At Library and Archives Canada (LAC), we receive questions about why materials in our collections are organized the way they are.

When it comes to the census returns, typically we explain that, as an archive, we acquire the census returns as they are—even when the writing is blurry or unreadable—as historical records. We also strive to maintain the records’ original order and context.

For the 1931 census returns, maintaining the records’ original order is not difficult. Rather than receiving 234,678 pieces of paper, we received 187 microfilm reels. On the microfilm, the imaged census returns are organized, overall, by province (east-to-west) and then by northern area (west-to-east). This is because the Dominion Bureau of Statistics imaged the census returns in order of census district number, and, within each census district, in order of census sub-district number. When we digitized the archival records here at LAC, we worked to ensure that the digital access copies would reflect the archival records’ original order and context to the extent possible. For instance, we grouped digitized images according to the title cards used in the microfilmed images. To each group of digitized images, we added additional metadata, extracted from the Dominion Bureau of Statistics’ listings accompanying the census returns.

Example of a title card used to organize 1931 census returns on microfilm. This title card is for the eight pages of returns from Prince Edward Island, census district 3 (Queens), sub-district 10 (MIKAN 5744023)

But this explanation does not answer the original question: why are the returns from the 1931 census of population organized geographically—by census district and sub-district—in the first place?

For answers to this original question, we turn to the Administrative Report included in the Dominion Bureau of Statistics’ 1936 publication Seventh Census of Canada, 1931, Volume 1, Summary.

Answer 1: Because of the original purpose of the census of Canada. The census returns are organized geographically because of the decennial census’ role in shaping representation in the House of Commons.

“In Canada the immediate, legal raison d’être of the census is to determine representation in the […] House of Commons. Under the [Constitution], the province of Quebec is given a fixed number of seats[…] while the number assigned to the other provinces is pro rata on a population basis as determined by the census[….] The Canadian Census is thus taken primarily to enable a redistribution bill to be passed through Parliament” [Page 32; emphasis added]

It is worth getting a bit technical here. In the early part of the 20^th century, redistribution bills updated the number and boundaries of federal electoral districts based on changes in population—as established by the previous decennial census—among other considerations established by law.

“Map of Federal electoral districts of Manitoba, Saskatchewan & Alberta” from an atlas created in 1924, prepared by the Department of the Interior (e011315903)

The updated boundaries of federal electoral districts informed the census districts used for enumeration purpose in the subsequent decennial census. In other words, the relationship between federal electoral district and census district had elements of a chicken-and-egg scenario.

Egg: Federal electoral district boundaries, as established in the Representation Act of 1924.

The Representation Act of 1924, sometimes referred to as the Redistribution Act, established the electoral districts to be used in subsequent federal elections. This redistribution of electoral districts was based, in part, on the population count and distribution established by the previous decennial census, namely, the Sixth Census of Canada in 1921.

The Representation Act established the official boundaries of federal electoral districts in form, not in maps.

• To view transcriptions of these descriptions of electoral districts, consult the Library of Parliament’s online resource “Elections and Candidates” for the 17th parliament; the description of each electoral district can be viewed by clicking on a constituency title listed for the general election of July 28, 1930, and then scrolling below the header “Information” to the description under the subheader “S.C. 1924, c.63,” which refers to the Representation Act.
• Although the Representation Act did not include maps, the Department of the Interior prepared an atlas with maps depicting the updated federal electoral district boundaries. To view digitized images of the maps in the 12-volume atlas, there are two options. The catalogue record for the Federal electoral district maps, 1924 displays thumbnail images. A 1931 Census Maps research tool supplies links to individual, higher-resolution maps.

Chicken: Census districts used for 1931 census.

1931 census districts used in enumeration mostly corresponded to the federal electoral districts established in the Representation Act of 1924, because

“[f]or the purposes of the census, the Statistics Act requires that the country be first divided into “census districts” corresponding as nearly as possible with the federal electoral divisions or constituencies for the time being—this in view of the association of the census with parliamentary representation.” [Page 51]

Nevertheless, at least eight electoral districts were “too large or too varied in physical or economic character” for the purposes of census work, and so each was split into two or three census districts (in Quebec, Charlevoix–Saguenay, Gaspé, Labelle and Pontiac; in Ontario, Port Arthur–Thunder Bay; in Alberta, Peace River; and in British Columbia, Cariboo and Comox–Alberni). Other areas to be enumerated fell outside a federal electoral district (e.g., Northwest Territories, Royal Canadian Navy ships).

• To view a listing of census districts used for the 1931 census (minus ships), consult 1931 Census Maps research tool.

The population established by the Seventh Census of Canada in 1931 was then used to inform the next redistribution of federal electoral districts.

Egg: Federal electoral districts, as established in The Representation Act, 1933…

This chicken-and-egg scenario is why, at LAC, we often use federal electoral district maps as working stand-ins for census district maps to help us navigate the census returns from the early part of the 20^th century. To navigate the 1931 census returns, we use maps depicting the federal electoral districts as established in the Representation Act of 1924. In the year 2028, we will likely be using the maps depicting federal electoral districts according to The Representation Act of 1933 to help us navigate the 1936 census returns from the prairie provinces after those census returns are transferred to LAC.

Answer 2: Logistics was another reason why the census returns were organized geographically.

The 1931 Census of Canada, like most population censuses, was meant to enumerate each person within the boundaries of the Dominion of Canada once, and only once. For this purpose, land within the borders of Canada, and navy ships, were divided into 15,167 units of enumeration. Each geographical unit of enumeration—called a census sub-district—was assigned to a single enumerator (more or less), tasked with recording each person residing within the census sub-district.

Enumeration proceeding in 1961: An R.C.M.P. member talks to three people from an Inuit community to collect census information (e011177562)

“The census enumerator is the only census official coming into direct contact with the general public; [s/he] is who makes the house-to-house and farm-to-farm canvass and who is primarily responsible for the details collected on the census schedules. The necessity of providing that no more or no less than a suitable amount of work should be assigned to each enumerator (experience has demonstrated this to be a population of 600–800 in ordinary rural districts, and of 1,200–1,800 in urban) […] renders departure necessary in many cases from the electoral boundaries; […] and the polling subdivisions are not always convenient as census sub-districts. In all such cases, however, the division is effected in a way that permits compilation of the results in the form required for the purposes had in mind by the Act.” [Page 51]

Establishing the geographical boundaries to be used in the census was no minor feat:

“The drawing up of the scheme of census districts and sub-districts is a task of considerable magnitude; it is put in hand about two years in advance of the census date, and is carried out not only in the light of conditions revealed in the preceding census, but in consultation with local officials, so that no inhabited area may be overlooked or left unprovided with the organization best suited.” [Page 51]

Ninety-two years later, the task of finding a particular census return among the 1931 census returns that are organized geographically—organized according to census district and sub-district—may seem a little overwhelming. The blog posts Puzzling through 1931 Census sub-districts – Part 1 and Part 2 describes the approaches we use at LAC to navigate the 15,167 sub-districts used for enumeration in the 1931 census.