September 14, 18S3.] 



SCIENCE. 



373 



courts. Assiimiiii; what I think all will ailinit, that 

 census inquiries shoiilil be made entirely free from 

 any suspicion that some tax or some personal liahil- 

 ity is to be incurred, it is evident that an assessor 

 cannot question an ignorant man about Iiis property 

 and his crops, without exciting his fears that some 

 tax is to be laid. The slierifforthe constable seldom 

 makes a professional call, except to serve some papers 

 or make an arrest. There is, therefore, a strong reason 

 for .appointing persons who are to make the census 

 inquiries their oidy business, and formaking it widely 

 known that there is no taxati(m, enrolment, or other 

 liability incurred by giving full and true returns, and 

 that there is no sectarian or political end to be served 

 by the Inquiry. 



This excellent end is now well enough secured 

 under the national law, and in several of the st.ates. 

 They have a still bettermelhod in Great Britain, where 

 a system of registration of births, marriages, and 

 deaths, has its districts and its agents under constant 

 organization, and to which, once in ten years, the 

 census can be assigned, without creating new offices. 

 In Sweden, where a system of registration, including 

 also a record of change of residence, is in charge of 

 the i)arish-clerks, they take a census whenever they 

 choose to post the books, without any special inqui- 

 ries being made, more than what these records con- 

 tain. 



In the national census before 1S50, — in New York 

 before 18.5.5, — and in some of the states still, each 

 family had one line upon the blanks; and the number 

 of persons of different ages, sexes, and colors, was 

 entered in cohnnns provided. The limit of classifica- 

 tion was of course restricted to these columns; and, 

 although the totals of each class were easily obtained 

 by adding, the results were meagre and unsatisfac- 

 tory. 



The change that allowed a line for e.ich name, one 

 column for the exact age, and other columns for native 

 counlrj-, profession or occupation, etc., while it sim- 

 plified the Labor of taking, allowed ample field for 

 classification; and U m.ade it necessary to employ a 

 large force of clerks, in a central office, for the reduc- 

 tion of the returns for publication. 



By a method now generally used in Europe, the 

 census is taken upon "householders' schedules,' 

 which are distributed one to each family, some days 

 beforehand, filled out by the head of the family, and 

 collected upon one day. The only instance in which 

 this has been done in th(! United States, within my 

 knowledge, was in the District of Columbia, upon 

 the Uth of November. 1807. This census was taken 

 by the metropolitan police, uiuier my own direction, 

 and with entire success. It was attempted in the 

 city of B.altimore some months afterwards, and failed, 

 apparently from want of proper management on the 

 part of those in charge. 



For all kinds of official inquiries, relating to busi- 

 ness, as well as to personal statistics, I think the true 

 anil proper method is, by me.ans of upecUtl blanks, 

 carefully prepareil, simple, and fully explained. 

 These should be distributed some little time before- 

 hand, and should be taken up, if not in one day. 



within a short period of time, hut with reference to 

 a given day. The chief difficulty to be encountered 

 is the illiteracy of tli ^se who should fill the blanks; 

 but in the District of rolumbia, which in 18(i7 con- 

 tained a large number of colored families, but recently 

 freed from slavery, the blanks had been, in almost 

 every case, filled out by some one to whom they had 

 been carried. 



In follow-ing our subject, — the ' methods of statis- 

 tics,' — we may notice some points in the condensa- 

 tion .and arrangement of facts that may be of interest. 



With a vast amount of information before us, as, 

 for example, in the returns of a census, let us con- 

 sider what is to be done, and how it can be done with 

 the least labor and greatest certainty. After inspec- 

 tion to make sure that the work is all together, in 

 proper order and condition, it will be found that sev- 

 eral distinct operations are necessary, in preparing 

 the results for the pres-s. Columns of figures must of 

 course be added, and carefully revised. As the totals 

 of several sheets will often be consolidated into one 

 sum, it is best to use spare sheets of the same sched- 

 ule fur entering the totals of pages, so that these 

 partial totals can be ciisily combined. It is always a 

 good practice, where long columns, of many figures 

 in eacli, are to be entered for adding, to provide paper 

 with narrow vertical ruling, that shall allow of but 

 one figure in a space. In cases where the first two 

 or three right-hand characters are generally ciphei-s, 

 they may be left out altogether, the significant figures 

 only being entered in their proper places. It saves a 

 little time and labor, and does not lead to error. 



Where a great amount of statistical material is re- 

 ported, — as, for example, the names in a census, — the 

 blanks should always be plainly divided by horizontal 

 and vertical lines, printed in preference to ruled. The 

 horizontal lines should be numbered from the top 

 downward, upon both margins. This numbering is 

 the more important where an entry is carried across 

 to another page. Each line should contain but one 

 entry, and there should be, if possible, no blank lines 

 except at the end. Then, with a little multiplica- 

 tion-table at hai.d, showing the number of lines in a 

 full sheet, and for each number up to the highest that 

 are likely to be found in a return, the totals can be 

 rapidly and accurately ascertained, as follows: The 

 number of sheets is first counted on the back edge, 

 and the number of entries they should contain, if 

 full, is set down. Then, by glancing over each page 

 separately, it is easy to notice whether there are any 

 lines with two entries, any blanks, or any lines in 

 excess. The deficiencies are set down in one column, 

 the excess in another, and their difference is added or 

 subtracted, as the case may require, when the true 

 sum is at once found. This operation, which is the 

 first thing done, should be repealed by another per- 

 son; and, when found to agree, it should be kept as a 

 test-nnmberfor verifying the accuracy of much of the 

 work th.at is to follow. In me.isuring parts of pages, 

 a scale made of a strip from the margin of a blank 

 schedule, and pasted upon a card, will save all labor 

 of counting. 



lu statistical labors, where the same returns afford 



