4 BULLETIN 1458, U. S. DEPARTMENT OF AGRICULTURE 
SOURCES AND CHARACTER OF THE DATA USED 
The period for which crop-yield statistics are available is rela- 
tively short, not only in the United States, but in Europe. Modern 
crop-reporting systems are largely a development of the second half 
of the last century. During the last 50 years the application of 
statistical science to the yield of crops has furnished us with a basis 
for determining the changes in acre yields realized on many of the 
large agricultural areas of the world. 
Prior to 50 or 60 years ago the literature of crop yields is scant 
and fragmentary (except for official statistics for France since 1815). 
The small amount cf material available consists largely of the per- 
sonal observation of certain men. These individuals based their 
estimates on observed yields for relatively small areas and on a 
smal] number of cases. In addition to this there are also records 
from some farm estates in various countries. 
In the United States the two principal sources of information on 
crop yields are the estimates of the United States Department of 
Agriculture and the reports of the United States census. It was not 
until after the Civil War that the Department of Agriculture began 
the publication of annual estimates on crop yields. ‘The question- 
naire of the Tenth Census in 1880 included questions on both acreage 
and production of crops, making available for the first time crop- 
yield data based on actual enumeration. 
During the first years of the publication of the yield estimates of 
the Department of Agriculture, owing to various imperfections of 
the system and an inadequate number of reports on which the esti- 
mates were based, the estimated crop yields were probably larger 
than the actual yields. However, the degree of this bias can not 
be determined with any degree of certainty, although the results of 
the Federal census of 1880 indicate that previous to that time the 
yield estimates published by the United States Department of Agri- 
culture were too high. The existence of this bias prior to 1880 was 
discussed and graphically shown by Milton Whitney (74) in 1909, 
and also was indicated in a chart showing the comparative crop yields 
from 1866 to 1912 for the States east of the Mississippi River, in a 
study published by the New York State Agricultural Experiment 
Station (72, p. 191). Since 1885 these estimates have been much 
more accurate, and at the present time are based on numerous reports 
from every county in the United States of any agricultural impor- 
tance. 
Another source of crop-yield statistics in the United States, cover- 
ing a considerable period, is that available from the annual reports 
of the agricultural or statistical organizations of various States. 
Among these, Ohio, Kansas, and Iowa have furnished a continuous 
series of annual reports for the longest periods. The accuracy of 
some of these reports for the period extending at least up to 1890 
has been questioned by many on the ground of imperfections in the 
system of getting the county returns and the making of the esti- 
mates. But the quality of these reports has improved in recent years 
and in some cases are now considered to be free from serious error. 
These reports are valuable in that they furnish a basis for determin- 
ing some of the changes that have occurred in the agriculture within 
the boundaries of these States. Through cooperation of the United 
