But in all these varying conditions of time and space there is one fact which must control our conclusions in considering most of the lands of the Arid Region, namely: any district of country which we may be studying is liable for many seasons in a long series to be without rainfall, when the whole supply must be received from irrigation. Safety in agricultural operations will be secured by neglecting the rainfall and considering only the supply of water to be furnished by artificial methods; the less favorable seasons must be considered; in the more favorable there will be a surplus. In general, this statement applies throughout the Arid Region, but there are some limited localities where a small amount of rainfall in the season of growing crops seems to be constant from year to year. In such districts irrigation will only be used to supply deficiencies. The complicating conditions arising from soil and subsoil are many. Experience has already shown that there are occasional conditions of soil and subsoil so favorable that the water may be supplied before the growing season, and the subsoil will hold it for weeks, or even months, and gradually yield the moisture to the overlying soil by slow upward percolation or capillary attraction during the season when growing crops require its fertilizing effect. When such conditions of soil and subsoil obtain, the construction of reservoirs is unnecessary, and the whole annual supply of the streams may be utilized. On the other hand, there are extremely pervious soils underlaid by sands and gravels, which speedily carry away the water by a natural under drainage. Here a maximum supply by irrigation is necessary, as the soils must be kept moist by frequent flowing. Under such conditions the amount of water to be supplied is many fold greater than under the conditions previously mentioned, and between these extremes almost infinite variety prevails. Practical agriculture by irrigation has also demonstrated the fact that the wants of different crops are exceedingly variable, some requiring many fold the amount of others. This is due in part to the length of time necessary to the maturing of the crops, in part to the amount of constant moisture necessary to their successful growth. But by excluding the variability due to rainfall, and considering only that due to differences of soils and crops, and by taking advantage of a wide experience, a general average may be obtained of sufficient accuracy for the purposes here in view.