TWINS AND HEREDITY 481 



and not on the other, or else it may express itself much more exten- 

 sively on one side than on the other. For example, there are certain 

 finger-print and palm-print patterns that may occur on the right hand 

 of an individual and not on the left, or may be present in one form on 

 one hand and in another form on the other hand. Now, if identical 

 twins are derived from the separated right and left halves of an em- 

 bryo that typically would produce a single individual, such twins may 

 be expected to be only as similar as are the two sides of a single indi- 

 vidual. In view of this circumstance, we would be inclined to modify 

 the working hypothesis of Siemens to the extent of saying that charac- 

 ters which are always the same in both individuals of a pair of identical 

 twins are undoubtedly hereditary, but that the failure of certain 

 characters to be the same in both members of a pair does not prove 

 that such characters are non-hereditary. That this point is well taken 

 is demonstrated by a study of armadillo quadruplets, which are un- 

 questionably monozygotic. In these animals it frequently happens 

 that a well-defined peculiarity, such as a local doubling of a band in 

 the armor of the back, a rare anomaly known to be inherited from 

 the mother, does not appear in all four offspring from a single zygote. 

 One of the four may have it on the right side, another on the left side, 

 a third on both sides, and a fourth may fail to show any sign of it. 

 Yet it would be foolish to claim that, because the character failed to 

 appear in one out of the four quadruplets, such a character is non- 

 hereditary. 



Using the working hypothesis, however, in its original form, Sie- 

 mens has been able to show that a large number of pathological condi- 

 tions are hereditary. He has specialized on birthmarks (naevi) and 

 has shown that various types of naevi constantly appear in both twins, 

 if they appear at all. He has shown that even moles and freckles are 

 hereditary. 



Bonnevie, Wilder, and I, working independently, have shown that 

 various peculiarities of ringer prints and palm patterns are hereditary. 

 This has been done by the method of determining the coefficients of 

 correlation between right and left sides of the same individual. Com- 

 parisons of the extent of correlation in identical twins is made with 

 those in fraternal twins, sibs, and unrelated persons. For example. 

 I have studied statistically the numbers of ridges in the finger patterns 

 of fifty pairs of identical and of fifty pairs of fraternal twins, and the 

 coefficients of correlation obtained are as follows: 



