Of the thousands of serious genetic diseases, the vast majority are rare, affecting fewer than one in ten thou- sand people. Most of these diseases result from reces- sives, genes that don’t cause any trouble except in individuals unlucky enough to get two copies, so there is no normal allele at that locus. This misfortune becomes more likely if you marry a relative, who will have more genes identical to yours than a non- relative will. This is why marriages between close relatives are more likely to produce abnormal babies.
It is hard for natural selection to eliminate a deleterious recessive gene. If, as is likely, people heterozygous for a rare recessive have no disadvantage, the rate of adverse selection may be so small that nat- ural selection cannot depress the gene frequency further. If a gene is present in one in a thousand individuals and people normally marry nonrelatives, then on average only one in a million will be homozy- gous. Even if all of these unfortunate people die early in life, the effect of selection is weak. In this situation, new mutations can often create the defective gene as fast as natural selection eliminates it, because as the gene frequency decreases, the prevalence of homozy- gous individuals decreases even faster. A lethal recessive gene that is created by mutation in one out of a million pregnancies will stabilize in frequency at about one in a thousand individuals. This is indeed a situation in which the power of natural selection is limited.