The idea of "mirror-image" twins always gets me excited, but maybe not quite like most folks. This idea goes back to a biologist named Spemann who published most of his work in German in the 1920's. He worked with newt embryos. Newts are small amphibians with legs, a kind of salamander. In some parts of this country they are called 'lizards' and used as fish bait. Lizards are reptiles, with dry scaly skin. Newts and other salamanders are amphibians, with smooth moist skin, who lay their eggs in water and are tadpoles before they grow legs. They are an attractive subject organism for developmental biology research, because the fertilized eggs can quite readily be watched closely while developing in an ordinary dish of water on an ordinary lab bench.

Spemann experimented by tying fine hairs around some newt embryos. If he pinched an embryo all the way in two, with only very rare exception it died. If, however, he stopped tightening the hair while the two halves were still held together by a bridge of cells, and put it back in the water, some of those didn't die. Some of the ones that lived developed into twins. In some of those twin pairs, the normal asymmetries of the internal organs were reversed in one of the animals; heart right- instead of left-of-center, stomach to the right and the larger lobe of the liver to the left, and so on. So he had produced monozygotic twin embryos, by a process which caused a reversal of normal developmental asymmetries in one member of each pair. This is where the whole idea of "mirror-image twins" comes from. There is no reason to think natural human twinning has anything to do with any process at all like that. The cells of one embryo have to separate to become two separate people, but there is every reason to believe that is a consequence of twinning and not a cause. It happens after the cellular event that decides twinning, not before.

Newman and some other researchers who were contemporaries of Spemann took the newt results to represent a mechanism of twinning which could be substituted for our frighteningly complete ignorance of the mechanisms of human twinning. So, from that time forward, "everybody knows" that monozygotic twins are produced by the "splitting" of an embryo, presumably mechanically, in some process that must somehow closely resemble tying it in half with a child's hair. And, if it happens late enough in development, it should result in mirror-image reversal of normal developmental asymmetries. That's what "everybody knows", and it's nonsense.

In all of the records of medicine and science, the entire number of twins (or any other humans) ever found to have inverted positions of the internal organs like Spemann's twin salamanders had is very small. On the other hand, a great many pairs of twins do differ in motor hand preference - handedness. That includes many pairs who otherwise resemble each other very closely. So, drawing from Spemann's results, the logic goes, sometimes the "identical" twinning process must occur late enough to result in "mirror-image" handedness. It got even crazier from there. When I got into this business in the mid-1970s, the literature said that twins were more often lefthanded than singletons, and that this was primarily due to the monozygotics, because of their mirror-imaging when their "split" was late. After a fair amount of careful work, it turns out that every paper which claimed a statistically significant excess of monozygotic twins among lefthanders had included data gathered in the 1920s and 1930s. The special thing about those data was that any twin pair who differed in handedness was assumed to be - not just monozygotic, but "late-splitting" - monozygotic. If such twins had different- colored hair and eyes, clearly different heights or weights or body types, even different sexes, these oddities were considered to be the obvious consequences of the embryonic disruption caused by the late splitting. Blood typing was just barely becoming understood and was not available for use in the zygosity typing of twins. If we count every dizygotic pair which includes only one lefthander as monozygotic instead of dizygotic, then yes indeed we compute a huge excess of lefthanders among the monozygotic .... and a shortage among dizygotics!, compared to the general population.

What My Studies Showed...