I took a look at the article Joe referred to. On the surface it appears reasonably credible, and to have been written by a technically astute person. However, I do question several things:
1)The noise components in question are so low in level (on the order of 0.01% of the 120 volt 60 Hz component, corresponding to about 70 db down) that I would expect differences to be observed simply as a result of minor changes in the physical positioning of the probe (and especially its return lead) relative to the nearby chassis and/or transformer or other circuit components.
2)In that regard, some of the fuses, apparently the power amp mains fuses, were tested "in a Radio Shack 20 Amp in-line fuse holder with 12 gauge stranded wire leads." "Alligator clip leads were used" to connect to the holders of other fuses. Again, I would expect in both of these cases that the physical manipulation involved in changing fuse orientation would affect the results by introducing small changes in the physical positioning of the holder leads and/or the probe and/or its ground lead.
3)All of the foregoing increases the importance of doing what I talked about earlier, namely going back and forth between the two orientations several times, and verifying that the results are consistent. There is no indication that this was done.
4)I note that all of the measurements were confined to frequencies of less than 500 Hz. And (as one would hope) the amplitude of the 60 Hz component appears to be the same in all of the spectrum photos being compared. Which leads me to wonder what sort of magic enables the fuse, much less its orientation, to be able to distinguish between 60 Hz and other frequencies that are so close to 60 Hz. Aside from power regenerators (which generate a completely new AC waveform), I am unaware of how even a sophisticated and expensive power conditioner could be designed to affect frequencies which are so close to 60 Hz differently than they would affect 60 Hz, much less a fuse, much less the orientation of a fuse.
I am not trying to be argumentative either, and I appreciate that Joe brought this seemingly well done paper to the table, but those are my comments. Mapman makes good points as well, IMO. The article refers to "the direction of current (energy) flow." Current flow, which the fuse "knows" about, is back and forth. Energy transfer is from source to load, but as I indicated earlier in the thread I have yet to see a credible explanation of how a fuse would have any "knowledge" of the direction in which energy is being transferred.
Regards,
-- Al