Four days of testimony last week by the first child to testify in the McMartin Pre-School molestation case illuminated problems that experts say are inherent in prosecuting such cases: The key witnesses are young children who are easily confused, tire quickly and, if the alleged molestations occurred, are deeply troubled.

Some of the child's answers on the witness stand appeared contradictory, and defense attorneys say the testimony proves their contention that the alleged incidents never took place. One psychiatrist who specializes in the treatment of sexually assaulted children, however, said the boy on the stand is being "tricked."

The youngster, 7, who is to resume his testimony in the pretrial hearing today, has testified that teachers at the McMartin school made the children play "naked games" and "touched" their genitals.

But later, the boy, called John Doe No. 7, seemed to retract those statements, and some news accounts reported that he said he had never been afraid at the school. In fact, the child says very little in response to questions from adults. His replies, more often than not, are "Yes," "No," or "Don't remember."

After a long series of questions about "games," including such childhood games as ring-around-the-rosy and hide-and-seek, defense attorney Daniel Davis asked the boy, "You weren't ever afraid when you went to the preschool, were you?"

"No," the boy said.

Dr. Roland Summit, a psychiatrist at Harbor-UCLA Medical Center and specialist in the psychology of sexually assaulted children, said in an interview, "It isn't fair to trip a child up with leading questions where the outcome is obvious. When you say, 'You weren't scared, were you?' it appeals to both the child's wish to please and the child's wish to be strong, and to his desire to come across well."

Summit said children who have been molested feel "shamefully complicit, and sure they're going to get in trouble." Therefore, "the most universal characteristic among these children is denial. Getting a child to tell you about sexual assault is very difficult . . . and, having told, the child may feel so anxious and guilty and fearful that he'll immediately take it back. Retraction is just about as universal as denial."

Defense attorney Bradley Brunon opened his cross-examination last week with questions that drew giggly "no's" from the witness: Had he ever ridden a bear, or seen 77-year-old defendant Virginia McMartin roaring around Manhattan Beach in her wheelchair? The boy, still giggling, agreed "Yeah" to Brunon's statement that the therapist treating the child "thought Ray Buckey, another defendant touched you on the penis?"

"She kind of made you say that, didn't she?" Brunon then asked.

"Yeah," the child agreed.

Defense attorney Forrest Latiner said this proves that the alleged molestations were nothing more than "children's imaginations running wild," after "bizarre adventures" had been suggested to them by psychologists.

But Summit contends that a child on the witness stand has "a thoughtless wish to please the questioner. It's very hard to argue with an adult, and relatively easy to accede to what he wants you to say." He added, "If you ask leading questions, the answers to which have pleasant connotations, you get the child more into denial, more into a boy's natural macho. Then you say, 'You weren't scared, were you?' And then you've got him. You've tricked him."

Pediatricians have testified that they found physical evidence of molestation in 37 of the 41 children who are complaining witnesses in the McMartin case. One doctor said she had found scars indicative of sodomy in John Doe No. 7.