Of all possible IF frequencies, why did so many consumer radio and amateur HF radio manufacturers (and filter component vendors, etc.) end up converging on 455 kHz as the IF frequency, from the vacuum tube era, and onward? Is there something special about the number 455000?
Piecing together information from a number of sources on the web, I believe this is the correct combination of circumstances:
Early amplifiers were not stable much above 500 kHz, so the IF had to be less than that frequency.
Also, the broadcast band itself begins around 530 kHz (in the US), and you don't want any chance of having a station on or near the IF.
On the other hand, making the IF as high as possible improves image rejection.
In addition to the desired station mixing with the local oscillator to produce an IF signal, two strong stations whose frequencies happen to be separated by the IF will also produce an interfering signal.
In the US, stations are spaced by multiples of 10 kHz.
Elsewhere, stations are spaced by multiples of 9 kHz.
Therefore, 455 kHz was chosen because it is not possible to have stations separated by that amount in either system: 455/10 = 45.5 and 455/9 = 50.555. This is the highest frequency less than 500 kHz for which this is true.
Specifically, we're looking for frequencies that are as close as possible to odd multiples of both 10/2 = 5 kHz and 9/2 = 4.5 kHz. This puts the beat notes produced by this particular phenomenon as high as possible in the audio output of the mixer, where they can be filtered out. As it happens, 445 kHz would have worked equally well. The nearest frequencies to these two that would have worked just as well are +/- 90 kHz away -- 355/365 kHz or 535/545 kHz.