Frequency and Length of Drought 107 



sin, we have found it necessary to water all of the crops grown, 

 in them as often as once in seven days; and even this period has 

 been found too long for the soils which are coarse and sandy. 

 So, too, in our field irrigation we have found that as much as 2 

 inches of water may be applied to corn, cabbages and potatoes as 

 often as once in 10 days, with decided advantage unless, in the 

 interval, there has been a rain of from .5 to a full inch, falling 

 nearly at one time, so as to penetrate the ground deeply. To 

 what extent and to what advantage tillage may take the place of 

 irrigation, or make it undesirable, we shall discuss in the next 

 chapter. Starting with the soil well supplied with moisture at 

 seeding time, and then a uniform distribution of rains equal to 1 

 inch once in seven days through the growing season, we shall have 

 all the moisture that would be needed for very large crops. On 

 the average of years most parts of the United States east of the 

 97th meridian have this amount of rain during the growing season. 

 It is true, however, that in many parts of the humid districts the 

 distribution of the rainfall in time and in quantity i? such as to 

 cause severe suffering from drought. 



To show just why it is that in Wisconsin the irrigation of 

 ordinary farm crops does produce a very marked increase in the 

 yield, we have made a study of the distribution of the rainfall at 

 Madison for the years 1887 to 1897, inclusive. The results 

 are here given in a condensed form, as an illustration of the type 

 of rainfall conditions under which, in a humid climate, it may be 

 desirable to irrigate where water privileges are such as to permit 

 it. to be done cheaply. 



It is generally true that a rain of .05 or even of .1 of an 

 inch, when it conies alone, separated by two or three days 

 from any other rain, benefits ordinary farm crops but little ; but 

 in order that we shall not undervalue the rain which falls, we 

 have included everything, large and small alike, and have con- 

 structed a table for these years, 1887 to 1897, which shows the 

 length and number of periods in each year between April 1 and, 

 September 30, when there were consecutive days having a rain- 

 fall whose sum did not exceed .05, .1, .5, 1, 1.5, 2, and 2.5 

 inches. The table is given below: 



