South Africa studying the ancient mammal-like reptiles of the
Permian Period, a colleague from Johannesburg Museum took me
for a weekend outing to one of their famous parks. There were
camels—though the species isn't native to South Africa. They had
been imported for use in crossing desert regions and are popular
exotic displays in the outdoor parks. Camels have thick cushiony
pads under their toes, and these pads spread out under their own-
THE CASE OF THE DUCKBILL'S HAND | 157
er's foot as they walk. In the course of our visit I happened upon
a dead camel in an unkempt corner of the park. It lay like a desic-
cated mummy, all its natural juices evaporated by the hot Trans-
vaal sun. The camel's mummification is not uncommon in such dry
climates. Beneath its outstretched feet, its cushions, plump and
elastic in life, were now dried-out bags of skin, which had flat-
tened against the dusty soil surface. A spark of recognition shot
through my brain. If camels were extinct and this carcass were found
covered by flood-borne sand, wouldn't paleontologists conclude that
the camel had webbed toes? The flattened skin of its paws created
the perfect imitation of a web.
The skin of the duckbill's paws was not marked by calluses
the way camels' paw skin is. But the way the duckbill mummies
are preserved permits the hypothesis that in life those flattened
hands were in reality plump, rounded cushions of connective tis-
sue—elastic shock absorbers for the impact of the ground on the
wrist when the animal moved fast over hard ground. Duckbill fore-
paws were so narrow and compact that a paw cushion would do
invaluable service by lessening the load of impact within the joints
of the toes. Fossil duckbill trackways, just now being excavated in
158 | THE HABITAT OF THE DINOSAURS
Foreclaw cushions,
inside view
Canada, suggest that indeed this line of reasoning may be correct.
The forepaw impressions resemble smooth crescents, as though the
individual toes were all imbedded within a single, insulating mit-
ten. There is definitely no sign of a spreading ducklike web.
This may be the true solution to the century-old mystery of
the mummy's hand. That brown withered paw may have misled
four generations of paleontologists into believing in a series of
nonexistent adaptations for swimming. The mummy's hand, when
alive and full of healthy tissue, may have worn a shock-absorbing
glove, an earth-mitten entirely designed for walking on dry ground.
THE CASE OF THE DUCKBILL'S HAND
159
8
DINOSAURS AT TABLE
Orthodox paleontologists insist most of their dinosaurs ate
mush. They condemn both of the great tribes of plant-eat-
ers—the brontosaurs and the duckbills—to a way of life at the
water's edge, forced to eat nothing but soft water plants. In its own
way, this theory epitomizes the traditional view of most dinosaurs
as swampland creatures, virtual dead ends in evolution's race to
develop lively, active species. In 1915, William Diller Matthew, a
very respected mammal paleontologist, wrote, a highly influential
book, Climate and Evolution, which argues that evolution bogs down
in the soggy lowlands. Matthew believed that only on the high,
dry soil of plains and plateaus did evolutionary forces create the
most vigorous, most advanced creatures. There's a lot of truth in
Matthew's thesis. It has been ascertained, for example, that water-
loving turtles and crocodiles evolve most slowly, changing so little
on average through geological time that a single genus can be fol-
lowed for thirty million years or more. So the orthodox concept
of a mush diet is consistent with the overall theory of sluggish di-
nosaurs: soft, plant food was all they required for their sluggish
metabolic needs, and the consequent swampy habitats limited di-
nosaurs to slow rates of evolution.
There may be some ground for believing the brontosaurs ate
such soft foods. If the possibility of gizzard stones is ignored, the
160 | THE HABITAT OF THE DINOSAURS
The duckbill Kritosaurus:
life portrait and skull
of tougher plants. But there are no grounds whatsoever for be-
lieving it of duckbills. The mouth of a duckbill dinosaur contained
one of the most efficient cranial Cuisinarts in land-vertebrate his-
tory. Duckbill teeth and jaws were incomparable grinders, de-
signed to cope with foods right inside the duckbill's oral
compartment.
The myth of mushy foods for duckbills began with a single
error by one of the great pioneering American dinosaur hunters.
Edward Drinker Cope discovered a fragmentary duckbill jaw in
1885. His specimen had cracks running through the row of teeth,
so that individual teeth fell out of the fossil jaw when he exam-
ined it. Cope mistakenly assumed this condition was natural and
DINOSAURS AT TABLE I 161
jumped to the conclusion that a duckbill's teeth would break off
whenever the beast tried to chew tough food. This error should
have been corrected by 1895, when complete skulls and jaws re-
vealed that duckbill teeth were firmly packed together and no one
tooth could possibly fall out before it was totally worn down. Even
then, whenever a worn tooth dropped out, a new tooth already
stood beneath it ready to take over chewing duties. Duckbills ap-
parently never ran out of teeth. No one has ever discovered a se-
nile duckbill mouth; not one specimen exists with all its teeth either
The head of Edmontosaurus,
a duckbill. Life portrait at
top, skull in the center, and
skull cut through the tooth
rows at the bottom.
162 | THE HABITAT OF THE DINOSAURS
worn out or fallen out. To all appearances, from the day they
hatched out of the egg to their last breath, the duckbills enjoyed
the use of healthy dental machinery, continually renewed by young
teeth growing in to replace the old.
Not only were the duckbills' teeth never-ending, their ar-
rangement was designed especially for powerful grinding. At any
one moment many rows of young teeth were growing into the
mouth, providing the animal with grinding surfaces made up of
hundreds of closely packed teeth. Each tooth was built up from
two different biological materials: a thick layer of very hard enamel
and a central core of softer dentine. Since many rows of teeth were
packed together in each jaw, and all the rows together partici-
pated in chewing action, the chewing surface was a mosaic of enamel
ridges and dentine. Enamel ridges always protruded a little higher
than the dentine cores, because the enameled parts of the teeth
got worn down a bit more slowly than the softer cores. This ar-
rangement was very effective. No matter how hard the duckbill
How duckbill teeth work
DINOSAURS AT TABLE I 163
chewed or how hard its food was, the
enamel stuck up further than
the dentine, young teeth kept replacing the old, and the duckbill
maintained a grinding surface that worked much like a self-sharp-
ening vegetable grater.
Although Professor Marsh of Yale clearly illustrated the real
qualities of the duckbills' chewing equipment in 1896, most pa-
leontologists retained the mistaken theory and ignored the ob-
vious adaptations for tough food. It required yet another Yale
professor to set matters straight. In 1961, John Ostrom published
his heretical interpretation of duckbills. He defined them as land
creatures and emphasized the mechanical—ecological implications
of their dental Cuisinart. He pointed out that the teeth of duck-
bills had a pattern that virtually necessitated tough food. Their
characteristic bills were also consistent with a tough-food diet, de-
spite a superficial resemblance to the bill of modern water-feeding
ducks. Way back in the 1880s, Cope had already found fossil rem-
nants of the horny edge that had lined the bony beak of duckbill
dinosaurs while alive. This horny edge was sharp and deep from
top to bottom, more like the edge of a cookie-cutter than the soft,
sensitive rim of a mud-dabbling duck. After Cope's initial discov-
ery, other horny fossils turned up, making it clear that all duck-
bills possessed deep, sharp-cutting edges along the entire upper
and lower beak. Such sharp edges were obviously for cropping
tough plants—not for grazing on mush. So soft-beaked ducks were
never good analogues for duckbill dinosaurs, but modern tortoises
are; the tortoise's beak is tall and sharp-edged, and constantly used
to cut through tough blades and stems.
If duckbill dinosaurs were truly efficient shredders of tough
fodder, they would also have required good tongue—cheek coor-
dination. Consider what it takes to chew something as recalcitrant
as a piece of celery—your tongue contributes by moving the fi-
brous lump between palate and teeth. Your cheeks play their role
by retaining the mass of celery and preventing it from slipping.
Tongue-in-cheek skill is characteristic of the best shredders among
today's Mammalia—horses, cows, elephants, rabbits, kangaroos. All
these herbivores possess large, active tongues and strongly mus-
cled cheeks. Incidentally, that lump of food while being chewed
in the mouth has been dignified with a technical scientific label:
"bolus."
164 I THE HABITAT OF THE DINOSAURS
All of today's Reptilia are cheekless. Their open mouthline
extends all the way back to the joint of the jaw before the ears.
There is no skin to hold any food being chewed. Consequently,
herbivorous reptiles—tortoises and iguana lizards, for example—
are sloppy eaters; when their jaws slice off a piece of leaf, the part
sticking out of the mouth simply falls to the ground. Each time
they chew, they lose nearly half their mouthful, quite a wasteful
business. Primitive meat-eating dinosaurs had similar wide-open
mouthlines.
Traditionally, duckbill dinosaurs have been portrayed as
cheekless, with the mouthline running from chin to ear like a liz-
ard's. A dissenting voice was raised by Yale Professor Richard
Swann Lull (Yale's tradition of duckbills seems to have been con-
sistently heterodox). In 1942, Lull restored duckbills with cow-style
cheeks walling the sides of the oral space. But most of Lull's col-
leagues rejected the idea because everyone knew dinosaurs were
reptiles, and reptiles, by definition, didn't have cheeks. Such ob-
jections were specious. No living reptile has cheeks. But no living
reptile has grinding teeth anything remotely resembling those of a
duckbill. If the duckbills could have evolved such unreptilian teeth,
why couldn't they have evolved unreptilian cheeks?
The final Yale duckbill—cheek conclusion was joined in the
late 1960s. Peter Galton, an English paleontologist resident at Yale
as a research associate, reinvestigated the question of the duck-
bill's oral tissue. He concluded that Lull's reconstruction of cheeky
duckbills was almost certainly correct. All duckbills had deep re-
cesses in their skull and jaw bones running parallel with their
mouthlines above and below where their teeth came together. This
recessed zone resembled the deep hollowed-out areas found in the
jaws of gophers, chipmunks, and other rodents which have capa-
cious cheeks for holding food while they chew. A slightly rough-
ened ridge often marks the top and bottom of the duckbill recess,
and some sort of skin or muscle or both must have attached to it.
Peter Galton drew diagrams of the cheek—pouch recesses in mod-
ern species such as pigs, horses, elephants, and rodents, which
demonstrated how duckbill pouches must have been as well de-
veloped as any of these.
What, then, did duckbills eat? Considering their prodigious
dental powers, the flip answer might be "anything they wanted."
DINOSAURS AT TABLE I 165
But in terms of serious theory, those powers expand the boundary
conditions of any hypothesis about their diet very widely. The
duckbills might have masticated extremely tough leaves, stems,
twigs, pinecones, even roots and tubers. Some clues as to their ac-
tual dining habits can be gleaned from their very curious body
posture. All duckbills had much longer and stronger rear than fore-
legs and probably moved semibipedally, striding on their hind legs
and using their forepaws only to touch down lightly for balance.
Old restorations showed duckbills standing in a tripodal posture,
their hind legs and tail supporting their weight, with their back
and neck nearly vertical. Such a posture would have permitted the
duckbills to feed high in the pine trees of their habitat. Yet that
upright body posture was wrongly conceived. The build of the
duckbill was clearly designed for low, near-the-ground feeding, not
for tree-browsing. If duckbills had specialized in high-level feed-
ing, they would have had shoulders and necks designed for reach-
ing upward. But that is not the case. Instead, in the region of the
shoulder their backbone bends permanently downward. This sharp
flexure locates the base of the neck and the head on a very low
anatomical level. The downward bend in the chest area is so marked
that even when a duckbill raised its neck as far as it could go, the
head was still below the level of the topmost point of the shoulder.
Some mammals today exhibit this same downward curve of
the backbone. In the American buffalo, for example, the line of
the vertebral column curves sharply downward as it passes from
shoulder to neck. Thus they must always hold their heads low, with
muzzles close to the ground. As the song says, buffalo roam where
the deer and the antelope play, but deer and buffalo represent di-
vergent tactics for eating plants. Deer can carry their heads much
higher than can buffalo, and can reach up into the trees to nibble
on twigs, leaves, and bark. Buffalo stick to grou
nd level and use
their strong, wide snouts to pull up tough grass deer cannot deal
with. Clearly, the duckbills were more like buffalo than like deer.
And the entire tribe of duckbills must have spent most of their
time feeding at or near the ground.
These considerations dramatically narrow the boundary con-
ditions for any hypothesis about their diet. The duckbills' pre-
ferred food must have been low-growing herbs or shrubs (grasses
had not yet evolved in Cretaceous times). These boundaries still
166 I THE HABITAT OF THE DINOSAURS
Permanent downflex in the American buffalo and a duckbill (dashed line
shows line of backbone)
allow for a wide selection of Mesozoic roughage and greens:
horsetails, ground pine, ferns, low tree ferns, seedling evergreens
(pines, cypress, etc.), cycads and other tough-frond types, low-
growing palms, magnolialike shrubs, and so on.
It's probably barking up the wrong herb to try to find the one
duckbill food. Duckbills were so varied in snout design that it's
unlikely all species fed on the same plant stuffs. Today the ante-
lope family demonstrates how snouts can be custom-tailored to fit
each species' method of feeding. Cape buffalo (cows and buffalo
are members of the antelope family) have very wide muzzles, fine
for biting a wide swath through the sward but much too clumsy
for picking out individual succulent tidbits. Royal antelope have
slender snouts which they can use to pick and choose. Among the
duckbills, Edmontosaurus had a huge, blunt muzzle and must have
cropped wide batches of leaves with each bite. Duckbills with hol-
low head crests, Lambeosaurus and its kin, adopted a totally differ-
ent approach; their muzzles were narrow, and allowed them to poke
around for a more discriminating bite.
Everywhere on the Late Cretaceous deltas the duckbills' con-
stant companions were the great horned dinosaurs. Side by side,
three-horned Triceratops and wide-mouthed Edmontosaurus cropped
the greenery. Duckbills and horned dinosaurs were distant cous-
DINOSAURS AT TABLE I 167
ins—both had beaks and traced their ancestry to the same ancient
little dinosaur of the Triassic Period. But what an extraordinary