The determination of twin zygosity by bloodtyping is reconsidered, and the model for the individual case is reformulated. The crucial diagnostic question may be phrased as follows: Given the particular array of bloodgroup phenotypes that the twins display and are concordant for, how might this array have been obtained by a pair of dizygotic twins, and how might the array have been obtained by a monozygotic pair? The solution yields a differential probability value that is uniquely tailored to the actual phenotype array shown. The procedure offers a coherent and more direct method for arriving at the needed probability figures, and it is recommended to supersede previous methods. Some similarities and differences between the methods are discussed.

The formulas needed in the determination of monozygosity in twin pairs using genetic markers are derived and presented. A general formula for the calculation of concordance probabilities independent of gene frequencies and allele number is derived, enabling either manual computation or computer programming for any Mendelian markers usable in twin zygosity diagnosis.

More and more genetic markers usable for twin zygosity determination have become available. A relatively small number of markers is sufficient to achieve a satisfactory probability of correctly classifying a twin pair. Previously only the genetic properties of markers have been considered when choosing the markers to be determined. A cost-effectiveness analysis, which considers both genetic properties and relative determination costs of markers, is presented and illustrated with data from the Finnish Twin Registry studies.

The dependency of probabilities of phenotypic concordance of gene frequencies in three-allele genetic systems is presented. A graphical display enables the rapid comparison of the relative effectiveness of different systems, taking into account dominance relationships within each genetic system. Four or more allele systems can also be approximated, while two-allele systems are considered to be special cases of three-allele ones.