Just weeks before its demise . . . the Committee was given new and startling information. Some time earlier critic Gary Mack, among others, had drawn the attention of Committee staff to the possibility that the noise of gunfire might have been inadvertently recorded on Dallas Police Department dispatch transmissions made on November 22, 1963. The original recordings of these transmissions, made over two separate police radio networks, were located in the possession of a Dallas official. Police transmissions had been recorded on Department Channel 1 by means of a Dictaphone belt recorder and the day of the assassination this channel was used primarily for normal police activities. Channel 2 was used that same day as a communications link for the presidential motorcade. It was voice-activated and recorded on a Gray Audiograph Disk at headquarters. Though Channel 2 was apparently not in use during the period when the actual assassination occurred, by a fluke of a microphone transmitter on a motorcycle or other vehicle being stuck on the "On" position, approximately 5.5 minutes of the noises in and around the vehicle were recorded by the Dictaphone belt, including around the time of the shooting.

Though unclear to the unaided ear what among the various noises recorded on the Dictabelt meant, several critics postulated that among the clatter were a number of possible gunshots. The Committee decided to give this problem over to acoustics experts. These respected acoustics scientists would analyze the nature and origin of the suspect sound impulses on Channel 1 to determine if sounds of shots had been recorded; and if so, how many, the time interval, and point of origin. In May 1978 the Committee contracted with Bolt, Beranek and Newman, Inc. [BBN], to attempt the analysis. By means of sophisticated and, to the layman, complicated scientific analysis of the recordings, chief scientist Dr. James Barger located 6 impulse sequences which could have been caused by a loud noise such as a gunshot. The Committee was urged to conduct an acoustical reconstruction of the assassination at the Dallas site. Realizing that Barger's initial findings, if true, pointed to a probable assassination conspiracy, the Committee sought an independent review of his analysis by Queen's College, New York, professor Mark Weiss and his research associate, Ernest Aschkenasy. Barger's analysis and methodology for the reconstruction were concurred by the two others, and on August 20, 1978, an elaborate test in Dealey Plaza was conducted. Microphones were strategically located at 36 separate positions to record test shots fired from the sixth floor southeast corner of the Texas School Book Depository Building window and from the area behind the stockade fence on the grassy knoll from where various witnesses believed shots to have been fired. A total of 432 shot impulse sequences were recorded. These "acoustical fingerprints" were then laboriously compared with the six impulses noted on the original recording for a total of 2592 comparisons. This analysis was not completed until days before Barger's public hearing on September 11. Though cautioning that the match he had found did not prove conclusively that the impulses on the 1963 recording represented gunfire, Barger testified that his studies showed that the 1963 recording contained four sounds attributed to probable gunshots. Three of the impulses matched an origin point at the Texas School Book Depository sixth floor, and one impulse, the third in the sequence, matched an origination point on the grassy knoll. He further cautioned his findings of the grassy knoll sound to a probability of 50 percent.

Asked by the Committee to further study Barger's work to obtain more certain results of his possible grassy knoll shot, Weiss and Aschkenasy put together an analytical extension to refine the estimate. They studied Dealey Plaza determining which structures were likely to have caused echoes received by the microphones. By identifying these echo-generating sources around the vicinity of the knoll, there were able to predict what "sound fingerprints" would have been created by a shot from the grassy knoll location when picked up by an open microphone. Each location of a microphone relative to a shooter's location would, by echoes generated off constant structures, produce a unique sound travel pattern which they referred to as a "sound fingerprint." The experts were confident that their precise calculations, taking numerous variables including air temperature in 1963 and buildings structured after 1963 into consideration, gave them a certainty factor of 95 percent or better, that impulse number 3, previously identified by Barger, was in fact a shot fired from the grassy knoll. (Trask 131-132)

Indeed, there were important additional elements that corroborated the conclusion of Barger, Weiss and Aschkenasy. The positions they determined for the motorcycle at the time of the four shots traced out a path on Houston Street that fit the actual course and speed of the motorcade. . . . Moreover, an "N-wave," characteristic of supersonic gunfire, appeared in each dictabelt impulse for which the police microphone was in an appropriate position to detect it, including the recorded sound of the third shot [identified by the Committee as coming from the grassy knoll]. The most striking find, however, was the exact location of the grassy knoll gunman. According to the acoustical calculations, this firing position was behind the picket fence, eight feet west of the corner. That was just TWO TO SEVEN FEET from where S. M. Holland, a dozen years earlier, had placed the signs observed by himself and fellow railroad workers: the puff of smoke, muddy station wagon bumper, cigarette butts, and a cluster of footprints. (25:36, original emphasis)

In clear and forceful terms, Weiss and Aschkenasy reviewed their work. They had become involved at first only to review Barger's proposed test in Dallas in August. After Barger's September testimony, they had been asked to try to move the uncertainty off center "either way." They started their work in October and finished in late December. The result was "a probability of 95 percent or better [that] there was indeed a shot fired from the grassy knoll." They were, they said, able to place the shooter within a five-foot "circumference." They could pinpoint the location of the microphone to within a foot and a half. They knew that the weapon had fired a super-sonic bullet, since a shock wave had preceded the sound of the muzzle blast (they could detect both phenomena on the tape). The weapon could have been a rifle or a pistol, since either could fire super-sonic ammunition. They explained how they knew the [policeman's] microphone had been shielded at various points by the windshield of the motorcycle.

In their calculations, they had made allowances for a possible small error on the scale map they had used (less than one foot); air-temperature and humidity variances; and the characteristics of the type of radio equipment used by the Dallas police in 1963. They had double-checked their calculations, and, yes, they were satisfied with their conclusions "beyond a reasonable doubt." In addition, the scientific principles they had employed were little more than "high school physics and geometry." Anyone who had heard an echo could understand what was involved.

No, the sound [of the grassy knoll shot] could not have been a motorcycle backfire since it was preceded by the supersonic shock wave. In any event, there was no motorcycle behind the picket fence. Obviously, the bell tolling on the tape had come from somewhere other than Dealey Plaza. Could the sound of the grassy knoll shot also have come from a different area? Only if the other area were an exact acoustical replica of Dealey Plaza, and shots had been fired there too. "If somebody were to tell me that the motorcycle was not in Dealey Plaza," Aschkenasy noted, "[that] he was transmitting from some other location. . . . I would ask to be told where that location is, . . . I would go there, and, . . . I would expect to find a replica of Dealey Plaza. . . ." The shot had not been fired up in the air, they said, but at the presidential limousine. No, the sound of the grassy knoll shot could not have been an acoustical mirage. The distance was too short, and the sequence of echoes was inconsistent with a mirage. The conclusion was, they said, inescapable; it was not a matter of interpretation; there "didn't seem to be any way to make those numbers go away, no matter how hard . . . [they] tried." (Blakey and Billings 116-117)

The impulses selected for the BRSW study [i.e., part of the HSCA study] were not always the largest impulses. Frequently, large impulses were omitted and some impulses close to the noise level were retained. There are far more impulses that do not fall into the BRSW classification of "probably sounds of gunfire" than do. Since the results of the correlation coefficient calculations are highly dependent on the impulse and echo pattern selection process, it is especially critical that the scheme used to distinguish these sounds stand up to close scrutiny, with the process being spelled out in detail so others can duplicate the analysis. From the published reports, IT IS IMPOSSIBLE TO DO SO. Furthermore, weak spikes on the Dictabelt often are selected to correspond to strong patterns, in the test patterns and vice versa. (Livingstone 363, original emphasis)

While the panel offered some valid criticisms of the methodology used in the House acoustical studies, it introduced complex and controversial assumptions and made several errors of its own. In a letter of February 18, 1983, Dr. Barger noted enigmatic features in a recording upon which the National Academy of Sciences panel relied and pointed out that it "did not examine the several items of evidence that corroborated our original findings." Barger stood by the acoustical determination of a grassy knoll shot as accepted by the House Select Committee on Assassinations. (Scheim 35-36)

. . . the critical Weiss-Aschkenasy conclusion of a 95-percent probability of a grassy knoll shot was treated only in a sketchy three-page appendix [in the NAS panel's report] that made one outright error--there was only one degree, not two, of freedom associated with the position of the shooter along the grassy knoll fence. This appendix also recalculated the probability by subtracting degrees of freedom adjusted in the Weiss-Aschkenasy analysis from matches obtained, an arbitrary approximation to a complex mathematical calculation, akin to computing the volume of a cube as three by adding its dimensions. The appendix itself included the admission that this critical calculation was "possibly overconservative" and "may be unduly conservative." (Scheim 431 n 120)

The findings of Bolt, Beranek and Newman--like almost everything in the Kennedy case--have subsequently been questioned by the FBI, and by a panel assembled by the National Research Council (whose members are drawn from the Councils of the National Academy of Sciences. . . .). According to a "Notice" on the first page of the NRC report, the committee that studied the BBN findings "was chosen for their special competence and with regard for appropriate balance"--not because they were acoustics experts, which they were not.

I personally found it interesting not only that the NRC found that it had conclusively disproved the Select Committee's acoustical report and that there was no need for further study, but also that--remarkably, and just as with the findings of the Warren Commission--there was not a single dissent among any of the panel's members. (It may or may not also be relevant that, among the most vocal of the panel's members was a scientist who, before joining the panel and reviewing the acoustical study in detail, had taken strong positions in support of the Warren Commission's findings. . . .)

The NRC's principal rationale for rejecting the findings of Bolt, Beranek and Mark Weiss was that the Channel I tape contained "cross-talk" from Channel II that indicated that the portion of the Channel I tape containing the four impulse patterns identified as gunfire occurred at least 30 seconds after the actual assassination. The NRC offered possible (plausible) explanations as to various ways that such cross-talk could have gotten onto Channel I, including that the stuck microphone on Channel I was positioned near another microphone that was monitoring Channel II, and that the words being transmitted over Channel II were picked up (very faintly) by the stuck Channel I microphone, and transmitted and recorded on the Channel I Dictabelt in the police station. Subsequent re-recording is another possible explanation. The NRC in the end was not able to definitely state the cause. Nor were they able to verify that the Channel I tape they analyzed was the original DPD tape, and thus could not say for sure that the cross-talk had been recorded on November 22, 1963. Finally, subsequent private analysis as well as further review by Dr. Barger has revealed that the NRC's tests appear to have been conducted with the tapes being run at an improper speed, thus invalidating their calculations of when the impulse patterns at issue actually did occur in relation to the assassination.

And the NRC essentially ignored, and never did explain how, if these impulse patterns were not gunfire, their timing, sequencing, and qualitative characteristics were so extensively corroborated by the other physical and scientific evidence in the case. Was all of the meshing of such evidence simply a coincidence? . . . Several witnesses testified that one shot came from the grassy knoll, just as the acoustics indicated. Just a coincidence? The shock waves and windshield distortions were present on the shots where they should have been, and absent on the others. One more coincidence? Since the NRC described their findings as conclusive and not subject to question, one must wonder why the NRC ignored all of this evidence that corroborated the Barger and Weiss findings, but is totally inconsistent with the NRC findings that these impulses are not the actual sounds of gunfire. One might also wonder why the NRC never addressed, never discussed, and never attempted to explain other "cross talk" on the Channel I tape that is totally inconsistent with the NCR conclusion that impulse patterns evidencing four shots occurred 30 seconds after the actual assassination. (Cornwell 112-114)

In his testimony before the HSCA on December 29, 1978, Dr. Barger said that, having slowed down, the level of engine noise remained constant for 30 to 40 seconds. It then rose to an even higher pitch than it had earlier reached, and then remained at this high level for at least two minutes. In a letter to Bob Cutler dated February 2, 1979, Dr. Barger disclosed that the motorcycle engine was, in fact, IDLING after the shots were fired, before apparently moving off at high speed.

A detailed study of the photographic evidence showed that the actions of police motorcyclist Bobby W. Hargis were entirely consistent with these facts. Hargis, who was riding approximately 10 feet behind and immediately to the left of the President, can be seen in many photographs taken in Dealey Plaza. His testimony before the Warren Commission can be verified by reference to these photographs. . . .

Film of the motorcade on Elm Street shows that Hargis stopped his motorcycle immediately after the final shot. Bond 4 and Bothun 4, two still photographs taken 20 and 26 seconds later respectively, show Hargis returning to his motorcycle and remounting it, exactly as he testified. . . .

Unlike those of McClain, the actions of Officer Hargis correspond exactly with those which must have occurred in order to generate the motorcycle engine noises found on the radio recording. . . .

Researcher Stephen Barber has recently analyzed these sounds in great detail. His study confirms the sound of the idling engine, and also the sound of another motorcycle passing the one which is stationary. This second motorcycle is, in fact, the one ridden by Officer McClain, who can be seen riding past Hargis' stationary motorcycle in the Bond photograph taken 20 seconds after the final shot. Just after McClain goes by, Barber detects the sound of Hargis snapping his kick-stand into place as he prepares to leave the scene. During the next 15 seconds Barber notes the echo of Hargis' engine as he passes through the Triple Underpass. For the next 20 seconds, according to Barber, the engine noise reverts to an "idling" sound. It may well be, however, that this simply the drop in engine noise level which would follow as Hargis emerges from the tunnel and scans the area, exactly as he testified. [Hargis indicated that he slowed down after going through the triple underpass in order to scan the area to see if anyone was running away from the plaza.] The noise level rises again, however, as Hargis circles round and returns to the TSBD, from where he transmitted on Channel 2 of the police radio between 12:34 and 12:35 pm. It was during the latter part of the return journey that Hargis recorded the sound of sirens approaching and receding, and it is quite possible that these sirens were on vehicles which passed Hargis, going in the opposite direction towards Parkland Hospital.

This reconstruction of Hargis' movements is consistent with Dr. Barger's finding that, following the period of idling immediately after the shots were fired, the motorcycle accelerated and remained in motion for 2 to 3 minutes. It also answers most, if not all, of the outstanding questions about the identity of the motorcycle policeman with the jammed transmitter. (Scally 39-42, original emphasis)

HARGIS. A passer-by states the shots came from the Texas School Book Depository Building.

DISPATCHER. Get all the information.

- Hargis realized his transmitter was switched to Channel 1, and he simply turned it back to Channel 2. This possibility was not discussed during his testimony before the Warren Commission, nor is it mentioned in the HSCA's published evidence.

- The fault in his radio caused it to alternate between Channels 1 and 2, and it had reverted to Channel 2 automatically by the time he returned to the Book Depository and spoke to the dispatcher.

- His Channel 2 transmission was made over a different radio. This is at least POSSIBLE, since Hargis undoubtedly parked his motorcycle near the TSBD and then moved around the area on foot. He could therefore easily have used another radio to make his transmission on Channel 2. Once again, however, neither Warren Commission testimony nor the HSCA's final report address this possibility. (Scally 43-44, original emphasis)

Beginning at 262 seconds and lasting until 299 seconds [on the tape] are the sounds of several sirens. These sirens are very important as to determining the location of the cycle with the open mike. The sirens are the sirens mounted on the motorcade vehicles. They were all turned on following the assassination. The sirens on the tape sound as if they are passing a stationary microphone not in the motorcade but on Stemmons Freeway. The sirens rise in intensity, fade, rise again, and fade again. This continues and suggests that several vehicles are passing the open mike. It is clear that the open mike is not in the motorcade but somewhere on Stemmons Freeway. (Livingstone 351)

Why, now, are there no shots audible on the tape? Rifle shots in such an enclosed urban space, echoing off buildings, would be very loud and certainly were heard by everyone in Dealey Plaza. Was there some technical reason that they might not record through an open mike located somewhere in Dealey Plaza? The answer is that there was no microphone open anywhere near Dealey Plaza, and so the gunshots could not have been recorded. (Livingstone 357)

Not once in its final report did the HSCA address how a gunman firing from the knoll might have missed, nor did it speculate on where the bullet hit. Having rushed to judgment on the issue of the dictabelt and the recording showed, the Committee dissolved itself with the admonition that it was unable to identify the other gunman and that the Department of Justice should examine the audio evidence to see if it concurred with the Committee's findings.

Of course, the Justice Department did not concur, nor did the National Academy of Sciences. The NAS review, in particular, discovered the existence of "bleedover" on the Dictabelt at the time of the assassination [actually, the bleedover was brought to the NAS panel's attention by a private researcher]. This bleedover [i.e., the Decker transmission] contained dialogue between law enforcement officers that indicated the portion of the recording where shots were found had actually been recorded a minute or so after the assassination. (Moore 142-143)

During the broadcasts, the Dallas dispatcher notes the time every so often, but is subject to some error by the angle at which he had to look at his clock, and the FBI and Dallas Police found in a test with stopwatches in 1964 that the time was often out by as much as one minute. There were three different dispatchers working at the time of the assassination, and they used three different clocks. The clocks got out of synchronization by as much as one minute over the course of a month before they were reset. During the assassination, there was a great deal of confusion and pressure, and the time checks could not be expected to be perfectly accurate. And the Channel 2 recording had stopped twice at Curry's two transmissions: his "at the underpass" report and his "Parkland Hospital" transmission. The BBN report states clearly that there is no way of knowing how long the machine was off during the two recordings.

Vaughn and Barber equated the two tape times to each other nevertheless, using the simultaneous transmissions by Decker on Channel 2 and both faint and incomplete on Channel 1, a few seconds after the last shot. The "gap" is not on Channel 1. (Groden and Livingstone 254-255; Note: Livingstone no longer accepts the validity of the HSCA's conclusions about the recording.)

Said Groden, "In all likelihood, the fatal shot did not come from the Book Depository, but rather from the grassy knoll; whether or not Lee Oswald was firing, someone else had actually killed the President." He went on to describe how when the fourth shot was matched up to the pictures of the President's head "exploding," none of the other shots were in alignment with the [Zapruder] film. But when the THIRD shot was advanced to match up with those pictures "EVERY OTHER IMPULSE MATCHED AN ACTION ON THE FILM EXACTLY." In HIGH TREASON, Groden recounted how Professor Blakey took him aside and ordered him not to express to the Committee any conclusions that he had drawn from his study of the film and tapes. The Congressmen (and the world) were to be told that the fatal shot came from the rear, and the fourth shot was the only one to be considered the head shot. (Smith 147, original emphasis)

The HSCA's decision to conduct test firings from the Book Depository and the knoll alone had serious repercussions, because in ignoring other possible firing points, they ruled out the likelihood that any of the unmatched sounds on the police radio tape could be impulses caused by shots from other locations such as, for example, the Dal-Tex building. (Scally 35)

. . . examination of the correlations between the test shots and the sounds on the Dallas police tape shows that all of the matches with shots at this target were ultimately rejected as "false alarms." It seems there is NO acoustical or other evidence to prove that the second shot was fired from the Book Depository! (Scally 35, original emphasis)