By Jared Diamond

Genetic diseases seem an evolutionary blunder, an unjust fate decreed from the moment of
conception. Yet they may be our protection against hidden dangers.

What comfort is it to me that cause follows effect? I must have justice, or I’ll destroy myself. All
religions are built on this longing, and I=m a believer. But then there are the children: What am
I to do about them? That=s the question I can=t answer. It=s beyond all comprehension why
children should suffer to pay for a divine system of justice. Justice isn’t worth the tears of even a
single tortured child. God demands too high an admission price for his divine system. And so, I
respectfully give back to God my entrance ticket. - Dostoyevsky, The Brothers Karamazov

         The first person I knew who died of a genetic disease was a four-year old named Sara. I
was seven then, and my pediatrician father was caring for Sara and her younger brother, Tommy.
 Both at first had been normally adorable and happy children, but Sara=s life after age two was
made increasingly miserable by the blood disease she had inherited. She passed her last months
in constant pain from the enlarged liver pressing against her diaphragm. At the time she died
Tommy still seemed normal, but eventually he too was hit by the same disease that had taken his
sister; he died even more quickly and miserably.
         As a youngster I identified with the dying children, not with their parents. I could make
no sense of a world in which children died without having done anything to deserve it. Not until
more than 40 years later could I feel even pan of what Sara and Tommy's parents must have gone
through, as I watched my own son Joshua in the hospital, his life in doubt, and the terrified look
of a wounded animal in his eyes. Joshua survived, but two of the children in neighboring beds
didn't, and I understood then the grief and outrage in the eyes of their parents. As had
Dostoyevsky, I found myself wondering how anyone could believe in a system of justice that
requires children to suffer.
         Genetic diseases uniquely offend us. Yes, we grieve when someone dies for other
reasons, from an accident, say, or a contracted illness. But we routinely try to understand such
events by asking whether the victim did anything, or neglected to do anything, that led to his
death. In genetic diseases that path to understanding is foreclosed. We cannot impute
responsibility to the victim, because the cause of his death was present at the time of his concep-
tion. Someone with a genetic disease is a walking time bomb, set to explode at any time from
infancy until late adulthood, and often there is little or nothing the victim can do to prevent it.
Why does God or fate play these supremely dirty tricks on us?
         Our sense that the blow is undeserved is compounded by our sense of its
meaninglessness. The death seems to be nothing more than the consequence of a purposeless
genetic mistake, a need-less mutation blindly passed on to a child who will be summarily
disposed of by the relentless workings of natural selection. We can accept the notion of a mistake
when we buy a pen that proves defective and we have to discard it in the trash. But we can't face
saying, "My child was a mistake, so she ended up in the trash, and there's no more meaning to it
than that."
At the root of this anguish lies a big, and unresolved, scientific question: If it=s true that the filter
of natural selection acts to eliminate deleterious genes, then how is it that genetic diseases

persist? The answer may be that, contrary to appearances, genetic diseases are not mere mistakes,
that they are in fact sustained by some positive evolutionary purpose. It's an answer that may
seem cruel, at least as it pertains to any one victim. And yet it is the only way I've found to make
sense of the dilemma that arises when genetic diseases confront our longing to believe in a world
of justice.
        The term genetic diseases conjures up a few rare, genetically simple conditions; such as
Duchenne's muscular dystrophy and Tay-Sachs disease, in which everyone affected by the gene
invariably dies. But genetic diseases are actually a much broader phenomenon, encompassing the
commonest causes of human death. For example, there are very common conditions – such as
diabetes, ulcers, heart disease, and probably manic-depression – that depend on multiple genes,
and in which only some people bearing those genes fall prey to the disease. Also, we usually
think of infectious diseases as being completely different from genetic diseases. In fact, some
well-studied infectious diseases – tuberculosis, for one – prove to have a genetic component that
affects whether an exposed person will succumb to the infectious agent.
        Still other deaths that we don't normally think of as genetic include some of those
resulting from miscarriages. Traditionally the frequency of miscarriages was considered to be 15
percent, but this was in pregnancies that lasted long enough to be recognizable by the mother.
When hormonal methods for early detection of pregnancy were devised, it became clear that the
actual frequency of miscarriages was considerably higher, around 50 percent. Recent studies
suggest that many other fertilized embryos don’t even implant or survive long enough to become
detectable hormonally; adding these to the total yields a miscarriage rate of perhaps 80 percent.
That is, for each person born alive there may have been four phantom brothers or sisters who
died before they could be born. Genetic factors contribute to these prenatal deaths also.
        So, in the end, genetic diseases are involved in most human deaths. We are all of us time
bombs, differing from one another only in the triggering agent, the length of the fuse, and the
inevitability of the explosion.
        Yet why should this be so? Surely we can't all be merely the unwitting inheritors of past
genetic mistakes; for there's no doubt that natural selection does indeed tend to eliminate deleteri-
ous genes – a cold-blooded euphemism meaning that people with such genes may well die before
having children. Natural selection can't eradicate a deleterious gene permanently, of course,
because the gene keeps reappearing in new human embryos by mutation. Could new mutations
then account for all cases of a lethal genetic disease?
        If this were the case, then each year the number of people dying of that disease would, on
average, just equal the number of people in whom that mutation arose independently. Each
victim would carry a fresh mutation; the victim's parents wouldn't have carried that genetic
mistake, and the victim would have no children to pass it on to. Genetic diseases would then be
like scattered fires; each would be extinguished by the fire department, but every day there would
be new fires to put out.
        Something like this seems to be true for a congenital malformation called Apert's
syndrome, in which babies are born with deformed heads, hands, and feet. Most such babies
have normal parents, and few grow up to have children of their own. Muscular dystrophy, a more
familiar genetic disease, also seems to persist as a result of fresh mutations. However, muscular
dystrophy differs from Apert's syndrome in that only men die, while women survive to pass the
gene to their children. Still, if fresh mutations ceased, the deaths of all male children inheriting
the gene would eventually cause muscular dystrophy to disappear.

        Clearly, these and some other diseases really are maintained just by purposeless genetic
mutations. But they're all rare diseases, because genetic mutations themselves are rare. Muscular
dystrophy is the result of perhaps the most common mutation known in humans, and even it
arises only once in every 20,000 births; most other human mutations arise only once in 10
million births. It follows then that any genetic disease that kills many more than one in 20,000
people can't be sustained by mutations alone in the long run and must instead have some other
explanation for its persistence.
        I said "in the long run" because in certain circumstances a genetic disease can occur
frequently for a while without any purposeful factor to sustain it: the gene can become common
in a small population as a result of what's known as the founder effect.
        Suppose that ten couples settle a sparsely inhabited area and found a new population.
Suppose also that, by chance, one member of those ten founding couples carries some rare
deleterious gene. If the small group now increases rapidly in numbers, that gene will initially
occur in about one-tenth of the population, even though the gene is much rarer in populations
that have already reached genetic equilibrium.
        Of course, the people carrying the harmful gene will tend to leave fewer descendants than
other people, and eventually, after enough generations, the frequency of the gene's appearance
will decline to the rate that can be sustained by mutation alone. But that may take centuries, and
until then the gene will be present at an above-normal frequency.
        There are a number of modern populations with many members descended from a small
number of founding couples in the recent past – just think of the white population of Australia.
Studies of such populations regularly detect certain genetic diseases that occur frequently in those
populations but rarely in the rest of the world. With good genealogical records it's often possible
to prove that all known victims of the disease are descended from just one of the population's
early founders.
        Huntington's chorea, for example, is a fatal inherited neurological disease that occurs with
above-normal frequency in southeast Australia. Some painstaking record searching shows that at
least 432 Australian victims were directly descended from one English widow named Miss
Cundick, who took her 13 children by two marriages and immigrated to Australia in 1848. Evi-
dently Miss Cundick had the gene for Huntington's chorea and passed it on to some of her
numerous offspring. They in turn were able to leave many descendants, including hundreds with
Huntington's chorea.
        As another example, consider a genetic disease called osteodental dysplasia, which causes
all one's teeth to fall out by age 20 but otherwise doesn't markedly impair one's ability to attract a
mate and have children. While known cases are scattered around the world, a disproportionate
number are from South Africa. Is this because something in South Africa's climate makes
toothless people especially fertile? Not at all. It turns out that at least 71 known South African
cases are descended from a polygamous immigrant named Arnold who, assisted by his seven
known wives and 53 grandchildren, propelled the gene to its present high frequency.
        Thus the founder effect can explain some locally common genetic diseases – but only
under special circumstances. It applies mainly to populations that multiplied greatly from a small
number of founding couples in the recent past. And it applies to genes that are eliminated only
slowly by natural selection – either because the gene isn't fatal (as the case of osteodental
dysplasia), or because it appears in late adulthood, after the victim could already have had
children (as in the case of Huntington=s chorea); or because it's recessive, a term I'll explain in a

         All this still leaves unexplained genetic diseases that are wide spread in long-established
populations. Here, one would think, the frequencies of lethal genes should have long ago
declined to the frequencies of mutations. If a fatal genetic disease is nevertheless common, then
it=s unlikely that the gene=s frequency is being maintained by mutations alone. The conclusion
seems inescapable that since the gene is not being eliminated by natural selection, it must be
fulfilling some positive evolutionary purpose, conveying some benefit that offsets its harmful
effect. By that I mean that the gene, while making its bearers more likely than other people to die
of condition X, may in fact make them less likely to die of condition Y. Evolution could thus
cause the gene to persist, if the children whom it saved at least balanced the children whom it
         This conclusion may seem grisly. But this scientific explanation for the persistence of
genetic diseases could help us accept the ethical dilemma they pose. The disease gene would not
be a pure injustice, since it would also bring advantages - albeit at a price.
         There are two ways in which it could actually work. One is that the same person who
risks the gene's disadvantages may also stand to reap the gene's advantages: for example, a gene
that is bad for you when you're an adult could be good for you when you're a baby. Alternatively,
the gene could bestow its harmful effects only upon certain carriers, while blessing others with its
benefits: a gene could be good for people who inherit it from just one parent, for example, but
bad for people who inherit it from both. That explanation would also help us in our ethical
dilemma. The victim's death would not be in vain. In effect, he would be sacrificing his life so
that brothers and sisters might reap the benefits of the gene and live.
         My guess is that these agonizing bargains are what maintain the high frequencies of our
familiar genetic diseases. But it=s hard to test this speculation; there=s still only one disease –
sickle cell anemia – for which protection at a price is well established. There's suggestive
evidence for a few more diseases, speculative theories for several others, and no good clue for
         To see how these trade-offs might operate, we’ll first need a quick review of genetics.
Recall that we carry two copies of most of our genes: one that was inherited from your mother,
and one from your father. If your maternal and paternal forms of a particular gene are identical,
you're termed a homozygote for that gene; if they're different, you're termed a heterozygote.
         Suppose that a certain form of a particular gene causes a disease even when you're a
heterozygote with only one copy of that form (your other copy being the normal, non-disease-
causing form). We then say that the disease is dominant. If, on the other hand, you get the disease
only when you're a homozygote with both copies of the gene being the disease-related form, then
we say the disease is recessive.
         Sickle-cell anemia is a recessive disease. It arises from an abnormal form of the
hemoglobin pigment which carries oxygen and gives red blood cells their red color. The
abnormal hemoglobin turns red cells from disks into crescents, causing them to be destroyed, and
affected homozygotes often die in childhood from a severe red-cell deficiency. Heterozygotes,
however, are completely healthy under most circumstances.
         The sickle-cell gene is virtually confined to black Africans and their descendants and
some peoples of the Mediterranean, the Arabian Peninsula and India. About 10 percent of U.S.
blacks carry the gene, but in some areas of Africa its frequency is as high as 40 percent, which
means that in these regions the abnormal form of the gene appears almost as frequently as the

normal form and far, far more frequently than could be accounted for just by new mutations.
         Until the advent of modern medicine, most sickle-cell homozygotes died without leaving
children. Heterozygotes survived and reproduced, but whenever two heterozygotes produced
children, one-quarter of their offspring on the average were sickle-cell homozygotes and likely to
die. It would seem then that over many generations natural selection should have caused the
frequency of the sickle-cell gene to drop to low levels. Why hasn't it?
         The explanation proves to be that the gene is double-edged; while it condemns some to
death, it confers life on others. For it turns out that people with a single sickle-cell gene have a
significant advantage over normal people in regions where malaria is common. The
heterozygotes are less likely to die of the disease because their red cells are poor at supporting the
growth of malaria parasites.
         In those areas, as a result, more sickle-cell heterozygotes than normal people live long
enough to have children of their own. This excess of heterozygous children balances the deaths
of homozygous children, such that the sickle-cell gene persists in frequency from generation to
generation. In effect, the few homozygotes die of anemia so that the more numerous
heterozygotes will have a better chance of surviving malaria. The population as a whole has
bought partial protection of many of its children against its worst infectious disease, but at a
price. Without consulting them individually, evolution has struck a grim bargain.
         Among genetic diseases, sickle cell anemia is unique in that it offers very strong evidence
of an underlying evolutionary trade-off. Many other diseases may involve similar bargains, with
homozygotes paying the price and heterozygotes getting the protection, but the evidence is much
more speculative. Apparently the protection is either weaker, and thus harder to recognize, or it
was amassed against some infectious disease that was important in the past but less common
         After sickle-cell anemia the most plausible examples are some other blood diseases (such
as thalassemias) that may offer Africans, Asians, or Mediterranean peoples weaker protection
against malaria. Much more speculative is a theory for the persistence of Tay-Sachs disease, a
fatal neurological condition of children that is especially common in Jews of Eastern European
origin. The Tay-Sachs gene may formerly have protected heterozygotes against tuberculosis in
the days when it was still a leading killer. If so, one could rationalize why the Tay-Sachs gene be-
came most frequent in Eastern European Jews, formerly confined to urban ghettos teeming with
tuberculosis. Similarly, the gene for cystic fibrosis, another childhood killer and the most com-
mon recessive genetic disease of whites of European descent, may protect heterozygotes against
lethal diarrheas from bacterial infections – once a far worse risk of childhood than cystic fibrosis.
         There may also be evolutionary bargains in which the costs and benefits are apportioned
between the sexes, rather than between homozygotes and heterozygotes. A speculative example
is hemochromatosis, which may represent the most common abnormal gene in the United States.
The disease arises from unusually rapid intestinal absorption of iron, which can eventually cause
body stores of the metal to rise to ten times normal levels. The victim in effect dies of self-
inflicted iron poisoning.
         Men normally lose very little iron and can't get rid of any excess, unless they're addicted
to leeches and bloodletting. Women, on the other hand, are far more likely to suffer from iron
deficiency rather than surfeit, because they lose iron through menstrual bleeding, pregnancy, and
nursing. Not surprisingly, then, symptoms of hemochromatosis are seen much more often in men
than in women.

         Perhaps the hemochromatosis gene is actually good for women, by enabling them to
absorb iron more efficiently and thus combat their risk of iron deficiency. When, someday, a
textbook on genetics comes to be written by a liberated woman physician, the current view of
hemochromatosis as a disease may be dismissed as one more distorted product of male
chauvinism. Instead, we'll read that a gene for increased iron absorption is a wonderful
evolutionary blessing that enables women to overcome one of their leading metabolic problems.
A footnote will then mention, by the way, that a few older men bearing the same gene pay with
their lives so that women may benefit.
         In the cases I've discussed so far, one person reaps the gene's benefits, while a different
person bears the risks. But there are at least two examples in which the beneficiary and the victim
are one and the same.
         Both examples involve diabetes mellitus, a common condition that is often classified as
either insulin-dependent or non-insulin-dependent. The insulin-dependent variety is caused by
damage to the pancreatic cells that produce insulin, a hormone controlling our blood sugar
1evels, and it is treated by insulin injections. In contrast, non-insulin-dependent diabetes arises
from the body's developing a resistance to its own insulin; treatment usually involves controlling
the patient's diet. (These two types of diabetes were formerly known as juvenile-onset and adult-
onset diabetes, respectively, because of their tendencies to arise earlier or later in life.) Both
forms are genetically influenced, though not in such a simple way as the other diseases I've
         Only one out of five people genetically at risk actually develops insulin-dependent
diabetes (the outcome may depend on infection by a virus that damages the pancreas).
Nevertheless, until insulin injections became available, those who did develop diabetes were
likely to die, so one would have expected that eventually the frequency of diabetes-related genes
would have declined to a level low enough to be sustained by mutations. How did the disease
manage to remain common?
         Surprisingly, the answer may be that the risk of diabetes is offset by a genetic protection
offered to a fetus against the risk of miscarriage. Consider a parent who is heterozygous for a
diabetes-related gene. That is, of the parent's two copies of the gene, one is the normal form,
while the other is the form that predisposes to diabetes. You'd then expect that half the babies
born to such a parent would inherit the normal gene, and half would inherit the diabetes related
gene. In fact, at birth up to 72 percent of the children carry the diabetes-related gene. This
suggests that the frequency of miscarriage was higher for fertilized eggs carrying the normal gene
than for those carrying the diabetes-related gene. Since four out of the five fetuses with diabetes-
related genes will not develop diabetes, the improved chance of prenatal survival keeps the genes
common despite the postnatal deaths of one-fifth of the carriers.
         Protection against miscarriage will surely prove to be important in explaining the
persistence of some genetic diseases, whether or not it proves valid for diabetes in particular.
Recall that many or even most conceptions end in miscarriage. Any genetic advantage that helps
the fetus survive death before birth is evolutionarily valuable, for its whole childbearing career
lies in the future. Evolution may have given us many genes that favor fetuses and babies while
causing diseases in adults past childbearing age.
         My other example involves the type of diabetes in which the body develops resistance to
its own insulin. Normally we release insulin after a meal to help our body store the food we
ingest. In this disease, resistance is thought to develop as a result of the body releasing insulin too

quickly, too often or in too large amounts after a meal. While the condition has a genetic basis,
it's also obviously influenced by diet. Examples can be found in Pima Indians, Pacific islanders,
and Yemenite Jews; in all these peoples the occurrence of diabetes shot sky-high within a decade
or two of their rapid transition from an active, low calorie, low carbohydrate life to a life of little
exercise and a high-calorie, high-carbohydrate diet.
         Why would genes that were good for people living meagerly suddenly bring on diabetes
when those people became Westernized? It may have something to do with the availability of
food. For many people in the West today, food is easily acquired and eaten three or more times
daily at scheduled hours. But for many others food is something they eat whenever they are lucky
enough to get it. This was certainly true for almost all humans in the past. Our former life was
one of feast or famine. Those who could pack away lots of food when it was available, and
whose bodies could store it for later use, were able to survive subsequent famines.
         Many stories illustrate the ability of people accustomed to a Spartan life to gorge
themselves when food was abundant. The British Arctic explorer Sir Edward Parry once did the
experiment of offering an Eskimo man named Tooloak as much food and drink as he could
consume overnight. In 13 hours Tooloak packed away more than ten pounds of meat and bread,
and washed it down with almost two gallons of soup, alcohol, and other liquids.
         Throughout most of our evolutionary history, those individuals who released lots of
insulin gained an advantage by retaining the calories they gorged. That enabled them to survive
subsequent times of starvation, while their insulin poor comrades succumbed. But put the same
people on a Western diet, where they are challenged to pour out insulin all the time, and they
develop resistance to insulin and hence diabetes.
         This, then, is the scientific sense we're able to make of genetic diseases. They somehow
manage to persist at high frequencies, despite the tendency of natural selection to eliminate
deleterious genes. While mutations or the founder effect explains some cases, the main
explanation for common widespread genetic diseases is that the genes somehow help as well as
harm our survival. Where genetic diseases differ is in how the costs and benefits are allocated
among people. Sickle-cell anemia and possibly cystic fibrosis and Tay-Sachs disease, help
heterozygotes while harming homozygotes. Hemochromatosis may help women while harming
men. Insulin dependent diabetes may help a person as a fetus, then harm him later in life. And
non-insulin-dependent diabetes may help a person on a Spartan diet but harm the same person on
a generous diet. In all these cases, disease genes bring concealed blessings – at a price.
         Do these answers to the scientific paradox suggest an approach to the more difficult
ethical paradox? Can they tell us how to maintain our faith in a universe that permits genetic
diseases to exist? It's hard to formulate a response that will sound convincing when you're on a
pediatric ward, listening to sick children crying. Anyone who has heard those cries understands
why Ivan Karamazov spurned God's entrance ticket to his divine system.
         But perhaps there is an answer, incomplete though it may be. We live in a world that
exposes us to danger at every moment from conception onward. First there is the danger of
miscarriage. Next, the risk of succumbing to childhood infectious diseases; and then the risk of
other infections and starvation through out all of adult life. In this perilous world we can only try
to solve each risk as it arises. Some of our genetic solutions catch up with us later and become
labeled as genetic diseases, but that term is a misnomer. Those genes enabled us, or our relatives,
to survive earlier dangers. In that sense, genetic diseases are agonizing Faustian bargains in
which natural selection plays the role of Mephistopheles.


To top