TWINS AND HEREDITY 479 



process of twinning in man must be essentially the same as in the 

 armadillo. Now, in the armadillo, every step in the double twinning 

 process has been observed, and we have some rather well-supported 

 theories as to why the egg divides into four individuals instead of 

 developing into but one. There is in the armadillo a well-defined 

 hitch in the process of the attachment of the young embryo to the 

 membranes of the uterus. For a month or more after the embryo 

 descends the egg tube into the uterus it lies free in the uterine cavity, 

 whereas it should normally have attached itself immediately to the 

 maternal tissues so as to receive food and oxygen from the maternal 

 blood. During this period the young embryo is completely stopped 

 down and is unable to develop at all. Its vitality becomes so lowered 

 that, when finally the mechanism of attachment does right itself and 

 the embryo is afforded the chance to feed and respire, the fires of life 

 have burned so low that the embryo cannot maintain its unity, but 

 starts to develop first at two regions and then at four. These four 

 growing points develop into the four separate embryos. 



That a lowering of the rate of vital activity during early develop- 

 ment is able to induce twinning has been demonstrated experimentally 

 in several species of animals, notably certain fishes and echinoderms. 

 Apparently some embryos cannot maintain their singleness under 

 conditions that lower vitality. On this account right- and left-hand 

 sides cease to co-operate to form a single individual, and each half 

 grows independently into a whole embryo. In armadillo quadruplets, 

 some hundreds of which have been examined, there are frequent cases 

 of mirror-imaging, that is, peculiarities found in the right side of one 

 twin occur on the left side of the other. Again, it has been found that 

 armadillo quadruplets are about 90 per cent identical, on the average, 

 and are always all of the same sex in a litter. 



Applying these known facts to the case of human twins, we find 

 the following correspondences. Identical twins in man average about 

 90 per cent identical; they are always of the same sex in a pair; they 

 show frequent instances of mirror-imaging. Thus, in about one-third 

 of the cases one twin is right-handed, the other left-handed; also in 

 about one-third of the cases the whirl of hair at the crown is clockwise 

 in one twin, counter-clockwise in the other. Minor peculiarities of 

 features — skin, teeth, finger prints, and palm patterns — are commonly 

 on opposite sides of the two individuals. In addition to these evi- 

 dences for the one-egg origin of human identical twins, it should be 

 said that many prenatal human twins have been observed by obste- 



