There are myths about women giving birth to children from two different fathers. A look at genetics today, though, shows that those myths might have originated in truth. It's called superfecundation.
The full term for it is heteropaternal superfecundation. The Greeks probably got the idea from cats and dogs, which can bear litters in which each offspring is from a different father. Dogs, especially, look so dramatically different from breed to breed that a discrepancy in paternity was bound to be noticed. But it wasn't until relatively recently that superfecundation in humans was discovered.
Superfecundation isn't necessarily heteropaternal. Superfecundation is a general term for when two eggs get fertilized in two different, to use the scientific term, "coitions." (I learned a word, today!) Generally, these coitions happen within a space of twenty-four to forty-eight hours. Since most incidents of this happen with the same father, there's no practical way to separate this type of superfecundation from regular fraternal twins. Some physicians consider them the same thing. At times, though, there can be dramatic differences. One woman went in for an ultrasound at two and a half months pregnant only to find out she was really three months pregnant, with one fetus two weeks more developed than the other.
Such an extreme time difference in conception is almost vanishingly rare. Scientists obviously want to know how often superfecundation happens. The only way to test how often superfecundation occurs is performing genetic or blood tests on fraternal twins, and seeing how many have two different fathers. The data varies hugely. One test shows that while there are only about three cases of heteropaternal superfecundation in an overall test database of 39,000 records, but the rate goes way up when the parents of twins are in a paternity suit. About 2.4 percent of tested twins, in those cases, were heteropaternal. Another study suggests heteropaternity among twins in the general population to be as high as one pair in four hundred and estimates that one in twelve sets of fraternal twins are the result of superfecundation, if not superfecundation by two different fathers.
It's interesting to think about exactly how this might cause an evolutionary twist. The traditional view is that males could have offspring with many females at once, but females could only get one genetic mate at a time. With superfecundation, a female has a chance to combine her DNA with many partners as well. Who knows what kind of evolutionary effect that has had on animal, and perhaps human, development?