of fossil faces
from immature hominids (members of the family Hominidae, which includes
mankind and extinct related forms). One of the most celebrated in the collection
was the Taung child – the first early hominid to be found in Africa, named
as Australopithecus africanus by Raymond Dart in 1925. I arranged to analyse
the replicas in Boyde’s laboratory in London.

First impressions were thoroughly disappointing. The replication technique
had worked well, but it seemed there was nothing worth looking at. The precious
fossils bore little resemblance to fresh, developing bones. It seemed that
their history – dragged around by scavengers, say, or washed downstream
to a lake’s edge or the mouth of a cave and then buried in tons of sediment
for hundreds of thousands if not millions of years – had all but obscured
the information on their surfaces. The only way around the impasse was by
controlled experiments, abrading the surfaces of modern bones, then looking
at the damage and for surface features beneath the damage. Gradually we
were able to separate damage from biology and return to the fossils and
begin to read the message on their surfaces.

Interpretations of early hominids (the australopithecines) have swung
back and forth over the decades, sometimes emphasising their apelike qualities
and sometimes their resemblance to humanity. In the early 1980s the second
interpretation was in vogue, and under the spell of this idea I half expected
the face of the Taung child to bear a surface pattern rather like a human’s.
It did not. Instead, the pattern for the Taung face was typical of a monkey
or ape. The parts of the face that were directed forwards were areas of
deposition of bone, not of partial resorption which is the pattern in modern
humans. The Taung pattern would have resulted in jaws that jutted forwards.

This result was soon joined by other evidence that pointed in the same
direction. The evidence concerned the rate and pattern of dental development.
By studying lines of growth in tooth enamel, Chris Dean and I, working at
University College London, showed that the teeth of many early hominids
developed at a rate reminiscent of apes.

The Taung child was rather difficult to fit into this picture at first
because most of its permanent teeth were still developing inside its jaws
at the time of death. However, Glen Conroy and Michael Vannier of the University
of Washington devised a way of studying these developing teeth by means
of the technique known as computerised tomography. They confirmed that the
teeth were indeed developing in a distinctly apelike manner.

It now seems that the Taung child died at the age of three years, not
six as had previously been assumed. Its rate of growth and maturation was
fast – much more like that of an ape than that of a human.

Further evidence has come from Holly Smith of the University of Michigan,
who has drawn attention to the close relationship between various life history
traits – in all primates – such as the size of the brain in adults, body
size, features of dental development and so on. These relationships have
shown independently that the development of early hominids must have been
more like apes than modern humans.

One of the central aims of anthropology is the reconstruction of a family
tree that summarises our evolutionary history, or phylogeny. Information
on facial growth patterns has an important bearing on these reconstructions,
because similar patterns of growth can be a sign of shared ancestry (see
Box 2 for a cautionary tale on this subject). Again, the evidence from fossil
faces runs counter to what might be described as the ruling paradigm in
the science.

That paradigm can be expressed in its simplest form as a Y-shaped family
tree (see Figure on opposite page). The trunk of the tree represents Australopithecus
afarensis, most fossil remains of which have been found by Don Johanson
and his colleagues at sites in Ethiopia that are around and older than three
million years old. One branch represents an early species of Homo – the
forerunner of modern humans, Homo sapiens. The second represents species
of Australopithecus and the related genus Paranthropus, in which brain size
remained small and the size of the cheek teeth and jaws increased, creating
a characteristically robust appearance. This second branch became extinct
about one million years ago. (Some anthropologists do not use the name Paranthropus;
they put all the earliest hominids into the genus Australopithecus.)

The Y-shaped tree implies that A. afarensis was the last common ancestor
of all later hominids. This conclusion rests on two main arguments. The
earliest representatives of both branches display suites of characters that
researchers interpret as signs of descent from A. afarensis. Furthermore,
because A. africanus and Paranthropus both evolved in a different direction
to early humans – especially as far as their grinding teeth were concerned
– neither can be directly on the line that led to mankind. A. afarensis
is left at the fork of the tree. An analysis of facial growth and architecture
points to an alternative view.

The developing anatomical relationship between the brain case and the
face in early hominids is particularly revealing. The skulls of most mammals
are organised in such a way that a line can be drawn connecting three points:
the base of the brain, a protuberance on the upper jawbone called the maxillary
tuberosity, and a point between the upper central incisors. Furthermore,
if lines are drawn from the ear opening to the maxillary tuberosity and
to the centre of the eye socket, they make an angle of about 45 degrees.
These two patterns result from the way the skull develops at important growth
sites and boundaries, and they are shared by most mammals, including modern
humans. However, the apes and the earliest hominids are exceptions to this
rule. They do not conform to the normal mammalian pattern.

In the chimpanzee, for example, the front of the upper jaw is high up
above a line connecting the base of the brain and the maxillary tuberosity
(see Figure on previous page). The same is true of A. afarensis and some
A. africanus, but Paranthropus and some early Homo (ER1813) begin to reapproach
the normal mammalian condition as in humans. Furthermore, in A. afarensis
and in some A. africanus (Sts5), the angle formed at the ear opening is
within the range found in chimpanzees, while in other A. africanus (Sts71),
all Paranthropus and early Homo, the angle increases away from the generalised
mammalian condition and less like humans. These similarities between Paranthropus
and Homo create a problem for the conventional view of hominid evolution,
because they imply that similar traits were acquired independently in both
branches of the Y-shaped tree. Another explanation is that Homo and Paranthropus
inherited those traits from a common ancestor. Who is this ancestor?

THE KENYAN CONNECTION

Before answering this question, it is important to consider a fossil
from East Africa that dates from the same period as the South African australopithecines
– a specimen 2.5 million years old from Kenya, known as WT 17000, or the
Black skull on account of its colour. Researchers have noted some specific
similarities between WT 17000 and A. afarensis. However, WT 17000 combines
these qualities with a facial structure and robustness that are typical
of Paranthropus. Some specimens of Australopithecus africanus (for example,
Sts71) also show such a combination. It seems wherever we look, we find
fossils of this period, 2.5 million years ago, with a mixture of features
characteristic of Australopithecus and Paranthropus. But because we know
that Paranthropus and early Homo share so many characteristics, perhaps
we should start to reconsider the standard view that these have evolved
independently. Paranthropus and Homo may be related more closely than has
been supposed – and may have shared a common ancestor similar to A. aethiopicus
WT 17000.

Because Paranthropus is so extreme in its anatomy – with enormous cheek
teeth and jaws – there has been a tendency to deny it any connection with
the ancestry of Homo. This may be incorrect. One of the most complete (and
best known) of the early specimens of Homo is ER 1470, a fossil cranium
and face 1.9 million years old from east of Lake Turkana, in northern Kenya.
When it was first reconstructed, the face was fitted to the cranium in an
almost vertical position, much like the flat faces of modern humans. But
recent studies of anatomical relationships show that in life the face must
have jutted out considerably, creating an ape-like aspect, rather like the
faces of Australopithecus.

With this new perspective on ER 1470 there comes a provocative resemblance
to WT 17000, the much older Black skull. To be sure, ER 1470 is Homo in
many respects and it has a phenomenally large brain for its time. Normally
this would be enough to rule out any comparison with robust forms like WT
17000. Yet could WT 17000 have been close to the evolutionary threshold
at which the early hominids reacquired the standard mammalian skull pattern
– by dropping down the front of the jaw and increasing the angle at the
ear opening – a pattern shared by Paranthropus and early Homo? This could
put it on, or close to, the main line of human evolution.

If these conclusions are correct, then A. afarensis would not be the
last common ancestor for Homo and Paranthropus. This is because that ancestor
must be a form that was already evolving the characteristics that occur
in both Homo and Paranthropus. WT 17000 represents the best small-brained
candidate for that common ancestor in East Africa. In South Africa, A. africanus
may serve as the most appropriate candidate for this ancestor.

The following picture emerges from these studies of our ancestors’ faces.
A. afarensis, from East Africa, is the most primitive early hominid and
its development was similar in many respects to that of modern chimpanzees.
A. africanus is South Africa’s gracile form, representing southern dispersions
of the genus, which did eventually evolve a more ‘robust’ morphology – by
altering rates of facial growth and remodelling. At the same time WT 17000
was evolving along similar lines at the northern extent of the australopithecines
range. A. africanus in the south and WT 17000 in the north provide tantalising
glimpses of human origins. I say glimpses because the relationships between
the various candidates are unclear. Whether Homo and Paranthropus sprang
from one common ancestor, or from more than one in various parts of the
Great African Rift, is a matter for conjecture.

Most researchers agree that there were at least two species of Paranthropus
and many think that there must have been at least two species of early Homo.
To be credible, any theory, however robust, must acknowledge this wide range
of variation. New fossil faces will emerge – and the story will definitely
change – but at the moment I think it is an explanation worthy of the available
facts.

Tim Bromage studies facial development at Hunter College in New York
City.

* * *

1: THE CHANGING FACE OF BONE

Bones grow and change shape by means of two coordinated processes: deposition
of new bone by cells called osteoblasts, and resorption of existing bone
by another set of cells called osteoclasts. These processes, together referred
to as remodelling, help to create the pattern of facial growth that characterises
each species.

The cells responsible for remodelling leave characteristic microscopic
signs on the bone surfaces. Fresh bone comes in three textures that can
be identified in samples removed from the surface. Surfaces on which resorption
has occurred are marked by the sharp-edged excavations of osteoclasts. Surfaces
where deposition is taking place lack these excavations and show evidence
of mineralisation – the process in which mineral salts are laid down in
a protein matrix secreted by osteoblasts. Surfaces on which deposition has
stopped are overlaid with inactive osteoblasts.

* * *

2: A TALE OF TWO FACES

When two species share an anatomical feature, researchers must decide
whether the resemblance is the result of common ancestry or independent
evolution. Methods for distinguishing between these two explanations are
very important in evolutionary studies, as the following example makes clear.

There are many anatomical differences between the faces of Australopithecus
afarensis and Homo sapiens. But there are also a number of apparently primitive
features common to both faces, which we may have inherited from A. afarensis.
One striking similarity is the way the cheekbones are swept back relative
to the upper jaw – a point that could indicate that A. afarensis is ancestral
to humankind.

I have analysed the way in which the cheekbones are built during development
in the two species. If the human species inherited the characteristically
shaped cheekbones from A. afarensis, then the bones would develop in an
identical way in the two species. They do not.

In A. afarensis, all forward facing surfaces of the growing face are
sites where bone is being deposited. The muzzle projects beyond the cheekbones
for reasons related to the size of the teeth and the method of chewing,
as in a living chimpanzee. In modern humans on the other hand, resorption
is taking place over most forward facing surfaces of the face. The cheekbones
are resorbed back to a level that exposes the jaws in front.

So although the two faces have some similarities, they are built in
very different ways during development. This particular characteristic cannot
be used in support of an ancestral relationship between A. afarensis and
humankind.