fossils of those elusive first whales. The vertical wall of hard reddish-
purplish rock rises five feet out of the ground and the monsoons have
washed it for us, exposing delicate, beautifully preserved fossils. The
braincase that I saw some years ago is still there. The wall was originally
not vertical, it was horizontal. The movements that formed the Himala-
yas pushed it up and superimposed a pattern of crisscrossing cracks,
which make the wall look like it was built from carefully fitted, jagged
stones. The fossils stand out in bright white and often run across a
crack: after all, they were there before the rocks cracked. We take the
wall down block by block, keeping track of adjacent blocks so as to not
separate two parts of a fossil. We have a bit of a conveyer belt for fossils
going. One of us numbers all the blocks when they are still in the wall;
the next takes the individual blocks and brushes the dirt off them; then
they are handed to me. I sit with a heavy hammer, chisel, and hand lens.
I note where there are fossils and mark them, and someone else keeps
the fieldbook up to date: “Five-pointed star; humerus, matches five-
pointed star in Block 23.” Two more people label and wrap blocks. The
hunt goes well. We will be paying dearly in excess weight at the airport.
If there are no fossils visible on the outside of a block, I smash it to see
127
128 | Chapter 10
if there are some inside. If there are fossils, I trim the block, getting rid
of unnecessary weight and thus saving money on shipping.
By far most of the teeth we find are whale teeth, pakicetids, but there
are a few others. There is a tiny artiodactyl called Khirtharia, but we
have only a few jaws of it. Artiodactyls, or even-toed hoofed mammals,
include pigs and camels, goats and cows, hippos and giraffes, but
Khirtharia is much smaller than those—the size of a raccoon, but totally
unrelated to that carnivore. Teeth are actually not the most diagnostic
part of an artiodactyl. All artiodactyls are characterized by the particu-
lar shape of a bone in the ankle, the astragalus. In all mammals, the
astragalus is the bone on which the ankle pivots. To allow that, the bone
has a hinge joint called the trochlea that articulates with the shin bone
(tibia) above it. The other side of the astragalus is an area called the
head. It faces the foot and has different shapes in different mammals. It
is globular in most mammals, flat in horses, and has the shape of another
trochlea in artiodactyls. This double-trochleated astragalus is very dis-
tinctive, characterizing all artiodactyls from the smallest mouse deer to
the largest giraffe, including all the fossil ones. An astragalus for Khirth-
aria was among the bones collected by Davies and sent to Pilgrim in the
British Museum, long before I was born, and before Dehm went and
collected the first whale in Pakistan.
We also find lots of limb bones, and it is easy to identify those as
tibias, femurs, or humeri. It is not so easy to figure out to which animal
they belong. Given that most of the teeth are whales, most of the skele-
ton bones are probably also whales, but one cannot be sure. Size helps
some—given the big size difference of the teeth, it is not possible to
confuse a Khirtharia femur with that of a whale. Complicating this
could be a shadowy species for which no teeth have been found yet, but
only bones. That appears to be our problem here. There are a number
of large double–trochleated astragali at this locality. They are clearly
artiodactyl, based on their shape, but they are much bigger than Khirth-
aria. The species must have been pretty common, given that we have so
many bones of it, but we have never found teeth of that artiodactyl. It is
an enigma—but I do not worry about it. With the collection from this
site growing, the problem will go away, and we will eventually have
teeth and bones for all animals represented.
I muse about such matters as block 7 reaches me. It is so large that I
have trouble lifting it, but parts of several fossil bones are immediately
obvious on the outside. My hammer hits a corner of the block. Another
bang, and I gasp. The rock breaks, and the crack exposes part of a
The Skeleton Puzzle | 129
braincase. It looks like the pakicetid braincase that Philip Gingerich
found in 1981 (note 2 of chapter 1). That was a nice fossil, but the parts
with the eyes, nose, and jaws were missing. As a result, we don’t know
what the face of Pakicetus looks like. The areas in the front of the brain-
case cannot be seen in this one either, but it is possible that it is in the
adjacent block that is still in the wall.
I brush the new skull with water and a toothbrush, scraping dirt out
of the cracks, so that it can dry and I can glue the weak points. This
takes time. The others keep working, and a pile of blocks ready for
smashing forms next to me. I ask the others to find the block that was
adjacent to this one and wash it. Sure enough, the skull goes on into the
next block. This could be exciting. Another bang, carefully placed, with
adrenalin going to my head. The block breaks in two. My heart stops.
The crack reveals the eye sockets of the whale, perched on top of the
skull. The whale stares straight at me across forty-nine million years, as
if the rock were muddy water and the whale were sizing me up as prey.
I sit back and drop my hammer, and call the others to come and look.
Here is the best skull of the first whale known to people—what a find.
As I gently brush it off, I consider that the rest of its bones are at this
locality too. It will just take time to extract them from these rocks.
how many bones make a skeleton?
The critical question that we hoped to answer with fossils from H-GSP
locality 62 was this: What are whales related to? More pakicetid fossils
would answer that question, and this skull is an important part of the
answer. For more than two decades, there was no controversy among
paleontologists: whales are related to a long-extinct group called mes-
onychians (Mesonychia in Latin). It was an idea proposed by the bril-
liant and eccentric paleontologist Leigh Van Valen.1 He observed that
mesonychian teeth looked just like those of early whales. In both, a
lower
molar has a high trigonid with a single cusp and a low talonid
with one cusp, very unusual for a mammal (figure 34). A lot of fossils
are known for mesonychians—dentitions, skulls, and skeletons from
North America, Europe, and Asia—and they lived at the right time to
have given rise to whales.2 Their dentition suggests a meat diet, and
their body is vaguely wolf-like, but they are hoofed mammals: five toes
per foot, with a tiny hoof at the end of each. However, the paleonto-
logical romance with the mesonychian-whale hypothesis encountered
trouble from molecular biologists who found that, in terms of proteins
130 | Chapter 10
and DNA, modern whales are very similar to modern artiodactyls. So
similar, actually, that it appears that cetaceans should be included in
even-toed ungulates: their closest relatives are hippos, and hippos are
more closely related to cetaceans than to any other artiodactyl. That is
called a sister-group relationship. Of course, mesonychians are extinct,
and their proteins and DNA cannot be studied, and that leaves the pos-
sibility open that cetaceans and mesonychians were sister groups but
that the two of them combined form a group that is the sister group of
hippos. However, that did not sit well with the paleontologists either:
the double-trochleated astragalus occurs in all artiodactyls, including
hippos, but not in mesonychians, which seemed to exclude mesonychi-
ans from the artiodactyl group. In cetaceans, it is impossible to tell what
the astragalus looked like since all modern and nearly all fossil ceta-
ceans have lost their hind limbs. In basilosaurids, the ankle bones are
fused into an unrecognizable lump, and in remingtonocetids, no ankle-
bones are known. Ambulocetus was disappointing in this regard too:
we found half of an astragalus, but not the part that would have solved
the problem.
This is why a pakicetid skeleton is needed. It would provide a skeleton
of a cetacean sufficiently primitive to allow us to make direct compari-
sons to artiodactyls and mesonychians. Ankles would be of particular
importance to solve the artiodactyl-mesonychian riddle.
So, back in the United States, that is what we are going for. Ellen is
extracting the bones from the blocks from locality 62 in the hope of
finding enough of them to build a skeleton of a pakicetid. The trouble
with the locality is that there are no single skeletons: the bones of lots of
individuals and species are jumbled together here. Ellen has prepared
drawers full of locality 62 bones, and there are lots of whales, given the
teeth and skulls, but I cannot directly recognize which limb bones and
back bones go with those teeth.
I open the drawers frequently, fitting humeri on radii and tibias on
astragali. As I play with that unique jigsaw puzzle, missing pieces haunt
me. I pull some of the most common bones out of the drawer. They belong
to an animal that must be similar in size to the pakicetids. I put the bones
on a table. The foot bones together make a well-proportioned foot, but it
is not that of a whale, it is an artiodactyl instead: it has a double-troch-
leated astragalus. The two middle toes are similar in length, and much
longer than the side toes. That is another artiodactyl feature: each foot
has even numbers of equally sized toes (usually two long ones and two
shorter ones, or just two toes of similar length). This foot belongs to a
The Skeleton Puzzle | 131
figure 38. The skeleton of Eocene whale Pakicetus, put together from the bones
of a number of different individuals, all washed together at Locality 62 in the Kala
Chitta Hills of Pakistan, approximately forty-nine million years old. Study of stable
isotopes confirmed that these all represent bones of this early whale species. The
marker between the legs is 13.5 cm in length.
common locality 62 beast that really is an artiodactyl, so it is frustrating
that I do not have any large artiodactyl teeth from here. Ellen takes more
fossils out of the rock daily, but there are no large artiodactyl teeth.
The bones become an obsession. I leave them out on the table. The
vertebral column; the shoulder, forelimb, hind limb. But there is no skull
or teeth. I take one of the pakicetid skulls and put it at the front of the
skeleton. It fits the first vertebra (the atlas) very well, and size-wise, it
matches the skeleton (figure 38). It would solve the problem of the mys-
tery artiodactyl: the mystery artiodactyl is actually a whale. Ellen walks
in to show me a new bone that she just extracted from the rock. She sees
what I did and blushes, which she does easily. The skeleton on the table
is making a reckless statement about whale evolution: if that beast has
a double-trochleated astragalus, Van Valen’s great insight that whales
are derived from mesonychians would be wrong. Disturbingly, it would
mean that the teeth were lying to us—the detailed similarities between
mesonychian teeth and pakicetid teeth would be convergences, not
related to having a common ancestor.
Ellen and I ponder what to do next to see whether the fossil evidence
supports the idea. To give this molecular biology–inspired idea a chance,
we first have to study the relative abundance of fossils at locality 62. We
count all the bones and teeth. Of the teeth that I can identify without
doubt, 61 percent pertain to pakicetid whales. The bones are harder to
count—there are so many of them, and all the different kinds have to
be counted separately. After all the counting, the bones that are of the
132 | Chapter 10
correct size to fit an artiodactyl with that mystery astragalus are more
common than any other bones, just as the whale teeth are more com-
mon than any other animal’s teeth. One would expect the most com-
mon teeth to belong to the same animal as the most common bones at
a fossil site, so that supports the match between whale skull and artio-
dactyl skeleton.
Then, we look at the other animals known at locality 62. First is
Khirtharia, the raccoon-sized artiodactyl, which makes up 14 percent of
the identifiable dental fossils, the second-most common beast. We com-
pare its teeth to those of artiodactyls from Messel, the to
xic lake site in
Germany, where entire skeletons are preserved, articulated as if their
owners could jump up from the rock and run off. The mystery artiodac-
tyl bones are much bigger than the bones of a Messel artiodactyl that
has teeth the size of Khirtharia. Clearly, whales cannot be confused with
Khirtharia. The whale-artiodactyl hypothesis passes another test.
About 11 percent of the jaws and teeth at locality 62 belong to an
anthracobunid, the putative elephant-manatee relative, and it is the
third-most common mammal. Their jaws are bigger than those of paki-
cetids, and there are several large bones at locality 62, much larger than
the bones of the mystery artiodactyl. At a different place in the Kala
Chitta Hills, we found a partial skeleton of an anthracobunid: teeth,
skull, and bones, all of one individual. The bones are short and squat,
and match the proportions of those at locality 62. They are very unlike
the long and gracile bones of the mystery artiodactyl. Another road-
block eliminated.
I feel good about this, but it would be nice to confirm it with another
line of evidence. Isotope geochemistry comes to the rescue. Stable iso-
topes of carbon in pakicetid teeth and jaws at locality 62 are very differ-
ent from those of the teeth of the other mammals. Do they match the
bones? I eagerly await the results of Lois’s isotope study. They are a
match! Pakicetid dental isotope signatures match those of the bones of
the mystery artiodactyl, and they are different from the teeth of the
other mammals. The conclusion now becomes inescapable. Ellen and I
lay the skeleton out once more. It is the same skeleton, but now the
discomfort is giving way to a sense of amazement, and of victory.
The Walking Whales Page 20