Why The Fragments Weren’t Planted

Kenneth A. Rahn
30 August 2001

      Since Dr. Vincent P. Guinn announced the results of his NAA studies on the bullet fragments some twenty years ago and showed that all the fragments recovered and tested came from two bullets fired from Lee Harvey Oswald’s rifle, critics have been trying to prove that the analyses were invalid because some or all of the fragments had been planted or tampered with. The provenance of the fragments is crucial to understanding the assassination because it starts a chain of reasoning that ends with a simple, direct nonconspiratorial explanation for all the physical evidence. If the fragments were not genuine, the NAA data are worthless, and the many other important conclusions that flow from it are nullified. If the fragments were genuine, the NAA data are legitimate, and those other conclusions can follow, perhaps the two most important being that only bullets from Oswald’s rifle hit the two men and that the rest of the physical evidence is genuine. The latter conclusion then opens the door to additional higher-level conclusions about the assassination. In other words, a huge amount is riding on the genuineness of the fragments that were analyzed by NAA.
The critics’ idea is that Oswald was innocent, or at least that he was part of a conspiracy with another shooter. The planted fragments were designed to shift the entire guilt to Oswald and hide the existence of conspirators firing from the front or the rear. The critics often cite CE 399, claiming that it had to have been planted because it was not damaged enough, that Ruby put it on the stretcher á la Oliver Stone’s JFK, or that O. P. Wright remembered that it was “pointy-headed” rather than round. Some critics also claim that the two large front-seat fragments were planted to make it look like the head bullet had come from the rear. Many of these same critics dispute the chains of custody of the fragments as well. The flames of these critics were fed by some unfortunate comments by Dr. Guinn to the effect that he didn’t test the same “specific little pieces” that the FBI had tested in its NAA 13 years earlier. Guinn knew what he meant, but everybody unfamiliar with NAA misinterpreted him in a conspiratorial way. The final criticism of the NAA results is that they just seemed too good to be true—something had to have been amok to produce them.
The specter of planting has hung over the NAA data for two decades. It is commonly used to attack the NAA without having to work hard. But this line of thought is very superficial, and does not survive scrutiny. It is surprisingly easy to debunk the debunkers. This document shows how to do that, by combining older evidence and the newest interpretation of the NAA data. Its logic has two main steps, first that the fragments could not have been planted, and second that even if for the sake of argument they are considered to be planted, they were not because conspirators could not have created or matched the observed groups by chance. The joint force of the two arguments eliminates any chance of planting. The first argument contains at least three subarguments—the fragments could not have been planted because there was no time, no knowledge of how to do it, and no opportunity. Of course, this kind of negative argument—that something could not have happened—is risky in principle, and should generally be avoided. But in this case its components are strong enough to justify it.
In short, every line of reasoning for planting runs into roadblocks immediately. None of the fragments were planted because it was impossible logically, practically, and mathematically.

The fragments could not have been planted
Too early and too little time. The first reason that the fragments could not have been planted is timing—false fragments would have to have been planted within hours of the assassination. The fragments were not altered after May 1964 because the two sets of NAA results (May 1964 and September 1977) agreed with one another. The fragments were not altered after the afternoon or evening of 22 November 1963 because the fragments had been officially placed into evidence by then. (See comments on chains of custody below.) Thus the only time that they could have been planted or altered was in the first few hours after the assassination, if that. (See illustrative details for CE 399 below.)
No knowledge of how and where to plant fragments. The second reason why the fragments could not have been planted is that the window of opportunity came well before anyone knew how what kind of fragments to plant and where. This impossible situation is almost always overlooked. Even if planted, the fragments could not have been planted intelligently because the special properties of WCC/MC bullets (that their composition differs from bullet to bullet, even in the same batch and the same box) were not known until Guinn discovered them years later, in the early 1970s. So any planting of MC bullets immediately after the assassination would have been “dumb,” i.e., without knowing that they had to match the Sb in groups of fragments. Thus the observed groups would have been created purely randomly.
Imagine the plight of the purported conspirators. They would have had only an hour or so to learn about the exact nature of the wounds to both men, determine what kinds of fragments to plant in order to make the shots appear to have come from Oswald’s rifle, and actually plant them while the evidence was still in Parkland Hospital. In that quest they were doomed, for the hour expired long before anyone understood the full pattern of the wounds in Kennedy and Connally. No one yet knew of the double-bullet hit (formerly known as the SBT, or single-bullet theory), and would not fully grasp what had happened for several months. Perhaps Connally’s doctors could have connected some sort of bullet and previous trajectory with his shallow thigh wound, but no one, absolutely no one, saw both the wound and the actual bullet while in Parkland. (The doctors saw the wound, and the Secret Service saw the bullet.) So any conspirators would have needed super-knowledge—that of the doctors plus the Secret Service—which was patently impossible. Furthermore, conspirators would have to have anticipated the full range of possible injuries to Kennedy and Connally so that they could fire a large-enough suite of test bullets into a medium like gelatin in order to re-create all possible types and degrees of damage, including the extremely rare case of a bullet passing first through a neck and then through a torso, sliding along a rib, fracturing a wrist, and winding up partly in a thigh. The conspirators would have to have come to the hospital loaded with a whole array of fragments, get in, find out exactly which ones to plant and where, plant them, and not be stopped during any of this by the tight security. “Let’s see—I think that fragment 15A from test 27 might work here. No, fragment 17G from test 31 seems better. Uh-oh, time is getting short. What do I do? Are those footsteps I hear coming down the hall? Does this white coat make me look enough like a doctor? How am I going to explain this case full of fragments?” The conspirators would previously have to have stolen Oswald’s rifle from the Paines’ house, where it was kept in an attached garage, then put it back without anybody noticing, all without the benefit of knowing which day Oswald would decide to go home to retrieve it. The probability of all these events happening is beyond infinitesimal. It hardly matters that superimposed on all this would be the negligible chance of accidentally matching the planted fragments chemically.
No opportunity to plant fragments. The third reason why the fragments could have not have been planted, alluded to above, is that Parkland Hospital was sealed off as soon as the motorcade arrived. During the hour or so that conspirators had available to comprehend the situation and to plant the appropriate fragments, they could not have entered the building or moved freely inside it.
Prediction that chains of custody are valid. Since the fragments could not have been planted, all their chains of custody must be valid. This is exactly what is seen from the testimony to the Warren Commission. The chains of custody are so tight that only one has been seriously challenged—for CE 399. But that challenge is easy to refute. CE 399 was found by Darrell Tomlinson within an hour of the time that Kennedy and Connally arrived at Parkland (as estimated by O. P. Wright). Tomlinson pointed it out to Wright, who picked it up from the stretcher, where it had rolled out from under something when Tomlinson bumped it against the wall while moving it out of the center of the hallway. Wright gave it to SS Agent Johnsen, who prepared a receipt for it when he passed it to his chief, James Rowley (also in Dallas). Rowley, Chief of the entire Secret Service, carried the bullet to the FBI laboratory in Washington, D.C., where he gave it that evening to FBI Agent Elmer Lee Todd, who initialed it. That same night of the 22nd, SA Todd gave the bullet to SA Robert A. Frazier, who also initialed it. All this was sworn to in Warren Commission testimony.
On 31 March 1964, SA Frazier identified CE 399 in WC testimony (3H428 ff.) by his initials plus those of SA Todd (before the FBI’s NAA). On 24 June 1964, SA Todd identified CE 399 by his markings on it (after the FBI’s NAA). This means that CE 399 had remained physically undisturbed from when it was initialed the night of the assassination through the two times 4–5 months later that it was identified from these initials, which bracketed the dates when the FBI ran its NAA on the fragments. Thus CE 399 was not disturbed when the FBI used it for NAA. Since the FBI’s results were later confirmed by Guinn, this means that CE 399 had remained undisturbed from 22 November 1963 through September 1977. Any tampering would have had to have been before the night of 22 November. How much before that night? Since CE 399 was given to the Chief of the Secret Service, James Rowley, while he was still in Dallas, and probably shortly after it was found at 1:30 or earlier, it would have to have been tampered with before about 1:30 p.m. or so. That reduces the window of opportunity to between 12:30 and 1:30 on the 22nd.

The fragments were not planted
The fragments were not planted because the probabilities of matching the composition do not allow it. Recall that the fragments were grouped into two tight, physically meaningful groups by the concentration of the element antimony (Sb) in the lead. The probability of a fragment’s having been planted is the chance that its Sb matched that of the other fragment(s) in that group purely randomly, i.e., the chance that a fragment chosen without knowing its Sb just happened to fall into the proper group. The simplest way to get a feeling for these probabilities is to imagine that the Sb in a set of bullets or fragments is evenly distributed over its entire range. We then compare the Sb in a fragment to the Sb in another fragment of the same group. The probability that the first fragment fell within that “distance” from the first fragment is the difference in Sb divided by the total range of Sb for that kind of ammunition, the full range being 1200 ppm for WCC/MC bullets and 44,000 for all kinds of bullets (according to Guinn’s tests). Say for example that the reference fragment had 600 ppm Sb and the questioned fragment had 630 ppm Sb. The probability that the questioned fragment had fallen by chance between 600 and 630 ppm is 30/1200, or 0.025. To be conservative, we allow the same range on the other side of the reference fragment, which corresponds to the probability that the questioned fragment had fallen within that distance either way. These conservative, bi-directional probabilities will be double the unidirectional probabilities. The conservative probability for our questioned fragment would then be (2 x 60)/1200, or 0.05.
Here are some examples of the resulting probabilities that fragments were planted with WCC/MC lead, meaning that the conspirators knew enough to plant the same kind of lead as would have been fired from Oswald’s rifle. (This assumption is probably wrong, however—conspirators could not have known about WCC/MC ammunition in 1964 because Sb in different kinds of ammunition was first measured several years later.) According to Guinn’s results, the best set available, the three tiny fragments differ in Sb from the “anchors” of their groups by the same amounts as in our example—an average of 30 ppm out of 1200 ppm, which when doubled makes their ranges 60 ppm and their probabilities 5%. We can then use this 60 ppm and 5% for estimating the probability of any fragment or group of fragments being planted. For example, the probability that the infamous CE 399 was planted would be about 5%, or 1 chance in 20. (The actual value is 6%.) This probability is right at the level of scientific decision-making; most scientists would conclude from this alone that CE 399 was not planted. The probability for the large tip found in the front seat would also be 5%. The probability that both large fragments were planted (on the theory that it makes no sense to plant CE 399 alone unless the head shot that produced the front-seat fragment really did come from the rear, which most critics dispute) is 5% x 5%, or 0.25%, which is one chance in 400. The probability that the three tiny fragments found in the rear carpet were planted will also be about 5%, and the probability that all three of these fragments (CE 399, tip from front seat, rear carpet) were planted will be 5% x 5% x 5%, or 1.25 x 10-4, which is about 0.01%, or more precisely one chance in 8000. In other words, we can state with extreme confidence that the full set of three fragments was not planted.
The probabilities of 0.01% to 5% for planting with WCC/MC lead are the most optimistic possible. The values become far smaller if one allows for fragments to have been planted from the full range of bullet lead, which scenario is entirely possible because the special properties of WCC/MC lead were not understood in 1963. (Conspirators could not have known that they had to use WCC/MC lead for planting the tiny fragments in order to match the big fragments chemically.) If the purpose of planting fragments was to falsely link Oswald to the shooting, that would mean that somebody else was shooting from behind. If Oswald was up there shooting, too, there would have been two shooters in the Depository, and Oswald would have known about the other one. If there was only one shooter, then the two large front-seat fragments from the head shot, one of which had a lead core, would have also come from a different rifle, and they would have to be fixed. But if the conspirator shooting from the rear was not using a Mannlicher-Carcano rifle, which was the probable scenario because there would have been little reason for a professional assassin to use such a cheap weapon, the range of Sb in his bullets expands from the 0–1200 ppm of WCC/MC ammunition to the 0–44,000 ppm of other ammunition tested by Guinn (which may still underestimate the true range because Guinn didn’t test every kind of ammunition). That weapon would leave behind tiny fragments in Kennedy’s head and Connally’s wrist that could not have been tampered with. Their starting range of Sb would be 0–44,000 ppm. The chance of Sb in those tiny fragments falling into the much lower WCC/MC range and matching the planted fragments then becomes the product of the chance of falling into the 0–1200 range and the chance of overlapping a particular fragment within that range. The first of these terms can be approximated by the fraction of non-MC bullets whose antimony is low enough to fall into the WCC/MC range (7/36 according to Guinn’s measurements). The second term is the usual 60/1200. The product of these two terms, the chance that a non-MC fragment matched an MC fragment, is then (7/36) x (60/1200), or 0.01. (This figure is five times lower than the chance of two random WCC/MC fragments matching each other.) The value of 0.01 means that if Kennedy and Connally were hit by non-WCC/MC bullets and the larger fragments had been replaced by WCC/MC plants, the chance of single tiny non-WCC/MC fragments matching one of the big ones (such as in the group composed of the wrist fragment and CE 399) would be 1%, whereas the chance of two tiny non-WCC/MC fragments matching  a big one (as in the group with the front-seat fragment, the brain fragments, and the rear-carpet fragments) would be a minuscule 0.01%. If all three tiny fragments were planted non-WCC/MC, the probabilities decline further to 0.0001%, or one part in a million. Thus we can again state with great confidence that none of these three scenarios happened.
In summary, none of the probabilities of planting false fragments are large enough to make them reasonable. The absolute best scenario calculated in this simplified manner, for the case of a single WCC/MC fragment accidentally matching another WCC/MC fragment, is 5%. Probabilities for scenarios with two or three fragments being planted decline to 0.25% and 0.0125%, respectively. For non-WCC/MC fragments matching WCC/MC fragments, these three probabilities virtually disappear, sinking to 1%, 0.01%, and 0.0001%. So these scenarios didn’t happen, either.
A more rigorous way to calculate the probabilities has been developed by Larry Sturdivan. It uses the log-normal distribution of Sb in Guinn’s 14 test bullets to derive a geometric mean and the Anderson-Darling P-statistic. The results show that the chance of two random WCC/MC fragments matching each other are 2%–3%, which is only half the value of the simplified probabilities calculated above. They also suggest that the chances of non-WCC/MC fragments matching WCC/MC fragments are one to two orders of magnitude below that, which are also lower than the simplified values for the same scenario. In other words, Sturdivan’s more-rigorous analysis takes a situation that is already highly improbable and makes it even more improbable. If the fragments weren’t planted the first way, they surely weren’t planted the second way.

In spite of the fact that critics have cried “plant” for the bullet fragments over the last twenty years, simple analysis reveals impenetrable roadblocks for this idea at every turn. Conspirators would have had available only an hour or so after the assassination to do the job, because the fragments recovered at Parkland Hospital were in custody by then. That amount of time was grossly insufficient because the basic nature of the wounds was not established until that night, and the full wounding scenario until months later. The conspirators would have to have learned more in that hour than the full team of doctors and investigators was able to piece together in months. They would have to have anticipated all possible scenarios and arrived at the hospital with a satchel full of fragments for all of them, including the highly improbable DBH. Further, they would have to have done all this without being detected by the intense security, and in a way that later fooled all the medical personnel and the government scientists brought in by the Warren Commission. They would also have to have stolen and replaced Oswald’s rifle from virtually inside the Paines’ house without anyone noticing. And because they were doing this in the early 1960s, before Sb was known to be the best tracer element for bullets and before the unusual properties of WCC/MC bullet lead were known, the conspirators would not have known anything about the need to match types of lead in general and WCC/MC fragments in particular. All of this amounts to the kiss of death for planting. It is time to abandon that failed notion, and recognize that planting is only for people who haven’t thought things through.

      The writer gratefully acknowledges the jump start provided by Joel Grant’s 1996 manuscript “The Neutron Activation Analysis.”

Back to NAA