I don't claim to know as much about bullets as anyone here. But, I do know about testing, and without knowing the details of the above testing, it is hard to know what this chart means.
E.g., how many bullets of each type were in each sample? How many lots? What is the standard deviation in 10 runs on the same bullet? Between 10 bullets in the same lot? SD between lots?
I'm guessing that what we are seeing is the limitations of accuracy in testing, either in the equipment, procedure, or operator. It makes no sense, for example, that some GMX bullets vary so greatly in purity. Are they buying copper a pound at a time?
The other thing that is suspect is reporting accuracy to 5 decimals. Unless you know the variation between the same sample is within +/- 0.00001 you can't report that degree of accuracy. Just because the machine reads it doesn't mean it's so. Significant figures, and stuff like that.
That's why Steve's suppliers only report to 2 decimals.