Ground rules directions are given to children in forensic interviews to explain what is expected of them, and to reduce their tendency to acquiesce to erroneous or incomprehensible questions. Ground rules may also be necessary when children provide testimony in court. Drawing on research conducted for the Australian Royal Commission into Institutional Responses to Child Sexual Abuse, the present study examined the use of ground rules directions delivered in court in 52 trials by 24 presiding judges in three jurisdictions to 57 child complainants (aged 7-17.5 years). Eleven categories of rules were identified. The number of words spoken to deliver each rule was counted, and grade-level readability scores were calculated as a proxy for the complexity of the ground rules. When judges asked comprehension or practice questions, the question types were coded. More than one third of the children (35%) received no ground rules directions from the judge; the remaining 65% received directions on an average of 3.5 types of ground rules out of a maximum of 11 types. While comprehension questions were common, practice questions were rare. Comprehension questions were most often presented in a yes/no format that implied the expected response, although this form of question is unlikely to provide an effective assessment of a child's comprehension. Neither the number of rules delivered nor the number of words used was related to children's age. Implications for children's court testimony are discussed.