Next, the statement is carelessly made that the recent success in dry-farming is due to the fact that we are now living in a cycle of wet years, but that as soon as the cycle of dry years strikes the country dry-farming will vanish as a dismal failure. Then, again, the theory is proposed that the climate is permanently changing toward wetness or dryness and the past has no meaning in reading the riddle of the future. It is doubtless true that no man may safely predict the weather for future generations; yet, so far as human knowledge goes, there is no perceptible average change in the climate from period to period within historical time; neither are there protracted dry periods followed by protracted wet periods. The fact is, dry and wet years alternate. A succession of somewhat wet years may alternate with a succession of somewhat dry years, but the average precipitation from decade to decade is very nearly the same. True, there will always be a dry year, that is, the driest year of a series of years, and this is the supposedly fearful and fateful year of drouth. The business of the dry-farmer is always to farm so as to be prepared for this driest year whenever it comes. If this be done, the farmer will always have a crop: in the wet years his crop will be large; in the driest year it will be sufficient to sustain him. So persistent is the half-expressed fear that this driest year makes it impossible to rely upon dry-farming as a permanent system of agriculture that a search has been made for reliable long records of the production of crops in arid and semiarid regions. Public statements have been made by many perfectly reliable men to the effect that crops have been produced in diverse sections over long periods of years, some as long as thirty-five or forty year's, without one failure having occurred. Most of these statements, however, have been general in their nature and not accompanied by the exact yields from year to year. Only three satisfactory records have been found in a somewhat careful search. Others no doubt exist.