The right time to sow in the fall can be fixed only with great difficulty, for so much depends upon the climatic conditions.  In fact the practice varies in accordance with differences in fall precipitation and early fall frosts.  Where numerous fall rains maintain the soil in a fairly moist condition and the temperature is not too low, the problem is comparatively simple.  In such districts, for latitudes represented by the dry-farm sections of the United States, a good time for fall planting is ordinarily from the first of September to the middle of October.  If sown much earlier in such districts, the growth is likely to be too rank and subject to dangerous injury by frosts, and as suggested by Farrell the very large development of the root system in the fall may cause, the following summer, a dangerously large growth of foliage; that is, the crop may run to straw at the expense of the grain.  If sown much later, the chances are that the crop will not possess sufficient vitality to withstand the cold of late fall and winter.  In localities where the late summer and the early fall are rainless, it is much more difficult to lay down a definite rule covering the time of fall sowing.  The dry-farmers in such places usually sow at any convenient time in the hope that an early rain will start the process of germination and growth.  In other cases planting is delayed until the arrival of the first fall rain.  This is an certain and usually unsatisfactory practice, since it often happens that the sowing is delayed until too late in the fall for the best results.

In districts of dry late summer and fall, the greatest danger in depending upon the fall rains for germination lies in the fact that the precipitation is often so small that it initiates germination without being sufficient to complete it.  This means that when the seed is well started in germination, the moisture gives out.  When another slight rain comes a little later, germination is again started and possibly again stopped.  In some seasons this may occur several times, to the permanent injury of the crop.  Dry-farmers try to provide against this danger by using an unusually large amount of seed, assuming that a certain amount will fail to come up because of the repeated partial germinations.  A number of investigators have demonstrated that a seed may start to germinate, then be dried, and again be started to germinate several times in succession without wholly destroying the vitality of the seed.

In these experiments wheat and other seeds were allowed to germinate and dry seven times in succession.  With each partial germination the percentage of total germination decreased until at the seventh germination only a few seeds of wheat, barley, and oats retained their power.  This, however, is practically the condition in dry-farm districts with rainless summers and falls, where fall seeding is practiced.  In such localities little dependence should be placed on the fall rains and greater reliance placed on a