It is vain to attack statistics to-day. Every society publishes statistical sheets. Every society by publishing them shows that it recognises the value of statistics. The difficulty to-day is not that the societies do not publish statistics, but that the statistics which they publish are not related to any aim or purpose, and do not include factors or standards which enable us to measure progress. (ii) It may also cause surprise that we ask for estimates in some cases where exact information is not immediately accessible. It may be said that statistics are misleading, but estimates are hopelessly misleading: let us have correct figures or none. That attitude is easily understood, but under the circumstances it is vain. "Correct figures," that is, meticulously exact figures, are unattainable. An estimate is in nearly all matters of daily life and business the basis, and rightly the basis, of our action. It will be noticed that in that letter which we quoted above concerning the statistics of the candy trade in the United States of America, estimates had a place, and foreign missions involve matters about which "correct figures" are more difficult to obtain than the candy business. An estimate carefully made and understood, a deliberate statement expressed in round numbers, is not unscientific: it is only unscientific to mistake such figures for what they do not profess to be. When men object that the figures are not exact, if the figures do not profess to be exact, it is the objector who is unscientific, not the statistics. Nevertheless, we must acknowledge that the admission of estimates and round figures does open the door to serious error. Men will be tempted to mistake an estimate for a guess. An estimate is a statement for which reasons can be given, a guess is--a mere guess. The great safeguard against guesses, as against all slipshod statistical entries, is the assurance that the statements made will be used. At present missionary statistics are untrustworthy mainly because so few people use them, and consequently those who supply them do not feel the need of revising them carefully. Furthermore, it is important to bear in mind that the field for estimate in statistics of the kind proposed is limited; it only embraces figures for which exact totals are unobtainable, for instance, area, population, and figures of societies which refuse to give statistics, etc., and in every case precision in these statistics is not of vital importance.