REPORT OF THE SECRETARY. 91 



month and year after year without money compensation; and some 

 of these men have served the department 36 years. These vohmtary 

 correspondents, numbering about 130,000, consist of township corre- 

 spondents, reporting for the townships in which they reside; county 

 correspondents, reporting for the county in which they reside, 

 from personal knowledge and upon two or more reports made by as- 

 sistants living in other parts of the county; and special correspond- 

 ents, supplying special information, such as crop yields, farm prices, 

 cotton acreage and ginning, concerning grain in mills and elevators, 

 the live stock on farms, and the tobacco industry. 



The salaried reporters are State statistical agents, one residing in 

 each State and rendering monthly reports to the bureau bused on 

 reports received by him from correspondents throughout the State 

 and on his own personal knowledge and observation, and the special 

 field agents assigned to duty in groups of States, performing travel 

 throughout their territories, examining crops, interviewing farmers, 

 country merchants, implement dealers, and others from whom de- 

 pendable information can be obtained, and reporting each month to 

 the bureau the conditions as ascertained by them. 



COiMPILING THE REPORTS. 



The work of finally making the bureau's crop estimates each month 

 culminates at sessions of a board whose personnel, with the exception 

 of the chief of the bureau, wdio presides, and two regular members, 

 is changed each month. The meetings are held behind locked doors, 

 and all telephone or other communication is effectively prevented 

 until the report is handed to 'the Secretary. 



No other Government attempts to make so elaborate reports nor 

 has so widespread or numerous crop correspondents. But the reports 

 issued from month to month do not purport to be other than esti- 

 mates; they are not the results of actual enumerations, as are the 

 figures reported decennially by the Bureau of the Census. Every 

 quantitative estimate of this bureau, whether relating to acreage and 

 production of crops or numbers of live stock, is nothing more than 

 a consensus of judgment of many thousands of correspondents and 

 a limited numl)cr of agents. 



The annual estimates regarding acreage of crops and numbers of 

 live stock are based on corresponding estimates for each preceding 

 year, there being no other bases to which can be applied the per- 

 centages of increase or decrease indicated by reports received from 

 correspondents and agents, except once in 10 years, when census 

 figures become available. 



It is, of course, out of the question that an agricultural census be 

 taken every year; the expense would be prohibitive. The only way 

 in which the constant and increasing demand for information can be 



