EXPANDED VERSION
A N A M AT E U R ’ S PE R S PE C T I V E
Evolution in the Devonian
By John A. Catalani
As you probably know, I dabble mainly in Ordovician rocks
Live Birth in Materpiscis
In the first paper, John Long and colleagues described a new
of the Midwest. I therefore have an intense interest in the
ptyctodontid placoderm – a jawed fish with an armored head
fossils (and you know which ones I’m talking about) and geologic history of that period of the geologic column. However,
– found near Gogo Station in Western Australia in rocks that
are early Frasnian in age (approximately 380 million years
the Devonian Period has always fascinated me because the
old). They named the animal Materpiscis attenboroughi –
pace of vertebrate evolution increased dramatically toward
Materpiscis is Latin for “mother fish” and attenboroughi honthe end of the period resulting in the appearance of a veriors the renowned naturalist Sir David Attenborough (classy
table gaggle of unique and interesting animals. Recently, several papers were published in the British journal Nature that
choice). The reason for the name and for the significance of
the specimen is that in the upper body cavity of the fossilized
caught my attention. The first concerned the earliest record of
adult fish is preserved a partial skeleton, consisting of unbrolive birth in the fossil record and the second described, more
ken fragile bones,
completely
than
that appeared to the
previously, another
authors, based on
“intermediate” in
similar dentition,
the stem-group linto belong to the
eage of tetrapods.
same species as the
In both reports, the
adult. The authors
animals described
concluded that this
lived during the
is a preserved emLate
Devonian,
bryo – evidence for
which is divided
live birth of fullyinto the earlier
formed
juveniles
Frasnian Stage and
in the Paleozoic.
the later Famennian
Because the position
Stage (approximateof the embryo in
ly 385-360 million
the adult was deteryears ago).
mined to have been
As any collecat the normal locator will tell you,
tion of the uterus
a fossil that tells a
and not the stomstory is the most
ach, it is unlikely
interesting, be it a
that the partial skelhorseshoe crab foseton is the remains
silized at the end of
of an ingested meal.
its trail in the mud,
The presence of
a monoplacophoran
various preserved
jammed into the
living chamber of a Model of Materpiscis, a placoderm fish, in Museum Victoria. Illustration courtesy of Sularko
soft tissues, some
interpreted as an
nautiloid, or a Green via Wikimedia Commons.
River fish fossilized
umbilical cord between the adult and the partial skeleton, whereas others are
in the act of eating (or choking on) another fish. And we are
regarded by the authors as representing a possible yolk sac,
all familiar, I’m sure, with those incredible ichthyosaur fossils
further supported their conclusion. Several additional previfrom the Jurassic of Holzmaden, Germany, some displaying
ously-collected specimens of fish from the same locality were
a carbonized outline of the body and at least one capturing
found, under closer examination, to also contain embryos.
the process of live birth, tail first, of a juvenile. Well, a fossil
These embryos were originally interpreted as simple clusters
has recently been uncovered that pushes back live birth in
of scales, presumably the remains of a last meal. The relative
vertebrates by some 200 million years.
AMERICAN PALEONTOLOGIST 16(4) Winter 2008 online supplement
position of the scales in the body of the adult fish, however,
These intermediate forms, most found and/or described
is the same as the embryo in Materpiscis indicating that these
in the last 20-or-so years, although not always in a direct
are, most likely, also fossilized embryos. At the end of the abline to fully-developed, land-based tetrapods, displayed the
stract, the authors concluded that it was the placoderms that
changes in body form, eye position, respiratory structures,
evolved, for the first time in vertebrates, an internal fertilizaauditory receptors, and, of course, limbs required to gradution/live birth form of reproduction.
ally transform fish-like animals into tetrapod-like animals.
As mentioned above, the tempo of vertebrate evoluThese fossils have provided us with a virtual “time-lapse” view
tion was significantly accelerated beginning around the late
of how evolution works: adaptations gradually accumulated
Middle Devonian. During the rest of the Devonian, fish beover millions of years through a series of intermediate forms
gan to change by experimenting with new adaptations and
producing a new group of animals adapted to take advantage
exploiting new habitats. The result was the initial appearance
of a previously unexploited environment. The specific adapand diversification of tetrapods, the group to which all of
tations illustrated by these fossil specimens were pretty much
us land vertebrates belong. It was assumed that during this
what we expected should occur in the stem-group lineage as
time period, the water-to-land transition took place but the
fish evolved into animals with four limbs that ended in digfossil record has shown, so far anyway, that true five-toed,
its instead of ray fins. Whether or not these individual anifully-terrestrial
tetrapods
mals were on the direct line
did not appear until the folto true amphibians, we are
lowing geologic period, the
fortunate that such a lovely
Carboniferous. However, that
transitional “sequence” (addoes not diminish the signifimittedly a somewhat simpliscance of the fish-to-tetrapod
tic view but adequate for this
transition which was achieved
essay) of stem-group memduring the Devonian. Now, I
bers is available to illustrate
am by no means as familiar
the concept of transitional
with vertebrates as I am with
forms sometimes referred to
nautiloids but I have always
as “missing links,” a concept
been curious about this evothat is often misunderstood.
lutionary event (we amateurs
When evaluating such transiare an inquisitive lot). Those
tions, researchers are primarof you that belong to my genily concerned with how each
eration can probably rememof the intermediates relates
ber the traditional explanation
to the forms that both prethat we were told in school:
ceded and succeeded them in
air-gulping lobe-finned fish,
the sequence so that adaptasuch as Eusthenopteron, with
tions can be tracked through
bones articulated in their pectime from inception to full
toral fins crawled from pond
development as well as idenMuseum of the Earth’s own Acanthostega model takes a (admittedly
to pond during droughts in inaccurate) stroll in the Gorge Garden.
tifying which structures are
the Devonian thus strengthplesiomorphic (primitive),
ening their developing “legs.” (The extant coelacanth, often
which are apomorphic (derived), and which are autapomorreferred to as a “living-fossil,” is a lobe-finned fish.) The rephic (unique). In fact, I used this evolutionary event and sevsult was a tetrapod, such as Ichthyostega discovered in eastern
eral of the intermediate taxa previously to refute the opinion
Greenland, which was assumed to have been fully adapted to
held by some that transitional forms do not exist.
life on land. When I began teaching, I conveyed that same
Ventastega
story to my students during our evolution unit and showed
In the second paper, Per Ahlberg and colleagues re-described,
them the film “This Land” (narrated by the incomparable
using new specimens, one of these transitional forms,
Peter Thomas) that presented viewers with a picture of a fosVentastega curonica, that had been originally described in
sil of Eusthenopteron and, using models, illustrated the water1994 from fragmentary fossils found in rocks of the Ketleri
to-land transition. Problem was, transitional forms beyond
Formation (late Famennian) from western Latvia. The new
Eusthenopteron were lacking and Ichthyostega was viewed at
material, collected from the original site, allowed the authors
the time as a more-or-less fully-functional, land-living tetto not only more fully describe the critter but also re-evalurapod. Fortunately, continued aggressive collecting resulted
ate its importance and position in the tetrapod stem-group
in the discovery of a number of intermediate forms collected
sequence. The new fossils consisted of additional skull elefrom many locations around the globe.
online supplement AMERICAN PALEONTOLOGIST 16(4) Winter 2008
ments that were previously missing as well as various clavicle
and pelvic bones. Although structures in the skull, such as
a shovel-shaped snout and large eye holes positioned at the
top of the skull, gave it an early-tetrapod shape, morphometric analysis showed that the proportions of the skull were
more fish-like. Another interesting feature in the skull was
the large size, compared to other stem-group members, of the
spiracular opening, which was probably accompanied by a
greater volume of the spiracular chamber – features that were
at first used to facilitate air-breathing and were eventually instrumental in the development of the vertebrate middle ear.
This trend, enlargement of these structures through time, is
interpreted as evidence for a developing dependence on air
for respiration in the stem-group tetrapods. The authors also
concluded that changes in skull morphology were gradual
during the transition from fish to tetrapod. The lower jaw
had a tetrapod-like structure but retained fish-like features,
such as coronoid fangs, that were lost in limbed tetrapods
located more crownward from Ventastega. The morphology
of clavicle and pelvic structures were also tetrapod-like and
the authors conjectured that Ventastega probably possessed
limbs with digits and would, therefore, be classified as a true
tetrapod. As to the position of Ventastega in the sequence of
intermediate forms, the authors stated that the suite of characters possessed by Ventastega are what one would expect in a
transitional animal that is positioned between Tiktaalik and
Acanthostega. They cautioned, however, that detailed analysis
of these intermediates, both those known and those yet to be
discovered, could reveal that they possessed morphological
characters different from more crownward positioned taxa
without these features having been uniquely derived. Such
a scenario has the potential to significantly modify our assessment of the stem-group lineage and the position of taxa
on that sequence. At the very least, the authors asserted, evidence suggests that early tetrapods possessed substantial morphological diversity.
A very brief description of several of these intermediate forms will serve to illustrate some of the morphological
changes that have been observed in the fossilized specimens
during the fish-to-tetrapod transition. Besides Eusthenopteron,
there are several fish that have been described that display
increasingly tetrapod-like structures. In another paper, John
Long and colleagues reported on Gogonasus, a marine fish
collected in rocks of the same formation and location as
Materpiscis, which possessed slightly more derived features
than Eusthenopteron. For example, some of the pectoral fin
bones and their orientation, as well as the enlarged spiracular
opening in the skull, were similar to those exhibited by the
crownward-located fish Tiktaalik.
A Fish with Fingers
Another stem-group member was the fish Panderichthys
found in rocks of the Lode Formation of Latvia, which is interpreted as being either late Givetian or early Frasnian in age.
Panderichthys, a Devonian fish whose pectoral fins were recently shown to have the precursors of digits. Illustration by Dmitry Bogdanov, via Wikimedia Commons.
AMERICAN PALEONTOLOGIST 16(4) Winter 2008 3
Panderichthys does occur, however, in rocks definitely known
to be early Frasnian in age located in other Baltic States and
Russia. Panderichthys presented researchers with an interesting
combination of primitive (pelvic girdle) and derived (pelvic
fin) structures. As reported by Catherine Boisvert, the pelvic
fin of Panderichthys was more primitive and smaller than its
pectoral fin indicating that it was the pectoral fins that first
began to change from a fin to a true limb. Locomotion along
the shallow-water substrate or even on land was feasible for
Panderichthys and would have been facilitated mainly by the
pectoral fins, similar to the method of locomotion on land
undertaken by the extant “walking catfish,” in contrast to
true land-tetrapods, which rely primarily on the hindlimbs.
In another study, Martin Brazeau and Per Ahlberg examined
a well-preserved skull of Panderichthys, which had a large spiracular opening and an expanded spiracular chamber, and
concluded that middle-ear structures were evolved for and
were first used in the respiration process.
As I was writing this essay, a new study of the pectoral appendage of Panderichthys was published online by Catherine
Boisvert and colleagues. A CT scan of a well-preserved pectoral fin revealed the presence of distal radials that the authors concluded were akin to digits. According to the authors, these structures strengthen the argument that digits (at
least fingers) evolved from pre-existing structures possessed
by lobe-finned fish and were not just a tetrapod “invention.”
The authors also made the case that these structures were
more limb-like than those of the more crownward-placed
Tiktaalik. If so, this would call into question the position
of either Panderichthys or Tiktaalik or both along the stemgroup lineage.
Tiktaalik Leaves the Water
Possibly the most celebrated stem-group individual found
recently is Tiktaalik roseae first reported in a Nature article
by Edward Daeschler and colleagues one of which (Neil
Shubin) is a professor at the University of Chicago. Collected
in rocks of the Fram Formation (early Frasnian) of Ellesmere
Island, Tiktaalik possessed several significant tetrapod-like
derived features. These features included a flattened skull and
elongated snout suggestive of a shift from suction feeding
to a carnivorous lifestyle, eyes located on the dorsum (top)
of the skull, wider spiracular opening and expanded spiracular chamber (similar to that of Ventastega) indicating more
advanced respiratory abilities, and pectoral fins with articulated bones more advanced than in Eusthenopteron but still
retaining fish-type fin rays. Additionally, the loss of extrascapular bones meant that the shoulder of Tiktaalik was not
Life restoration of Tiktaalik roseae, a transitional fossil between sarcopterygian fishes and tetrapods from the Late Devonian Period of North
America. Illustration by Zina Deretsky, courtesy of the National Science Foundation, via Wikimedia Commons.
online supplement AMERICAN PALEONTOLOGIST 16(4) Winter 2008
firmly attached to the skull – in other words, Tiktaalik was
the first vertebrate that enjoyed the advantages of a flexible
neck. Along with these derived features, Tiktaalik also possessed primitive features such as a lower jaw similar to that
of Panderichthys. The authors have placed Tiktaalik in the
“morphological gap” between Panderichthys and Acanthostega
and stated that the discovery of transitional animals, such as
Tiktaalik, tends to progressively blur the fish/tetrapod line,
making a clear distinction between them virtually impossible to determine. The same can be said, we are discovering,
about the “line” between aquatic and fully terrestrial tetrapods. Further analysis of the pectoral appendage of Tiktaalik
by Neil Shubin and colleagues revealed a set of derived tetrapod-like structures that placed the appendage between a
limb and a fin in terms of its functional morphology. The
flexible neck, decoupled shoulder, and more advanced pectoral fin endowed Tiktaalik with a greater range in motion of
the shoulder and wrist joints. This increased flexibility would
have facilitated not only complex locomotion along the substrate but also a more upright stance to lift and support the
body above the substrate as well as to lift its head above water
to gulp air. It might even have been possible for Tiktaalik to
make brief forays onto land to, say, avoid predators. Thus, the
discovery of Tiktaalik has provided us not only with an additional transitional form along the stem-group sequence but
also with significant insights into the evolution of tetrapodlike limbs and mobility.
The First True Tetrapod
The earliest recognized true tetrapod is Elginerpeton from the
late Frasnian of Scotland, although the recovered material
is fragmentary with some elements only tentatively identified as belonging to this taxon. According to Per Ahlberg,
the hindlimbs of Elginerpeton were more paddle-like than
leg-like, beginning a trend seen in the more crownward positioned members of the sequence, which points to a predominantly aquatic form of life. The skull was elongated,
narrow, and triangular with a pointed snout. The lower jaw
was primitive and retained fish-like coronoid fangs.
It is at this point (a node in paleo-speak) in the stem lineage that Per Ahlberg and colleagues positioned Ventastega.
The two succeeding crownward members of the stem lineage are Acanthostega and Ichthyostega both from the late
Famennian of East Greenland and known from extensive,
well-preserved material. As Jennifer Clack pointed out, these
two taxa had much in common such as backward-directed
hindlimbs shaped more like paddles for swimming than
legs for walking, fish-like tails that facilitated movement in
water, and multi-digit limbs – eight digits on the limbs of
Acanthostega and seven digits on the limbs of Ichthyostega.
They were contemporaries in space and time but inhabited
different environments – Acanthostega was almost exclusively aquatic whereas Ichthyostega was at least partially terrestrial. Several structural limitations confined Acanthostega
to the water, including internal gills, wrist and ankle joints
incapable of bearing weight on land, and primitive pectoral
structures. However, the forelimbs of Acanthostega probably
both aided the animal in its movement on stream/lake bottoms and assisted it in lifting its head out of stagnant, poorlyoxygenated water to gulp air. Analysis of the limb structures
of Acanthostega forced researchers to reject the previously
accepted scenario for limb development – lobe-finned fish
crawling onto land, thus strengthening fins that eventually
evolved into legs – and to replace it with a new sequence of
events in which the evolution of limbs with digits occurred
while these animals were still essentially confined to a water
environment. The developing limbs allowed the early tetrapods to move more efficiently on the lake or stream bottom
and, once more-or-less fully developed, were finally adapted
for movement on land. Ichthyostega had a more developed
pectoral region suggesting that the animal was better adapted
to locomotion on land than other stem-group members although its rigid spinal column limited its flexibility to vertical rather than horizontal, as in Acanthostega, body motion
which would have restricted its movement on land. However,
it also possessed specialized ear structures adapted for use under water, indicating that Ichthyostega was also, possibly primarily, aquatic.
Finally, Losing the Gills
Finally, there is Tulerpeton, known from well-preserved but
fragmentary remains, mainly the shoulder, pectoral and
pelvic limbs, and some skull elements. Tulerpeton fossils
were collected in rocks of the late Famennian from Russia.
In a review of the fish/tetrapod transition, John Long and
Malcolm Gordon indicated that the limbs of Tulerpeton bore
six digits each and its pectoral girdle was more robust than
that of other Devonian tetrapods but, as with Acanthostega
and Ichthyostega, the large hindlimbs were more paddle-like
than foot-like. They also state that, although mainly aquatic,
Tulerpeton might have relied more on air for breathing than
water because gill-supporting structures were absent and
therefore it probably lacked internal gills.
These and other intermediates are providing us with information on both the specific adaptations required for constructing a true tetrapod and the timing of the appearance
of these adaptations in the stem-group lineage. This, then, is
the developing story of how fish evolved into early tetrapods;
tetrapods that eventually left the water for a life on land and
diversified into all subsequent terrestrial vertebrates – a dynamic story of evolution in action. Analysis of the derived
features possessed by these intermediates and assessment of
the environments each inhabited reinforces the view that
most tetrapod-like structures evolved in the water and not on
land as had been previously assumed. And, as I mentioned
in that earlier essay, aggressive collecting and state-of-the-art
analysis of an ever increasing diversity of fossil specimens has
forced us to altered our concept of evolutionary transformaAMERICAN PALEONTOLOGIST 16(4) Winter 2008 online supplement
tions from a traditional view of “missing links” as individual
species or forms that link two well-known animal groups to
a more realistic and holistic view that involves a suite of transitional forms – members of a continuum that delineate an
evolutionary trend often resulting in a lifestyle change such
as the water-to-land transition.
Further Reading
Ahlberg, P. E. 1995. Elginerpeton pancheni and the earliest
tetrapod clade. Nature, 373: 420-425.
Ahlberg, P. E., J. A. Clack, E. Lukševičs, H. Blom, & I.
Zupinš. 2008. Ventastega curonica and the origin of tetrapod morphology. Nature, 453: 1199-1204.
Boisvert, C. A. 2005. The pelvic fin and girdle of Panderichthys
and the origin of tetrapod locomotion. Nature, 438:
1145-1147.
Boisvert, C. A., E. Mark-Kurik, & P. E. Ahlberg. 2008. The
pectoral fin of Panderichthys and the origin of digits.
Nature, published online 21 September 2008.
Brazeau, M. D., & P. E. Ahlberg. 2006. Tetrapod-like middle
ear architecture in a Devonian fish. Nature, 439: 318321.
Catalani, J. 2006. Intelligent? Design, Part II. American
Paleontologist, 14(2): 35-37.
online supplement AMERICAN PALEONTOLOGIST 16(4) Winter 2008
Clack, J. A. 2006. The emergence of early tetrapods.
Palaeogeography, Palaeoclimatology, Palaeoecology, 232:
167-189.
Daeschler, E. B., N. H. Shubin, & F. A. Jenkins, Jr. 2006.
A Devonian tetrapod-like fish and the evolution of the
tetrapod body plan. Nature, 440: 757-763.
Long, J. A., & M. S. Gordon. 2004. The greatest step in
vertebrate history: a paleobiological review of the fishtetrapod transition. Physiological and Biochemical Zoology,
77: 700-719.
Long, J. A., K. Trinajstic, G. C. Young, & T. Senden. 2008.
Live birth in the Devonian Period. Nature, 453: 650652.
Long, J. A., G. C. Young, T. J. Senden, & E. M. G. Fitzgerald.
2006. An exceptional Devonian fish from Australia sheds
light on tetrapod origins. Nature, 444: 199-202.
Shubin, N. H., E. B. Daeschler, & F. A. Jenkins, Jr. 2006.
The pectoral fin of Tiktaalik roseae and the origin of the
tetrapod limb. Nature, 440: 764-771.
John Catalani is retired from teaching science at South Hill
High School in Downers Grove, Illinois. His column is a regular
feature of American Paleontologist. Email fossilnautiloid@
aol.com.