rhinos and other large modern mammals and from the "sprawling
gait" of the ground-hugging lizards.
Alligators sprawled at the elbow much less than Professor Lull's
Centrosaurus, and yet the horned dinosaur was supposed to be a
much more advanced evolutionary design than the 'gator. Some-
thing was deeply wrong here. Why would an advanced dinosaur
exhibit a more sprawled posture than its more primitive relative?
I needed evidence from the shoulder-bone structure which I could
use to evaluate dinosaur forequarters. Two pieces of evidence came
immediately to hand: First, the shoulder socket's shape. An ele-
phant or rhino's shoulder socket is shaped like an oval saucer. It
is a hollowed-out joint surface, elongated fore to aft, which faces
downward and backward to fit over the top of the upper arm bone.
But lizards and crocs, whose elbows sprawl, have a saddle-shaped
shoulder joint, concave from bottom to top and convex from the
inside out. This saddle-shaped notch lets the upper arm swing out
and back and twist around like an axle, a complicated set of move-
ments required by the sprawling and semi-erect gaits. Now, what
kind of shoulders did dinosaurs have?
I spent a year digging into museum drawers, and covering
myself with dust while I diagrammed the shoulder sockets of the
Dinosauria. Almost all had rhino-type joints. When properly
mounted, dinosaur shoulder joints were concave sockets facing
downward and backward. Markings on the bones showed clearly
that the joint didn't curve around to face sideways as it did in 'ga-
tors or lizards. Professor Lull's Centrosaurus had a misaligned front
end, as did the mounts of most other horned dinosaurs.
The second piece of evidence reinforced the first. Crocodil-
ians and chameleon lizards had a semi-erect gait, and when I mea-
sured their shoulder joints oriented to a side view, I found that
both of these reptiles displayed a joint which slanted so that it faced
slightly downward as well as outward and the upper edge of the
joint overhung the lower edge. Fully sprawling lizards didn't ex-
hibit a trace of this downward slant. On the other hand, dinosaurs
THE TEUTONIC DIPLODOCUS: A LESSON IN GAIT AND CARRIAGE | 209
Shoulders designed for sprawling. Lizard shoulder joints are doubly curved
notches shaped like a saddle, and the normal walking posture is with the
elbows stuck far out to the side. (The upper-arm bone—the humerus—is
pulled out of the socket a bit in the diagram to show the fit.)
Horned dinosaur
shoulders were designed
for fully upright posture.
The upper edge of the
shoulder socket overhung
the lower edge a great
deal, even more than in
crocodilians. And, viewed
from the rear, the
shoulder socket faced
mostly downward, not
outward.
Shoulders for a more upright gait. Alligator shoulder joints are saddle-shaped
but face more strongly downward than do those of lizards, and so the gator
can hold its body higher off the ground.
all manifested very strong downward slants, so that their entire
shoulder socket had been reoriented from the primitive arrange-
ment. This strong downward orientation meant that the dinosaur's
upper arm could swing fore to aft in an upright stride. And the
upward force of this limb's stroke would be braced against the
downward-facing shoulder socket.
Finally, there was the acid test of fossil footprints. Quadru-
pedal dinosaur footprints aren't as common as those left by bipe-
dal types, but each and every set of four-legged footprints showed
forepaws working on a very narrow track. Triceratops and the rest
of the four-legged Dinosauria did not splay their forelimbs. Marsh
had been right in the 1890s, Lull wrong in the 1930s.
Lull's own account of why he mounted the Centrosaurus with
wide-set forepaws was quite surprising. Lull wrote that he had
THE TEUTONIC DIPLODOCUS: A LESSON IN GAIT AND CARRIAGE | 211
carefully studied the fossil footprints of big quadrupeds found in
illustrations of those prints in 1930, several years before Lull
mounted his sprawl-elbowed beast. But Sternberg's diagrams
showed right and left forepaws quite close to the centerline, and
not spread widely apart. Lull simply ignored this, because he was
so convinced, a priori, about splayed forelimbs that the obvious
facts simply didn't register, as they still don't for some. Several large
quadrupedal skeletons have been erected in various museums during
the last decade, and some still faithfully cling to the traditional stance
with the widely splayed forepaws, despite the publication of doz-
ens of footprint diagrams.
I was pretty proud of myself when I finished my undergrad
thesis on posture evolution. I published a couple of articles argu-
ing that the dinosaurian fully erect gait was superior to the sprawl-
ing gait because erect posture didn't waste as much muscular effort.
It seemed like a logical idea, and Al Romer had used it way back
in the 1920s. For example, if you do push-ups on the floor, you
can put your arms in the lizard-style posture by bending your el-
bows at right angles and holding your body halfway off the floor.
In this position, you feel very uncomfortable strain in your arm
muscles. If you hold your arms straight up and down, in a fully erect
Footprints don't lie.
All dinosaur tracks
show that the forepaws
were put down right
under the body with
only a little space
between the line of
march of the left and
the right set of prints.
But many museum
reconstructions still
show dinosaurs with
widespread forepaws
that would have left a
sprawling-style
trackway. (This
drawing is from a
model in the National
Museum of Canada.)
212 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
posture, you can keep your body off the floor with less effort.
When I got to Harvard, I had fun chatting with Romer about
how my theories agreed with his. But then I got my comeup-
pance. As part of my Ph.D. work, I had to run lizards on minia-
ture treadmills inside micro-environmental chambers to measure
just how hard they had to breathe to run at different speeds. (Hot,
boring work for me and the lizards—each run was thirty minutes
and I needed twenty runs per lizard.) When the results came tick-
ing out of the oxygen analyzer, I was devastated—and my theory
was totally deflated. My sprawling lizards were more efficient than
fully erect mammals and birds. All the lizards used less energy to
run at any given speed than did birds or mammals of the same
size. As the old laboratory saying goes "The theorist proposes,
Nature disposes."
I trotted into Romer's offi
ce the next day and sadly an-
nounced, "Our theory is dead." Then I plopped the computer
printout on his desk. Romer scrutinized it. Then with a twinkle in
his eye and a mock inquisitorial tone in his voice he said, "Your
data are probably correct. But they must be suppressed. Our
beautiful theory has got to be preserved." I felt better. If Romer
could chuckle, so could I.
So what advantage is the fully erect gait? Probably it allows
for much higher speeds even if efficiency is sacrificed. Having a
Correct stance. Here's the proper
reconstruction of a horned dinosaur
(genus Chasmosaurus) made to fit
the fossil trackways.
THE TEUTONIC DIPLODOCUS: A LESSON IN GAIT AND CARRIAGE
213
vertical limb stroke means that you can exert more of a thrust
downward onto the ground with your paws. And the speediest gaits
require such thrust to propel the body when all feet are airborne.
When I finally arrived at Harvard in 1972, I was still inter-
ested in the gait of dinosaurs. All the anatomical footprint evi-
dence vindicated Marsh's light-footed and lively postural restorations
of the 1890s. The forelimbs of dinosaurs were aligned quite per-
fectly to match with the stride of the hind limbs. I now asked
myself, "How fast might the big dinosaurs have been?" Most
twentieth-century paleontologists had been willing to concede lively
locomotion to the small, long-legged ostrich dinosaurs and to the
smaller predators, but the big two-ton-plus species were always
reconstructed as slow shufflers. But large mammals can gallop.
While in South Africa I observed three-ton white rhino bulls at a
full gallop with all four huge feet off the ground simultaneously in
mid-stride. In fact, rhinos can accelerate and turn faster than horses,
though in the stretch a horse can outdistance the short-winded
rhinos. Perhaps big quadrupedal dinosaurs could also quick-start
off into their own clomping high-speed charge.
A useful piece of evidence about the speed of dinosaurs can
be extracted from the angles in their joints. Seen from the side, a
running rhino always exhibits greater flexure at the elbow, knee,
hip, and shoulder than does an elephant. Elephants run straight-
legged, thigh lined up with shank and upper arm with lower arm,
so their legs look rather like mobile Doric columns. Rhinos run
with a more bent-legged stride and are consequently faster than
elephants—top speeds are thirty-five miles per hour for the rhino,
twenty-two for the elephant. The rhino owes its greater velocity
precisely to the bounce it gets from the stretching tendons at its
joints each time its feet plant down. Flexing joints provide more
of this bounce, and all the big mammals that gallop are so jointed.
Elephants can never get all their feet off the ground simultane-
ously, even at top speed, and their fastest gait can best be labeled
a running walk. If we could compare the angles in dinosaur joints
to those in these living mammals, we would have an important clue
to the bounciness of their gait and hence their speed.
Brontosaurus has a reputation for being a relatively slow di-
nosaur, and here orthodoxy is correct—all the brontosaurs had
rather straight elephantine legs that didn't flex very much and must
214 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
Swinging shoulders and bouncing knee joints. Big modern gallopers—like
rhinos—have more flexure at their joints than do elephants. Brontosaurs,
such as Camarasaurus, had little flexure and must have run like elephants.
But horned dinosaurs had much more bend in each joint and must have been
more rhinolike in gait. Both brontosaurs and horned dinosaurs had very long
shoulder blades.
Immense power at the dinosaur calf
and knee. A Ceratosaurus set of hips
and hind legs are shown in running
configuration. The extraordinarily
long upper-hip bone (ilium)
supported a huge knee-opening
muscle that attached to the enlarged
crest on the shin. This crest also
was the attachment site for birdstyle
calf muscles bulging backward and
sideways.
Giant calf muscles of Triceratops as
seen from the front.
Triceratops shin,
front view
have limited them to a running walk. But the bipeds and the
quadrupedal horned dinosaurs display much sharper joint flexures
and probably bounced quite a bit as their thick tendons stretched
out and snapped back with each stride. How strong, then, was the
bouncing stroke of such a limb? Big gallopers today possess strong
knee muscles that attach to the kneecap and shank in such a way
that the knee joint opens and closes under tremendous muscular
power. A bony ridge, the cnemial (silent c here: "nee-mee-al") crest,
marks the point of attachment for the knee tendons, and one can
directly gauge the muscle power of a knee from the size of a cne-
mial crest. Elephants, turtles, and salamanders are all slowpokes in
their body-size classes and all have puny knee muscles and low
cnemial crests on the shank bone. Rhinos have big cnemial crests,
as do other large-bodied gallopers, such as water buffalo, giraffe,
bison, and gaur. Big crests would also mean big calf muscles.
All dinosaurs had bigger cnemial crests than do elephants, even
those groups with relatively straight hind legs—the giant horned
dinosaurs, stegosaurs, and brontosaurs. When these systems of
oversized knee muscles contracted, the power exerted on the hind
THE TEUTONIC DIPLODOCUS: A LESSON IN GAIT AND CARRLAGE I 217
paw would have had no equal today. The biggest meat-eater, three-
ton-plus Tyrannosaurus, had an absolutely huge cnemial crest, even
by dinosaurian standards. At full speed, a bull Tyrannosaurus could
easily have overhauled a galloping white rhino—at speeds above
forty miles per hour, for sure. The consistent pattern of huge cne-
mial crests is documentary evidence of super-powerful knees and
calves that gave fast top speeds in most big dinosaurs.
A quite different approach to the question of dinosaur speed
is provided by calculating the maximum strength of the bone shafts
of the limbs. Legs do break in nature, and evolution usually outfits
a species with bone shafts strong enough to withstand the highest
strains imposed when muscles contract. Rhinos have relatively stout,
thick-shafted legs. Elephants feature a more spindly design. To
measure the shaft strength of dinosaur limbs, I constructed scale
models in clay of the life appearance of various species. I then cal-
culated the live weight by measuring the volume of the model (most
land animals are a little less dense than water, so live weight is
about 95 percent of the body's volume in water). Brontosaurs and
stegosaurs were somewhat thin-thighed, and in cross section their
bones are about as thick as we would expect in an elephant of sim-
ilar size. But Triceratops, Tyrannosaurus, and the other predators
were much more massively shafted, far stronger in girth of bone,
and these dinosaurs could exert positively prodigious force through
their limbs without fear of fracture.
Tyrannosaurus moving at forty-five miles per hour is a hor-
rendously heretical concept, and when I began to publish recon-
structions of galloping dinosaurs, the shrill voice of outraged
orthodoxy rose to deafening heights. The advocates of slow di-
nosaurs had two strong arguments. They pointed out that the di-
nosaurs' joint surfaces usually weren't smooth and polished as are
those in mammals, but were roughened and pitted. Those pits held
cartilages. It was therefore alleged that dinosaurs had too much
gristle in their knees to stand the strain of fast trots and gallops.
But this argument is flawed.
In point of fact, cartilage is excellent biological material for
absorbing shocks—better than dense, brittle bone, because carti-
lage will compress under load, building up hydrostatic pressure in
its fluid-filled micropores and springing back when load is re-
leased. Adult mammals and birds have only a thin film of cartilage
218 | DEFENSE, LOCOMOTION, AND THE CASE FOR WARM-BLOODED DINOSAURS
over their joint surfaces, but their young often possess thicker pads
of cartilage, which fill pits in their bones like the pits found in di-
nosaurs'. And adolescent animals usually display greater locomo-
tor vigor than adults, not less. The pitted limb bones of dinosaurs
would be no handicap to high speeds.
The other argument against galloping concerns the question
of long shanks versus short shanks. Many fast mammals have long
shank bones in comparison to the length of their thighs and even
more elongated ankle bones (called metatarsals). Gazelles and most
other fast-running antelope show bones and shanks that are very
long relative to the thigh. Ostriches are fast runners and also have
long shanks and ankles and short thighs. According to the tradi-
tional theory of shanks, to estimate the top speed of an extinct
creature, one measures the length of shank + ankle and divides by
length of thigh. If the resulting number is over 1.5, the animal is
moderately fast; if over 2, the animal is in the gazelle category.
Very few dinosaurs possessed shanks and ankles as long and thin