During these 63 days, Millikan recorded in his notebooks data for about 100 separate drops. Of these, about 25 are obviously aborted during the run, and so cannot be counted as complete data sets. Of the remaining 75 or so, he chose 58 for publication. Millikan s standards for acceptability were exacting. If a drop was too small, it was excessively affected by Brownian motion, or at least by inaccuracy in Stokes s law for the viscous force of air (more about this later). If it was too large, it would fall too rapidly for accurate measurement. He also preferred to have a drop change its charge a number of times in the course of an observation, so that he could have changes in charge, as well as a total charge, which had to be integer multiples of a single unit of charge. None of this could be determined without actually taking and recording data on a candidate drop. Thus, it should not be surprising that Millikan chose to use the data on only 58 of the drops he observed during the period when he and his apparatus had reached near perfection. Furthermore, he had no special bias in choosing which drops to discard. A modern reanalysis of Millikan s raw data by Allan Franklin (see following page) reveals that his result for the unit of charge and for the limits of uncertainty in the result would barely have changed at all had he made use of all the data he had, rather than just the 58 drops he used.