Since the time of the House Select Committee on Assassinations (HSCA), Oswald-did-it-advocates have trumpeted the neutron activation analysis test as the crown jewel of their case against the accused assassin. Former Chief Counsel of the HSCA, Robert Blakey leaked the results of the NAA testing to the press in advance of its actual presentation in the public hearings in a clear attempt to influence media coverage of his verdict against Oswald. Let me quote from The Assassinations in this regard:
Guinn's findings were very important to Blakey. He leaked them to the press early in 1978 as the final nail in the HSCA's verdict against Oswald. It was the rigorous scientific analysis that he so much admired and enthroned. And it showed that the single bullet theory was not just possible but that it actually happened.
Yet today, after the peer reviewed and published work of Erik Randich and Patrick Grant (Journal of Forensic Science, July 2006), Blakey is singing a different tune. The work of these two men has been so destructive of both the HSCA analysis and their NAA interpretation that Blakey now has termed the whole exercise "junk science". Further, the FBI has made the decision they will not use the process in court again. To understand why this astonishing retreat has taken place in broad daylight, let's go back to the beginning.
According to the Warren Commission, the FBI had done what was called "spectrographic analysis" on some of the ballistics evidence in the JFK case. According to Henry Hurt's discussion of this in his book Reasonable Doubt, both the FBI and the Commission were maddeningly vague about the results of the analysis. According to Hurt, this issue was to be addressed by the last witness called by the Commission, who was involved in the spectrographic analysis. Yet, during his interview, the commissioners never asked him a question on the issue. The Warren Report then noted that there were similarities in the metal composition of some of the bullet fragments. With the actual analysis not present and these vague generic terms in play, most considered that what the FBI did was not of any forensic value.
But it was later revealed that the FBI had gone beyond spectography to a much finer testing of the bullet fragments for trace metal testing of what the lead cores were made out of: namely neutron activation analysis. Yet although this kind of testing is much more exact for analysis of what metals are in the bullet lead and to what degree, according to Hurt, there is no mention of it in the Warren Report or the accompanying volumes of evidence and testimony. But in a later declassified letter from J. Edgar Hoover to the Commission, the so-called 1964 NAA tests were noted. Although Hoover tries to put the best face on the results, the sum total of his letter was that they were inconclusive. (Hurt, p. 81)
The whole NAA issue seemed to be a dead end. But that did not discourage HSCA Chief Counsel Robert Blakey. Blakey decided to do a "retest" of the compositional analysis of the lead cores of the bullets involved in the case for the HSCA in September of 1977. Dr. Vincent Guinn, a nuclear chemist at Cal Irvine, did the testing and the HSCA called him as a witness at its public hearings. Guinn was called upon on September 8, 1978. The difference between what Hoover had reported, or not reported, in 1964 and what Guinn and the HSCA declared in 1978 was startling. The scientific test called NAA went from being "inconclusive" to showing that:
- Only two bullets struck the presidential limo and its occupants, thereby upholding the Warren Commission.
- All bullet lead trace metal analysis showed that the ammunition came from Western Cartridge Company's Mannlicher-Carcano (WCC MC) manufacture. This would seem to link the ammunition to the alleged rifle found on the sixth floor.
- Fragments from Connally's wrist were "matched" with CE 399, or the stretcher bullet that allegedly went through Kennedy also. This would seem to indicate the Warren Commission was right about the single bullet theory.
In fact, the HSCA went out of its way to take a crack at Cyril Wecht here using Guinn's testimony. Wecht had passionately argued that one bullet could never have remained as intact as CE 399 if it had done all the damage in two men this bullet was supposed to have done. On the basis of the new NAA testing, counsel Jim Wolf asked Guinn if Wecht was right on this point. Guinn replied: "Well, I think that is his opinion; but like many opinions and many theories, sometimes they don't agree with the facts." (HSCA, Vol. 1, p. 505) This was where the matter was left at the conclusion of the HSCA in 1979. Bullet lead analysis by NAA was taken seriously and Guinn and Blakey's "findings" bolstered the Warren Commission. Advocates like Gerald Posner in his book Case Closed summarized the NAA as such: "Guinn's finding ended the speculation that CE 399 had been planted on the stretcher, since there was now indisputable evidence that it had traveled through Connally's body, leaving behind fragments." Ken Rahn and Larry Sturdivan wrote in an academic journal that, "The NAA results were the most important new physical evidence that surfaced as a result of the House Select Committee on Assassinations investigation. It knits together the core physical evidence into an airtight case against Lee Oswald."
The first serious broadside against Guinn and the NAA was issued by Wallace Milam at the COPA Conference in Washington in 1994. Milam questioned the very basis of Guinn's conclusions. Guinn had said that the metallic make-up of WCC MC could vary widely from bullet to bullet, the term for this being "heterogeneous" or irreproducible. But Guinn also added that the metallic make-up of a single MC bullet was not; the term for this was "homogeneous". But Milam showed that Guinn's data did show large variations within a single bullet, especially in the measure of the trace element antimony, which Guinn placed much weight on. He also noted that his testimony on this issue seemed to contradict a paper he wrote that very same year in Transactions of the American Nuclear Society. There he wrote that: "In the U. C. Irvine INAA background studies of the Mannlicher-Carcano ammunition, it was found that this bullet lead is remarkably heterogeneous somewhat within a given bullet." (Emphasis added) Yet, for the HSCA, Guinn seemed to place the efficacy of the findings on the intra-bullet lead being homogenous or uniform and consistent throughout. And really, if this were not the case, then Guinn's whole testimony would dissolve.
Milam's logic was penetratingly simple on this point. In the 1964 tests the FBI had taken microscopic samples from the same bullet and come up with different concentrations of antimony ranging from 636 parts per million (PPM) all the way up to 1125 PPM. With this wide range of data within one bullet, then it would be possible to match varying PPM values to differing fragments if one were to allow a large enough variance. And it appears that this variance is what led the Bureau to declare the earlier NAA results "inconclusive".
Milam's discussion was well informed, pointed, and well documented. (It is available online at the site Electronic Assassinations Newsletter, under the title "Blakey's linchpin".) Later on Art Snyder, a physicist, questioned the statistical analysis used by Guinn. Interestingly, although Guinn prepared such an analysis, when challenged to assign a probability number for the certainty of his work, he declined. (For instance, the two acoustical analysts for the HSCA gave their work a 95% probably certainty statistic.) Snyder later commented that Guinn probably did not assign such a figure because the number would have been too high, signaling a high probability of error due to the high variables involved in his findings.
In the face of these two trenchant attacks on the Guinn/Blakey analysis, Warren Commission defenders like Rahn and Sturdivan continued to defend the viability of the NAA analysis, saying that somehow Milam had misinterpreted Guinn's work. (Yet, when Milam challenged Rahn to a simple test comparing heterogeneity and homogeneity in WCC MC bullets, Rahn declined.) Rahn and Sturdivan will not be able to ignore the content of the new work by Randich and Grant. It goes beyond the estimable critiques of Milam and Snyder to question the underlying tenets of Guinn's NAA analysis.
This new analysis shows that although the critiques of Milam and Snyder are valid, they don't go far enough. The reason being that neither of them, or Guinn for that matter, was a metallurgist. Which Randich is. (And as Randich told me, Guinn should have had a metallurgist consulting him in his work.) The following startling facts have been left out of the debate over NAA.
Randich and Grant declare that a major fault in the HSCA work is Guinn's tenet that WCC MC lead "was found to differ sharply from typical bullet leads." This is not the case: WCC MC bullets do not differ sharply from most bullet leads. They are much like other metal-jacketed leads. The MC lead seemed different to Guinn because he compared it to unjacketed handgun rounds. In his talk in San Francisco of July 15, 2006, Randich said that outside of .38 and .22 handguns, most bullet manufacturers use the same lead alloy. (He put the figure at about 75%.) This is shocking. What it says is that the lead alloy for MC ammunition, far form being unique, is the same that say, Remington would use. So one of the pillars of Guinn's work, the singular identifiability of MC ammunition, has now fallen.
Randich and Grant also discussed a crucial phenomenon in lead smelting called "segregation", i.e. how the lead and trace elements distribute themselves through the heating and cooling process. During this process, the lead, because it is heavier, stays in the center, while the antimony "floats" to the edges. So depending on where one draws the sample from, that particular location will determine the levels of antimony. Further, copper tends to coagulate in clumps, so if you drew a sample from just one spot you might get a high concentration of copper. If you drew it from a few millimeters away, you could get a very low concentration. In fact, this is precisely what happened to Guinn. Which is why he tended to ignore his copper findings in favor of antimony and silver.
Randich and Grant also concluded that Guinn's sampling number for his conclusions was way too small to allow for the possibility of random matches. Randich said that the FBI held until recently that you could not get random matches with NAA analysis. It was later determined that they had never looked for any. After 2004, when Randich became a witness against them, they did look and they found one.
Perhaps the most arresting piece of evidence produced by the duo was a chart measuring the trace elements of the five pieces of evidence Guinn analyzed: the bullet left in the rifle at the so-called "sniper's nest", the bullet that Oswald allegedly fired at General Walker, and the three samples from the Kennedy assassination. The values varied widely, but especially for the "sniper's nest" bullet and the Walker bullet. Peter Dale Scott looked at this chart and surmised that it looked like those two bullets came from a different gun or type of ammunition. Right before this chart was placed on the overhead, Randich was talking about how trace metal values can vary widely in a particular run if one of the ingots used for the metals has been replaced on the production line. I asked him if he was then saying that theoretically all of those bullets, with their wide trace metal values, could come from one box. He replied that yes, they could. I then asked him, "Then what's the basis for this science?" He replied, "You're talking to the choir." Today, this new analysis is so convincing that the man that originally sponsored it and then advocated it to convict Oswald, Robert Blakey, has now joined the choir. Unfortunately, like many things in this case, the truth has emerged 28 years too late.
It is relevant to remind the reader of two other relatively recent discoveries in the bullet evidence field. The work of Josiah Thompson and Gary Aguilar has shown that CE 399 was not positively identified by the two witnesses who the Warren Commission relied upon for their alleged identification. So there is no positive proof that the "magic bullet" was the one found on the stretcher at Parkland Hospital. Secondly, the painstaking work of John Armstrong in his book Harvey And Lee raises the most serious questions about whether or not Oswald actually ordered the alleged murder weapon. That work, when combined with the writings of Randich and Grant, has undone the so-called "core physical evidence" against Oswald. Far from composing an "airtight" case against Oswald, the most serious questions need to be asked about the origins of this evidence, and where the linkage to Oswald can be found. Like the medical evidence in the nineties, these new developments in the bullet and rifle evidence have left both the HSCA and the Warren Commission foundering in a sea of questions and doubt.
(Stuart Wexler will be writing another article for this site on the dubious validity of Guinn's NAA study with new evidence and analysis questioning the statistical basis of his work.)