As I remember from second year engineering school, susceptibility of a material to fatigue failure could be predicted by a "notch sensitivity" test. A test where an ingot of the material to be tested has a standard notch machined out of it and is given a whack by a standard pendulum to see what it takes to break it off at the notch. A fatigue crack is most likely to start at a surface imperfection or a sudden change in section of a component e.g. injudicious scraping with the tuning file or chisel (especially across the reed!)or at the junction of the base of the reed where it's section is reduced after the rivetting flange. This suggests that 'tuning passes' would be best made along the reed rather than across it and a rotary grinding tool might give less surface discontinuity than the regular file or scraper does. Hot rodders have a fetish for highly polishing highly stressed components like engine conrods to try and ensure that there are no surface imperfections or stress raisers to breed fatigue cracks. They also compress the outer surface of such components by shot peening them (bombarding them with steel shot)but this does not look like a feasible strategy for harmonica reeds. Aongus Mac Cana
_______________________________________________ Harp-L is sponsored by SPAH, http://www.spah.org Harp-L@xxxxxxxxxx http://harp-l.org/mailman/listinfo/harp-l