In the absence of an unambiguous test for identifying Duchenne muscular dystrophy (DMD) heterozygotes, methods are needed for combination of the results of individually equivocal tests as effectively and rationally as possible. Tw used logistic discrimination to assess the effectiveness of measurements of serum creatine kinase, hemopexin, pyruvate kinase, and lactate dehydrogenase alone and in various combinations in identifying DMD carriers. We analyzed 127 serum samples from 63 normal female controls (20-40 years old) and 67 from 38 obligate DMD carriers. The best two tests to use in combination were creatine kinase and hemopexin, and these two, with lactate dehydrogenase, were the best three. t the 95% level (with 5% of controls misclassified), 54% of the carriers were identified by CK alone, whereas 88% were identified by means of the four tests. Although a small proportion of known carriers still cannot be identified, application of the four tests to a group of 45 possible carrier mothers of isolated cases of DMD resolves the population into fairly discrete "normal" and "abnormal" subgroups. Thus, if bias of selection can be eliminated, application of logistic discrimination may permit a direct estimate of the proportion of mothers of affected boys who are homozygous normal.