Witness accounts

Penny Elgas

.../... In that split second, my brain flooded with adrenaline and I watched everything play out in ultra slow motion .../... At the point where the fuselage hit the wall, it seemed to simply melt into the building. I saw a smoke ring surround the fuselage as it made contact with the wall. It appeared as a smoke ring that encircled the fuselage at the point of contact and it seemed to be several feet thick. I later realized that it was probably the rubble of churning bits of the plane and concrete. The churning smoke ring started at the top of the fuselage and simultaneously wrapped down both the right and left sides of the fuselage to the underside, where the coiling rings crossed over each other and then coiled back up to the top. Then it started over again -- only this next time, I also saw fire, glowing fire in the smoke ring. At that point, the wings disappeared into the Pentagon. And then I saw an explosion and watched the tail of the plane slip into the building.

Tom Hovis

It was as if he leveled out at the last minute and put it square into the building. The wings came off as if it went through an arch way leaving a hole in the side of the building it seems a little larger than the wide body of the aircraft. The entry point was so clean that the roof (shown in news photo) fell in on the wreckage.

Frank Probst

The plane's right wing went through a generator trailer "like butter. The starboard engine hit a low cement wall and blew apart.

Tim Timmerman

Tim Timmerman : Pilot. I was looking out the window; I live on the 16th floor, overlooking the Pentagon, in a corner apartment, so I have quite a panorama. .../... I saw it hit right in front of -- it didn't appear to crash into the building; most of the energy was dissipated in hitting the ground, but I saw the nose break up, I saw the wings fly forward, and then the conflagration engulfed everything in flames. It was horrible. It was a Boeing 757, American Airlines, no question. It was so close to me it was like looking out my window and looking at a helicopter. It was just right there.

An attempt of crash explanation

A lot of things have been written on this subject, including on the previous versions of this web site. A geometrical analysis of the destructions caused by the plane (lamp poles, damage before impact, damage on the front) makes it necessary to plot sketches of the scene. This gives an engineer's view of the situation which avoids to make mistakes and/or allows to get rid of some hypotheses on what stroke the Pentagon on september 11 2001. The first diagram shows the scene before the attack, with the items already listed upper on a previous page. All the coordinates of items have been seized on a satellite photo, then analytically transformed to coordinates in the diagram. This reconstitution has been done on a scale double of the scale of the images presented below, i.e. one pixel for 10 cm (3 pixels / foot). The errors resulting of the overall process are around 0.3 m (1 ft), with a maximum due to the fuzzy quality of the original image of 0.5 m (1 ft 7").

The "737" hypothesis

The following diagram shows the arrival of a Boeing 737-200.

The plane is figured, at right scale, on the path where it is after having struck the lamp poles. This axis is confirmed by the damage inside the building. The purple central axis impacts building between pillars 13 and 14. The two purple dotted lines correspond to the damage made to the wire netting fence on starboard and ventilation structure low wall on port. They are supposed to be the paths of the two engines. It is obvious on this diagram that the two engines of a 737-200 are too close from the centerline to account for these damage. The yellow line touching the building's front is the extension of the damage on the building, from pillar 8 to pillar 20. Two dotted lines have been drawn from it's extremities, parallel to the trajectory. It seems that, if the port wing could be supposed to have damaged the front on this extension, the starboard cannot. But this is not an evidence : admitting that the central axis could be shifted a little, a 737 could be responsible of the damage seen on the Pentagon's front. In a previous version of this web site, I even wrote that the damage fitted exactly the size of a 737, and I was not very far from the truth if considering that the plane kept it's structural integrity when hitting the building. It allowed me to state the hypothese of the "junked 737" attacking plane. It is, indeed, reasonning at the limit, and excludes smaller planes such as a commuter plane, a fighter like a F 16 or - still smaller - a cruise missile.

The "757" hypothesis

Now the following diagram shows the arrival of a Boeing 757-200, which is officially the hijacked plane of flight 77.

The plane is figured, at right scale, on the same path as for the 737. The purple central axis and supposed paths of the two engines are the same. It is obvious on this diagram that the two engines of a 757-200 are just at the right spacing from the centerline to account for the damage on the generator and on the ventilation structure. The yellow lines are also the same as previously. It is clear that, if the starboard wing can approximately be supposed to have damage the front on this extension, the port wing should have damaged on a much wider zone. In other words, a 757 is too large to be responsible of the damage seen on the Pentagon's front if we assume that the plane didn't break into pieces before hitting the building.

The starboard engine has damaged and moved the generator (not shown). The nose impacts on the building. The port engine is destroying some of the parts above ground of the ventilation structure. A geometric construction is possible at this precise time assuming the plane as a solid object. But from this time, the extreme deceleration applyed to the plane's body implies that it must be considered as a set of soft materials. Witnesses account for the fact that "the wings went forward" when the plane hit the front of the Pentagon. This is normal, on a structural analysis. The deceleration applied to the body of the plane when the nose hits the wall is tremendous. The wings are not build to support such efforts and, pushed by their own inertia, "want" to keep their speed : containing the landing gears, engines, fuel tanks, the wings are indeed heavy parts of a plane. They must have taken their autonomy to bend forward along the body of the plane, acting as whips. This "whip effect" had two consequences :

This made the destructions important on the wall of the first floor, on port side, (wall ripped away) due to the shock energy, and did also that, if measured from the assumed impact point, the damage on the port side are shorter than on the starboard side.

The roots of the wings stayed probably attached to the plane, at least because the main structural beam, located approximately at the first third on the wing depth, was deformed and not sheared. We can consider that the translation energy of the wing was transformed into a movement of rotation around a pivot which can be located approximately at the junction of the wing with the fuselage, at the first third of wing's depth. If we "detach" graphically the wings from the body of the plane, and plot on successive sketches what the movement could have been, we get :

The body is shifted, the nose towards the port side, by the shock against the front of the building. The wings, taken forward by their inertia, fold along the axis of the body. The port engine impacts the low cement blocks which had been put there to protect the ventilation structure. The nose of the plane is depicted entering the building, as if it was strong enough to pierce the external wall upon impact. This is naturally false, the front of the plane acting much probably as an accordeon in this phase. This graphical presentation doesn't show it, just because I'm lazy when using graphics softwares and don't have sophisticated tools to mix bitmap / vector graphics in the way I would like to...

The body is shifted a little more. The wings keep on folding along the body. The two engines are near to touch the belly of the plane. The starboard wing is drawn according to the hypothesis that it broke against the building. The wing tip has probably bounced against the building and, due to this shock, did not fold as much as the port wing.

The process is going on. The starboard wing is partly inside the building, partly outside. Two effects must occur, not figured on the sketches :

The two engines are inside the belly of the plane. They will join their masses to the plane's body to create the important damage at the "entry hole" in the front of the building, around pillar 14. The port wing tip has now taken a very important speed (supersonic) by "whip" effect. It will go through the contractor's trailor located on it's trajectory and explode it. The other trailor is destroyed by the port engine and part of wing nearer from body. The wing tip is going to raze pillar 8 (or even touch it a little) before entering the building.

Both engines are inside the building. The port wing has hit the front and, due to it's speed, destroyed the pillars and walls. The whole plane will now enter the building through the holes done mostly by the two engines and the port wing. It is possible that the tip of the starboard wing, having bounced on the building, could collide with the tail of the plane, for example right on the vertical tail fin.

This process is coherent, only under an hypothesis : the damage left at second floor level by the tip of the starboard wing must be considered in the maximum extent. According the the upper sketches, these damage should extend up to pillar 22. As a matter of fact, the impact point of the wing tip would be just after this pillar.


The following interactive diagram summarizes the sketches shown above. Click to show plane's progression. There is approximately 15 ms beetween two of the sketches here above. The fraction of the crash shown on these diagrams took place in a time approximately equal to 250 ms (a quarter of a second). It was not possible for humans to survive this brutal deceleration.

The bent columns on the front

On the front line of the building, on line "AA", equally spaced columns support the building's front. The distance beetween them is approximately 3 m. On the impact zone, on the first floor, the following damage were created to the columns :

From the clean - not impacted - state of column 8, it is clear that the port wing entered the building on the right of this column. The odd state of column 18 will be examined in the next page. It is however possible to draw conclusions from the state of columns 9AA, 16 AA and 17 AA. The following pictures show these columns.


The sketch below shows the trajectory of the plane (plain purple line), the resulting speed direction around columns 16-17 and column 9 (purple arrows), and the real bending direction of these columns (green arrows).

The difference beetween the expected columns bending, i.e. parallel to the path of the plane, and the real columns bending, is the key to understand how the plane's wings folded during the crash. The two sketches below show the vector speed composition which allows to estimate the direction of impact on these columns. Three colors have been used :

On the starboard wing (left on the sketches), the part which impacts pillars 16 and 17 is the engine. It's probable that this engine was itself colliding with the ground. It disconnected the foot of columns 16 and 17 from the slab, leaving them hanging from the second floor slab bent towards north. On the port wing (right on the sketches), pillars 10 to 12 are impacted by the wing and totally broken. Pillar 9 is impacted by a thiner part of the wing and resists, though bending towards the direction of impact, i.e. towards East (green arrow). Note the greater size of rotation-induced speed on pillar 9 (whip effect). It would not be surprising that beetween pillars 8 and 9, the wing tip speed was supersonic, which would explain the totally "exploded" state of the wall there.


The crash reconstitution here above is coherent as well with the witness accounts as with the laws of physics. The traces on the building, including the "apparently odd" bending direction of columns 9, 16 and 17 or "apparently too small hole" account for the "wings folding forward" effect that was described by some witnesses. The hypothesis which fits the best to the whole damage done by the attacking plane is that this plane had the size of a Boeing 757. It is logical and straightforward to conjecture that it was B 757 number 644 AA of American Airlines, fleet number 5 BP, which had taken off from Washington Dulles airport one hour before with sixty persons on board.