cause only thus can the water present travel to the roots of the plant. Lastly, a loose layer must be maintained on the surface, which, though dry itself, acts as a screen and a barrier to prevent loss of water from the effective soil below by any other channel than that of the plant. Granted these methods of cultivation, the new feature about "dry-farming," which has been introduced by settlers in the arid districts of Australia and North America, is the use of a year of bare fallow in which to accumulate a supply of water for the next year's or two years' crop. This raises the fundamental question of how much water is necessary for the growth of an ordinary crop. The first investigation that Lawes and Gilbert carried out at Rothamsted dealt with this very point; they grew the usual field crops in pots, protected the surface of the soil from evaporation so that all the loss of water proceeded through the plant, weighed the water that was supplied from time to time, and finally weighed the produce, expressing their results as a ratio between the dry matter produced and the water transpired by the plant. These experiments have been repeated under different climatic conditions by Hellriegel in Heidelberg, by Wollny in Vienna, by King and others in America, Now the two processes in the plant, carbon assimilation and transpiration, are not causally connected, though as both are carried out in the leaf and have some factors in common they are found to show some constancy in their relative magnitudes. Lawes and Gilbert obtained a ratio of about 300 lbs. of water transpired for each pound of dry matter harvested, but the other investigators under more arid conditions found much higher figures, up to 500 and even 700 to 1. Now, a crop yielding 20 bushels of wheat per acre will contain about a ton of dry matter per acre, so that, taking the high ratio of 500 to 1, no more than 500 tons of water per acre or 5 inches of rain will have been consumed in the production of this crop. It is, of course, impossible to ensure that all the rain falling within a year shall be saved for the crop, much must evaporate before it reaches the subsoil where it can be stored, and only when the crop is in full possession of the land can we expect that all the water leaving the soil shall go through the crop. What proportion the waste bears to that which is utilized will depend not only on the degree of cultivation but upon the season at which the fall occurs; summer showers, for example, that do not penetrate more than a few inches below the surface will be dissipated without any useful effect. When the climatic conditions result in precipitation during the winter, the water will be in the main available for crop-production; and it has been found by experience that cereals can be profitably grown with as small a rainfall as 12 inches. The necessary cultural operations consist in producing such a rough surface as will ensure the water getting into the subsoil, hence autumn ploughing is desirable. Where the precipitation is largely in the form of snow, a broken surface also helps both
