the possibilities of evaporation within Signal Service shelters over the whole country for an average wind velocity.

Daily evaporation=30[ 1.96 pw+43.9 (Pw—Ps)]

His results in this respect are platted on chart No. VI of the Monthly Weather Review, September, 1888, and show that the total annual depth of evaporation has its maximum of over 90 inches in southern Arizona, California, and New Mexico, whence it diminishes to a minimum of 20 inches annually in the northwest corner of the State of Washington and thence eastward to Maine. These figures, like his formula, take no account of the wind, because within. the Signal Service shelters the wind is reduced to a velocity far less than that in the open air. These figures, therefore, represent the evaporation in open air only when there is no wind above some small limit—say 6 miles per hour but may be adapted to strong winds by the use of the figures given in the first paragraph of this section.

CULTIVATION DIMINISHES SURFACE-SOIL EVAPORATION.

The general effect of cultivation is to pulverize the upper soil; this protects the capillary roots from surface exposure, it breaks up the capillary outlets of the moisture in the soil, checks the natural evaporation that goes on at the surface, and thus preserves the water within the soil for the use of the plants. Dr. E. L. Sturtevant's observations show that the extent to which the water is thus conserved by cultivation during the months from May to September, 1885, at Geneva, N. Y., may be thus expressed: With a rainfall of 14.42 inches the cultivated soil evaporated 1.4 inches less than the uncultivated naked soil and 2.25 inches less than the soil covered with sod. In other words, the efficiency of the soil to retain useful water is increased by cultivation to an extent equivalent to 10 per cent of the rainfall. If the capillary connections between the soil in the neighborhood of the roots and the supply of moisture lower down be broken no supply of moisture can come up from below, but if the soil be well rolled the compacting will aid the capillary attraction and the plants will secure moisture from below. Again, when weeds are allowed to grow freely the injury to the crops is not due to robbing the soil of nutrition nor to their shading the ground, but principally to their robbing the soil of its moisture. Those who can with impunity allow weeds to grow must have soils containing an excessive moisture, which they thus get rid of, while those who have a comparatively dry soil must destroy the weeds in order to reserve moisture for the use of their crops. (Agr. Sci., Vol. I, p. 216.)

PERCOLATION.

The permeation of water through soils of different qualities has been studied by Welitschkowsky (Wollny, 1888, X, p. 203.) He maintained a layer of water at a constant height above the material through which it permeated; therefore the pressure forcing the water through was constant. He found that the quantity of flow increased at first rapidly, then slowly for several days, depending on the thickness of the stratum of soil and the pressure of the water, until the permeation reached the maximum; then the rate of flow diminished slightly for a day or two until it became constant. He found that the quantity of water delivered in a unit of time has no simple relation to the pressure forcing it through the soil or to the thickness of the layer of soil through which it flows, but the relation is more nearly expressed as follows: If the pressure be increased by regular additions the flow of water increases in an arithmetical progression such that the quantity equals (A) plus a constant factor (D) times the pressure (P) less unity; A+D (P-1). The numerical values of these terms can be deduced from his extensive tables of experiments, of which the following table is an abstract:

a The capacity for water is expressed as a percentage of the weight of the dry soil.

The general laws of the flow of waters through soils of different natures have been elaborately investigated by Milton Whitney in a series of papers published in Agricultural Science, Volume IV, to which the reader must refer for the details.

The percolation of water through the soil, whether it goes downward as drainage or upward to be evaporated from the surface, depends not merely upon the degree of comminution of the soil and its compactness, but also, among other things, to a slight extent, upon the barometric pressure of the atmosphere, so that a falling barometer is, according to E. S. Goff, generally accompanied by a corresponding increase in the rate of drainage or of percolation downward. (Agr. Sci., Vol. I, p. 173.)

AVAILABLE MOISTURE.

In his investigations as to the relation of atmospheric precipitation, especially rainfall, to the plants and the soil, Wollny shows that the percentage of water in the layer of soil containing growing plants increases from above downward as soon as the downward movement of the rain water in the soil ceases, but that the percentage increases from below upward while the rain is falling and so long as the water continues to be penetrating downward. The frequency of rainfall is of even greater importance than the quantity. Slight rainfalls that only wet the soil to the depth of a few millimeters do but little good to the vegetation, because the greater part of the water is quickly evaporated back again into the atmosphere. If it should rain daily 2 millimeters during the three summer months, then, even with this abundant precipitation the plants might die for want of water, whereas if this total of 180 millimeters were uniformly divided into ten or twelve rains during the three summer months it would be considered a remarkably favorable growing season, since under these conditions the earth would be wet down to a considerable depth and the water thus stored up is protected from evaporation. Therefore, for equal quantities of rain its value for agriculture increases as the number of rainy days diminishes, and diminishes as the number of rainy days increases, at least up to a limit that varies with the character of the soil.

In order to attain precise ideas on this subject, Haberlandt set out a series of glass tubes full of dry earth; each received at the beginning a certain quantity of water, and by weighing these from day to day he determined the loss due to evaporation. These losses are given in the following table, in percentages of the original quantity of water, which latter may be considered as a rainfall whose depth is given at the top of the column:

These experiments give us some idea as to what percentage of the rainfall remains in the soil for the use of the plant in the case of large and small rains, but do not quite answer the question how one and the same quantity of rain is utilized in moistening the earth when it is distributed through a larger or smaller number of rainy days.

On this latter question Wollny has made the following experiment: A quantity of water corresponding to a rainfall of 60 millimeters was communicated to an experimental tub, No. 1, all at once, while in tub No. 2, 30 millimeters were given the first time and the remaining 30 after three days; in the third tub 20 millimeters were given at first and 20 millimeters every other day thereafter, and, finally, in the fourth tub, 10 millimeters were given every day, so that in six days all had received the same quantity of water. These experiments were repeated for different kinds of soil and the results show that in all cases the quantity of water lost by evaporation is larger the more frequently the water was communicated or the greater the number of rainy days. A fine illustration of the truth of this principle as applied to practice is narrated by Haberlandt, who found that in 1874 the farmers at Postelberg got much better crops than those at Lobositz, which could only be attributed to the fact that during that year Postelberg had received 246 millimeters of rainfall in forty days, or an average of 6, whereas Lobositz had received 309 millimeters in seventy-seven days, an average of 4, so that the usefulness of the greater quantity of rain in Lobositz did not equal that of the smaller quantity at Postelberg.

Wollny shows that since the period of the heaviest rainfall occurs throughout central Europe at the time of the largest evaporation from the soil we must conclude that for the naked earth the wetting of the soil during the warmer season of the year is controlled much more largely by the rainfall than by the evaporation depending on the temperature. His observations with the lysimeter show that the precipitation is principally concerned in the moistening of the naked soil during the warmer season, while the influence of the temperature and the resulting evaporation nearly disappears and is only observable in periods that are deficient in rain. In most cases the vegetation is injured when the atmospheric precipitation during the coldest season of the year is insufficient. The precipitation at this time of the year is therefore quite as important for the success of the harvest as that which falls during the period of vegetation. (Wollny's Forschungen, Vol. XIV, pp. 138–161.)

A. Seignette has shown that the law of levels propounded by Royer is confirmed. This law states that for given plants and for other uniform conditions the reserve nutriment in the earth is always found at a constant distance below the surface; thus the bulbs of

a plant under given conditions are found at a given level, and if we change these conditions as to moisture, temperature, etc., we shall change the distance from the surface down to this level. (Wollny's Forschungen, Vol. XIV, p. 132.)

TRANSPIRATION.

The quantity of water transpired by trees and plants depends upon the amount of water at their disposal, as well as on the temperature and dryness of the air, the velocity of the wind, the intensity of sunlight, the stage of development of the plant, the amount of its foliage, and the nature of its leaf. The following are some of the results of measurements at European experiment stations. (See Fernow, Report, 1889, p. 314.)

F. B. Hoehner found that the transpiration per day per 100 grams of dry weight of leaves is for conifers 4.778 to 4.990 grams, but for deciduous trees about ten times as much, 44.472 to 49.553. During the whole period of vegetation a unit weight of dry leaves corresponded to a total weight of evaporated water, as shown by the following table, for three different years.

Transpiration of water corresponding to growth of unit weight of dry leaves.

The variability of transpiration is shown by the action of a birch in the open air, which transpired on a hot summer day from 700 to 900 pounds, while on other days it probably transpired not more than 18 to 20 pounds. A beech about 60 years old had 35,000 leaves, whose total dry weight was 9.86 pounds; hence its transpiration, at the rate of 400 pounds of water per pound of leaves, would be 22 pounds daily. An acre containing 500 trees would, during the total period of vegetation, transpire nearly 2,000,000 pounds of water, or about 50 pounds to the square foot.

A younger beech, thirty-five years old, with 3,000 leaves and a dry weight of 0.79 pounds, would, under the same conditions, transpire 470 pounds per pound or 2 pounds per day from June to November. An acre containing 1,600 such trees would transpire about 600,000 pounds per acre or 15 pounds to the square foot from June to November.

Of the entire mass of wood and foliage on an acre of forest from 56 to 60 per cent of the weight is water and 44 to 40 per cent dry sub