"When geologists and astronomers first recognized that lunar and terrestrial craters were produced by impacts, they surmised that much of the impacting body might be found still buried beneath the surface of the crater floor. (Much wasted effort was expended to locate a huge, buried nickel-iron meteorite believed to rest under the famous Barringer meteor crater near Winslow, Ariz.) Much later, however, scientists realized that at typical solar system velocities--several to tens of kilometers per second--any impacting body must be completely vaporized when it hits.
"At the moment an asteroid collides with a planet, there is an explosive release of the asteroid's huge kinetic energy. The energy is very abruptly deposited at what amounts to a single point in the planet's crust. This sudden, focused release resembles more than anything else the detonation of an extremely powerful bomb. As in the case of a bomb explosion, the shape of the resulting crater is round: ejecta is thrown equally in all directions regardless of the direction from which the bomb may have arrived.
"This behavior may seem at odds with our daily experience of throwing rocks into a sandbox or mud, because in those cases the shape and size of the 'crater' is dominated by the physical dimensions of the rigid impactor. In the case of astronomical impacts, though, the physical shape and direction of approach of the meteorite is insignificant compared with the tremendous kinetic energy that it carries.
"An exception to this rule occurs only if the impact occurs at an extremely shallow, grazing angle. If the angle of impact is quite close to horizontal, the bottom, middle and top parts of the impacting asteroid will strike the surface at separate points spread out along a line. In this case, instead of the energy being deposited at a point, it will be released in an elongated zone--as if our 'bomb' had the shape of a long rod.
"Hence, a crater will end up having an elongated or elliptical appearance only if the angle of impact is so shallow that different parts of the impactor strike the surface over a range of distances that is appreciable in comparison with the final size of the crater as a whole. Because the final crater may be as much as 100 times greater than the diameter of the impactor, this requires an impact at an angle of no more than a few degrees from horizontal. For this reason, the vast majority of impacts produce round or nearly round craters, just as is observed.