It should also be noted that this interview was carried out years after the event. Had Hemphill been standing at his office window when he was interviewed he would not have made the mistake of thinking the path might have been over or north of the Citgo service station. His line of sight from his office window to the impact point passes directly over the service station (see Fig. 3), so he was particularly well placed to judge that the path he clearly remembered, and asserted in all his statements, was to the south. He insisted that the path was straight, so could not have deviated round the Citgo service station. Again we have an interview that was not carried out on location, with no mention being made that this is another exception to their claim of reliability due to location. Hemphill was irritated by the pressure of the questioning and remarked “I saw what I saw. That is where it stands.”

It is important to note that several of the witnesses Ranke was quoting “over and over” were near the Arlington cemetery. It is not reasonable to assert they could accurately judge from that distance that the plane was a little north or a little south of the Citgo service station. In contrast it would be easy for them to judge whether the plane was steeply banked, but all these witnesses stated the plane flew “flat” over the Annex and then banked only slightly, or made no mention of an unusual steep bank angle.[32] Hemphill’s words “it didn’t turn right, it didn’t turn left” correspond with Boger’s phrase “it didn’t veer” and with Morin’s assertion that the plane was “heading directly towards the Pentagon”. All these observations contradict the NOC path, as will become apparent below where we discuss the angle of bank required.

The evidence to this point leads us to infer that CIT has misled the public in regard to the witness testimonies, having failed to present the evidence of Lagasse, Turcios, Brooks, Paik, Morin, Hemphill and Boger fairly. It is our purpose to add another dimension to this discussion.


Let us examine the dynamic feasibility of a NOC flight path.

Radar could not provide useful information close to the Pentagon as, by then, the plane would be too low, but radar tracks from four different facilities corroborate each other, leading directly toward the Pentagon. The track from the nearest radar facility, at Ronald Reagan National Airport, reaches to a point only about 6 seconds prior to impact. This is close enough to the Pentagon to indicate that all the witnesses who mentioned the distance from themselves to the plane underestimated the distance, as can be seen in the following image (Fig. 1).[33] Recall, for instance, that Deb Anlauf at the Sheraton, about 500 feet from the radar track,  describes the plane as right outside her window saying, “You felt like you could touch it; it was that close.” Perhaps we underestimate distances in situations like this because we are not used to seeing large planes flying so low and interpret the large size as indicating closeness. From a statistical point of view it is unfortunate that there are many witnesses who were north of the path and few who were south of the path, close enough to form a clear impression of the position of the track, hence it is not surprising that there should be some northerly bias in the reporting. The FDR data extends the radar data for about 6 more seconds (Fig. 1) and shows no deviation right to the Pentagon.

Figure 1

Figure 1: Four radar tracks approach the Navy Annex. The yellow line is the last 20 seconds of the FDR data, aligned with the trail of damage. It shows no deviation over this distance.

The testimony of the witnesses cited above is in reasonable conformity with the path defined by the radar data, the FDR data and the damage trail. Some witnesses said the plane was coming along highway 395; some said it was coming along Columbia Pike, which runs close to the south side of the Navy Annex, nearly parallel with the closest section of highway 395. To be consistent with these witnesses the plane must have passed south of the Sheraton Hotel, south of the A-1 Car Repair shop of Ed Paik, and near Terry Morin. Morin may have been between the wings of the Navy Annex as the plane flew over, but said that he “ran to the outside” from between the wings to a “position where I could see it.”[34] “As he starts to descend … he basically starts to disappear … the engines disappear, the bottom of the fuselage, the wings…” He followed the plane as it dipped down over a row of trees on its approach to the Pentagon until all he could see was the tail. He does not mention bank angle.

Taken at his word, Morin witnesses a direct approach to the Pentagon along the south path. We will, however, loosen that assumption to enable our discussion to proceed further and in our initial analysis will simply assume the plane does not veer north until it has passed Morin. Already this conflicts with Morin’s testimony, in that it cuts short the distance that it would be visible to him, as will become evident below.

To favor the NOC hypothesis as much as possible, we assume that the plane passed barely north of Citgo, in fact flying directly over the northernmost corner of the station. Finally we assume that, in order for the alleged “magic trick” with the smoke to work, the plane flew directly over the impact point on the west face of the Pentagon. For the sake of argument, we set aside the numerous observations of impact, the observations of the plane hitting light poles, a fence and a generator, and observations that it flew straight and descended very close to the ground.

For the plane to follow a path that conforms with the remaining constraints, it must deviate from its initial path. It must first turn left and then turn right. Turns for aircraft involve banking and heightened g-force, which is the apparent gravity induced by centripetal acceleration. We make the assumption that the radius of curvature, and hence the g-force, is equal for the left and right banks. This is the assumption most favorable to the NOC hypothesis. Furthermore, for a large plane to transition from a hard left bank to a hard right bank requires a roll maneuver, which takes some time. For simplicity we assume that the plane flies straight during the brief transition from left bank to right bank. Fighters are small and have their mass close to the longitudinal axis of the plane, so they can roll quickly. Large planes, with their outboard engines and heavy fuel tanks in the wings, have a large moment of inertia and would require several seconds to make the transition. However, for the sake of argument we will consider the implausibly short left-to-right roll times of one second, and one half second. Shortening the assumed roll time allows more space for the bank maneuvers, thus favoring the NOC hypothesis.

Aircraft speed

Speed is important as it is one of the two factors which determine angle of bank. Some of CIT’s witnesses estimate speeds of 350 to 400 mi/hr. These low estimates are uncalibrated guesses. There is good reason to believe that the testimony by the same and other witnesses that the plane was accelerating is more reliable, since it was based on the sound of the motors revving up.  The distinctive sound of the engines would be more reliably assessed than the speed of the plane itself as our ears are sensitive to pitch. It is understandable that visual estimates of speed would be low, given that all the observers perceived the plane to be closer than it really was. If an object moving across the field of view is farther than the viewer estimates, the reduced angular motion across the field of view, due to distance, will be misinterpreted as being due to lower speed. Several witnesses use words like “spooling up,” “full throttle,” and “powered descent” to describe what they heard.[35] We recall that Morin said he heard a “… steady high-pitched whine” indicating to him that “the throttles were steady and full.” The plane was also diving, so it had assistance from gravity in gaining speed. The FDR data indicates an average ground speed of 552 statute miles per hour (mi/hr) for the last 4 seconds and the final speed measured prior to impact was 556 mi/hr. The official estimate is 530 mi/hr, which is presumably based on the final speed shown in the original improperly decoded FDR file, 465 knots (535 mi/hr). The FDR file also shows that the engines were suddenly set to full power for the last half minute, during which time the plane accelerated rapidly and uniformly. Measurements of the radar positions, recorded every 4.7 seconds, shown in Fig. 1, provide independent confirmation of the speed and acceleration, as shown in Fig. 2. Distances were calculated using the Haversine method for great circle arc length, spherical earth approximation, from latitude and longitude.

Figure 2

Figure 2: Radar data from Ronald Reagan National Airport (the nearest source) confirm the acceleration shown in the FDR file, in conformity with witness reports of hearing the engines “spooling up”.

A trend line through this period gives a last 4.7 second interval averaging 520.2 mi/hr, accelerating at 6.39 mi/hr/s. Extrapolation of these figures to the next interval would give an average speed of 550.3 mi/hr, and after a further 2.35 seconds to the end of the interval, which would be very close to the moment of impact, the speed would be 565.3 mi/hr.

The last measured speed in the FDR file was 556 mi/hr. This is recorded in word 94. The impact is recorded in word 225, thus 131 words later. Each word is 1/256th of a second hence 0.51 seconds elapsed prior to impact. If the calculated acceleration was maintained during this period the final speed would be 559 mi/hr. The radar data thus lead to a final estimated speed which corresponds well with the FDR file.

As this plane, known to be aerodynamically efficient, was clean and diving, it could not possibly slow down significantly in those last few seconds as it passed the Navy Annex, even if, contrary to witness reports, it was throttled right back. Given the weight of evidence from the FDR file and radar, and the witness reports of the engine power rising, we reject the low speed estimates of 350 to 400 mph as flawed visual estimates without evidence. We will base our calculations on the official speed, 530 mi/hr, as a low estimate, and the FDR speed, averaged for the last four seconds, 552 mi/hr, as the more realistic estimate.