You are on page 1of 8


the recent

of humans
Genetic studies
reveal that an
African woman
from less than
200,000 years ago
was our common
By Rebecca L. Cann
and Allan C. Wilson

POINT-COUNTERPOINT: For an opposing view of how humankind arose

around the globe, see The Multiregional Evolution of Humans, on page 46.

54 SCIENTIFIC AMERICAN Updated from the April 1992 issue

n the quest for the facts about had been right and they had been wrong. lived about 200,000 years ago, probably
human evolution, we molecular Once again we are engaged in a de- in Africa. Modern humans arose in one
geneticists have engaged in two bate, this time over the latest phase of place and spread elsewhere.
major debates with the paleon- human evolution. The paleontologists Neither the genetic information of
tologists. Arguing from their fos- say modern humans evolved from their living subjects nor the fossilized remains
sils, most paleontologists had archaic forebears around the world over of dead ones can explain in isolation
claimed the evolutionary split between the past million years. Conversely, our how, when and where populations orig-
humans and the great apes occurred as genetic comparisons convince us that all inated. But the former evidence has a cru-
long as 25 million years ago. We main- humans today can be traced along ma- cial advantage in determining the struc-
tained human and ape genes were too ternal lines of descent to a woman who ture of family trees: living genes must have
similar for the schism to be more than a
AFRICAN ORIGIN for all modern humans is indicated by the genetic evidence. A genealogy based on
few million years old. After 15 years of 182 current mitochondrial DNA types (outer edges) points to the existence of a common female
disagreement, we won that argument ancestor from Africa. The arrows on the map (center) indicate the route and the minimum number of
when the paleontologists admitted we unrelated females (red circles) who colonized various areas, as inferred from the branching pattern.

110 100 90 80 70
120 60



36 31





15 10

African Asian Australian New Guinean Caucasian


0 0.2 0.4 0.6 0.6 0.4 0.2 0

Divergence in DNA Divergence in DNA
Sequence (percent) Sequence (percent) SCIENTIFIC AMERICAN 55

ancestors, whereas dead fossils may not The fossil record, on the other hand, tory at the University of California at
have descendants. Molecular biologists is infamously spotty because a handful Berkeley, challenged a fossil primate
know the genes they are examining must of surviving bones may not represent the called Ramapithecus. Paleontologists
have been passed through lineages that majority of organisms that left no trace had dated its fossils to about 25 million
survived to the present; paleontologists of themselves. Fossils cannot, in princi- years ago. On the basis of the enamel
cannot be sure that the fossils they ex- ple, be interpreted objectively: the phys- thickness of the molars and other skele-
amine do not lead down an evolutionary ical characteristics by which they are tal characteristics, they believed that
blind alley. classified necessarily reflect the models Ramapithecus appeared after the diver-
The molecular approach is free from the paleontologists wish to test. If one gence of the human and ape lineages and
several other limitations of paleontol- classifies, say, a pelvis as human because that it was directly ancestral to humans.
ogy. It does not require well-dated fos- it supported an upright posture, then Sarich measured the evolutionary
sils or tools from each part of the fami- one is presupposing that bipedalism dis- distance between humans and chim-
ly tree it hopes to describe. It is not viti- tinguished early hominids from apes. panzees by studying their blood pro-
ated by doubts about whether tools Such reasoning tends to circularity. The teins, knowing the differences reflected
found near fossil remains were in fact paleontologists perspective therefore mutations that have accumulated since
made and used by the population those contains a built-in bias that limits its the species diverged. (At the time, it was
remains represent. And it concerns itself power of observation. much easier to compare proteins for
with a set of characteristics that is com- As such, biologists trained in modern subtle differences than to compare the
plete and objective. evolutionary theory must reject the no- genetic sequences that encode the pro-
A genome, or full set of genes, is tion that the fossils provide the most di- teins.) To check that mutations had oc-
complete because it holds all the inherit- rect evidence of how human evolution curred equally fast in both lineages, he
ed biological information of an individ- actually proceeded. Fossils help to fill in compared humans and chimpanzees
ual. Moreover, all the variants on it that the knowledge of how biological pro- against a reference species and found
appear within a population a group of cesses worked in the past, but they that all the genetic distances tallied.
individuals who breed only with one an- should not blind us to new lines of evi- Sarich now had a molecular clock;
other can be studied, so specific pecu- dence or new interpretations of poorly the next step was to calibrate it. He did
liarities need not distort the interpreta- understood and provisionally dated ar- so by calculating the mutation rate in
tion of the data. Genomes are objective chaeological materials. other species whose divergences could
because they present evidence that has be reliably dated from fossils. Finally, he
not been defined, at the outset, by any Molecular Clock applied the clock to the chimpanzee-hu-
particular evolutionary model. Gene se- ALL THE ADVANTAGES of our field man split, dating it to between five mil-
quences are empirically verifiable and stood revealed in 1967, when Vincent lion and seven million years agofar lat-
not shaped by theoretical prejudices. M. Sarich, working in Wilsons labora- er than anyone had imagined.

The Inheritance of Mitochondrial DNA

Egg Fertilized egg

37 genes

Mitochondrial DNA

Mitochondrion Nuclear DNA MOST OF AN INDIVIDUALS GENES are located on DNA molecules
in the cell nucleus. Mitochondria, the specialized structures that
provide cells with energy, also carry some genes for their own
manufacture on a ring of DNA. When a sperm and an egg cell unite,
they contribute equally to the DNA in the nucleus of the resulting
Sperm cell. All the mitochondria and the DNA they contain, however,

derive from the egg. Studies of mitochondrial DNA can reveal

an individuals maternal ancestry.


Mitochondrial DNA is inherited from the

mother alone, so all of it today had one female ancestor.

At first, most paleontologists clung convert food into a form of energy the chondrial DNA is inherited from the
to the much earlier date. But new fossil rest of the cell can use. Unlike the DNA mother alone, unchanged except for
finds undermined the human status of of the nucleus, which forms bundles of chance mutations. The fathers contribu-
Ramapithecus: it is now clear that Ra- long fibers, each consisting of a protein- tion ends up on the cutting-room floor,
mapithecus is actually Sivapithecus, a coated double helix, the mitochondrial as it were. The nuclear genes, to which
creature ancestral to orangutans and not DNA comes in small, two-stranded the father does contribute, descend in
to any of the African apes at all. More- rings. Whereas nuclear DNA encodes an what we may call ordinary lineages,
over, the age of some sivapithecine fos- estimated 100,000 genes most of the which are of course important to the
sils was downgraded to only about six information needed to make a human transmission of physical characteristics.
million years. By the early 1980s almost being mitochondrial DNA encodes For our studies of modern human ori-
all paleontologists came to accept Sarichs only 37. In this handful of genes, every gins, however, we focus on the mito-
more recent date for the separation of one is essential: a single adverse muta- chondrial, maternal lineages.
the human and ape lines. Those who con- tion in any of them is known to cause Maternal lineages are closest among
tinue to reject his methods have been re- some severe neurological diseases. siblings because their mitochondrial
duced to arguing that Sarich arrived at For the purpose of scientists studying DNA has had only one generation in
the right answer purely by chance. when lineages diverged, mitochondrial which to accumulate mutations. The de-
Two novel concepts emerged from DNA has two advantages over nuclear gree of relatedness declines step by step
the early comparisons of proteins from DNA. First, the sequences in mitochon- as one moves along the pedigree, from
different species. One was the concept of drial DNA that interest us accumulate first cousins descended from the mater-
inconsequential, or neutral, mutations. mutations rapidly and steadily, accord- nal grandmother, to second cousins de-
Molecular evolution appears to be dom- ing to empirical observations. Because scended from a common maternal great-
inated by such mutations, and they ac- many mutations do not alter the mito- grandmother and so on. The farther
cumulate at surprisingly steady rates in chondrions function, they are effective- back the genealogy goes, the larger the
surviving lineages. In other words, evo- ly neutral, and natural selection does not circle of maternal relatives becomes, un-
lution at the gene level results mainly eliminate them. til at last it embraces everyone alive.
from the relentless accumulation of mu- This mitochondrial DNA therefore Logically, then, all human mito-
tations that seem to be neither harmful behaves like a fast-ticking clock, which chondrial DNA must have had an ulti-
nor beneficial. The second concept, mo- is essential for identifying recent genetic mate common female ancestor. But it is
lecular clocks, stemmed from the obser- changes. Any two humans chosen ran- easy to show she did not necessarily live
vation that rates of genetic change from domly from anywhere on the planet are in a small population or constitute the
point mutations (changes in individual so alike in most of their DNA sequences only woman of her generation. Imagine
DNA base pairs) were so steady over that we can measure evolution in our a static population that always contains
long periods that one could use them to species only by concentrating on the 15 mothers. Every new generation must
time divergences from a common stock. genes that mutate fastest. Genes con- contain 15 daughters, but some mothers
trolling skeletal characters do not fall will not produce a daughter, whereas
Mitochondrial Clue within this group. others will produce two or more. Be-
W E C O U L D B E G I N to apply these Second, unlike nuclear DNA, mito- cause maternal lineages die out whenev-
methods to the reconstruction of later
stages in human evolution only after REBECCA L. CANN and ALLAN C. WILSON applied the tools of genetics to paleontology dur-

1980, when DNA restriction analysis ing many of their collaborations. Cann is professor of genetics and molecular biology at the
made it possible to explore genetic dif- John A. Burns School of Medicine of the University of Hawaii at Manoa. She received both
ferences with high resolution. Workers her bachelors degree in genetics and her Ph.D. in anthropology from the University of Cal-
at Berkeley, including Wes Brown, Mark ifornia, Berkeley. As a postdoctoral fellow, she worked at Berkeley with Wilson and at the
Stoneking and us, applied the technique University of California, San Francisco. Cann is using mitochondrial DNA to assay the ge-
to trace the maternal lineages of people netic diversity of birds in the Hawaiian Islands. Until his death in 1991, Wilson was profes-
sampled from around the world. sor of biochemistry at Berkeley. A native of New Zealand, he received his doctorate from
The DNA we studied resides in the Berkeley. Wilson also worked at the Weizmann Institute of Science in Rehovot, Israel, at the
mitochondria, cellular organelles that University of Nairobi and at Harvard University. SCIENTIFIC AMERICAN 57

er there is no daughter to carry it on, it homeland: the global distribution of mi- Working at Berkeley with Stoneking,
is only a matter of time before all but one tochondrial DNA types he saw could we expanded on Kochers work by ex-
lineage disappears. In a stable popula- then be explained most easily as the re- amining a larger genealogical tree made
tion, the time for this fixation of the ma- sult of no more than three migration up of 182 distinct types of mitochondri-
ternal lineage to occur is the length of a events to other continents. al DNA from 241 individuals. The mul-
generation multiplied by twice the popu- A crucial assumption in this analysis tiple occurrences of mitochondrial DNA
lation size. is that all the mitochondrial lineages types were always found among people
evolve at the same rate. So when Kocher from the same continent and usually in
Eve in Africa conducted his comparison of the human persons who lived within 100 miles of
ONE MIGHT REFER to the lucky wo- mitochondrial DNAs, he also included one another. Because the tree we con-
man whose lineage survives as Eve. Bear analogous sequences from four chim- structed had two main branches, both of
in mind, however, that other women panzees. If the human lineages had dif- which led back to Africa, it, too, sup-
were living in Eves generation and that fered in the rate at which they accumu- ported the hypothesis that Africa was
Eve did not occupy a specially favored lated mutations, then some of the 14 hu- the place of origin for modern humans.
place in the breeding pattern. She is man sequences would be significantly One noteworthy point that jumps
purely the beneficiary of chance. More- closer or farther away from the chim- out of our study is that although geo-
over, if we were to reconstruct the ordi- panzee sequences than others. In fact, all graphic barriers do influence a popula-
nary lineages for the population, they 14 human sequences are nearly equidis- tions mitochondrial DNA, people from
would trace back to many of the men tant from the chimpanzee sequences, a given continent do not generally all be-
and women who lived at the same time which implies that the rates of change long to the same maternal lineage. The
as Eve. Population geneticists Daniel L. among humans are fairly uniform. New Guineans are typical in this respect.
Hartl, now at Harvard University, and The chimpanzee data also illustrated Their genetic diversity had been suspect-
Andrew G. Clark, now at Cornell Uni- how remarkably homogeneous humans ed from linguistic analyses of the re-
versity, estimate that as many as 10,000 are at the genetic level: chimpanzees markable variety of language families
people could have lived then. The name commonly show as much as 10 times the usually classified as Papuan spoken on
Eve can therefore be misleading she genetic variation of humans. That fact this one island [see The Austronesian
is not the ultimate source of all the ordi- alone suggests that all of modern hu- Dispersal and the Origin of Languages,
nary lineages, as the biblical Eve was. manity sprang from a relatively small by Peter Bellwood; Scientific Ameri-
From mitochondrial DNA data, it is stock of common ancestors. can, July 1991]. On our genealogical tree,
possible to define the maternal lineages
of living individuals all the way back to 1
a common ancestor. In theory, a great 2
number of different genealogical trees
could give rise to any set of genetic data. 3
To recognize the one that is most prob- 4
ably correct, one must apply the parsi- 5
mony principle, which requires that sub-
jects be connected in the simplest possi-
ble way. The most efficient hypothetical 7

tree must be tested by comparison with 8

other data to see whether it is consistent
with them. If the tree holds up, it is ana-
lyzed for evidence of the geographic his- 10
tory inherent in elements. 11
In 1988 Thomas D. Kocher of Berke-
ley (now at the University of New Hamp-
shire) applied just such a parsimonious 13
interpretation to the interrelatedness of 14
the mitochondrial DNA of 14 humans 15
from around the world. He determined
that 13 branching points were the fewest 16
that could account for the differences he

UNIVERSAL MATERNAL ANCESTOR can be found for all the members of any population. The example
found. Taking the geographic consider- shown here traces the lineages of 15 females in a stable population. In each generation, some
ations into account, he then concluded maternal lineages proliferate and others become extinct. Eventually, by chance, one maternal
that Africa was the ultimate human lineage (dark blue) replaces all the others.


Huge levels of gene flow between early continentsvery
unlikelywould have been needed for multiregionalism.
New Guineans showed up on several dif- tochondrially speaking, races are not ed that constraint. The reaction makes it
ferent branches, which proved that the like biological species. We propose that possible to duplicate DNA sequences
common female ancestor of all New the anatomical characteristics uniting easily, ad infinitum; a small starting sam-
Guineans was not someone in New New Guineans were not inherited from ple of DNA can expand into an endless
Guinea. The population of New Guinea the first settlers. They evolved after peo- supply.
must have been founded by many moth- ple colonized the island, chiefly as the re- The polymerase chain reaction en-
ers whose maternal lineages were most sult of mutations in nuclear genes spread abled Linda Vigilant of Pennsylvania
closely related to those in Asia. by sex and recombination throughout State University to redo our study using
That finding is what one would ex- New Guinea. Similarly, the light skin mitochondrial DNA data from 120 Af-
pect if the African origin hypothesis color of many whites is probably a late ricans, representing six diverse parts of
were true: as people walked east out of development that occurred in Europe af- the sub-Saharan region. Vigilant traced
Africa, they would have passed through ter that continent was colonized by a genealogical tree whose 14 deepest
Asia. Travel was probably slow, and Africans. branches lead exclusively to Africans
during the time it took to reach New During the early 1980s, when we and whose 15th branch leads to both Af-
Guinea, mutations accumulated both in were constructing our genealogical tree, ricans and non-Africans. The non-Afri-
the lineages that stayed in Asia and in we had to rely on black Americans as cans lie on shallow secondary branches
those that moved on. substitutes for Africans, whose mito- stemming from the 15th branch. Con-
Thus, people who are apparently re- chondrial DNA was difficult to obtain in sidering the number of African and non-
lated by membership in a common geo- the required quantities. Fortunately, the African mitochondrial DNAs surveyed,
graphic race need not be very closely re- development of a technique called the the probability that the 14 deepest
lated in their mitochondrial DNA. Mi- polymerase chain reaction has eliminat- branches would be exclusively African is
one in 10,000 for a tree with this branch-
ing order.
Satoshi Horai and Kenji Hayasaka of
the National Institute of Genetics in
Mishima, Japan, analogously surveyed
population samples that included many
more Asians and individuals from few-
er parts of Africa; they, too, found that
the mitochondrial lineages led back to
Africa. We estimate the odds of their ar-
riving at that conclusion accidentally
were only four in 100. Although these
statistical evaluations are not strong or
rigorous tests, they do make it seem like-
African human ly that the theory of an African origin for
Non-African person human mitochondrial DNA is now fair-
Chimpanzee ly secure.
Pygmy chimpanzee
200,000 Years or Less
the chimpanzee data showed that the
human mitochondrial DNA clock has
ticked steadily for millions of years, we
knew it should be possible to calculate

INTERRELATEDNESS of 14 humans and four chimpanzees was inferred

from similarities discovered in their mitochondrial DNA sequences. The
when the common mother of humanity
chimpanzee data help researchers to measure when various evolutionary lived. We assumed that the human and
divergences in the human lineages occurred. chimpanzee lineages diverged five mil- SCIENTIFIC AMERICAN 59

lion years ago, as Sarichs work had might run by fits and starts, we ran a test occupied by archaic people who had mi-
shown. We then calculated how much to measure how much mitochondrial grated from Africa to Asia nearly a mil-
humans had diverged from one another DNA has evolved in populations found- lion years ago. Such famous fossils as
relative to how much they had diverged ed at a known time. Java Man and Beijing Man are of this
from chimpanzees that is, we found The aboriginal populations of New type. This finding and the hypothesis
the ratio of mitochondrial DNA diver- Guinea and Australia are estimated to that the archaic Eurasian population un-
gence among humans to that between have been founded less than 50,000 to derwent anatomical changes that made
humans and chimpanzees. 60,000 years ago. The amount of evolu- them resemble more modern people led
Using two different sets of data, we tion that has since occurred in each of to the multiregional evolution model:
determined that the ratio was less than those places seems about one third of similar evolutionary changes in separate
1:25. Human maternal lineages there- that shown by the whole human species. geographic regions converted the inhab-
fore grew apart in a period less than 1 25 Accordingly, we can infer that Eve lived itants from archaic small-brained types
as long as five million years, or less than three times 50,000 to 60,000 years ago, to modern big-brained types.
200,000 years. With a third set of data or roughly 150,000 to 180,000 years Huge levels of gene flow between
on changes in a section of the mito- ago. All our estimates thus agree that the continents, however, would be neces-
chondrial DNA called the control re- split happened not far from 200,000 sary to maintain human populations as
gion, we arrived at a more ancient date years ago. one biological species. The multiregion-
for the common mother. That date is Those estimates fit with at least one al evolution model also predicts that at
less certain, however, because questions line of fossil evidence. The remains of least some genes in the modern east
remain about how to correct for multi- anatomically modern people appear first Asian population would be linked more
ple mutations that occur within the con- in Africa, then in the Middle East, and closely to those of their archaic Asian
trol region. later in Europe and east Asia. Anthropol- predecessors than to those of modern
One might object that a molecular ogists have speculated that in east Africa Africans. We would expect to find deep
clock known to be accurate over five the transition from anatomically archaic lineages in Eurasia, especially in the Far
million years could still be unreliable for to modern people took place as recently East. Yet surveys in our laboratories and
shorter periods. It is conceivable, for ex- as 130,000 years ago [see The Emer- in others, involving more than 1,000
ample, that intervals of genetic stagna- gence of Modern Humans, by Christo- people from Eurasia and its mitochon-
tion might be interrupted by short bursts pher B. Stringer; Scientific American, drial DNA satellites (Australia, Oceania
of change when, say, a new mutagen en- December 1990]. and the Americas), have given no hint of
ters the environment, or a virus infects On the other hand, a second line of that result.
the germ-line cells, or intense natural se- evidence appears to conflict with this It therefore seems very unlikely that
lection affects all segments of the DNA. view. The fossil record shows clearly any truly ancient lineages survive unde-
To rule out the possibility that the clock that the southern parts of Eurasia were tected in Eurasia. We simply do not see
the result predicted by the regional mod-
el. Moreover, geneticists such as Masa-
toshi Nei of Pennsylvania State Univer-
sity, Kenneth K. Kidd of Yale Universi-
ty, James Wainscoat of the University of
100,000 Klasies Neandertal Ngandong Oxford and Luigi L. Cavalli-Sforza of
Dali Stanford University have found support
for an African origin model in their stud-
Age (years)

ies of nuclear genes.

300,000 Saldanha Petralona (Beijing) Sambungmachan Multiregional Mystery
P R O P O N E N T S O F the multiregional
evolution model typically emphasize
that they have documented a continuity
of anatomical morphologies between
700,000 Olduvai European Lantian Java
the archaic and modern residents of dif-
ferent regions; they insist that these mor-
phologies would be unlikely to evolve
Homo erectus
independently in any invading people.

ARCHAIC HUMAN GROUPS were gradually replaced throughout the Old World by modern humans who For that argument to hold true, howev-
arose in Africa. Archaic females do not seem to have contributed mitochondrial genes to the modern er, it must also be shown that the cranial
people of Europe, east Asia and Australia. features in question are truly indepen-


dent of one another that is, that natur-
al selection would not tend to favor cer-
tain constellations of functionally relat- Father Mother
ed features anyway. Yet we know that
powerful jaw muscles may impose
changes on the mandible, the browridge
and other points on the skull; circum-
stances that promoted the evolution of
these features in one population might
do so again in a related population.
Other paleontologists also dispute
the evidence for continuity. They argue
that modern populations are not linked
to past ones by morphological charac-
teristics that evolved uniquely in the fos- Male
sil record. Instead fossils and modern Female
populations are united by their shared
retention of still older ancestral charac- Mitochondrial DNA source
teristics. The continuity seen by believ- PEDIGREE of one individual illustrates the difference between the patterns of nuclear and
ers in multiregional evolution may be an mitochondrial inheritance. All 32 ancestors from five generations ago contributed equally to his
illusion. nuclear DNA. His mitochondrial lineage (blue line) leads back to only one person in every generation.
The idea that modern humans could
cohabit a region with archaic ones and Lucotte, while at the College of France, proved methods of extracting DNA
replace them completely without any and his colleagues have indirectly com- from still older fossilized bone now ap-
mixture may sound unlikely. Neverthe- pared such sequences in an effort to pear close at hand. With them, we may
less, some fossil finds do support the trace paternal lineages to a single pro- begin building the family tree from a root
idea. Discoveries in the caves at Qafzeh genitor Adam, if you will. Those that was alive when the human family
in Israel suggest that Neandertals and preliminary results also point to an was young.
modern humans lived side by side for African homeland, and with further re-
40,000 years, yet they left little evidence finements this work on paternal lineages Epilogue
of interbreeding. may be able to provide an invaluable S I N C E T H I S A R T I C L E was first pub-
How one human population might check on our results for maternal lin- lished, further genetic work on the mi-
have replaced archaic humans without eages. Unfortunately, base changes ac- tochondrial DNA sequences of three
any detectable genetic mixing is still a cumulate slowly on useful regions of the Neandertal specimens upholds our con-
compelling mystery. One of us (Cann) Y chromosome, making it technically clusions about the lack of a mixture be-
suspects that infectious diseases could difficult to conduct a detailed genealog- tween ancient and modern Homo sapi-
have contributed to the process by help- ical analysis. ens. Furthermore, whole mitochondrial
ing to eliminate one group. Cavalli- More progress can be expected soon, genome sequencing all 16,569 base
Sforza has speculated that the ancestors as molecular biologists learn to apply pairs from more than 50 donors gives
of modern humans may have developed their techniques to materials uncovered more precise resolution to the timescale
some modern trait, such as advanced by our friendly rivals, the paleontolo- of our emergence. It now seems that the
language skills, that effectively cut them gists. Preliminary molecular studies have earliest migration out of Africa is closer
off from breeding with other hominids. already been conducted on DNA from to 120,000 years ago than 200,000 years
This and related questions may yield as mummified tissues found in a Florida ago more recent, yet still within the
molecular biologists learn how to link bog and dated to 7,500 years ago. Im- range we had originally estimated.
specific genetic sequences to the physical
and behavioral traits that those se- MORE TO E XPLORE
quences ultimately influence. Mitochondrial DNA and Human Evolution. Rebecca L. Cann, Mark Stoneking and Allan C. Wilson in
Even before then, further studies of Nature, Vol. 325, No. 6099, pages 3136; January 17, 1987.
both nuclear and mitochondrial DNA Mitochondrial DNA. M. Stoneking and A. C. Wilson in The Colonization of the Pacific: A Genetic Trail.
Edited by Adrian V. S. Hill and Susan W. Serjeantson. Oxford University Press, 1989.
will render more informative genetic
Mitochondrial DNA Sequences in Single Hairs from a Southern African Population. Linda Vigilant,
trees. Particularly enticing are the se-

Renee Pennington, Henry Harpending, Thomas D. Kocher and Allan C. Wilson in Proceedings of the
quences on the Y chromosome that de- National Academy of Sciences USA, Vol. 86, No. 23, pages 93509354; December 1989.
termine maleness and that are therefore Sequence Evolution of Mitochondrial DNA in Humans and Chimpanzees. T. D. Kocher and A. C.
inherited from the father alone. Gerard Wilson in Evolution of Life. Edited by S. Osawa and T. Honjo. Springer-Verlag, Tokyo, 1991. SCIENTIFIC AMERICAN 61