NINTH CENSUS OE THE UNITED STATES. 
COMMUNICATIONS FROM MR. EDWARD JARVIS. 
Dorchester, Massachusetts, February 17,1869. 
Dear Sir : In your favor of the Bth instant you proposed four spe- 
cific and one general question for my consideration : 
1. Whether the census can be taken more thoroughly and economi- 
cally through the officers of the internal revenue than through the 
marshals. 
2. Whether any changes should be made in regard to compensating 
the census-takers, whoever they are. 
3. In how short a time do you believe it practicable to take the cen- 
sus, and to take it well. 
4. What modifications ought to be made in the tables. 
5. Whether any other suggestions occur to me to be made as to this 
whole matter. 
I will answer these in their order; yet the Ist and 2d must be com- 
bined in some degree; also the Ist and 3d, and the sth may.apply in 
some measure to all. 
1. Whether the marshals and their appointees or the internal reve- 
nue officers and their appointees shall be employed. 
The use of the latter, the assessors, promises the best chance of a 
thorough and reliable enumeration of the peoj>le. 
There is but one marshal in most States, and at most one in each 
judicial district—s2 in all the United States according to the register of 
1808; and there is one assessor in every congressional district, making 
near five times as many assessors as marshals. Each marshal appoints 
deputies—four in Massachusetts—and if this is the average, there are 
208 in the United States. The assessors appoint about one assistant 
assessor for every 15,000 or 20,000 inhabitants, or about seven for each. 
If so, there are between 1,600 and 1,700 of these officials in the whole 
country, probably eight times as many assessors as deputy marshals. 
The assistant assessors are scattered through all the country, each 
resident in the district of his labor, and presumptively acquainted 
with the character and condition of the people, their occupations, and 
their means. 
The deputy marshals are more in the centres of population, many, 
perhaps most, and in some districts all of them resident at the place 
of the court—Boston, Providence, New York, Trenton, &c. 
The assessor’s office requires them to have some knowledge of, and 
some direct or indirect intercourse with, every family. The deputy mar- 
shals have business only with few—with those having business that may 
come under the cognizance of the national courts, or with offenders 
against the national laws. In either case these standing officials can do 
but a part, and properly only a small part, of the work of the census, 
but must appoint deputies, who, if the plan I shall hereinafter propose 
be adopted, must be very many. 
The assessors, principals and assistants, being among and nearer the 
people and more acquainted with them, will be able, first, to select for 
the enumerators men of proper character, who are themselves residents 
of their respective sub-districts and know the localities and the condi- 2 
NINTH CENSUS OF THE UNITED STATES. 
tion and circumstances of the people; and, second, having a general 
knowledge of the people, these officials—the assessors—can better judge 
of the character of their reports, whether they are consistent with the 
known or probable facts, and to discover any error. 
In the revenue department will be found a better class of minds—men 
of better education, much better training, more used to business, to 
accounts, to analysis and combination; they are presumptively discip- 
lined in accurate habits of thought and accustomed to close and truth- 
ful dealing with men—with facts, and representing these with mathe- 
matical precision. 
The marshal’s office does not necessarily require these qualities, for 
either principal or deputy, nor does their administration call them into 
daily and constant use. 
In either case the appointing officers would naturally select for their 
deputies, or the enumerators, according to their own ideals of charac- 
ters—the assessors, men of precision and rigid discipline; the marshals, 
simply men of fair character, or those personally acceptable. The gov- 
ernment, therefore, will probably obtain a more thorough and reliable 
enumeration through the internal revenue than through the marshal’s 
department. 
The economical view refers to the value of the work rather than the 
actual expenditure. That which will produce the most complete and 
reliable census is the cheapest; and therefore, I suppose,the census can 
be more economically taken by the revenue officers and their co-operators. 
In either case the enumerators—those who must visit and take account 
of every dwelling, shop, and farm, and their population and wealth— 
should be paid sufficient to secure the requisite skill and faithfulness; 
and such service cannot be obtained by the government at less cost than 
the same degree of skill and amount of service are obtained for in the 
general business of society. 
2. The present manner of compensation, or rather that adopted in the 
law of 1850, and practiced in the seventh and eighth censuses, seems to 
be just and expedient, but the rate should be increased to suit the present 
state of financial values. The law of May, 1850, offers the marshal one dol- 
lar for every thousand persons enumerated under his direction, provided 
there are one million or more people in his district, and one dollar and 
a quarter if the population be less. But in no case shall he receive less 
than $250; and if he have less than $5OO, he may have allowance for clerk 
hire. By section five, the marshal “ shall determine the rate of compen- 
sation to each assistant marshal,” and by section 12, “each assistant 
shall be allowed as compensation for his services, after the rate of two 
cents for each person enumerated, and ten cents a mile for necessary 
travel,” and by section 13, in addition, “there shall be paid for each 
person fully returned ten cents; for each establishment of productive 
industry fully taken and returned fifteen cents; for the social statistics 
two per cent, upon the amount allowed for the enumeration of the popu- 
lation, and for each name of a deceased person returned, two cents. 
If this compensation was no more than sufficient in 1850 and in 1800, 
it is not now. All other labor, whether skilled or unskilled, now receives 
nearly or quite double the wages that were given to it 10 and 20 years 
ago, and the government must pursue the same course, or be content 
with inferior service, and inferior worth of the work when done. 
The British government paid the enumerators of England and Wales 
ten dollars and ninety-six cents ($lO 00) per thousand, in 1841, eleven 
dollars and eighteen cents (sll 18) in 1851, and eleven dollars and forty NINTH CENSUS OF THE UNITED STATES. 
3 
cents (sll 40) in 1861. Wages are very much lower in England than in 
this country: and the districts, as I shall hereafter show, were much 
more easily surveyed than ours. 
I see no reason to change the plan of payment, that is, so much per 
person, house, farm, or shop, &c., enumerated and recorded. The rate, 
however, of 1850 and 1860 should be increased, and there should be a 
discrimination in favor of larger payment of the enumerators, in the 
sparse districts of the newest States and Territories, whose labors and 
difficulties are much greater than in the old States. This should be left 
to some superior officer, as the principal assessor in the States, or the 
chief of the Census Bureau, or some other proper officer to be determined. 
3. In how short time can the census be taken properly? The answer 
to this depends: 1. On the plan adopted for the enumerators. 2. On 
the kind of men employed as enumerators. 
If the former plan of large districts and few enumerators be adopted, 
a month, at least, should be allowed them to collect the facts and prepare 
their returns. 
In the early censuses the enumerators seem to have been allowed long 
periods to collect their facts and make their returns. 
1. 1790. The marshals of most of the States dated their reports, and 
these dates are printed in the government report. These dates vary 
from May 4 to September 19, 1791, and the South Carolina report was 
dated February 5, 1792; nevertheless the Secretary of State published 
his reports of the whole, except South Carolina, October 24, 1791. 
2. 1800. The State reports are dated, North Carolina, August, 1800; 
the others vary from June 2to December 21, 1801. The report of the 
Secretary is dated December 8, 1801. 
3. 1810. I have not a copy of this report, but I consulted that in the 
library of the Massachusetts Historical Society, which unfortunately 
lacks the title and some prefatory pages. I am unable therefore to say 
when the government completed and published it, but some of the State 
reports are dated January and February, 1811. 
4. 1820. Census takers began the first Monday of August. They were 
allowed eight months, to April 1,1821, to complete the enumeration and 
transmit their report to the Secretary of State; and by act of March 3, 
1821, this time was extended to September 1, 1821. The government 
report is dated December 18, 1821. 
5. 1830. The facts relate to June 1. Enumerators allowed to February 
1, 1831, to collect and return their facts. This was extended to August 
1, 1831, by act of February 3, 1831. The general report is dated Jan- 
uary 10, 1832. 
0. 1840. There is no date nor certificate of State or national officers 
attached to the general report or the compendium; nothing to indicate 
its progress except the printer’s date, 1841, on the title-page. 
7. 1850. The first short summary report was made to Congress Decem- 
ber 1, 1851. The second, December 1, 1852; printed 1853. The princi- 
pal report is dated November 10, 1853. The compendium is dated Sep- 
tember 1, 1854. The mortality volume was dated November 20, 1855. 
I find by my letter-book that I corrected the last proof of the part that 
I contributed to this volume, November 17, 1855. 
8. 1800. The preliminary report was dated May 20, 1852. The first 
large volume, population, was printed 1854. I prepared the mortality 
report in the last volume. I wrote the last chapter January, 1800. I 
wrote the index iu March and corrected the last proof in April, 1800. 
Thus you see, that iu all the eight enumerations aud preparations of 4 
NINTH CENSUS OF THE UNITED STATES. 
the reports a long time has been taken, which, with the means and force 
allowed by the government, was necessary. 
The districts were very large, and the houses and places of business to 
be visited, and people to be counted, and all the circumstances, ages, 
sex, conditions, &c., to be recorded, were very many. The marshal in 
the seventh and eighth censuses carried his six schedules, with his 141 
questions, beside those concerning slaves, to be asked and answers 
recorded. He must find at each house, shop, farm and store, a trustwor- 
thy and willing informant. He must patiently go through all these items, 
and to many offer much explanation in order to elicit the desired infor- 
mation. 
In many cases, the marshal, not finding a suitable witness in or about 
the house, was obliged to make moie than one visit, travelling over the 
same ground, perhaps long distances, several times before he could 
accomplish his purpose. 
As in the progress of years the people have become more intelligent 
and more familiar with the manner and objects of the census, they are 
better prepared with the facts and feel more ready to co-operate with 
the public officers. It has been, therefore, growing less and less difficult 
and laborious to get the facts: besides, there has doubtless been an 
improvement in these officers. They are more intelligent and disciplined, 
more courteous and persuasive, and win the popular confidence more 
and more easily, and hence they do their work within a month, while 
their predecessors required eight and ten months, and more, to gather all 
that was required. 
It must be remembered that even now, or, at least, at the enumera- 
tions of 1850 and 1860, as well as previously, the assistant marshals 
were new to their work. They have been appointed afresh every 10 
years. They had no previous training for, or knowledge of, this business. 
Many of them were mere hungry men, wanting any employment, with 
no exact mental habits, no discipline, no mathematical acquirement, no 
facility of explanation to make the questions clear to the dull, and no 
courtesy to win over the unwilling. With the chance of such as these, 
with the average of such as probably will be selected if the old systems 
s hall be continued, and especially if political affinities are deemed essen- 
tial qualifications, we cannot expect that they can complete the enume- 
rations in less than a month’s time. 
The British system is better than ours has been. 
England and Wales are permanently divided into 031 registration dis- 
tricts, and these into 2,101 sub-districts, with a permanent superintendent- 
registrar or sub-registrar in each. These officers are constantly employed 
in gathering, recording, and returning the facts connected with "the 
births and deaths. Thus there is spread a constant corps of trained, 
disciplined, intelligent officers, used to this kind of work, in continual 
intercourse with the people and making inquiries as to their domestic 
condition and history. 
To this corps of near 3,000 established officers is the census committed. 
They are required to divide their sub-districts into others so small and 
containing so few houses and families that an enumerator can visit each 
and collect all the facts in one day. There were 31,144 of these small 
districts, and as many enumerators appointed by the sub-registrars, at 
the census of 1861, in England and Wales. 
The General Report, census 1801, vol. 3, p. 1, says: 
The enumerators were required to be intelligent and active, able to read and write well, 
not younger than 18 years of age nor older than 60. They were to be respectable per- 
sons, likely to conduct themselves with strict propriety and courtesy in the discharge of their 5 
NINTH CENSUS OF THE UNITED STATES. 
duties, and well acquainted with the district in which they were to act. No difficulty was 
experienced in procuring the services of a highly respectable body of enumerators, including 
clergyrnen and many other professional men, who undertook the work from public motives. 
In another place it is stated that many schoolmasters were employed 
for this purpose. 
Schedules were prepared at the General Eegistry Office in London, 
one at least for each family in the kingdom, containing lines and 
columns, places for the record of all the names, and facts, age, sex, 
domestic condition, &c., of each person in the kingdom. 
These blanks were sent—over 40 tons—in 1861, from the central office 
to the 631 superintendent registrars; by them sent to the 2,101 sub- 
registrars, and by these distributed to the 31,114 enumerators through- 
out the kingdom. 
Each enumerator left one of these blank schedules with each family 
in his district, within the week next preceding the Bth of April. The 
family was directed to fill the schedule, with answers to each inquiry, as 
to each member, referring to the night preceding the Bth April. 
On the Bth of April the enumerator called on each family, took the 
schedule, and examined it carefully, item by item, with some responsible 
member of the family, to make it certain, by renewed affirmation, that 
each answer was fully and correctly made. But if the schedule was not 
filled, or only partly filled, then lie took the most intelligent member 
of the family, or witness, that he could find, and with this aid he made all 
the requisite record. 
In this manner, and with this plan, so carefully drawn and so system- 
atically carried out, the census of England and Wales was taken in one 
day. The same was done in Scotland in 1861, and also in all Great 
Britain in 1851. 
We can adopt and carryout the same plan in this country. With the 
aid of the assessors and their assistants the whole country may be divided 
into small districts, which one man can traverse and canvass in one day. 
We have a very much more intelligent population than any European 
country, and our people are very much more used to co-operating with 
the government. They have more sympathy with public operations; 
have more regard for law, and take more pleasure in carrying out national 
and State measures than the people of old countries. 
If there be any doubt raised, or exception made, it can be only in some 
of the new States and Territories, where possibly, but not probably, there 
may be found no fitting and willing man to enumerate the people in some 
districts. If so, an officer can be sent; some member of the army, or 
some soldier, under the direction of his superior; and, also, in some dis- 
tricts the dwellers may be so widely separated that a day is insufficient 
for the work. Provision of more time and payment should be made for 
all these cases. 
In all cases the enumerators should be held responsible for the cor- 
rectness of their records, and for the probable correctness of their facts. 
They should see that no statements inconsistent with each other, no such 
gross improbabilities, or manifest impossibilities, as have been allowed 
to come into some of our former censuses, should now be permitted to 
appear in their schedules and returns to the assistant assessors. 
Then these assistant assessors should revise all these returns for the 
same purpose, comparing them with their more general knowledge of the 
people and their condition ; to see if every village, hamlet, street, road, 
and dwelling, so far as they know, has been visited, and the statements 
are all consistent with each other. 
Again, all these returns should be re-examined at Washington by men 
competent for the purpose, to see that every class of facts has been 6 
NINTH CENSUS OF THE UNITED STATES. 
obtained in every State, county, and town; and no such omissions as 
whole counties without a death, or a farm, or a store, that have formerly 
been made, should be again permitted. 
In all these cases, if any such error, fault, or omission, appear to the 
assessor, or to the authorities in Washington, the papers should be 
immediately returned, or exact copies of them, to the subordinate officers, 
assessors or enumerators, and they be required to re-examine, and cause 
correction to be made. 
USE TO BE MADE OF THE FACTS AFTER BEING GATHERED IN THE CEN- 
TRAL OFFICE AT WASHINGTON. 
This ’is a matter to be considered in connection with the limit of time 
allowed for the publication of the report of the census. This is very 
various. It was as little as possible in the early censuses. The work 
filled four large quarto volumes in the eighth census. There may be a 
great amount of profitable tabulation, calculations, and deductions, that 
will show what we are as a nation, what progress we have made, and 
what we may be. There may be many valuable lessons of philosophy, 
of political economy and humanity drawn, and much instruction in regard 
to health, happiness, and prosperity, derived from the census. 
The character of the men who are to do this work on the gathered 
facts, the qualifications of mind, experience and heart, constitute a very 
important element in the consideration of the time when the report shall 
be made, after the enumerators have done their work. 
Yearly, I think quite, all other civilized nations have permanent sta- 
tistical bureaus connected with their governments. These are managed 
by permanent officers who are scholars, trained mathematicians; men 
eminently fitted for their work. They are generally, perhaps always, 
retained long in their office, for their governments find it for their interest 
to keep and profit by the talent which they develop and cultivate. Dr. 
Farr, master spirit of the English registry office, has been there from 
1837. Mr. Hammich and others have been there nearly as long. They 
are among the accomplished and practical scholars, mathematicians and 
calculators of the nation. I met them in the London Statistical Society, 
where they are among the leading members, and their articles in that 
Society’s Journal are authority for the world on questions of vitality, of 
population and political economy. 
Berg of Sweden, Engel of Prussia, Baumhauer of Holland, Quetelet 
of Belgium, Legoyt of France, chiefs of their bureaus of statistics, have 
been long, and some of them very long in this branch of the public ser- 
vice. So also Yon Hermann of Bavaria, Yarrentropp of Frankfort. 
They are retained as long as they can be persuaded to serve. I met 
these and many others at the International Statistical Congress in 
London, 1860. I find their names in the reports of the same body at 
their meetings in Belgium, Prussia, Paris, Italy, Anemia, &c., years 
before and after 1860. Their names appear also in the periodical reports 
on population and mortality of their respective countries, and in their own 
statistical journals. 
With such trained and cultivated men, and with co-operators of simi- 
lar character and experience, the European nations are ever ready to 
take their census, most of them once in three or five years, and to send 
forth reports full of wisdom and instruction, of reliable statements and 
deductions profitable to their peoples and to the world. 
Their reports are not only more valuable, but also prepared in a shorter 
period than those written and published by extemporized corps of unused 
men, with no especial fitness for their work. 7 
NINTH CENSUS OF THE UNITED STATES. 
Yet even these statistical bureaus of men so fitted and ready could 
not prepare their reports of census in very short periods. 
In England, census of 1851: The schedules were collected by the enu- 
merators March 31. The first volume of the report, folio, 221 pages, was 
dated July 21, 1852. The second and third volumes, folio, 1,438 pages, 
population tables, ages, &c., were dated April 24, 1854. 1861: The 
schedules were gathered April 8. The returns were all received in Lon- 
don in May. A general statement of the population was made June 7. 
First volume of report, with summary tables, was dated September 29, 
1862. The large report was dated July 23, 1863. 
Scotland, 1851: Schedules left at houses first week in April; collected 
Bth. Examined, corrected, and sent to registrars 16th April, lie-exam- 
ined by registrars and sent to clerk of sheriffs 30th. Sent to regis- 
trar general, Edinburgh, June 1. Some four distant districts were not 
received until July, and those from St. Kilda not until 28th July. First 
summary sent by registrar general to secretary of state July 22. First 
volume of report dated June 22,1862; second volume, population tables, 
dated February 10, 1864. Both folio, large. 
Ireland, 1851: Schedules left at houses 20th to 29th March ; schedules 
collected 31st; abstract submitted by registration board to the lord lieu- 
tenant August 4; results of census by “townlands” printed; first vol- 
ume, March 15, 1852; second volume, November 30; third volume, 
status of disease, March 30, 1854; fourth volume, ages and education; 
fifth and sixth volumes, deaths and causes, January, 1856. 1861: 
Schedules collected April 8; report dated June 24, 1862; four other 
volumes on agriculture, and whose dates I have not time to examine. 
France, 1856: Taken July 1: published 1859. 1861: Taken Juuel; 
published 1864. 
Holland: Taken 1859 ; published 1864. 
Prussia: Taken December 3, 1864; printed 1867, dated July 15. 
Hanover: Taken December 3, 1858; printed 1860. Taken December 
3, 1861; printed 1863. 
Berlin, Prussia: Taken December 3,1851; printed 1863; dated Octo- 
ber 11, 1863. Taken December 3, 1864; printed 1866, dated April 22. 
Bavaria: Taken 1858; printed 1861. Taken 1861; printed 1844. 
Taken 1864; printed 1865. 
Sweden: Taken December 31,1866,; dated February 29,1868; printed 
1868. 
Portugal: Taken January 1, 1864; report published 1868. 
Norway: Taken, facts of, January 1,1866; report dated October 1,1868. 
South Australia: Facts of April 7, 1861; census taken iu one day, 
Bth April; report published March 15, 1862. 
The censuses of Belgium do not enable me to give you so accurately 
the dates of enumeration and publication. I find the same difficulty iu 
the Spanish census and Swiss. 
Thus you see that even the nations best prepared require considerable 
periods for the preparation of their reports, and those that do this in 
short periods give the most meagre and least valuable reports. Sweden, 
which published in 14 months, sends forth the fewest and the least 
profitable tables. 
The most valuable documents of this nature that I have seen are the 
English of 1851 and 1801; the Irish of 1841, 1851, and 1801. That of 
1841 was said to have been, in its time, the best census in the world, 
and with good reason. That of 1851 was in 10 folio volumes, which is, 
perhaps, my richest treasure. The censuses of France and of Belgium 
are very full, and worthy of the study o± political economists in every 8 
NINTH CENSUS OF THE UNITED STATES. 
nation. Most of these contain many very important tables on the la w 
of life and the law of mortality, which, we hope, our own will imitate 
hereafter. Certainly it is to be hoped that our government will so 
measure, analyze, and tabulate the population and mortality as to furnish 
the life tables for our life insurance companies, and not permit them to 
do their work on the European life tables, and, from the experience of 
life and death in other countries, conjecture the value of life in America. 
The last English life-table prepared in the registry office, by and 
under the superintendence of Dr. Farr, is a quarto volume of 005 pages, 
and full of the most important calculations, tables, and instructions. It 
is valuable to the world, to us in America as well as to the people of 
Europe ; but it is based on English and not on American facts. Its 
principles, applied to our experience of life and death, would furnish to 
our government, life insurance companies, and people most important 
aids, in their business and self-management. 
The last census—eighth—of the United States is an improvement on 
its predecessors. It contains much that is valuable, and nothing, or 
next to nothing, that should be omitted. If further time could have 
been granted several other chapters would have been added to the mor- 
tality report: 
1. On the comparative value of life in the northern and southern St ates. 
2. Qn the comparative cost of producing, developing, and maturing 
men, and their comparative returns made to the country in labor 
through their self-sustaining years, in the north and south. 
3. On the comparative value of life, longevity, and productive force 
and value to the nation of foreigners and natives in the United States. 
4. On the comparative value of life and production, or laboi iug power, 
in this and other countries. 
5. Refutation of the opinion, held by some abroad, that human life 
and power deteriorate in course of generations in America. 
6. On the cost of development of a man, or building the human pro- 
ductive machine, and his worth to the body politic. The interest of the 
nation and government in the health and preservation of the laboring 
power of each individual. This is the political economy of health. 
7. On the effect of cities or the condensation of population on human 
life. (This is soon to appear in the Atlantic Monthly Magazine.) 
8. On the effect of prosperity and poverty on the development and 
preservation of life. 
0. On the effect of education and ignorance on the same. 
10. The connection of low health with crime. 
11. The increase of human life in the progress of ages. (This is soon— 
March, April, and May—to be in the Atlantic Monthly.) 
12. The comparative interest in human life and health manifested by 
the governments of the several nations by legislation, &c. 
13. The law and practice of the several States of this country in 
measuring the life and vitality of their several peoples. 
14. The comparative spread of consumption and pneumonia in the 
north and south, in old and new States. 
These were partly prepared, the facts gathered, and calculations mostly 
made, in reference to them, in 1804 and 1805. I wrote them, as on the 
preceding page and half, 21 and 22 of this letter, to Mr. Harlan, the 
Secretary ot the Interior, and proposed that they be completed and 
added to the report, but the government thought it not expedient to 
delay the publication of the volume. 
Your fourth question: What modification should be made in the 
tables, remains to be answered. 9 
NINTH CENSUS OF THE UNITED STATES. 
I am not sure whether you refer to the schedule of primary inquiries, 
or to the tabulation of these facts, in the reports. I will, however, 
ans\Ver to the first, 
I have here copies of the schedules that were used by the marshals in 
18G0. 
The plan of this was drawn up by the American Statistical Associa- 
tion, in answer to a circular from the census board in 1849 or spring of 
1850. It was adopted, with little alteration, by the census authorities. 
The basis was the English census. Columns 8 and 9 of this schedule 
indicate inquiries that, I fear, will elicit no satisfactory informa- 
tion. On the contrary, it is probable that these will irritate and 
offend, and render the informants less disposed to answer the other 
questions. Yet Mr. Shattuck, whose opinion on these matters I regard 
very highly, thought much of ascertaining the fact of the number of 
real-estate owners, as persons who had the most interest in the prosper- 
ity of the country. Yet I would omit these queries, and, if need be, 
obtain the facts of ownership of real estate from the public registers. 
Schedule I.—population. 
Column 11 should rather refer to the standing fact of the married state 
than to the event of marriage, within the year. That is of but little 
consequence, taken alone. But the broader fact of the number of per- 
sons who are single, who are married, and who are widowed, is of very 
great consequence, and is one of the most valuable evidences of national 
prosperity and happiness. 
Most, and I think nearly all, European nations ascertain and publish 
this fact. 1 trust we shall do likewise. If so, this column should be 
headed, “Single, married, widowed,” and against each person should be 
written S, M, or W. 
Column 13 would better indicate the education of the people if the 
age were 15, instead of 20, as few, or more probably none, ever learn to 
read after 15. 
Column 14 is a singular and most uuphilosophical medley of incongru- 
ous matter, which may confuse the informants and trouble the enumer- 
ators. As it reads, “ which misfortune befalls the man, deaf-muteness, 
blindness, insanity, idiocy, poverty, or criminality?” the answer may 
be impossible, or very difficult. He is probably neither, or he may be 
all. I know of some who are criminal, insane, and pauper; some who 
are blind, insane, and pauper; some who are deaf and dumb, blind, and 
pauper. 
The sanitary facts, the questions of mental and bodily disease or dis- 
ability, should be determined in regard to every person in the nation. 
For the main purpose of the census, which is to measure the power 
of the nation, it is not enough to merely count the people. They differ 
very greatly in their worth. A human being of any definite age is not 
a constant quantity in all cases. Men of the same age are not alike in 
force and effectiveness. The terms man of 30 and teaman of 30 do not 
represent a definite and invariable element of population. One is strong, 
another is weak; one is a supporter of the community, another is a bur- 
den on it; one contributes to the common power and commonwealth, 
another lessens them. One community of 10,000 people, with all its 
members in full health, is stronger and more effective than another, in 
which 10 per cent., 20 per cent,, or 30 per cent., are sick or weak. 
The government, then, wants to know, as precisely as possible, the 
extent and kinds of the disabilities that rest upon the people, in order, 
first, to know what discount should be made in the apparent amount of 10 
NINTH CENSUS OF THE UNITED STATES. 
its force; and second, to know the kind and place of these burdens, 
and then, so far as it can, find a way of relief. 
It would be well, then, that this column should be headed: Is he or 
she in good health for Ms or her agef If not, What is his or her disease or 
disability f 
If well, the record may be simply a mark, to show that the question 
was asked and answered in the affirmative; if not, then the record will 
be according to the fact—-fever, consumption, dysentery, broken leg, loss of 
arm, cut wound, insane, idiot, deaf and dumb, blind, the. 
As this inquiry is to be made in respect to each individual who, at the 
time, is under consideration and in the mind of both enumerator and 
informant, it cannot be evaded. There must be an answer specific and 
direct. Whereas, if the question be asked in general terms, are any 
insane or idiotic persons in your family ? frequently it is not answered, 
or in the negative, when such are really there. In this way the govern- 
ment will obtain much fuller and more reliable information, especially in 
regard to diseases and disabilities that are not pleasant to the sensibili- 
ties nor creditable to the pride of families. 
With the answer to all these questions, in respect to all the people, the 
census officials at Washington would first determine the exact number 
of healthy, able-bodied, and sane persons in each State and all the 
States; and also the exact amount of each disease prevailing in each and 
all parts of the country. 
This would be a most important contribution to the vital statistics of 
the United States. This plan was proposed in the international statis- 
tical congress at London, in 1860. The report recommending it is in the 
transactions of that congress, page 497. The plan was partly adopted 
in Ireland at the census of 1861, and produced tables of great worth, 
showing the extent of each kind of disease and injury in that country 
when the enumeration was made. 
With this information the government would know the exact measure 
of the power on which it rests, and by which it may effect its purposes. 
This, moreover, offers a standard of the national power at each suc- 
cessive census, by which the progress or decline of the country may be 
determined through the decades and ages of its being. 
The other matters in column 14—the number of convicts and pau- 
pers—cannot be ascertained at the houses of the people. The family will 
not confess the criminality of any of its members, nor do they love to 
confess their pauperism. But these facts can be fully and satisfactorily 
learned from the public authorities, or at the institutions where these 
classes are confined or cared for. 
CONVICTS AND PAUPERS. 
Schedule 2. 
This is useless now. 
Schedule 3.—mortality. 
This is very important. All its inquiries should be made and answered 
with very much more care than they were in 1850 and 1860. The enume- 
rators should be particularly enjoined to pursue this matter in every house 
and family, and not be satisfied with the simple statement of ignorance, 
as heretofore in many cases. In 1850 some whole counties returned no 
deaths. 
In both 1850 and 1860 this schedule brought out much valuable infor- 
mation ; but this was far short of what should have been obtained. It 
afforded no means of calculating the rate of mortality in any State. NINTH CENSUS OF THE UNITED STATES. 
11 
Column 5 is now useless. 
Column 7 will not produce information of any value, and will only 
burden the inquirer and witness. 
Schedules 4 and s.*—agricultural and manufactured pro- 
ducts. 
I have no suggestions to make in regard to these. 
Libraries.—The instructions of the superintendent limited these to 
large, generally public, libraries, containing not less than 1,000 volumes. 
It would be a most important indication of the extent of education and gen- 
eral intelligence if this inquiry should lie extended to every family. If there 
was a column on schedule 1, headed, “ Number of books in the house, 
belonging to the family or persons,” some, but not many, would say, 
“None;” few—possibly many—would answer, “The Bible and a few 
school books.” There are many, especially in the northern States, that 
have 100 volumes; many that have 200 or 300. Within my knowledge 
there are not a few who have 1,000 to 2,000, or 3,000. 
Schedule 6.—social statistics. 
This inquiry would develop a state of culture honorable to our nation, 
and showing a great advancement in civilization in comparison with the 
people of the older world. 
I am not sure that I would put this into the schedule to be asked of 
every family, yet I would like to see it there, and the results of the 
experiment, and I commend it to your consideration. 
I have yet to speak of your fifth and last topic—any other sugges- 
tions? This I must put in another letter, which may go hence to-mor- 
row. 
Till then, believe me, with high respect, very truly, yours, 
EDWARD JAEYIS. 
General James A. Garfield, 
Chairman Committee on Census, 
House of Representatives, Washington, 1). C. 
Dorchester, Mass., February 19, 1809. 
Dear Sir : In my last to you, mailed this morning, I answered four 
of your propositions, and now I will answer your fifth and last, asking 
for any suggestions in regard to the whole matter. 
In course of my remarks I have indicated all that now occur to me, 
and even pointed to that to which I now beg leave to ask your particu- 
lar attention—the character of the agents that shall be employed in all 
the stages of this work. 
o , 
About 20 years ago, just preceding the 7th census, the American 
Statistical Association received a'Similar request after they had sent 
their plan of enumerating the population. They answered “they would 
then refer to the men who should do the work, and advised that men of 
exact mental habits, book-keepers, school-teachers, accountants, mathe- 
maticians, naturalists, men who were accustomed to look at and accept 
of facts in their exact measurement and character and to made records 
corresponding.” And they advised further, “to avoid mere politicians, 
and all men of loose mental habits, who look upon facts as raw material 
to be manufactured for other purposes, to be curtailed or enlarged or 
shaped to suit occasions; such men and such habits could not be relied 
* These were proposed bj the late Mr. Lemuel Shattuck, of Boston, and Mr. Archibald 
Russell, of New York, both very eminent statisticians. Also Schedule 6. 12 
NINTH CENSUS OF THE UNITED STATES. 
upon for any work of collection or record, for however honest their in- 
tentions, they had not the power nor the discipline to see things in their 
exact form, still less to remember and represent them always as they 
originally appeared/ 
This advice is still worthy of consideration, for reasons found in the 
history of several of our censuses. 
I will go no further back than the census of 1840. In that, several 
counties in the rural districts of the agricultural States are represented 
as without farms. Five hundred and eighty-nine counties are reported 
as having no person engaged in trade or commerce. Albany, New York, 
a large commercial city of New York, reports 35 engaged in trade, while 
her neighbor and rival, with less than half the population, reports 736 
so engaged. 
In the same census many towns are reported on one page as having no 
colored inhabitants, and on another as having one, two, three, four, five, 
seven, even eight colored insane or blind persons; many others show one-third, 
one-half, three-quarters, and larger proportions, even all of their colored 
people, to be insane, or blind, or deaf and dumb. The manuscript copy 
of the marshal’s returns shows that all the 133 patients in the State 
Lunatic Hospital at Worcester, Massachusetts, are colored, and this is 
reaffirmed in the printed report, which says that out of the 151 colored 
inhabitants of Worcester, of whom only 113 are over 10 years old, 133 
are insane. So, according to that document of the United States, Massa- 
chusetts has the credit or the honor of devoting her only lunatic hospi- 
tal to the exclusive use of the negroes, a statement which would have 
gratified the abolitionists if it had been true. 
In the same census it is shown or stated that of all the colored popu- 
lation in Massachusetts 1 in 44, in New Hampshire 1 in 28, in Michigan 
lin 27, and in Maine lin 14, were insane. Through all the free States a 
similar though less proportion of colored insanity was offered for the 
world’s credulity to believe. 
Mr. John Quincy Adams told me that Mr. Calhoun urged this evi- 
dence of the inability of the colored race to endure the burdens of self- 
direction as a strong reason for the admission of Texas. The south was 
overcrowded with slaves, and if more room could not be found for them 
they must go free, and freedom was their unavoidable danger of insanity; 
“therefore we must have the new lands of Texas, where they may go and 
live with undisturbed brains. Humanity demands it.” 
These discrepancies of the statements, of no colored population yet 
many colored insane in these towns, were not noticed by the marshals 
who first reported, by the clerks who digested their returns, nor by the 
superintending and responsible officers of the census department. More- 
over, Mr. De Bow, in the Compendium of the Seventh Census, 1850, page 
76, reaffirms the same tale, and with much sophistry tries to show how 
it should be so. 
But Mr. Berrien, then senator from Georgia, told me that it was un- 
doubtedly an error; but he added, “ Congress will not correct it—it 
was too good an argument for the south to give up, and there were too 
many speeches based upon this to be made in support of slavery to let 
this be lost.” 
The Statistical Association memoralized Congress, showing the untrust- 
worthiness of the sixth census, and asking, for the honor of our country 
and the credibility of our national documents, that it be revised and cor- 
rected. Mr. Adams had it in charge ; Congress printed the memorial, 
but, as Mr. Berrien said, nothing could be done beyond this. 
The census department was unfortunate in its co-operators in 1850 as 
well as in 1840. Mr. De Bow, tire superintendent, says, in his Compen- NINTH CENSUS OF THE UNITED STATES. 
13 
dium, page 18, “ Each census has taken care of itself. Every ten years, 
some one at Washington will enter the hall of a department, appoint 50 
or 100 persons under him who have perhaps never compiled a table be- 
fore, and are incapable of combining a column of figures correctly. 
Hundreds of thousands of pages of returns are placed in the hands of 
such persons to be digested, in 1840 the returns were given out by 
the job, to whoever would take them. In 1850, such was the pressure 
of the work that almost any one could, at times, have had a desk.” 
Mr. Kennedy, in his frequent correspondence with me in regard to the 
eighth census, spoke sorrowfully of the incompetence of the force that 
was granted to him, and the difficulty he had in getting the work done 
properly. 
That Mr. Kennedy complained with good reason was manifest from the 
works that were sent forth. The ordinary clerks failed in accuracy, and 
there was not a sufficiency of the supervising intelligence to detect and 
correct their errors. In the Preliminary Report, which Mr. Kennedy 
wrote me “ was prepared in haste, and without opportunity for revision,” 
on pages 22 and 23 is a table showing the number of deaths in each State. 
On pages 138 and 139 there is another table showing the same facts in each 
month, and the total of the year in each State. These two statements of 
the same facts agree only in 18 of the States, and disagree in 22. 
Mr. Kennedy sent me manuscript sheets of tables showing the deaths 
from each cause in each State and Territory, in one set of tables classified 
according to age, and in another according to the season and month of 
death. The totals from each were given in connection with each classi- 
fication in every State of the whole country. There were 118 of these 
causes, besides those of violence. 
The number of deaths from 31 causes were the same in these manu- 
script tables as those given in the tables of the same on pages 158 to 161 
of the preliminary report. But those from 87 causes differed. 
In the tables of the violent deaths the numbers given in the manu- 
script agree with those in the printed book, page 167, in regard to only 
three causes, and differ in regard to 10 causes. 
The number of deaths at each of the 18 periods of life, given in pages 
140 and 141 of the preliminary volume, are different in every case from 
the statements of the same facts in the population volume, page xlv. 
The manuscript tables sent to me were prepared in the census office; 
they were presumptively complete, both in their specific and their general 
statements, and ready for publication. These were the sole and entire 
basis of my report. The columns of deaths at each age, in each month, 
and from each cause, were added, and the tables given for each State. 
These would have been accepted as received, and transferred to the 
printer, if no suspicion had arisen. On examination it was found that 
the number of deaths from several of the causes was not the same when 
classified as to age as when classified as to months of death, and those 
of the ages differed from the statements in the printed volumes. The 
whole was then revised, and about COO corrections made. 
So far, I could not go behind the record; I could only use one part to 
correct another. But in the catalogue of causes there were statements 
so contrary to the law of disease, to all experience and probability, that 
I made corrections there. Infants under one year were given as dying 
from drunkenness and from delirium tremens; persons 50 or 60 years 
old dying of teething; persons frozen to death in the extreme south in 
July and August; persons sun-struck in the farthest north in January. 
Finding that the same number of deaths from intemperance and teeth- 
ing was given when classed by seasons as by ages, I concluded that the 
diseases were reported correctly and the error was in the age. I there- 14 
NINTH CENSUS OF THE UNITED STATES. 
fore put all these in the column of unknown age. For a similar reason 
I put thesedeaths from heat and cold into the unknown months. 
I cannotsay where these errors originated. Probably the last might 
have been the misunderstanding on the part of the enumerators of the 
reports of the informants at the houses, and the officer, without think- 
ing of the absurdity and impossibility, recorded it without further 
inquiry. Others must have occurred in the central office. Some of those 
in regard to insanity in 1840 were merely the misplacement of figures, 
the long and narrow columns of white insane and colored insane being 
side by side, and it was difficult for the eye to follow them down from 
the headings. The errors in the addition were from incompetent or 
careless accountants; and it seemed to be but an accident that two or 
more digesting and summing the same facts, under different classifica- 
tions, or even one man doing this at different times, should come to the 
same conclusion and find the same sum in the end. 
Whatever these errors may have been, and wherever originated, they 
passed or should have passed through several revisions, and these by 
several hands. Yet through all this ordeal of security for truth they 
escaped notice, and were sent forth to the government and the country, 
and the world was asked to believe them. 
What credibility these documents, thus imperfect and in some parts 
so self-convicting of error, can gain for themselves or for the authorities 
that present them, you can easily conceive. 
There is a remedy—in the selection of the men to do all the work, from 
the first collection of the facts to the last revision of the printers’ proof. 
You want men of such general intelligence as will understand the nature 
of things, probabilities of the reports or statements, of such acute per- 
ceptive faculties and ready reasoning as will detect at once any error, 
any improbability, and of such faithfulness, that will lead them to pur- 
sue the inquiry until the truth shall be known. The enumerators must 
do this in the first place. The assessors’ assistants must again revise 
and scrutinize each record with the same acumen ; the principal asses- 
sors must again do this; and then competent officers in the census office 
must do the same; and then, through all the stages of analysis, combi- 
nation, tabulation, calculation, inference, and philosophy, no part must 
be taken for granted as true until it has had all the sifting that the best 
intelligence can give it. 
I will beg leave to speak again of the topics of inquiry and the sched- 
ules. It would be well not to fix them unalterably in the law, but 
leave room or liberty to add more by proper authority—say the Secre- 
tary of the Interior and superintendent of census, or the President with 
them 5 certainly some trustworthy authority that Congress may desig- 
nate. 
I have much fear that if the enumeration of the population and 
mortality, and the gathering of all the other statistics specified in the 
schedules, be done at the same time, the work will be embarrassed and 
imperfect. 
1 would suggest for your consideration the question of separating 
these, and of taking account of the people and mortality at one time, 
and the amount of property, business, and finance at another, but all 
by the same officers. 
Very respectfully, yours, 
EDWAED JARVIS. 
General James A. Garfield, 
Chairman Committee on Census, 
House of Representatives, Washington, I). G. NINTH CENSUS OF THE UNITED STATES. 
15 
Doechestee, Mass., Marsh 1,1869. 
Deae Sie : In my letter of Saturday I said I would send you what- 
ever I should have to say on the bill which you sent me. I intended to 
send it this morning, but I have not had control of my time. 
I have now gone carefully over the law, and made my remarks on 
another sheet, which I enclose with this. 
To give you further proof of the importance of employing a class of 
disciplined and absolutely reliable minds in the enumerators, and in all 
who have anything to do with the census—whose errors or negligences 
might long, perhaps forever, escape detection—l beg to ask your atten- 
tion to some omissions which I discovered in some of the former reports 
in 1868. 
I compared the numbers of persons iu each period of life, in 1850, with 
the numbers of the same persons ten years older, in 1860, in order to 
find the loss by death, and some light on the law of mortality in this 
country, its different parts, and among the different races. 
. I compared the numbers who were under 5 years old in 1850 with the 
numbers of the same persons iu 1860, then 10 to 15 years old; those 
5 to 10 at the former year with those 15 to 20 at the latter ; and those 
20 to 30 in 1850 with those 30 to 40 in 1860, and so on through all the 
ages. 
According to the natural and universal law of mortality there must 
have been a loss from age to age, until the last wrould show none to be 
numbered; and, according to the same law, this loss would be the 
greatest in the first period of childhood. 
If there were no disturbing element—if none came in except by birth, 
and none went out except by death, and if all had been counted and 
recorded at each census—the difference, thus ascertained, would show 
the exact number of deaths, and the rate of mortality easily determined. 
If this were done for different peoples and different races, states, or 
nations, the rates, thus deduced, would be good ground for comparing 
the longevity and mortality at different ages of one race with another: 
the white with the colored, or one State or nation with another 5 Ohio 
with Massachusetts; America with Great Britain, &c. 
In making this calculation, I included all the persons, of every age, 
who were reported to be living in the United States iu 1850 ; but in the 
other element of the calculation, the population of 1860,1 excluded, Ist, 
all those who were under 10, who were born since 1850. I excluded also 
all the foreigners, of every age, who had arrived since 1850. Thus, this 
last element included only those who were here in 1850 and had survived 
to 1860. 
In these public reports it was plain that the numbers who were in the 
age 10 to 15 in 1860 represented the survivors of those who were 0 to 5 
in 1850; that those who were 15 to 20 in 1860 were the survivors.of the 
class stolo iu 1850; and those who were 40 to 50 in the latter date 
were the sole representatives of the class that were 30 to 40 years old 
at the former enumeration, &c., through all the ages. 
My hopes of light from this source were disappointed, by finding- that 
these documents—the censuses of 1850 and 1860—represented not a 
decrease as people grew older, but rather an increase in some of the 
early ages. 
Thus, the 7th census states that, in 1850, there were 802,337 males 
under sin the southern States. According- to the law of mortality these 
should have lost at least 16.8 per cent, in the next ten years, and been 
represented by 667,545 survivors, 10 to 15 years old, in 1860. But the 
Bth census says that, in 1860, there were 816,354 males between 10 and 16 
NINTH CENSUS OF THE UNITED STATES. 
15 in those States, showing a gain of 15,017, instead of a loss of 134,792, 
which should have happened. 
In all the United States the whites 0 to 5 years old, in 1850, gained 
almost two per cent, in the 10 years, and showed, in the census of 1800, 
so many more of the age 10 to 15 than they had in childhood. 
If there were any suspicion that these whites in 1860 included any 
who had come into the country since 1850, certainly there can be none 
as to the colored. To them none were added by immigration, none were 
lost by emigration; all their gains were by birth, all their losses by death. 
According to the 7th and Bth censuses there were in this 
country, in 1850, colored children, under 5 284,538 
And, in iB6O, there were colored boys and girls, 10 to 15 ... 294,097 
Showing an apparent gain of 9,559 
Instead of a probable loss of 47,712. 
In these cases, I presume, the error is in the omission of the young 
children in the census of 1850, and that the last, 1860, is more nearly 
correct. Assuming, then, that the reports of the number of boys and 
girls, 10 to 15, in 1860 were true, then, by the law of mortality, there 
should have been 995,951 instead of 802,337 children under 5 in the 
southern States in 1850—an omission of 193,614. And among the 
colored, in all the United States, there should have been 358,798 children 
under 5 in 1850, instead of 284,538, as reported. 
There were other manifest inconsistencies in these reports—all, how- 
ever, in the earlier ages. The statements of the more advanced ages 
had an appearance of probability and even of truth; nevertheless these 
omissions and negligences prove a want of due care and faithfulness on 
the part of some of the officers, probably the enumerators, and cast a 
shade of doubt as to the absolute reliability of all the rest. 
Certainly these offer a strong reason for the government to employ 
the best classes of minds in all the steps of the work—the most clear, 
disciplined, and faithful agents that can be obtained. 
The government must do as men of business in the world, bank man- 
agers, merchants, manufacturers, who, if they succeed in their under- 
takings, employ only such as are sure to do their work in the way 
appointed, and pay them for their talent, skill, and faithfulness. 
Very respectfully, yours, 
EDWARD JAEYIS. 
General J. A. Garfield, 
Washington.