Editor: What Scott means, I think, is that individual grains of powder are not all exactly the same in weight or shape. The predicted properties of the powder are a function of the aggregated mass–with some kernels slightly heavier, some slightly lighter. If you tried to use a lab scale to “hand-pick” virtually identical kernels you would then end up with an aggregate that would be different than what the manufacturer delivered in the powder jug. The collection of hand-sorted kernels could have greater or lesser density or energy. This, in turn, could perform very differently than the powder “as delivered” by the manufacturer.

The conclusion is that, even if you are tempted, you should NOT use an ultra-precise scale to hand-sort kernels, trying to select only those that are most uniform.

Good point Scott has raised here — that this super-precise weighing capability might tempt someone into trying a procedure (weight-sorting individual kernels), that could cause problems.