Â鶹´«Ã½

The first steps taken by an animal on to land was one of the most
significant events in the history of life on Earth. But the process of
pinpointing the time when footsteps were made and discovering the type of
animal that made them is fraught with problems. For a long time we have
lived happily with the idea that the sequence of colonisation of land from
the sea and rivers paralleled the evolutionary tree of life – first the land
was colonised by plants, then by small herbivorous animals that grazed on
the plants and finally by larger, carnivorous animals that fed on the
herbivores. But a series of recent fossil finds dating from between 450 and
370 million years ago reveals a more complex sequence of events.

One of the first problems in trying to sort out which animals first
colonised land stems from the nature of the fossil record. Moving from a
watery environment, which was very conducive to fossilisation, early animals
arrived on a bare, hostile surface. The land was whipped by winds, and
constantly shifting rivers and sandstorms repeatedly covered, uncovered and
moved dead bodies. The rocky, sandblasted ‘soils’ were low in organic
matter, and thus in cohesion. These were hardly ideal conditions for fossils
to form. But at a few sites, a number of fortuitous circumstances, such as
rapid burial in sediments and quick mineralisation, combined to preserve the
remains of an animal’s body, or the burrows and trails left behind as it
moved about on land. These fossils allow us to glimpse the first attempts at
colonisation of the land by animals.

One important question is whether the land the animals first walked on
supported plant life. The early fossil record of land plants, however, is
almost as indecipherable as that of the first animals. The earliest evidence
comes from spores that Jane Gray, of the University of Oregon, and
colleagues found in 1982 in Libya. These were preserved in rocks from the
Ordovician period that may be up to 470 millions year old, and resemble the
spores of some ‘lower’ plants, such as living bryophytes (mosses and
liverworts). This suggests that some of the earliest land plants were
liverwort-like and able to withstand long periods of drought. The only
earlier evidence of life on land before this is in the form of mats of
bacteria and algae, which have recently come to light in 1200
million-year-old cherts found by Robert Horodyski in Arizona (‘When algal
mats ruled the land’, Â鶹´«Ã½, 2 January 1993).

Some of the earliest evidence for terrestrial animals comes from fossil
soils in Pennsylvania dating from about 450 million years ago during the
late Ordovician. Greg Retallack and Carolyn Feakes of the University of
Oregon found deep, vertical burrows, ranging from 2 to 21 millimetres in
diameter, which they believe were made by soil animals, possibly millipedes.
Other, indirect, evidence for terrestrial animals comes from late Silurian
deposits in Sweden, where Gray and Martha Sherwood-Pike, also of the
University of Oregon, found fossilised faecal pellets dating from 410
million years ago that contain remains of fungi. This implies the existence
of a fungivorous microarthropod such as a mite or millipede that fed on the
decomposing remains of simple land plants: a decomposer rather than a
herbivore.

The simplistic view of the progressive colonisation of land by organisms
from successively higher levels in the food chain must be modified to take
account of such decomposers. These creatures would have played a vital role
in soil production. By breaking down dead plants, they would have added
nitrate and phosphate into the upper layers of the soil, thereby providing
the necessary conditions for colonisation by ‘higher’ plants – the vascular
plants equipped with tissue that allows them to conduct water and other
nutrients taken in through roots to the rest of the plant, and frees them
from the need to live in wet habitats.

The earliest known vascular plants are all from about 420 million years ago
during the late Silurian. These were plants such as Baragwanathia, Salopella
and Hedeia, which consisted largely of a branched stem and, sometimes, small
spine-like ‘leaves’. Perhaps the most primitive-looking of all the plants
was Cooksonia – which looked rather like a stumpy sort of sedge with a
simple stem and no leaves. Doubts about the terrestrial status of Cooksonia
were dispelled only recently by Dianne Edwards and colleagues from the
University of Wales, Cardiff, when they found water-conducting vessels
(tracheids) and stomata in Cooksonia from late Silurian rocks in
Shropshire.

Baragwanathia, which resembles a club moss in form, and Cooksonia are
usually preserved in marine sediments, which suggests that these plants grew
near the shore. The extent to which inland sites were vegetated at this time
is unknown. But we do know that around 400 million years ago there was a
rich flora and fauna – including fairy shrimps, spider- like trigonotarbids,
mites and springtails, and the simple, early vascular plants Rhynia and
Aglaophyton – occupying a well- established hot-spring ecosystem at Rhynie,
a classic early Devonian site in Aberdeenshire that was definitely inland.

So it seems likely that though the land was probably green when the first
animals emerged from the water, it was hardly a place of bountiful food
supply. Fossil evidence indicates little in the way of substantial plant
life that could support herbivores. In which case, why did it take about 100
million years from the first major explosion of life in the seas some 540
million years ago, at the beginning of the Cambrian period, for animals to
pluck up the courage to move on to land?

Pressing problems

They must have faced immense problems: a terrestrial environment is hostile
to organisms adapted to living in water. The transition to a dry terrestrial
habitat, where diurnal temperature changes were much greater than in aquatic
environments, the unpredictable water supply and the necessity for
respiration in air, required profound physiological changes.

One of the most pressing problems for a potential terrestrial animal was
avoiding desiccation. The fossilised traces of the 450 million-year-old
millipede burrows found by Retallack and Feakes in Pennsylvania point to
arthropods being the first group of animals to colonise the land. Perhaps
these early terrestrial colonisers coped with problems of temperature and
desiccation by burrowing in the ‘soil’. Any animal that already possessed
characteristics which suited them to terrestrial conditions in some way
would have been the ones most likely to succeed in the terrestrial lottery.

Another reason for the success of arthropods would lay in their tough
cuticle, which had evolved at the beginning of the Cambrian period some 540
million years ago. In addition to providing defence, this exoskeleton gave
strength to the body for moving and feeding. On land, it helped overcome the
impact of gravity – a major concern to any animal moving from water to land
as anyone who has ever crawled out of the sea after a long swim can testify.
Without strengthened legs, movement on land would have proved difficult
(even now, the limbs of land arthropods are vulnerable to mechanical failure
during times of moulting while the new cuticle is hardening). Equally
important was the ability of arthropods to secrete a waterproof outer layer
on their cuticle to prevent water loss.

Direct evidence for arthropods emerging from water has been uncovered at
Kalbarri in Western Australia, where the Murchison river flows through a
deep gorge and has exposed 420 million-year-old late Silurian sandstones
formed from sands deposited by an ancient river and by the wind. At various
levels in the gorge, the surfaces of red sandstones, often covered by
shallow-water ripple marks, are crossed by a wide range of trackways. The
most spectacular are those which, together with Nigel Trewin of the
University of Aberdeen, we interpret as being made by eurypterids. These
extinct arthropods, related to and resembling scorpions (hence their common
name sea scorpions) grew to 2 metres in length. Their distinctive tracks are
up to 20 centimetres across and in places extend for many metres. The form
and number of the individual footprints clearly points to the tracks being
made by eurypterids. These arthropods possessed six pairs of appendages, but
usually only three pairs functioned as legs for walking, with the hindmost
pair commonly paddle-shaped for swimming. The front three pairs of
appendages were adapted for feeding. In some species they evolved into
pincers and in others into ferocious-looking cage-like structures.

While it is relatively straightforward to ascribe the eurypterid tracks to
their producers, many other tracks at Kalbarri are not so easy to identify.
One set appears to belong to an animal of perhaps 5 to 6 centimetres long
that had 10 or 11 pairs of legs. We believe these tracks were made by a
euthycarcinoid – the only arthropod to have yielded a body fossil so far in
these deposits (see ‘Is Australian fossil the ancestor of all insects?’,
Â鶹´«Ã½, 17 August 1991). It appears that the euthycarcinoids, a group
until recently considered to be entirely aquatic, were amphibious.

These extinct arthropods, resembling a multilegged cockroach, were most
closely related to the centipedes and insects. Their presence in rocks of
this antiquity has also pushed back the age of euthycarcinoids by about 120
million years, and suggests the euthycarcinoids may have, after all, been
the ancestors of hexapods, the group that contains insects. Until recently,
it was thought myriapods (the group that contains centipedes and millipedes)
gave rise to hexapods because the earliest known euthycarcinoid was 310
million years old, while the first insect dates from about 370 million years
ago.

Predators at large

So, what were the various animals doing walking around on Kalbarri’s sandy
river flats? Often, interpretations of the evolution of organisms into major
new ecological niches suffer from what can only be termed the Star Trek
syndrome: the implication that the species are ‘boldly evolving where no
species has evolved before’. In other words, there is a vacant ecological
niche, so it must be filled. In the case of these early arthropods it was
one thing to have a vacant niche to evolve into, another to possess the
morphological and physiological adaptations needed to cope with the effects
of gravity, low oxygen levels, large diurnal temperature ranges and the
ever-present danger of desiccation. If there was little in the way of plants
to attract animals on to the land, we must turn to other causes.

In recent years, palaeontologists have recognised predation as an important
driving force in the evolution of a number of marine invertebrates, in
particular sea urchins, barnacles and some molluscs. We believe predators
were also the driving force behind the colonisation of land. In addition to
eurypterids and numerous other arthropod predators that flourished in the
seas, lakes and rivers, a major new group of predators were evolving in the
oceans and rivers at the time – the fishes. The great explosion in
diversity of fishes took place between 400 and 420 million years ago. The
appearance of this new group of predators, many of which would have fed upon
aquatic arthropods, would have increased the predation pressure felt by some
of the small arthropods.

This would explain the various trackways at Kalbarri. Many were made by
animals moving from one pool to another as the water dried up in the dry
seasons. In such a situation, with density of animals increasing in the
pools as the water levels shrank, predation levels would have been very
high. The stomach contents of eurypterids from elsewhere show them to have
been carnivores, and they would have followed the euthycarcinoids and
smaller arthropods out of the water hoping for a meal. The longer any small
arthropod could survive out of water and outrun any eurypterid predator, the
better its reproductive chances – and the more likely that its genes would
flourish.

While the burrows in Pennsylvania provide indirect evidence of the existence
on land of soil animals as early as 450 million years ago, a number of new
discoveries of body fossils suggest it was between 420 and 380 million years
ago, during the late Silurian and early Devonian, that the major
colonisation of the land by animals took place.

The oldest body fossils of land animals, dated at 414 million years old,
were found in 1990 at Ludlow on the Welsh Borders (‘Oldest creepies crawled
the land in Shropshire’, Â鶹´«Ã½, 10 November 1990). The famous Ludlow
Bone Bed has yielded fragments of fish and other fossils for many years, but
in 1990 Andrew Jeram, then at the University of Manchester, discovered a
large quantity of arthropod cuticle in rocks just above the bone bed. On
sorting the mass of tiny leg and body segments and a little articulated
material from the muddy siltstone, Jeram identified the fragments as
belonging to at least two sorts of extinct centipedes and a trigonotarbid
arachnid – a group of extinct arthropods that possessed six pairs of
appendages, four pairs for walking and two pairs for feeding. They differ
from spiders in lacking silk-producing spinnerets, and have armour plates
on their abdomens.

How can palaeontologists be so certain that the Ludlow remains are of
terrestrial animals? First, all trigonotarbids found at other undoubtedly
terrestrial sites, such as Rhynie in Scotland and the 375 million-year-old
site at Gilboa in New York State, bear structures called book-lungs. These
distinctive air-breathing organs are possessed by spiders today, suggesting
that trigonotarbids were also air-breathers. Secondly, all modern centipedes
are terrestrial, and the structure of the legs of the two Ludlow centipedes
is that of an animal which moved on land, not in water. Thirdly, land plants
were also found with the remains, including Cooksonia.

Rhynie and Gilboa are rich sources of terrestrial body fossils from the late
Silurian and early Devonian periods. At Gilboa centipedes, trigonotarbids,
mites, pseudoscorpions and the earliest known spider, Attercopus, have been
found, while the 395 million-year-old deposits at Rhynie have yielded
springtails, mites and trigonotarbids. An interesting fact to emerge from
the fossils found at all three sites is that early terrestrial arthropods
were predominantly carnivorous. No herbivores are known. The predators
probably fed mainly on microarthropod decomposers, such as mites, millipedes
and soft-bodied worms, that were rarely preserved but which probably
comprised a substantial part of the fauna that inhabited the soil.

Strangely, there is little evidence of herbivory in the fossil record until
well into the Carboniferous period, about 330 million years ago. It is only
in fossilised leaves of this age and younger that there is unequivocal
evidence of the kind of damage caused by animals. William Shear of
Hampden-Sydney College, Virginia, who has done much of the work on the
Gilboa arthropods, has suggested that by-products from the synthesis of
lignin, which was present in early vascular plants, were toxic to early
terrestrial animals. True herbivory only really evolved when animals
developed enzymes and a gut microflora of symbiotic bacteria which could
bypass the decomposer and break up fresh plant material, returning it to the
ground as excrement.

A second group of animals to step on to land during the late Silurian and
early Devonian were the tetrapods – the first terrestrial vertebrates which
had four limbs and had evolved from the crossopterygian fish with its four
body fins. Anne Warren and colleagues at La Trobe University, Victoria, have
found tetrapod trackways in the Grampians hills about 250 kilometres from
Melbourne, dating from about 385 million years ago.

Driven to land

We believe their appearance on land may have again been driven by predators
because the timing corresponds to the rapid expansion in the diversity of
fishes. The fact that one of the most abundant groups – the placoderm fishes
– evolved a heavy armour to protect themselves in the early Devonian
supports the argument for high levels of predators within the aquatic
environment. John Long of the Western Australian Museum, Perth, found
stunning fossil evidence of this recently in Devonian rocks in the Kimberley
region of Western Australia. It was a specimen of the predatory
crossopterygian fish Onychodus preserved with a smaller placoderm in its
mouth. It is presumed that the Onychodus choked to death on its meal.
Devonian rocks elsewhere have yielded other fossilised examples of fish
predation. So, for fishes that evolved fins capable of supporting the body
weight out of the water and were able to breathe air, the land would have
offered a refuge from the predation pressures present in the seas, lakes and
rivers.

The earliest skeletal remains of tetrapods are known from younger rocks than
those in which tracks are preserved. Per Ahlberg at the University of Oxford
has recently identified a tibia, humerus and incomplete jaws of a tetrapod
from late Devonian rocks, some 370 million years old, near Elgin in
Scotland. There is no doubt that these early tetrapods had well developed
legs with digits (up to eight on each forelimb in Acanthostega, seven in
Ichthyostega and six in Tulerpeton).

Mike Coates and Jennifer Clack of the University of Cambridge, however, have
recently questioned if some of these early tetrapods preserved in late
Devonian rocks in Greenland, were fully terrestrial animals. From their
analysis of well-preserved material they have discovered that Acanthostega
retained fish-like internal gills and other structures associated with
aquatic respiration. Coates and Clack suggest that the legs may well have
evolved primarily for use in water. The forelimbs were, they believe, more
likely to have been paddle-like. Furthermore, compressed lower leg bones in
Ichthyostega closely resemble the pectoral flippers of whales, while the
fossilised remains of Tulerpeton are only found in marine sediments far from
the shoreline. But these well-developed legs, like the hard exoskeleton of
arthropods, meant that tetrapods were superbly equipped for life on land.

For those animals able to withstand the rigours of a non-aqueous existence,
the land was the way out. Here, at least early on, was a host of new niches
to be exploited, free from the great pressure of predation existing in the
seas, lakes and rivers. When all the morphological, physiological and
environmental problems of a terrestrial existence were solved, it may have
taken only the build-up of sufficient aquatic predation pressure to lead to
the conquest of the land. Who knows, maybe animals didn’t jump – perhaps
they were pushed.

Ken McNamara is Senior Curator of Invertebrate Palaeontology at the Western
Australian Museum, Perth, and Paul Selden is Senior Lecturer in
Palaeontology at the University of Manchester.

* * *

Conquering a new land – plants or animals first?

The success of predatory arthropods in colonising virtually barren
landscapes, analogous to the situation on Earth about 420 million years ago
during the Silurian, is seen today in the sequence of ecological succession
at sites such as Mount St Helens and Krakatoa, following violent volcanic
explosions.

Zoological expeditions to the Krakatoa islands in 1984 and 1985, led by Ian
Thornton of La Trobe University, Victoria, revealed that arthropods quickly
colonised areas of the bare lava formed from volcanic activity in the 1950s
and 1970s, well before any plant became established. Thornton found
arachnids, springtails and a range of insect groups.

Just two months after the explosion of Mount St Helens in 1980, long before
any plant life had become established, 43 species of spiders had arrived by
‘ballooning’ – floating through the air on gossamer threads. Similarly,
crickets and wolf spiders are among the first colonisers on newly formed
lava flows on Hawaii, arriving well before plants.

Springtails, present in some of the earliest terrestrial fossil deposits,
are known to be very early, opportunistic colonisers of new islands. On both
Krakatoa and the volcanic island of Surtsey in the North Atlantic,
springtails were present at a very early stage. The early arachnids and
springtails on these islands quickly established an ‘aeolian’ ecosystem,
existing on other windborne arthropods, either as scavengers or predators.

Consequently, the ability of arthropods to establish ecosystems in regions
devoid of plant life supports the view that the earliest terrestrial
ecosystems on Earth may have involved arthropods and ‘lower’ plants, such as
liverworts and mosses, to the exclusion of ‘higher’ plants. Furthermore, the
evolution of the ‘higher’ vascular plants may have postdated the conquest of
land by animals. Herbivorous animals that fed on the higher plants may have
been the last to arrive on the scene.