Content uploaded by Thure E Cerling
Author content
All content in this area was uploaded by Thure E Cerling on Jan 25, 2021
Content may be subject to copyright.
Nature © Macmillan Publishers Ltd 1997
NATURE
|
VOL 389
|
11 SEPTEMBER 1997 153
articles
Global vegetation change through
the Miocene/Pliocene boundary
Thure E. Cerling*, John M. Harris†, Bruce J. MacFadden‡, Meave G. Leakey§, Jay Quadek,
Vera Eisenmann¶& James R. Ehleringer#
*Department of Geology and Geophysics, University of Utah, Salt Lake City, Utah 84112, USA
†George C. Page Museum, 5801 Wilshire Boulevard, Los Angeles, California 90036, USA
‡Florida Museum of Natural History, University of Florida, Gainesville, Florida 32611, USA
§The National Museums of Kenya, PO Box 40658, Nairobi, Kenya
kDepartment of Geosciences, University of Arizona, Tucson, Arizona, USA
¶Laboratoire de Pale
´ontologie, Musee
´National d’Histoire Naturelle, 8 rue de Buffon, 75005 Paris, France
#Department of Biology, University of Utah, Salt Lake City, Utah 84112, USA
............................................................................................................................................................................................. ...........................................................
Between 8 and 6 million years ago, there was a global increase in the biomass of plants using C
4
photosynthesis as
indicated by changes in the carbon isotope ratios of fossil tooth enamel in Asia, Africa, North America and South
America. This abrupt and widespread increase in C
4
biomass may be related to a decrease in atmospheric CO
2
concentrations below a threshold that favoured C
3
-photosynthesizing plants. The change occurred earlier at lower
latitudes, as the threshold for C
3
photosynthesis is higher at warmer temperatures.
The C
3
and C
4
photosynthetic pathways fractionate carbon isotopes
to different degrees; C
3
and C
4
plants have d
13
C values ranging
from about −22‰ to −30‰ and −10‰ to −14‰, respectively1–5.
(Isotopic ratios are reported relative to the isotopic standard PDB,
where d
13
C (in ‰) = ½ð13C=12 CÞsample=ð13C=12 CÞstandard 21ÿ31;000Þ.
So the carbon isotope composition of fossil tooth enamel reflects
the C
3
/C
4
composition of mammalian diet6–9. Soil organic matter
preserves this isotopic distinction with little or no isotopic frac-
tionation (,2‰), but both soil carbonate and carbonate in
biogenic apatite from large mammals are significantly enriched in
13
C compared to source carbon7,8,10. Cerling et al.9have studied fossil
soils and tooth enamel from Pakistan and North America and
concluded that C
4
ecosystems underwent a global expansion
between about 7 and 5 million years (Myr) ago in the late Miocene
and early Pliocene epochs; they suggested that atmospheric CO
2
concentrations could have fallen below a threshold critical to C
3
photosynthesis. However, others11–13, also studying fossil soils and
tooth enamel, concluded that there was not a global expansion of C
4
biomass in the late Miocene, and that there was no link between
changes in C
3
/C
4
biomass and atmospheric chemistry. Hill13 sug-
gested that observed dietary changes in Africa ,7 Myr ago need not
signify a vegetation change, but may be explained by faunal
immigration or in situ speciation.
Here we report the results of stable carbon isotope analyses from
more than 500 equids and other hypsodont (that is, having high-
crowned teeth) large mammals from Asia, Africa, North America,
South America and Europe. Our studies emphasized equids because
modern equids are thought to be predominantly grazers, and equids
are abundant in the fossil record. We also analysed other hypsodont
large mammals because the record of equids is more limited in
Europe, Asia, Africa and South America than it is in North America.
Thus, from Europe, Asia and Africa we also report the results of
analyses of fossil proboscideans (elephants and their allies), and
from South America those of notoungulates (an extinct order of
endemic South American mammals). There are several advantages
of using mammalian teeth rather than soils for indications of C
4
biomass; first, identification is straightforward, and second,
mammals enhance the isotope signal by selective feeding. In most
cases C
4
diets are indicators of grazing although some C
4
dicots (for
example, Chenopodiaceae) can be important components of the
diets of certain mammals, especially in regions having saline soils.
C
3
grasses today are important in some ecosystems, so that a C
3
diet
does not necessarily indicate a browsing diet.
We first show that bioapatite (tooth enamel) in large mammals is
,14‰ enriched in
13
C compared to their diet. Using this discri-
mination factor we then show that large mammals with ages greater
than 8 Myr from a global population all had diets compatible with a
pure C
3
,orC
3
-dominated, diet. We then show that by 6Myr ago
equids and some other large mammals from low latitudes (,378)
had a C
4
-dominated diet in Africa, South America, North America
and southern Asia. No evidence is found suggesting a significant C
4
component in the diets of large mammals from western Europe at
any time. Comparison of the quantum yields of C
3
and C
4
mono-
cots, which are primarily grasses and sedges, indicates that C
4
monocots are favoured at atmospheric CO
2
concentrations less
than 500 parts per million by volume (p.p.m.v.) when accompanied
by high growing-season temperature. The persistance of significant
C
4
biomass beginning about 6–8 Myr ago and continuing to the
present is compatible with atmospheric CO
2
levels in the late
Miocene declining below the ‘crossover’ point where C
4
grasses
are favoured over C
3
grasses or other C
3
plants.
C
3
- and C
4
-dominated diets
The unambiguous detection of the presence of the C
4
signal is an
important issue. C
3
plants have a considerable range in d
13
C; water-
stressed ecosystems are enriched in
13
C (as high as −22‰) com-
pared to the average C
3
value of about −27‰, whereas closed
canopies are depleted in
13
C, having values as low as −35‰ (refs 5,
14, 15). C
4
plants have a much more restricted d
13
C range, where
plants using the NADP-me and NAD-me sub-pathways have
average d
13
C values of about −11.4‰ and −12.7‰, respectively16.
The isotopic fractionation between diet and bioapatites (such as
tooth enamel) is not well established for large mammals. After
reaction with H
3
PO
4
and cryogenic purification, samples were
reacted at 50 8C with silver wool to remove trace amounts of SO
2
gas which was occasionally identified in both modern and fossil
samples; trace amounts of SO
2
in CO
2
can result in positive
13
C
shifts greater than 4‰ (unpublished data). Table1 shows the results
of analyses from the hypergrazer alcelaphine bovids (hartebeest and
wildebeest) from Kenya that have a diet of NADP-subpathway
grasses (about −11.4‰; ref. 16), and from restricted feeders from
the Hogel Zoo in Salt Lake City, Utah with a diet of meadow hay
Nature © Macmillan Publishers Ltd 1997
and alfalfa (−26.5‰). These results indicate an isotope fractiona-
tion factor for both C
3
and C
4
diets in large mammals: a
enamel-diet
is
1.0143 to 1.0148, or d13Cenamel 2d13Cdiet <14:3‰, where aenamel-diet ¼
ð1;000 þdenamelÞ=ð1;000 þddiet Þ. This enrichment in
13
C of ,14.3‰
for tooth enamel in large mammals compared to their diet is greater
than observed in laboratory experiments on very small mammals
(mice)17. Therefore a d
13
C value for enamel of −8‰ would corre-
spond to a dietary intake of −22‰ to −22.5‰, which is within the
range of observed pure C
3
ecosystems and plants1,2,5,14. Water stress
or high light conditions (or both) causes an enrichment of
13
C in C
3
plants5,14 so that −8‰ for the d
13
C of enamel can be taken as
conservative ‘cut-off’ value to exclude the possibility of a ‘false
positive’ indicating a significant C
4
biomass in diet. For fossil
samples, yet another correction should be considered: Friedli and
others18 and Marino and McElroy19 have shown that the d
13
C of the
atmosphere and plants, respectively, have become 1.5‰ more
negative in the past 150 years because of fossil-fuel burning.
Therefore, the ‘cut-off’ for a pure C
3
diet may be even more positive,
perhaps even −7‰. Others11 have used a ‘cut-off’ of −10.5‰ for
tooth-enamel d
13
C values to indicate significant C
4
biomass, which
we believe is too
13
C-depleted for the reasons discussed above.
We analysed 226 different mammals (bovids, camelids, equids,
proboscideans, rhinocerids, suids, tapirids) older than 8 Myr and
find no evidence for a significant C
4
component in diets of
mammals from Europe, Africa, Asia, or the Americas. The average
d
13
C value for this suite was 210:661:3‰ and only a single
sample gave d13C. 2 8‰ (d13C¼27:5‰). These data are
compatible with all the animals having diets from −22‰ to about
−28‰, with an average diet of −25‰ which is in the range of the
carbon isotopic composition for modern C
3
plants. Figure 1 shows
the d
13
C values for 825 modern plants; also shown are d
13
C for tooth
enamel for 309 modern mammals, and 226 fossil mammals with
ages older than 8 Myr. The modern mammals show a distinction
between C
3
-dominated and C
4
-dominated diets, and the .8-Myr
mammals indicate diets compatible with an essentially pure C
3
diet.
The d
13
C of the primary dietary signal is preserved in the fossil
record in tooth enamel and does not seem to be affected by
diagenesis. This is illustrated in Fig. 2 where we show the d
13
C
values for east African deinotheres (elephant-like ungulates of the
order Proboscidea), other proboscideans, and equids through the
past 20 million years. Deinotheres always have d
13
C values consis-
tent with a pure C
3
diet, whereas the equids and proboscideans have
d
13
C values consistent with a C
3
diet before 8 Myr, but consistent
with a C
4
-dominated diet after about 7 Myr. The deinotheres were
collected from the same sedimentary deposits as the other fossils. In
addition, palaeosols and other sedimentary carbonates from the
Koobi Fora (Kenya) sequence20 have d
13
C values intermediate
between the d
13
C values for C
3
and C
4
endmembers.
C
4
ecosystem development in Neogene times
The striking change from C
3
to C
4
ecosystems was first noted in
palaeosol carbonates in the Siwalik sediments of Pakistan21 which
showed a change in d
13
C starting about 7 Myr ago with values
averaging about −10‰ and reaching about 0‰ byabout 5 Myr ago.
This can be compared to the record of equid and proboscidean
tooth enamel from the same time interval (Fig. 2). These data show
a significant C
4
component in both the equid and proboscidean diet
between 8 and 7 Myr, but that the C
4
endmember diet was not
reached until about 5 Myr (perhaps as early as 6 Myr). The transi-
tion begins at about 7.8 Myr using the palaeomagnetic timescale of
Cande and Kent22, or 7.3Myr using the older palaeomagnetic
timescale of Berggren23. Equids first appear in the Pakistan sequence
about 10.5 Myr, the time of the ‘Hipparion datum’ and become
widespread throughout much of Europe, Africa and Asia. Notably,
the earliest equids in the Siwalik sequence have a C
3
-dominated diet.
East Africa has an abundant fossil record of proboscideans and
equids. We report data from Maboko, Fort Ternan, the Turkana
basin and the Suguta depression, and include in our discussion
previously published data from the Baringo basin11. Both elephan-
tids and equids changed from a C
3
-dominated diet to a C
4
-
dominated diet between about 8 and 7 Myr, while deinotheres
retained a C
3
-dominated diet (Fig. 2). Equids appear in east
Africa by about 10 Myr, and in two sites older than 9 Myr equids
have a C
3
diet. The equids have transitional diets at about 8 Myr as
recorded in the Samburu Hills, and have largely adapted to a C
4
-
dominated diet by the time of the oldest sediments in the Lothagam
sequence, estimated to be about 7.5 Myr. Elephantids show a similar
pattern although they seem to lag the equids in making the
transition to a C
4
-dominated diet.
The South American record was sampled from deposits in
Argentina and in southern Bolivia. Equids entered South America
very late, so notoungulates were also included in our analysis. The
ages of the samples in Fig. 2 are based primarily on the South
American Land Mammal Ages (SALMA), although several samples
are also included from well-dated Neogene (Pliocene + Miocene)
deposits in northern Argentina24. Securely dated notoungulates
have a C
3
diet before 8 Myr, but show evidence for a significant C
4
articles
154 NATURE
|
VOL 389
|
11 SEPTEMBER 1997
-30 -20 0-10
δ13C (‰)
>8 Myr mammalian
tooth enamel
(
n
= 226)
δ13C = -10.6 ± 1.3 ‰
-10-30 -20 0
-20 -10 0
-30
C4-dominated
diet
C3-dominated
diet
Modern mammalian
tooth enamel
(
n
= 309)
C3 grasses
δ13C = –26.7 ± 2.3 ‰
C4 grasses
δ13C = -12.5 ± 1.1 ‰
1.5 ‰
atmospheric shift
δenamel-δdiet
~14 ‰
δenamel-δdiet
~14 ‰
Modern grasses
Figure 1 Histograms of d
13
C for modern grasses (compiled from refs 5–8, and
unpublished University of Utah data), modern tooth enamel, and fossil tooth
enamel .8 Myr. The d
13
C for modern and fossil mammalian enamel is from
Europe, Asia, Africa, North America and South America, and represents all
samples analysed in our laboratory of these age intervals. Fossil samples include
bovids, equids, giraffids, notoungulates, proboscideans and rhinocerids. The
d
13
C axis of fossil tooth enamel is shifted by an additional 1.5‰to adjust for the
anthropogenic shift in d
13
C in the atmosphere (and therefore diet and enamel)
resulting from fossil-fuel burning18,19.
Nature © Macmillan Publishers Ltd 1997
component (−4.8‰) by 7.6 Myr (ref. 24).
We divided the North American data in a ‘low-latitude’ (,378N)
and a ‘high-latitude’ group (.378N), 378N represents a convenient
dividing line placed at the northern boundaries of Oklahoma, New
Mexico and Arizona, and is the approximate boundary between the
southern Great Plains and the northern Great Plains. We report data
from more than 300 fossil equids from North America. The low-
latitude group includes samples from Mexico, Florida, Texas,
Oklahoma, New Mexico, Arizona and southern California. It
shows a significant isotopic change in the late Hemphillian. All
sites with ages older than 7 Myr have d
13
C values between −8‰ and
−15‰, but are as high as −2.7‰ at 6.8 Myr at Coffee Ranch in
northern Texas. Late Hemphillian sites in Mexico have d
13
C values
up to +1.7‰ by 5.7 Myr. Equids in the low-latitude region of North
America show considerable scatter in the d
13
C values, probably
indicating a reliance on both C
3
and C
4
grasses possibly during
different times of the year.
High-latitude sites from North America included Alaska, north-
ern California, Idaho, Nebraska, Nevada, North Dakota, Oregon,
South Dakota, Washington and Wyoming. Equids from these sites
consumed a smaller fraction of C
4
biomass than did the low-latitude
equids. This is to be expected because of the lower abundance of C
4
grasses in northern North America compared to southern North
America25. Of the high-latitude sites, only those in Nebraska show
significant C
4
biomass in the diet where one d
13
C enamel value
reaches −3‰.
European sites in Fig. 2 are from Spain and France, between 388
and 488N. Neither equids nor proboscideans show any evidence of a
significant C
4
biomass in their diets at any time during the past
20 Myr. This is consistent with the dominance of C
3
plants in the
region today, and agrees with data obtained from additional
samples of equids and other ungulates from the eastern Mediterra-
nean (for example, Samos, Pikermi, Pasalar)26,27 and from Morocco
and Algeria in North Africa (unpublished data). These data suggest
that C
4
plants have not been a significant component of the biomass
in western European or Mediterranean ecosystems at any time.
There is now evidence from four different widely separated
regions (Pakistan, East Africa, low-latitude North America, and
South America) for a significant expansion of C
4
biomass between
about 8 and 6 Myr. All samples older than about 8Myr have d
13
C
values between −8‰ and −15‰, yet by 6.8 Myr regions have at least
some d
13
C values that indicate a C
4
-dominated diet (.−4‰), and
reach d13C<0‰ by about 5 Myr or earlier. Meanwhile, fossil
hypsodont herbivores from high-latitude North America show a
subdued increase of C
4
biomass in their diets, whereas those from
Europe exhibit no increase. The pattern of dietary change with
latitude (Fig. 3) in equids and proboscideans is compatible with
conditions that would favour C
4
biomass in hotter regions, but
conditions that also promote C
4
biomass expansion simultaneously
in widespread parts of the globe. Figure 3 shows that we sampled
both the northern and southern limbs of the C
3
/C
4
transition, which
is between about 258and 408latitude in both hemispheres. The high
variability in the C
4
component of diet in such intermediate
latitudes may be the result of several factors, such as the variability
in growing season for different regions (for example, Mediterranean
climates at about 358have fewer C
4
plants than monsoonal climates
at the same latitude), variability in C
4
biomass during different parts
of the growing season (for example, spring versus summer con-
ditions) or long-term climate fluctuations (for example, glacial
versus interglacial). Equatorial sites show low isotopic variability for
equid diets (Fig. 3).
Faunal change in the latest Miocene
The period in the late Miocene and Pliocene when we have
identified significant change in diet was also a period of worldwide
faunal change. Significant faunal turnover is observed in Pakistan,
articles
NATURE
|
VOL 389
|
11 SEPTEMBER 1997 155
North
America
( >37 °N)
Equid
Deinothere
Elephantid
Equid
East Africa
15 10 0
5
20
Age (Myr)
Western
Europe
Deinothere
Elephantid
Equid
15 10 0
5
20
Age (Myr)
0
5
-5
-10
-15
Pakistan
Proboscidean
Equid
15 10 0
5
20
Age (Myr)
Equid
Notoungulates
South
America
radiometric
Equid
North
America
(< 37 °N)
0
5
-5
-10
-15
15 10 0
5
20
Age (Myr)
15 10 0
5
20
Age (Myr)
15 10 0
5
20
Age (Myr)
0
5
-5
-10
-15
δ
13
C (‰)
0
5
-5
-10
-15
SALMA
δ
13
C (‰)
δ
13
C (‰)
δ
13
C (‰)
Figure 2 Changes in d
13
C of equid and some other hypsodont mammals in the
Neogene. Although most of the data in this are new, we also include previously
published data from Pakistan8,9,41
, North America42, South America24,43, and
Africa11,36. South American samples in bold are samples from a well-dated site24
and the others43 are shown as the average age according to their respective
South American Land Mammal Age (SALMA)44.
Nature © Macmillan Publishers Ltd 1997
North America, South America, Europe and Africa. There has been
debate as to whether such changes were in response to local climate
change, immigration or other factors13,28, but it is now clear from
stable carbon isotope studies that an important global ecological
change was underway at this time.
In Pakistan, many woodland-adapted mammals were replaced by
more open-habitat representatives between 8 and 7 Myr (refs 29,
30). Tragulids are replaced by hypsodont artiodactyls, and true
giraffes appear in the post-7.5 Myr assemblages, along with hippo-
potamid species30. After 7.4 Myr, local assemblages are dominated
by hypsodont ungulates. Among the primates, Sivapithecus (a large-
bodied hominoid) and lorisids became extinct in Asia between 8
and 7 Myr ago, their place eventually being taken by cercopithecids
(Old World monkeys) that appeared in the latest Neogene31. Late
Miocene changes among the small mammals include extinction of
dormice, and the appearance of more open-adapted advanced
rhyzomyids and hares31.
In North America, equids reached their maximum diversity in the
middle Miocene but their diversity was greatly reduced in the
Hemphillian (late Miocene and earliest Pliocene, or about 7 to
4.5 Myr ago)32,33. Camelids, antilocaprids, palaeomerycids and
gomphotheres were likewise greatly reduced in diversity during
this interval. In general, the more hypsodont lineages from these
families were favoured in the Pliocene. This Hemphillian episode of
extinction was the most severe to be documented in the North
American Neogene, exceeding in extent the late-Pleistocene extinc-
tion event32.
East African mammal faunas showed a marked shift in their
community structure during the Neogene34,35. Early Miocene mam-
malian faunas in east Africa had a tropical-forest character with
common taxa including hominoids, hyraxes, suids, rhinos and
proboscideans. The Pliocene witnessed a sharp increase in season-
ality with the faunas evolving a savanna-mosaic character. Grazing
antelopes and hippos replaced chevrotains and anthracotheres as
the dominant artiodactyls. Among the perissodactyls, three-toed
equids replaced the browsing rhinos and hyraxes. High-crowned
elephantids replaced bunodont long-jawed gomphotheres. Mon-
keys underwent a major radiation, replacing the diverse early and
middle Miocene hominoid assemblage. During the terminal Mio-
cene, open wooded-grassland habitats replaced the earlier less
seasonal woodland/forest habitats; the Lothagam fauna seems to
be transitional between the archaic earlier Miocene and the
advanced Plio-Pleistocene faunas36.
It is now clear that the expansion of C
4
grasses was a global
phenomena beginning in the late Miocene and persisting to the
present day. It was accompanied by important faunal changes in
many parts of the world. It is not likely that the expansion of C
4
biomass in the late Miocene is due solely to higher temperature or to
the development of arid conditions. There have always been some
parts of the Earth with hot, dry climates yet it seems that the C
4
expansion was triggered by a single phenomenon as this expansion
occurred simultaneously in widespread regions of the world that
were separated by oceans (for example, the Old World, South
America, North America). Significantly, C
4
plants in the cooler
parts of the planet did not respond as effectively as in the hotter
regions. Thus the C
4
expansion is not documented in the enamel of
equids and proboscideans from the late Miocene and early Pliocene
of western Europe, although by the lower Ruscinian of Europe
hipparions (that were so abundant in the Miocene) have virtually
disappeared, probably because of changing climate conditions37.
Quantum yields, temperature and atmospheric CO
2
Plant metabolism responds directly to atmospheric CO
2
concentrations38,39. C
3
plants respond to changes in atmospheric
articles
156 NATURE
|
VOL 389
|
11 SEPTEMBER 1997
Table 1 d
13
C values for modern mammals and their diet
d
13
C (‰) Name Name
Wild grazers, Athi plains, Kenya (year of death,1969): estimated d13 Cdiet ;211:4‰*
.............................................................................................................................................................................
3.1 Coke’s hartebeest Alcelaphus buselaphus cokii
.............................................................................................................................................................................
1.9 Coke’s hartebeest Alcelaphus buselaphus cokii
.............................................................................................................................................................................
3.0 Coke’s hartebeest Alcelaphus buselaphus cokii
.............................................................................................................................................................................
3.2 Coke’s hartebeest Alcelaphus buselaphus cokii
.............................................................................................................................................................................
3.8 Coke’s hartebeest Alcelaphus buselaphus cokii
.............................................................................................................................................................................
2.9 Wildebeest Connochaetes taurinus albojubatus
.............................................................................................................................................................................
3.9 Wildebeest Connochaetes taurinus albojubatus
.............................................................................................................................................................................
3.2 Wildebeest Connochaetes taurinus albojubatus
.............................................................................................................................................................................
3.9 Wildebeest Connochaetes taurinus albojubatus
.............................................................................................................................................................................
3.7 Wildebeest Connochaetes taurinus albojubatus
.............................................................................................................................................................................
3:260:6 (d13Cenamel 2d13 Cdiet ¼14:6)‡
.............................................................................................................................................................................
Hogel Zoo animals: estimated d13 Cdiet 226:5‰(n¼5)†
.............................................................................................................................................................................
−12.8 African elephant Loxodonta africana
.............................................................................................................................................................................
−12.7 African elephant Loxodonta africana
.............................................................................................................................................................................
−13.7 Bactrian camel Camelus bactrianus
.............................................................................................................................................................................
−12.9 Bactrian camel Camelus bactrianus
.............................................................................................................................................................................
−13.0 Giraffe Giraffa camelopardalis
.............................................................................................................................................................................
−12.0 Pigmy hippopotamus Choeropsis liberiensis
.............................................................................................................................................................................
−12.8 Pigmy hippopotamus Choeropsis liberiensis
.............................................................................................................................................................................
−12.0 Zebra Equus burchelli grantii
.............................................................................................................................................................................
−12.0 Zebra Equus burchelli grantii
.............................................................................................................................................................................
−12.4 Zebra Equus burchelli grantii
.............................................................................................................................................................................
212:660:5 (d13Cenamel 2d13 Cdiet ¼13:9‰)‡
.............................................................................................................................................................................
Hartebeest and wildebeest of East Africa are hypergrazersand have d
13
C values 1– 2‰more
positive than zebra of the same year of death and from the same location (unpublished data).
The large mammals fromthe Hogel Zoo in Salt Lake City, Utah, all have a diet that is primarily
meadow hay.
*The average d
13
C of C
4
grasses using the NADP sub-pathway is −11.4‰(ref.16). C
4
grasses
in the Athi plains predominantly use this subpathway.
†Five samples of meadow hay and alfalfa pellets collected in 1991 give
d13C¼226:560:7‰. Year of death for these animals was between 1980 and 1990.
‡The ,14.3%difference between diet and enamel is compatible with the data in Bocherens
et al.50
60 30 0 30 60
Equus
South North
Latitude (degrees)
5
0
-5
-10
-15
δ13C (‰)
Figure 3 d
13
C of modern and fossil Equus versus latitude (all samples below
2,000 m elevation); we include data from Thackarey and Lee-Thorp45 and data
from Fig. 2. The equatorial dominance of C
4
grasses, the transition to C
3
grasses
in intermediate latitudes (30– 408), and the dominance of C
3
grasses at high
latitudes (,.458), can be seen.
Nature © Macmillan Publishers Ltd 1997
CO
2
with decreased maximum net photosynthetic rates that are
related to lowered CO
2
levels because of both inherent CO
2
substrate limitations and higher photorespiration rates. C
4
plants
are less sensitive to atmospheric CO
2
levels.
The quantum yield (photosynthetic efficiency) of C
3
grasses
relative to C
4
grasses varies with both atmospheric CO
2
levels and
temperature (Fig. 4). The crossover point favouring C
3
over C
4
grasses is dependent on temperature and partial pressure of CO
2
(p
CO2
) such that C
4
-dominated ecosystems are favoured under low
p
CO2
conditions when accompanied by elevated temperature. The
modern world is a ‘C
4
-world’ where C
4
plants make up an important
biomass in tropical, sub-tropical and some temperate ecosystems.
When atmospheric CO
2
levels are high, above about 500 p.p.m.v.,
the C
3
photosynthetic pathway would be favoured in all conditions
except those with extremely high temperatures. The ‘C
3
-world’,
where C
4
plants do not make up a significant fraction of the biomass
even in tropical regions, is predicted to have been the more
productive pathway from the origin of terrestrial vascular plants
at about 400 Myr ago until the late Miocene between 8 to 6Myr
when the ‘C
4
-world’ became established. A possible exception to
this could have been during the late-Carboniferous to Permian
glaciation if p
CO2
levels were low enough and some plants indepen-
dently evolved the C
4
pathway, which was subsequently lost in the
Mesozoic when CO
2
levels were again high.
The model of Fig. 4 explains some interesting features of the
temporal change in diets of mammals shown in Fig. 2 and in the
spatial change in diets shown in Fig. 3. The change from C
3
to C
4
diet in equids occurred somewhat earlier in tropical regions than in
higher latitudes. In East Africa (38S to 58N), the transition is very
rapid and is complete between 8 to 7.5 Myr; in Pakistan (32–338N)
the transition is slightly more gradual and occurs between about 7.8
and 6 Myr; in southern North America (20– 378N) it takes place
between 6.8 and 5.5 Myr; in central North America (Nebraska; 40–
438N) the oldest sample analysed so far with a definite C
4
signal is
about 4 Myr and no samples have d
13
C values above −2‰; in
western Europe (between 408and 508N), there is no indication of
a C
4
diet at any time. This is compatible with a history where the
‘crossover’ of quantum yields favouring C
4
plants over C
3
plants was
reached first at low latitudes, and at later times at successively higher
latitudes (because of lower temperature) as atmospheric CO
2
levels
declined during the Neogene. It further implies we are unlikely to
find evidence for widespread C
4
plants in periods of the Earth’s
history where p
CO2
was higher than about 500 p.p.m.v.
The modern spatial pattern of C
4
and C
3
grasses shows that C
4
grasses dominate in tropical and subtropical regions, that the
transition to C
3
grasses takes place between about 308and 458
latitude, and that C
3
grasses dominate at high latitudes (Fig. 3).
Figure 4 shows that the crossover for the modern atmospheric level
of CO
2
(280 p.p.m.v. for the pre-industrial value of CO
2
) is between
16 8C and 20 8C (daytime growing-season temperature), with C
3
grasses being favoured in cooler regions (such as high latitudes and
high altitudes).
This model is compatible with gradually decreasing CO
2
in the
atmosphere during the Tertiary, and crossing a threshold important
to C
3
photosynthesis near the end of the Miocene. Changes in
atmospheric CO
2
levels are related to continental weathering;
increased weathering rates during the past 40 Myr, especially in
the tectonically active Himalayan –Tibetan region, have resulted in a
lowering of CO
2
(ref. 40). The culminating effect is a world where C
3
plants are increasingly starved by decreasing atmospheric CO
2
levels
in the late Neogene, a world where C
4
plants have an advantage over
C
3
plants in many environments.
This model also has important implications concerning the
present glacial–interglacial period of Earth’s history, and the
future of the Earth where atmospheric CO
2
concentrations are
increasing because of fossil fuel burning. First, this model implies
that at very low CO
2
conditions, such as during the glacial periods,
C
3
grasses would have been at a great disadvantage worldwide so
that at intermediate and low latitudes an expansion of C
4
grasses
would be expected. The CO
2
concentration minima of about 160 to
180 p.p.m.v., reached during the Last Glacial Maximum, seems to be
near the limit for successful competition of C
3
grasses with respect
to C
4
grasses. Second, this model implies that C
4
grasses will be at an
increasing disadvantage as CO
2
levels increase owing to human-
kind’s energy appetite, in agreement with other models. This has
further evolutionary implications because the past 7 Myr of evolu-
tion, including the evolution of hominids, has been in the ‘C
4
-
world’. By increasing atmospheric CO
2
concentrations humans may
be changing the Earth’s atmosphere to conditions not favourable to
a ‘C
4
-world’, which were the conditions in which they originally
evolved. M
Received 26 December 1996; accepted 23 July 1997.
1. Bender, M. M. Variations in the
13
C/
12
C ratios of plants in relation to the pathway of photosynthetic
carbon dioxide fixation. Phytochemistry 10, 1239– 1245 (1971).
2. Winter, K., Troughton,J. H. & Card, K. A. d
13
C values of grass species collected in the northern Sahara
Desert. Oecologia 25, 115–123 (1976).
3. Brown, W. V. The Kranz syndrome and its subtypes in grass systematics. Mem. TorreyBot. Club 23, 1 –
97 (1977).
4. Vogel, J. C., Fuls, A. & Ellis, R. P.The geographical distribution of Kranz grasses in South Africa. S. Afr.
J. Sci. 74, 209– 215 (1978).
5. Farquhar, G. D., Ehleringer,J. R. & Hubrick, K. T.Carbon isotopic discrimination and photosynthesis.
Annu. Rev. Plant Physiol. Mol. Biol. 40, 503–537 (1989).
6. Quade, J. et al. A 16-Ma record of paleodiet using carbon and oxygen isotopes in fossil teeth from
Pakistan. Chem. Geol. 94, 183–192 (1992).
7. Lee-Thorp, J. & van der Merwe, N. Carbon isotope analysis of fossil bone apatite. S. Afr. J. Sci. 83, 712 –
715 (1987).
8. Wang, Y., Cerling, T. E. & MacFadden, B. J. Fossil horses and carbon isotopes: new evidence for
Cenozoic dietary, habitat, and ecosystem changes in North America. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 107, 269–279 (1994).
9. Cerling, T. E., Wang, Y. & Quade, J. Expansion of C
4
ecosystems as an indicator of global ecological
change in the late Miocene. Nature 361, 344–345 (1993).
10. Cerling, T. E., Quade, J., Wang,Y. & Bowman, J. R. Carbon isotopes in soils and paleosols as ecological
and paleoecologic indicators. Nature 341, 138–139 (1989).
11. Morgan, M. E., Kingston, J. D. & Marino,B. D. Carbon isotope evidence for the emergenceof C
4
plants
in the Neogene from Pakistan and Kenya. Nature 367, 162–165 (1994).
12. Kingston, J. D., Marino, B. & Hill, A. Isotopic evidencefor Neogene hominid paleoenvironments in
the Kenya Rift Valley. Science 264, 955–959 (1994).
13. Hill, A. in Paleoclimate and Evolution with Emphasis on Human Origins (eds Vrba, E. S., Denton, G. H.,
Partridge, T. C. & Burckle, L. H.) 178–193 (Yale Univ. Press, New Haven, 1995).
14. Ehleringer, J. R., Hall, A. E. & Farquhar,G. D. (eds) Stable Isotopes and Plant Carbon/Water Relations
(Academic, San Diego, 1993).
15. van der Merwe, N. J. & Medina, E. Photosynthesis and
13
C/
12
C ratios in Amazon rain forests. Geochim.
Cosmochim. Acta 53, 1091–1094 (1989).
16. Hattersley, P. W. d
13
C values of C
4
types of grasses. Aust. J. Plant Physiol. 9, 139–154 (1982).
articles
NATURE
|
VOL 389
|
11 SEPTEMBER 1997 157
40302010
200
300
400
500
600
700
C3 grasses
favoured
C4 grasses
favoured
Daytime growing-season temperature (°C)
Atmospheric CO2 (p.p.m.v.)
Figure 4 Results of a model for predicting C
3
/C
4
dominance of grasses related to
temperature and partial pressure of CO
2
according to which photosynthetic
pathway has the greater quantum yield; here ‘temperature’ is the daytime
growing-season temperature. This model is based on the equations of Farquhar
and von Caemmerer39 using constants determined by Jordan and Ogren46.
Parameters of the model, including quantum yield-patterns of C
3
and C
4
grasses,
have been verified by comparing the model results with published observa-
tions47–49.
Nature © Macmillan Publishers Ltd 1997
17. DeNiro, M. J. & Epstein, S. Influence of diet on the distribution of carbon isotopes in animals.
Geochim. Cosmochim. Acta 42, 495– 506 (1978).
18. Friedli, H., Lotscher, H., Oeschger, H., Siegenthaler, U. & Stauffer, B. Ice core record of the
13
C/
12
C
ratio of atmospheric CO
2
in the past two centuries. Nature 314, 237–238 (1986).
19. Marino, B. D. & McElroy,M. B. Isotopic composition of atmospheric CO
2
inferred from carbon in C
4
plant cellulose. Nature 349, 127–131 (1991).
20. Cerling, T. E., Bowman, J. R. & O’Neil,J. R. An isotopic study of a fluvial-lacustrine sequence: the Plio-
PleistoceneKoobi Fora sequence, EastAfrica. Palaeogeogr. Palaeoclimatol. Palaeoecol. 63,335– 356(1988).
21. Quade, J., Cerling, T. E. & Bowman, J. R. Development of Asian monsoon revealed by marked
ecological shift during the latest Miocene in nor thern Pakistan. Nature 342, 163–166 (1989).
22. Cande, S. C. & Kent, D. V. A new geomagnetic polarity time scale for the Late Cretaceous and
Cenozoic. J. Geophys. Res. 97, 13917–13951 (1992).
23. Berggren, W. A., Kent,D. V.,Flynn, J. J. & van Couvering, J. A. Cenozoic geochronology.Geol. Soc. Am.
Bull. 96, 1407–1418 (1985).
24. Latorre, C., Quade, J. & McIntosh, W. C. The expansion of C4 grasses and global change in the late
Miocene: Stable isotope evidence from the Americas. Earth Planet Sci. Lett. 146, 83–96 (1997).
25. Teeri, J. A. & Stowe, L. G. Climatic patterns and the distribution of C
4
grasses in North America.
Oecologia 23, 1–12 (1976).
26. Quade, J., Solounias, N. & Cerling, T. E. Stable isotopic evidence from paleosol carbonates and fossil
teeth in Greece for forest or woodlands over the past 11 Mya. Palaeogeogr. Palaeoclimatol. Palaeoecol.
108, 41–53 (1994).
27. Quade, J., Cerling, T. E., Andrews, P. & Alpagut, B. Paleodietary reconstruction of Miocene faunas
from Pasalar,Turkey using stable carbon and oxygenisotopes of fossil tooth enamel. J. Hum. Evol. 28,
373–384 (1995).
28. Hill, A. Causes of perceived faunal change in the later Neogene of East Africa. J. Hum. Evol. 16, 583–
596 (1987).
29. Barry, J. C., Johnson, N. M., Raza, S. M. & Jacobs, L. L. Neogene mammalian faunal change in
southern Asia: correlations with climatic, tectonic, and eustatic events. Geology 13, 637–640 (1985).
30. Barry, J. C. in Paleoclimate and Evolution, with Emphasis on Human Origins (eds Vrba, E. S., Denton,
G. H., Partridge, T. C. & Burckle, L. H.) 115–134 (Yale Univ. Press, New Haven, 1995).
31. Barr y,J. C. & Flynn, L. J. in European Neogene Mammal Chronology (eds Lindsay, E. H. et al.) 557– 571
(Plenum, New York, 1988).
32. Webb, S. D., Hulbert,R. C. & Lambert, W. D. in Paleoclimate and Evolution, with Emphasis on Human
Origins (eds Vrba, E. S., Denton, G. H., Partridge, T. C. & Burckle, L. H.) 91–108 (Yale Univ. Press,
New Haven, 1995).
33. MacFadden, B. J. Fossil Horses: Systematic, Paleobiology and Evolution of the Family Equidae
(Cambridge Univ. Press, New York, 1992).
34. Andrews, P. & Van Couvering, J. A. H. in Approaches to Primate Paleobiology (ed. Szalay, F.) 62–103
(Karger, Basel, 1975).
35. Van Couvering, J. A. H. & Van Couvering, J. A. in Human Origins (eds Isaac, G. L. & McCown, E.)
155–208 (Benjamin, Menlo Park, 1976).
36. Leakey, M. G. et al. 1996, Lothagam: A record of faunal change in the Late Miocene of East Africa. J.
Vert. Paleontol. 16, 556–570 (1996).
37. Eisenmann, V. & Sondaar,P. Y.Hipparions and the Mio-Pliocene boundary. Boll. Soc. Paleontol. Ital.
28, 217–226 (1989).
38. Ehleringer, J. R., Sage, R. F., Flanagan, L. B. & Pearcy, R. W. Climate change and the evolution of C4
photosynthesis. Trends Ecol. Evol. 6, 95–99 (1991).
39. Farquhar, G. D. & von Caemmerer, S. Modeling of photosynthetic response to environmental
conditions. Ency Plant Physiol. 12D, 549–587 (1982).
40. Raymo, M. E. & Ruddiman, W. F. Tectonic forcing of late Cenozoic climate. Nature 359, 117–122
(1992).
41. Stern, L. A., Johnson, G. D. & Chamberlain, C. P. Carbon isotope signature of environmental change
found in fossil ratite eggshells from a South Asian Neogene sequence. Geology 22, 419–422 (1994).
42. MacFadden, B. J. & Cerling, T. E. Mammalian herbivore communities, ancient feeding ecology, and
carbon isotopes: a 10 million-year sequence from the Neogene of Florida. J. Vert. Paleontol. 16, 103–
115 (1996).
43. MacFadden, B. J., Cerling, T. E. & Prado, J. Cenozoic terrestrial ecosystem evolution in Argentina:
evidence from carbon isotopes of fossil mammal teeth. Palaios 11, 319–327 (1996).
44. Flynn, J. J. & Swisher, C. C. in Geochronology, Time Scales, and Global Stratigraphic Correlation (eds
Berggren, W.A., Kent, D. V.,Aubry, M.-A. & Hardenbol, J.) 317–333 (Soc. Sedimentary Geol., Tulsa,
1995).
45. Thackeray, J. F. & Lee-Thorp, J. A. Isotopic analysis of equid teeth from Wonderwerk Cave, northern
Cape Province, South Africa. Palaeogeogr. Palaeoclimatol. Palaeoecol. 99, 141–150 (1992).
46. Jordan, D. B. & Ogren, W. L. The CO
2
/O
2
specificity of rubulose 1,5-biphosphate carboxylase/
oxygenase. Planta 161, 308–313 (1984).
47. Monson, R. K., Littlejohn, R. O. & Williams, G. J. The quantum yield for CO
2
uptake in C
3
and C
4
grasses. Photosynth. Res. 3, 153–159 (1982).
48. Ehleringer, J. R. & Pearcy, R. W.Variation in quantum yields for CO
2
uptake among C
3
and C
4
plants.
Plant Physiol. 73, 555– 559 (1983).
49. Brooks, A. & Farquhar, G. D. Effect of temperature on the CO
2
/O
2
specificity of rubulose-1,5-
biphosphate carboxylase/oxygenase and the rate of respiration in light. Plants 165, 397–406 (1985).
50. Bocherens, H., Koch, P. L., Mariotti, A., Geraads, D. & Jaeger, J.-J. Isotopic biogeochemistry of
mammalian enamel from African Pleistocene hominid sites. Palaios 11, 306–318 (1996).
Acknowledgements. Wethank W. Akersten,R. Anderson, T. M. Bown, J. D.Bryant, B. Engesser, J. Fleagle,
J. A. Hart, J. Hearst, H. Hutchison, L. L. Jacobs, E. H. Lindsay, E. L. Lundelius, H. G. McDonald,
N. Mudida, M. Voorhies, A. Walker, D. Whistler, D. Winkler and M. O. Woodburne for assistance in
obtaining samples. Wealso thank J. Kappelman and G. Farquhar for comments. This work was supported
by the US NSF.
Correspondence and requests for materials should be addressed to T.E.C. (e-mail: tcerling@mines.utah.
edu).
articles
158 NATURE
|
VOL 389
|
11 SEPTEMBER 1997
