Thanks Gordon for your helpful post and picture. That makes perfect sense. Can you describe the "heat treating" for those of us who have very little back ground in metallurgy?
I just used a rough and ready approach, with rather little control over anything, with a risk of the surface cracking to uselessness...
First, I knew that this particular piece of metal was heat treatable steel, i.e. sufficient carbon (or equivalent) content in the alloy... I bought it as a cheap off-cut from a specialist steel supplier. Probably cheap because of its odd shape.
I used oxy-acetylene to heat the curved surface to cherry red. Most of the rest was not hardened because it never got to cherry red.
I then waved stirred it around in a large bucket of water, to quickly cool it. The fast cooling from cherry red is what makes the affected metal very hard (and also brittle).
The size and shape, (and unhardened interior) meant that brittleness probably won't affect it during use.
However during the rapid cooling the surface shrinks more, and faster than the not-so-hot, and not-so-quickly-cooled interior, so I risked severe splits in it. I struck it lucky, with only slight fracture, and not in an area that mattered. I have used this for a few decades, with barely a mark on that curved area that I use.
As I said, this was a very rough approach that happened to work well this time. If I wanted it done better I would most certainly take it to a specialist heat treatment plant, 3 km away from here. I happen to know enough to know that heat treatment is actually a very complicated area. Such a plant would possibly use case hardening, such as that done with the gears in a gear box. There are a variety of ways to do this, not at my disposal in my workshop.