SCIENCE. 



[N. S. Vol. XVII. No. 440. 



area aggregate about .8 per cent., thus 

 leaving the final report less than a com- 

 plete exhibit. 



In Ohio, Indiana and Illinois, and in 

 most of the older states, the margin of 

 omissions exceeds the duplications and 

 errors exaggerating results of not less than 

 2 or 3 per cent., and in the range states 

 extending from western Kansas to the Pa- 

 cific the margin of omission is from 20 to 

 25 per cent., and for the nation not less 

 than 5. 



The margin of error in staple crops, such 

 as corn and wheat, is not, however, much, 

 if any, greater than 1 per cent., and for 

 minor crops, dairy products, poultry, etc., 

 much greater, approximating 10 per cent. 



The margin of error in the office work of 

 the census due- to errors in schedule revi- 

 sion, tabulation, result work, etc., varies 

 from .1 to .5 per cent. 



Requisites in Crop Reporting: H. Parker 



Willis, Washington, D. C. 



The subject of crop reporting is of spe- 

 cial interest at this time because of some 

 dissatisfaction in the business world con- 

 cerning the service now available. Within 

 recent years there has been a lack of faith 

 in the work done by the government offices, 

 and this has rendered it of more than usual 

 importance to study the methods of de- 

 veloping such a service along proper lines. 



The first requisite in establishing a crop- 

 report service is honesty in the officials in 

 charge. This statement implies not mere 

 personal honesty, but scientific truthful- 

 ness, freedom from bias and a display of 

 the scientific spirit. As incidents to the 

 attainment of these qualities the officials in 

 charge of any crop-report service should 

 of course be free from any pecuniary in- 

 terest in that service, and devoid of bias in 

 favor of any particular class in the com- 

 munity. They should not be permitted to 



speculate on the exchanges where those 

 products which are affected by their re- 

 ports are listed. Furthermore^ the force 

 working under them should be free from 

 bias, and should be selected upon civil ser- 

 vice principles, promotions being made 

 after a non-partisan method. This applies 

 not merely to the force in the office, but 

 also to the outside force of correspondents, 

 who ought not to represent any particular 

 class, but should be selected in such a way 

 that any bias manifested by any one class 

 will be offset by the bias shown in the re- 

 turns furnished by another. The returns 

 should be absolutely public and should be 

 furnished to all persons simultaneously. 

 They should be public, not only as regards 

 results, but as regards methods. In col- 

 lecting the figures there should be an effort 

 constantly to look for actual facts rather 

 than for opinions. In other words, crop 

 returns should, if possible, not be estimates 

 at all, but should be representative of exact 

 facts. Correspondents and all employees 

 should be paid, for in this way more accu- 

 rate and reliable results are obtained. The 

 governmenii should certainly not publish 

 estimates. If, however, it is to do so there 

 should be such a relationship between the 

 office through which the basis for the esti- 

 mates is obtained and the office furnish- 

 ing those estimates, as to insure harmony 

 of result. 



Some Vieivs of Recent Sociology: James 



H. Baker, president of the University of 



Colorado. 



One conclusion is justified, namely, that 

 social progress can not rely upon natural 

 selection alone, but must bring to its aid 

 all the forces of material and physical bet- 

 terment, of public opinion, law, morality 

 and religion. 



Democratic government is the servant of 

 the people; the will of the people can con- 



