It has been suggested that translocations, and perhaps other chromosome rearrangements, disturb meiotic disjunction of uninvolved chromosome pairs and predispose to trisomic offspring. If so, then one would expect an excess of translocations not involving chromosome 21 among the parents of regular trisomic Down's syndrome patients. Such translocations have been reported, but mostly as anecdotal single case reports or very small series. In an attempt to collect a larger series, a collaborative study of regular Down's syndrome families was made in southern England. This was retrospective, and covered periods of 7 to 10 years since 1970. The number of regular trisomy families investigated was 1454. Only 945 of the 2908 parents were karyotyped, and 10 balanced reciprocal translocations not involving chromosome 21 were identified, together with one Robertsonian (13q14q). Expressing these as percentages of the parents tested (945), prevalences are as follows: reciprocals 1.06%, Robertsonians 0.11%, and all translocations 1.16%. Expressed as percentages of the total parents (2908), tested and untested, the prevalences are 0.34%, 0.03%, and 0.37% respectively. The 'true' prevalences, that is what would have been found had all parents been tested, must lie between these two sets of figures. The prevalence of reciprocal translocations exceeds that found for consecutive banded newborn infants, which is 0.16%, and this excess may reflect a real interchromosomal effect. Robertsonian translocations in the banded newborn series are at a frequency of 0.11%, identical to that found in the tested parents of regular trisomics. Interpretation of these figures is critically dependent upon the real prevalence of translocations among the newborn, estimates of which increase as technical methods are improving.