Still have not found the article, but last evening I did some simple tests--certainly not the final word and not trying to control all variables, but: Some reports have it that the powder was poured over the ball in a cupped hand [leading to variable hand size problems]--however Horace Kephart in 1918 [Outing Mag] reported that a FLAT palm was used and Audubon's 1810 description just says that the ball was laid in the palm and covered. So I did the reverse experiment and placed a ball from three calibers [one at a time] of mine, a .440, .490 and .530 in both cupped hand and open flat palm and covered them with the measured fffg loads I use for hunting for each [65 gr for .45, 75 gr for .50 and 85 gr for .54]. In all cases in the cupped hand the amount was "too much", that is overcovered the ball---so, if I had "just covered" each of these the charge would have been lighter than my standard hunting load, though possibly adequate as a close range target load [need to check]. However, in the flat palm, the charges appeared to "just cover" the three ball sizes--that is, such a measure would work for me in those three calibers. Consistency and spillage are two variables that come to mind. The old time cooks did very often use a cupped hand to measure ingredients quite successfully and I expect after a time a muzzleloader would be able to "eyeball' a proper load. Whatever we think of this method, it has been reported in period writings as a standard way the old timers loaded. It has in more recent writings been relagated to a "first try" method to get close to a load which thereafter is charged with a horn, antler or brass charger. Speed loading methods in the heat of battle or the chase often used "less optimum" methods such as spitting a bare ball down the barrel on a guessed charge dumped in from the hand or horn. I am still looking for the article I first asked about as it is a summary of a more extensive examination of covering the ball....