134 THE IRRIGA TION A GE. 



roots have only to advance so far as is needful to allow capillarity to 

 bring them the water they need from the store which the soil has re- 

 tained. With these physical principles and conditions to throw light 

 upon the problem and with the other fact that 6 inches of water, 

 when the crop can have it to use to the best advantage, is enough to 

 produce 20 bushels of wheat to the acre, we can see its outlines with 

 sufficient clearness to feel sure that more study in the field would give 

 us its full solution. As the matter now stands, the case is sufficiently 

 clear that we may not conclude, because 9 to 12 inches of rain in Cali- 

 fornia has produced abundant crops of wheat that a similar rainfall in 

 the sub-humid belt ought to produce like results. It should be suffici- 

 ently evident also, that even with the best modes of tillage we can 

 hope to adopt, there will still be much more water required per pound 

 of dry matter produced all through the sub- humid region, than is de- 

 manded under the conditions of the lower San Joa^uin valley. 



Let me address myself specifically to the question To what extent 

 may tillage conserve soil moisture? We have succeeded the present 

 year in carrying out a field experiment under normal field conditions 

 which appears to demonstrate that a piece of uncultivated fallow 

 ground lost by evaporation at the surface, 7.96 inches of water, or 

 63.49 per cent more than adjacent soil cultivated 3 inches deep once 

 per week, enough for 50 bushels of oats per acre. This saving appears 

 to be very large and may not be correct but it is supported by labora- 

 tory trials published in the Annual Report for 1898 that a 3 inch mulch 

 on a similar soil diminished the evaporation 63.13 per cent. 



In another series of experiments which will appear in our next 

 Annual Report, just leaving the press, we have measured the loss of 

 water, by surface evaporation, from four columns of soil 10 feet long 

 and of two types, one a sandy loam and the other a clay loam, and 

 have determined the final distribution of moisture in them in 6 inch 

 sections as it was found after a period of 314 days. On one column in 

 each set was developed a soil mulch 3 inches deep, while the surface 

 in the other was left unstirred. The columns stood where they were 

 continuously dry in front of a ventilating stack which maintained a 

 continuous strong draft across their tops. During the 314 days of ex- 

 posure the sandy loam, unmulched, lost 4.638 inches of water and the 

 mulched surface 4.314 inches or at the rate of 1.463 inches and 1.369 

 inches respectively per 100 days. The clay soil unmulched lost during 

 the same period and under the same conditious 9.639 inches and the 

 mulched 6.57 inches, or at the rate per 100 days of 3.04 inches and 2.07 

 inches respectively. 



The sandy loam was a little wetter in the lower 6 inches at the 

 close of the experiment than when it started, from downward percola- 

 tion, but in both columns 9. 5 feet below the surface water had been 

 lost upward. The clay loam was 1.38% drier in the lower six inches 



