It is probably appropriate that the CIA's massive mind control and drug experimentation program was launched because of a mistake, and a ludicrous one at that. In 1951 a U.S. military attache reported that the Russians had purchased 50 million doses of LSD from the world's only manuacturer of the hallucinogen, the Swiss pharmaceutical firm Sandoz. The problem was that the attache got his numbers scrambled and the real Russian drug buy was 50 doses of LSD. By the time the mistake was noticed, MK-ULTRA, as the CIA's code-named program was known to a select handful, was under way. From all the available evidence things never got much better.

The key here is the evidence. Despite its bulk, the 16,000 pages of Xeroxed material the CIA chose to share with the public about its decades-long mind control fixation are just too skimpy to get a clear picture of what the intelligence agency's scientists were up to. More ominous, they reveal almost nothing at all about what the CIA or the military did with the results of their mind control work.

John Marks, in his book "The Search for the 'Manchurian Candidate,'" readily acknowledges that critical information gap. After painstaking research among the CIA's heavily censored files and interviews with those principal figures who would talk, he admits "the final result is not the whole story of the CIA's attack on the mind." There are some who could tell, but they, for reasons of their own, have chosen to remain silent.

Still, in the absence of that information, what remains makes for fascinating reading, not only for the buffs who follow the CIA's doings the way people sometimes chase fire engines, but also for those who want to know the limits to which the Cold War of the 1950s and an emerging new science of behavior control pushed their government.

Marks' book is the best view so far of that area of marginal ethics where science and fantasy were allowed to merge. In its frantic search to uncover the workings of the human mind and to manipulate enemies, friends and strangers, the CIA tapped every source it could find, from magicians to psychiatrists. The results were predictably bizarre.

There were terrifying trips into the unknown world of LSD for those unsuspecting souls who chanced to share a drink or a woman with the CIA's operatives. There were psychiatrists who practiced "depatterning" -- a mild word for the horror of repeated drug and shock treatments aimed at rearranging the human mind. And there were -- and, according to Marks' research, may still be -- CIA-backed efforts at genetic engineering, the logical end to all this tinkering.

But there never was a "manchurian Candidate." Despite its efforts and huge outpouring of concealed funds, the CIA could never produce the programmed killer it sought. That is where the fantasy stopped. In the end the entire exercise was no more than the LSD scare that touched it off -- a mistake.

Perhaps, if something good had come out of it all, a spinoff like the synthetics or gadgets spawned by the space program, it might have had some value. But, as Marks points out, the best the CIA could spawn with its channeling of drug research into the nation's best colleges, was the LSD generation of the 1960s.

Last year, the CIA notified Marks that it had uncovered yet another cache of documents on its mind control work. They fill 130 boxes, nearly 20 times the amount the CIA found when it was searching for its MKULTRA records. Stretching into the 1970s they cover a second-generation effort by the CIA to reach its elusive target, this time using far more sophisticated methods than were available to the men from MK-ULTRA. When drugs failed them in the 1950s and 1960s, the CIA's scientists turned to computers and electronics in the 1970s to search for the Manchurian Candidate. Maybe this time, if we really want to know, we'll find out whether they achieved their goal.