The nature of the soil is another important condition regulating this loss. Some soils are very much opener and more porous than others; in such soils, of course, the loss by drainage will be greatest. We are apt at first sight, however, knowing the great solubility of nitrates, to overrate this source of loss. We have to remember that while nitrates are constantly being washed down to the lower layers of the soil, there is likewise an upward compensating movement of the soil-water constantly taking place. This is due to the evaporation of water from the surface of the soil, which induces an upward capillary movement of water from its lower to its higher layers.[85] This upward movement of water is very much increased, in the case of soil covered with vegetation, by the transpiration of the plants. The climate and the season of the year will affect the extent of this upward movement. Where there is a heavy rainfall it will be very much less than in dry climates. After a long period of drought the nitrates will be found to be concentrated in the top few inches of the soil; and in hot climates this sometimes takes place to such an extent that the surface of the soil has been actually covered with a saline crust, caused by the rapid evaporation of soil-water under the influence of a burning tropical sun. From this point of view it will be seen how very much less powerful a single shower of rain is--even although at the time it is heavy--in causing loss of nitrates by drainage, than a continuance of wet weather. In the former case, where the showers are separated by an interval of dry weather, the nitrates washed down into the lower layers of the soil are slowly brought up again by the capillary action caused by evaporation. _Amount of Loss by Drainage._ What the actual amount of loss is which takes place in this way it is wellnigh impossible to say. What it amounts to under certain definite circumstances has been discovered by actual experiment at Rothamsted. Taking the circumstances most favourable to extreme loss--viz., unmanured fallow land--the highest amount registered at Rothamsted for a year is 54.2 lb. per acre from soil 20 inches deep, while the smallest amount is 20.9 lb. In the former case, the drainage-water was equivalent to 21.66 inches, while in the latter, to 8.96 inches. The average for thirteen years on unmanured fallow soil has been 37.3 lb. (for 20 inches), 32.6 lb. (for 40 inches), 35.6 lb. (for 60 inches). The point of especial interest in this connection is that an annual loss of nitrogen, equal to over 2 cwt. of nitrate of soda, may take place from a comparatively poor arable soil lying fallow. The loss on cropped soils is of course very much less--in short, should amount to very little--especially in permanent pasture, where it is reduced to a minimum. Taking an average, Mr Warington is of opinion that the loss in England may be put at 8 lb. per annum per acre.[86] _Loss in Form of Free Nitrogen._