Conundrum?? Bramble and Lieberman (2004), in a much-discussed review article
in Nature, cite a number of derived Homo features they claim to be
adaptations for more efficient endurance running in arid, open habitats.
However, while some of these supposedly Œcursorial adaptations¹ appear first
in the fossil record in H. habilis, others appear first in H. erectus, and
others still in H. sapiens, suggesting a much more complex story than
proposed by Bramble and Lieberman. Their conclusions are reached without
systematic comparisons with other animals (including endurance runners) and
with general comparisons restricted to fossil hominids and Pan. Since
convergent traits are strong indicators of evolution in similar environments
(Bender 1999), a systematic comparison with a broad range of animals with a
variety of locomotor strategies would have been more informative.
In addition, discussion of possible locomotion styles is restricted to
walking and running, with no consideration at all given to activities such
as wading, swimming or underwater foraging, yet humans are regular waders
and more accomplished swimmers and divers than other primates. Most of the
list¹s Œadaptations¹ for walking could just as easily be explained by
wading. One of the frequent Œexplanations¹ in the list is ³stress
reduction², a reference to the vertical posture of humans with the weight
resting on two legs. But this says nothing about endurance running, with
standing, wading, walking or short distance running all using a similar
posture, and therefore all requiring stress reduction. Other Œexplanations¹
include ³counter rotation², ³thermoregulation² and ³stabilization², but no
comparative data to corroborate these interpretations are provided. In other
words, their Œexplanations¹ are ad hoc suppositions, applied to one example
(human ancestors) without any consideration as to whether these supposed
adaptations are seen in other animals, which means their Œexplanations¹ are
statistically invalid (n=1). Long legs, and possibly shortened forearms,
could be seen as running adaptations, but these are just as typical of
wading and swimming species compared with runners (Hildebrand 1974: 584,
Bender 1999).
In a waterside scenario, wading and swimming would be preadaptative to the
humanlike Œvertical¹ locomotion that Bramble and Lieberman (2004) believe to
be a direct adaptation to endurance running. In our view, frequent
terrestrial locomotion, whether for walking or for (relatively slow)
running, was more recent (Homo sapiens) and could not be derived directly
from an ancestral locomotion in forests, whether on the ground or in the
branches, because in that case a more baboon-like locomotion would be
expected (the Œbaboon paradox¹).

Table 4. Bramble and Lieberman¹s (2004) list of supposedly derived features
of the human skeleton with so-called cursorial functions

Functional role in running & walking according to Bramble & Lieberman
(2004) W = walk R = run Earliest evidence Comparative data. More
likely alternatives in our opinion. NSS = not seen in savannah animals. NSC
= not seen in cursorial animals. NUL = not unexpected in littoral animals.
Enlarged posterior & anterior semicircular canals Head/body
stabilization R H. erectus NSS as far as known. NUL, e.g., for
equilibrium during descent & ascent in diving. Requires more comparative
data.
Expanded venous circulation of neurocranium Thermoregulation R>W H.
erectus NSS. NSC. NUL. Skull base & paravertebral venous networks are
typical of diving species.
More balanced head Head stabilization R H. habilis NSS. NSC. Could
be advantageous in frequent standing rather than running. Alined build NUL.
Nuchal ligament Head stabilization R H. habilis NUL, e.g., in
pronograde swimming.
Short snout Head stabilization R>W H. habilis NSS. NSC. Snout
shortening has to do with mastication rather than head stabilisation.
Tall, narrow body form Thermoregulation R>W H. erectus NUL: long
legs are typical of wading species.
Decoupled head & pectoral girdle Counter-rotation of trunk vs head R
H. erectus? NUL: waterside as well as a mosaic milieus require versatile
locomotions.
Low, wide shoulders Counter-rotation of trunk vs hips R H. erectus?
NUL: Œlow¹ could be for wading as well as for underwater swimming. No
relation to running.
Forearm shortening Counter-rotation of trunk - H. erectus NUL:
typical of frequently swimming species.
Narrow thorax Counter-rotation of trunk vs hips R H. erectus?
Dorso-ventrally narrow. NSS, NSC, NUL: typical of shallow water dwellers,
e.g., platypus, hippo, beaver.
Narrow & tall waist between iliac crest & ribcage Counter-rotation of
trunk vs hips R H. erectus? NUL: waterside as well as mosaic milieus
require a wide range of locomotions.
Narrow pelvis Counter-rotation of trunk vs hips Stress reduction R

abduction: NSS, NSC, NUL.
Expanded lumbar central surface area Stress reduction R>W H.
erectus Suggests vertical body. NUL, e.g., for wading.
Enlarged iliac pillar Stress reduction R>W H. erectus Idem.
Stabilized sacroiliac joint Trunk stabilization R H. erectus
Idem.
Expanded surface area for mm. erector spinae origin Trunk stabilization
R H. erectus Idem.
Expanded surface area for m. gluteus maximus origin Trunk stabilization
R H. erectus Idem.
Long legs Stride length R>W H. erectus NUL, typical of wading
species.
Expanded hindlimb joint surface area Stress reduction R>W H.
erectus Suggests vertical body. NUL, e.g., for wading.
Shorter femoral neck Stress reduction R>W H. sapiens Not seen in
H. erectus. Presumably post-littoral.
Long Achilles tendon Energy storage Shock absorption R Homo?
Comparative data are needed. Typical cursorial species are not plantigrade.
NUL.
Plantar arch (passively stabilized) Energy storage Shock absorbtion
Powered plantar-flexion R R>W R>W Homo? NSS. NSC. NUL: plantigrady for
wading and swimming.
Enlarged tuber calcaneus Stress reduction R>W Homo? Cursorials
do not have enlarged heels. NSS. NSC. NUL.
Close-packed calcaneo-cuboid joint Energy storage Stability during
plantarflexion R>W OH-8 Comparative data are needed. NUL.
Permanently adducted hallux Stability during plantarflexion R>W
OH-8 NUL: wading, swimming.
Short toes Stability during plantarflexion Distal mass reduction R>W
OH-8 NSS. NSC. NUL: metatarsal lengthening and toe shortening is to be
expected in swimming & wading.

Most of Bramble and Lieberman¹s Œadaptations¹ are not what we would expect
in a cursorial (running) animal. For example, their list includes ³enlarged
posterior and anterior semicircular canals², but there are no comparisons
with, for instance, giraffes (heads high above the ground), gibbons (fast
and versatile locomotion), kangaroos (cursorial bipeds), or swimming or
diving species. It is conceivable in fact that the frequent change of
posture seen when diving for seafood (descending and ascending) required a
different labyrinth structure, and that the larger Homo erectus labyrinth
was adapted to terrestrial walking and running as well as to wading,
swimming and diving locomotions.
There is no indication that an ³expanded venous circulation of neurocranium²
had anything to do with thermoregulation, but there is long-standing
evidence of expanded venous networks in diving species (Slijper 1936).
More balanced heads and short snouts are not seen in cursorial species,
whether bi- or quadrupedal, and low shoulders are to be expected in wading
and underwater swimming.
What Bramble and Lieberman refer to as ³narrow body form², ³narrow thorax²
and ³narrow pelvis² is not clear to us: compared to most primates, humans
have a relatively broad thorax and pelvis (laterolaterally), and this was
even more so in the case of australopithecines. In our opinion, the
combination of Œflared¹ iliac blades and long and relatively horizontal
femoral necks as seen in Homo erectus indicates well-developed ad- and
abduction, which is obviously not an adaptation for running, but would not
be unexpected and indeed would be advantageous for a species that had to
regularly wade, tread water, swim or climb. In Homo sapiens the pelvis
(bi-iliac diameter) did become narrower and the femoral necks shorter and
more vertical, and we agree with Bramble and Lieberman that this could be
related to more frequent terrestrial locomotion.
Plantar arches, enlarged tubera calcanei, close-packed calcaneo-cuboid
joints and short toes are not seen in cursorials, whether bi- or quadruped,
to the contrary: running species are typically unguli- or digiti-, not
plantigrade, and typically have elongated toes.
In conclusion, comparative data suggest that none of the features described
by Bramble and Lieberman (2004) are typical either of savannah dwellers or
frequently running animals, whether slow or fast. Until the features are
considered in the context of swimming and wading as well as terrestrial
movement, their interpretations should be considered with extreme caution.
As it is, there is no obvious reason why any of the features cited could not
have been of advantage in a littoral environment. We do not deny that
humans today are adapted to terrestrial locomotion including walking and
moderate running, but in our opinion the peculiar human anatomy is not
directly derivable from a typical primate ancestor who moved from closed to
more open, arid habitats.
At least two conspicuous anatomical features of Homo erectus are notably not
included in the list of features cited by Bramble and Lieberman (2004).

1) Homo erectus typically has a more robust, and therefore heavier,
skeleton than all other (fossil and extant) primates, including H. sapiens
and the other apes. One of its defining characteristics is the shape and
size of the femoral bone, which shows cortex thickening and densening
(pachyostosis) and a narrow cavity of the bone marrow (medullary stenosis).
The cranial bones, especially the posterior part (the occiput), are also
notably thicker than in other primates including H. sapiens. Unusually heavy
bones would be a disadvantage for a species relying on endurance running,
and are not seen in running mammals such as dogs or horses, whereas for a
species collecting sessile food from the water¹s edge, including underwater
foraging, they could have been a significant advantage. Human divers such as
the Ama of Korea frequently use weights to help them descend (Hong and Rahn
1967). Slow-diving mammals for sessile foods typically have medullary
stenosis and pachyostosis to a higher degree than in H. erectus (walruses,
dugongs and fossil littoral species such as Kolponomos, Odobenocetops and
some Thalassocnus species), while fast-diving mammals for mobile prey have
light-weight bones (dolphins and sealions).

2) Archaic Homo had a lower and longer brain skull than H. sapiens, with
generally less flexed cranial base and with the eyes somewhat more in front
of the brain (requiring a supraorbital torus for eye protection) rather than
fully below the frontal brain as in H. sapiens, meaning that the eyes would
have been more naturally oriented towards the sky if they were standing with
an upright posture, rather than directed more towards the horizon as is the
case when H. sapiens stands upright. This would be a disadvantage for a
species relying on endurance running because, among other things, more
energy would be needed to look at where the feet were making contact with
the ground. In a diving position, as well as in a more procumbent body
position while wading for food, for example, this would have resulted in the
eyes being more naturally oriented in the direction the individual was
moving (i.e., in the case of swimming and diving, head first through the
water). We are not aware of any models that suggest early Homo ran with a
bent hip posture, but we do note that human sprinters generally run with the
body leaning forward.

Within many contemporary H. sapiens populations there are individuals who
are capable of long distance running, but compared to typical savannah
species, humans are slow and inefficient (Figure 4). Moreover, recent
research suggests that endurance training in athletes sometimes causes
cardial arrhythmias and sudden death (Ector et al. 2007). Even Bramble and
Lieberman (2004) admit that ³humans are mediocre runners in several
respects² and ³running is more costly for humans than for most other
mammals². And since H. erectus generally had, for instance, heavier bones
than H sapiens and longer femoral necks, it must have been an even less
efficient cursorial than extant H. sapiens.