The diagnosis of zygosity in twins

Uniovular and diovular twin pairs can be classified with at least 90% accuracy by careful questioning or by ratings by observers of similarity in appearance, eye color, ear conformation, and the like. This article outlines the methods required for highly accurate zygosity diagnosis using blood typing, fingerprints, and anthropometry. The logic behind the use of discrete Mendelian phenotypes is explained and tables are provided which give the odds against a dizygotic twin pair showing concordance on eight blood group systems, four serum proteins, and six red blood cell enzymes. These tables are suitable for use with American twins of European ancestry; the method of generating such tables is explained so that similar tables based on different gene frequency estimates or for additional genetic markers can be easily constructed. The logic behind the use of metric, polygenic traits in zygosity diagnosis is also presented, together with tables giving the odds against a dizygotic twin pair showing various degrees of similarity in ponderal index, in cephalic index, and in Slater's Z score, a measure of fingerprint similarity. Taken together, the 18 serological markers should misclassify dizygotic twins as monozygotic fewer than two times in 1000. With the three anthropometric variables, this average probability of misclassification drops about 1 order of magnitude.