Your sample is adequate. Maybe I was too narrow in my thinking. Fuse wire can fail under many different conditions and when I wrote earlier, I was thinking in very simplistic terms. I was simply thinking of an ideal situation with a perfect resistance wire brought somewhat slowly, maybe over 3 seconds or so, to the failure point. Too slow and the terminals and wires start warming and while the overall approach remains the same, it would get more and more difficult to know all the factors.
If you pulse the fuse wire, i.e. just flip the switch on with a much higher than normal current, other factors could be involved. There is a magnetic component as the current thru the fuse tends to mechanically flex the wire. Very rapid heating of the wire may cause failure at the connector due to crimping or other termination factors. I suspect that due to the mass manufacturing nature of the cartridge fuse, there is probably not a lot of care given to uniform termination of the fuse wire to the end cap of the cartridge, which is connected in various ways including a solder ball.
I suspect that this experiment would be more reliable if done with bulk fuse wire somewhat longer, maybe 2 to 3 inches with careful attention to the uniformity of the termination and a soft approach to failure rather than a pulse.
Some final thoughts. While the labs I'm associated with, often deal with pulse currents of 10s and 100s of thousands of amps, failures at even a 2 thousand amps can fling a burned cable end across the bench and damage lab equipment. Your test current was far, far less than that but still the effect is there. Another possible factor is that when you pulse the wire with a fairly high di/dt, the wire expands throughout it's length, very rapidly, while the highest temperature is still likely to be near the center, the greatest mechanical stress is at the terminating points. You may notice a fuse that is pulsed (short duration pulse) well beyond it's normal failure point but not long enough to cause it to fail often exhibits a permanent bow in the wire. This is not so much due to the, sag, melt scenario I mentioned earlier, but to the fact that the fuse wire rapidly grows in length but it can't push the ends out.
I can't think of a good reason why fuse manufacturers would want to force a fuse to fail necessarily in the center. Interruption is interruption as long as it's not an exotic application and in most cases they are not reusable. I rather suspect that by narrowing the wire at some point is a more accurate way to force the fuse to fail at a fairly precise temperature (current) while providing adequate mechanical support during normal moderately high current operation.
Maybe we will attract the attention of a someone from Ferraz Shawmut or Bussman who would probably have a far more educated answer than my theoretical musings.