The first version of the story: Oscar Pistorius' bedroom contained two boxes of steroids, found after the murder of his girlfriend, Reeva Steenkamp.

At the prompting of the prosecutor, the lead detective changed that testimony: No, the substance was not steroids; it was basic testosterone.

The third version, from the Olympian's defense: No, it was an herbal remedy.

The fourth, from a police spokesman: The substance has not been identified. A lab is still reviewing it.

This dizzying volley will end soon enough. The debate about 'roid rage, a story with countless versions, will go on past the trial, apparently without end.

Even if the prosecutor could show that Pistorius possessed illicit hormones, used them in huge quantities, exhibited horrible mood swings while under their influence, and often acted violently while taking the drugs, the existence of 'roid rage would still be a theory, not an established fact.

Consider a 2005 study that assessed six men incarcerated for homicide and violent assaults, all of whom reported using anabolic steroids. "These men were described as experiencing psychotic symptoms, with 'stereotypic qualities of irritability, aggressiveness and grandiosity,'" said a summary of the work in a 2009 article by a group from Harvard Medical School. "The authors report that all six men returned to a normal mental status within 2 months of stopping AAS [anabolic-androgenic steroids]."

Psychiatrist Ryan C.W. Hall and his father, Richard, did that research. Both work as forensic psychiatrists in the criminal-justice system. Getting even six subjects, a small sample size, to discuss steroid use honestly can be difficult. For some, the drugs represent shame. For others, they might represent the most cynical of opportunities -- a mitigating factor in a crime and a lighter prison sentence.

"Sometimes, they [the subjects] kind of taint the facts a little to fit their situations. You have to look for as much medical history and [lab results] as you can find,'' Hall said. "And even with people who are being honest with you, a lot of times they don't know what they're injecting. They're buying stuff that's a yellow vial that has fluid in it, and they're told what's in it, but they don't really know. I've seen people who thought they were taking one steroid and tests showed they had taken another one. So this is a very hard area to study."

Just last month, a defense based on "steroid abuse and its impact on the defendant's behavior'' gave way to a first-degree murder plea from a former sheriff's deputy in Virginia who shot his ex-wife in a parking lot. A search of the killer's home after the shooting turned up synthetic testosterone.  His attorney said he planned to introduce evidence linking the hormone and violence during the sentencing phase.

The anecdotal evidence also includes:

• James Batsel IV, a police officer in an Atlanta suburb, who shot and killed a nightclub owner while participating in a spree of burglaries with other members of the force. Steroid use was said to be rampant among the police, and Batsel's attorney blamed the drugs for the murder. He received a life sentence, and his father later told ABC News that his son had become extremely volatile. "He had a temper you would not believe," the father said. "He had a dog that he just loved -- and he took that dog out and shot it."

• A 16-year-old boy, bullied most of his adolescence, who started taking the drugs to bulk up. He ended up murdering his girlfriend, according to research compiled by psychiatrist Harrison Pope, of Harvard and McLean Hospital in Belmont, Ma. In a 2003 article, Pope and co-authors quoted from a South African study of 12 bodybuilders using steroids. "One subject repeatedly threatened to shoot his mother while taking [steroids]; another kicked car doors in a parking lot continuously for six minutes. The authors noted no comparable behavioral changes in a control group."

• The ultimate guinea pig, Stuart Stevens, who wrote a piece for Outside Magazine in 2003 about taking a course of PEDs for eight months. Stevens found his introduction to synthetic testosterone very rage-free.

I walked out of Dr. Jones's office smiling broadly, then waited for a werewolf surge. And I waited. But the truth is, I didn't feel much of anything. No irresistible bursts of lust or rage, no particular feelings of omnipotence. That afternoon I went home and celebrated my newfound energy and aggression with a long nap.

Later, in an interview, he did describe feeling uncharacteristically cranky in traffic jams. "I could feel myself getting upset in a way I normally wouldn't have," he said. Stevens said he didn't worry about health effects, because he took such low dosages that his visits to PED message boards turned into sessions of mockery for his wimpy approach. One regular called him a "girlie man.''

Last year, Stevens acted as the chief strategist for Mitt Romney's presidential campaign

• Former Steelers lineman Steve Courson, who openly discussed his steroid use after retirement and testified in front of Congress. He once told me that he always felt irritable during his cycles of drug use, and that he worried about the effects on people predisposed to mental illness and emotional disorders. Courson did not tend toward overstatement, even though he made a big leap just by telling the truth while others hid their drug usage.

He had a heart condition that put him on a transplant list until daily medication quelled the worst of it. He wondered whether the steroids might have damaged his heart, but he also believed that just bulking up to nearly 300 pounds could have been  responsible.

"People wanted me to say that these things were like cyanide killers," Courson said in 2003, more than 15 years after he first blew the whistle on himself and his sport. "And I wouldn't do it, even back then."

Even Courson's concerns about mood disorders did not advance the idea that, with only a moderate steroids dose, a placid young man would instantly transform into The Terminator.

"Steve, I thought, articulated it best," said retired Penn State professor Charles Yesalis, a longtime friend and co-author of a book with Courson (who died in 2005 when a tree he was cutting down fell on him). " 'If you were an a--hole before you used steroids, you'd be a bigger a--hole after you used them.' I think that says as much as any clinical study."

***

In 1996, the New England Journal of Medicine cited a 10-week study in which 43 men received either 600 milligrams of testosterone per week or placebos. Psychological tests of the users and interviews with their spouses indicated no elevated aggression.

Among steroid users who are mentally healthy, "testosterone doesn't turn men into beasts," declared researcher Shalender Bhasin of L.A.'s Drew University. The study also became one of the earliest to establish a link between the use of steroids and muscle growth. Remember, this was in 1996.

The medical profession came very slowly to the belief that these drugs accomplished what many gym rats and East Germans and Steve Courson already knew they could. In the late '70s, the American College of Sports Medicine issued a position paper stating that the drugs were ineffective at building muscle mass. A decade later, realizing that lab dosages did not approach the amounts consumed by athletes, the organization had to retract its official skepticism.

"The failure of the medical community to recognize the efficacy of (steroids) for such a long period has represented an embarrassment even to this day,'' said a medical journal article co-authored by Pope in 2009.

The same article cites four laboratory studies, involving a total of 109 male volunteers dosed at 500 milligrams or more, and says that a total of 4.6 percent "developed hypomanic or manic syndromes during blinded (steroid) administration whereas none showed such syndromes on placebo."

Less than 5 percent may seem negligible, but as Dr. Hall pointed out, this is a slippery subject. For one thing, the Drew University subjects are part of 109, and the original press coverage of that research made the effect seem nonexistent.

"The 4.6 percent figure likely underestimates the true prevalence of such syndromes among illicit … users because illicit users may reach far higher dosages than those administered in the laboratory," the article in the journal Hormones and Behavior states. Also, " … laboratory studies typically screen volunteers to exclude individuals with existing psychopathology or substance abuse, but illicit [steroid] users do not screen themselves with such care."

The researchers cannot ethically dose their subjects at the level they believe elite athletes reach. "Naturalistic" studies take another tack, reviewing the experiences of people who choose to dope on their own.

A 1994 Harvard study gathered information from 88 users or former users and 68 non-users found at gyms in Boston and L.A. Using diagnostic interviews, the researchers determined that 23 percent of the active users "displayed a major mood syndrome.'' The number, as described in a 2003 article in Pyschoneuroendocrinology, dropped to 6 percent when the users went off the drugs and to 4 percent among the non-users.

Other naturalistic studies diverged greatly from that finding, including one that recognized "virtually no psychiatric changes in [steroid] users." But Pope and his colleague, David L. Katz,  determined that the studies showing no effect had collected information from people using lower dosages than most athletes would.

In the end, their paper recommends that "the possibility of (steroid) use should always be considered and, if possible, assessed via urine testing" whenever the suspects in violent crimes are athletes or have unusually muscled physiques.

In the case of pro wrestler Chris Benoit, who killed his wife and young son before ending his own life in 2007, the police tested Benoit's body and discovered synthetic testosterone in his system, along with the painkiller Vicodin and anti-anxiety medication Xanax. They also found steroids in the house. However, the deliberate nature of the murders suggested premeditation, rather than uncontrollable rage.

Then Benoit's brain was examined by researchers who found severe damage, suggesting that repeated blows to the head in his wrestling career could have produced a form of dementia.

The case had too many answers, and no absolute proof of anything other than the fact that Benoit had killed three people.

In the end, the South African court will be primarily focused on what Pistorius thought as he shot through the bathroom door at his girlfriend, and on whether the angle of the entry of the bullets indicates that he put on his artificial lower legs before he left the bedroom. He says he did not.

Regardless of what turns up in the lab analysis of the herbal remedy/testosterone/steroids, one American doping expert believes that hormones could not possibly have played a role in this violence.

"For any drug-tested athlete, the use of testosterone would have to be very low to fly under the [the initial screening for disproportionate levels],'' said Yesalis, who has testified in front of Congress a half-dozen times. "So no, I'm sorry, I would argue strongly that that wouldn't cause a fit of rage."