cellent example. Early nineteenth-century anatomists believed thick-
skinned mammals were all somehow related. Consequently, horses,
tapirs, rhinoceroses, elephants, and hippos were all formally iden-
tified as the Order Pachydermata. But in the 1880s, an avalanche
of fossil data poured in demolishing the notion that "pachyderms"
constituted a natural evolutionary unit. Rhinos, tapirs, and horses
are indeed closely related; they stem from an ancestor much like
446 | DYNASTIC FRAILTY AND THE PULSES OF ANIMAL HISTORY
the little Dawn Horse, Eohippus. But that group is not at all close
to hippos, which trace back to a pig ancestor, and neither the horses
nor the hippos are close to elephants. The case against the Pachy-
dermata was overwhelming. And the name was stricken from the
list of acceptable terms. The "pachyderms" were an unnatural
mixture of three distinct pedigrees, sharing only a few unimpor-
tant resemblances in their skin and digestive systems, resem-
blances clearly resulting from independent evolutionary events.
When I was in college in the 1960s, it was commonly ac-
cepted that the case against the Dinosauria was equally strong—
the dinosaurs were in fact two entirely separate groups, each trac-
ing its ancestry to a different Triassic reptile (both ancestors were
pseudosuchian reptiles of the Triassic). The implication was that
each group of dinosaurs was no more closely related to the others
than they all were to crocodiles or to birds. The public might have
great affection for the Dinosauria, but the label was without sci-
entific foundation. This was the accepted wisdom, but the evi-
dence for it was not much discussed. It was, in fact, yet another
piece of orthodoxy, rarely debated, never seriously challenged. And
the notion that the term "Dinosauria" should not even be used
matched nicely with the generally dismissive attitude at large toward
the Mesozoic monsters.
Into the mid-nineteenth century, "Dinosauria" was accepted
as a respectable scientific term. Paleontologists believed there were
three major subgroups: the carnivorous Theropoda; the long-necked
Sauropoda (brontosaurs); and the beaked herbivores, now called
Ornithischia ("bird hips"). Out of these three groups, the Orni-
thischia had the most unusual skull and hip. It therefore became
fashionable to divide the Dinosauria into two groups: the Order
Saurischia for the carnivores and brontosaurs combined, and the
Order Ornithischia. At first, this division did not undermine the
belief that all dinosaurs were one natural group. Paleontologists
wrote as though the Saurischia and the Ornithischia had evolved
from a single, very primitive ancestor, the Ur-dinosaur.
By the 1920s, however, the view had become quite different.
Paleontologists began to embrace the idea that the ancestry of the
Ornithischia had been totally different from that of the Saurischia.
Formal classification listed the Order Saurischia and the Order
Ornithischia as distant cousins, alongside the Order Crocodilia, and
the Order Pterosauria (flying reptiles). All these independent or-
DINOSAURS HAVE CLASS I 447
Example of an unnatural taxonomic act—the "Order Pachydermata." To
qualify as a natural group, a clan of species must trace its ancestry back to
one single common ancestor. In the early nineteenth century, naturalists
lumped all the big, thick-skinned mammals into the Order Pachydermata. If
pachyderms were truly a natural group, then rhinos, hippos, and elephants
must have had a common ancestor. But fossil discoveries in the last half of
the nineteenth century proved that the Pachydermata was hopelessly
unnatural. In fact, rhinos, hippos, and elephants belong to three separate
orders, and each of the three "pachyderms" evolved their big size and thick
skin from separate small, thin-skinned ancestors.
ders were grouped together into the reptile subclass, Archosauria.
Obviously, the Dinosauria was an unnatural group.
The evidence for this change in classification was in fact ex-
tremely scanty, and the idea was never thoroughly thought out or
debated. An altogether rather half-baked idea established itself as
the orthodox view of the relationships among the dinosaurs. Even
at its inception, this belief was logically flawed: Just because the
Order Ornithischia was more specialized in its skull and hip than
the Order Saurischia did not prove that the two orders did not
share a common ancestor. If the defining characteristics of the Or-
nithischia and Saurischia, all well known in 1920, were compared,
more than a dozen were shared by both groups—and none of them
were present in any other "reptile" group. But all of this was ig-
nored in the press to adopt the view that ornithischians and saur-
ischians were two separate groups issuing from two hypothetical
ancestors. The change in conception was part of a general concep-
tual crisis that afflicted the whole field of evolutionary biology at
that time. Paleontologists had somehow arrived at a view of evo-
lution which I call the hub-and-spoke syndrome. Each major evo-
lutionary line supposedly originated in one primitive, unspecialized
stock, an evolutionary hub. Subsequently, all the advanced lines
evolved outward—like separate spokes of a wheel—from that hub.
Under the influence of this conception, paleontologists tended to
invent wholly imaginary groups to serve as the ancestors for their
grand theories. Poorly known fossils, represented by fragmentary
skeletons, were often elevated to the status of "common ancestral
stocks." The crimson crocodiles were treated this way. Some of
those little-known Triassic creatures were lumped together as a
hypothetical archosaurian ancestral stock, called the Pseudosuchia.
As all this theory crystallized into textbook form, the Pseudosu-
chia became firmly established as the ancestral hub from which ra-
diated all the separate archosaurian lines to become saurischians,
ornithischians, crocodiles, pterodactyls, and birds.
Nearly everything written about the Pseudosuchia as the cen-
tral hub for Archosaurian evolution is hypothetical. And owing to
recent discoveries in South America and China, many pseudosu-
chian families are now far better known. Almost without excep-
tion, each family possessed strong distinctive specializations that
disqualified it as a "generalized" ancestor for any more advanced
group. A good illustration here is provided by the ornithosuchids,
DINOSAURS HAVE CLASS | 449
Myth of the hub-and-spokes. Since the 1920s, most textbooks show the
pseudosuchians as the central archosaur hub, with two or three separate
spokes representing the different dinosaur clans. If this view is correct, the
Dinosauria is an unnatural group.
the "bird-crocodiles." Both technical and popular books define them
as the essential hub for all the Archosauria, the perfect common
ancestor. As far as they were known in the 1930s, from poorly
defined skeletons, the ornithosuchids did fit
preconceptions about
what the ancestor of the dinosaurs should look like. Their hind
legs were longer than their forelegs and there was the vague sug-
gestion of the bipedal locomotion common in primitive dinosaurs.
By the 1970s, however, good, clear skeletons from Argentina re-
vealed that the ornithosuchids possessed bizarre specializations—
450 | DYNASTIC FRAILTY AND THE PULSES OF ANIMAL HISTORY
for example, a huge, drooping snout, pinched in the side-to-side
dimension—a development totally unexpected in an ancestor of the
dinosaurs. Ornithosuchids were not dinosaur ancestors or anyone
else's. They spent their entire evolutionary history evolving into
more and more specialized ornithosuchids. They did not evolve
into true crocodiles, pterodactyls, birds, or anything else among
the ranks of the more advanced archosaurian groups.
The tale of the ornithosuchids also illustrates the underlying
theoretical weakness of the hub-and-spoke conception of evolu-
tion. Common ancestral stocks may be posited at the base of an
hypothetical family tree. But the real-life ecosystem is a mean, cruel
place. In order to survive, any group must evolve to keep pace
with the threats from new predators and competitors. It is simply
not possible for a family of species to wait around for millions of
years, twiddling its evolutionary thumbs as it were, until it re-
ceives the signal to evolve into something more important.
The system of nomenclature employed by biological scien-
tists represents their understanding, their organization of the life
they study. The system is clear and hierarchical. In zoology, the
Kingdom embraces the widest group. Next beneath that comes the
Phylum—and the term "Phylum Chordata" embraces all verte-
brate animals. The Phylum is then divided into Classes and Sub-
classes, and each Class or Subclass is divided into Orders. Orders
are then divided into Families, Families into Genera, and finally
Genera into Species. It must be emphasized these are man-made
classifications. And sometimes this system of nomenclature can warp
our perception of the real evolutionary events. For example, the
Archosauria is traditionally ranked as a subclass which is divided
into the Order Saurischia, the Order Ornithischia, the Order Croc-
odilia, the Order Pterosauria, and as a primitive hub, the Order
Thecodontia—reserved for the pseudosuchians and other crim-
son-crocodile groups. It makes for a tidy arrangement, everything
in its proper place, especially if we plot each advanced order aris-
ing independently, as a spoke, from the hub order, the Thecodon-
tia. But this orderliness produces a seductive bias toward
reconstructing the history of evolution in oversimplified ways.
How can the pitfalls of such overly tidy thinking be avoided?
Over the last several decades, scholars of evolution have devel-
oped methods for tracing the probable pattern of branching in the
DINOSAURS HAVE CLASS | 451
families they investigate. All their data indicates that evolution rarely
follows the hub-and-spoke pattern. Evolution rather follows a pat-
tern whereby a family emerges by early branchings, and each early
branch branches again, so that the family tree finally resembles a
tangled blueberry bush, a maze of ever smaller branchlets, bifur-
cating and ramifying in many directions. Is it then possible to make
any sense of such a complicated pattern from the scattered fossils
available for studying the history of the dinosaurs?
The most reliable method employed today consists of the
search for what are called sister groups. Sister groups are quite easily
defined. Suppose, for the moment, that the Ornithischia and Saur-
ischia were sister groups. That would imply that they had de-
scended from a common ancestor that was already specialized in
certain distinctive ways. It would then be proper to expect an evo-
lutionary indicator for this relationship, some feature that had
evolved in the common ancestor and was then passed down to all
the subsequent branches. Now, if orthodoxy were correct and the
two orders of dinosaurs really did evolve independently from a
primitive archosaurian hub, then no common heritages could be
expected. And the only course of action for clarifying the relation-
ships of the dinosaurs is to examine them thoroughly for shared
adaptations that might mark each branching point.
Until 1971, I was myself a firm believer in the hub-and-spoke
theory and the orthodox view of the dinosaurs as a group. The
Dinosauria really consisted of three separate spokes—the Sauro-
poda, Theropoda, and the Ornithischia—since, in my opinion, even
the Order Saurischia was artificial and the brontosaurs (Sauro-
poda) were a totally separate line from the carnivores (Thero-
poda). Each was ranked as an Order, and each supposedly evolved
separately from the pseudosuchians. But my convictions were about
to change.
In 1971, Peter Galton, then a postdoctoral fellow, was en-
gaged in work on the very primitive anchisaurid dinosaurs. In the
process, he made some really important discoveries on how the
dinosaurs' claws worked—he was the first to discover the peculiar
thumb-twists. Primitive meat-eating dinosaurs had huge, curved
claws on their thumbs whose tips pointed inward when the fingers
were flexed upward, but downward when the fingers were flexed
downward—a most unusual arrangement. Galton also found ex-
452 I DYNASTIC FRAILTY AND THE PULSES OF ANIMAL HISTORY
actly the same sort of thumb-twist in one group of primitive plant-
eating dinosaurs, the anchisaurs, which were evolutionary uncles
of the giant brontosaurs.
This twist-thumb was a blow to my belief in the artificiality of
dinosaurs as a group. I had even been lecturing that the meat-eat-
ing dinosaurs had evolved quite independently from the anchi-
Kinship revealed in jaw, chest, and hand. Primitive beaked dinosaurs like
Heterodontosaurus had a double breastbone, twist-thumb claw, and a low jaw
joint just like that of anchisaurs. So anchisaurs must be very close relatives of
the ancestors of all the Ornithischia.
DINOSAURS HAVE CLASS | 453
saurs-brontosaurs. But that twist-thumb was simply not to be found
in any potential ancestor from the Triassic. The twist-thumb is a
good example of what paleontologists call a "high-weight nov-
elty," an adaptation so complex and unusual that it is highly un-
likely it evolved twice in totally separate lines. The twist-thumb
was therefore a very strong argument that the anchisaurs-bronto-
saurs were a very closely related sister group of the carnivores,
and thus, that both groups had descended from one common
ancestor—an original species that had first evolved the thumb.
It went against the grain in 1971 to be convinced that the
theropods and sauropods were in fact a single evolutionary group.
But I was about to embrace a far worse heresy. Orthodoxy main-
tained, as a kind of holy tenet, that the beake
d dinosaurs, the Or-
der Ornithischia, were in no way, shape, or form closely related
to the other dinosaurs, the Order Saurischia. But in 1972, a little
dinosaur from the Connecticut Valley was to blast this belief as
well. The dinosaur in question was Anchisaurus, a five-foot herbi-
vore originally discovered by Othniel Charles Marsh in the
Brownstown quarries near Portland, Massachusetts. Peter Galton
was at work on this skeleton, cleaning off areas still covered with
rock, though the specimen had first been put on exhibition in 1880.
As he removed the very hard sandstone, some quite unexpected
bones came to light. The Anchisaurus % upper hip bone (ilium) dis-
played unmistakable characteristics of the Ornithischia. Unlike other
primitive saurischians, whose ilium was very short from front to
back, Anchisaurus had an extra-long prong that stuck forward, which
in life would have increased the size of the major muscle running
from the ilium to the kneecap. But primitive ornithischians did have
a very similar iliac prong. Was this dinosaur the ancestor of the
ornithischians, or at lease a close cousin of the real ancestor? This
was a very unorthodox question since it would imply that Anchi-
saurus was a missing link, an evolutionary bridge between the sup-
posedly unlinkable Order Saurischia and the Order Ornithischia.
If so, then the orthodoxy of 1970 had it wrong and the view of
1880 had been right: the entire Dinosauria constituted a single
natural group.
Galton and I scrutinized every inch of the anchisaur skeleton.
We constructed a list of the characteristics which might link an-
chisaurs to the Ornithischia, and in a brief time it grew to a for-
454 | DYNASTIC FRAILTY AND THE PULSES OF ANIMAL HISTORY
midable length. Dozens of adaptations linked these dinosaurs and
distinguished them from all other archosaurs:
1. A slender, graceful neck shaped into an easy S-curve.
2. A shoulder structured with no collarbone or at best a very
reduced one.
3. An upper arm bone (humerus) with a long shelf for the
chest muscles.
4. A wrist structured without any prominent prong of bone
on the rear inside (this is called the pisiform bone and is very ev-
ident in most reptiles and mammals, including man).
Robert T Bakker Page 46