HEAR
NO EVIL
THE
ACOUSTICAL EVIDENCE IN THE KENNEDY ASSASSINATION
Presented
by
Dr.
Donald B. Thomas
November
17, 2001
Dallas,
Texas
(Copyright
2001 by Donald B. Thomas--reprinted here with author’s
permission)
In
1978 the House Select Committee on Assassinations was presented
with acoustical evidence that multiple shooters had been involved
in the murder of President John F. Kennedy. During the hearing,
staff members played a tape recording for the Committee with the
explanation that they were about to hear a rifle shot fired from
the Grassy Knoll. After listening to this tape the
ranking Republican member of the Committee, Representative Samuel
Devine of Ohio, rose in the chamber to declare that he had a
great deal of experience with firearms and familiarity with rifle
fire. He knew a gunshot when he heard one, he said, and the sound
alleged to be from the Grassy Knoll could be many things, but it
was clearly not a rifle shot. The staff then explained to Mr.
Devine that the tape recording was of a test shot fired from the
Grassy Knoll that summer; not the Dallas Police tape from 1963.
The incident suggests two things. First, that one cannot
determine that a recorded sound is or is not gunfire merely by
listening with the naked ear. Secondly, it suggests that
Congressman Devine may not have been completely open-minded to
the concept under investigation by his Committee.
Over
the last year I have discovered that there are others who are
less than receptive to this evidence. While experiencing my
fifteen minutes back in March, appearing on television and radio,
the producer for ABC's Nightline
program asked me if I realized that the article I
had published (in the British forensic journal Science &
Justice) had made a lot of people very angry. I said that,
yes, I understood that. He then remarked that, on the other hand,
I had undoubtedly made a lot of conspiracy buffs very happy. "Well,
no, not really," I said. The acoustic evidence does
contradict the official version of events which holds that there
was no more and no less than exactly three shots. But, I
explained, most conspiracy buffs are convinced that JFK received
at the very least, a frontal shot through the throat, and another
through the head as well. The acoustical evidence indicates only
one shot from the front. Moreover, when one synchronizes the
acoustical evidence with the filmed evidence, the shot from the
front aligns with the head shot. The lack of evidence for a
frontal shot for the throat wound tends to support the single
bullet theory, and the single bullet theory is anathema to most
conspiracy buffs.
The
producer told me that in journalism, when those on opposite sides
of an issue are both unhappy with the reporting, they like to
think that they are probably doing something right. I am not sure
that this is a perfect analogy, but the point is that, as I stand
before you today, I am perfectly well aware that many if not most
of you are not yet convinced by this evidence. And I hasten to
add that I am not going to try to convince you. I have a good
friend who teaches Biology at a college in Georgia, a course
which includes instruction in Evolution. Naturally, his students
include many who are devoutly religious. He tells them, what I
wish to tell you now. I don't care what you believe, but, I do
care that you know the facts.
When
I first wrote my article on the acoustics, I submitted it to the Journal
of Forensic Science here in the United States for
publication. The editor kicked it back, stating that it was their
editorial policy not to publish articles on the Kennedy
assassination. He defended this policy on the grounds that no
amount of reanalysis was going to change anyone's mind. As one
who routinely reviews scientific articles for publication, I must
say that this seemed like an odd position for a scientist to
take. And I reiterate that I am not trying to change anyone's
mind. My mission is to present the facts and let people make up
their own mind. Having said that, I am going to avoid as much as
possible making an overly technical presentation of the
acoustical evidence today. The reports of the acoustics studies
are available in the HSCA proceedings if anyone needs
those details. Rather, in my talk today I am going to address the
criticisms of the acoustical evidence that have been brought to
my attention, and to show how the acoustical evidence meshes with
the other crime scene evidence, particularly the Zapruder film.
One
of the criticisms that I have been personally subjected to was
brought up, among other places, on the Fox Morning News
program. The host of that show said, "You're just an
entomologist, why should anyone believe you?" Now you should
understand that these shows are rehearsed. The producer of the
program likes to know ahead of time what the guest is going to
say, in part so they have interesting discussion, and in part so
the host doesn't look stupid. So I knew the question was going to
be asked. I was tempted to say that, "No, I am not an
acoustical expert, but I did stay in a Holiday
Inn last night." Instead I pointed out
that even an entomologist knows that a scientific hypothesis
stands or falls on the evidence behind it and not on the status
of the person who makes it. I might also have pointed out that if
expertise were the issue -- then I win.
When
the House Select Committee on Assassinations was first
confronted with this evidence, they asked the Acoustical
Society of America for a short list of the top acoustics
laboratories in this country. At the top of the list was the
expert consulting firm of Bolt, Baranek & Newman (BBN)of
Cambridge, Massachusettes. They had done the Watergate
tapes for the Ervin Committee and the acoustics study of the Kent
State shooting for the Department of Justice. These
experts determined that the assassination gunfire was on the
Dallas police tapes and they were the experts who found the "fingerprint"
of a gunshot from the Grassy Knoll.
Because
that finding was politically incorrect, and because there was an
element of uncertainty with regard to the alleged grassy knoll
shot, a second expert opinion was sought. Back to the short list,
the next laboratory was the Computer Science Department of Queens
College, New York, where Professor Mark Weiss and his
assistant Arnold Aschkenasy wrote computer programs with sonar
applications for the military. They had also published on methods
for detecting and separating real sounds from noisy backgrounds.
Using the principles of sonar analysis (echo location)
they eliminated the cause of the uncertainty and concurred that
there was scientific evidence of a shot from the Grassy Knoll
on the police tapes.
So,
if expertise is what one requires, the top acoustic experts agree
that there was scientifically valid evidence for a shot from the Grassy
Knoll. Moreover, there has never been a direct
challenge to the acoustical evidence, or its analysis, or the
methods which were used to determine that shots were present on
the police tapes.
What
about the FBI report? A report by the FBI is
sometimes cited as an expert refutation of the acoustical
evidence and I have been been criticized for ignoring the FBI
report, which indeed I did. The criticisms raised in the FBI
report, published in two installments of the Law
Enforcement Bulletin in November and December 1983, have no
scientific validity. The article was written by Special Agent
Bruce E. Koenig of the Technical Services Division, a forensic
expert with the Signal Analysis unit. Given a chance to present
these criticisms to a joint meeting of the NRC panel and
the HSCA committee's experts, most present agreed that
the FBI's arguments were "irrelevant." More
to the point, they were an embarrassment.
The
FBI report states that Weiss and Aschkenasy had
testified that they had identified the gunshot from the Grassy
Knoll on the basis of the supersonic shock wave and the
muzzle blast impulse. In point of fact, neither Weiss nor
Aschkenasy (nor anyone else) ever made any such
statement in their testimony (or their reports).
Moreover, Koenig added,
"It
is not known which characteristic Weiss and Aschkenasy actually
used in their analysis."
[Koenig
52(11):8]
Weiss
and Aschkenasy had relied on neither the muzzle blast nor the
shock wave in concluding that there was valid evidence for a
gunshot from the Grassy Knoll. Weiss and Aschkenasy's
conclusions were based on echo correlation. That is, on the high
degree of match between the echo delay pattern of recorded test
shots and an impulse pattern on the police evidence tape. Koenig
never did explain in his articles that Weiss and Aschkenasy had
used echo location to determine that there was a Grassy Knoll
gunshot, nor the methods by which BBN had identified the
suspect noise in the first place. These omissions prevented the
reader from knowing the actual data and logic behind the
conclusions of the acoustics experts.
Preliminary
screening of the police tapes by the BBN lab led to the
identification of five impulse patterns that were considered as
candidate gunshots because they had some of the acoustic
characteristics of gunfire. These suspect noises occurred
together in a cluster about two minutes into the open microphone
segment. All five suspect patterns were found to match test shots
fired in Dealey Plaza, and all five were ultimately identified as
gunshot patterns by the experts. I realize that this fact is
widely misunderstood. But no one misunderstood the analysis as
badly as the FBI. The FBI critique
states that six impulse patterns were rejected by the BBN analysis,
either because they were inconsistent with the motorcycle
position, or, they were inconsistent with the target location [Koenig
52(12):5]. This statement was completely divorced from
reality. BBN had rejected six mathematical correlations
to test shots on those grounds, but no evidence patterns were
rejected because they were inconsistent with the motorcycle
position or the target location. This misstatement by the FBI
is particularly prejudicial because it would lead a reader
to conclude that there were many other candidate impulse patterns
on the police tape besides the ones alleged to be gunshots.
The FBI
report complained that the patterns should not have been
judged as matches because,
"...tests
performed by BBN on a radio system similar to that used by the
DPD showed considerable distortion of loud impulsive sounds such
as gunshots, which resulted in the elimination of impulse peaks,
change in the position of peaks, and even the production of new
peaks where no impulse peaks previously existed."
[Koenig
52(11):5]
If
one compares these two oscillographs (Figure
1) it can be seen that just as the FBI complained,
they don't look anything alike. The graph on top is the echo
pattern of a test shot fired from the Grassy Knoll.
Below is an impulse pattern found on the police tape. The Dallas
police recording was not a high fidelity record. The automatic
gain control and limiting circuitry does distort the amplitude
and waveform of loud impulses. But, it was absolutely essential
to the understanding of the acoustical matching procedure that
impulses were not moved or eliminated, nor were impulses produced
where none had existed. Without that fundamental understanding it
would be impossible to comprehend the analyses that were
conducted on the DPD recording by BBN and Weiss and
Aschkenasy, and exactly why the FBI opinion was invalid.
Koenig missed, misunderstood, or ignored the following statement
in the Assassination Committee's report,
"The
time of the arrival of the impulses or echoes, in each sequence
of impulses was the characteristic being compared, not the shape,
amplitude or any other characteristic of the impulses or
sequence."
[HSCA
Final Report p. 70].
I
have to admit that my reaction to reading the FBI report
was one of anger. I felt that the FBI report was
prejudicial and dishonest. I now realize that the FBI
report was the product of incompetence. In that FBI report
Mr. Koening cited as evidence of his expertise in acoustical
principles, his work on the case known as the "Commie-Klan
shootout." Koenig was indeed called as an expert
witness in this case in which the Ku Klux Klan had been
given a permit to march in Greensboro, South Carolina. The
Communist Workers Party arranged for a counter-demonstration,
attracting members of the Nazi party, and anticipating trouble,
some had brought weapons. Trouble did ensue, fighting broke out
and several persons were shot dead. Koenig was asked to use
acoustical principles to determine who had fired the first shot.
An
account of the trial can be found in the textbook "The
Acoustics of Crime," by a real acoustics
expert, Professor Harry Hollien of the University of Florida,
who was also consulted during the case. He recounts,
"...a
government agent testified that he used the sounds of the gunfire
(recorded by several television crews) and the known reflective
surfaces of the environment to identify the source of the
gunfire. From these data, it was concluded that the communists
had shot first. The Klan members were acquitted of murder.
However, during the second trial (involving violation of civil
liberties), the agent indicated that he had recalculated his
measurements and now believed that it was the Klan members who
had first opened fire. Obviously,this reversal in position
resulted in some confusion and I was asked to reanalyze the
data."
[Hollien
pp. 310].
Hollien's
investigation revealed that Koenig had based his analysis on the
wrong microphone location. (In case you are wondering,
Hollien subsequently determined that a communist had shot first)
[Hollien p 311]. This criticism of Koenig is not
just an ad hominem argument. Koenig specifically cited his
performance in the Commie-Klan Shootout as evidence of his
expertise [Koenig 52(11):2]. The truth suggests
that Koenig's performance was less astute than his article had
asserted.
Koenig
has also asserted that his work in the Commie-Klan Shootout was
the first instance that acoustical principles had been used to
identify a shooter in a criminal case. This is not so. The first
such case was the Kent State shooting. That case was worked by BBN's
lead scientist, James Barger. And in that case, it should be
noted, that using the exact same procedure - echo location - and
the audio record of the incident, Barger had identified the
physical locations to within ten feet of where the first several
gunshots were fired. Using photographs and films of the incident,
the Department of Justice identified the individual National
Guardsmen in those specific locations. When these men were
arrested - all admitted that they had fired their weapons.
I
mention this evidence for the benefit of those who cite the claim
of the NRC panel that this technology is unproven.
Although its use in criminology is somewhat novel, the same
technology had been used for many years by the army for locating
enemy gun emplacements, by the Navy to navigate underwater, and
by geologists searching for oil.
I
should now clarify an issue which I earlier stated is widely
misunderstood, and that is with regard to the number of candidate
sounds examined and the number determined to be shots. By
preliminary studies in the laboratory, BBN identified six
segments of tape which contained impulse patterns which had the
acoustical characteristics of gunfire and which were then
subjected to echo correlation analysis. But, one of the six
evidence patterns was included even though it had failed to pass
the preliminary screening tests. It failed because it did not
have as many impulses (putative echoes) as the other patterns. It
was included primarily because it was in close proximity to the
other candidates, a sort of guilt by association. Secondly, BBN
wanted to test the supposition that a gun fired with the barrel
withdrawn inside the window of a building might be attenuated,
and thus have fewer impulses (echoes) than a shot fired with the
muzzle outside. Ultimately, this test failed and the pattern was
rejected (again). All of the five segments of tape identified in
the preliminary screening tests had patterns which matched to
test shots. The ultimate conclusion of the HSCA Committee was
that there was evidence for only four shots (only?). It should be
clarified that the acoustics experts employed by the HSCA never
came out and stated boldly that there were any particular number
of shots. In fact, those of you familiar with his testimony know
that Dr. Barger went to great pains to resist making any such
final conclusion for the Committee. Rather, he insisted that it
was up to the committee to decide if any of the sounds really
were assassination gunshots, and that such a conclusion should be
made by those with access to all of the evidence, including the
acoustical evidence.
I
endorse that sentiment whole-heartedly, and the object of my
lecture today is to show that the non-acoustical evidence is in
close accord with the acoustical evidence and for that reason it
would be perverse to insist that the assassination gunshots are
not on the police tape. Moreover, in spite of the Assassinations
Committee's finding, BBN had found acoustical evidence for five
shots. That is because five evidence patterns had matched to the
echo patterns of test shots. The reduction from five to four in
the final report was made not because of acoustical evidence but
because of a perversion of the non-acoustical evidence.
In
detection theory one has to allow for false alarms. It is
reasoned that because acoustical principles were used to make the
detections, only non-acoustical evidence should be used to judge
whether a detection is true or false. In this instance unreliable
non-acoustic evidence and the non-expert advice of the
Committee's Chief Counsel resulted in the judgment that one match
was a false alarm. Oswald's rifle could not be fired in less than
2.25 sec. The problem was that the sound in question occurred too
close to the previous sound identified as a shot, only 1.1 sec,
and thus, inasmuch as both could not have been fired from the
Book Depository rifle, the latter was judged a false alarm [8
HSCA 65]. This reasoning was tantamount to saying that if
Oswald didn't fire it, it wasn't a shot. Logically, the same
reasoning should have been applied to the spacing between the
first two putative shots (1.6 sec). When the weapons testing
evidence is applied objectively, the shot that should have been
identified, as the rogue shot was the second, not the third. But
the second could not be reasonably dismissed as a false alarm
because four test matches were achieved. I have consulted with
Professor G. Robert Blakey, the former chief counsel for the
Assassinations Committee, and Dr. Barger on this issue. Evidently
there was a misguided perception that the Committee members might
be more easily convinced of the acoustics evidence if there were
not a rogue shot. Dr. Barger admitted to me that the criteria for
judging a "false alarm" in this instance was "ad
hoc," -- which is Latin for "bull-oney." Some
matches were judged to be false alarms because it would require
an unrealistic microphone trajectory. That is not the case for
the third noise. On the contrary, it falls exactly into the order
required by the working hypothesis.
THE
ORDER IN THE EVIDENCE
For
those who would hear no evil, the matching of the test shots to
the sounds on the police tape is dismissed as a case of random
noises which by chance happened to resemble gunshots. Yes, it
could happen. Perhaps there was a solar flare just eight minutes
before the President turned on to Elm Street and a burst of
electromagnetic radiation struck the police radio system giving
rise to static clusters that resembled three shots from the book
depository, one from the Grassy Knoll, and another from one of
the buildings behind the President. My article in S&J
provides a formal statistical calculation against that
probability. For the sake of argument, let us suppose that there
were extenuating circumstances that are beyond our present
knowledge and the finite probability was actually reasonably
high. Even so, if the evidentiary and test patterns had no real
commonality and the matches were entirely spurious, then matches
would occur at any of the 36 microphone locations,
and
in no particular order. But they occurred in exactly the right
order. That is, if we number the noises on the police evidence
tape in chronological order, 1-2-3-4-5, the matches were found at
microphones that line up along the path of the motorcade in the
same 1-2-3-4-5 order. There are 125 ways to order five numbers,
all of which have an equal chance of occurring and only one of
which is, 1-2-3-4-5.
The
similarity in the spacing is also remarkable. The spacing between
the noises on the police tape is: 1.6, 1.1, 4.8 and 0.7 seconds.
The array of test microphones were set at 18 foot spacings (Fig. 2). The first three matches were found
at three consecutive microphones on Houston Street. The last two
matches were found at two consecutive microphones on Elm Street.
The distance between the third and fourth matching microphones
was 72 feet. Thus, the spacing between the microphones is a close
match to the spacing of the noises on the police evidence tape.
Moreover, the distance from the first matching microphone
location (no. 5 of array 2) to the last (no. 5 at array 3) was
143 feet. The time separating the first and last shot was 8.3
seconds. For a motorcycle to travel 143 feet in 8.3 seconds its
trajectory would have to be 17.2 feet per second; equivalent to
11.7 mph. This is an impressive coincidence when one compares it
to the FBI's reconstruction in 1963 which estimated that the
President's limousine was traveling at an average speed of 11.3
mph on Elm Street. Thus, the topographic order in the matching
data is in remarkably close accord with the working hypothesis
that a police motorcycle with an open microphone was traveling in
the motorcade, northerly on Houston Street and westerly on Elm
Street at a speed of around 11 mph when the President was killed
by gunfire. That seems like a lot to ask of sunspots. It was this
order in the data that caused the acoustics experts to conclude
that there was valid scientific evidence that the assassination
gunfire had been captured on the police tapes.
FILMED
EVIDENCE OF THE MOTORCYCLE.
The
obvious procedure to follow at this point is to examine the
newsreels and photographs of the Presidential motorcade in Dealey
Plaza to see if there was a motorcycle in the location predicted
by the acoustical evidence, and if there was, did that motorcycle
have a sticky microphone relay switch. Some of you already know
the answer to that question. Richard Trask's otherwise excellent
book, "Pictures of the Pain," states that
there was not. So let us clarify the record.
If
there is any validity to the acoustical evidence then there had
to have been a police motorcycle with an open microphone in the
vicinity of the intersection of Elm and Houston during the
assassination. More precisely, the open microphone had to have
been on Elm Street 141 feet behind the President when an assassin
fired from the Grassy Knoll, if one infers, as I do, that the
Grassy Knoll shot was the fatal shot. Our chronograph of the
assassination is the Zapruder film. So we begin our analysis by
synchronizing the putative shots to specific frames in the
Zapruder film. This synchronization is shown in Table 1.
Table
1.- Synchronization of Putative Shots to Zapruder Frames
ACOUSTIC
TAPE TAPE-TIME REAL
Z-FRAME
SHOT
EVENT
TIME INTERVAL
TIME EQUIVALENT
ORIGIN
A
136.2 - 8.7
9.1 Z - 147
No Match
B
137.7 - 7.2
7.5 Z - 175
TSBD
C
139.2 - 5.6
5.8 Z
- 204 Rogue Shot
D
140.3 - 4.6
4.8 Z
- 224 TSBD
E
(145.1) 144.9
0
0
Z - 312 KNOLL
F
145.6
+ 0.7
0.7 Z - 326
TSBD
Tape
Times from BBN Report Table 2.
Event
E time correction at 8 HSCA 115.
Tape
speed correction factor 1.043 [8 HSCA 27].
Zapruder
Film speed 18.3 fps.
In
the mid-section of the motorcade there were four motorcycle
patrolmen: Marion Baker, Clyde Haygood, J.W. Courson and H.B.
"Buddy" McLain. Of these, Baker and Haygood stopped to
search for the assassins in Dealey Plaza. Because the motorcycle
motor noise on the police tape does not stop, only Courson and
McLain are viable candidates for the source of the broadcast, if
the broadcast originated in Dealey Plaza. In testimony to the
House Select Committee on Assassinations, McLain acknowledged
that he had a chronic problem with a faulty microphone relay on
his unit that caused it to stick open from time to time.
This
photograph (Fig 4) taken by Wilma Bond,
shows McLain and Courson on Elm Street in front of the Grassy
Knoll where patrolman Bobby Hargis had stopped to search for the
assassin. Hargis' motorcycle was parked on the south of the sixth
pair of roadstripes from the intersection at Houston Street, just
beyond where President Kennedy received the fatal shot. According
to Richard Trask [p. 208] the Bond Photo was taken
"within 20 sec" of the shooting. However, the scene has
to happen later than that. A discontinuous film taken by Mark
Bell also shows McLain and Courson passing Hargis. But an earlier
sequence shows a witness in the background named Charles Hester
rising from the ground. At the sound of gunfire Hester had pushed
his wife to the ground and covered her body with his own. Hester
is seen standing up in other films, in particular, a newsreel
shot by Dave Wiegman. Wiegman's film is a clock because it can be
connected to the pivotal Zapruder film. A brief instant of the
Wiegman film shows the President's limousine approaching the
underpass. In the Zapruder film, the President's limousine
arrived at the underpass at frame 463, which is 8.2 sec after the
head shot. The Wiegman film is 27.3 sec long and the frames
showing the limousine approaching the underpass appear 11 sec
into the film. Therefore, the Wiegman film begins no later than
This
map (Fig. 5), uses the acoustic evidence
coupled with the aforementioned film evidence, to plot a
trajectory for McLain's motorcycle through Dealey Plaza at the
time of the assassination. Point (c) is McLain's position as seen
in the Bond photo. Point (b) is the acoustically determined
position of the motorcycle at the time of the Grassy Knoll shot.
It is defined as 97 ft south of the TSBD and 27 ft east of the
south west corner of that same building [8 HSCA 28]. The
path of the motorcycle is shown in the south lane of Elm Street
because the films show McLain there. From point (b) to point (c)
the plotted trajectory would require McLain to idle at about 4
mph. This is consistent with the police tape which indicates that
the motorcycle idled for about 30 sec after the last shot [5
HSCA 714].
The
five circles represent the 18 ft radii of the microphone
positions that recorded test shots that matched impulse patterns
on the Dallas Police tape. The acoustical reconstruction requires
that the motorcycle passed through each of these circles in
succession. This distance, from first to last, is 143 feet, and
the time lapse was 8.3 seconds. This required a velocity of
approx. 11 mph during this sequence.
Point
(a) on this map corresponds to McLain's position as seen in a 5
sec film sequence taken by Robert Hughes, the last sequence shot
by Hughes prior to the assassination. Hughes wrote in a letter to
his parents,
"About five seconds after I quit taking
pictures we heard the
[Trask
p. 265].
The
frame showing McLain (Fig. 6-see bottom) is the 20th
frame of the sequence, and thus occurs four sec before the end of
the film. Added to Hughes' five sec estimate places the frame
nine sec before the first shot. I have estimated the time to be
closer to six or seven sec before the first shot. The audio
record shows the sound of the motorcycle motor at a constant
level until 3 sec before the first putative shot, when the motor
noise drops to one fourth of that level [8 HSCA 11]. This
requires the motorcycle to travel at a faster speed prior to the
shots than during or after the shots. By setting H-20 at 6-7 sec
prior to the first shot a trajectory of about 20 mph is projected
for McLain's motorcycle on Houston Street. This speed is about
twice the speed of the motorcade itself and is required because
at this time, the President's limousine has already turned on to
Elm Street, 220 ft away. Recall that the acoustically defined
position at the time of the head shot is only 141 ft behind the
limousine, thus, McLain has some ground to make up.
The
proposed trajectory forms a hypothesis which is subject to
testing. There have been assertions that films and photographs
prove that McLain was not or could not have been in the
acoustically predicted
According
to the critics, this action corresponds to frame 160 of the
Zapruder film. This map (Fig. 7), is a
plot of the position of the cars seen at the intersection of Elm
and Houston at frame 160. If this interpretation is correct; that
it corresponds exactly or even closely to H-20, then McLain
cannot possibly be in the locations required by the acoustical
evidence because Z-frame 160 synchronizes to less than one sec
before the first shot (at Z-175). McLain would have to travel
about 180 feet in less than one sec, requiring a velocity of 140
mph.
But
this interpretation is probably not correct. An important factor
in this analysis is the speed of the motorcade. One can measure
the speed of the cars by counting the number of frames it takes
for the vehicle to pass objects in the background. In this case,
Car-5 can be seen passing a lady in red from Z-frames 144 to 180.
That is, it required 36 frames, or 2 sec, to travel its own
length which was about 16 ft. This calculates to a speed of only
5.5 mph. This means that a small error in the car's position can
mean a big error in the estimated time lapse between events.
On
this map (Fig. 8) I have plotted an
alternative interpretation of the car positions seen in H-20.
This interpretation suggests that Car-5 began its turn prior to
the point where it came into view of Zapruder's lens, and then
because of the sharpness of the turn, had to make a second
steering adjustment while in the intersection that is seen in the
Zapruder film. The question is, which of these two
interpretations is more likely to be correct. I would argue that
a connection between the Hughes film and the Zapruder film can be
made more accurately by relying on the cars closer to Hughes,
because their position is less equivocal than the one farthest
away. Such a car is car No. 8.
This
map (Fig. 9) plots the positions of Car-8
as seen in the Zapruder film and as it is seen in the Hughes
film. We can plot its position in the Zapruder film exactly,
because it comes into alignment with an oak tree between Zapruder
and Houston Street at frame 220. Moreover, we can measure the
velocity of this car in the same manner that we measured that of
Car-5, by measuring its change in position relative to stationary
objects. In this instance it required 21 Z-frames to travel its
own length. Because Car-8 is a two-door Chevrolet Impala its
chassis length was 15 ft (180 inches according to the 1964
Chilton's manual). Its speed thus calculates to 8.5 mph (=12.5
ft/sec). The difference in speed between Cars-8 and -5 is
understandable because the turns through Dealey Plaza were
causing an accordion effect, such that the cars would bunch and
slow down through the intersections, but then on leaving the
intersection, space would open and the cars could speed up on the
straight-aways.
The
distance between Car-8's measured position at Z-220 and its
estimated position at H-20 is 88 feet. At a speed of 12.5 ft/sec,
the car would need 7 sec to cover the distance. At 18.3 frames
per sec, H-20 would be equivalent to Z-90, not Z-160. This would
place McLain's position at the Main Street intersection about 4.6
sec prior to the first shot. But, if one factors in the accordion
effect, Car-8 was probably averaging between 6-8 mph and thus
McLain's position in H-20 is likely closer to six to seven sec
prior to the first shot, in accord with Hughes memory that the
shots occurred several seconds after he stopped filming.
Another
mistaken assertion that filmed evidence discounts the acoustical
evidence, traces back to the days of the Assassinations
Committee. The Committee published a frame from the Dorman film
showing a motorcycle officer at the corner of Elm and Houston
which was supposed to be officer McLain. To the officer's right
was an automobile asserted to be Car Number 8 and it was further
asserted that this time and location was coincident with the
predictions of the acoustical evidence. Both assertions were
wrong. In the first place, in order to be in the right place,
McLain should have rounded the corner in the proximity of Car-6.
It was subsequently realized that the automobile partially
visible in the Dorman film was actually the eleventh car in the
motorcade and this places the motorcycle well back of where it
must be to have the microphone that recorded the assassination
gunfire.
But
the officer in the Dorman film is not McLain; it is Clyde
Haygood. This can be seen by examination of the newsreel footage
taken by Malcolm Couch. This still (Fig 10)
is a frame from this newsreel, which from context we can see was
taken a few seconds before the Bond photo. Couch's film shows all
four of the motorcycle patrolmen at the mid-section of the
motorcade. In this single frame we can see three. McLain is way
in the distance approaching Hargis's parked motorcycle, Courson
is about half way to McLain, and here is Haygood. In the running
film one can see Haygood passing Couch on the left. Couch was in
the tenth car of the motorcade. To orient the situation I have
prepared this plot of the vehicle positions (Fig.
11). Car-10 is at the first road stripe on Elm Street when it
was passed by Haygood. This means that Car-11 is at or near the
corner and this means that the sequence seen in Couch immediately
follows the sequence seen in the Dorman film where Car-11 is
approaching the corner. Therefore the motorcycle officer next to
Car-11 in the Dorman film has to be Haygood.
When
Officer McLain learned that his testimony to the Assassinations
Committee was supportive of evidence that a motorcycle radio had
recorded the assassination gunfire, McLain claimed that he had
stopped on Houston Street, and that moreover, he claimed that he
watched Mrs. Kennedy climb out on to the trunk of the limousine
by looking through holes in the reflecting pool wall between
himself and the other side of the Plaza. However, McLain's memory
is contradicted by the recollections of J.W. Courson. Perhaps
unaware of McLain's statements, Courson related to researcher
Larry Sneed in "No More Silence," that
just as he turned the corner on to Elm Street he saw Mrs. Kennedy
out on the trunk of the President's limousine. This event is seen
in the Zapruder film about 2-4 sec after the head shot. Thus,
Courson must have reached the corner at about the time of the
head shot, and because McLain is well ahead of
Although
no other films or photographs show McLain between the Hughes film
and the newsreels taken after the shooting, there are films
showing portions of Elm Street and Houston Street during this
sequence. These at least show us that McLain was not where he was
not supposed to be. It is sometimes asserted that McLain should
be visible in the Altgens photograph and his absence is cited as
evidence that he is not where he should be. Here is the
photograph in question (Fig. 12). It is
equivalent to Zapruder frame-255. Altgens was about 60 ft in
front of the President's limo. Note that the vice-president's car
is aligned with a shadow thrown by a tree on the south side of
Elm. Most importantly one can see that a portion of Elm Street is
missing, although a portion of the south curbside is visible. The
field of Altgen's lens-eye view ends to the right at a point on
the Dal-Tex building between the Houston Street entrance and the
corner of the building. A plot of Dealey Plaza shows this line of
sight (Fig. 13) and shows how, based on
the proposed trajectory, McLain should be out of sight, to the
right and behind the fifth car in the motorcade.
Sequences
in the Dorman film show both Car-5 and Car-6 at or near the
intersection of Elm and Houston, which is about the time that
McLain is predicted to have rounded the corner. Unfortunately,
Dorman's film is stop and go. Moreover, her camera had a
telephoto lens which greatly narrowed the field of view when she
did film Elm Street. In this case, Dorman stopped filming just as
the nose of Car-5 entered the last frame of one sequence. When
she began filming the next sequence, the view just captures the
rear half of Car-6, the Mayor's car, passing to the west. Had she
filmed continuously we might have had evidence that McLain was,
or wasn't, where he is supposed to have been.
Similarly,
it has been alleged that McLain should be visible in the Zapruder
film. There is a brief sequence, between frames 175 and 190,
about one sec, where the Mayor's car, is undoubtedly in the scope
of view, but hidden behind the crowd on the corner of Elm and
Houston. McLain should be between the Mayor's car and the crowd,
but if he is, he is also hidden from view. Some researchers
believe that one object visible in gaps between the bodies is
McLain's helmet and another is the wheel of his cycle. I am not
convinced one way or the other.
The
bottom line is that the film evidence is not definitive with
regard to whether McLain was or was not in exactly the right
positions required by the acoustical evidence because we simply
do not have pictures showing these positions when Mclain is
predicted to be there. However, if McLain was in continuous
motion between where the motion pictures by Hughes and Couch show
him to be, he would have been at least close to the predicted
positions - and he did have a sticky microphone relay. In itself
that is a remarkable coincidence if the gunshot sounds are
nothing more than solar flares.
THE
DOUBLE DECKER
Unable
to find significant errors in the acoustical analysis, the
National Research Council's panel on ballistic acoustics relied
on an artifact to raise doubts about the validity of the
acoustical evidence. If there was reasonable doubt, then one
could argue that the acoustical evidence was not proof that there
was a gunshot from the Grassy Knoll. The artifact,
actually discovered by researcher Steve Barber and dubbed the
Double Decker, suggests that the noises on the police tape might
not be synchronous with the time of the shooting.
The
evidence for this assertion is a barely audible fragment of
garbled speech that occurs at the time of the putative shots on
police channel one. The graphic I am showing you now (Fig. 14) depicts the timeline of events
that occurred in the minutes immediately following the
assassination. The Ch-2 recording shows that Sheriff Decker made
a broadcast that included the phrase, "Hold everything
secure..." almost exactly one minute after Chief Curry
broadcast his order, "Go to the Hospital..."
Curry would not have given his order unless he knew there was a
medical emergency in the motorcade. Because a portion of the
Decker broadcast crossed over to channel one and is essentially
simultaneous with the alleged shots, the sounds reputed to be
gunshots must actually occur well after the assassination.
But
this hypothesis is valid only to the extent that crosstalk can
give us an unequivocal synchronization of events recorded on the
two channels. Although it is true that alignment of the Decker
broadcasts fails to match the putative gunshots to the time of
the assassination, in point of fact, matching of the Decker
broadcasts fails to align any events on the two channels.
Conversely, using any of the other instances of
The
assertion that Curry made his broadcast "almost
immediately" following the shots has been criticized on the
grounds that Chief Curry would not have known that anyone had
been shot so soon after the shooting. In an interview with
researcher Gary Mack, Curry recalled that he learned that the
President was shot only when the President's limousine and escort
caught up with his car at the underpass, an event which is seen
in films to have occurred at least 15 sec after the shooting.
However,
there is better evidence that establishes an earlier contact.
Secret Service Agent Roy Kellerman was in the limousine with the
President. That evening at Bethesda Naval Hospital, Kellerman
related his experience to FBI agents Siebert and O'Neil, who were
detailed to observe the autopsy. Kellerman repeated his story in
a formal interview four days later, and in his testimony to the
Warren Commission. Kellerman stated that after the first volley
of shots he clearly heard the President say in his Bostonian
accent, "My God - I'm hit." Kellerman related that he
had his microphone in his hand, that he was in radio contact with
secret service agent Winston Lawson, sitting next to Curry, the
driver of the pilot car, and that he immediately called with
orders,
"Lawson,
this is Kellerman...We are hit; get us to the hospital
that
I talked just now, a flurry of
[2
WCH 73-74].
Kellerman's
account of events recorded on the day of the assassination must
be given more weight than Curry's recollection of events many
years later. One could also argue that the reason that the lead
car stopped under the triple underpass and waited for the
Limousine and escort was because they already knew that something
was wrong. Kellerman's testimony explains why Curry could have
made his call so close after the time of the shooting. According
to Kellerman's testimony his call to the lead car was made before
the shooting was over, not after. Interestingly, although there
is no recording of the Secret Service channel, James Altgens, who
was standing next to the Limousine on Elm Street at the time of
the head shot, claims that he heard Kellerman radio "We've
been hit, get us to the nearest hospital" [Trask p.
315].
Thus,
the alignment of the channels placing the time of the shooting
immediately before Curry ordered his officers to go to the
hospital is in accord with Kellerman's contemporaneous account of
events. Nonetheless, it is also true that the timeline is not
completely reliable. If it were true, as the NRC panel implied,
that one can use crosstalk to synchronize events on the two
channels, then one should be able to use any instance of
crosstalk. but, it can be seen that whichever crosstalk one
chooses, the others will not align. Several reasons have been
offered to explain why the events on the two channels are out of
synch.
Firstly,
the original recordings were made on machines which utilized a
stylus which etched an acoustic groove into a soft polyvinyl
surface. The recording instruments were useworn and had developed
idiosynchracies. Jim Bowles, head of police communications unit
at the time, writes that the needles would sometimes not
"groove" properly, that parts of messages would not be
recorded, or a "ghost" signal would be recorded. He
also states that it was a common experience to observe noticable
changes in speeds between units [Bowles 1979].
Secondly,
the original recordings had become scratched and worn from
multiple playbacks during the transcription process that was
applied during the Warren Commission's investigation. As a
result, the tendency for the stylus head to skip was exacerbated.
Thirdly,
both recorders had a sound-actuation feature which was designed
to save space on the recordings by pausing whenever there were
periods of dead-air. To the extent that this happened during the
critical sequence of events there would be disagreement between
tape time and actual time.
Fourthly,
the electronic recordings which have entered into evidence
involved the use of separate playback and/or recording
instruments, sometimes both, which inevitably results in a time
warp because playback speed and original recording speed are
unlikely to match precisely.
The
NRC panel undertook a study of the recordings for the purpose of
identifying the discrepancies caused by these problems and to
eliminate them to the extent possible. The timeline which you see
here in Fig. 14 resulted from that study
and are the time intervals presented in the NRC report at their
table C-1. The reason that misalignments still remain is largely
due to their use of the Bowles tapes. Bowles recorder ran out of
tape half way through the critical sequence of broadcasts. Skips
and repeats are numerous. It is suspected that the recorder
stopped at times during the sequence and the duration of these
silences cannot be measured directly. There is also evidence that
either the audograph machine or Bowles tape recorder was varying
in speed. The NRC panel was unable to resolve this issue
completely. During the deliberations of the NRC panel, its
members met with scientists from the HSCA investigation and
attempted to resolve the differences. Each favored a different
timeline and, for whatever reason, they failed to reach a
consensus.
But
there have been new developments on that front. In the last few
months a new investigation of the channel 2 timeline has been
undertaken by an expert named Michael O'Dell. O'Dell has
graciously given me permission to show you this new channel 2
timeline (Figure 15). O'Dell acquired a
copy of the channel 2 recording that was originally made by the
FBI for the NRC panel directly from the original audograph disc.
This recording was mentioned by the NRC panel but the timeline
from this recording was not relied on because of a warp
introduced by using a standard turntable (which plays at uniform
revolution speed) instead of an audograph disc recorder (which
plays at uniform track speed). The advantage of this recording is
that it eliminates the skips and repeats which occur in playback
with an audograph machine. What O'Dell has accomplished in his
studies was to find a way to correct for the warp in speed caused
by the difference in turntables.
The
reasons for believing that this new timeline is a closer
approximation of reality, or at least, has fewer artifacts,
include, 1) it actually agrees closely with the same correction
that the NRC panel used, but ultimately decided not to rely on,
and 2) it agrees closely with the timeline preferred by BBN,
arrived at using a completely different method. Over the six
minute period of interest they agree within four seconds.
Now,
as to the significance of this new timeline. The NRC panel and
its defenders argue that a simple explanation for the
discontinuities in the timeline, which we can see by comparing
the crosstalk events, is the action of the sound activation
feature which would cause the recorder on channel 2 to stop
whenever there was dead air. The problem with that hypothesis, as
I noted in my article, was that the dispatcher time notations are
in close agreement to the actual elapsed time. As one can see in
this graphic, the 12:32 time notation occurs exactly two minutes
after the announcement of 12:30 and there is similar agreement
with some of the others. This time notation at 12:36, is almost
exactly six minutes after the 12:30 notation. If one assumes that
the time discontinuity is due to a stoppage of the recorder and
puts back a full missing minute between HOLD and YOU on channel
2, it throws the time notations unreasonably out of whack. I
therefore argued that a stoppage of the recorder for a full
minute in this sequence was implausible. An alternative
explanation was that the stylus head had skipped during the
recording process and that no significant amount of time had been
added or lost.
O'Dell's
new timeline however, by eliminating most if not all of the speed
warps, skips and repeats, changes the juxtaposition of these
events. As you can now see, if one uses the Bellah broadcast YOU
to synchronize events, the shots no longer align with Curry's
broadcast. Of course, because we should be free to use any
crosstalk to synchronize the channels, if one uses the Henslee
broadcast - "Attention all units...", which was
actually a deliberate simulcast, not an accidental crosstalk, we
re-arrive at synchroneity between the putative shots and the time
of the assassination.
But
importantly, O'Dell's new timeline reduces the time discrepancy
between the two Decker broadcasts from one minute to only about
35 seconds. This reduction in the time discrepancy makes the
hypothesis of the recorder stopping much more plausible. To put
it in lay terms, my objection to the NRC's hypothesis is largely
blown away.
But
that doesn't mean that the hypothesis is correct, only that the
hypothesis is now plausible. It is still not strong enough to
refute the acoustic evidence. First of all, there is no direct
evidence that the recorder stopped. Secondly, the jumping stylus
theory is still in better agreement with the time notations than
with the recorder stoppage theory.
Note,
for example, that there is about a 20 sec offset between the
Bellah and Henslee broadcasts on the two channels. One can fairly
assume that this offset is due to missing time, perhaps due to a
pause in the recorder. We can correct for the discrepancy by
adding 20 sec to the timeline at a point just before the Henslee
broadcast on channel 2. By doing so we achieve alignment between
all of the crosstalks, except the
The
value of the slope will be one if there is a perfect agreement.
This second value, called the intercept, will be zero if there is
a perfect agreement. I have also shown a third value,
"r" the regression coefficient, because statisticians
would look for it, but the value doesn't change as you can see.
With our proposed correction the slope moves closer to one and
the intercept moves closer to zero. This gives us confidence that
our proposed correction is reasonable. Similarly, you might have
noticed that the 12:31 time notation on ch-2 was offline. This
Table 2.- Regression Analysis of Time Notations: and test
of
Time Y
X
X
X
Notation (sec) (sec)
Corrected Hypothetical
12:30 0
0
0
0
12:31 60
95.8 95.8
95.8
12:32 120
121.2 121.2
156.2
12:34 240
212.8 212.8
247.8
12:35a 300
268.9 268.9
303.9
12:35b 300
300.3 300.3
335.3
12:36a 360
329.9 349.9
384.9
12:36b 360
358.8 378.8
413.8
Slope .902
.944 1.04
Intercept 14.8
10.7 16.0
r
.99 .99
.99
w/o
Slope .947
.991 1.07
12:31
Intercept 0.2
-4.8 6.6
r
.99 .99
.99
might
well have been due to the fact that the dispatcher had to wait
for Sheriff Decker to stop talking on channel two before he could
make his own broadcast. If we recalculate the agreement between
the timeline and the other notations without the 12:31 notation,
the slope again moves closer to one and the intercept moves
closer to zero. Therefore, let us make both corrections. Though
the intercept misses the mark a little, the slope becomes nearly
perfect at 0.99.
One
can apply the same reasoning to the hypothesis that the recorder
stopped between the Decker and Bellah broadcasts. But when one
adds the missing 35 seconds, the opposite happens. Both the slope
and the intercept move away from agreement with the dispatcher's
time notations. This evidence suggests that there is no time
missing between the Decker and Bellah broadcasts, or at least,
not as much as 35 sec.
Now
we all remember what Mark Twain said about lies, damned lies and
statistics. The regression analysis above does not prove one
hypothesis over the other. Numbers only force us to be logical,
not factual. There is no reason to believe that the dispatcher's
time notations (other than the station identification at 12:30)
were chronometrically exact. Moreover, a modern study, such as
that applied to ch-2 by O'Dell has not yet been applied to ch-1.
The point is that the uncertainties inherent to these timelines
preclude them from being definitive proof that the putative
gunshot sounds are or are not precisely synchronous with the time
of the assassination, even if the regression analysis were to go
the other way.
Our
bottom line with regard to the timeline evidence is much like the
bottom line with the microphone trajectory evidence. The evidence
is not sufficiently definitive to say that the putative shots
were or were not exactly synchronous with the assassination. One
can fairly say that there is evidence that they might not be. But
one can also say that they are certainly close - within the same
minute of time.
ZAPRUDER
FILM EVENTS
Earlier
I endorsed the concept that all of the crime scene evidence
should be considered in an assessment of the validity of a shot
from the Grassy knoll. Abraham Zapruder's home movie is the
single most important piece of evidence in the Kennedy murder. Of
the many films of the assassination it is the most comprehensive.
The 26 sec film has been used by all official investigations,
including the Warren Commission and the House Select Committee on
Assassinations as the standard chronometer of events. Although
bullet impacts are gruesomely depicted in the film, the Warren
Commission was unable to arrive at an explanation for the
sequence and spacing of the gunshots coherent with the
presumption of a single assassin. Thus, the official version
posits that the apparent wounding of Governor Connally was a
delayed reaction; that, against all medical expectations, a
victim shot through the neck would raise his arms; that the
assassin chose to shoot at, and managed to strike, a target he
couldn't see; and it insists that a person shot from behind would
be somehow thrust backwards. Then, just to top it off, a spent
bullet devoid of any connection to any of the wounds, was
declared to be responsible for nearly all of them!
Noted
historian Barbara W. Tuchman once wrote an essay on "Practicing
History." She wrote,
"It
is wiser, I believe, to arrive at a theory by way of the
Events
depicted in Zapruder's Film
A
forensic study of the Zapruder film was conducted by the HSCA
Photographic Evidence Panel. The film shows the passage of the
President's limousine from the east end of Elm Street until it
disappeared under the railroad underpass at the west end of
Dealey Plaza. It is known that the shooting occurred during this
sequence. A major impediment to the interpretation of the
Zapruder film is that an intervening freeway sign blocks the view
of the President for a critical second from frames 207 to 224.
With that proviso, the panel found four episodes in the film in
which the passengers of the President's limousine appeared to be
reacting to a severe external stimulus.
A)
Beginning at frame 165 Governor John Connally, sitting directly
in front of the President, makes a rapid head movement 90 deg
to his left, then turns completely around in the opposite
direction to glance back over his right shoulder. In his
testimony to the Warren Commission the Governor stated that he
turned to look back in response to hearing what he believed was a
gunshot [WCR:112].
B)
In the sequence of frames 194 to 207 (at which point he
disappears behind the sign) President Kennedy suddenly froze his
waving hand and abruptly raised his right elbow which had been
resting on the car windowsill. He then shook his head from right
to left. During the same
C)
Between frames 224 and 240 Governor Connally's posture stiffened,
then contorted. His right shoulder dropped, his cheeks puff out
and a look of anguish appeared on his face. He then collapsed
into the arms of his wife seated next to him. During this same
sequence, the President's arms, which were in a position folded
in front of his body, appear to flap up at the elbows.
A
1967 photogrammetric study of the Zapruder film by the ITEK
corporation, under contract to CBS News, revealed that
following a stiffening of the Governor's posture at frame 224,
his white Stetson hat, held in the right hand, flipped up and
rapidly down at frames 227-229. In 1992 a study of an enhanced
version of the film commissioned by the American Bar Association
revealed that the right lapel of the Governor's jacket flapped
outward at frame 224. Medical examination of the Governor
immediately after the shooting revealed that he had been shot
through the right side of his chest, the bullet exiting his rib
cage and then striking his right wrist. The flap of the
Governor's lapel, accompanied by his stiffened posture at frame
224, followed by clear evidence of anguish, indicates impact of
the bullet at or about frame 224.
