Actions
  Alerts
  Letters

About CTKA
  What CTKA Is

Resources
  National Archives
  Links

CTKAformerly published Probe Magazine. Most of the articles on this site first appeared in Probe.

If you would like to submit an article to be considered for publication on this site, please send mail to us at here.

Jim DiEugenio's Upcoming appearances and radio Interviews:

April 13th, Barnes and Noble, Metro Pointe, 901 B South Coast Drive Ste 150, Costa Mesa, CA
714-444-0226, 12-3PM

May 4th, Barnes and Noble, Orange Town & Country 791 South Main Street Suite 100, Orange, CA
714-558-0028, 12-2PM

NEW DATE! May 18th, Barnes and Noble Bookstore in Manhattan Gateway Shopping Center 1800 Rosecrans Avenue Building B, Manhattan Beach, CA 90266
310-725-7025, 12-4 PM

October 16-19th Passing the Torch Conference, at Duquesne University in Pittsburgh

November 21-24, November in Dallas, at the Adolphus Hotel in Dallas

Hear No Evil: Social Constructivism and the Forensic Evidence in the Kennedy Assassination
by Donald Byron Thomas

A Comprehensive Review: Part Two
by David Mantik

Chapter 9: Terminal Ballistics (p. 297)

Thomas claims that JFK would not have been able to raise his arms after his spinal cord was damaged. He believes this damage was caused by the bullet through the back, i.e., the SBT. What he seems to forget, though, is that only those nerves inferior to such a level are knocked out, whereas those superior to it are not affected. There are actually two issues here: (1) Was the spinal cord actually damaged? (2) If so, what was the level? For the first question, there is no answer: the pathologists simply did not examine the cord. For the second, the level is in serious doubt, as the discussion above about the SBT has already suggested—it might have been as low as T4. Thomas wants to place the cord injury at about C7, which would impact the central brachial plexus (the plexus encompasses C5 to T1). However, the deltoid muscle (see my Figure 10) is the principal abductor (that means movement away from the body) of the upper arm and it is supplied by nerves from C5 and C6, which clearly lie superior to C7. Therefore, they would not have been damaged by trauma at C7. Furthermore, initiation of abduction is achieved by the supraspinatus muscle, which is also supplied by C5 and C6. However, without help from the supraspinatus, the deltoid cannot abduct by itself, unless the shoulder is first dipped a good deal. Based on these considerations, therefore, Thomas is wrong to conclude that JFK could not have elevated his arms after this gunshot.

Figure 10. Posterior view of neck and shoulder. Ulnar and radial should be reversed. The horizontal scapular spine can be faintly seen (red arrow), inferior to the level of T1.

Thomas waffles about the exact site of the minimal vertebral trauma (C7 vs. T1), although both are cited on p. 237. (As an excellent reference for this question and for many others in this review, Google “How Five Investigations into JFK’s Medical Autopsy Got It Wrong” (May 2003), by Gary L. Aguilar, MD, and Kathy Cunningham.) See Figure 7.5 by Thomas for the chest X-ray, where he incriminates C7. On two pages (pp. 317, 713) Thomas states that the vertebral process was knocked out of place, while at another (p. 242) he implies the opposite. On yet another page, he states that the bullet was merely deflected. In any case (in my opinion), this trauma was quite trivial. It is not even certain that it occurred during the assassination, so no serious time should be wasted on this issue.

There is a compelling alternative scenario to the SBT. If a frontal projectile caused the throat wound, and especially if this wound lay above the necktie and collar (as the eyewitnesses stated), then the level would have been close to C7, consistent with tracheal injury between the 2^nd and 4^th rings. If this wound had been caused by a shard of glass from the windshield, then it might well have gone undetected at the autopsy (and on the X-rays). It could even have caused the bruising seen at the top of the right lung. Furthermore, the shard would likely not have exited through the back, but rather would have stopped within the body. The shock wave from this shard might even have triggered the reaction of the deltoid and supraspinatus muscles (via C5 and C6), i.e., JFK’s splayed arms as seen in Z-224. That both arms are elevated is perhaps a bit unexpected. The shock wave, though, might have reached (or nearly reached—which is good enough) both left and right spinal foramina (the openings in the vertebrae, where the spinal nerves emerge from the spinal cord) for C5 and C6, so that both arms were affected. There is, of course, a corollary to this: the back wound would then be viewed merely as a superficial shrapnel wound (totally independent of the throat wound). I have already noted the generous evidence on the skull X-rays for shrapnel. In addition, multiple witnesses recalled events that likely produced such shrapnel. Thomas does not address this scenario.

Thomas seems to believe that the deltoid and supraspinatus muscles did not cause JFK’s arm reactions. Instead, he proposes this: “...exaggerated splaying of the arms and the elbow flap can be interpreted as a myotactic reflex; a reaction induced by the pluck of the tendons of the trapezius muscles” (p. 320). But then he sows confusion:

A myotactic contraction of the trapezius..., after [emphasis added] his arms were raised, would pull on the shoulder (deltoid) muscle, causing the upper arms to ‘flap’ as seen in the Zapruder film.

The first quote seems to explain why the arms became elevated, but the second quote implies that they were already elevated (from what, he does not say). So, I emerged quite confused from all of this. Thomas does make one item clear though: JFK’s arms must have been elevated at Z-224, before he was shot through the “base of the neck.” (This level of trauma, of course, would please the budding pathologist Gerald Ford, who elevated the wound to this level.) Now here follows yet another puzzle created by Thomas. The trapezius muscle is very long—it attaches to all twelve of the thoracic vertebrae. For this large muscle to respond over much of its length, merely from localized pressure on a tendon in the lower neck (i.e., as in the SBT), seems far fetched. Furthermore, the action of the upper trapezius is to draw the shoulders strongly backward and also to extend the arm, neither of which is seen in Z-224. Lastly, unless the single bullet struck both the left and right trapezius muscles, Thomas’s proposal (of a tendon “pluck”) cannot explain why both arms were elevated. This is serious trouble indeed. But there is even more: the tendons are in the wrong location for Thomas’s effect—they are located too near the midline, much too far from the back wound.

But Thomas does not stop there—he suggests that the trapezius muscle actually triggered the deltoid muscle (p. 320). This is most peculiar. I have never encountered anything like this before in anatomy or physiology: muscles are not triggered by other muscles. On the contrary, they are triggered by specific nerves.

My Summary. JFK’s splayed arms probably resulted from a frontal projectile. A shard of glass (from the windshield) is one possibility.

Chapter 10: The Rearward Head Snap (p. 333)

Thomas claims that the report of the forward head movement originated with CBS news, presumably meaning Dan Rather (p. 371, footnote 15). However, this overlooks the fact that Hoover’s top aide, Deke DeLoach, also reported, in his autobiography (Hoover’s FBI: The Inside Story of Hoover’s Trusted Lieutenant 1995, Cartha DeLoach, p. 139), that he saw a forward head snap in the film that same weekend. Like Rather, he says nothing about a backward snap. To further contradict the extant Z-film, no one who saw the original film that weekend reported the backward head snap. And, finally, Dealey Plaza witnesses did not report it either. In the face of all of this evidence I am surprised that Thomas is certain that federal authorities knew about the (backward) head snap that weekend. No one else seemed to know. Furthermore, I have never met anyone, who, after viewing the extant film, said that the head went forward—but at least two presumably sane individuals (DeLoach and Rather) described precisely that. Thomas states that Jackie could not have pushed JFK’s head backwards that quickly. If the film is authentic, that would be true, but if frames have been excised (and the film reconstructed) then it’s anyone’s guess. It is likely that Jackie did indeed push JFK’s head backward, though not at the speed seen in the extant film.

Both Larry Sturdivan and ITEK concluded that a frontal shot from an M-C could not account for the backward head snap. Thomas (apparently approvingly) cites controlled experiments by Olivier and Sturdivan (p. 363) with human skulls that make this point:

a Mannlicher-Carcano will not impart kinetic energy to the brain case in a manner and quantity sufficient to propel the head in the direction and velocity seen in the Zapruder film (p. 370).

He does, however, argue strongly that a .30-06 bullet, fired from the GK, would yield a backward head speed of 2.6 ft/sec, well above the observed value seen in the Z-film of 1.6 ft/sec (see Josiah Thompson’s analysis for the latter figure). His calculations, however, include important assumptions: (1) he assumes that 80% of the momentum is deposited, and (2) he ignores JFK’s torso, which also moves, but the required energy (and especially the mass) is not even estimated. Thomas does quote Sellier (Thomas, p. 356; also see my Appendix 5 here), who claimed that 50% of a bullet’s energy is thermalized, which leaves only half for the final kinetic energy. I suspect that Thomas may be correct in both of his conclusions: (1) a M-C bullet won’t work and (2) a different bullet just might do the trick. However, more experiments would be required to verify this. G. Paul Chambers (Head Shot 2010, pp. 210-212) also agrees that most of the bullet’s energy would be lost to heat. More to the point, he essentially agrees with Thomas that a frontal bullet (e.g., a Winchester .220 in his case) could do the trick. Incidentally, he also notes that the energy required to lift JFK’s body against gravity would require a negligible amount of the bullet’s energy. After repeating his calculations, I agree with him about this (which changes my prior opinion on the importance of gravity). If the M-C bullet cannot produce the visible head snap in the Z-film, then film alteration would be left as the best explanation for anyone who still believes in the M-C bullet. However, a different bullet from the front, with more momentum (and especially a higher speed), might well work, but that conclusion would, of course, immediately destroy the WC’s conclusion (they only permitted a M-C bullet).

Thomas briefly revisits the issue of multiple head shots, but quickly dispenses with this option, saying that evidence in recent years has convincingly failed to show multiple shots and that, furthermore, the brain only shows one track. I shall not browbeat this issue much; however, I have already cited the very recent work of Horne (and my own work, which is not yet too old), neither of which Thomas cites. Likewise, the brain photo is very likely not authentic, although Thomas avoids this serious concern.

The “Kronlein Schuss” effect is Thomas’s primary explanation for the marked destruction of JFK’s skull. Though he may have Lattimer’s experiments in mind in this specific discussion, curiously, he does not cite them until later in this chapter. Instead, he cites a textbook, Wound Ballistics (Beyer 1962). I agree that Lattimer’s skulls did show a great deal of skull trauma (see Figure 10.3 by Thomas). It should be emphasized though that they did not, even remotely, duplicate JFK’s skull wounds. Gary Aguilar has repeatedly made this case.

After some detailed calculations Thomas concludes that, contrary to the work of Alvarez, there was no jet effect. (I agree.) In addition, though, he accuses Alvarez of tinkering with his experiment in order to get the results he wanted. (He did.) He also states that Alvarez should have used a M-C and a coconut. (I agree.) See Thomas’s footnotes for more finely honed comments about Alvarez. (I liked these!)

In the section on Lattimer’s stepladder, he notes that the work of Alfred Olivier and Larry Sturdivan, using ten skulls shot with LHO’s carbine, did not support a jet effect either—in fact, all ten skulls moved away from the shooter. He finishes this chapter with a severe critique of Lattimer’s experiment and a similarly harsh one of the Discovery Channel’s program, Inside the Target Car (2008). I agree with both and have, in fact, issued my own critique of the latter.

Thomas has a daunting challenge to explain the right occipital skull defect as deriving from a GK shot. I have already cited Cyril Wecht on this question—he regarded such a large deflection as most unlikely. Paul Chambers, too, opposes Thomas on this point (Chambers 2010, p. 160): “Unless Kennedy had a metal plate in his head, the bullet could not have been deflected sideways and would have followed a straight path, one way or the other.” Thomas bases his case solely on an article by Karl Sellier, “Bullet Deflection and Bullet Deformation.” Although Thomas summarizes this article (written in technical German) with generic comments, nothing specific is offered to support such a large deflection, especially not for a bullet striking a skull. See my Appendix 5 for a summary of pertinent points from this article.

Chapter 11: The Magic Bullet (p. 375)

Even though the HSCA made the magic bullet (CE-399) the key to its case, Thomas does not believe in it. He does, however, believe in the SBT (single bullet theory), which is defined as the magic bullet theory stripped of CE-399. In other words, some other single bullet (not in evidence) did pass through both JFK and JBC. Thomas believes this occurred at Z-224. He bases his conclusion on JBC’s lapel flap, JFK’s arm elevation, and the subsequent motion of his wrist and hat. (Thomas says JFK’s arms did not elevate due to a shot through the neck, although he does not state what the true cause was.)

Thomas then recounts the WC’s treatment of James Tague and the magic bullet theory, which even Humes had labeled “extremely unlikely.” Of course, the SBT also directly contradicted the FBI report. Richard Russell even reportedly said that he would not sign onto the SBT (see p. 417, footnote 24). The curbstone (stored at NARA), which Tague’s projectile had apparently struck, was discovered by Harold Weisberg to have been repaired (!), thus obscuring the original evidence. Thomas concludes that, within the framework of just three shots, neither the FBI nor the WC was able to offer a coherent explanation for Tague’s wounds—or for the curbstone mark. On the contrary, they offered only bewilderment. I agree that Thomas is right on target here.

Regarding the magic bullet (CE-399), supposedly found on a Parkland stretcher, Thomas states that there is no evidence that it had ever harmed anyone (except, of course, the WC). Darrell Tomlinson, who found the bullet, adamantly refused to be coerced by Specter into supporting the SBT. Thomas emphasizes that the bullet was found on a stretcher that also held a stethoscope, but JBC’s stretcher had no stethoscope. Given Specter’s passion for truth, though, he did not want to know about the stethoscope.

Thomas then tackles the receipt from the autopsy for a “missile.” After reviewing the Dealey Plaza evidence (including an interview with “Jean Lois Hill”—curiously, not called Jean Lollis Hill) he concludes that a bullet had been plucked from the turf there and carried to Parkland, where it was eventually dropped into JFK’s coffin (perhaps by Richard Johnsen), only to reappear during the autopsy. Captain David Osborne, Chief of Surgery at the Bethesda Naval Medical Center, recalled that an intact bullet had fallen out of JFK’s clothing and that it had been collected by the SS. The FBI, however, maintained that the “missile” receipt was for two tiny slivers of metal taken from JFK’s brain. Dennis David (whom I have met and interviewed) also later recalled that he had typed a receipt (requested by the SS) for four metal fragments. Even Commander Stover, whose name was on the receipt, admitted to Lifton that the bullet had indeed been present at the autopsy. Finally, a memorandum written by SS Chief James Rowley stated that Elmer Todd (FBI)

...was given the bullet [emphasis added] enclosed in an envelope.

Thomas concludes that this bullet most likely was CE-399. It is, after all, the only intact bullet in the JFK case. Furthermore, Thomas suggests that the appearance of this bullet led Humes to postulate that this bullet had fallen out of JFK’s back. It should also be noted that Special Agent Richard Johnsen [DM: apparently spelled as “Johnson” on p. 407; Richard Johnsen, but not Richard Johnson, is in the index] said he had not personally carried any bullet from Parkland to Bethesda. If true, that would make the bullet’s journey via the coffin more likely. Thomas tells a fascinating story, but it has been mortally wounded by the independent analyses of Martin Hay and John Hunt (discussed below).

Regarding the skull fragments, Humes said that only one piece of bone was brought into the autopsy room. On the other hand, James Jenkins recalled several bone fragments lying in the coffin. Vincent Drain (FBI) told author Larry Sneed (No More Silence 1998) that a security officer at Parkland probably gave him a skull fragment; presumably this was then placed into the coffin. Thomas suggests that the donor of the fragment was Doyle Williams (FBI), who got into a scuffle at Parkland with Andrew Barger (SS). Thomas wonders if Doyle had actually placed the bullet into the coffin. I found this proposed reconstruction by Thomas fascinating. If true, who was directing Doyle Williams?

My Summary. Thomas does not believe that CE-399 (the magic bullet) caused any injuries. Instead, he proposes that some other single bullet did the same job. He suggests that CE-399 arrived at Bethesda via the coffin, but Hay and Hunt strongly disagree. (So do I.)

Chapter 12: Bullet Trajectory (p. 421)

This chapter is an effective annihilation of Thomas Canning’s trajectory work for the HSCA. Thomas goes into exquisite, and welcome, mathematical detail to expose Canning’s biases (social constructivism at work again, though Canning probably never heard that phrase). Thomas concludes that Canning’s analysis was thoroughly unscientific, even though in real life Canning was a rocket scientist. Canning’s analysis assumes, of course, that the magic bullet followed a straight line. However, the Forensic Pathology Panel had specifically advised the HSCA to avoid an analysis like Canning’s—because the data was too uncertain. Canning then proceeded to ignore this very advice. [DM: I have criticized Dale Myers for his computer reconstruction on these same grounds, i.e., uncertain data.] Thomas explicitly, and appropriately, chastises Canning for his bizarre concept of “the minimum reasonable margin of error,” a concept that does not exist in standard practice. Canning’s chief concession though was this: the circle of potential error included parts of another building behind JFK! Of course, this statement was omitted from the final report. That would, after all, have opened the door to another gunman.

One of Canning’s chief uncertainties for the head shot was the degree of deflection. Thomas asks whether 5º was reasonable. If so, why not 10º or even 15º? (Of course, this logic applies also to Thomas’s own proposal of a GK shot exiting JFK’s right occiput—more later about that.) He returns to the paper cited above by Karl Sellier (1971). But now Thomas is a bit more specific: deflection increases “...with the hardness of the bullet, the hardness of the target, and the angle of strike.” Deflections as great as 65º were reported by Sellier when bullets struck metal plates. Since Thomas has just cited the “angle of approach” to JFK’s head (from the GK) as 50-60º, he seems to suggest that this 65º deflection (with metal plates) is both noteworthy and pertinent to the JFK case. Thomas then introduces the notion of yaw, which is the angle between the bullet’s axis of spin and its momentum vector (direction of travel), with the implication that a significant yaw may have contributed to the large deflection of a GK shot. I found this entire argument less than convincing, especially in view of Cyril Wecht’s verdict, already cited above.

Thomas next tackles, in commendable detail, the trajectory of the neck wound. He concludes that the forward tilt of JFK, cited by Canning in order to facilitate the SBT, would require a man with the stature of Shaquille O’Neal (spelled “O’Neil”). His final conclusion though is that this type of trajectory analysis can neither prove nor disprove the SBT: “No precise alignment can be proven” because the data are simply too imprecise and the bullet track through tissue cannot now be known. I liked this conclusion very much—that’s just the way it is, but someone should tell Dale Myers.

My summary. The magic bullet is a fiction. So is any other single bullet. The throat wound and the back wound were caused by separate events.

Chapter 13: Comparative Bullet Lead Analysis (CBLA) (p. 451)

Thomas is at his best here. He totally destroys the use of NAA (neutron activation analysis), as it has been historically employed. It had been Vincent Bugliosi’s matchless bad luck to applaud this antiquated approach, only to see several peer-reviewed papers immediately appear that decimated his entire discussion. Of course, the ultimate irony is that Blakey had once called NAA the “linchpin” of the assassination. If that had ever been true, then that linchpin is now permanently cracked—it should be exhibited at the Sixth Floor Museum. And someone really should ask Blakey about this. I shall be brief here: it is now simply a waste of time to discuss NAA in the context of the JFK assassination. Thomas makes one final point though (which I have also emphasized): the number and size of the NARA bullet lead specimens (which I have seen at NARA) do not match the specimens removed from JFK in 1963. TO say the least, this is very queer.

Chapter 14: The Tippit Case (p. 491)

I am not an expert on the Tippit homicide, so although I found Thomas’s discussion informative, original, and interesting, I cannot pass verdict on it. He concludes that LHO shot Tippit [DM: Maybe he did]. But the circumstantial evidence also strongly implicates the Dallas police. According to Thomas, the physical evidence “fails to pass muster.” The two Remington-Peters casings do not show the marks supposedly placed on them by officers Joseph Poe and Pete Barnes, who swore that they had indeed marked the bullets while at the scene. The two Winchester-Western casings turned up only after the scene had been searched by crime scene detective Barnes—many hours after LHO and his pistol had been in custody! Thomas therefore suggests that these items had first been removed from LHO’s pistol and then later planted at the scene. Officer M. N. McDonald (a relative of a former patient of mine), swore that six live cartridges had been taken from LHO’s revolver, including one with a dented primer, and that he had indeed marked all six. However, none of the six in evidence has a dented primer, or nor do any of them carry McDonald’s mark.

Three different expert examinations of the firearms evidence led to three different, but contradictory, conclusions. None of these experts even considered that the evidence might be unreliable. [DM: This is precisely the problem that has bedeviled the medical evidence.] A billfold was found at the scene, but it was not dropped by LHO.

The radio log shows that Tippit had been dispatched to LHO’s area and that he was not on routine patrol (which is contrary to the official version). Furthermore, the dispatchers had linked suspects in a station wagon to both the JFK shooting and to the Tippit shooting. The FBI learned that this wagon was owned by Carl Mather, a close friend of Tippit and that the suspects had been lurking nearby at a Mexican restaurant. Moreover, that same restaurant was identified as LHO’s destination. Thomas concludes: “The evidence clearly implicates Tippit and his colleagues as accessories after the fact.”

Chapter 15: Lie Detection (p. 537)

This is another delightful summary by Thomas. He notes that the polygraph had well known flaws, while the problems of the PSE (Psychological Stress Evaluator), employed by George O’Toole (The Assassination Tapes 1975) were intractable. Did Jack Ruby know LHO? Ruby said that he did not, and the WR said that he (Ruby) had passed the polygraph. But the WC actually discounted these results; after all, in 1964, Ruby carried a diagnosis of psychotic depressive. The HSCA reviewed this lie detector work and cited “gross abuses of basic polygraph principles” and “Total disregard of basic polygraph principles.” Thomas then details exactly how all the basic rules had been flaunted. He concludes that Bell Herndon, the examiner, had purposely sabotaged the exam and that he (Herndon) then gave false testimony to the WC. The HSCA concluded that the Ruby exam had been invalid and unreliable. Thomas says that evidence points toward collusion between the FBI and the staff of the WC—in order to manipulate the exam results. As summarized here, several footnotes have some tantalizing tidbits:

Footnote 16. Sociopaths are not susceptible to the polygraph. (One can only wonder how LBJ would have done.)

Footnote 19. According to a Mailer novel, John Paisley (CIA) may have been killed because he had gotten too close to the infamous Soviet defector, Yuri Nosenko.

Footnote 90. Ruby’s friend, Al Maddox, produced a note, given to him by Ruby, which said that he (Ruby) had shot LHO in order to silence him, i.e., nothing about sparing Jackie inconvenience here.

Chapter 16: Ballistic Acoustics (p. 559)

Now, at last, we come to the heart of Thomas’s expertise—the acoustics case for conspiracy. Here his relentless pursuit of the truth and his dogged shaking of every bone of contention show him at his best. For clarity, I first summarize the basic facts. [Those who want to skip ahead may preview our final destination in Appendix 8.]

1. Three audio recordings are germane to the case.
2. The first was made by radio station KBOX, but analysis to date has been inconclusive. It is not discussed further in this review.
3. The HSCA obtained a set of Dictabelts (Ch-1), and a magnetic tape copy, from former Captain Paul McCaghren, who had taken them home from Chief Curry’s office after Curry retired. The Dictabelt was in such poor condition that BBN (Bolt, Baranek & Newman) concentrated their work on the magnetic tape copy. But, for comparison and authentication, they also made their own magnetic tape copy from the Dictabelt. McCaghren also submitted the Audograph discs (Ch-2) to the HSCA; he had also taken these home for safekeeping.
4. Channel one (Ch-1), the primary DPD frequency, was used for normal activity. Recordings were made at headquarters on a sound-actuated Dictabelt (shown in Figure 16.1 by Thomas). This is the channel that contains the suspect gunshots.
5. Channel two (Ch-2) was devoted to talk within the motorcade that day. Recordings were made at police headquarters on a sound-actuated Gray Audograph disc. This channel does not contain suspect gunshots.
6. Silence (after about four seconds) in either channel would stop the recording.
(See Appendix 7 for possible sources of uncertainty in establishing a correct time line.)
7. Time announcements were made only to the nearest minute—there was no second hand, so exact concordance between the two channels cannot be expected.
8. At 12:34 PM the Ch-2 dispatcher said:

   Unknown motorcycle; up on Stemmons with his microphone stuck open on channel one. Could you send someone up there to tell him to shut it off?

9. On Ch-1, for 5 1/2 minutes (12:28 PM – 12:34 PM), a motorcycle dominated the sounds.
10. After about two minutes, the motor noise dropped dramatically, suggesting a deceleration. After idling for about 34 seconds, the motor noise increased again.
11. Towards the end of this recording (between 260 and 300 seconds) sirens from two other vehicles are heard. A Doppler effect suggests passing vehicles.
12. On these channels, like a party line, more than one person could speak at the same time. So even on the stuck Ch-1, another officer could have been recorded.
13. It was known that motorcycle microphone buttons could get stuck.
14. The first hint of shooting was on Ch-2, when Chief Curry said, “Go to the Hospital!” Six seconds earlier, Curry had announced his location near the triple underpass; JFK should then have been in mid-Elm St. (p. 607, footnote 20).
15. According to the HSCA, no discernible gunshots can be heard by the human ear (i.e., they may be audible, but they cannot certainly be recognized as gunshots). What Gary Mack heard as gunshots were not gunshots. Sound technicians Mark Weiss and Arnold Aschkenasy reported that the human ear only senses static at the critical moments of the suspect gunshots. (To listen to these Dictabelt recordings, click here.)
16. AGC (automatic gain control) decreases the amplitude of the original sound, but it may increase the sound of the echoes. In short, it tends to equalize amplitudes.
17. Five major buildings around Dealey Plaza typically lead to a total of ten echoes for each sound.
18. The initial sound plus echoes lasts less than 0.5 seconds.
19. BBN isolated suspect sounds on Ch-1, at 135-148 seconds, just seconds after the motorcycle noise decreased. Five sounds made their final cut, as possible authentic gunshots; ultimately only four were considered. Thomas discusses six, and explains why the sixth one was discarded.
20. After 12:30 PM for Ch-2, a graph (Figure 16.3) of time announcements vs. playback time (the latter is measured via a stopwatch while listening to the recording, i.e., no corrections are made) yields a slope of about one, which suggests that moments of tape silence were minimal. (Linsker, et al., do not agree with this, as I discuss later; this is a major issue.)
21. A similar graph for Ch-1 shows excellent correlation between the time announcements and playback time—that’s because the motorcycle engine kept the tape from stopping.
22. A study of the 60 Hz background hum proved that the police tape playback speed was 5% too fast. Therefore the measured time on Ch-1 needs to be increased by 5% to obtain the actual elapsed time.
23. After this was corrected, the difference in times on the two channels was still 35 seconds. However, because of the inexactness of dispatcher announcements of time, Thomas does not consider this to be a problem.
24. The above analysis suggests to Thomas that the cluster of suspect shots had been recorded on the Dictabelt (Ch-1) within a minute of the actual shooting, i.e., likely consistent with conspiracy, especially considering the uncertainty in time as announced by the dispatchers.
25. As a control, the entire tape (Ch-1) was searched for similar sounds, but no other similar sequence was identified. Actually, this should be qualified: an impulse train of four seconds duration was found (thirty seconds after a bell sound), but it failed to meet the screening criteria, so it was discarded. However, as Thomas notes, this anomaly proves that something besides gunshots could produce impulse trains on the Dictabelt.

1. Representative Samuel Devine of Ohio, a self-proclaimed expert on gunshot sounds, stated that the sounds on the tape could not be gunshots, but he had just listened to the test tape, not the Dictabelt! On the other hand, Blakey and Bowles said that the test shots were easy to hear on the test tape.
2. Test shots were fired from only two sites: the TSBD and the GK. No test shots were fired from the storm drain, on the overpass—or anywhere else.
3. Only four targets were used for these test shots: the limousine position at Z-160, at Z-200, and at Z-313, as well as Tague’s position (where he was struck by a projectile while near the base of the triple underpass). The selection of targets is the least sensitive of the choices made in this reproduction (p. 577).
4. To mimic possible sites for the stuck microphone, microphones (36 total) were placed at intervals of 18 feet along Houston and Elm streets. However, each array had only twelve microphones. Therefore, in order to cover all 36 microphone sites, three different arrays were required. This entire sequence had to be performed separately for the two GK shots (a M-C and a .38 pistol) and for the test shots from the TSBD shot (a M-C, in all cases).
5. A M-C was fired from the TSBD at the four targets. From the GK, the M-C and a .38 pistol were each fired at just three targets (Z-160 was omitted). No other shooting sites were tried. (See test patterns in Thomas’s Figure 17.1).
6. This entire sequence led to 432 test patterns. Of these, 144 were produced by firing the M-C at the four targets from the TSBD, i.e., 36 microphones x 4 targets = 144 test patterns. Then another 144 patterns were obtained with the M-C withdrawn inside the window. At the GK site, for the M-C only three targets were employed (Z-160 was omitted). This yielded another 108 patterns (36 microphones x 3 targets = 108 patterns). According to the HSCA, for the GK the .38 pistol was fired only at the target for Z-313. (But I could not find where Thomas had stated this clearly.) This yielded just 36 patterns for the .38 pistol. Altogether then the total number of test patterns was 144 + 144 + 108 + 36 = 432.

1. The task now was to find matches, if any, between the 432 test shot patterns and the six evidence patterns. Such matches would presumably determine both the shooter locations and the target sites. For this exercise, the reader must imagine a very large matrix, consisting of 432 entries vertically and six entries horizontally. For each element of this matrix there is an evidence pattern and a test pattern, which are to be compared to one another. So a total of 432 x 6 = 2592 comparisons must be made.
2. Matches for a specific shot were decided based solely on the time between spikes, i.e., amplitude was ignored (except, of course, for the already completed, initial selection of suspect gunshots).
3. A deviation of eight milliseconds (msec) was permitted, since the microphones might not precisely match the motorcycle position. Even air movement might change the matches.
4. The statistical formula for detecting a match was this:

   Binary Correlation Coefficient = r =   i  ,
             √ N x n
   
   where i = number of coincident events
             N = number of spikes in the evidence pattern and
             n = number of spikes in the test pattern.

   For a perfect match, r = 1, while r = 0 means no match. But, partly because of so much noise, a perfect match could not be expected. Results of interest were for r > 0.6; however, it should be emphasized that this is an arbitrary value. Some other value could have been chosen, with a likely different final outcome, possibly even wildly different.
5. Values for r < 0.5 were ignored; that left only 15 possible matches (see Table 13 by Thomas). These 15 had the generic pattern of gunshot echoes in Dealey Plaza. The reader must understand that this does not mean 15 shots! After all, duplicate test shots had been fired from the TSBD (inside and outside the window). Furthermore, matches sometimes occurred at adjacent microphones—from the same shot—as might well be expected if the motorcycle had been between two adjacent microphones. Only four actual shots were proposed.
6. Table 13 lists these four shots, in chronological order: (a) TSBD, (b) TSBD, (c) GK, and (d) TSBD. (The shots at 136.20 and at 140.32 seconds were discarded by the HSCA.) The first two shots (from the TSBD) were recorded only in microphone array #2, while the last two shots were recorded only in microphone array #3. (Positions of each microphone on the street are shown in parentheses, after the number that identifies the microphone array.)
7. The first two suspect shots were separated by only 1.6 seconds (and both were from the TSBD), which is an absurdly short interval for a M-C carbine. Because of this extremely short interval, Thomas believes that Blakey should not have permitted these two suspect shots to be accepted as official shots. Nonetheless, Blakey did it. Despite the ridiculously small time interval, they were both attributed to LHO.
8. The initial three suspect shots (Table 13 by Thomas), which were matched to the TSBD at t = 137.70, 139.27, and 140.32 sec, covered a total interval of only 2.6 sec, with individual intervals of 1.6 and 1.1 sec. Therefore, if LHO did it, he could not have fired the 1^st and 2^nd shots, nor could he have fired the 2^nd and 3^rd shots. If very skilled, he might have been able to get off the 1^st and 3^rd shots. The Army’s best marksman took 2.3 sec between shots.
9. At t = 140.32 seconds, the analysis suggested a shot from the TSBD, fired at the limo. It was recorded by a microphone at the corner of Houston and Elm. This shot was tossed out merely because the M-C could not be fired that quickly! (The FBI had concluded that 2.3 seconds was the minimum interval (p. 141).) The logic here, naturally, is that if the shot was not from the M-C then it could not be a gunshot! Thomas calls this logic by its simpler name—“bullshit.” Thomas believes that this shot should have been recognized as an authentic shot, although not fired by LHO.
10. Here is what is striking to Thomas: these four matches occurred in the order expected as the motorcycle progressed on the street. As a contrary example, if the matches had been random in sequence, then the first shot might have matched to a motorcycle well down on Elm St. (which makes no sense) and the last shot might have matched to a site on Houston St., both of which are out of chronological order. In fact, all four matches occurred in the proper order: time and space marched together, as they must in the real world. The acoustical experts and the HSCA found this order particularly impressive. (See Figure 16.5 by Thomas.)
11. All of these matches occurred between 12:30 and 12:31 PM. They are found within a nine second tape segment, with these intervals between suspect shots: 1.65, 1.1, 4.8, and 0.7 seconds.
12. The time between the first and last suspect shots was 8.3 seconds. For a motorcycle to travel 143 feet (a distance that was assumed) in this interval implies a speed of 11.7 mph. Independently (based on the film), the FBI had calculated the limousine speed at 11.3 mph—a lovely match for Thomas between the audio and visual data.
13. A shot from almost anywhere in Dealey Plaza was likely to mimic shots from many (perhaps most) sites—if it was recorded at the same microphone.
14. Twice as many test shots were fired from the TSBD; therefore TSBD matches were twice as likely as GK test shots.
15. BBN finally concluded that there was a 50% chance of a GK shot.

1. Mark Weiss and Arnold Aschkenasy (WA) were asked to move BBN’s estimate of 50% off of the fence (that’s the statistical fence of 50-50, not the one on the GK). In particular, they needed a more precise location for the motorcycle and its stuck microphone.
2. WA focused solely on the GK shot. They ignored the TSBD.
3. To achieve their goals of high resolution, the site on Elm St. (where the motorcycle was thought to be) would have required very closely spaced microphones (about 180 total). This promised a resolution of 1 msec. Such a detailed approach would have been extremely tedious and ultimately impractical. Furthermore, during the time of the echoes (about 1/3 second) the motorcycle would have moved several feet.
4. So they did not do another round of test firing. Instead, they just calculated how the echoes would look, based on the old BBN data. They got a surveyor’s map that showed all important reflecting surfaces. Using pins on their map, they precisely measured distances—to the buildings, to the proposed GK shooter, and to the pertinent microphones.
5. In all they correlated 22 structures, which accounted for 26 major echo patterns.
6. Beginning with the impulses from the BBN GK test shot (see Figure 16.6 by Thomas) they then calculated where those impulses should have occurred had the microphone been moved to various (nearby) locations. (They could also calculate the pattern if the shooter had been displaced by a short distance.) In short, they performed a virtual simulation—but no more test shots were fired. The next step was to compare these newly calculated impulse patterns (for this single GK shot) to the Dictabelt. And that’s where the remarkable match was made—to a motorcycle on Elm St. located five feet southwest of microphone #4. It was a perfect fit, with all 26 test echoes (for this one shot) matched to a precision of 1 msec. WA took this one step further: they had identified a specific loud impulse as the muzzle blast. So they used amplitude to screen for loud impulses (in order to eliminate irrelevant noises of lower volume); they selected 13 impulses on the test tape and 15 impulses on the Dictabelt. Comparison of these echo peaks yielded eleven coincident impulses, with an impressive binary correlation coefficient of 0.79. They calculated the odds of such a random match as 5%. On that basis, the GK gunman was now 95% alive (definitely off the fence) instead of only 50% alive. They had also located the gunman behind the stockade fence, at eight feet (+5 feet) west of the corner.

1. Weiss told the HSCA that the GK shot probably came from a rifle, rather than from a pistol. This was because of the likely presence of a shock wave (from a supersonic bullet), which was more characteristic of a rifle than of a pistol.
2. Thomas estimates the muzzle velocity at 2455 ft/sec, which helps to identify the weapon. He offers the Winchester .30-30 as a tentative fit: it was a popular weapon with a muzzle velocity of 2410 ft/sec and used a 150 grain bullet (the M-C bullet is about 160 grains.) Furthermore, within 15 minutes of the shot, Inspector Herbert Sawyer radioed headquarters, and his comments were recorded in the police log:

   The wanted person in this is a slender white male about thirty, five feet ten, one sixty five, carrying what looked to be a 30-30 or some type of Winchester.

   A radio reporter (not identified by Thomas) stated:

   Police are looking for a young white man wearing a white shirt and levis carrying a lever type action rifle.

   Thomas notes that all .30-30s are lever action rifles. Furthermore, the Winchester is a logical choice: it was specifically designed for hunting in bushy terrain where aiming and firing quickly (at moving targets) is required. Finally, it has power, accuracy, and reliability.
3. Thomas concludes this section by emphasizing the excellent correlation now achieved between the audio sequence and the Z-film. The two anchors in time were Z-313 (the headshot) and Z-224 (the SBT event). The time interval between these two events on the Z-film was 4.8 sec. On the Dictabelt, the pertinent time interval was 4.8 sec (after the 5% correction).

My summary. Suspect gunshots were identified on the police Dictabelt (Ch-1). They were attributed to an open (stuck) microphone on McLain’s motorcycle in Dealey Plaza. The HSCA fired test shots from the TSBD and from the GK. Two successive acoustic teams finally concluded that a shot had been fired from the GK.

Chapter 17: The Acoustics Challenged (p. 613)

1. Steve Barber discovered crosstalk on Ch-1 (the Dictabelt), which came from Ch-2. This implied to the NRC that the suspect gunshots had been recorded well after the actual shooting (perhaps a minute afterwards). If so, the acoustic analysis would be irrelevant—despite the statistical correlations achieved.
2. The FBI published its own analysis in December 1983. Thomas concludes, however, that no serious refutation of the HSCA was offered in that paper.
3. The NRC panel did not include a single expert on acoustics, but it did include Alvarez, who had already staked his reputation on the lone gunman.
4. The NRC report is loaded with false statements, according to Thomas. Its errors of omission were particularly bad, e.g., it excluded the crucial evidence that supported gunshots. The preliminary screening studies were not cited—these demonstrated that the suspect patterns had the expected amplitude, duration, and numbers of impulses. The NRC never stated that the suspect sounds were specific to about a ten second interval. Thomas cites several additional and critical misleading statements by the NRC, which demonstrate an obvious bias by them. (Additional discussion follows below.)
5. Here is an especially serious misrepresentation, according to the author:

   The identification of shots and impulses by BRSW was completely different from that by WA as demonstrated by the more than 200 millisecond (or more than 200 ft.) displacement between the two identifications...the BRSW analysis missed the identification that WA considers to be the primary one. [DM: BRSW is not defined by Thomas, but likely refers to the actual study authors—Barger, Robinson, Schmidt, and Wolf—of the BBN study. BRSW is not in the index.]

   Thomas states that Richard Garwin (of the NRC) argued that the acoustic teams could not both be correct. But Thomas counters that both had been correct— because the contradiction raised by the NRC did not exist.
6. Thomas explains what had happened (see Figure 17.1 by Thomas). The first echoes to reach the microphone at (3)-4 on Elm St. were unique. [DM: This is the microphone from array #3 (of 3), at position 4, the site where the GK shot had been picked up by a microphone. See Table 13 by Thomas.] The second echo cluster arrived 190 msec later, from buildings farther away (on Houston). And that is indeed the pattern on the Dictabelt (see the bottom of Figure 17.1 by Thomas). The problem was with the AGC—it reduced the volume of the echoes (for an interval of 100 msec) after an initial loud sound, and this led BBN to miss this early part of the pattern. They did, however, lock onto the second part of the pattern (190 msec later). But even using only half of the data, the correlation coefficient (by BBN) identified this as a GK shot. That only half of the data was required for the correlation is powerful evidence of their methodology, according to Thomas. Despite this, the NRC concluded that an error had been made.
7. The NRC reported another “error:”

   Essential tests to confirm both the analysis procedure and the interpretations were omitted.

   The specific procedure not tested was the WA sonar model, specifically for the TSBD shots. But WA had not been asked to test any TSBD shots, nor had they been funded for this! Thomas asks the most reasonable question: If the NRC truly believed that more tests were essential, why didn’t they just do them? After all, the NRC was not constrained by either time or funding.
8. There is actually a good reason why WA were able to analyze the GK shot. On the other hand, they would have found the TSBD shot a much greater challenge: the GK analysis had required only two dimensions, whereas the TSBD analysis would have required three (i.e., the TSBD was elevated well above Dealey Plaza).
9. Thomas states that only three of the four TSBD matches likely came from the TSBD. The fourth likely came from a nearby building. More data would have been needed to nail this down, but these were never obtained.

The NRC published its critique in Science, where its authors accused BBN of computational errors. However, they relegated those details to their appendix. Thomas protests, however, that these were not errors; they were merely the NRC’s opinion that a different set of assumptions should have been used for the tests. The NRC had not even noted the order in the data (as discussed above). In the end, Thomas insists that the NRC actually documented zero computational errors.

NRC’s central criticism focused on the assumptions that led to the WA’s 5% estimate (that a match to a GK gunman appeared randomly). Of the 13 impulses on the test shot, 11 were coincident with 15 impulses on the Dictabelt (during the expected time interval). Here then is the statistical question: Given 15 static bursts (or other random noises), what is the probability of finding 11 of them in the same time slots (of + 1 msec) as the Dictabelt candidates? This assumes a microphone located about 72 feet west (i.e., downhill toward the overpass) of the Elm-Houston intersection. Within an interval of 90 msec, BBN had calculated the odds as 20 to 1 against such an occurrence by random chance, i.e., there was a real gunman. Thomas notes that these calculations are identical to those for lottery predictions, i.e., the hypergeometric probability function, p, determines the outcome:

where M = the number of time slots (or windows)
N = the number of impulses in the evidence pattern
n = the number of impulses in the test pattern
i = the number of coincident events.

A coincident event is counted when an impulse from the evidence pattern (the Dictabelt) falls into the same time window (±1 msec) as an impulse from the test pattern (i.e., a BBN shot from the GK). The factors in the above equation are the usual binomial coefficients although, rather mysteriously, Thomas does not tell us that useful fact. I would add that, in general, where n and k are integers or zero,

For those who have forgotten their high school math:

Here is a simple example: 5! = 5 x 4 x 3 x 2 x 1 = 120. Tables of these binomial coefficients can be found online. They also form the entries for Pascal’s triangle, for which a small sample is shown here:

There is even an online site for automatic calculation of the probability. In general, this hypergeometric function is used for comparing two finite populations (the Dictabelt and the test shots in this case) which are sampled without replacement, i.e., each match decreases the remaining number of impulses that are available for further matches. Since Thomas is unnecessarily terse in his explanations here (although he clearly knows what he is doing), several points should be emphasized about p: (1) it is the probability that random chance could explain the event in question and also (2) it is the probability of a match for just one position of the microphone. A bit later (sooner would have been better), Thomas recalls that the WA model employs 180 microphone positions (2 feet apart), so that the cumulative probability for a random match (i.e., taking into account all 180 microphone positions) is 180p. (It’s like buying 180 lottery tickets instead of just one.) Suppose the final result (for random chance of a match) is 5%; then the final probability of a live GK gunman becomes 95% (i.e., what is left over after randomness is subtracted out: 100% - 5% = 95%).

Thomas does report, though, that BBN had made a major mistake (in interpreting data—see p. 628), which had been pointed out to them by WA. BBN thought that impulses had arrived in a 90 msec window, but it should have been 180 msec. Therefore BBN had overestimated the probability of a random match. (If done correctly, the probability of a GK gunman would have increased.) The NRC caught this error, which implies that they did understand the 190 msec gap between the two 90 msec clusters of sounds, and that these were not two completely separate patterns—as they had actually (and paradoxically) asserted in their critique.

Thomas then details the extremely (perhaps even radically) conservative approach taken by the NRC, by means of which they successively whittled away at the number of matches. One example was the set of assumptions in the WA model: the NRC identified seven. This supposedly allowed them (the NRC) to deduct seven coincident pairs that WA had found. The NRC also deducted two degrees of freedom because the motorcycle was free to vary in two dimensions; so was the shooter, the NRC said. But this latter was incorrect. It should have been one dimension—the fence constrained movement to one dimension (i.e., the gunman did not stroll in front of the fence, nor did he amble backwards away from the fence). Thomas correctly argues, however, that the seven assumptions were not really free variables. In fact, without including motorcycle speed, air temperature, and tape recorder speed, their work would have been criticized. Furthermore, these items were not derived from the acoustical data but rather from independent evidence, and had nothing to do with the acoustical analysis. Thomas next notes another error by the NRC: they had mistakenly included the same correction twice. This was the erroneous adjustment for M (the number of time slots or windows), an error that BBN had indeed made. However, WA had already corrected for this, so the NRC should not have corrected for it a second time.

Thomas then tells us how degrees of freedom should be counted: two for the motorcycle, one for the shooter, and one for the muzzle blast. This leads to a final probability of 100,000 to one against random chance as the explanation, meaning that the chances of a GK gunman increase a lot.

Ultimately, according to Thomas, the NRC had engaged in statisticulation—dressing up minor details as serious errors and also mischaracterizing the strengths of the HSCA analysis as though they were flaws. Ultimately, though, Steve Barber, with his crosstalk proposal, had provided the NRC with their one valid argument. And, as we shall soon see, this is where the sequel article to the NRC (Linsker, et al.—see my Appendix 4 for the citation) bases its case against conspiracy, and where Thomas ultimately runs into trouble.

Listening to a copy of the Dictabelt recording (Ch-1), Steve Barber had heard a background voice saying, “...Hold everything secure....” But Barber had also found this same phrase by Bill Decker (who had ridden with Chief Curry) in a transcript of Ch-2. So how did Decker’s broadcast (from Ch-2) get onto Ch-1? (The appearance of this spoken phrase on both channels is the source of the pun in “Double Decker.”) The answer is crosstalk, an occasional event in radio communication, which occurs when multiple persons use two different frequencies at the same time. For example, suppose that two radio units (on different channels) are close spatially and that one channel is already open and broadcasting. When the second operator initiates a call on the second channel, then the live microphone on that second channel may overhear the first channel broadcast. If so, then that will be rebroadcast from the second channel—it’s a simulcast, a simple concept. Typically, only fragments of the conversation are picked up though. In any case, such simulcasts might serve to synchronize events between Ch-1 and Ch-2.

But how did that happen in this case? It could not have been McLain’s microphone, because no other motorcycle was close enough. Thomas (p. 634) favors an open microphone (on a motorcycle) at the Trade Mart, although he names no particular officer. (Jim Bowles may have been the first to suggest this.) Some siren sounds were picked up by this same open microphone. The main clue occurs at about 12:33 PM on Ch-1, when there was a faint broadcast, probably overheard from a nearby radio, “Attention, all units.” Thomas states that this was a deliberate simulcast (on Ch-1 and on Ch-2) by Henslee (p. 639).

About 12 seconds after the dispatcher had announced the time as 12:30 PM, Curry (p. 567) had said on Ch-2, “Go to the hospital!” That had been the first overt clue (on the tapes) of a shooting. But here is the catch: on Ch-2, Decker’s broadcast (“...Hold everything secure...”) occurred a full minute after Curry’s broadcast, whereas the critical Decker phrase on Ch-1 (the one that Barber could barely hear) had overlapped the last of BBN’s gunshot impulses. The first conclusion from this is simple: there was no synchrony between the channels. But the deeper problem, if Barber had heard right (on Ch-1), was that the entire BBN analysis had been rendered useless—because their suspect shots had occurred far too close to Decker’s, “...Hold everything secure...”. In other words, because of inconsistent timing, the suspect gunshots could not derive from the actual shooting, but must have some other cause, quite unrelated to the crime. Barber, of course, proceeded to advise the NRC of his observation but, ironically, the NRC at first did not even know that the police radio system worked like a party line.

Because the sounds that Barber heard were so faint, they were subjected to an objective analysis by the NRC: a digital computer sampled the double Decker sounds at 20,000 readings per second, comparing Ch-1 with Ch-2 (the latter was a copy from the DPD). The recordings from the two channels were permitted to slide across one another in time, while the computer looked for overlap of identical or nearly identical segments. This was not conclusive, so the NRC made its own recording of Ch-2 directly from the Audograph disc. This led to a correlation coefficient of 0.5 (Figure 17.2 by Thomas), which was only modestly convincing (that crosstalk had occurred for the “double Decker”).

The same tests were done for the S. Q. Bellah communication, “YOU want me to...”. He had led the motorcade and had broadcast over Ch-2, but there had been some crossover to Ch-1. This correlation coefficient was 0.8 (Figure 17.3 by Thomas), much higher than 0.5, so this suggested that the NRC method had merit. It also implied that Bellah’s phrase was an example of crosstalk.

But the key question was this: Could any crosstalk be relied upon for synchronizing phrases between the two channels? For example, if the Bellah crosstalk was used for synchrony (Thomas would like that), then the suspect shots on Ch-1 fall right between Curry’s two broadcasts on Ch-2, i.e., the triple underpass announcement and the “Go to the hospital!” utterance (p. 567). In that case, Thomas could claim that the suspect sounds were the real shots. However, we already know that using Decker for synchrony means that the suspect sounds occur about one minute too late. So what is going on here? The NRC had stated that crosstalk could reliably identify synchrony. In this, they were almost certainly wrong, according to Thomas. (However, see below regarding the question of missing time on Ch-2.) Thomas suggests possible sources of asynchrony: a recording pause on one channel, stylus displacement (movement from one track to an adjacent one), and/or over-recording (i.e., an overdub, a recording more than once on the same track). Most likely the cause cannot now be pinned down, but clearly something was not kosher. Arriving at this conclusion may have pleased Thomas—after all, if the “double Decker” had been accepted for synchrony, then the acoustic case was dead. But now, knowing that crosstalk could not guarantee synchrony, the acoustic case was still alive.

Thomas had earlier (p. 560) stated that the HSCA had acquired (and used for its acoustical data) the original Dictabelt. This (supposedly) original Dictabelt No. 10, located at NARA, is shown in Figure 16.1 by Thomas. My first question regarding the acoustics data had always been: Was the original Dictabelt used? (This JFK case already has too many proposed duplicates: autopsy reports, autopsied brains, brain examinations, skull X-rays, posterior skull photos, Minox gear, two contradictory FBI reports for the brown paper wrapper, two Oswalds, two contradictory affidavits for CE-543, etc.) Now, suddenly, near the end of Thomas’s discussion, this specter (a pun) raises its ugly head again—perhaps the original dictabelt was missing, after all.

The evidence for a missing original is rather remarkable (although Linsker et al. disagree—see more discussion below). The BBN study had found two powerline hums on Dictabelt No. 10: at 60 and at 120 Hz (Figure 17.6 by Thomas). The 60 Hz hum is not too surprising, but how did the 120 Hz hum get there? Oddly, for Dictabelts Nos. 11 and 12 (both at NARA) there is scarcely any hum at all. An FBI study of No. 10 found only the 120 Hz hum (Figure 17.7 by Thomas), whereas the DPD copy of No. 10 showed multiple 120 Hz hums as well as the 60 Hz hum. One explanation is that the machines used for playback themselves caused some hums; since different ones had been employed, different hums may have resulted. Thomas notes that the Dictaphone amplifiers only covered 300 to 3000 Hz (the natural range of the human voice), so the 60 Hz hum lay below that range. Furthermore, the Dictaphones had no outlets for detachable speakers or for any other devices. Therefore, in order to transfer sounds from the Dictaphone to another device, a microphone for the new recording was merely placed (in the air) near the (built-in) Dictaphone speaker, surely a rather antiquated method. But that meant that any sounds in the room (e.g., overhead fans) would be picked up in the process. So here is Thomas’s verdict: “...the presence of prominent hums on Belt 10 which are not present on the other belts suggest the belt is a copy and not the original.” Dictabelt No. 10 at NARA is in such poor shape that it can no longer be used for playback. In 2004, NARA contracted with the Lawrence Livermore Lab in Berkeley to determine if optical scanning might be feasible (as had been achieved with old Edison cylinders). I asked Les Waffen (at NARA), via e-mail in late September 2010, about this issue, but got no reply. I suspect that funding has not been forthcoming.

Chris Scally conducted an exhaustive study of the custody of these recordings. Based on that work, Thomas states that the

convoluted chain of possession of the various recordings and his [Scally’s] discoveries do not reflect well on the authorities charged with safeguarding this historical evidence.

If that is anywhere near the truth it would be very difficult to introduce this tape at a criminal trial. Thomas summarizes the Dictabelt story with a surprising table (Table 16), which shows 14 different Police Dictabelts from November 22, 1963, including two labeled as #1! The start and stop times vary widely and so does the total time interval. From this, Thomas infers that a duplicate set of belts once existed; it also suggests that some of NARA’s belts are duplicates. His final comment is highly pertinent: “If the Dictabelt analyzed for the acoustical studies is a copy, then the scenarios for impositions, whether deliberate or accidental, also multiply.” Thomas’s comments had actually been preceded by Barger’s comment:

The original Dictabelt could be studied more extensively for possible evidence...of being a copy...Further studies could include a careful search for a second hum...which would characterize a copy.

However, see Linsker’s powerful argument below that the original had indeed been employed in the analysis.

After his 2001 article, Thomas was reprimanded by surviving members of the NRC—they accused him of making significant errors. According to Thomas, however, they actually only cited one: his mistake in using their adjusted timeline for Ch-2, i.e., the problem of skips and repeats on the Gray Audograph. The NRC (and Ralph Linsker, too) claims that this contains no skips. But Thomas states that it is actually missing at least four broadcasts that can be heard on the DPD tapes. In short, Thomas does not even accept this one reproof from the NRC. Since Linsker and colleagues still disagree strongly with Thomas about this matter, more discussion follows below.

Linsker also analyzed possible crosstalk that had been overlooked by the NRC. For example, here is Deputy Chief Fisher on Ch-2: “Naw that’s all right, I’ll check it.” According to Thomas (who abbreviates this as CHECK), the last three words had crossed over to Ch-1. Furthermore, on Ch-2, it occurs just two seconds before the triple underpass broadcast (by Curry) and, on Ch-1, just two seconds before the suspect shots. This case had also been cited by Jim Bowles as the exemplar of crosstalk. Bowles had been the supervisor of communications at the DPD and had prepared a transcript in March 1964 for the WC. Oddly, his name is not in Thomas’s index, although it is cited many times in Chapter 17 and it is heavily represented in Thomas’s endnotes for that chapter. For Bowles’s own account see JFK: First Day Evidence 1993, Gary Savage, pp. 313-370. (Another source is No More Silence 1998, Larry Sneed, pp 169-193.) Even the NRC had been aware of this case, but did not discuss it. Thomas emphasizes that these crosstalk sounds were analyzed by the computer, just as the Bellah and Decker cases had been, and it (the Fisher case) was indeed positive, even though Linsker (according to Thomas) never acknowledged this result. This issue will be revisited soon—Linsker disagrees with Thomas. Thomas’s significant conclusion is this:

The truth is that the Fisher crosstalk supports the acoustical analysis in that it establishes exact synchrony between the acoustically determined gunshots and the time of the assassination when used as the tie-point between channels.

He adds that the proximity of the Fisher crosstalk (two seconds before the suspect shots) makes it the most reliable case for aligning Ch-1 with Ch-2. Linsker and colleagues strongly dispute this, as we shall see below.

Another erratum: Near the top of p. 646, Figure 17.4 is cited; it should instead cite Figure 17.5.

My summary. Steve Barber discovered a paradox—Decker’s words were superimposed over the suspect shots. The NRC agreed with this. They concluded that these suspect sounds had arrived about a minute after the actual shooting, so the sounds were not gunshots after all. Thomas wonders if the original Dictabelt had been used in the acoustic studies.

Don't miss Part One and Part Three of David Mantik’s review of Hear No Evil.

FLASH! This book is now available on KIndle for the lowest price ever, of $10.99

New Edition, Updated!

Destiny Betrayed: JFK, Cuba, and the Garrison Case by James DiEugenio

Order Now! Amazon.com

Enemy of the Truth: Myths, Forensics and the Kennedy Assassination
by Sherry G. Fiester

Forensics can be a complicated subject, yet Fiester provides the reader with easily understood, accurate, information. Enemy of the Truth: Myths, Forensics and the Kennedy Assassination is so comprehensive in its approach, this work should be used in the instruction of all new crime scene investigators nationwide. William LeBlanc, CFCSI

$24.95