Hitting bottom of an axe to seat the axe head I have been thinking about a certain step in hafting an axe which I can't seem to explain physically. As a reference, I've found a forestry manual which explains the process, in case my explanation is not clear.

Basically, after ensuring a close fit, you put the head onto the handle. Then, to fully seat the head, you hold up the axe (with the head down) and strike the bottom of the handle. This draws the head tight down onto the handle.
Why does this work? My initial thought was that it has something to do with Newton's Law of equal and opposite reaction, but I'm not really sure.
EDIT: The axe is being held in midair and the head is not resting against anything when the handle is hit. See, e.g., https://youtu.be/D8CJ2u8oDI4?t=214.
 A: Ok, I understand from a previous comment that the axe head is not resting against anything and I think I understand what is going on.
There is some very nice physics here involving inertia and shock waves.
When you strike the top of the handle, it sends a shock wave down the handle. That is, there is an abrupt motion in the wood, travelling at the speed of sound, slightly deforming the wood as it goes, in a kind of ripple. When this ripple arrives at the axe head, the head itself does not move so much because it has a higher density and therefore a higher inertia. Also, the speed of sound in metal being higher, the head tends to move more as a whole. So this, combined with its inertia, means than the metal head tends to stay in one place while the ripple runs down the wood. Thus the wood 'inches' its way further into its place in the close-fitting hole in the metal. This ripple effect does a much better job of bedding the handle than you could do merely by trying to move the handle as a whole and overcoming the friction, because the wood creeps forward one bit at a time: as the wave passes along, the local force between wood and metal is first reduced in the contractive part of the ripple, allowing the local wood to move forward a little, and then when the wood expands it simply grips the metal and does not move one way or the other relative to the metal. So the net result is to bed the handle.
I should add that the above is simply what I, as a physicist with a fairly long experience, suspect is what is going on. It is not something I have read about and I am sure there is somewhere a more thorough discussion. So I hope I am right; I think I have a good argument. As I have described it above, I have in mind mainly the last part of the process where the wood only moves a little relative to the metal. In the earlier part, when the wood moves through a larger distance, it is inertia that is the main consideration, just like in the party trick where you abruptly whisk away a table cloth and the dishes on the table stay where there are. The more abrupt the better.
Added remark
It occurred to me that there is another thing worth mentioning here, that makes this method preferable to resting the axe head on something, or supporting the handle on a work top and hitting the head. It is that by hitting the end of the handle, with the head just hanging, you are going to deliver the force more accurately at the join, because it travels along the handle in exactly the direction you want. If instead you strike the head then there is a danger it will be knocked slightly obliquely, introducing a random tilt with each blow, which is liable to deform the wood and thus loosen the fit.
