Thanks for the "Reader's Digest" version, breakeyp!
I'm thinking there's out of spec. materials dramas here, as well as heat treating issues. The cocking pieces above seem to have failed in two cases via slow crack propagation ("fatigue"), whilst the other went all at once. As it occurred in an area w/ rapid section change, it's possible heat treatment was a contributing factor, but I suspect a bit of "notch sensitivity" as well. Fairly roughly machined, most are. The Enfield example (Early type cocking piece) shows the classic crack propagation features the best (I think...they're going to be re-photoed in a few days, so maybe we'll have a better illustration.)
Yah, I get that! I still don't know of one "bible" for this subject. Most of my best references seem to be small papers and studies. ASM and ASTM affiliated, the majority; old AWS magazines as well.