In the nearly last two years, using ten different media combinations in our research for the fragile X, we have observed cytogenetic variations, the knowledge of which may be useful for correct diagnosis. This experience, based on more than 6000 metaphases in 41 cases, includes 5 typical probands, 4 obligate and 4 possible heterozygotes, 19 deficient psychopaths, and 9 controls. Three main techniques were retained as shown in the text. In this study a fragile site was documented on 239 chromosomes of the C or X group, among which 180 could be specifically identified by trypsin Giemsa banding. Of these 180 fragile sites, X involvement was shown in 101 cells, and 55 other lesions were found to affect a chromosome 6. In our experience, none of 1180 cells from 9 control individuals were positive. It thus may be that under rigorous culture conditions the occurrence of one single fragile X at q27 or q28 is suggestive of the presence of the underlying mutation. In 19 cases studied because of mental and psychotic problems, nonetheless considered as clinically negative, only 2 out of 2510 cells had a fragile X, whereas frequency among cytogenetically verified X in our positive cases varied from 4 to 20% cells. We have come to distinguish five variations in the cytogenetic aspect of this site on the X, shown from A to E in a figure, two of which (D and E) may not have been described before. In rare cases this fragility usually seen as a chromatid or isochromatid gap may be present as a break with a resulting "double minute" found elsewhere in the metaphase field.(ABSTRACT TRUNCATED AT 250 WORDS)