The approach to the case material will now be different, as the two samples will be compared for various specific characteristics. Complete individual histories will no longer be presented but the material will be scrutinized for the presence or absence of certain features (or of constellations of these) that emerged from the pilot study and that were also observed in the three case studies just presented. \Vherever possible, the findings are set out in the form of simple tables extracted from the case histories, and typical examples will illustrate these findings. This inevitably involves, for purposes of description and measurement, the artificial isolation of items of behaviour that in reality are interwoven with others. But only by defining certain characteristics and examining their interrelations can such comparisons be made. If sixty-four complete case histories were presented such comparisons would of course not be possible. However, by comparing first some characteristics of the mothers and fathers in both samples, followed by a discussion of various phenomena of their marital relationships and of the young men's childhood, it is hoped to build up gradually a picture of the 'D u family'. Although tables are presented for clarity and for the indication of trends, tests of significance have not been carried out. The numbers are small and the variables are highly interrelated and such tests might in fact give an impression of spurious precision.