According to Wikipedia (because I'm too lazy to find my Machinery's Handbook), 10.9 grade hardware should have a hardness of Rockwell C33-39. It's fairly unscientific, but I'd see how easily a dull file bites into the metal. A sharp file will skate off of fully hardened tool and spring steels, will just barely bite into something tempered back to a cutting edge hardness Rockwell C 50-60ish, and will easily cut steel in the Rockwell C 30-something range. An old, dull file should mostly skate off of a bolt in that range, but will easily bite into, say, a class 5.8 bolt.

The grain of the steel where it broke holds clues as to hardness. If i really wanted to know, I would get three new bolts of the same size; one class 4.6, one class 8.8, and one class 10.9, then drill and tap a piece of steel, drive in the bolts, and torque until they break. Compare grain structure of the broken metal. Should be able to see a difference with the naked eye.

I have always thought the spark test to be too hard to read.