perfect or incomplete. This is the case with all agricultural statistics, and especially so with the returns of crops, always an estimate of uncertain foundation. No two fields of grain are exactly the same, and the variation of field to field is exaggerated when it is made the basis for an estimate of the condition or yield of crops in a township, a community, or a State. The observers are different, and each observer would look at the same field through the medium of a personal equation. No general rules can be laid down for preparing these estimates, and much latitude must be given to the agent. In our own crop service, as conducted by the Department of Agriculture, the estimates of crop condition are based upon reports from 56,700 regular correspondents, reporting monthly, and 140,500 special correspondents, reporting at particular seasons of the year. Even the Secretary of Agriculture is not satisfied with this machinery. "I am much impressed," he writes in his report for 1897, "with the extreme cumbrousness of the system of crop reporting that has been in use in this division" (of statistics) "during the last few years. Instead of conducing to completeness and accuracy, it would appear from the report of the statistician to in some measure defeat its own object by its unwieldiness, and by the fact that the indefinite multiplication of crop reporters weakens the sense of individual responsibility." The defects of the system have been recognized by others, who are obliged to be informed on the crop conditions, and who have been compelled to look elsewhere than to the Department of Agriculture for crop returns and estimates. The "commercial estimates" prepared by experts, and checked by actual receipts or movement of grain, find a more ready acceptance than do the "official" estimates. In the last season the commercial estimate placed the crop of wheat at 550,000,000 bushels, while the department forecast a yield of only 450,000,000 bushels, a difference too large to be admissible in a statistical examination of the same subject. The reporters for the department, practical farmers as most of them were, did not wish to report a heavy crop, lest the market be influenced and prices fall. A short crop appeals more to their interest, and their estimate inclines, consciously or unconsciously, to an understatement of conditions. "The best authorities are now agreed that in 1891 the bureau (of agriculture) underestimated the American wheat crop by 73,000,000 bushels, in 1892 by 64,000,000 bushels, and in 1893 by 79,000,000 bushels."[1] "In 1894 the Government estimate in December of the wheat crop was 460,000,000 bushels, while the best commercial estimates of that year were 525,000,000 bushels."[2]
