Wakayama Prefecture, where the village of Taiji is located. Ours is the
From Embryos to Evolution | 175
only plane. The runway is hewn out of black rock that erodes in steep
cliffs, and it is so narrow that the plane taxis to the terminal on the
runway—there’s no room for a taxi lane. A big whale is carved into the
hillside across from the terminal building, and an airport store sells
canned whale meat. On the two-hour drive along the rocky, winding
coast, we stop at a restaurant which sells whale-shaped incense burners
and piggy banks. Later, we drive by a giant waterfall in the shape of two
whale flukes.
It finally dawns on me that this is the heart of Japanese whaling
country. Most Japanese don’t eat whale meat, and I’ve never seen it on
any of my many previous trips to Japan. However, the government tries
to encourage its consumption, and in some areas, such as here, whale
and dolphin hunting is part of the local culture. These people fiercely
protect it.
the marine park at taiji
The next morning we go to the Ocean Park, where the skeleton of a
large blue whale is mounted outside. The white bones contrast with the
black rocks that frame the valley, creating a very Japanese image, ready
for a woodcut artist to immortalize. A curator meets us, but all the con-
versation is in Japanese, and Jim and I just walk along. Then the direc-
tor—the man who apparently pulled the plug on our talks—comes out.
He too is very courteous, and introduces us to the veterinarian and
trainers. Business cards are exchanged in a very Japanese ritual, and the
stack of my own cards diminishes as my stack of received Japanese
cards rises higher and higher. The director’s card has a picture of Haruka,
and I realize that this is an important attraction for them. We walk past
a large building with two whales painted on it and along a sea arm that
reaches between the cliffs. As I look at the water, a large black fin
emerges from it, followed by a loud whoosh. It startles me, but I recog-
nize that a killer whale has come up to breathe. Why is such a large
animal swimming in this narrow cove? I scan the bay and realize that
the entrance to the cove is dammed, and forms a natural aquarium for
their captive killer whale. Farther down, there are rectangular wooden
structures floating on barrels in the water. The water inside occasionally
ripples with something dark: these are pens with cetaceans in them. I
count at least six pens, some with multiple animals. We walk down a
winding path to a round concrete building, built around a large tank. As
we walk in the door, I realize that we are in a giant plexiglass tube at the
176 | Chapter 13
figure 56. Haruka, the dolphin with small rear flippers, seen from below. The rear
flippers are its hind limbs. Normal dolphins do not have rear flippers; the structures
developed in this individual as a congenital anomaly.
bottom of the aquarium, and three bottlenose dolphins are lying on top
of the tube. One is Haruka (figure 56). We all crowd around with our
cameras. The dolphins are not bothered. Two of them start to play. The
third stays; it is Haruka. She seems to know that she is the star of the
show, and wants to please the fans with photo ops.
Haruka has real little flippers in the back, one slightly larger than the
other, but nicely formed, like miniature front flippers. The trainer says
that one of them is much more loosely attached than the other. The
director, the veterinarian, and the curators have stayed with us. They
motion us away: “Please, please.” Do they want us to leave because we
have antagonized them, I wonder?
No, quite the opposite. They lead us to the roof of the building, to the
top of the aquarium. The trainers take coolers and go onto a platform
in the tank, and the dolphins show up immediately. The trainers feed
fish to the dolphins. One of the trainers gestures, and Haruka rolls over,
exposing her belly. The trainer gently puts his hand underneath the flip-
per and we shoot pictures and video. All captive dolphins are taught to
go belly-up so that a trainer or vet can examine their genitals and anal
region, and take their temperature rectally. It lasts a few minutes, and
then the dolphin receives another signal, rolls over, and loudly snorts.
When she is on her belly, she cannot breathe. She gets another fish, and
is told to roll over again, to the sound of more camera clicks and more
movies. An hour passes before I realize it.
From Embryos to Evolution | 177
shedding limbs
Haruka is not the first modern cetacean born with hind limbs. There are
more than half a dozen reports of whales and dolphins with such struc-
tures,2 varying in size from the little bumps on the abdomen of a Rus-
sian sperm whale to the four-foot-long appendages in a humpback
whale caught near Vancouver Island, Canada, in 1919.3 But unlike all of
the others, Haruka is alive.4 She might be able to teach us how hind
limbs affect swimming, and why whales lost their hind limbs in the first
place.
To understand why Haruka is different, we need to understand how
limbs develop in other mammals. Limb embryology is relatively similar
across vertebrates, from fish to birds to mammals. Early in development,
an embryo looks a lot like a worm, with a head, segments along its back,
and a tail, but no limbs. Human embryos look like this too, until in the
fourth week after fertilization, when the embryo is still smaller than a
pea, it grows two small bumps in the chest region. Later in the fourth
week, two small bumps appear at the base of the tail. Limb buds form in
this way in all vertebrates, but there are many differences too. Timing is
one: in a mouse, limb buds form much earlier, around day 10. To make
it easier to compare embryos of species with different gestation times,
embryologists have divided development into so-called Carnegie stages.
Limb buds in most mammals start forming at Carnegie stage 13.5
Limb buds consist of two kinds of cells. The outside is covered by a
single layer of flat cells that cover the embryo like the pavement on a
street. These are called epithelial cells. Inside, the entire bud is filled with
undifferentiated cells that make up the mesenchyme. The epithelial cells
on the top of the limb bud form a crest called the apical ectodermal
ridge, or AER. The limb bud grows, and inside of it clusters of cells
clump together (or condense) and form cartilage bars that will turn into
bone. One such bar forms between shoulder and elbow—it will be the
humerus. Two bars form between elbow and wrist: the radius and ulna,
the lower arm bones. Five distinct bars of cartilage take shape in the
hand, the precursors of the fingers. In most mammals, roughly the same
bars form in the lower limb, to make the leg and toe bones. Once the
cartilage bars have formed, cells that will form muscles migrate into the
developing limb from the body. Initially, there are no separate fingers in
the hand or distinct toes in the foot. The cartilage bars are embedded in
a flat pad of tissue, and the hands and feet look like mittens without
even a thumb. As the hand and foot develop, the tissue between the
178 | Chapter 13
AER makes FGF8
AER makes FGF8
AER makes FGF8
ZPA
AER makes FGF8
makes
epithelium
ZPA makes SHH
SHH
mesenchyme
Limb bud
body of embryo
figure 57. Diagram of embryology of the limb in a vertebrate. The forelimb and hind
limb initially form as a small limb bud (left diagram) that projects from the body wall.
Over time (drawings further right), the limb bud will grow, and eventually a skeleton
will form inside it. AER, apical ectodermal ridge; FGF8, fibroblast growth factor 8 (a
protein); ZPA, zone of polarizing activity; SHH, sonic hedgehog (another protein).
digits thins and eventually disappears to make five independently mov-
able fingers and toes.
The genes regulating all this are reasonably well known (figure 57).
Initially, the AER produces a protein called FGF8,6 which leaks into the
mesenchyme underneath it. An area like this, which produces a protein
that is used to communicate to other areas, is called a signaling center.
A group of mesenchymal cells in the rear of the limb bud also becomes
a signaling center, the zone of polarizing activity (ZPA). The ZPA now
starts to produce a protein called sonic hedgehog (after the videogame
character), a name usually shortened as SHH. SHH diffuses into the tis-
sue around it and is necessary to keep the AER alive and working at this
stage. The mesenchymal cells immediately underneath the AER divide
and as the limb bud grows longer and longer, cartilage bars form and
these divide the limb into segments: in the forelimb, shoulder to elbow,
elbow to wrist, and hand; in the hind limb, hip to knee, knee to ankle,
and foot.7
The ZPA plays a role early on, working with the AER to accomplish
limb-bud growth. The ZPA again plays a role later in development, dur-
ing the formation of the fingers and toes. SHH produced by the ZPA
oozes into the surrounding mesenchyme, and, because the ZPA is
located on the pinky side (posterior side) of the hand (the little-toe side
of the foot), concentrations of SHH drop as you go toward the thumb
(big-toe) side. As growth continues, cells farther toward the thumb side
will receive both a lower concentration and a shorter duration of SHH
exposure. That is the signal that makes different fingers and toes: the
index finger is exposed only briefly to low concentrations of SHH, the
pinky for a longer time and higher concentrations, and the remaining
fingers are intermediate.8 This controls the particular shape of the fin-
gers and toes.
From Embryos to Evolution | 179
In many mammals, forelimbs and hind limbs follow the same
trajectory, but this is not true in cetaceans. In the forelimb,9 develop-
ment initially proceeds as in most other mammals, until the soft tissue
that connects the digits fails to disappear and make separate fingers.
In addition, unlike many other mammals, many cetacean species form
more than three phalanges per finger or toe (figure 13). This all makes
a smooth, asymmetrical flipper, a blade that cuts through the water
and is used in steering.
The trajectory for the cetacean hind limb is quite different from that
of the forelimb. Although, in living cetaceans, external hind limbs appear
only if development goes wrong—in animals like Haruka—every living
(and presumably fossil) cetacean had hind limb buds as an embryo (fig-
ure 58). Those buds form but eventually disappear long before birth, as
their developmental trajectory is cut short. This leaves only some internal
structures that are attached to the genitals (figure 15).
Hind limb buds in cetacean embryos were discovered a long time ago,
and that story holds some lessons for us in the present. In The Origin of
Species, Darwin did not make much of embryology’s contribution to evo-
lution, but mainland European embryologists enthusiastically adopted
his ideas about evolution, because these allowed them to interpret previ-
ously inexplicable features of embryos. In 1893, three decades after Dar-
win’s book, the German embryologist Willy Kükenthal interpreted the
two bumps low on the abdomen in a porpoise embryo as hind limb buds.
Kükenthal’s publication10 was much to the chagrin of his Norwegian col-
league Gustav Guldberg, who had talked about hind limb buds in prena-
tal cetaceans in a lecture a few years earlier. However, Guldberg’s buds
were not in the same place as those of Kükenthal, and also at a different
time in development. Guldberg, who worked in a small institute in Ber-
gen, off the academic beaten path, had failed to publish his finding. He
felt scooped by the professor from the famous university in Jena, but
remained civil: “I was thus rather surprised . . . that in the work . . . of my
friend Professor Kükenthal, . . . he appears to believe to show the pres-
ence of the early beginning for hind limbs in a porpoise embryo of 25 mm
length.” Kükenthal’s bumps were too far forward on the body to be hind
limb buds, according to Guldberg, and instead he described buds in addi-
tional porpoise embryos of 7, 17, and 18 mm that document earlier stages
of development.11 Kükenthal dismissed Guldberg’s buds as the beginnings
of the mammary glands instead of the limbs, writing: “Grave reservations
arose in me, when I read that Guldberg considers the hind limb buds to
be two prominences in embryos of 17 and 18 mm.”12
180 | Chapter 13
Carnegie stage 13
Carnegie stage 14
Carnegie stage 16
length: 6 mm
length: 9 mm
length: 11 mm
LACM 94656
LACM 94738
LACM 94722
forelimb
bud
penis
hind limb bud
hind limb
al squares have
bud
1 mm sides
penis
nipple
hind limb
bud
Carnegie stage 17
LACM 94650
umbilical
cord
penis
nipple
anus
hind limb
penis
bud
hind limb bud
Carnegie stage 19
Carnegie stage 17
length: 72 mm
Carnegie stage 19
length: 22 mm
LACM 94743
LACM 94743
LACM 95041
figure 58. Dolphin embryos ( Stenella attenuata) at different stages of
development and not at the same scale (black squares have 1 mm sides for all
specimens). These photos show the increase in size of the forelimb bud and its
development into a flipper. They also show the increase in size of the hind limb bud
(first and second photo) and its subsequent reduction in size, and disappearance.
Nipples in dolphins are located next to the genitals, and they are visible in the
lower two photos (and their enlargements on the right). Embryonic stages are
referred to as Carnegie stages. Length of the embryo (CRL, crown-rump length) is
a good indicator of age as embryos grow rapidly.
The confusion is not as silly as it seems. In an early embryo, a bud-
ding hind limb and a mammary gland are not so very different. Both
begin as a swelling filled with mesenchyme and covered by epithelium,
with one patch of the epithelium thickened. In addition, the nipples of
modern cetaceans are in fact located on either side of the genitals, near
where limbs would be. Land mammals such as moles and squirrels have
nipples in their groin, and these form in the embryo as low welts between
From Embryos to Evolution | 181
genitals and hind limb.13 However, nipples start to form long after the
limb buds. Human nipples form in the seventh week of development,
The Walking Whales Page 27