SUMMARY We previously found that nearly 35% of the Bantu had no antigens at the first locus detectable with the antisera available to us. The sera of 1,004 Bantu women were therefore screened and those containing antibodies were tested against 50 unrelated Bantu doners in parallel with known antisera, using the 2-stage microlymphocytotoxictest. Antisera for Te 63 and Te 66 were obtained from the National Institutes of Health (NIH). These two specificities almost completely filled the gap previously found at the first locus. Two hundred two of the 1,004 Bantu sera contained HL-A antibodies. but only one had the specificity anti-Te 63. One hundred twenty selected sera were then used to test a further 100 Bantu and 100 Caueasians. We tested for antibodies at the first locus. HL-A1,2,3,9,10,11, W28, W19, Te 63, and Te 66; and at the second locus we tested for 12 antibodies. HL-A5,7,8,12,13,W5, W22, W15, W17, W10, and W27. HL-A1 has a very low frequency in the Bantu (5%) and no Bantu were found with HL-A11, while HL-A3 had a lower frequency (12%) than in Caucasians W28 (19%), HL-A9 (17%), HL-A10 (23%), Te 63 (13%), and Te 66 (31%) all had higher frequencies in Bantu than in Caucasians. At the second locus, the frequency of HL-A7 was only 11% but W22 was found in 34% of the Bantu (5% in Caucasians). Thirty-five anti-HL-A12 sera could be divided into two groups, one reacting as a short anti-HL-A12