The following article was written by my good friend, William Dollarhide.

If you are a genealogist, you are probably an avid user of the federal censuses, 1790-1940. But even experienced census users may not know about some of the obscure aspects of the censuses. Here’s one of them. It’s called the “census day.”

Beginning with the 1790 federal census — continuing with every census thereafter — each enabling law authorized by Congress specified a census day for gathering the census information from every household in America. From 1790 to 1820, the census day was the first Monday in August.

The census day was not the day the enumerator arrived at a household, it was the day for which all the statistics of the census were collected. The actual instructions given to all the U.S. Marshals right before the 1790 through 1820 censuses explains:

“….all the questions refer to the day when the enumeration is to commence; the first Monday in August next. Your assistants will thereby understand that they are to insert in their returns all the persons belonging to the family on the first Monday in August, even those who may be deceased at the time when they take the account; and, on the other hand, that they will not include in it, infants born after that day.”

The above quote was part of the first set of instructions, originally written in 1790 by Thomas Jefferson, Secretary of State; and directed to the U.S. Marshal in each federal court district. For the next three census years, 1800, 1810, and 1820, the text above was repeated exactly. Similar instructions have been given for every census since 1820, but with different census days. The table below shows the census day for each census, 1790-1940, and the time allowed to take the census:

Census Year Census Day Time Allowed
1790………..2 Aug………9 months
1800………..4 Aug………9 months
1810………..6 Aug……..10 months
1820………..7 Aug……..13 months
1830………..1 Jun……..12 months
1840………..1 Jun……..18 months
1850………..1 Jun………5 months
1860………..1 Jun………5 months
1870………..1 Jun………5 months
1880………..1 Jun………1 month
1890………..1 Jun………1 month
1900………..1 Jun………1 month
1910……….15 Apr………1 month
1920………..1 Jan………1 month
1930………..1 Apr………1 month
1940………..1 Apr………1 month

Census Day Differences
On the above table, note that the census day changed from the 7th day of August in 1820 to the 1st day of June in 1830. If one is researching families appearing in the 1820 and 1830 censuses, looking at these families again may be important. Since the census days for 1820 and 1830 are not exactly ten years apart, the two-month difference may reveal some surprising results.

For example, if a child indicated in a family were born between 1 June 1820 and 7 August 1820, that child would appear in the 1820 census in the “under 10″ category. But in 1830, that same person would appear in the “of 5 and under 10″ rather than the “of 10 and under 15″ category, since the person had not turned 10 yet. Comparing the 1820 age categories for a person appearing ten years later and not in the “correct” age category may actually give a clue to a person’s date of birth within a two-month period.

Other differences occurred from 1910 to 1920 and from 1920 to 1930. Both reflect a difference of about four months, and both will impact the age of certain persons listed. Knowing the census day changed between certain censuses should help explain the slight differences in some ages reported from one census to the next. (But, unfortunately, it will not explain why some women aged only five years in the ten years elapsed between censuses).

Time Allowed to Take a Census
On the above table, note the time allowed to take each census. All of the states complied with this provision, except South Carolina in 1790. South Carolina could not complete its 1790 enumeration in nine months. The U.S. Marshal complained that he was having great difficulty finding people to take the job because of the resistance to the census being taken. A Charleston jury met to decide the fate of six persons who had “refused to render an account of persons in their households as required by the census act.” A South Carolina census taker was brought on trial for neglect of duty. He did not complete the census in his district. These and other problems led to South Carolina being granted a nine-month extension by Congress. The last of the South Carolina census returns were dated 5 February 1792, a full eighteen months after the census day.

Note that the time allowed for the 1840 census was eighteen months, while the 1850 census was allowed five months. That was quite a change. But, the reason for the change was that in early 1850, the first U.S. Census Office was visited by a couple of young inventors from upstate New York with a new “counting machine.” The Census Office loved it, bought one, and determined to use more machines to compile the 1850 census statistics. After a trial run, they decided they could complete the work in five months, which was the amount of time they requested from Congress. With a contract from the Census Office, the two inventors formed a company in Poughkeepsie – which later became known as IBM.

Differing Census Days
In a couple of cases, there have been census days assigned to certain states that were different than the rest of the U.S. for that year. When Vermont entered the Union as the 14th state in 1791, the 1790 census was already underway. Vermont’s “1790” census was taken with a census day of the first Monday in April 1791, with five months allowed to take the census there. And Utah, which became a territory in September 1851 had an “1850” census taken with a census day of 1 April 1851. The visitation dates on the Utah census pages are mostly in October 1851. Thus, the 1851 census enumerators probably asked Utahans questions like, “Six months ago, back on April 1st, who was the head of this household?”

Census Day vs Enumeration Date
Genealogists should record two dates when copying information from the censuses: the census day and the enumeration date. No matter how many months it took for an enumerator to reach a house, he was supposed to gather the information as if time had stopped on the census day. Every person whose regular abode was in a particular household on the census day was to be enumerated, even if a person were away at the time of the enumeration.

Understanding the impact of the census day versus the enumeration date may explain why certain people appear in a census listing, even though you have other evidence to show the person died before the household was enumerated. If a person were alive on the census day, that person was to be included — even if it took some time for the enumerator to get around to the house to take the census. The person could have been dead for several months.

Or, you may wonder why that youngest child in a family was not listed in a census. If a child were born after the census day, that child was not to be included — even if the census taker had visited the house and was aware of a playful little toddler crawling around in front of him.

Now, some of you may have to go back to all of those census lists you have copied down and confirm the date of enumeration AND the census day. Any missing people? Any extra people?

For more information, see: