=45. Computation.= The most direct method of expressing the water-content is by per cents figured upon the moist soil as a basis. The ideal way would be to determine the actual amount of water per unit volume, but as this would necessitate weighing one unit volume at least in every habitat studied, as a preliminary step, it is not practicable. The actual process of computation is extremely simple. The weight of the dried sample, _w_^1, is subtracted from the weight of the original sample, _w_, and the weight of the can, _w_^2, is likewise subtracted from _w_. The first result is then divided by the second, giving the per cent of water, or the physical water-content. The formula is: (_w_ − _w_^1)/(_w_ − _w_^2) = _W_. The result is expressed preferably in grams per hundred grams of moist soil; thus ²⁰⁄₁₀₀, from which the per cent of water-content may readily be figured on the basis of dry or moist soil. =46. Time and location of readings.= Owing to the daily change in the amount of soil water due to evaporation, gravity, and rainfall, an isolated determination of water-content has very little value. It is a primary requisite that a basis for comparison be established by making (1) a series of readings in the same place, (2) a series at practically the same time in a number of different places or habitats, or (3) by combining the two methods, and following the daily changes of a series of stations throughout an entire season, or at least for a period sufficient to determine the approximate maximum and minimum. The last procedure can hardly be carried out except at a base station, but here it is practically indispensable. It has been followed both at Lincoln and at Minnehaha, resulting in a basal series for each place that is of the greatest importance. When such a base already exists, or, better, while it is being established, scattered readings may be used somewhat profitably. As a practical working rule, however, it is most convenient and satisfactory to make all determinations consecutively, i. e., in a series of stations or of successive days. Under ordinary conditions, the time of day at which a particular sample is taken is of little importance, as the variation during a day is usually slight. This does not hold for exposed wet soils, and especially for soils which have just been wetted by rains. In all comparative series, however, the samples should be taken at the same hour whenever possible. This is particularly necessary when it is desired to ascertain the daily decrease of water-content in the same spot. In the case of a series of stations, these should be read always in the same order, at the same time of day, and as rapidly as possible. When a daily station series is being run, i. e., a series by days and stations both, the daily reading for each place should fall at the same time. While there are certain advantages in making readings either early or late in the day, they may be made at any