every detail of their shoulder, hip, thigh, and ankle. As a conse-
quence, questions about dinosaurs which had long baffled the
scholars could be solved when Archaeopteryx was used as a stan-
dard of comparison. Deinonychus, for example, had had a strange
lower shoulder-blade bone (coracoid), unique among dinosaurs
because of its great depth from shoulder socket to breastbone. But
Archaeopteryx had had a coracoid of the same deep pattern as Deino-
nychus. Clearly both the bird and the dinosaur had evolved the
ARCHAEOPTERYX PATERNITY SUIT: THE DINOSAUR-BIRD CONNECTION | 313
unusual shape to increase the size of their breast muscles. Deino-
nychus had had a peculiar upper hip bone (ilium) compared to other
predatory dinosaurs, but Arcbaeopteryx exhibited the same pecu-
liarities. And in its wrist Deinonycbus's similarities to birds were
nothing short of astounding.
The wrist bones and hand—the structures that had first caught
John Ostrom's attention—were, in fact, the single most persuasive
part of his argument. The wrists of primitive reptiles were simple
devices, merely a flexible mosaic of squarish bones held together
by ligaments. When a primitive reptile pressed its wrist against the
ground, the mosaic of bones could bend or twist but couldn't pro-
duce any precisely controlled movement. Arcbaeopteryx and all
modern birds are different. The joint surfaces of wrist bones are
complexly curved, and the large central bone has an elegantly de-
signed joint surface of semicircular shape that guides the animal's
hand in a precisely controlled flexing movement. Prior to Os-
trom's discoveries, most scientists tended to believe that the bird
type of wrist had evolved rather suddenly, in the first true birds.
Neither primitive dinosaurs nor the false-croc reptiles Heilmann
favored as the ancestors of birds possessed anything like the bird
type of wrist. But Ostrom found that Deinonycbus's wrist bones
were identical to those of Arcbaeopteryx, and that Deinonycbus's wrist
would therefore deliver the very same sort of precise flexing
movement in the entire set of fingers. Moreover, the long fingers
so distinctively characteristic of Deinonycbus had been designed
identically in Arcbaeopteryx. Both Arcbaeopteryx and the dinosaur
had had three fingers only—not the five found in primitive dino-
saurs. And the proportions of the fingers had been the same: A
short, stout thumb and two longer outer fingers, with the outer-
most of the three very slender, bowed outward, and closely bound
by ligaments to the middle finger. This unique pattern can still be
recognized in a modern bird's wing; the three fingers are all firmly
fused together in an adult bird, but in an unhatched chick, the bones
are not yet fused. In a chick the separate wrist and hand bones are
clearly discerned, exactly as they had been in Deinonycbus and Ar-
cbaeopteryx.
There exists today one species of bird that retains its finger
bones unfused and flexible into the first weeks of life in the nest.
This bird, the hoatzin of South America, allows us to surmise how
314 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
the Archaeopteryx worked. As birds go, an adult hoatzin exhibits
nothing special in the anatomy of its wing. But the young nestling
is a genuine evolutionary throwback, an ugly little chick that climbs
through the vegetation by grasping with its three-fingered, claw-
tipped hands designed to the Archaeopteryx blueprint. The hatch-
ling can thus escape predators—snakes and hawks—by using its
wing-claws to climb out of its nest and work into the labyrinth of
vines surrounding it. Heilmann had drawn diagrams of wrists of
these hoatzin chicks next to those of Archaeopteryx in his book—
the anatomical identity was so stunning. But Heilmann didn't have
Deinonychus for comparison. If he had lived to see Ostrom's dis-
coveries, I'm certain that Heilmann would have converted to the
dinosaur-bird theory.
Hoatzin chicks also force a rethinking of the idea that there
could be no big reversals in the evolution of birds. Evolutionary
reversals unquestionably were necessary to make a hoatzin. Hoat-
A hoatzin chick climbing
with its wing claws
ARCHAEOPTERYX PATERNITY SUIT: THE DINOSAUR-BIRD CONNECTION | 315
zin's relatives all have much weaker wing claws in the chick stages
of life than hoatzins themselves have. Most ornithologists there-
fore conclude that hoatzins evolved from some ancestor with the
"normal" pattern of growth in which the chick never possesses
strong, flexible, unfused fingers for climbing. According to this view,
the hoatzin chick evolved by means of a Darwinian U-turn—the
strong, Archaeopteryx-like flexible fingers were recalled from ge-
netic storage.
Genetic storage is a nuance of evolution too often ignored.
Many paleontologists believe that when a bone disappears in evo-
lution, the genetic blueprint for that bone is also erased. Hence,
when dinosaurs lost their clavicle, their genetic code also suppos-
edly lost the instructions for making collarbones. If evolution really
occurred in this fashion, Heilmann would unquestionably have been
right when he maintained that re-evolution of a lost clavicle was
most implausible. Re-evolution of a lost set of genes for making
clavicles would entail a highly unlikely swarm of mutations and
natural selections.
But in fact evolution does not occur in this fashion. Hoatzin's
ancestors never lost the genetic blueprint for producing Archaeop-
teryx-style clawed fingers. In essence, they merely turned off the
physiological switch that ordered genes to produce organs accord-
ing to the encoded information. Recent advances in genetic re-
search reveal that most species carry such blueprints that are
"switched off" and can't express their code as fully formed tissue.
In other words, when an organ has been "lost," most of the time
its blueprint is still there, in genetic storage. Hoatzin's ancestors
were "normal" modern birds that employed a modern blueprint to
produce a wing in their nestlings that was like a chicken's, with
stiff, fused fingers. Hoatzins evolved their distinctive Archaeopteryx-
like clawed fingers by the process of turning off that blueprint
for its nestling and turning back to the older one to reexpress
itself.
A wealth of evidence supports this theory of reexpression by
genes that have been turned off for millions of years. Most of it
occurs in throwbacks (what nineteenth-century scientists called
atavisms), the rare appearance of ancient organs in species that, as
a whole, had lost the anatomical features millions of generations
earlier. A good example is multi-toed horses. Modern horses be-
316 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
long to the same general group as tapirs, and tapirs have four toes
on each forefoot. The single-toed modern horse evolved from a
four-toed ancestor. Every so
often a healthy, normal, single-toed
mare gives birth to a colt that has little extra toes sticking out be-
side the big main toe. Zoologists point to this multi-toed foal as a
case where natural processes allow a bit of the ancestral blueprint
to show through, letting ancient ancestral traits reexpress them-
selves.
Whales offer a more spectacular case. Modern whales have no
hind legs at all, and even when all the blubber and muscle are
flensed from the hip region, there is no remnant of the hip bones
except a small splint representing the ilium. Even the oldest-known
fossil whales display only slightly enlarged hip bones and some
remnants of thigh and knee. But way back in their ancestry whales
did have big hind legs, at a stage when they were land-living pred-
ators. And every once in a while a modern whale is hauled in with
a hind leg, complete with thigh and knee muscles, sticking out of
its side. These atavistic hind limbs are nothing less than throw-
backs to a totally pre-whale stage of their existence, some fifty
million years old.
Such throwbacks even occur in human infants. Hospitals oc-
casionally register an entirely modern-looking baby characterized
by all the expected organs, plus an unexpected tail, a long, caudal
appendage protruding beyond the buttocks for two or three inches.
Some of these tails are even bigger than the average caudal rem-
nant displayed by our close kin, the chimps, gorillas, and orangu-
tans.
Genetic experiments have revealed that these throwbacks are
controlled by suppressor genes. We now know that most complex
pieces of anatomy—such as the clavicle and its muscles—are con-
trolled directly and indirectly by scores of genes that interact and
can suppress each other. We also know that the full genetic blue-
print in any single species is rarely, if ever, fully expressed. In-
stead, much of the genetic information is stored in the "inactive
file," genes that don't produce their potential impact because some
other gene prevents them from turning on. When an anatomical
feature disappears during evolution, its genetic blueprint is not
erased. Some new combination of genes has evolved to suppress
the still-present blueprint.
ARCHAEOPTERYX PATERNITY SUIT: THE DINOSAUR-BIRD CONNECTION | 317
Birds with teeth may have appeared ridiculous to creationists,
but in point of fact modern birds do carry the ancestral genetic
code for making teeth tucked away in their inactive file. No living
species of bird manufactures teeth. But recent surgical manipula-
tions of bird embryos demonstrate clearly that the potential is still
there. In 1983, experimenters transplanted tissue from the inner
jaw (dental lamina) of an unhatched chick to an area of the body
tissue, where the graft could grow. In the transplanted position,
the chick's dental lamina started to produce tooth buds! Birds with
teeth could grow right in the twentieth century.
Suppressor genes solve Heilmann's paradox, the problem of
evolving birds with big collarbones from dinosaurs with atrophied
collarbones. Evolution at some point must have been able to re-
move the genes that suppressed collarbones from the dinosaur that
was ancestor of the birds. This is not farfetched, nor even mildly
implausible. The scenario might have run like this: A long-armed
dinosaur, such as Deinonychus, might evolve extra-long arms with
extra-long scales (feathers are modified scales) and begin to jump
from branch to branch, using its arm scales to gain a few extra feet
of glide, much like a flying squirrel. This proto-bird has no collar-
bone, but its ancestors long before did. The genetic code for a
collarbone remains in the dinosaur, stored in its inactive genefile.
Once the proto-bird uses its forelimbs for gliding, a bony strut in
front of the shoulder blade becomes an advantage. Any mutation
that removes genes suppressing the clavicle now becomes favored
by natural selection. In a few hundred generations, the proto-bird
could therefore re-evolve a collarbone, rearranging it a bit to pro-
duce the distinctive V-shaped wishbone characteristic of birds.
Why would dinosaurs begin to fly and thus cross the thresh-
old into the avian class? What was Archaeopteryx's niche? Most pa-
leontologists have leaned toward an analogy with flying squirrels.
Proto-birds are supposed to have been tree climbers who evolved
wings first for gliding, then for powered flight. But there's an al-
ternative possibility—the speedy jogger. Birds might have evolved
flight first by running at high speeds over the Mesozoic landscape,
employing their arms, outfitted with protofeathers, as airfoils for
increasing ground speed. According to this theory, hypothetical
proto-birds finally evolved a speed fast enough to become air-
borne. John Ostrom champions this speedy-jogger theory. He was
318 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
unhappy with the traditional restoration of Arcbaeopteryx as a tree-
climbing glider and flier. Among other details, he had observed
that Arcbaeopteryx's foot couldn't get the same grip on a branch as
can modern birds. Climbing birds have an inner toe that faces
backward and flexes forward to grasp a branch against the other
three toes. For the most efficient performance, all four of these
toes must be long and their base joint must be at the same level,
located at the very bottom of the long ankle bones (metatarsals).
Arcbaeopteryx's foot was not so built. The toe facing rearward was
too short and too high up on the ankle, so that its grip on a branch
wouldn't be anywhere near as effective as a modern bird's.
As an alternative, Ostrom suggested that perhaps wings first
evolved as catching devices. Today, small birds and bats use strokes
of their wings to sweep prey into their mouth. Arcbaeopteryx wasn't
a strong flier—its major feathers weren't fused to its arm and wrist
the way they are in modern flying birds. So maybe Arcbaeopteryx
had been a land-running predator that used its feathered, Deino-
nycbus-type arms to coerce prey.
I accepted this hypothesis in an article about the renaissance
of dinosaurs I published in Scientific American in 1975. But accu-
mulating evidence has forced me back to the orthodox view of Ar-
cbaeopteryx as a climbing and gliding flier. The aerodynamic shape
of its flight feathers is the first consideration. Flying birds today
have asymmetrical feathers—the leading edge is narrower and
stronger than the trailing edge. This is a necessity for powered flight
because air pressure is greater along each feather's front edge. Re-
cently, an ornothologist from North Carolina took the obvious step
of carefully examining Arcbaeopteryx's feathers—the first time any-
one had done so since the initial discovery in 1861. There was no
doubt, the wing's main feathers were asymmetrical. Therefore Ar-
cbaeopteryx very probably did indulge in powered flight, even though
it must have been a no
isy, slow, and inelegant performer in the
air. Furthermore, even though Arcbaeopteryx's foot didn't have as
precise a grip as the most specialized modern perching birds do,
it did have as much grasping power as many modern birds that
climb adequately. And Arcbaeopteryx wouldn't have had to rely on
its hind feet alone for effective climbing because its wings also had
hooklike claws. Arcbaeopteryx certainly could have clambered
through the ancient Bavarian vegetation as efficiently as any hoat-
ARCHAEOPTERYX PATERNITY SUIT: THE DINOSAUR-BIRD CONNECTION | 319
zin chick. Finally, if Archaeopteryx were a ground jogger, its hind
claws would have been blunt like those of a modern ground bird.
In fact, the Archaeopteryx's feet ended in needle-sharp claws. And
if it had run about on such pointed hind claws, it would have worn
down their horny outer sheath. Yet the fossils display hardly any
wear even on the delicate points of the claws.
More clues as to how Archaeopteryx developed flight come from
considering its teeth and claws combined. Archaeopteryx and the
Cretaceous birds from Kansas had teeth that terminated in thick,
barrel-shaped roots, like crocodiles'. Teeth so shaped are special
Bony claw cores and
horny sheaths
320 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
adaptations and are evidence for a seafood diet. Over a century
ago, Sir Richard Owen demonstrated that such teeth were a trade-
mark of the fish-eaters of the Mesozoic seas—the ocean lizards
(mosasaurs), and the fast-swimming fish lizards (icthyosaurs). Ar-
chaeopteryx and the Kansas birds were preserved in saltwater de-
posits full of fish, squid, shrimplike crustaceans, and other seafare.
The strong-winged Ichthyornis probably dove at fish from the air,
while the flightless loon-footed Hesperornis must have chased fish
underwater. Archaeopteryx is usually portrayed as a land feeder,
swooping down on oceanside prey along the beach from its roosts
in the seashore trees. But the shape of its teeth requires a differ-
ent hypothesis, a fishier one. Many modern fish-eating birds—puf-
fins, penguins, snakebirds—swim with their wings. Hoatzin
fledglings also swim underwater with strong strokes of their wings.
Archaeopteryx's hoatzinlike wings would have been fine for sub-
Robert T Bakker Page 32