Content uploaded by Michael Crawford
Author content
All content in this area was uploaded by Michael Crawford on Jun 10, 2015
Content may be subject to copyright.
British
Journal
of
Nutrition
(1998),
79,
3-21
Review article
3
Rift Valley lake fish and shellfish provided brain-specific nutrition for
early
Homo
C.
Leigh Broadhurst'*t, Stephen C. Cunnane2 and Michael
A.
Crawford3
122"d Century Nutrition Inc., 1315 Harding Lane, Cloverly, MD 209054007, USA and Visiting Scientist, Nutrient
Requirements and Functions Laboratory, Building 307, Room 224, USDA Beltsville Human Nutrition Research Center,
Beltsville, MD
20705,
USA
2Department
of
Nutritional Sciences, Faculty
of
Medicine, FitzGerald Building,
150
College Street, University
of
Toronto,
Toronto, Ontario
M5S
IA8, Canada
31nstitute
of
Brain Chemistry and Human Nutrition, University
of
North London, 166-222 Holloway Road,
London N78DB,
UK
(Received
3
January
1997
-
Revised
2
September 1997
-
Accepted
3
September 1997)
An abundant, balanced dietary intake of long-chain polyunsaturated fatty acids is an absolute
requirement for sustaining the very rapid expansion of the hominid cerebral cortex during the
last one to two million years. The brain contains
600
g lipid/kg, with a long-chain poly-
unsaturated fatty acid profile containing approximately equal proportions of arachidonic acid
and docosahexaenoic acid. Long-chain polyunsaturated fatty acid deficiency at any stage of fetal
and/or infant development can result in irreversible failure to accomplish specific components
of brain growth. For the past fifteen million years, the East African Rift Valley has been a
unique geological environment which contains many enormous freshwater lakes. Paleoan-
thropological evidence clearly indicates that hominids evolved in East Africa, and that early
Homo inhabited the Rift Valley lake shores. Although earlier hominid species migrated to
Eurasia, modem
Homo
supiens
is believed to have originated in Africa between
100
and
200
thousand years ago, and subsequently migrated throughout the world. A shift in the hominid
resource base towards more high-quality foods occurred approximately two million years ago;
this was accompanied by an increase in relative brain size and a shift towards modem patterns of
fetal and infant development. There is evidence for both meat and fish scavenging, although
sophisticated tool industries and organized hunting had not yet developed. The earliest
occurrences of modem
H.
supiens
and sophisticated tool technology are associated with aquatic
resource bases. Tropical freshwater fish and shellfish have long-chain polyunsaturated lipid
ratios more similar
to
that of the human brain than any other food source known. Consistent
consumption of lacustrine foods could have provided a means of initiating and sustaining
cerebral cortex growth without an attendant increase in body mass.
A
modest intake of fish and
shellfish
(6-12
%
total dietary energy intake) can provide more arachidonic acid and especially
more docosahexaenoic acid than most diets contain today. Hence, 'brain-specific' nutrition had
and still has significant potential to affect hominid brain evolution.
Fish: Brain-specific nutrition: Long-chain PUFA: East Africa Rift Valley Lakes
Marked expansion
of
the hominid cerebral cortex took
place only in the last one to two million years. During this
small evolutionary window, genus Australopithecus be-
came extinct while Homo greatly expanded. Sophisticated
tool manufacture, organized hunting, culture, and speech
followed rapidly. We hypothesize that the unique geolo-
gical and ecological environment of the East African Rift
Valley provided an equally unique nutritional resource base
for the enlargement of the Homo brain, culminating in
Homo sapiens.
How in this remarkably short stretch of
evolutionary history did our intelligence arise? While many
physical (i.e. development of bipedalism, speech), ecolo-
Abbreviations:
AA,
arachidonic acid; CNS, central nervous system;
DHA,
docosahexaenoic acid; EPA, eicosapentaenoic acid; EQ, encephalization
quotient;
FD,
high-fish mixed-diet population group; LA, linoleic acid; LC-PUFA, long-chain PUFA; LNA, a-linolenic acid; PUFA, polyunsaturated fatty
acids;
VD,
vegetarian-diet population group.
*Corresponding
author: Dr
C. Leigh Broadhurst, email cleigh@cais.com
?Author
is
not USDA employee.
4
C.
L. Broadhurst
et
al.
gical (adaptation to omnivorous diet, drier climatic condi-
tions) and cultural adaptations (use of tools, living in
groups) have roles, these adaptations alone are apparently
not sufficient to account for the unique intelligence and
culture we have today. If these adaptations alone were
sufficient, then in all cases we must ask ourselves why no
other primates developed as such.
Previous authors (Martin, 1983; Harvey
&
Clutton-
Brock, 1985; Blumenschine, 1991; Foley
&
Lee, 1991;
McHenry, 1994) have considered an ecological approach to
human evolution. In these arguments, the high metabolic
energy requirements of the brain require that hominids
must have accessed relatively-high-quality and abundant
food resources. Various selective pressures affecting
hominid evolution might be necessary, but are not
sufficient conditions for cerebral expansion. Whatever the
selective pressures, they can only be satisfied in a context
where sufficient dietary energy and essential nutrients are
available to fuel the added brain growth. Sufficient protein,
vitamins, and trace elements are certainly required, but
dletary essential polyunsaturated fatty acids (PUFA) are
probably the most limiting nutrients for neural growth
(Crawford
&
Sinclair, 1972). If hominid diets were
consistently deficient in long-chain (LC-) PUFA, the
uniquely complex human neurological system could not
have developed, regardless of the diverse stimuli that may
have been involved.
We are in agreement with the ecological approach to
evolution of human intelligence, and propose that nutrition
played a more crucial role in the rapid neural development
of genus Homo than has been considered previously. We
concentrate specifically on the nutrition provided by the
unique ecosystem of the East African Rift Valley lakes.
The fossil evidence clearly indicates that Homo arose in the
vicinity of these lakes, which are geologically better
classified as ‘proto-oceans’
.
The diverse alkaline-fresh-
water fish species within those lakes provided, either
directly or indirectly, a source of both protein and PUFA. In
particular, the freshwater-fish lipid profile has a docosa-
hexaenoic acid (DHA)
:
arachidonic acid
(AA)
value that is
closer to that in our brain phospholipids than any other food
source known. We hypothesize that consistent consumption
of fish, crustaceans, molluscs, and other lacustrine species
from the lake margins provided a facile means of both
initiating and sustaining growth of the cerebral cortex
without an attendant increase in body mass.
We are aware that the origin of human intelligence is one
of
the most fundamental questions ever posed by man, and
we do not propose
to
answer it completely in this brief
discourse. Many genetic, environmental and climatic
conditions almost certainly conspired to allow for the
selection and expansion of our large, complex brains. We
discuss some of these relevant conditions in the following
sections. However, we are not fully satisfied with the slow
pace and passivity of evolutionary models based entirely on
selective pressures. For example, we can hypothesize two
end-member conditions for the eventual occurrence of
H.
sapiens:
Did hominids ‘become’ intelligent enough to
begin fishing, or did they fish and then become intelligent?
Since these two end-members are not mutually exclusive,
and would in fact reinforce each other, the answer
is
likely
to lie between. We envision something such as: horninids
scavenged fish and/or fished opportunistically, which
helped increase intelligence enough for them to fish more
often and more successfully. We aim to bring attention to
the fact that the latter end-member, in which nutrition plays
a crucial role in the origin and maintenance of intelligence,
has not been given adequate consideration.
Encephalization quotients
(EQ):
quantifying cerebral
expansion
The term EQ was introduced in the 1970s to account for the
influence
of
body size on brain size, thereby permitting a
scale of comparison between species that would identify
relatively larger-brained species independent of body size
(Martin, 1983). The EQ compares brain weight
:
overall
body weight for all species of interest, and is scaled
so
that
the comparison among brains is in effect done for constant
body size. It has become a valuable tool for quantifying the
remarkably larger brain size in
H.
sapiens
v.
other extant
primate and extinct hominoid species (Table 1). By
providing a quantitative and reproducible scale, the EQ
has helped evolutionists focus on developing an adequate
explanation for the enlargement of the evolving human
brain. Exact EQ values vary among sources, depending on
the method of calculation and the database used; however,
all sources consistently show that adult
H.
sapiens
has an
EQ about 2.4-2.7 times larger than adult
Pan troglodytes
(modem chimpanzee). They also consistently show that as
hominids evolved, EQ increased from about 1.4 in
Australopithecus africanus
to about 2..4 in
Homo
erectus.
EQ values are not as different between neonates of
various primate species as they are between adults,
suggesting that both brain and body growth postnatally
have an influence on the resultant EQ of adults. It should be
noted that overall body weight is still a key reference point
for establishing EQ values, and modem-day primates are
significantly leaner than humans. Hence, the presence of
15-30
%
body fat in humans actually reduces the EQ
difference between primates (or hominids) and
H.
sapiens.
This is especially true for neonates, since human infants
have much more body fat than infants of other primate
species.
Table
1.
Mean brain volumes and encephalization quotients (brain
weight
:
overall body weight;
EQ)
for selected hominoid species
Species Brain volume (ml)
EQ1’
EQZt
Australopithecus:
afarensis
africanus
boisei
robustus
habilis
rudolfensis
erectus
sapiens
Homo:
Pan troglodytes
384
420
488
502
579-597
709
62-44
1250
41
0
1.23
1.31
1.37
1.49
1.74-1.79
1.41
1.59-1.63
3.05
1.25
1.45
1.62
1.72
1.92
2.1 0-2.29
2.1 1
2.38-2.44
4.26
1.57
*
From
calculations
of
Martin
(1983).
t
From
calculations
of
Harvey
8
Clutton-Brock
(1
985)
Fish provided brain-specific
nutrition
5
East African Rift Valley geological summary
The Red Sea, Gulf of Aden, and East African Rift Valley
are the only current examples of what is termed
geologically as a ‘failed ocean’. Rifting began about thirty
million years ago, thinning and stretching the continental
crust, but significant uplift did not begin until fifteen
million years ago. Bohannon
et
al.
(1989) proposed that
rifting was passive, doming postdated rather than preceded
uplift, and was caused by adjacent asthenosphere (plastic,
flowing mantle) and deep continental lithosphere (partially
rigid mantle) flowing into the area of thinned crust. The
Red Sea axis has thin longitudinal strips of oceanic crust
that are only approximately five million years old; 2.5-
4
km
of uplift has occurred in the continental areas
adjacent to the Red Sea in the past 13.8 million years.
In
East Africa, faulting related to the crustal extension
and uplift formed a series of half-graben basins which link
together to form the Rift Valley (Fig. 1). Large lakes
formed in the basins, with inputs from both interior
drainage and river systems. On the border fault side
of
the lakes, cliffs may rise to
>
2
km
above the lake level.
Some of the lakes were so extensive during Cenozoic
(sixty-five million years ago to present) highstands that
they are termed proto-oceans. Lakes Malawi and Tangan-
yika presently have water depths up to 1500 and 600 m
respectively. Many of the deep water channels in Lakes
Malawi and Tanganyika are similar in form and scale to
those observed in the deep ocean (Scholz
et
al.
1990).
Lakes with interior drainage have little or no input from
major rivers, the water levels are highly dependent on the
geology and climate. For example, Lake Victoria currently
loses 90
%
of its water input by evaporation; therefore the
lake level is very sensitive to regional temperature and
rainfall (Leeder, 1995). During the Miocene and Early
Pliocene, lakes covered
>
1000
000
km2
that are now
desert or savanna. By comparison, Lake Victoria is
presently 69 000
km2,
and is the world’s largest tropical
lake (Stager
et
al.
1997).
Permanent lake levels fluctuated widely during the
Middle to Late Pliocene and Pleistocene between shallow,
alkaline and saline, and deep and stratified. Smaller lakes
tended to become more alkaline and saline, or even
ephemeral. Associated with the lakes are fine-grained
lacustrine, shoreline sand, coarse-grained fluvial or chan-
nel, and alluvial-fan or delta deposits. These sedimentation
patterns indicate prolonged lowstands in the Pleistocene
(Baker
et
al.
1972; Scholz
&
Rosendahl, 1988; Scholz et
al.
1990). Lake Victoria, for example, is at least 043-1.6
million years old, but its level has fluctuated greatly. The
lake was completely
dry
in the Late Pleistocene, until 12.4
thousand years ago (Johnson
et
al.
1996).
Even in a passive rifting model, there is broad agreement
that the Afar area (Ethiopia; Fig. 1) overlies a mantle
plume. Extensive alkalic volcanism and plutonism (molten
rock intruded into the crust, not erupted) along the East
African Rift has occurred in the past twenty-six million
years (Dawson, 1992). The Rift is the largest peralkaline
volcanic province in the world, and has the only volcano
known to have erupted carbonatite tephra and lava in
historic times (Dawson
et
al.
1996).
Carbonatite/peralkaline magmas (magma is molten rock
below surface, actual or hypothetical) are very unusual in
that they are relatively poor in SOz, and water, but
relatively rich in C02, halogens, Na, Ca, and trace
elements. They are associated with continental rifting or
oceanic islands over mantle plumes. Large stratovolcanoes
(typical cone-shaped volcanoes with alternating layers of
lava, ash and pyroclastic flows) are common, and much of
the Cenozoic strata in the Rift Valley consists of lava, tufa
and pyroclastic flows, or ash fall. These carbonatite-type
volcanic rocks are rich in alkali and incompatible (i.e. rare
earth elements, Zr, Rb, halogens) elements (Bailey
&
Macdonald, 1987; Bestland
et
al.
1995). In summary, the
East African Rift Valley, with its extensive proto-oceanic
lakes and intracontinental volcanic activity, is a unique
tectonic province unmatched elsewhere on Earth in both
type and extent.
Divergence
of hominids from hominoids
The phylogenetic designation ‘hominoid’ refers to the
Superfamily Hominoidae, which contains the extinct
common ancestors of apes and humans, as well as extinct
precursors to, and extant genera of both apes and humans.
The designation ‘hominid’ refers to the Family Hominidea,
including only bipedal primates considered to be in the
lineage leading to modem humans. Combined geological
and environmental conditions are thought to have initiated
the divergence of Hominidae from Hominoidae.
Before divergence, protosavanna (evergreen woodland
with some open patches) appeared in Southwest Africa,
Arabia, and North Africa 17 million years ago, but did not
appear in equatorial Africa until 14 million years ago
(Harris, 1993), roughly coincident with the onset of
continental uplift. Protosavanna existed until the Late
Miocene, although there was a mosaic of environments,
with at least some rainforest present in the Eastern Rift 12.2
million years ago (Jacobs
&
Kabuye, 1987).
Extensive uplift and faulting occurred along the rift axis
starting at eight to nine million years ago. The tectonic
activity may have geographically isolated two groups of the
common ancestor of Ponginae (great apes) and Hominidae
(Coppens, 1994). The population on the eastern, more arid
and open side of the rift evolved into the hominids. Those
on the western side remained in a more humid, arboreal
environment and continued along their evolutionary track
to great apes (e.g. Pan, Gorilla, Pongo).
Near the end of the Miocene, Africa became cooler and
drier. Savanna appeared in equatorial Africa seven million
years ago. Between four and seven million years ago, 75
%
of fifty-nine known land mammal genera were new,
including the first leporids and hominids, new felids, extant
hyaenids, new hippopotamids, extant giraffids, extant and
diverse elephantids, and diverse extant bovids (Harris,
1993; Vrba
et
al.
1995). Growth of the Antarctic ice sheet
lowered sea levels, and was a fairly major factor in a
complete dehydration
of
the Mediterranean Sea approxi-
mately six million years ago. With these major changes in
local oceanic circulation, cold Antarctic ocean currents
flowed along the west coast of Africa and drew moisture-
6
C.
L.
Broadhurst
et
al.
Fig.
1.
Map
of
East Africa and South Africa giving major hominid fossil localities and lakes discussed in text. The Red Sea and the Gulf of Aden
are considered to be two ‘arms’ of a tectonic plate triple junction (i.e. the intersection of three tectonic plates). The third ‘arm’ is the East African
Rift
Valley, extending
from
the Afar area of Ethiopia
(see
inset) to near the eastern Zambezi River, below Lake Malawi. Lake Victoria is the
worlds largest tropical lake; many other lakes are present in the Rift Valley which are not shown on this scale.
Fish provided
brain-specific
nutrition 7
laden air from the land, increasing the aridity of coastal
southern Africa. The equatorial forest began to shrink, and
more drought-resistant flora spread. The transitional
ecological zone between forest and adjacent savanna
increased in extent (Conroy, 1990; Sikes, 1994; Vrba et
al.
1995).
This phenomenon was intensified in the Eastern Rift due
to the combined effects of global climatic changes,
rainshadow and altitude effects from the continued uplift,
and periodic volcanic eruptions which temporarily reduced
surrounding areas to wastelands. Several authors have
hypothesized that adapting to these transitional zones,
neither forest nor savanna, drove the shift from arboreal
quadrupedalism to terrestrial bipedalism in the Late
Miocene primates (Lovejoy, 1981; Conroy, 1990; Coppens,
1994).
As
noted in the previous section, the first oceanic
crust appeared in the Red Sea Rift approximately five
million years ago. Oceanic crust appearance indicates near-
maximum lithospheric attenuation (thinning of the upper-
most brittle layer of crust plus mantle), and in response
volcanic activity in both Africa and Arabia increased.
Australopithecus afarensis
and precursors
About four million years ago, the land mammal generic
diversity nearly tripled
(Harris,
1993). Included in this
expansion are the oldest hominids known (Fig. 2).
Currently
Ardipithecus ramidus
is recognized as the oldest
hominid, dating to
4.4
million years ago (White
et
al.
1995;
fossils from
Ararnis,
Ethiopia only). It cannot be deter-
mined conclusively from the fossils recovered whether this
hominid was bipedal. The oldest bipedal hominid
Austra-
lopithecus anamensis
dates to 3.9-4.2 million years ago
(Leakey
et
al.
1995; fossils from Allia Bay and Kanapoi,
Lake Turkana, Kenya).
Australopithecus afarensis,
the
oldest undisputedly widespread hominid, is found in many
localities dating to 3.8-3.9 million years ago. (One recent
fossil has been dated to 4.18 million years ago, which if
confirmed may make this species contemporaneous with
A.
anamensis
(Kappelman
et
al.
1996).)
Australopithecine fossils are confined to Africa, and
collectively demonstrate that bipedalism preceded dramatic
enlargement of cranial capacity by over 2 million years.
A.
afarensis
was primitive, with a brain volume of approxi-
mately 400
ml,
well within the size for extant apes (Table
1). AL-288 ‘Lucy’ had a birth canal which was not
designed to allow the passage of an enlarged fetal cranium
(Lovejoy, 1981; Rak, 1991; Rosenberg, 1992). Major
dentition differences between pongids and
A. afarensis
are indicative of a shift towards an omnivorous diet,
including: (1) decrease in canines and
loss
of canine gap;
(2)
increase in premolar and molar size; (3) overall more
gracile dentition (Rak, 1983; Kimbel
et
al.
1984; Suwa
et
al.
1994). Dental wear indicates that
A. afarensis
ate mostly
plant foods, but occasional small vertebrates and insects are
not ruled out; neither would be eggs and invertebrates such
as molluscs (Johansen
&
Edgar, 1996).
Australopithecus africanus
A. afarensis
remains have not been found in any strata
younger than three million years. The next hominid to
appear in the fossil record is
A. africanus
at 2.8 million
years.
A. africanus
was about the same height and weight,
and is thought to have eaten a similar vegetarian or
omnivorous diet (Skelton-& McHenry, 1992; Wood, 1992;
Sikes, 1994; Johansen
&
Edgar, 1996).
A. africanus
shows
no evidence for the delayed maturation characteristic of
human infants, or selection for a pelvic structure designed
to accommodate an enlarged fetal cranium (Rak, 1991;
Rosenberg, 1992; Conroy
&
Kuykendall, 1995; this would
be true for
A. aethiopicus
also,
but this species is very
poorly represented in the fossil collections). For example, it
is estimated that Australopithecine young erupted their first
adult molars at about 3-4 years of age, similar to pongids,
and dissimilar to human children, who erupt these teeth at 6
years (Smith et
al.
1995).
In general,
A.
afarensis
is thought to be ancestral to two
distinct lineages, one Australopithecine and one Homo
(Fig. 2), but there is no agreement on the detailed
phylogeny, and discussion of such is both beyond the
Dresent scoDe. and subject to change each time a new fossil
hiscovery made. It
is
agreed that
A. boisei, A. robustus,
and
Homo
spp. share
a
co-on
ancestor,
which
is
likely
to
be
A. afarensis,
but may also be a hCminid as yet unknown
Fig. 2.
Proposed phylogenetic tree for hominid evolution. Question-
able lineages represent differing opinions among researchers as well
as incomDlete fossil records. fAdaoted from Wood.
1993
and
>.
Johansen’& Edgar,
1996.)
(Skelton
&
McHenry, 1992; Foley, 1994).
8
C.
L.
Broadhurst
et
al.
Further evolution
of
hominids:
an
overview
Three million years ago the Northern hemisphere climate
was generally warmer, with sea levels at least 25 m higher
than present levels. The most significant warming was in
the high latitudes. Equatorial East Africa was actually
cooler than at present, but was wetter (Dowsett
et al.
1994),
and had evergreen forests (Bonnefille
et al.
1987). From 2.5
to 2 million years ago, coincident with the onset of major
northern hemisphere glaciation (mostly growth of polar ice
caps), the climate became cooler and drier (Shackelton
et
al.
1984; Versteegh
et
al.
1996). The environment was
more similar to that of the present (Bonnefille, 1983), but
was not yet dominated by savanna grassland (Cerling,
1992). There was a mosaic of forest, bush, savanna and
patchy wooded grassland, with a general trend towards a
more open, arid, grassland-dominated environment (Vrba
et
al. 1995). Mammalian generic diversity remained high
and relatively unchanged, but guilds deepened (a guild
being a group of closely related but distinct species that
have very similar ecological requirements and also occur
together in particular habitats; deepening of a guild
indicating that the number of species in the guild is
generally increasing), and some extant genera appeared for
the first time (Harris 1993; Vrba
et
a1.
1995; including
especially grazers such as Equus and Oryx (horse, oryx),
camivore-omnivores such as Vulpes and Ichneumia (fox,
mongoose), and the browser
Loxodonta adaurora
(modern
elephant)).
Australopithecus robustus, A. boisei
and
Homo habilis
Homo was one of these new genera, appearing in East
African deposits dating to 2.3-2.5 million years ago, and
South African deposits dating to approximately two million
years ago (Conroy, 1990; Wood, 1993; Foley, 1994;
Schwarz
et
al.
1994; Johansen
&
Edgar, 1996). Also
approximately 2.3 to 2 million years ago, the gracile
Australopithecines
A.
africanus
and
A.
aethiopicus
were
replaced by the robust species
A.
robustus
and
A.
boisei.
Robust australopithecines coexisted with
H.
habilis,
and
later
H.
erectus,
for over one million years,
so
there must
have been subtle ecological differences between genera
(Conroy 1990; Wood, 1992; Johansen
&
Edgar, 1996).
Fossils in South Africa are confined to cave breccia (an
accumulation of angular rocks from cave roof falls that are
loosely cemented with CaC03; since the breccia accumu-
lates over time by numerous cave collapses and cementing
episodes, it cannot be accurately dated) formed in
Precambrian (older than 580 million years) limestone,
and are not generally used to directly infer phylogenetic,
paleoenvironmental, or chronometric relationships, but
rather to support or repudiate inferences made from East
African fossil localities. Fossils from these limestone caves
are not
in situ,
and are considered to have accumulated via
accidental deposition by predators and/or other hominids,
which makes sample dating very difficult (CONOY, 1990;
Schwarz
et
a1.
1994). In addition, the rock types in which
the fossils are found are not amenable to K
:
Ar
or
U
:
Th
radiometric dating. Schwarz
et al.
(1994) reported an
electron-spin-resonance date for bovid teeth recovered at
the Australopithecine site at Sterkfontein, South Africa
of
2.1 million years ago, but this
is
an average of dates from
1.72 to 2.37 million years ago. Further confusing the South
African fossils in the past was a one million years
radiometric date for the
Dart
‘Taung baby’;
A.
africanus
skull, which has recently been revised to 2.3 million years,
a date consistent with faunal evidence (Tobias
et al.
1993).
In contrast, East African early hominid sites are
contemporaneous with the geological strata. Numerous
stratigraphically coherent layers of lava and volcanic ash
allow for
K
:
Ar dating of many East African localities, and
cross-referencing between localities. East African sites are
associated with watercourses, mostly ancient lake margins,
but also riverene forests (for discussion, see below). Every
site contains both sedimentary and igneous strata which
record continual uplift, faulting, and volcanic activity
(Baker
et al.
1972; Dawson, 1992; Sikes, 1994). The
Miocene
to
Pleistocene lakes were typically 10
000-
100
000
km2
in area.
Hominid localities and paleoenvironments
As depicted in Fig. 1, numerous hominid fossils have been
recovered in Ethiopia (e.g. Hadar and Omo River), Kenya
(e.g. Lake Turkana Basin) and Tanzania (e.g. Olduvai and
Laetoli).
H. habilis
and
H.
erectus
have been found at Omo,
Turkana and Olduvai. (Fossils from Koobi Fora, Turkana,
classified as
H.
habilis
may represent another species
H.
rudolfensis,
but this is not generally accepted (Wood,
1993).) The Hadar locality was mainly a marshy lake
margin with rivers flowing in from the Ethiopian escarp-
ment; Paleo Lake Hadar periodically filled the whole basin.
However, there was a mosaic of microenvironments,
including bush, grassland and wooded areas. The Omo
River locality had both riverene fluvial environments and
swampy lakes. Allia Bay records evidence of the proto-
Omo river system flowing into the Turkana basin, with
bordering gallery forest (Bonnefille
et
al.
1987; Leakey
et
al.
1995; Vrba
et
al. 1995). Overall, Turkana was an
enormous
(>
15000
km’)
lake basin with wide marshy
lake margins, and extensive mud flats which were covered
with grasses in the
dry
season. Lake levels fluctuated
significantly during the Plio-Pleistocene and Lake Turkana
was a closed, alkaline lake for at least part of the time
(Abell, 1982).
The Olduvai locality was also on the margins of a
fluctuating lake, probably with no outlet. The perennial lake
was alkaline and saline, but there was periodic flooding of
the lake basin. Alluvial-fan and plain deposits are also
present, indicating significant sedimentation derived from
continual Rift fault uplift and associated river downcutting
(Leakey, 1971, 1979; Plummer
&
Bishop, 1994; Behrens-
meyer
et
al.
1995). Lateoli was more arid, upland savanna,
not necessarily near a permanent water course (Leakey
&
Harris, 1987; Andrews, 1989; Cerling, 1992).
A.
afarensis
fossils are the only species found at Lateoli and Hadar.
Localities under development include Manoga Valley,
Tanzania (Harrison, 1994), and Semliki Valley, Zaire
(Boaz
et al.
1994) both of which were also large lake
basins. At 2500 km west of the Rift Valley,
A.
afarensis
Fish
provided
brain-specific nutrition
9
fossils dating to
3
to 3-4 million years have been found in
Chad (Brunet et
al.
1995). The paleoenvironment was also
lakeside, with both perennial and permanent streams and a
mosaic of gallery forest, wooded savanna, and open
grassland.
Although Australopithecines were evidently widespread
and existed for two to three million years, their EQ
increased little (Table 1). In an analysis of
A.
boisei, Wood
et al. (1994) found little evidence for gradual modification
of this species. They remarked that ‘evolutionary stasis is
the predominant signal coming from the masticatory
morphometric data
. . .
and is the predominant signal
throughout the time span of the lineage’. How did
H.
habilis and
H.
erectus gain an adaptive and/or intellectual
edge over
A.
boisei and
A.
robustus that resulted in the
dominance of the former, and the extinction of the latter?
The answer may lie in the adaptation of Homo to the lake
margin environment. Clearly all hominids would benefit
from proximity to permanent water courses, especially
since the overall climate pattern was one of progressive
drying and seasonal precipitation. However, certain lines of
evidence indicate that Homo was more likely to inhabit
areas that were relatively open and arid, but that were near
the lake shores.
Behrensmeyer (1975) was one of the first to propose that
H.
habilis may have been more restricted ecologically to
the lake margin than was
A.
boisei. At Turkana,
Behrensmeyer (1975) assigned eighty-four hominids to
major depositional environments, thirty-nine to fluvial, and
forty-five to lake-margin deposits.
A.
boisei was more
abundant in fluvial environments, while
H.
habilis was rare
there. Both are represented in comparable numbers in lake-
margin environments; however
A.
boisei fossils are more
common than Homo in both channel and flood-plain
deposits. The fluvial channels were probably bordered with
gallery forest, as is the case today, while the lake margins
had wide mudflats, swampy in the rainy season and grass-
covered in the
dry
season.
Sikes (1994) reviewed over seventy studies concerning
the reconstruction of Plio-Pleistocene hominid paleoenvir-
onments. While there was no strong consensus regarding
habitats, reflecting perhaps the overall generalistic and
opportunistic nature of Hominidae, trends can be recog-
nized. An overall preference for more closed habitats as
opposed to open savanna was characteristic of all
Australopithecus and Homo species. Very few hominid
localities are reconstructed as open grasslands, with Laetoli
and Swartkrans being the major exceptions.
A.
robustus and
A.
boisei were found in montane, riverene, and closed
forest localities, while Homo was not. In contrast, only
H.
erectus and
H.
sapiens were found on the lake margins per
se, but they also utiIized open arid and closed wet habitats.
Almost all species were found in localities featuring patchy
woodland. Both
A.
boisei and
H.
habilis had among the
widest diversity of habitats, from savanna to riverine forest.
These two species are contemporaneous at Olduvai, the
most investigated hominid locality, and evidence indicates
that hominids in general accessed a full range of habitats
around Paleo Lake Olduvai (Plummer
&
Bishop, 1994).
Arguments related not to archaeology but to the change
in skeletal morphology between Australopithecus and
Homo provide support for the general paleoenvironmental
trends. Ruff (1991, 1994) considered the relationship
between body morphology and ambient climate. In the
past, ambient climate was a more powerful selective force,
since we now modify our microenvironments considerably.
Briefly, there are adjustments that can be made in body
morphology in order to conserve or dissipate heat in
different environments. In hot climates such as tropical
East Africa, it is desirable to dissipate heat and to keep the
body surface area
:
mass value constant as overall body
mass increases. Body breadth cannot change much,
so
the
alternative is an increase in height. Compared with
A.
robustus and
A.
boisei,
H.
erectus was taller and had a
narrower pelvis, the implications
of
which will be discussed
further (see pp. 10-1 1).
H.
erectus is considered to have a
morphology that was adapted to drier, more open environ-
ments, while Australopithecus could have inhibited open
dry
or closed wet environments. A modem analogy can be
drawn by comparing East African natives such as the tall,
slender Masai with the small, stocky ‘pygmy’ natives of the
Central African rainforest. Ruff (1991) comments: ‘It is
clear that
H.
erectus would not have inhabited a forested
environment, whether or not Australopithecus did’. A
change towards a higher-quality nutritional base may also
have a role in this change in body shape, since it was
accompanied by an increase in EQ (Table 1).
The skull and dental morphology of Australopithecus
provide good evidence that these hominids adapted early to
more arboreal niches, and that this adaptation was
reinforced with time.
A.
boisei is at least 2.3 million years
old, and thus must have diverged from a more gracile
Australopithecine ancestor relatively early (Skelton
&
McHenry, 1992). Australopithecus species from
A.
afar-
ensis to
A.
boisei have an increasing level of post-canine
megadontia; along with this comes larger jaws and stronger
muscles to move the jaw. The robust Australopithecines
developed prominent sagittal and nuchal crests for the
attachment of powerful chewing and neck muscles
respectively. They had bony struts in the face to withstand
the powerful chewing stresses set up through the massive
jaws and teeth. This type of molar enlargement and jaw
structure indicates that the robust Australopithecines were
chewing tough, fibrous vegetation (Rak, 1983; Ramirez-
Rozzi, 1993; Suwa et al. 1994; Johansen
&
Edgar, 1996).
In particular,
A.
boisei fossils found contemporaneously
with
H.
habilis are the most ‘robust’, with extremely large
jaws, teeth, zygomatic arches, and lateral ptyeroid plates.
Both molars and premolars are enormous compared with
incisors and canines, providing good evidence for a diet
including a high percentage of tough plant foods, with seed
crushing and nut cracking, etc. (Rak, 1983; Kimbel
&
Rak,
1993).
However, stable-isotope studies of remains from Swartk-
rans indicate that
A.
robustus was not completely
vegetarian (Lee-Thorp et
al.
1994). About 25-30% of
plants consumed by
A.
robustus were from C4 (initial
product
of
photosynthesis is a C4 molecule; grass) plants,
and
the
remaining 70-75
%
were from C3 (woody, broad-
leaf) plants. (C3 plants are the source of tubers, roots,
corms, leaves, nuts and fruit that provide the majority of
plant foods in the diet of hunter-gatherers observed
10
C.
L.
Broadhurst
et
al.
ethnographically.) It could not be determined whether 25-
30% of
C4
grass-type plants were consumed directly, or
whether the flesh of grass-eating animals was eaten instead.
Overall, dental morphology and a lack of grass phytolith
microwear on the fossil teeth support the latter case. Bone
Sr
:
Ca values provide independent evidence for the latter
case, as they fall between those for leopards and baboons,
and are unlike those of grazers (Schwarz
&
Schoeninger,
1991; Sillen
et
al.
1995). It is noted also that Hominoidea is
generally omnivorous and adaptable, as even chimpanzees
will occasionally hunt in groups, and often eat insects,
small mammals, and reptiles. Stable isotope studies are just
beginning to be a major tool for investigating the diets of
hominids and other extinct fauna, and may in the future
provide some insight into the proportion of aquatic foods
consumed by hominids.
Fish remains are associated with many early homind
sites, but since these sites tend to be near watercourses, the
fish bones have mostly been considered as background
noise. Stewart (1994) found that most fish faunal
assemblages were natural. Statistical evidence of fish-bone
size and frequency at Olduvai Beds I and I1 did yield some
discrepancies from natural scatter, which appear to indicate
sites where large numbers of fish were stranded and then
consumed by carnivores and/or hominids. The beds with
discrepancies are associated with
H.
habilis
and
A.
boisei
(contemporaneous species) and
H.
erectus
fossils and/or
artifacts.
Homo erectus
Homo
erectus
first appears approximately 1.2 million years
ago, and a lesser-know transitional species
H.
ergaster
dates from 1.8 million years ago. This time frame is
coincident with the first major continental glaciation in the
northem hemisphere (Shackelton
et
al.
1984). Due to polar
ice build up, the East African climate became significantly
hotter and drier, and there were massive local East African
extinctions. Those hardest hit were the wetter and more-
closed-habitat species, but the extinctions were not due
entirely to an increase in dry habitat. A number of open, dry
habitat genera did become extinct or locally extinct,
so
the
ecological change appears to be more complex (Harris,
1993; examples of
dry
open habitat genera reduction
between the Turkanan (2.5-1.0 million years ago) and
Natronian (1.0-0.15 million years ago) are as follows:
medium to very large herbivores dropped from thirty to
twenty-two, very small to small primarily-herbaceous
omnivores from twenty-nine to ten, and carnivores from
sixteen to eight).
H.
erectus
has been considered to have originated in
Africa, but the most recent interpretation of the fossil
record hints at an alternative explanation.
H.
ergaster
may
have been the common African ancestor of
H.
erectus
and
H.
heidelbergensis.
H.
ergaster
migrated from Africa to
Eurasia, and subsequently evolved into
H.
erectus,
which
was then an evolutionary ‘dead end’ (Tattersall, 1997).
African
H.
erectus
fossils are then reclassified as
H.
ergaster,
and possibly other transitional species.
A.
boisei
and
A.
robustus
did not become extinct until
about 0.4 million years later, but neither did they undergo
speciation events (or gradual change c.f. Wood
et
al.
1994)
as did Homo. Like others who have investigated this period
of prehistory, we consider this to be very significant,
transcending arguments based mainly on selective pres-
sures. As discussed by Foley (1994) and Vrba
et
al.
(1993,
climatic changes during the period of hominid evolution
can be fairly clearly and consistently related to extinction
events, but not necessarily to speciation events. Speciation
triggers in higher mammals are more complex than climatic
changes, and we believe that the role of brain-specific
nutrition is one of these complexities.
The time period about 16-1.7 million years ago saw
dramatic changes in hominid anthropometry. Female body
size increased dramatically, thus the marked sexual
dimorphism that characterized previous hominid species
diminished (McHenry, 1994). Birth canal enlargement and
related changes in the pelvic-femoral anatomy had not
been initiated or had not proceeded very far. However, the
modem pattern of infant secondary altriciality (extended
neonatal helplessness) and fetal growth was almost as fully
derived as that for modem humans (Ruff, 1991; Rosenberg,
1992).
McHenry (1994) recognized that a major change in the
hominid food supply must have been occumng: ‘Given the
energetic costs of brain size increases, this remarkable
change in brain size implies a major alteration in
subsistence’. However, McHenry (1994) did not suggest
specific food types, but rather hypothesized that greater
mobility by
H.
erectus
and
H.
ergaster
allowed access to a
wider range of environments and food sources. Similarly,
Foley
&
Lee (1991) concluded that about two million years
ago, the energetic costs imposed by increasing encephali-
zation would require both substantial quantities of higher-
quality foods and increased foraging efficiency. Foley
&
Lee (1991) note that the incorporation of 100-200
g
meat/kg in the hominid diet could have had a profound
evolutionary influence. We propose that at least some of
this ‘meat’, either scavenged or hunted, could have been
from fish and shellfish. Stewart (1994) found possible cut
marks on fish bones associated with
H.
erectus,
which
makes ‘a strong but not absolute case of early hominid fish
procurement’.
Homo sapiens
Periodic advances of continental glaciers continued
throughout the Middle and Upper Pleistocene (1.6 million
years to 10 thousand years ago), and severely affected
climate in the tropics as well as
in
higher latitudes. For
example, the tropical deep Atlantic Ocean cooled 4” on
average during the last glacial maximum (Schrag
et
al.
1996), and the cold ocean currents drew moisture-laden air
off the African continent, as was the case at the end of the
Miocene. As the climate became more arid, lake basins
shrank, rivers dried, and forest diminished. Paleosol
carbonate isotope data indicate that relatively pure savanna
grasslands similar to present conditions became established
about one million years ago (Cerling, 1992), just after the
last appearance of Australopithecus. The Rift Valley lakes
became virtually the only permanent sources of fresh water.
Fish provided brain-specific
nutrition
I1
In another geological setting, without the deep and
numerous Rift fault basins to retain water, a climate of
such aridity would not have permanent lakes (Leeder,
1995; Stager
et
al.
1997).
Despite the climate, around the lake margins Homo not
only survived, but again underwent a speciation event,
which featured astonishing cerebral expansion (Table 1).
During the period 500 thousand-200 thousand years ago,
cranial capacity expanded greatly, and the pelvic-femoral
complex characteristic of early Homo was replaced by the
modem anatomical complex, confirming that large fetal
crania, relatively difficult childbirth, and infant secondary
altriciality were present (Rosenberg 1992; Ruff, 1995;
Smith
et
al.
1995).
H.
erectus
appears in Europe, China, Java, and
possibly Siberia (Waters
et al.
1997) by
800
thousand
years ago. Nevertheless, the new species
H.
sapiens
is
thought to have originated in Africa between 100 and
300 thousand years ago (Stringer, 1992; Foley, 1994;
Johansen
&
Edgar, 1996; Swisher
et al.
1996; Tishkoff
et
al.
1996).
H.
sapiens
populations then migrated out of
Africa to the rest of the world about 120 thousand years
ago, rather than independent evolution of separate
H.
erectus
or other ancestral populations (Harrison 1993;
Lahr,
1994; Tishkoff
et al.
1996). In addition, a very young date
of 27 thousand to 53 thousand years ago for
H.
erectus
in
Java was recently reported (Swisher
et
al.
1996).
If
confirmed, this would require coexistence of
H.
erectus,
H.
neanderthalensis,
and
H.
sapiens
and would preclude
independent evolution of disparate
H.
erectus
populations
into
H.
sapiens.
Fully anatomically modem humans may have migrated
to the Middle East before 100 thousand years ago, but were
definitely widespread throughout Africa, Europe, and Asia
by forty thousand years ago (Schwarz
&
Grun, 1992). The
many controversies surrounding
H.
neanderthalensis
can-
not be discussed here (for a review, see Shreeve, 1995), but
there is general agreement that
H.
neanderthalensis
is a
different species which lived between 300 thousand and
30
thousand years ago.
H.
neanderthalenesis
coexisted with
H.
sapiens
for about
50
thousand years (Mercier
et al.
1991;
Stringer, 1992; Shreeve, 1995; Johansen
&
Edgar, 1996).
H.
neanderthalensis
apparently never developed the
sophisticated types of tools that
H.
sapiens
utilized,
including Aurignacian Industry stone tools, bone points,
projectile weapons, harpoon-type spear points, and fish-
hooks. The origin of
H.
neanderthalensis
may be analogous
with that of
H.
erectus
(Tattersall, 1997).
H.
heidelbergen-
sis
may have been the common African ancestor of
H.
neanderthalensis
and
H.
sapiens.
H.
heidelbergensis
migrated out of Africa to Europe and there evolved into
H.
neanderthalensis,
who also became an evolutionary
‘dead end’.
The earliest evidence for modem
H.
sapiens
is found in
Africa. Lake-shore sites in the Rift Valley have yielded
fairly sophisticated stone tools as old as 260 thousand years
ago associated with
H.
sapiens
remains. These Homo
skeletons have varying mixes of archaic and modem traits
(Clark, 1992). Two other African localities have yielded
early modem human remains associated with tools which
have not been found elsewhere until the Upper Paleolithic
(40 thousand to 10 thousand years ago). In both of these
cases, the ‘precocious’ cultures are associated with the
consumption of fish and shellfish.
At the Klasies River Mouth area, along the southern
coast of South Africa (Fig. l), a record of hominid
occupation for 60 thousand years beginning 120 thousand
years ago is recorded. Modern human fossils dating to
about 100 thousand years ago have been recovered in
Klasies River Mouth and Border Caves in the area. The
numerous occupation sites are littered with the shells of
mussels, turbans, and periwinkles (molluscs that can still be
picked up in abundance today). Some of the shells are
bumt, indicating the shellfish were cooked. The remains of
penguins, fur seals, eland, and small terrestrial mammals
have also been recovered; these relatively docile animals
were either hunted or scavenged (Deacon, 1992; Shreeve,
1995). Klein
&
Cruz-Oribe (1996) found that large,
dangerous game was not typically utilized. However, the
most recent analysis of Klasies River Mouth faunal remains
found that nearly one in five bones bears incisions from
butchery. Animal remains from heavily-fleshed body parts
are found in the caves with few signs of carnivore tooth
marks. In addition, a broken spear point tip was found in
the neck of an extinct giant buffalo, one of the largest game
animals in southern Africa at the time. Evidently these
Middle Stone Age humans could hunt efficiently and
collectively, perhaps driving game over cliffs or into pits
(Milo, 1997).
About 70 thousand years ago, a remarkably advanced
Upper Paleolithic-type tool technology (Howiesons Poort
Industry) emerged. The Howiesons Poort Industry features
sharp blades and projectile points which were hafted onto
shafts. Material for the points was not local, but came from
nearby areas with more suitable stone. Also about 70
thousand years ago, glaciers were advancing; the climate
became cooler and drier and.the coastline receded several
miles. One explanation for the Howiesons Poort Industry is
that superior tool technology was invented in order to adapt
to the harsher environment, which had less available for
scavenging and gathering, and required more hunting.
Shellfish remains are scarcer during this time (Deacon,
1992; Shreeve, 1995).
We would consider that regardless of the deteriorating
climate, the effects of 30 thousand to 50 thousand years of
brain-specific nutrition on the emergence of the technology
should not be ignored, especially since (1) the modem
human remains predate the Howiesons Poort Industry, and
(2) there is general agreement that the hominids must have
been travelling around and ‘taking note’ of superior stone
deposits before adopting the technology. Despite its
advancement, this culture was evidently not successful.
As time passed, the Howiesons Poort Industry petered out,
and was actually replaced by a less-sophisticated technol-
ogy more typical of elsewhere in the world. The area was
then abandoned 10 thousand years later until
50
thousand
years ago, when Upper Paleolithic hunters arrived and
began exploiting both marine and terrestrial food resources
with sophisticated hunting and fishing techniques (Shreeve,
1995; Klein
&
Cruz-Oribe, 1996).
At Katanda, in the Semliki River Valley, Zaire, there is
evidence that bone harpoon points were made as early as
12
C.
L.
Broadhurst
et
al.
100
thousand years ago, but such tools have not been
observed at other sites until 18 thousand years ago. The
harpoon points are associated with catfish and mollusc
remains, and one fragment
of
a human skull (Boaz
et
ul.
1992; Shreeve, 1995; Johansen
&
Edgar, 1996). It again
appears that a ‘precocious’ culture developed on an aquatic
(riverene and lakeshore) resource base, but remained
isolated for at least 40 thousand years. The explanations
for the precocious Katanda culture also invoke a response
to deteriorating climate, but could just as well incorporate
the influence of the aquatic brain-specific diet. Unlike the
Klasies River Mouth area, Katanda is a single site,
so
the
isolated culture remains anachronistic until confirmed at
other localities. Since archaeological exploration of the
general area
is
practically non-existent, the future may hold
some answers.
Upper Paleolithic
H.
supiens
in the East African Rift
Valley definitely utilized fish. Stewart (1989, 1994) des-
cribes numerous sites with vast fish faunal assemblages
dating from 40 thousand years ago to the present. Barbed
spear points, and evidence for fish-trapping dams and weirs
have also been recovered. The huge numbers of bones and
their distribution profiles provide evidence that foraging
H.
supiens
groups returned to certain areas year after year,
probably to take advantage of spawning runs or dry-season
fish stranding. Extensive use
of
diverse fish, shellfish, and
marine mammal and bird food resources is also recorded in
the Upper Paleolithic of South Africa (Buchanan, 1988). In
the following sections, we present research data indicating
that freshwater fish, shellfish, and similar lacustrine or
marine foods can and do provide brain-specific nutrition.
Finally, we will discuss the procurement
of
fish and other
animal food resources with limited technology, and
summarize our ideas.
Saturates and monounsaturates
16:O
16:ln-7 18:O
___+
18:ln-9
20:o
20:
ln-9 20:2n-9
1
5.
1
20:3n-9
-1
22:o 22:ln-9 22:3n-9
.1
1
24:O 24:ln-9
Long-chain polyunsaturated fatty acid composition
of
mammalian neural tissue
Unlike the protein- and mineral-rich musculo-skeletal
system, the major structural component of mammalian
neural tissue is lipid. The dry weight of the brain comprises
about 600 g lipid/kg, and has a unique profile of LC-
PUFA. PUFA are ‘essential’, which means that a portion of
the lipids which comprise the mammalian central nervous
system (CNS) cannot be synthesized and must come from
dietary sources. Precursor dietary essential PUFA in the
strict sense are linoleic acid (LA; 18
:
2n-6) and a-linolenic
acid (LNA; 18:3n-3; Fig.
3).
These CI8 PUFA are
alternately desaturated and elongated to form mainly Cz0
and CZ2 LC-PUFA with four or more double bonds. AA
(20
:
4n-6) and DHA (22
:
6n-3) are the main n-6 and
n-3
series end-member LC-PUFA respectively.
Good modem-day sources of LA are nuts (e.g. walnuts,
peanuts, pistachios, almonds, pumpkin seeds) and seed oils
(e.g. cotton, maize, sesame, sunflower, safflower, soya-
bean). Due to the widespread food use of agricultural oil
seeds, LA is far more prevalent in current diets than in the
past (Broadhurst, 1997). LNA is relatively uncommon in
modern diets, and is found in green leaves and walnuts as
well as flaxseed, mustard, rapeseed, and soyabean oils.
Foods richest in
AA
are egg yolk and organ meats and
muscle meats from land animals and tropical fish. Foods
richest in DHA and its precursor, eicosapentaenoic acid
(EPA; 20:
5n-3),
are marine fish and shellfish from cold
waters. Fish and shellfish from warmer marine or fresh
water have ubiquitous DHA and EPA; however, the content
of
AA
can also be high (Table 2).
In the forty-two mammalian species studied
so
far
(Crawford
et
al.
1976, 1992; Armstrong, 1983), the PUFA
Polyunsaturates
16:2n-6
1
1
1
1
1
-1*
18:2n-6(LA)
18:3n-6
20:3n-6
20:4n -6(AA)
22:4n-6
22:5n-6
16:3n-3
-1
CE
18:3f~-3(LNA)
ID
18:4n-3
-1
CE
1.
1
CE
20:4n-3
20:5n -3(EPA)
22:5n-3
l*
CE,D,RC
22:6n-3(DHA)
Fig.
3.
Long-chain fatty acid pathways in mammals. Palmitic acid
(16:
0)
is the starting point for saturates
(18:
0-24:
0)
and monounsaturates
(16: 1n-7-24: ln-9).
Linoleic acid
(18:2n-6;
LA)
and a-linolenic acid
(18:3n-3;
LNA) are the main precursors to the polyunsaturates
(n-3
and
n-6
polyunsaturated fatty acid series). Note also that the CI6
n-6
and
n-3
polyunsaturated fatty acids are present in the humandiet (i.e. selected
green vegetables) and can be metabolized to
18: 2n-6
and
18: 3n-3.
Pathways proceed through chain elongation (CE) and desaturation
(D).
The final step in polyunsaturated fatty acid metabolism involves retroconversion (RC) from CZ4 intermediates (*not shown). Preference by
desaturase enzymes is also shown
-b
>
-b
>
4.
EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid.
Fish
provided brain-specific
nutrition
13
Table
2.
Total fat content (g/kg) of representative fish and invertebrates, and arachidonic acid (AA) and
docosahexaenoic acid (DHA) in total lipid
(g/lOO
9). These fat ratios are typically not present in the terrestrial
food chain.
Fish and habitat
Lake Malawi African freshwater
Mbelele (catfish)
Njenu (carp)
Mfui (local sp.)
Kambale (local sp.)
Bream meat
Bream fat
Tropical marine
Australian barramundi
Indian halibut
Australian tropical freshwater
Freshwater temperate
Higher-latitude marine
Unspecified
Atlantic salmon (skinless)
Herring
Menhaden
N.
Atlantic mackerel
Temperate invertebrates
Mollusc
Squid
103 4.3 8.6
Pauletto et
a/.
(1 996a)
49 1
.a
7.8
11
8.0
19.1
18 5.9 13.3
16 5.3 5.6
Sinclair
(1992)
91
0
2.0 1.5
3 14.5 16.2
Mann
et
a/.
(1 995)
17 63 10.4
Pauletto
et
a/.
(1 996a)
(Oil)
3.3
8.0
lnnis
eta/.
(1995)
98' 1.2 17.2
Mann
et
a/.
(1995)
(Oil)
0.6 23.0
Cunnane
et
a/.
(1 993)
(Oil)
0.9 7.3
lnnis
etal.
(1995)
(Oil)
0.4
7.7
Pauletto
et
a/.
(1 996a)
>10 2.3 22.0
Cunnane etal.
(1993)
>10
5.8
21.3
'Salmon
with
skin contains
up
to
199
g
fat/kg.
content of brain ethanolamine phosphoglycerols is fairly
similar, and consistently dominated by
AA,
docosatetra-
enoic acid
(22:4n-6),
and DHA, with a
n-6:n-3
PUFA
value of
1-2:
1
(Table
3).
The
n-6: n-3
PUFA value in
most other cells is
3-5
:
1
(Horrobin,
1995),
and is variable,
depending on dietary intake and metabolic factors.
Mammalian brains all contain similar proportions of the
same basic phosphoglycerols; hence, the human brain
differs from the other mammalian species in a quantitative
rather than a qualitative sense (human brains are relatively
much larger, especially the frontal lobes, and have a more
sophisticated and diverse regional organization). While the
interspecies brain compositions are similar, the human EQ
is much larger (Table
l),
and humans devote a significantly
greater proportion of metabolic energy to the brain,
Table
3.
Mean polyunsaturated fatty acid composi-
tion of ethanolamine phosphoglycerols (g/lOO g) in
brain motor cortex grey matter of thirty-two mam-
malian species (Data from Crawford
et
a/.
1968,
1969, 1976)
Fatty acid
%
18:2 m6
20:3 n-6
20:4 m6
22:4 m6
20:5
m3
22:5 m3
22:6 m3
i8:3
n-3
0.9
1.7
15
a
0.3
0.9
2.3
21
especially neonatally (Martin,
1983;
Foley
&
Lee,
1991;
Cunnane
et
al.
1993).
The CNS is unique in not using the C18 precursors LA
and LNA, only the desaturated and chain-elongated LC-
PUFA.
Thus,
the mechanisms and efficiency
of
chain
elongation and desaturation as well as the dietary intake are
crucial for neural development. In humans, the intake of
preformed
AA
is a significant source of tissue
AA,
because
very little
AA
appears to be formed from LA in normal
individuals consuming mixed diets (Emken
et
al.
1992;
Mann
et
al.
1994).
This conversion is slow, especially when
compared with rats and mice, and can be impaired by many
physiological and pathological processes. When AA is
consumed it is readily incorporated into tissues (Whelan
er
al.
1992, 1993;
Mann
et
al.
1994).
Adam
et
al.
(1993)
found that at intakes normal for Western diets, LA does not
contribute to the formation of AA. Subjects on vegetarian
diets very low in preformed
AA
but with abundant LA
show correspondingly low
AA
levels in erythrocyte lipids.
Similarly, the conversion of LNA to EPA and finally to
DHA is slow and inefficient in many species (Cunnane,
1992;
Gerster,
1995).
LNA desaturation and chain elonga-
tion is especially weak in humans, and subject to
competition from
n-6
and
n-9
fatty acids (Fig.
3).
LNA
may not be converted to DHA and EPA in any significant
amount unless there has been a long-term deficiency of
n-3
PUFA, or if LNA levels are consistently low (Lands
et
al.
1991;
Sprecher,
1991;
Cunnane,
1992, 1995).
Caughey
et
al.
(1996)
found that
4
weeks of flaxseed oil supplementa-
tion increased plasma mononuclear cell LNA by 3-fold and
EPA by 2-3-fold, but did not raise DHA levels. However,
14
C.
L.
Broadhurst
et
al.
fish-oil supplementation for the following
4
weeks
dramatically increased both EPA and DHA. It must be
noted that although conversion of LA and LNA is thought
to be inefficient, it
is
variable among populations, and its
control is not fully understood. The majority of dietary LA
and LNA is likely to be oxidized or stored, not converted
(Cunnane
&
Anderson,
1997).
Abundant LC-PUFA is an absolute requirement for
advanced neural growth, and it
is
unlikely to be accidental
that the nutrient base of the Rift Valley lakes is an excellent
example of a rich dietary source of balanced, preformed
LC-PUFA (Table
2).
Long-chain polyunsaturated fatty acids during neural
growth and development
Research concerning fetal and infant growth and develop-
ment is relevant in considering the origin of human
intelligence overall. In order to sustain the rapid expansion
of the cerebral cortex, generation after generation of early
Homo must have had access to sources of abundant,
balanced PUFA, mostly probably in the form of AA and
DHA at a ratio of about
1
:
1.
In the sequential neural
development
of
infants, LC-PUFA deficiency during
critical growth periods results in irreversible failure to
complete components
of
brain growth. If LC-PUFA
are
the
limiting nutrients for the neural development of a
population, then EQ cannot increase much beyond that
observed in Ponginae, who have largely vegetarian diets.
Research with small mammals has also shown that with
increasing demands on maternal lactation, such as increas-
ing the number of pups in a litter, AA and DHA in the milk
are depleted faster than LA and LNA, which remain fairly
stable (Crawford
et
al.
1986).
Primates are noted for their
relative infrequence of multiple births, which may be a
prerequisite for brain expansion.
In some nutritional or ecological niches (i.e. savanna,
patchy woodland) LC-PUFA are greatly limited while
protein and minerals are not; subsequently body mass
increases greatly as compared with brain mass (Foley
&
Lee,
1991;
Cunnane
et
al.
1993).
This is the case in
herbivores, whose dietary PUFA consists entirely of LNA
and limited LA. The herbivore devotes a large proportion
of metabolic energy to grazing and digestion as opposed to
brain function. For example, a young rhino grows very fast,
reaching
1
tonne at age
4
years. The milk provided by its
mother has sufficient protein, minerals, LA and LNA, but
very little DHA and AA. The food supply onto which the
rhino is weaned also has no DHA or AA (Crawford
et
al.
1986;
Crawford
&
Marsh,
1995).
The rhino’s diet is
associated with a brain
:
body weight value
(g/100
g)
0.04,
as compared with
0.4
for the gorilla, and approximately
2
for
H.
sapiens.
Deposition of LC-PUFA in the CNS
is
rapid during
mammalian prenatal and postnatal brain growth (up to
approximately
18
months in humans), and is dependent in
part on the quantity and balance of fatty acids delivered by
the placenta prenatally and in the diet postnatally. The
stages of brain growth and maturation proceed in a fixed
temporal sequence (Innis,
1991).
The placenta does not
desaturate or chain elongate PUFA, but instead actively
concentrates and transfers to the fetus more AA and DHA
than LA or LNA. During the most active phase of fetal
growth, human brain development uses as much as
70
%
of
the energy delivered to it by the mother; postnatally, this
figure drops to
60
%,
and in adults is about
20
%
(Crawford,
1993).
Polyunsaturated fatty acid balance
and
long-chain
polyunsaturated fatty acid requirements
Maternal PUFA intake should be both adequate
(448%
total dietary energy intake) and balanced in order
to
ensure
a child’s normal CNS development. At
a
minimum, it is
considered that the
n-6
:
n-3
PUFA value should be kept in
the range
5
:
1-1
:
1
from conception to age
2
years. Human
milk normally contains both precursor and LC-PUFA with
a
n-6
:
n-3
value of
4-5
:
1.
However, human milk contains
less
LC-PUFA if they are chronically deficient in the
maternal diet (i.e. vegetarians; Holman
er
al.
1991;
Simopoulos,
1996).
The lack of LC-PUFA in infant formulas is of deep
concern, because the infant has a restricted ability to utilize
LA and LNA (Carlson
et
al.
1993a;
Nettleton,
1995).
Adult
humans differ greatly in their ability to elongate and
desaturate LA and LNA, and although full-term infants can
synthesize AA and DHA (Camielli
et
al.
1996),
it is not
known whether this occurs efficiently enough to accom-
modate their rapid brain growth (Farquharson
et
al.
1992).
The presence
of
quantitatively significant amounts
of
preformed LC-PUFA in breast milk and in the CNS
suggests that these metabolites are indeed essential for the
neonate. Preterm infants are in effect still dependent on the
placenta and definitely cannot utilize LA and LNA
effectively (Crawford,
1993;
Simopoulos,
1996).
An imbalance favouring
n-6
or
n-3
PUFA, or lack
of
LC-PUFA is a potentially serious problem for fetuses,
infants, and growing children. Infant formulas are supple-
mented with significant LA (from several vegetable oils),
and lesser amounts of LNA (usually from soyabean oil).
Infant formulas in some countries such as Japan now
contain supplemental
n-3
LC-PUFA, particularly to provide
DHA. The decision to supplement all formulas with AA
and DHA is still under heated scientific scrutiny (Makrides,
1997);
however, it is generally agreed that formulas for
premature neonates need such supplementation (Crawford,
1993;
Simopoulos,
1996;
Hansen
et
al.
1997).
Formulas supplemented with marine fish oil have been
utilized in animal studies and for preterm human infants,
and
do
significantly increase plasma and tissue DHA
(Arbuckle
et
al.
1991;
Carlson
et
al.
1992;
Clandinin
et
al.
1992;
Innis
et
al.
1994).
However, the very high EPA and
DHA and low AA contents of marine fish oils (Table
2)
pose a risk for infants, because AA levels in tissues decline
to the point where development is affected (Carlson
et
al.
1992, 1993~).
AA requirements are highest during early
postnatal growth, and EPA in particular competes with
n-6
PUFA for desaturation and chain elongation, thereby
interfering with AA production.
Fish provided
brain-specific
nutrition
15
In a situation where dietary PUFA are both restricted in
quantity and unbalanced, the optimum AA
:
DHA in the
infant brain cannot be maintained (Farquharson
et
al.
1992). Supplementing formulas with both
AA
and DHA
normalizes neurodevelopment in full-term infants (Agos-
toni
et
al.
1995; Makrides
et
al.
1995; Gibson
et
al.
1997)
and eicosanoid production in neonatal pigs (Huang
&
Craig-Schmidt, 1996), and has no reported negative effects
on growth or neurocognitive development. Innis
et
al.
(1995) compared the effect of freshwater-fish oil (EPA
:
AA
1.8) on growing rats as compared with cold-water marine
fish oil (EPA
:
AA 16.8). In the brain and other organs, the
freshwater-fish oil generally increased n-6 PUFA (includ-
ing
AA)
while slightly decreasing but still maintaining n-3
PUFA levels. Both AA and DHA were high, indicating that
freshwater-fish oil can prevent the decline in AA imposed
by marine-fish-oil feeding.
Tropical
fish
diets: current examples
Modern analogues for pre-agricultural diets based on
tropical marine and lacustrine resource bases exist, and
have been shown to be exceptionally healthy. Stewart
(1989, 1994) describes modern ethnographic and personal
observations of East African traditional fishers. As the dry
season ends and the rainy season begins, some hunter-
gatherer groups follow the fish spawning migrations.
During the
dry
season, fish became stranded in progres-
sively smaller pools as waters recede, and hunter-gatherers
will camp for extended periods near swampy lowlands to
take advantage of the naturally-high fish concentration.
These fishing societies have been observed personally by
one of us while working as a ‘bush doctor’ in East Africa
(M.A.C; c.f. Crawford
&
Marsh, 1995), and their habits and
excellent health noted.
Fish rich in fat, especially catfish, are highly prized by
these traditional fishers. Most of the fish caught today are
catfish, with the remainder mainly cichlids. Like other
tropical freshwater and marine fish, catfish have a relatively
high AA:DHA, which would favour brain expansion
(Table 2). Catfish also comprise over 90% of the fish
fauna recovered from over forty Late Pleistocene Nile
River sites described by Stewart (1989, 1994). Other sites
have more equal proportions of catfish, cichlids (known as
Tilapia
spp.), and a large minnow-like fish, Barbus.
Near the end of the dry season, game is relatively scarce,
and the fat content of game is exceedingly low (10-
40 g/kg; Crawford
et
al.
1976; Speth, 1989; O’Dea, 1991;
Crawford
&
Marsh, 1995). Plant foods are scarcest during
this time also. Fish are preparing to spawn at the first rains
of the season, and may actually have an increase in fat
content at the end of the dry season. A diet based on scarce,
lean game may be severely deficient in both protein and fat
(Speth, 1989). As will be discussed further (see p. 16),
scavenged mammalian bone marrow and brain tissue have
been proposed as options for fat procurement, especially
during times of climatic stress (Blumenschine, 1991;
Blumenschine
&
Cavallo, 1992; Bunn
&
Ezzo, 1993).
Similarly, Cunnane
et
al.
(1993) and Stewart (1994)
proposed that fish and shellfish may have served as an
important source of protein, trace minerals and fat.
Pauletto
et
al.
(1996a,b) investigated two groups of
genetically-homogeneous native Bantus near Lake Malawi
(Nyasa), Tanzania. One group lives on the lakeshore and
consumes a fish-based mixed diet
(FD,
n
622). The other
group lives approximately
75
km
from the lake, and
consumes a grain-based vegetarian diet (VD,
n
686). Both
FD
and VD eat strictly locally-available foods, as processed
foods are not available.
The proportions of plasma n-3 LC-PUFA were three to
four times higher in
FD
v.
VD. Plasma
AA
was higher in
FD
v.
VD, despite the fact that VD had overall higher levels
of
n-6
PUFA. Most of the VD dietary and plasma n-6
PUFA was in the form of LA. The
FD
Bantus consume
300-600 g freshwater fish daily throughout most of their
lives; however, the percentage of total dietary energy intake
from fat is approximately 12, the majority of which is
derived from the fish (see Table
2).
Despite a higher
cholesterol and saturated fat intake, their blood cholesterol
and triacylglycerols were lower than those of the VD
consuming an exceedingly-low-fat (approximately 7
%
total dietary energy from fat) vegetarian diet. Overall,
FD
had significantly lower mean blood pressure, total choles-
terol, triacylglycerols and lipoprotein than VD. Before the
very recent introduction of agriculture, these two groups
would have existed on wild plants alone or on wild plants
plus fish. The lakeside group has a clear advantage in terms
of cardiovascular fitness, high-quality protein intake, and
brain-specific nutrition. If we imagine the isolation of these
groups from one another for one million years or
so,
perhaps we may gain a glimpse of the divergent pathways
taken by coexisting Homo and Australopithecus.
Aborigines in northwestern Australia traditionally con-
sume a diet rich in tropical coastal fish. Typical marine-fish
diets lower plasma
AA
significantly; however, this tropical-
fish diet increased
AA
by 3-fold and EPA and DHA by 2-
fold as compared with controls (O’Dea
&
Sinclair, 1982).
Although tropical and subtropical fish species have
ubiquitous DHA and EPA, the content of
AA
can also be
high (Table 2).
In order to investigate the aetiology of type I1 diabetes in
native populations, diabetic and non-diabetic Aborigines
were put on three diet trials (Sinclair, 1993):
(1)
40
%
total
dietary energy from fat with
75
%
total dietary energy from
beef; (2) 20
%
total dietary energy from fat with 80
%
total
dietary energy from coastal tropical seafood; (3)
13
%
total
dietary energy from fat, with 85-87
%
total dietary energy
from kangaroo meat, freshwater fish, and yams. The
kangaroo meat consumed was only 10-20 g fat/kg wet
weight since it was wild game. Most of the fat in wild game
is structural phospholipid, containing significant LC-PUFA
but quantitatively low amounts
of
saturated fat. Diets 2 and
3 were traditional diets, low in LA and saturated fatty acids,
and produced marked improvements in the metabolic
abnormalities associated with diabetes, and a reduction in
cardiovascular disease risk factors, including hyperlipidae-
mia and blood pressure. Plasma phospholipids in diets 2
and 3 had roughly equal values for AA:EPA:DHA.
Prostacyclin activity was estimated
in
vivo,
and tropical-
fish and kangaroo diets had evidence for high activity. High
16
C.
L.
Broadhurst
et al.
activity was not seen, in comparison, in the cold-water-
marine-fish (plasma AA being approximately one-fifth of
EPA plus DHA) or vegetarian (AA, DHA and especially
EPA much lower) diets.
In summary, the lipid profile of tropical and subtropical
freshwater fish and other aquatic species have a DHA
:
AA
that is closer to that in brain phospholipids than any other
food source known. These edible species are found in the
East African Rift Valley Lakes around which
H.
sapiens
arose and eventually dominated. This food source also
provides abundant protein, and is known to be a healthy
diet for humans. Humans in the Rift Valley have a tradition
of utilizing the lake resources, especially as a source of
dietary fat.
Soils and surface rocks in the Rift Valley area were and
still remain relatively rich in trace elements due to the
constant volcanism and uplift (Baker
et
al.
1972; Bailey
&
Macdonald, 1987; Dawson, 1992). Although the details are
beyond the scope of the present discussion, we note
particularly that the Rift Valley environment provides
abundant Zn, Cu, and
I,
trace elements necessary for PUFA
metabolism and for normal brain development and function
(Cunnane
et
al.
1993). Hence, the enormous proto-oceanic
lakes provided a plentiful, protected source of brain-
specific nutrition, even as the climate in the area changed
dramatically in the Pleistocene. As discussed later (below),
accessing the lake resource base was not strictly dependent
on the seasons or the intellectual status of the evolving
hominids.
Cultural considerations, scavenging, and 'low-tech' fishing
It is not necessary to invoke organized fishing or hunting by
early Homo; in fact the origin of our intellect was probably
the more humble occupation of grabbing small cold-
blooded creatures and scavenging. While hominids may
have eaten some fish directly, it is important to realize that
the freshwater fish and shellfish provide a major link in the
broader food chain. They are consumed by birds, small
mammals, reptiles, amphibians, etc. all of which in turn
could have been consumed by hominids (although perhaps
in the form of eggs). For example, Leakey (1971)
postulated that tortoise shells found without other skeletal
pieces in bed at Olduvai were evidence for consumption by
hominids.
Stewart (1989, 1994) described procurement of fish
without sophisticated technology. Hyenas, leopards, canids,
and (anecdotally) baboons have been documented to pull
fish
from the water and eat them. During the spawning runs,
catfish and
Barbus
move into very shallow waters and can
be clubbed, speared, or picked up bare-handed. East
African cichlids typically inhabit shallow, slow-moving
waters, especially when spawning, and are very territorial
(Riehl
&
Baensch, 1986; Loiselle, 1988). They often return
to the same shallow-water nesting areas year after year.
Modem fishers have been observed marking these spots for
future reference and fish capture. During the
dry
season, as
lake and stream water levels recede, large numbers of fish
become stranded in shallow pools or concentrated in
lowland areas. Fish stranding is particularly common in the
Rift Valley due to its unique, highly-variable fault-
controlled interior drainage system. Stranded fish may be
scavenged after death, or again, clubbed, speared, or picked
up live; fishing tackle is not required. Fish in the central,
deeper waters of major rivers are difficult to procure
without sophisticated equipment, including hooks, lines,
baskets, weirs, dams, and nets, and were probably not a
nutritional option until
40
thousand years ago. (It should be
noted that the present warm waters and high alkalinity of
the Rift Valley Lakes has resulted in impoverished fish
faunal diversity. Only a few families of 'hardy' fish are
represented. Some Rift Valley cichlids have adapted to live
in water nearly devoid of dissolved
02,
and at pH up to 10.5
and temperatures of 40" (Riehl
&
Baensch, 1986; Loiselle,
1988; Johnson
et
al.
1996).)
Scavenging of the remains of larger carnivore kills is
also a logical possibility for hominids, and could also have
provided protein and some AA and DHA if the opportunity
arose. The initiation of meat scavenging has been proposed
as a causative factor of the dramatic increase in EQ about
two million years ago (Speth, 1989; Foley
&
Lee, 1991).
We would agree that scavenging played a role, but would
broaden the scavenging resource base to include fish,
shellfish, reptiles, etc. in addition to game, particularly
since the proportion of fat in game meat is low.
Cheetahs and leopards often leave ample meat on their
kills which is then available for both primary and secondary
scavengers. Large cats have been observed to leave
carcasses unattended for many hours at a time (Blu-
menschine, 1991; Blumenschine
&
Cavallo, 1992; Bunn
&
Ezzo, 1993). The size of the carcass correlates positively
with the length of time it is left unattended. Scavenging
hyenas typically leave a carcass stripped, but prefer open or
lightly-wooded habitats. Carnivore kills in densely-wooded
areas, such as lake and river margins, were found to be less
likely to be scavenged by hyenas than were open-land kills
(Blumenschine, 1987). Scavenging was likely to have been
more prevalent in the
dry
season, when other resources are
scarcest, and dietary fat is at a premium. Scavenging kills
brought down by the savanna pursuit carnivores is not as
difficult as hunting, but is still fairly dangerous, and
requires skillful, intelligent observation of the environment
and the behaviour patterns of other animals (Blumenschine
&
Cavallo, 1992).
Scavenging hominids could have used bone-crushing
tools to extract bone marrow and brains. There is no
absolute proof of tool making by Australopithecus, but they
probably utilized opportunistic tools such as crushers and
digging sticks. The Oldowan tradition (2.4 to two million
years ago) of simple flaked tools first appears in strata
where
H.
habilis
and
A.
boisei
remains are contempora-
neous. There are no Oldowan sites with
A.
boisei
fossils
alone. After 1.8 to two million years, the Oldowan tradition
continued to develop, and the Aechulian tradition (approxi-
mately 1.4 million years ago) arose, but between 1.5 and
one million years ago, Australopithecus became extinct.
The Aechulian tradition is clearly associated with
H.
erectus,
although it has also been found with
H.
habilis
in
Olduvai and Sterkfontein (Clark, 198.5, 1992; CONOY,
1990; Shreeve, 1995). Oldowan and Early Aechulian tools
Fish
provided
brain-specific nutrition
17
are sufficient for extracting brains and marrow, and crude
butchering of carrion, but probably not for organized
hunting.
Culture, speech, and tool use are not prerequisites for the
expansion of the cerebral cortex, but rather result from
expansion. Based on the EQ evidence, it cannot be assumed
that early hominids possessed the creativity and hand-eye
coordination to manufacture or use a variety of tools which
have yet to be discovered. Organized hunting with effective
weapons is mainly an Upper Paleolithic phenomenon (forty
to ten thousand years ago), but rather than postulate a
vegetarian diet until this time, it is considered that
scavenging occurred, despite a lack of compelling archae-
ological evidence. Statistical evaluation of damage done to
faunal remains recovered at or near hominid localities has
found that, in a small minority of cases, cut marks from
stone tools can be identified. Most marks on the bones are
from carnivore teeth (Blumenschine, 1991; Bunn
&
Ezzo,
1993; Selvaggio, 1994). An analysis of faunal remains from
Olduvai Bed I1 also found that most bone cuts were done by
carnivores; however, when hominids were involved, it
appeared they were acquiring scavenged carcasses long
before all the meat was gone,
and
had a preference for
larger game. They were evidently eating more than just
bone marrow, irrespective
of
their ability to actively hunt
(Monahan, 1996).
However, the logic applied to terrestrial fauna has not
been transferred to aquatic foods. Since sophisticated
fishing tackle, harpoons, and fishhooks
are
usually not
found up to eighteen thousand years ago, it has been
assumed that earlier humans and hominids did not eat fish.
While they may not have actively fished, they may well
have eaten fish. Fish scavenging, shellfish gathering, etc., is
not possible in every environment, but is very plausible in
the unique East African Rift Valley, and in other areas of
Africa with long histories of hominid occupation.
Homo proceeded from an opportunistic tool user to a
premeditated tool maker, and ultimately to a fine craftsman
and organized hunter, with astonishing rapidity. EQ
increased significantly in
H.
habilis,
and even more
dramatically in
H.
erectus
(Table 1). It is difficult to
account for this in an evolutionary sense without at least
considering the sources of brain-specific nutrients available
to early Homo (Cunnane
et
al.
1993). Grabbing or trapping
fish and crustaceans by hand, and smashing mollusc shells
requires less sophistication than either hunting or scaven-
ging game, yet yields a far greater amount of preformed
DHA and AA for the effort. Modem fishers often smash
fish crania with rocks or sticks in order to extract the brains.
Based on the large number of fish cranial fragments found
in Olduvai Level 3, fish-skull crushing could have been
done by hominids also (Stewart, 1994).
It is not suggested by the relatively crude nature of the
archaeological evidence that
H.
habilis
or
H.
erectus
posed
a serious threat to the subsistence base of the existing
savanna pursuit carnivores. While early Homo may have
scavenged meat and bones, it is only the internal organs and
their associated fat depots, and brains of game that could
have provided a consistent, concentrated source of
preformed LC-PUFA on a par with freshwater fish. The
organs and depot fat are parts
of
a carcass that are likely to
be consumed first by carnivores and primary scavengers,
and may not have been consistently available. We note also
that high consumption of internal organs can have a price; a
H.
erectus
skeleton shows pathological changes indicative
of retinol hypervitaminosis, probably caused by too much
animal liver (Walker
et
al.
1982), or possibly honeybee
larvae (Skinner, 1991), although fish liver is an equally
plausible source. This leaves only brains and bone marrow,
which require the use of tools to extract. Since tool use is a
result of cerebral cortex expansion rather than a cause, this
begs the question of which factors were most responsible
for the initiation and expansion of the hominid intellect.
The Lake Malawi Bantu eat 9188 kJ/d with 23
%
dietary
energy from fish and 12
%
dietary energy from fat, most of
which is from the fish (Pauletto
et
al.
1996a,b);
12%
dietary energy equals 29 g fish oil. Half or even one-quarter
of this intake is still about 7-15 g fish oil/d, a much higher
intake than the majority of the population obtains today.
Fish intakes of the order of half or one-quarter of 23%
dietary energy intake are 6-12
%
of the diet, easily within
the 10-20 % range for evolutionary influence proposed by
Foley
&
Lee (1991). Working within the framework of
known ecological influences on evolution and prodigious
expansion of the cerebral cortex, it is in fact difficult to
argue that very moderate freshwater-fish intakes could not
have affected hominid evolution.
Conclusions
African Rift Valley lake margins provided a unique source
of brain-specific nutrition, namely abundant freshwater fish
and shellfish providing LC-PUFA.
If
this resource base was
consistently exploited by hominids, it could have helped
provide a means for rapid, sustained cerebral cortex
enlargement without
an
attendant increase in body mass.
We recognize that this enlargement is based on a pre-
existing primate genetic capacity for relatively high
intelligence, and is subject to numerous selective pressures
and cultural reinforcements. However, we believe that the
role of abundant brain-specific nutrient deserves considera-
tion in the past and current evolution of Homo. Although
H.
erectus
evidently migrated to Eurasia, the weight of
evidence from very diverse anthropological arguments
points to a single speciation event in Africa which produced
H.
sapiens.
H.
sapiens
populations then migrated out of
Africa to the world approximately 120 thousand years ago,
rather than independent evolution of separate
H.
erectus
populations (Harrison, 1993;
Lahr,
1994; Tishkoff
et
al.
1996; Tattersall, 1997).
In this case we suggest that the same basic lacustrine
environment was responsible for the successful evolution
from
H.
habilis
to
H.
sapiens.
Perhaps we could not
successfully emerge from the African Lake Cradle until we
became intelligent enough to adapt to very diverse
environments. We hypothesize that part of this adaptation
to diverse environments involves ensuring that the food
supply for males, females, and children contains adequate
PUFA. Accessing rich sources of dietary fat, especially LC-
PUFA, may have been unconscious or opportunistic at first,
but eventually became conscious and desirable.
18 C. L. Broadhurst
et al.
Additionally, faunal evidence from Lainyamok,
Kenya
indicates that in some
areas
of
East Africa, extant
mammalian taxa
were
present
as
far back
as
390
thousand
years
ago
(Potts
&
Deino, 1995).
Foley
(1994) and Vrba
et
al.
(1995) concluded that while climatic
change
may
be
clearly related to extinction, it is rarely if
ever
clearly
related to the creation
of
species. This indicates that
climatic changes
alone
120
thousand
years
ago
may
be
too
simplistic
an
explanation for migration out of Africa, and
the origin
of
modem
H.
sapiens.
However, we
would
agree
that increasing population, coupled with deteriorating
climate
and
extreme
lake lowstands in the Pleistocene
almost certainly influenced migration from the Rift Valley,
and
the adoption of organized hunting.
Postulating that LC-PUFA
are
limiting nutrients for
human
brain
evolution
leads
to the prediction that
chronically-inadequate
LC-PUFA
nutrition will result in
suboptimal brain development in both individual
cases
and
in populations
as
a
whole.
There
is
good
evidence today
that lack
of
abundant, balanced DHA
and
AA
in
utero
and
infancy leads to
lower
intelligence quotient
and
visual
acuity (Crawford
et
al.
1992;
Carlson
et
al.
19936;
Cunnane
et
al.
1993; Nettleton,
1995;
Simopoulos 1996),
and in
the
longer-term contributes to clinical depression
(Hibbeln
&
Salem,
1995;
Adams
et
al.
1996)
and
attention-
deficit hyperactivity disorder (Stevens
et
al.
1995).
We
are
not
so
far removed from
our
Paleolithic ancestors that
we
can
expect
our
present agricultural, processed-food-based
diet to provide indefinitely for our continued intellectual
development.
References
Abell PI (1982) Paleoclimates at Lake Turkana, Kenya, from
oxygen isotope ratios of gastropod shells.
Nature
297,
321-323.
Adam
0,
Taxacher G, Lemmen C, Wiseman M
&
Schattenkirch-
ner M (1993) Lowering of eicosanoid formation, amelioration
of clinical symptoms and laboratory findings in patients with
rheumatoid arthritis on a vegetarian diet. In
Omega
3
Fatty
Acids: Metabolism and Biological Effects,
pp. 285-291 [CA
Drevon,
I
Baksaas and HE Krokan, editors]. Basel: Birkhauser
Verlag.
Adams PB, Lawson
S,
Sanigorski A
&
Sinclair AJ (1996)
Arachidonic acid to eicosapentanoic acid ratio in blood
correlates positively with clinical symptoms of depression.
Lipids
31,
S157-S161.
Agostoni C, Trojan
S,
Bellu R, Riva E
&
Giovannini M (1995)
Neurodevelopmental quotient of healthy term infants at 4
months and feeding practice: the role of long-chain poly-
unsaturated fatty acids.
Pediatric Research
38,
262-266.
Andrews PJ (1989) Paleoecology of Laetoli.
Journal of Human
Evolution
18,
173-1 8
1.
Arbuckle LD, Rioux
FM,
MacKinnon MJ, Hrboticky N
&
Innis
SM (1991) Response of (n-3) and (n-6) fatty acids in piglet
brain, liver, and plasma to increasing, but low, fish oil supp
lementation of formula.
Journal of Nutrition
121,
1536-1547.
Armstrong E (1983) Relative brain size and metabolism in
mammals.
Science
220,
1302-1304.
Bailey DK
&
Macdonald R (1987) Dry peralkaline felsic liquids
and carbon dioxide flux through the Kenya rift zone. In
Magmatic Processes: Physiochemical Principles,
pp. 9 1-1 19
[BO Mysen, editor]. University Park, PA: The Geochemical
Society.
Baker BH, Mohr PA
&
Williams LAJ (1972) Geology of the
Eastern Rift System of Africa.
Geological Society of America
Special Paper
no. 136.
Behrensmeyer AK (1975) Taphonomy and paleoecology in the
hominid fossil record.
Yearbook of Physical Anthropology
19,
36-50.
Behrensmeyer AK, Potts R, Plummer T, Tauxe L, Opdyke N
&
Jorstad
T
(1995) The Pleistocene locality of Kanjera, Westem
Kenya: stratigraphy, chronology, and paleoenvironments.
Journal
of
Human Evolution
29,
247-274.
Bestland EA, Thackray GD
&
Retallack GJ (1995) Cycles of
doming and eruption in the Miocene Kisingiri Volcano,
southwest Kenya.
Journal of Geology
103,
598-607.
Blumenschine RJ (1987) Characteristics of an early hominid
scavenging niche.
Current Anthropology
28,
383406.
Blumenschine RJ (1991) Hominid carnivory and foraging
strategies, and the socio-economic function of early archae-
ological sites.
Philosophical Transactions
of
the Royal Society,
London
B
334,
211-221.
Blumenschine RJ
&
Cavallo JA (1992) Scavenging and human
evolution.
Scient$c American
267,
90-96.
Boaz NT, Bemor
RL,
Brooks AS
et al.
(1992). A new evaluation
of the significance of the Late Neogene beds, Upper Semliki
Valley, Zaire.
Journal of Human Evolution
22,
505-517.
Boaz NT, Pavlakis PP, McDonnell M
&
Gatesy J (1994)
Environments in the Late Pliocene of eastern Zaire: the Senga
13B excavation.
Research and Exploration
10,
124-127.
Bohannon RG, Naeser CW, Schmidt DL
&
Zimmerman RA
(1989) The timing of uplift, volcanism, and rifting peripheral to
the Red Sea: a case for passing rifting?
Journal of Geophysical
Research B
94,
1683-1701.
Bonnefille R (1983) Evidence for a cooler and drier climate in
the Ethiopian uplands towards 2.5 Myr ago.
Nature
303,
487492.
Bonnefille R, Vincens
A
&
Buchet G (1987). Palynology,
stratigraphy, and paleoenvironment of a Pliocene hominid site
(2.9-3.3 m.y. at Hadar, Ethiopia.
Paleogeogmaphy, Paleocli-
mutology, and Paleoecology
60,
249-28 1.
Broadhurst CL (1997) Balanced intakes of natural triglycerides for
optimum nutrition: an evolutionary and phytochemical
perspective.
Medical Hypotheses
49,
247-26 1.
Brunet M, Beauvilain
A,
Coppens
Y,
Heintz
E,
Moutaye AHE
&
Pilbeam D (1995) The first australopithecine 2,500 kilometers
west of the Rift Valley (Chad).
Nature
378,
273-275.
Buchanan WF (1988)
Shell$sh in Prehistoric Diets. Cambridge
Monographs in African Archaeology
no. 3 1,
BAR International
Series
no. 455. Oxford: BAR.
Bunn HT
&
Ezzo JA (1993) Hunting and scavenging by Plio-
Pleistocene hominids: nutritional constraints, archeological
patterns, and behavioral implications.
Journal of Archeological
Science
20,
365-398.
Carlson SE, Cooke
RJ,
Werkman SH
&
Tolley EA (1992) First
year growth of preterm infants fed standard compared to
marine-oil n-3 supplemented formula.
Lipids
27,
901-907.
Carlson
SE,
Werkman SH, Peeples JM, Cooke RJ
&
Tolley
EA
(1993a) Arachidonic acid status correlates with first year
growth in preterm infants.
Proceedings
of
the National
Academy of Sciences USA
90,
1073-1077.
Carlson SE, Werkman
SH,
Rhodes PG
&
Tolley EA (19936)
Visual-acuity development in healthy preterm infants: effect
of
marine oil supplementation.
American Journal of Clinical
Nutrition
58,
3542.
Carnielli VP, Wattimena DJL, Luijendijk IHT, Boerlage
A,
Degenhart HJ
&
Sauer PJJ (1996) The very low birth weight
Fish provided brain-specific nutrition 19
premature infant is capable of synthesizing arachidonic and
docosahexaenoic acids from linoleic and linolenic acids.
Pediatric Research
40,
169-171.
Caughey GE, Mantzioris E, Gibson
RA,
Cleland LG
&
James MJ
(1996) The effect on human tumor necrosis factor a and
interleukin lp production of diets enriched in
n-3
fatty acids
from vegetable oil or fish oil.
American Journal of Clinical
Nutrition
63,
116-122.
Cerling TE (1992) Development of grasslands and savannas in
East Africa during the Neogene.
Paleogeography, Paleoclima-
tology, and Paleoecology
97,
241-247.
Clandinin MT, Garg
ML,
Parrott A, Van Aerde J, Hervada A
&
Lien E (1992) Addition of long-chain polyunsaturated fatty
acids to formula for very low birth weight infants.
Lipids
27,
896-900.
Clark JD (1985) Leaving no stone untumed: archaeological
advances and behavioral adaptation. In
Hominids Evolution:
Past, Present, and Future,
pp. 65-88 [P Tobias, editor]. New
York: Alan Liss.
Clark JD (1992) African and Asian perspectives on the origins of
modem humans.
Philosophical Transactions of the Royal
Society, London
B
337,
201-215.
Conroy
GC
(1990)
Primate Evolution.
New York W.W. Norton.
Conroy GC
&
Kuykendall
K
(1995) Paleopediatrics: or when did
human infants really become human?
American Journal
of
Physical Anthropology
98,
121-131.
Coppens Y (1994) East side story: the origin of human kind.
Scientijc American
271,
88-95.
Crawford MA (1993) The role of essential fatty acids in neural
development: implications for perinatal nutrition.
American
Journal of Clinical Nutrition
57,
703s-710s.
Crawford MA, Casperd
NM
&
Sinclair
AJ
(1976) The long chain
metabolites of linoleic and linolenic acids in liver and brains of
herbivores and carnivores.
Comparative Biochemistry and
Physiology
54B,
395401.
Crawford MA, Costeloe K, Doyle W, Leaf A, Leighfield MJ,
Meadows N
&
Phylactos A (1992) Essential fatty acids in early
development. In
Polyunsaturated Fatty Acids in Human
Nutrition,
pp. 93-1 10
[U
Bracco and RJ Deckelman, editors].
New York: Raven Press.
Crawford MA, Gale MM
&
Woodford MH (1969) Linoleic acid
and linolenic acid elongation products in muscle tissue of
Syncerus caffer
and other ruminant species.
Biochemical
Journal
115,
25-27.
Crawford MA
&
Marsh D (1995)
Nutrition and Evolution.
New
Canaan, CT: Keats Publishing.
Crawford MA, Patterson
J
&
Yardley L (1968) Nitrogen
utilisation by the Cape Buffalo
(Syncerus caffer)
and other
large mammals.
Zoological Society of London Symposium no.
Crawford MA
&
Sinclair AJ (1972) Nutritional influences in the
evolution of the mammalian brain. In
Lipids, Malnutrition and
the Developing Brain. Ciba Foundation Symposium
pp. 267-
292 [K Elliot and J Knight, editors]. Amsterdam: Elsevier.
Crawford MA, Williams G
&
Hassam AG (1986) Do develop-
mental periods
of
high demand outstrip the rate of desaturation?
A reduction in milk arachidonic acid with successive, short
birth intervals.
Progress in Lipid Research
25,
413415.
Cunnane SC (1992) What is the nutritional and clinical
significance of a-linolenic acid in humans? In
Essential Fatty
Acids and Eicosanoids: Invited Papers from the Third
International Conference,
pp. 379-382, [A Sinclair and
R
Gibson, editors]. Champaign,
IL:
AOCS Press.
Cunnane SC (1995) Metabolism and function of a-linolenic acid
in humans.
In Flaxseed in Human Nutrition,
pp. 99-127 [SC
Cunnane and LU Thompson, editors]. Champaign,
IL:
AOCS Press.
21,
pp. 367-379.
Cunnane SC
&
Anderson MJ (1997) The majority of dietary
linoleate in growing rats is B-oxidized or stored in visceral fat.
Journal of Nutrition
127,
146152.
Cunnane SC, Harbige LS
&
Crawford MA (1993) The importance
of energy and nutrient supply in human brain evolution.
Nutrition and Health
9,
219-235.
Dawson
JB
(1992) Neogene tectonics and volcanicity in the north
Tanzania sector of the Gregory rift valley: contrasts with the
Kenya sector.
Tectonophysics
204,
81-92.
Dawson
JB,
Pyle DM
&
Pinkerton H (1996) Evolution of a
natrocarbonatite from a wollastonite nephelinite parent: evi-
dence from the June 1993 eruption of Oldoinyo Lengai,
Tanzania.
Journal of Geology
104,
41-54.
Deacon HJ (1992) Southern Africa and modem human origins.
Philosophical Transactions of the Royal Society, London
B
337,
Dowsett H, Thompson R, Barron J, Cronin T, Fleming
F,
Ishman
S,
Poore R, Willard D
&
Holtz T Jr (1994) Joint investigations
of the Middle Pliocene climate
I:
PRISM paleoenvironmental
reconstructions.
Global and Planetary Change
9,
169-195.
Emken
RA,
Adlof RO, Rohwedder WK
&
Gulley
RM
(1992)
Comparison of linolenic and linoleic acid metabolism in man:
influence of dietary linoleic acid. In
Essential Fatty Acids and
Eicosanoids: Invited Papers from the Third International
Conference,
pp. 23-25 [A Sinclair and R Gibson, editors].
Champaign,
IL:
AOCS Press.
Farquharson J, Cockbum F, Patrick WA, Jamieson EC
&
Logan
RW (1992) Infant cerebral cortex phospholipid fatty-acid
composition and diet.
Lancet
340,
810-813.
Foley
RA
(1994). Speciation, extinction, and climatic change in
hominid evolution.
Journal of Human Evolution
26,
275-289.
Foley RA
&
Lee PC (1991) Ecology and energetics of
encephalization in hominid evolution.
Philosophical Transac-
tions of the Royal Society, London
B
334,
223-232.
Gerster H (1995) The use of
n-3
PUFA’s (fish oil) in enteral
nutrition.
International Journal of Vitamin and Mineral
Research
65,
3-20.
Gibson
RA,
Makrides M, Neumann M, Hawkes J, Pramuk K, Lien
E
&
Euler A (1997). A dose response study of arachidonic acid
in formulas containing docosahexaenoic acid in term infants.
Prostaglandins, Leukotrienes, and Essential Fatty Acids
57,
198 Abstr.
Hansen J, Schade D, Harris C, Merkel
K,
Adamkin
D,
Hall R
&
Lim M (1997). Docosahexanoic acid plus arachidonic acid
enhance preterm infant growth.
Prostaglandins, Leukotrienes,
and Essential Fatty Acids
57,
196 Abstr.
Harris
J
(1993) Ecosystem structure and growth of the African
savanna.
Global and Planetary Change
8,
231-248.
Harrison
T
(1993). Cladistic concepts and the species problem in
hominoid evolution. In
Species, Species Concepts and Primate
Evolution,
pp. 345-371 [WH Kimbel and LB Martin, editors].
New York: Plenum Press.
Harrison T (1994) Paleoanthropological exploration in the
Manoga Valley, Tanzania.
Research and Exploration
10,
Harvey PH
&
Clutton-Brock TH (1985) Life history variation in
primates.
Evolution
39,
557-581.
Hibbeln
JR
&
Salem N (1995) Dietary polyunsaturated fatty acids
and depression: when cholesterol does not satisfy.
American
Journal of Clinical Nutrition
62,
1-9.
Holman RT, Johnson SB
&
Ogburn PL (1991) Deficiency
of
essential fatty acids and membrane fluidity during pregnancy
and lactation.
Proceedings of the National Academy
of
Sciences
USA
88,
48354839.
Horrobin DF (1995) Abnormal membrane concentrations of
20 and 22-carbon essential fatty acids: a common link between
177-183.
238-240.
20 C. L. Broadhurst
et a1
risk factors and coronary and peripheral vascular disease?
Prostaglandins, Leukotrienes, and Essential Fatty Acids
53,
Huang M-C
&
Craig-Schmidt MC (1996) Arachidonate and
docosahexaenoate added to infant formula influence fatty acid
composition and subsequent eicosanoid production in neonatal
pigs.
Journal
of
Nutrition
126, 2199-2208.
Innis SM (1991) Essential fatty acids and growth and
development.
Progress in Lipid Research
30, 39-103.
Innis SM, Lupton B
&
Nelson CM (1994) Biochemical and
functional approaches to study of fatty acids requirements for
very premature infants.
Nutrition
10, 72-76.