BACKGROUND: Do the population relationships obtained using DNA or blood group plus protein markers remain the same or do they reveal different patterns, indicating that the factors which influence genetic variation at these two levels of analysis are diverse? Can these markers shed light on the biological classification of the Aché, a Paraguayan tribe which only recently established more permanent contacts with non-Indians? SUBJECTS AND METHODS: To consider these questions we typed 193 individuals from four Amerindian tribes in relation to 12 Alu polymorphisms (five of them never studied in these populations), while 22 blood group plus protein systems were studied among the Aché. These data were then integrated with those previously available (blood groups plus proteins) for the three other populations. DNA extraction and amplification, as well as the other laboratory procedures, were performed using standard methods currently in use in our laboratory. The genetic relationships were obtained using the D(A) distance, and the trees were constructed by the neighbour-joining method, both developed by M. Nei and collaborators. Reliability of the trees was tested by bootstrap replications. Other population variability values were also determined using Nei's methods. RESULTS: Alu polymorphism was observed in all populations and for most of the loci; in the seven systems from which we could compare our results with those of other Amerindian groups agreement was satisfactory. Unusual findings on the blood group plus protein systems of the Aché were a very low (5%) HP*1 frequency and the presence of the C(W) phenotype in the Rh blood group. The intertribal patterns of relationship and other aspects of their variation were remarkably congruent in the two sets (Alu; blood group plus protein) of systems. CONCLUSIONS: The answer to the first question posed above is affirmative. However, the problem of whether the Aché derived from a Gê group that preceded the Guarani colonization of Paraguay, or are just a differentiated Guarani group, could not be answered with the genetic information available; the second hypothesis seems more likely at present, but the point to be emphasized is the striking genetic distinctiveness of the Aché as compared to other Amerindians.