Felix Bernstein and the first human marker locus.

A CCUSTOMED as we now are to thousands of polymorphisms useful as human chromosome markers, it is hard to realize that in the first quarter century of Mendelism there was only one good marker. I t is a l l the more remarkable that its simple mode of inheritance was not understood until the trait had been known for 25 years. The human blood groups were discovered at the turn of the century (LANDSTEINER 1900), the year of the rediscovery of MENDEL’S laws. Two others, MOSS and JANSKY, independently designated the four groups I , 11, 111 and IV. This was fine except that MOSS’S I was JANSKY’S IV, and vice versa, leading to considerable transfusion confusion in the early days. LANDSTEINER was astute enough to invent a nomenclature that reflected the A and B antigens (and genes), or their absence, and this gradually prevailedvery gradually in at least some benighted quarters, for I remember learning about the four groups I , 11, 111 and IV in the 1930s. I t was quickly recognized that the blood groups were inherited and the first plausible hypothesis was put forth by VON DUNGERN and HIRZFELD (1910). Studying 72 families with 102 children, they hypothesized that the A and B antigens were produced by two independent dominant alleles (Table 1). The correct hypothesis of multiple alleles at one locus was not demonstrated until BERNSTEIN did so in 1924 and 1925. Let 0, A, B and AB stand for the proportions of the four types. On the prevailing two-locus hypothesis with Hardy-Weinberg proportions and linkage equilibrium it is clear that the expected frequency of 0 X AB equals that of A X B, for both are equal to p:p:(l p:)( 1 p;). BERNSTEIN used the equivalent relationship (A + AB)(B + AB) = AB. Applied to the observed proportions i n Table 1, the left-hand quantity is (0.500)(0.284) = 0.142, about twice the observed proportion of AB, 0.078.