The amounts for 1736, three of 1743, and those of 1760, 1761, and 1762 (Nos. 9, 10, 11, 12, 15, 16, and 17), are all extracted from De Guignes, who took them from the _Mémoires sur les Chinois_. The last, that of 1762, is given in detail in the table. The discrepancy of sixty millions between that given by Amiot for 1760, and that by Dr. Morrison for the same year, is owing, there can be little doubt, to foreigners, and not to an error of the Chinese. The work from which Dr. Morrison extracted his estimate for that year was published in 1790, but the census was taken between 1760 and 1765. The same work contains the census of 1711 (No. 8), quoted by him, and there is good cause for believing that Amiot’s or Grosier’s estimate of 157,343,975 for 1743, is the very same census, he having multiplied the number 28,605,716 by five, supposing them to have been families and not individuals. The three ascribed to the year 1743, are probably all derived from the same native authorities by different individuals. The three dated in 1760, 1761, and 1762, are harmonious with each other; but if they are taken, those of 1753 and 1760, extracted from the _Yih Tung Chí_ by Dr. Morrison, must be rejected, which are far more reasonable, and correspond better with the preceding one of 1711. It may be remarked, that by reckoning five persons to a family in calculating the census of 1753, as Amiot does for 1743, the population would be 189,223,820 instead of 103,050,060, as given in the table. This explains the apparent decrease of fifty millions. All the discrepancies between these various tables and censuses must not be charged upon the Chinese, since it is by no means easy to ascertain their modes of taking the census and their use of terms. In the tables, for example, they employ the phrase _jin-ting_, for a male over 15 years of age, as the integer; this has, then, to be multiplied by some factor of increase to get at the total population; and this last figure must be obtained elsewhere. It must not be overlooked that the object in taking a census being to calculate the probable revenue by enumerating the taxable persons, the margin of error and deficiency depends on the peace of the state at the time, and not chiefly on the estimate of five or more to a household. The amount for 1736 corresponds sufficiently closely with that for 1743; and reckoning the same number of persons in a family in 1753, that tallies well enough with those for 1760, 1761, and 1762, the whole showing a gradual increase for twenty-five years. But all of them, except that of 1753, are probably rated too high. That for 1762 (No. 17), has been justly considered as one of the most authentic.