during Carnegie stage 17.
Guldberg fought back, now with an exhaustive description.14 He had
hoped that more embryos would become available, but since they did
not, he was left restudying those that he had already. “As the short essay
by Prof. Kükenthal may have cast a dark shadow over my work . . .
detailed restudy confirmed my findings in all directions” (Guldberg’s
italics). That is where the conversation ceased, until Marga Anderssen,
also Norwegian, confirmed Guldberg’s interpretation some twenty
years later.15 She studied a larger collection of embryos, covering a
greater part of porpoise development. She confirmed beyond doubt that
the outer welt, the hind limb bud, develops earlier and disappears even-
tually, whereas the inner welt begins its development later, and eventu-
ally becomes the mammary gland.
The broader lesson is that these ancient embryologists were ham-
pered by the lack of material. They had only a few embryos. What they
needed was an ontogenetic series: embryos that cover all developmental
stages of one species, from some that are just a few millimeters in length
to those that already look like a miniature of the adult (in cetaceans,
these are the size of a mouse). Critical embryos, those between 7 and 17
mm, were missing in the collections of both Guldberg and Kükenthal.
Knowing about the hind-limb loss in evolution, and reading about
their embryology and the genes controlling it, has me psyched. There is
some real potential here to understand evolution at a deeper level.
Apparently, the genetic program that makes hind limbs in other mam-
mals is still operational in cetaceans—starting up as usual in early devel-
opment, but then grinding to a halt a few weeks later. Add to that the
great variation among modern cetaceans. Dolphins have just one bone
near the area of the hind limb, the pelvis, whereas others, like bowhead
whales (figure 59), have pelvis, femur, and tibia, all embedded internally.
That suggests that the developmental program is switched off at differ-
ent times in different whales. With so much information on the genetics
of limb development, and spectacular fossils documenting the transi-
tion, it would be amazing if one could figure out which of the limb genes
were altered in development, and when that evolutionary change hap-
pened in cetacean history. But how do you get an ontogenetic series of
cetacean embryos?
It is Bill Perrin, a scientist at the National Oceanic and Atmospheric
Administration in La Jolla, California, who points me in the right
182 | Chapter 13
hemal
arches
orbit
(for eye)
left and
cartilage
right
supporting
pelvis and
blowhole
hind limb
Forelimb
humerus
Pelvis and Hind Limb
pubis
ilium
radius
5 cm
ulna
femur
1 inch
scale bar for
tibia
metacarpal V
metacarpal I
top image only
A Fetus of a Bowhead Whale
five phalanges
(2000B3F)
on longest finger
figure 59. Bowhead whale fetus, treated so the soft tissues are transparent, bone
is purple, and cartilage is green. This technique is called clear-and-stain. Compare
the hand bones to the hand of an adult whale, in figure 13. This fetus also shows
the pelvis and hind limb, which are embedded in the abdomen in life (see figures 14
and 15). Baleen will form in the gap between upper and lower jaw.
direction. Bill was the person who, back in the 1980s, started to sound
the alarm bell over how many dolphins died in the tuna fishery. He initi-
ated a movement that eventually made the world a safer place for dol-
phins. In the process, scientists got access to many of the dolphins that
drowned in tuna nets, and embryos of the pregnant ones were collected.
The embryos ended up with whale scientist John Heyning at the Natu-
ral History Museum of Los Angeles County. When I contact John, he is
immediately intrigued by my question about the genes that form the
dolphin hind limb, and off we go to the museum’s warehouse, where
entire ontogenetic series of several dolphin species are housed in little
vials of alcohol. It is just the resource we need.
To study which proteins are active at what time in development, we
cut embryos of all stages into very thin slices and search them for specific
proteins involved in making limbs. A pea-sized embryo yields about a
thousand slices, each 7 microns thick. Such slices are mounted on glass
slides and are so thin that light shines through them, so they can be stud-
From Embryos to Evolution | 183
Two horizontal sections slide
slide
through the embryo
45
95
Brain
Somites that will develop
into vertebrae
Heart
One forelimb bud
damaged and missing
slide
slide
45
95
Umbilical
cord
Spinal cord
Forelimb bud
Somites
Spinal cord
Hind limb bud
Mesenchyme
Hind limb
bud
Tear in specimen
Dolphin Embryo
Epithelium
Stenella attenuata
LACM 94617
AER
Length of embryo: 7.2 mm
figure 60. Dolphin embryo at the stage when hind limb buds are largest. Lines mark
areas where thin slices (shown on right) were taken from this specimen. These slices are
affixed to microscope slides. Two of such slices are shown with enlarged areas showing
the sectioned hind limb bud. As development proceeds, the hind limb buds will regress
and disappear.
ied with a microscope (figure 60). The organs can be easily recognized
now, and the limb bud and its AER studied. Proteins made by the embryo,
FGF8 for instance, are embedded in the organ in which they were ini-
tially, now stuck to the glass slide. Then we douse the slice with another
protein, an antibody, designed specifically to search out and bind to the
FGF8 protein and only that protein. A special dye binds to the antibody
and reveals, in brown, areas where this FGF8-antibody complex is
located. Other dyes (histologists call them stains) are used to color the
184 | Chapter 13
Two dolphin fetuses
LACM 94671
clear-and-stained
> ( Stenella attenuata)
1 cm
pelvis
blowhole
enlargement of flipper
blowhole
vertebrae of sacrum
LACM 94310
left and right pelvis
figure 61. Clear-and-stained dolphin ( Stenella attenuata) fetuses. Note that the
smaller individual has less bone (purple). The photo of the flipper shows that all
bones of the hand are already present even though this fetus is tiny and long before
birth. See figure 59 for explanation of the technique.
tissues in purples and pinks, blues and reds, until the tiny embryo slice
looks like an expressionist painting, gorgeous and full of information.
Working with our dolphin embryos, we find FGF8 in the AER of
both forelimb and hind-limb bud. That was expected, of course. We
next use an antibody for SHH, and it indicates the presence of a ZPA in
the forelimb, but no SHH-producing ZPA in the hind limb. Thus, in the
hind limb, SHH is a broken link in the chain of genes needed to make a
limb. We conclude that the AER was present and functional initially in
both forelimb and hind limb, but that it dies prematurely in the hind
limb of dolphins as a result of the absence of an SHH-producing ZPA.16
That observation in dolphin embryos helps to explain why different
living cetaceans vary in their limb bones. In dolphins, the pelvis remains,
but no hind limb bones are left (figure 61). That indicates that the AER
dies early, the result of the total absence of SHH in the hind limb bud.
Bowhead whales, however, always have a femur, a piece of cartilage or
bone that represents a tibia, and occasionally even parts of a foot (figure
15). That could be explained by the longer life of the AER, which may
keep going in the presence of the ZPA for a longer time. Of course,
you’d need some bowhead whale embryos to prove the idea.
If SHH is a protein that changes over the course of whale evolution,
we might also use it to help us understand the hind limb shapes of fossil
whales. Basilosaurus from chapter 2 comes to mind, with its tiny hind
limbs that still contained femur, tibia, and fibula, plus three toes with
two phalanges each. So how can SHH contribute to that foot shape? As
we saw, SHH helps direct mesenchyme to form into fingers and toes.
From Embryos to Evolution | 185
The three-toed feet of Basilosaurus remind me of certain lizards called
skinks, which were studied by a developmental biologist named Mike
Shapiro.17 In some species of skinks, the number of digits varies between
individuals: there can be two, three, or four. Mike found out that more
fingers and toes form as the hand and foot of the embryo are exposed
longer or to higher concentrations of SHH. So we know that SHH plays
a role in the unique hind limbs of dolphins, and we know that its absence
causes loss of toes experimentally in mice, and in nature in skinks. Taken
together, all this suggests that the reduction of toes in Basilosaurus more
than forty million years ago was caused by a drop in SHH in the foot, a
decrease that foreshadowed the loss of all of the hind limb elements in
younger whales. The data from genes, embryos, and fossils complement
each other beautifully here and are able to explain the evolutionary pat-
tern and the modern shapes of the cetacean hind limbs.
Interestingly, during the part of Eocene whale evolution before Basi-
losaurus, there were great variations in the shape of different parts of
limbs among whales. That variation was related to function in chapter
4. However, in spite of that variation, Eocene whales always retain the
femur, tibia/fibula, and foot, even as the function of hind limbs in loco-
motion is lost. So the function of the hind limbs was reduced with no
changes in SHH action early on. This indicates that the very fundamen-
tal changes in the limb development process that eventually reduced the
hind limbs to nothing in the modern species were actually not driving
the significant functional changes related to different locomotor behav-
iors (figure 62).
Some developmental biologists believe that evolution is driven by
minor genetic changes very early in the embryo. That is based on the
understanding that such changes have the opportunity to modify the
resulting individual in very fundamental ways. However, in the case of
whale hind-limb evolution, the ontogenetically early developmental
change did not occur in evolutionary time until their locomotor func-
tion had long been lost.
Given all this knowledge from experiments and nature about SHH,
one wonders what happened to Haruka, the Taiji dolphin. She appears
to have the bones that are normally present in land mammals: femur,
tibia, fibula, and some ankle bones. There are two or three toes, one
with multiple bones. The pattern is roughly similar to that of Basilosau-
rus and skinks, suggesting that some mutation involving when and
where SHH was expressed, when Haruka was an embryo, helped to
create its hind limbs.18
186 | Chapter 13
Artiodactyla
Odontoceti
Mysticeti
toothed whales
modern
baleen whales
includes dolphins
and porpoises
Tragulus
radius
femur
Caenotherium
femur
femur
humerus
basilosaurids
40 million
femur
years ago
humerus
humerus
remingtonocetids protocetids
Indohyus
ambulocetids
50 million
pakicetids
Oscil ation
years ago
Origin
of fluke ?
Aquatic wading
Pelvic paddling Dorsoventral undulation
emur and tibia
Change in SHH
Gradual reduction of length of f
No loss of toes or other bones
gene function
leading to loss of
toes and phalanges
figure 62. Cladogram summarizing evolution of swimming modes in early whales,
with changes in bone density, and inferred change of SHH gene function. Irregular ovals
are cross-sections of bones; black indicates dense cortical bone; lattice is spongy bone;
and white is the marrow cavity.
Haruka also reminds us that the developmental process that makes
hind limbs in other mammals is
still locked into the cetacean genome; it
is just switched off in normal individuals. Maybe Haruka’s AER kept
going, long enough to make small rear flippers. It would be amazing if
we could sequence the DNA of this animal and compare it to that of
other dolphins. It is unlikely that the nucleotide sequence of SHH differs
between normal dolphins and Haruka, since that protein is involved in
so many processes. A mutation in it would probably cause lethal defor-
mations in the embryo. It is more likely that some gene involved in
switching the SHH gene on and off in the hind limb is different, causing
a difference in the timing of SHH expression. Those genes are called
regulatory genes.
If we could sequence Haruka’s entire genome, there would be thou-
sands of differences with other bottlenose dolphins, and most of those
would not be related to hind-limb development. However, the differ-
ences that did cause its hind limb bud to keep on growing would be
there too. And on those regulatory genes would be the fingerprints of
cetacean evolution. The same regulatory genes may also have effects on
From Embryos to Evolution | 187
other parts of the dolphin’s anatomy, and possibly those same genes
were involved in shaping the other parts of the anatomy of the Eocene
cetaceans so that different features that were evolving in cetaceans were
not inherited independently. These thoughts go through my head as I
think about Haruka. As DNA-sequencing technology gets cheaper and
faster, it should not be that hard to sequence bottlenose dolphins and
figure out what their differences are. And as we learn more about devel-
opment, it would make a really compelling story when combined with
the fossil evidence. Of course, one would need access to DNA from
Haruka. And that, I do not have.
whaling in taiji
Back in Taiji, our visit with Haruka ends. In the main building, the
director of the museum goes to the gift shop and brings each of us a tie
clip in the shape of Haruka. A poster for the aquarium features her as
its main attraction. Outlines of a series of fossil whales— Pakicetus,
The Walking Whales Page 28