KIRKEGAARD-SORENSON AND MEDNICK ARRIVED AT THESE MAIN FINDINGS: (1) THE PERCENTAGE OF MOTHERS REGISTERED WITH THE POLICE WAS HIGHER FOR THE HIGH-RISK GROUP THAN FOR CONTROLS, (2) THE PERCENTAGE OF FATHERS REGISTERED FOR CRIMINALITY DID NOT DIFFER BETWEEN THE GROUPS, AND (3) THE PERCENTAGE OF CHILDREN REGISTERED WAS GREATER FOR THE HIGH-RISK GROUP. THE RESEARCHERS TRIED TO INCLUDE AS MANY SIBLINGS AS POSSIBLE IN BOTH GROUPS AND SUCCEEDED IN INCLUDING A GREATER NUMBER IN THE EXPERIMENTAL GROUP THAN IN THE CONTROL. THE INCLUSION OF SIBLINGS AS SUBJECTS INTRODUCES THE POSSIBILITY OF EFFECTS DUE TO INTERACTION BETWEEN SIBLINGS. SUCH INTERACTION WOULD NOT AFFECT EXPERIMENTAL AND CONTROL GROUPS EQUALLY BECAUSE THE NUMBER OF SIBLINGS PER FAMILY INCLUDED IS NOT CONTROLLED. FURTHER, SEPARATIONS OF CHILDREN FROM PARENTS FOR LESS THAN 5 YEARS WAS NOT CONTROLLED, ALTHOUGH THIS IS A FACTOR ON WHICH HIGH-RISK AND CONTROL SUBJECTS ARE LIKELY TO DIFFER. CONTROL ANCESTORS WERE SCREENED FOR MENTAL ILLNESS, BUT THE HIGH-RISK ANCESTORS WERE NOT. THUS, THE POSSIBILITY OF ANCESTOR INFLUENCES OTHER THAN THE MOTHER'S SCHIZOPHRENIA IS NOT RULED OUT. FINALLY, THE REJECTION OF CONTROL CHILDREN WITH MENTAL ILLNESS IS MISGUIDED AND COULD SELECT AGAINST POTENTIAL CRIMINAL CONTROLS. STUDIES HAVE SHOWN THAT A LARGE NUMBER OF ANTISOCIAL ADULTS WERE EARLIER REFERRED TO CHILD GUIDANCE CENTERS FOR JUVENILE ANTISOCIAL BEHAVIOR. REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--DAG)