The finding that the SD antigen HL-A8 is associated with myasthenia gravis has raised the question as to whether or not there is also an association between some specific LD antigen and myasthenia gravis and/or HL-A8. MLC technique was used to study LD antigens in 33 myasthenic patients. The cells of a myasthenic patient who was homozygous for both HL-A and LD products were used in MLC tests as stimulators with HL-A8 bearing cells from 24 myasthenics, their 17 relatives and 16 controls and to non-HL-A8 cells from nine myasthenics and 16 controls. At least three different LD genes were found to be associated with HL-A8. One of them, called LDm' was present in 17 (63 %) of the 27 HL-A8-bearing haplotypes of myasthenics, and in nine (47 %) of the 19 HL-A8-bearing haplotypes of controls. In our study LDm was not found in the 84 haplotypes devoid of HL-A8. LDm is strongly associated with HL-A8 and through this also with myasthenia. Calculated from phenotype frequencies for HL-A8 and its association to LDm' the LDm frequencies are 9 % in the control population, 30 % in myasthenics and 48 % in young females with onset of myasthenia below 35 years. LDm had no indendent correlation with any of the clinical parameters of myasthenia gravis, even though is was found more often in HL-A8+ females with the onset of myasthenia before 35 years than in the whole myasthenic group. LDm appears to be similar to LD8a antigen.