122 



SCIENCE 



[N. S. Vol. XL. No. 1021 



same oTer-statement in the important com 

 states. 



The largest error is in the case of Louisiana, 

 where the Bureau of Statistics report of com 

 is 97 per cent, above the census report for 

 1909, being an error of 25 million bushels, 

 but the crop of Iowa was underestimated by 

 52 million bushels. The corn crop of the 

 United States was oTerestimated by 9 per cent. 

 But a careful study of the methods of 

 enumeration makes this error less conclu- 

 sive. By the census method of enumera- 

 tion, corn grovm for silage is unfortunately 

 put with coarse forage crops. It ought to be 

 enumerated separately. There were over four 

 million acres of such crops, of which corn 

 certainly made up the larger part. By the 

 methods used by the Bureau of Statistics, much 

 silage com is doubtless included with other 

 corn. It is probable that this would reduce 

 the error to 5 or 6 per cent. 



A study of Table I. shows that of the 

 thirteen crops reported, the production was 

 underestimated on six crops, overestimated on 

 six crops and practically correct on one crop. 

 Of the six most important American crops, 

 three, hay, cotton and potatoes are underesti- 

 mated, oats were correctly estimated, while 

 only two, com and wheat were overestimated. 

 OertairJy there is no indication of wilful 

 exaggeration. The most serious error is in 

 the underestimate of the hay crop. Census 

 reports include salt-marsh hay and all wUd 

 hay. It is probable that many crop reporters 

 do not consider any of this as hay except that 

 portion that is used for stock food. But even 

 making an allowance for this difference, it is 

 certain that the Bureau of Statistics reports 

 are too low. 



Careful study of Table I. and of the reports 

 for individual states indicate that the errors 

 in iadividual states may be very large, but 



TABLE I 

 COMPARISON OF CENSUS AND YEAR-BOOK KEPOKTS OF CROPS IN THE UNITED STATES IN 19091 



Yields of grain are given in bushels, hay in tons,cotton in bales, tobaoco and hops in pounds. 



1 Year-book reports are from the Year-book of 

 the United States Department of Agriculture for 

 1909 except the acreage of cotton, which is as re- 

 ported in the 1910 Year-book. The production of 

 cotton is the estimate as reported by the Bureau of 

 the Census in the 1910 Year-book. 



2 The Census report for grasses, clover and al- 

 falfa^ These figures may not be exactly compar- 

 able with hay as reported by the Bureau of Sta- 

 tistics. 



that the results for the United States are 

 accurate enough to be very useful. 



The percentage error is most likely to be 

 high in states that grow little of the crop. 

 The same is true of census reports. The error 

 is also likely to be large in regions that are 

 making the largest change in the area or yield 

 of the crop. 



The errors are the result of cumulatiye 



