1921 Census Of England & Wales

The information given for a person varies according to, for example, their age and employment status. Typically, though, the 1921 Census records gives at least the following information about an individual:

• Name
• Relationship to the head of household
• Age in years and months
• Marital status or details of orphanhood
• Place of birth
• Schooling or occupation
• Address
• Who they were living with

Depending upon an individual's circumstances, the census return could also show other information such as:

• Nationality
• Employer and place of work
• Number of minor dependants (children and step-children up to the age of 15 years)
• Language spoken

For a typical household return, you will see a household members table, giving names and key details of individuals in the order in which they appear on the original census schedule. Where there are six or more individuals in the household, only the first five are displayed by default; if you click on “Show more rows”, the household table will expand to display the sixth and later individuals in the household on census night.

Beneath the household members table, you will see a full transcription for the individual whose row is highlighted in the table. Simply click on a different person in the household members table to highlight them and switch to their full transcription.

Our transcriptions include nearly every field shown on the original household schedules. This includes the information, newly requested in 1921, about orphanhood, schooling and employer.

Only two fields have not been transcribed, meaning you would need to view the image to see them. The first of these is the room count shown towards the bottom-left of the back of a household schedule. This is the count of living rooms – in other words, bedrooms, dining rooms, kitchens, parlours, sitting rooms etc; it excludes small functional spaces such as bathrooms, laundry rooms, sculleries and toilets. The second is the minor dependency grid which appears to the right-hand edge of the back of a household schedule. The dependency grid has boxes for ages between 0 and 15, into which the householder was supposed to enter an X (or two, in the case of twins, for example) for each child or step-child, whether or not they were at home on census night. This data was supposed to be entered in the row for each married man (not woman), widow or widower in the household. These instructions proved difficult for householders to comprehend and apply, and in many instances you will see, on the original images, how enumerators or Census Office officials had struck out householder mistakes and re-entered the crosses themselves. As a result, the grid is often untidy and unclear, and was not transcribed. In addition to these two fields, we have not transcribed the measure of housing density which may usually be seen as a handwritten single-digit number just to the left of the Schedule No box to the top-right on the back of a schedule.

The transcriptions also include administrative geographical and archival information extracted from elsewhere within the census.

If you scroll down below the full transcription, you will see further content. This may include, for example, a map of the local area, a gazetteer description, a link to a newspaper from 1921 and some infographics about the Registration District in which the household was situated. One thing you might well see in the infographic is that the Age Breakdown bar chart shows an excess of females over males in the 20-29 and 30-39 age group cohorts especially, due to deaths of men of fighting age in the Great War.

When you click through to view an image for a record of interest, for a typical household schedule you will be able to see the following set of images:

• The back of the household schedule, showing details of household members, as completed and signed by the householder

• The front of the household schedule, showing the address, as completed by the enumerator, and printed instructions on how to fill in the form

• The front cover of the volume into which the household schedule was bound, giving geographical information about the area in which the household was situated

• A description of the Enumeration District in which the household was situated, often in the form of an enumerator’s walk

• A map of the Registration District in which the household was situated (often two images of the same map, one with an attachment lowered and one with it lifted)

Note that there are maps only for England and Wales; not for Channel Islands or Isle of Man.

When you click through, you will land on the first image of those listed above – the back of the household schedule. To see any of the other images, click on Extra Materials.

You can also use the thumbnail images at the foot of the image viewer to flick through the filmstrip. The landing image (which has a highlighted border and is usually in the middle of the thumbnail strip) is the back of the household schedule. Usually the one to the right of your landing image will be the front of the same schedule, showing the address. Beyond that, you are likely to be taken to a different household, which means you will encounter the paywall. An exception to this is the extended schedule (e.g. types coded EE and WW) in which the household schedule comprises four pages and therefore four images – these are sequenced in the order back page 1, back page 3, back page 2 and front. If there are no more than 20 individuals in the household, you should be able to see all four images without encountering the paywall.

The 1921 Census of England & Wales was taken on the night of Sunday 19th June 1921. The Census was expected to be fully comprehensive, and endeavoured to enumerate all households, institutions, armed forces at home and overseas, and merchant navy and fishing fleets in port on census night or arriving in port in the days immediately after. As is customary, the 1921 Census of England & Wales also includes the so-called “islands in the British seas”: the Channel Islands and the Isle of Man.

The 1921 Census includes the original census schedules, as completed in 1921. It also includes Plans of Division, which scoped how the census would be taken and divided up the country into manageable units called Enumeration Districts which could be covered by a single enumerator.

For the 1921 Census, there are no surviving Enumerator Summary Books (as seen in the 1911 Census) and no Census Enumeration Books (as seen in all the decennial censuses from 1841 to 1901).

Read on to find out how the census was taken in the summer of 1921 and how Findmypast brought it to life 100 years later.

How the census was taken in 1921

The Census Act of 1920 put the census of England & Wales upon a permanent footing and enabled not just the 1921 Census but all future censuses to be taken. Before then, the census apparatus had had to be created anew before each census could be taken. From now on, the Census Office was a permanent part of the General Register Office, which was based in the north wing of Somerset House on the Strand in London. For 1921 Census operations, the Census Office took over the old Lambeth Union Workhouse in Prince’s Road (which, if you wish to find it on maps, was renamed Black Prince Road in 1939).

One of the first steps in the taking of the census is to divide up the country into manageable units – Enumeration Districts – each of which could be covered by a single person – the enumerator – in the course of a day. In urban areas, this might mean 300 or more households in just a handful of neighbouring streets. In rural areas, however, the enumerator might be expected to cover, on foot or by bicycle, a route of, say, five miles, visiting all the hamlets, isolated cottages and farmsteads within their Enumeration District.

The position of enumerator was a short-term contract. Enumerators were engaged for only for a period of a few weeks before and after census night. Most would have been teachers, office clerks and other literate and numerate individuals. Others were ex-servicemen – and of course in 1921 the teachers and office clerks could well have been veterans of the Great War too. Note that female enumerators had been engaged for each census since the 1891 Census.

Before census day, the enumerator had to identify all the households within their Enumeration District and address a census schedule (the familiar census return or form) for each one. A household was defined as an independent unit. For example, if there was a family living upstairs and a family downstairs in a two-storey terraced house, each was a separate household and needed its own census schedule. A boarder – an individual who received both bed and board within the family – was part of the household in which he or she boarded. On the other hand, a lodger – an individual who received bed only, and ate independently of their hosts – constituted their own household and required their own census schedule.

Note that it was possible, for the first time, to request a confidential census return. This would typically be asked for by someone within a household or institution who didn’t want their details, for whatever reason, to be visible to the head of household or chief residing officer. If an application for a confidential return was accepted, the individual would be given a regular census schedule to fill in and then pass it in person or by proxy to the enumerator, in a sealed envelope if they so wished. The schedule itself was indistinguishable from any other. Examples of confidential returns – identifiable by the associated allowed claims cards – can be found in ED 14 of piece RG 15/16867 (for Bucklow Registration District) – these appear to be for certain staff working at the David Lewis Colony at Marthall, outside Knutsford. Note, however, that the instructions to enumerators state that the name of the confidential person is to be completed on the main household schedule by the head of the household, and the rest of their details added by the enumerator in red ink, from the confidential census return, after receiving it back from the head of household. Therefore, if you see not the name but the rest of an entry completed in red, this should indicate a confidential request.

On the front of each census form, the enumerator wrote the name and address of the householder responsible for completing and returning the schedule. The enumerator also had to write the Registration District, Registration Sub-district and Enumeration District number. Some enterprising enumerators seem to have invested in bespoke rubber stamps to spare themselves the tedious task of writing out the same text hundreds of times.

The boundaries of the Enumeration Districts had been set earlier and defined by the so-called Plans of Division. An enumerator was appointed for each one – there were 38,563 EDs in total – and supplied with a stack of blank schedules believed to be appropriate to the task at hand. Then, between 11th and 18th June 1921 the enumerator went round his or her Enumeration District and distributed the census schedules according to need. A household in England with a headcount of up to 10 would receive the standard schedule type code E form. A larger household of 11 or more occupants would be given the next size up, being the schedule type code EE. However, if an enumerator ran out of Es, EEs might start to be given out to smaller households (and vice versa) – distribution of schedules was, in practice, both by need and by availability.

Each schedule was folded in four, so that the address panel appeared on the front. This may have been done for ease of posting through letter boxes, and/or convenience of carrying around the district if enumerators handed them to householders personally, while giving instructions on how to fill in the form, answering any questions and reminding them when he or she would return to collect the completed form.

The householder was responsible for writing in the personal particulars of each individual staying overnight in the household on Sunday 19th June 1921. You will see that this is not the date printed on the census schedules (which is 24th April 1921). The census was postponed from its originally intended April date because of industrial action. Coal miners were on strike and there was the prospect – later averted – of a national strike (involving transport workers) which led Lloyd George to call a state of emergency on 5th April 1921, which mobilised the Army Reserve and a new Defence Force from 8th April. The Census Office did not want to delay the taking of the census beyond June. If it were to do so, it would run into holiday season, which would distort the picture of population distribution. Whole factories in towns such as Bolton and Salford might close and workers decamp to seaside resorts such as Blackpool or Prestatyn, so producing misleading figures for the statisticians. Similarly, university students would break up for the summer vacation, and harvest season with its associated migrant labour flows would begin. Accordingly, unwilling to postpone further, the Census Office pressed ahead with the 1921 Census on 19th June 1921. The millions of forms printed for the intended April census were used rather than pulped.

You can give a person very precise instructions, both printed and verbal, on how to complete a form and he or she will still do it differently. This is the case with the 1921 Census, just as it must have been for all censuses before and after. For example, there is an instruction to fill in the form “in Ink”. Some householders did complete their schedules in blue or black ink, using fountain pens or dip pens; but others used indelible or laundry pencils, and others graphite or coloured pencil. Some householders misunderstood what was required in recording number and ages of children and in which row (it should have been against a husband and not additionally against his wife, for example). The child dependency grid on the right-hand side of the back of the schedule, where an X was meant to be entered against an age for each child and step-child up to and including 15 years, is often a mess, and where messy often redone by a clearly irritated enumerator or Census Office clerk.

From 20th June, the enumerator returned to their Enumeration District and started collecting the completed census schedules. You can imagine that this would not have been a single problem-free trip. Some householders would not have been in when the enumerator knocked. Some might not have been co-operative. Some might have filled in the schedule incorrectly, or have questions in need of answers, or concerns about the uses to which the asked-for information would be put – would it be used for conscription, for example? So we can imagine the enumerator having to make repeated excursions around their Enumeration District before they had gathered up all the schedules. At this point, the enumerator sorted the census schedules into a tidy and sensible order and numbered them, if they had not already been numbered, in the Schedule Number box on the back of the form.

By 27th June 1921, the enumerator should have checked and bundled up the census forms and taken them to the Sub-district registrar, who would be expecting such deliveries from all the enumerators for the Enumeration Districts within his Registration Sub-district. Further checking was done at this point and queries raised with enumerators, and also by enumerators with some householders at the instruction of the registrar. Only then was the Sub-district registrar comfortable binding and sending all the schedules for all their Enumeration Districts further up the chain, where they reached the Census Office in Lambeth.

The Census Office had a workforce of 550 at its peak in August 1922. This was made up of a relatively small number of permanent employees and a much larger number of contract workers brought in for the enormous task of processing the schedules. Like the 1911 Census before it, the 1921 Census used sophisticated mechanised data-processing instead of the old clerical “tabling and ticking” paper-based methods of the 1841 to 1901 Censuses. The largest cohort of staff employed were the 202 punchers, or punch-card girls. These were 15- to 17-year-old young women with manual dexterity who operated the machines that punched holes in tabulation cards to record coded information about each individual. These cards then went through separate sorting and counting machines to produce the tabulated raw data of interest to the statisticians and demographers of the Census Office. 37,886,699 cards were produced – one for every person enumerated in the 1921 Census.

Before the census schedules reached the punch-card girls, however, each one was examined by a clerical officer who made annotations in a distinctive green ink. Presumably, green was chosen as the available colour it was thought householders were least likely to use when filling in their forms. Principal among the green-ink annotations is the coding of occupation and employment, but you may see others in different parts of the forms. In particular, on schedules completed in the Welsh language, clerical officers, presumably using Welsh-English language dictionaries, went through and translated text into English, again using their distinctive green to do so. For example, you are likely to see translations of Welsh-language occupations and kinship terms.

The physical processing of the 1921 Census went beyond summer 1921 until at least 1923. The Census Office produced a variety of reports on their findings, beginning with a high-level summary Preliminary Report of 23rd August 1921 (just two months after the census was taken!), a series of county-level volumes dating from October 1922 to March 1924, thematic volumes in 1925, and finally culminating in a definitive General Report in 1927. Most of the temporary staff, such as the punchers, would have been laid off when the job of punching, sorting and tabulating was done, and presumably the permanent staff would then have vacated the former workhouse in Lambeth and gone back to the General Register Office’s headquarters at Somerset House. The original census schedules, bound into hard covers by the Sub-district registrars and making up more than 28,000 volumes, then went into storage. However, the Plans of Division were revisited in the early stages of the preparations for the ill-fated 1931 Census (which was destroyed in 1942 in a fire started accidentally by a dropped cigarette). The Census Office tweaked its administrative geography for 1931, especially the size and numbering of Enumeration Districts, in line with changes in population density and distribution. This is why you can see annotations on the Plans of Division dated all the way up to around 1928. The unfortunate thing for us is that at that time officials often struck out text that was valid in 1921 and overwrote it with updated text for 1931. As most of these textual alterations are not dated, it is not always clear what is a genuine correction dating to 1921 itself and what a subsequent change for 1931.

The surviving 1921 Census of England & Wales has none of the Census Enumeration Books which family historians are used to seeing in the decennial censuses from 1841 to 1901. We understand that Census Enumeration Books were not used in 1921 (nor in 1911). Neither does the 1921 Census have the Enumerator Summary Books of the 1911 Census, which give just the head of household – these certainly were compiled (an unused example may be found at the back of piece RG 15/24428) but they have not been retained. Instead, the surviving 1921 Census of England & Wales comprises the original household schedules (as in the 1911 Census) and the Plans of Division which enabled the Census Office to conduct its work with the maximum of efficiency during the long summer of 1921.

How Findmypast digitised the census in 2019-2021

The 1921 Census does not appear online at Findmypast as if by magic. It was three long hard years in the making. Over 28,000 volumes in boxes on shelves were converted into indexed digital images for researchers to enjoy. How did that happen? These are the steps we took in processing the physical material so that it could be published online.

As just mentioned, in the summer of 1921 enumerators distributed the blank schedules – the census forms to be completed by each household in the country – to all addresses before census night and then collected them, filled in by householders, the week after. As with the 1911 Census, it is these original householder returns from 1921 that have been kept and survive today. The enumerators checked and processed the schedules, and then passed them on to the registrar of their Sub-district who, in turn, bound them into hardback volumes and packaged them up for the Census Office in London. Census Office clerical staff went through each and every census return, annotating them with occupational codes and sometimes making revisions in their distinctive green ink. A task force of nimble-fingered young women, mostly aged between 15 and 17, known as punchers, extracted information about each person on each census return on to tabulation punch-cards. The dextrous punch-card girls also undertook basic first-aid on damaged schedules, re-attaching detached sections. Comptometer operators then fed the punch cards into sorting and counting machines, which digested them and spat out the raw data that was to be worked up into the statistical abstracts that the government wanted to understand the population and develop policy.

After that the 1921 Census was closed by law, under the Census Act of 1920, for 100 years. Some parts of it were in fact used internally by the Census Office itself in the late 1920s to prepare for the 1931 Census (sadly now lost), and limited access was granted to academics on rare occasions. Other than that, though, the bound census returns sat in government department storage, occasionally moving from one place of deposit to another. For a while they were held at Somerset House on the Strand; later, in the basement of Audit House in London; and eventually at Christchurch in Dorset. They were shelved standing, as a result of which there was a tendency for papers to flag a little under their own weight. At some point prior to 1976, a part of the collection suffered from water damage, followed by mould and booklice infestation, which were treated according to the standards of the time. Other than that misadventure, essentially the 1921 Census sat in suspended animation for 97 years, patiently waiting to come of age.

What happened next was that the 1921 Census was brought back to life.

The books were placed in archive boxes and transferred to a secure government location in Titchfield, Hampshire for digitisation. There, under the watchful eyes and careful guidance of the Office for National Statistics and The National Archives, Findmypast set about the immense task of bringing the 1921 Census to its public.

Digitisation is the umbrella term we use for imaging and indexing a record set so as to make it publishable in fully-searchable form online. For the 1921 Census, though, there was an additional process which preceded the imaging and indexing – conservation. As we were moving into an empty workspace, devoid of everything except lights and power points, first of all we had to set up from scratch a fully-equipped conservation studio. This meant all the necessary work-benches, stools for when sitting and fatigue mats for when standing at bench, temporary shelving for holding batches of materials being worked on, trolleys and roll cages for transporting them, tools and aprons and personal protective equipment. We designed a workflow to optimise the available space – in other words, worked out how best to accommodate not just the many conservation team members as comfortably as possible but also how the archive materials would travel and be tracked around the space from one stage to the next in the conservation process.

We then carried out an audit or stock-take, to make sure that all the boxes we were expecting were in the studio, and that all the volumes were in the right boxes, and nothing that was in-scope was missing, and nothing that was out-of-scope was present. To give an idea what this entailed, in just the main record series RG 15, which contains the collection of census schedules, there were a little under 14,000 boxes. Each one of those boxes had to be located on its shelf, barcode-scanned, opened, its contents checked and then re-shelved. 28,162 pieces were found to be present and correct, including one or two which had been thought missing but which showed up bound into the back of other volumes. We mapped the stacks so we knew exactly which boxes were on which shelves on which bays and could confidently go straight to the one we wanted.

Only once the studio was ready and the audit had been completed could we commence conservation work.

In the context of the 1921 Census, conservation means conservation for digitisation. This is an approach to paper conservation that focuses on preparing and stabilising the documents for imaging, rather than on full repair and restoration. The objective is to be able to get the very best possible digital image from the paper document. In practice, what this meant was that for the main RG 15 record series we undertook the following conservation tasks, following The National Archives’ guidelines.

• The straps holding each bound volume closed were untied or removed, depending upon their type and condition.
• The interior bindings on each book, which bound the individual census returns into the volume, were cut.
• Any loose surface dust and accumulated debris, such as chads (rounds of card punched from punch-cards), were carefully brushed from cover boards and from between schedules.
• Foreign objects between the pages, or pressed onto pages, including squashed invertebrates from the summer of 1921, were removed.
• Surface dirt was delicately cleaned using smoke sponges.
• Metal fastenings, such as pins and staples, were extracted.
• Edges were flattened and any folds that could conceal text, such as the householder signatures and schedule numbers towards the corners, were gently turned back using a teflon bone folder or heated spatula.
• Points of weakness along the main folds (created when the schedules were folded in four to fit through householder letter-boxes in June 1921) were inspected to ensure they were not likely to split during imaging.
• Minor damage, such as straight tears, was repaired using Japanese tissue (tengujo).
• Unsatisfactory historical repairs, especially those obscuring text, were carefully undone where it was safe to do so and the document then professionally repaired.
• Particularly damaged or vulnerable schedules were inserted into archival clear polyester sleeves – 87,000 of which had to be made bespoke for the project by specialist conservation suppliers, because of the outsize nature of the census schedules.
• Schedules were inverted, or placed in sleeves, if the normal leading edge was frayed, so as to provide a clean feed-edge for the scanner.
• Detached sections and fragments of schedules were reunited with their other parts wherever possible.
• Historical inactive mould was cleaned using museum vacuums in fume cabinets in a contamination room.
• Blocked pages (those horribly stuck together by mould and damp) were carefully separated and pages consolidated (reconstructed) as far as possible.

The work was a combination of heavy lifting (each box weighs 10kg and the racking was six shelves high, requiring the use of safety steps) and meticulous painstaking attention to detail. Our team of qualified conservators and enthusiastic conservation technicians with interests in archives, research and history worked on this project between January 2019 and October 2021. They had to contend not just with the massive task of preparing the census materials for digitisation but also with Covid-19 and social distancing. We thank them all for their invaluable contributions to the making of the 1921 Census:

Alex Cuschieri, Amy Thacker, Beatrice Brotto, Becky Dabnor, Bryn Lloyd, Carina Rosas, Catt Thompson-Baum, Charlotte Mace, Eleanor Towell, Elisa Schlarp, Ellie Smith, Eloïse Lovejoy, Emily Briffett, Frances McMullan, Hattie Dixon, James Pride, Jo Thompson-Baum, Katherine O’Kane, Lucy Daish, Lucy French, Maddie Leishman, Rachel Rhodes, Rachel Roberts, Rebecca Hayward, Rebecca Merrifield, Sarah MacLean, Sarah Melluish, Sophie Goode, Tanya Nakamoto & Vikki Groom.

All of us in the family history world owe them a debt of gratitude for their tireless work in making possible the publication of the 1921 Census.

For workflow reasons, conservation started several months before imaging. This is because conservation is necessarily a slower process – if the two processes were to start at the same time, the imaging team would catch up and not be able to work at optimal capacity. For a period of several months, therefore, the conservation team worked to build up a buffer – thousands of prepared volumes – sufficient to protect them from being caught up by the imaging team. Only when the buffer was in place did the imaging team begin its work.

In due course, each box of volumes prepared by our conservation team was collected by our partners in the imaging team. The requirement of the project was to create an exact digital surrogate, or facsimile, of the paper original, so that, if they wished to, researchers would be able to experience the census almost as if handling the original volumes, as far as that is possible online.

This meant imaging everything within the collection. And everything meant everything. As well as the actual census returns, we imaged the cover boards. We imaged all the blank pages. We imaged ephemera from the Census Office of 1921 which had inadvertently been left in the books, including internal office memos, the occasional punch-card, private notes passed between the punch-card girls, tram tickets… We photographed the stubs of pencils and petrified rubber thimbles, spent matches and roast chestnut shells. Some crushed and very faded flowers. A Belgian 5-centimes coin dated 1910. Everything in each volume was imaged from cover to cover. Over 18 million images were produced.

Two different technologies were used to create images:

• Sophisticated duplex feed scanners were used for those schedules which were in sound condition or, if showing vulnerabilities, had been placed in transparent sleeves by the conservation team. These machines scanned both sides of a document at the same time.
• Rostrum-mounted overhead cameras were used for cover boards, the most fragile or damaged schedules and schedule fragments, and the found objects.

Every image so produced was then individually quality-checked within the studio itself to make sure it was complete, fully in focus, not skewed and, in short, in every way as true as possible to the original document.

We would like to thank all members of the scanning and image QA team and in particular Daisy Owens, Emily Skelton, Hannah Kennedy and Suren Abrahamyan for their excellent work.

Following imaging, our conservation team members re-assembled each volume. This involved removing any polyester sleeves and any paper slips which had been used to flag vulnerable schedules. All the schedules were then checked to ensure they were the right way up and right way round, and were made neat and flush, before the cover boards were retied around them and the volume re-boxed and the box re-shelved.

The boxes were then transferred to a deep storage facility, where they are now archived in ideal environmental conditions. Each page in each book in each of those boxes now has its digital twin, and it’s these faithful counterparts which are published on Findmypast and will be available for use by researchers across the world for the foreseeable future.

These records conserved, imaged, transcribed and published by Findmypast come from three separate original archive series at The National Archives. These are:

• RG 15 – the main collection of 1921 Census schedules
• RG 18 – the maps of Registration Districts
• RG 114 – the Plans of Division

In respect of all three series, please note that copyright applies as follows:

© Crown Copyright. Images reproduced by courtesy of The National Archives, London, England.

https://www.nationalarchives.gov.uk

The National Archives give no warranty as to the accuracy, completeness or fitness for the purpose of the information provided.

Images may be used only for purposes of research, private study or education.

Applications for any other use should be made to:

The National Archives Image Library
Kew
Richmond
Surrey
TW9 4DU

Tel: +44 (0)20 3908 9131

Email: image-library@nationalarchives.gov.uk

The 1921 Census was closed by law for 100 years under the Census Act of 1920. It is now an open public record. All records are open by default. We hope that centenarians will be excited to see their own records in the Census. However, if a centenarian does not wish their record to be open, they can request its closure. Their individual record would then be suppressed and it would not appear in search results, and the relevant row on the image would be redacted with a black-out bar. Requests can, of course, be made upon behalf of centenarians by their next of kin.

To request the takedown of the record of a surviving person who does not wish his or her 1921 Census entry to be public, please follow these steps from their transcription page:

1. Highlight the relevant individual’s record in the household members table at the top of the transcription page, so that his or her full details appear below
2. Click on More Actions (above the household members table)
3. Click on Close Record (the third option)
4. Complete the form requesting the takedown and click on Submit

Alternatively, if you have not viewed the transcription and do not wish to, you can email our customer services team direct at the following address: support@findmypast.com. Please use “1921 Census takedown request” in the subject line of your email.

Firstly, note that you do not need to enter a name to undertake a person search.

There are many search fields which you can use, but this is not an invitation for you to fill in every field on every search! Please don’t, as it is unlikely to give you the results you want. The general rule when searching record sets on Findmypast is to start with little information, see how many results are returned, then refine your search by adding search terms in one or more extra fields. In this way, particularly if you are looking for a person with a common combination of names, you can gradually narrow down the range of candidates.

There are two main types of search field on the 1921 Census search screen.

1. Free text search fields. Examples: last name; occupation. In these search boxes, you can enter any search term you wish and search results will be returned where there is a match.

2. Faceted search fields. Examples: Registration District; county. In these search boxes, you must select a valid search term from a controlled vocabulary in a dropdown list. To do this, click on the text starting “Browse… “ to the right of the search box and select a valid value from the list. Alternatively, enter a search term in the search box itself and, if it is valid, the term will pop up below the box so you can select it.

The Location in 1921 search field works by means of an underlying geo-coded gazetteer, so you can search for places within a certain radius of a search term. For example, you can enter Pershore as a search term, click on the version that pops up beneath the search box and set the slider to, say, 10 miles – you will then see search results for all individuals who meet your search criteria within a 10-mile radius of Pershore.

The Nationality field has been partly standardised to overcome, for purposes of searching, the great variety of different ways in which foreign nationals and others have expressed their status. For example, one individual from Antwerp or Brussels might write that they are a Belgian subject, while a second would express that as being a subject of Belgium, while some of their Francophone compatriots used the terms Belge or Belgique. Where practicable, we have standardised all these variants to Belgian; you will see the original term used when you view an image.

In the same way, we have used the following standardised European nationalities: Austrian, Bulgarian, Czechoslovak (but retained Czech and Slovak when so used in the original), Danish, Dutch, Estonian, Finnish, French, German, Greek, Hungarian, Icelandic, Irish, Italian, Latvian, Lithuanian, Maltese, Norwegian, Polish, Portuguese, Romanian, Russian, Spanish, Swedish, Swiss and Yugoslavian.

Note that Estonian is usually written as Esthonian in the original, and that Latvian sometimes appears as Lettish.

Many of the Lithuanian, Polish, Romanian and Russian nationals – or ex-nationals, as many naturalised British subjects have included their nationality of birth – are Jewish. You may also see this expressed as Polish Russian or Russian Polish with or without the use of the word Jewish. Again, Jewishness is expressed in different ways in the original documents – as Jew or Jewess, for instance. For convenience, we have standardised these variants to Jewish where possible. However, we have retained Hebrew as a separate distinct value in the database.

Armenians are sometimes described as being Ottoman, Persian or Turkish and we have retained these qualifiers where they are present. It is possible that some Armenians may be mis-transcribed as “American” so please exercise caution when searching for Armenian ancestors.

Where possible, we have also standardised other nationalities in the same way as the European ones given above. For example:

Argentine, Brazilian, Chilean, Chinese, Japanese, Siamese, Sinhalese etc.

Less standardisation has been undertaken of Australian, Canadian, New Zealand, South African and United States citizens. For example, you may see the same concept written as America, American, United States, United States of America, US, USA and many other variants. Note that on occasion USA may denote the Union of South Africa rather than the United States.

Foreign nationals usually add “resident” (in England & Wales) or “visitor” (to England & Wales) or similar.

The Nationality field was only meant to be filled in by non-British nationals and those who had gone through the naturalisation process (and become naturalised British subjects). However, it was often completed unnecessarily by British nationals, and occasionally you may see such text as “British to the Backbone” and “True Briton”.

The Optional Keywords search field can be used to look for any search term you wish. It interrogates all transcription fields in the database in which the content is not purely numerical.

Wildcard searching uses the asterisk (*). Asterisks can be placed before or after a word.

Some fields contain text which is stemmed and tokenised by word (using Natural Language Processing). This means that, for example, searching for the term “farmer” in the Occupation field should additionally return you results with text before and/or after that search term – e.g. “Arable farmer” or “Farmer of 120 acres” – without the need to use a wildcard.

For address searching, you can go to our census address search page, which can be found here:

Findmypast census address search

Note that it is possible to search for an address using the advanced person search – it’s not necessary to insert a name to conduct a search, and you can search by an address. However, the search results are returned differently. Person search results show the names of individuals and their personal details (birth year, Registration District of residence etc). Address search results show the streets in their larger geography (district, county etc) and then, when you click through on the particular street that interests you, you see the houses listed for that street.

We estimate that about 90% of addresses have been standardised. By this we mean that the raw text string, shown in transcriptions as “Full address as transcribed”, has been analysed, parsed into it component fields (house name, house number and street) and standardised programmatically to correct errors, expand abbreviations and remove as many variables as possible. However, you may still see some variants of street names in your search results, which you should consider along with your main (standardised) results.

Schedule number is displayed in the free search results (in column 7 of 7). This helps you to understand the sequence of census returns in the original bound volume. Usually, therefore, it also indicates the enumerator’s walk, and you may see, for example, that the enumerator walked down one side of the street first, enumerating all the even numbered houses, and then crossed over to enumerate the odds.

Finally, detailed street addresses are not given in the original document for all settlements. For example, Appleby Magna in Leicestershire and Braunston in Northamptonshire both had street names but these were not used by the enumerators when completing the address panels on the front of the census schedules. It is likely that this occurred only in relatively small settlements where the enumerator was familiar with the streets and their inhabitants and didn’t feel a need to record more specific address details.

The coverage of the 1921 Census of England & Wales includes the following:

• The civilian population of England & Wales, from neonates to centenarians
• The population of the Isle of Man
• The population of the Channel Islands
• The British armed forces (Army, Royal Navy and Royal Air Force) in England & Wales and the offshore islands
• The British armed forces throughout Ireland
• The British armed forces overseas throughout the world
• H.M. Coastguard Service in Ireland
• Merchant navy and fishing vessels, including foreign ships, in port on census night or returning to port in the days immediately after the census
• Visitors and tourists from outside England & Wales, including from Scotland
• Trans-migrants from continental Europe passing through en route to, for example, North America or South Africa
• Foreign diplomats and consular staff not resident at embassy or consular premises but living in the community
• Some foreign diplomats resident at their embassies and consulates (but not all, on the basis that an embassy is foreign soil)
• British subjects employed by and resident at foreign embassy or consular premises

The following are not included in the 1921 Census of England & Wales:

• The population of Ireland
• The population of Scotland
• The British armed forces (Army, Royal Navy and Royal Air Force) in Scotland
• Merchant navy and fishing vessels which were not in port on census night and which did not return to port in the days immediately after the census

It is also likely that some individuals falling within the following categories may not have been enumerated:

• Homeless, vagrants and others sleeping rough on census night
• Men tramping on the road for work
• Romany and Traveller caravans or encampments
• Refuseniks who for various reasons declined to complete or return their census schedules

Note, however, that the police were charged with locating and enumerating vagrants and there are plenty of census returns for individuals discovered sleeping rough in barns or tents. These usually appear on schedules at the very end of the Enumeration District in which they were found, after the main sequence of schedules created by the enumerator.

Similarly, enumerators were expected to engather as much information as possible about all individuals within their Enumeration District and in some cases filled in forms for the reluctant or unwilling.

Population figures

The official population count for the 1921 Census of England & Wales was given as follows:

• 37,885,242, according to the preliminary report, dated August 1921
• 37,886,699 according to the final report, dated 1927

Note that neither of these is the size of the actual resident population of England & Wales. This is because the Census Office in England & Wales records the de facto rather than the de jure population on census night – those who actually happen to be there on the night in question and not those who, so to speak, belong there.

The figure is inflated by, for example, seamen on foreign vessels which happened to be in port; trans-migrants such as those in shipping line hotels; and visitors and tourists.

The figure is deflated by, for example, merchant marine and fishermen at sea; undercounting of the vagrant population; and people normally resident here travelling abroad.

Lost and damaged parts of the 1921 Census of England & Wales

Only one piece (bound volume) of the main collection in archive series RG 15 was described as “missing at transfer” when the collection was passed over from the custody of the holding government body to The National Archives. This is piece RG 15/1946, being one of those for the Registration District of Bermondsey in SE London. However, it seems possible that piece 1946 was created for an Enumeration District (ED 59) which was never used. If so, not only would the census for Bermondsey be complete, but the 1921 Census of England & Wales will be essentially complete.

However, this is not to say that census returns survive for everyone in England & Wales on census night. This is not the case. Some individual schedules have been lost over the course of the years between the summer of 1921 and the Findmypast digitisation project of 2019-2021. Notes from as long ago as the 1970s show that there was already an official awareness of this. For example, schedule number 2 was noted as missing from piece RG 15/1499 (Whitechapel Registration District), schedule numbers 1 to 5 inclusive from piece RG 15/1525 (St George in the East RD) and so on. The quantities suggested are very low – a handful of schedules at most from each piece – but, of course, that will be no consolation if it was your family that had filled in a missing schedule.

The collection was moved on several occasions over the decades and suffered some damage. At one location, water and mould damage affected especially the fore-edges of volumes, which we surmise must have been shelved up against a damp exterior wall or leaking pipework. As a result, some returns only survive as fragments and others are so badly damaged as to be illegible in whole or in part. The five worst mould-affected pieces are RG 15/7482 (for Oxford Registration District), RG 15/26640, RG 15/26641 (both for Pontypridd RD), RG 15/26538 and RG 15/26539 (both Cardiff RD). However, several runs of volumes for Lancashire, Yorkshire (all three Ridings), Co Durham and some parts of South Wales have experienced greater or lesser degrees of mould damage. Again, this problem was noted as long ago as the 1970s.

The last three pieces within the collection – RG 15/28154, RG 15/28155 and RG 15/28156 – contain the fragments of census returns which could not be reunited with the schedules or pieces to which they belong.

The Census Office was very interested in employment in 1921, and in particular in the distances people were travelling to work (this is why workplace was requested on the form). This means that the 1921 Census elaborates upon the information asked for previously in the 1911 Census. The 1911 Census had contained four fields related to employment. These were an individual’s occupation; the industry in which he or she worked; their employment status (such as employed or self-employed); and whether working from home.

For 1921, Occupation and industry were asked for in col k; details of the employer, or whether working on own account (col l); and place of work (col m). This means that, for the first time, an individual’s employer and their actual workplace were requested. On 1911 and previous censuses, you would see workers but without a sense of how they belonged together in the economic landscape. Suddenly, in 1921, you see how they are working for, say, a particular local farmer, draper’s, or factory.

The information relating to occupation and employer were also coded in green ink by Census Office clerks. We have indexed these codes as well as the text. An occupation code should be between 000 and 999 and is usually followed by a forward slash and an extra digit. These suffixes after the slash indicate employment status. The suffixed codes are as follows:

/0 = not employed - in education etc
/1 = not employed - unpaid domestic duties etc
/2 = employer - not working at home
/3 = employer - working at home
/4 = self-employed (own account) - not working at home
/5 = self-employed (own account) - working at home
/6 = employed - not working at home
/7 = employed - working at home
/8 = unemployed

Some occupation codes have an alpha prefix, which has not been indexed. This is as follows:

Y = retired or not gainfully employed

You may also see the following alpha code, which is used without a numerical code:

X = retired or not gainfully employed - but used only in special cases such as female armed forces pensioners (e.g. widows of ex-servicemen living on army pensions), convicts, asylum inmates, retired colonial civil service etc.

A table showing occupation codes can be accessed from the Useful Links & resources bar on the right-hand side of this page.

Occupation codes are searchable. Select one or more of the numerical codes from the Occupation Code dropdown on the advanced search screen. For instance, search under code 478 for French polishers, 643 for piano tuners, 820 for Anglican clergy or 890 for bookmakers (“bookies” or turf accountants).

Occupations given on the Welsh language returns (usually on schedule type codes C and CC) were translated into English by the Census Office and written in green on the schedules. It is these English language occupations which were transcribed for the database.

Enumeration District ED 4 in piece RG 15/24420 for Coundon, Co Durham (and possibly others) contains enumerator annotations such as "on strike" in the occupation field against striking coal miners. This information was not required of householders but makes for fascinating social history, especially given that the delay to the taking of the 1921 Census was primarily a consequence of industrial action in the coal-mining sector.

The Census Office does not appear to have exercised itself too much about unemployment. Its 1927 General Report states that the number of individuals who declared that they were out of work was 50,865 males and 23,479 females. However, these figures relate only to those who made no other comment, and the instructions for completion of schedules ask explicitly for the usual occupation to be given for the unemployed (e.g. “coal getter, unemployed”, or “dock labourer, out of work”). Official parliamentary records give 1,549,307 males and 477,627 females as at the 24th June 1921, although these almost certainly include Scotland as well as England & Wales.

Paid female employment was largely restricted to certain fields – what the Census Office calls “personal service” (domestic servants, charwomen, laundry workers and so on), textiles, clothing manufacture, shop work, clerical and teaching. However, women were entering the professions: there were, for instance, 1,253 female medical professionals (excluding nurses and midwives); 278 policewomen; 127 Non-Conformist ministers; 49 architects; 24 vets; and 20 female barristers.

Employment (industrial sector) codes appear as green-ink annotations in col l. These have been transcribed. A table showing employment codes can be accessed from the Useful Links & resources bar on the right-hand side of this page.

Around 2.25% of the population was recorded as being resident in institutions of one sort or another on census night.

The 1921 Census of England & Wales distinguishes between Larger Establishments, Institutions and Prisons. All are characterised by having a population of inmates as opposed to family as occupants on census night.

Public and charitable institutions include asylums, barracks, boarding schools, colleges, convalescent homes, hospitals, infirmaries, nursing homes, religious communities, sanatoria, workhouses and the like.

Larger establishments include hotels, hostels, registered lodging houses, Salvation Army shelters and other places where there were expected to be 100 or more inmates upon census night and which were not regarded as formal institutions.

In practice, the distinction between these two could be blurred. You may find institutional returns (with schedule type codes I, II and III, depending upon their size) or extended household returns (with schedule type codes EE, EEE, WW and WWW) used for a hotel, for example. Indeed, some hotels gave out schedule type code E forms to each room, so that effectively the hotel guests submitted returns as if the hotel room was their home. An example of this is the Hotel Cecil, ED 11 in piece RG 15/533.

Prisons and borstals are a special case and have their own P, PP and PPP schedule type codes. Industrial schools and reformatories often use the same P forms.

Hospitals of many different sorts appear in the census – general hospitals, fever hospitals, infectious diseases hospitals, isolation hospitals and so on. One type to look out for are the Ministry of Pensions Hospitals, of which there were 48 by the late 1920s, many catering for shell-shocked ex-servicemen.

Religious communities usually appear as their own institutional EDs when they are of a certain size. Note that this is not usually a function of the count of members of the religious order but of educational and care facilities that they operate onsite. For example, Nazareth Houses (run by the Roman Catholic Sisters of Nazareth) often have an associated home for the sick, elderly or orphaned; likewise, establishments run by the Little Sisters of the Poor.

Workhouses appear under a number of different guises. There were 631 workhouses at the time of the 1921 Census, essentially one per Registration District, given that most Registration Districts arose out of the Poor Law Unions, each of which was a union of parishes combining together to share a common workhouse. In 1921 a workhouse could be called a Board of Guardians’ Institution; a Board of Guardians’ Poor Law Institution; a Guardians’ Institution; a Poor Law Institution; a Poor Law Workhouse; a Union Workhouse; or any variant upon this theme. It may also appear under a given name (which may then make it sound like a country house): for example, Fairmile House (at Christchurch, then in Hampshire) and Sunnylands (at Chard in Somerset). Most workhouses had an associated infirmary, which may be enumerated together with it or as a separate Enumeration District; others had casual wards, cottage homes (often called scattered homes) for children and other more specialist facilities.

Barracks in depots, and all other Army (and Royal Air Force and onshore Royal Navy) establishments such as camps, drill halls and hutments, should be recorded on NM (armed forces) census schedules rather than institutional returns.

Enumerating inmates. Some institutional returns, such as those for asylums, appear to have been completed in non-standard ways by their governors. The instructions and illustrative examples printed on the schedules make it clear that the full names of inmates should be given. However, you may see inmates enumerated with initials (instead of forenames) and surnames (e.g. “A B Collins” might be used instead of Albert Bertram Collins). In some cases, initials are used instead of forename and surname (e.g. “A B C” might be used instead of Albert Bertram Collins), presumably at the discretion of a governor or other chief resident officer wanting to protect the privacy of, for example, asylum residents. This does not appear to have been in accordance with any Census Office guidance. Examples of initial-only census returns for institutions include the Brighton County Borough Mental Hospital (in piece RG 15/4841), the West Riding Asylum (piece RG 15/21145) and the Joint Counties Asylum in Carmarthen (piece RG 15/27335). There are many more. The use of initials in this way preserved the confidentiality of inmates of mental health facilities but, at the same time, anonymises them and contributes towards their under-representation and writing out of history. For the genealogist, the use of initials means, of course, that inmates in such institutions will be much more difficult to find and identify.

Special EDs. Some Enumeration Districts (EDs) are described as “Special” on the front cover board of the piece. These invariably relate to Institutions. However, there appear to be no obvious factors distinguishing the EDs marked as Special with those EDs also containing Institutions but not marked as Special.

Arrangement. Generally, the pieces (bound volumes of census schedules) for each Registration Sub-district (SD) are arranged in Enumeration District (ED) order, starting with ED 1. Institutions tend to appear as the highest numbered EDs within each SD and to be gathered together in one or more pieces at the end of the collection for the Sub-district. For instance, East Stonehouse SD in Devon contains 14 numbered EDs, from 1 to 14. ED 13 and ED 14 are bound within the same piece, RG 15/10674. ED 13 is the Royal Marine Barracks and ED 14 is the Royal Naval Hospital. However, in some cases, institutional EDs are bound in the backs of pieces of regular EDs where space permitted. For example, piece RG 15/10807 for South Molton SD in Devon contains EDs 1, 2, 3, 4, 5, 25 and 26. The first five EDs comprise regular household returns for small districts. There was space within the bindings to include ED 25 for the South Molton Poor Law Institution and ED 26 for its Children’s Home.

Note that there may be discrepancies between information recorded in the RG 114 Plans of Division and what is to be found in RG 15. This is because the Plans of Division were created in advance, and some pragmatic decisions appear to have been made at the taking of the census in terms of whether to leave institutions in the Enumeration Districts in which they were actually situated, or to extract them and create a new ED for them. The ED Numbers given for institutions in RG 114 therefore do not always match those shown in RG 15.

Generally, the Larger Establishment returns on schedule type codes EE, EEE etc are usually bound in with the regular household schedules for an ED in their expected street number sequence. However, this is not universally the case and may have varied by, for example, Sub-district registrar’s practice or by size. For example, the Rowton Houses (a philanthropic chain of seven working men’s hostels) always appear as their own EDs. Some of the largest common lodging houses were also extracted from the ED in which they were located geographically and made into independent stand-alone EDs. However, many similar but smaller lodging houses were not extracted and will appear in the correct house number sequence for the street in which they occur.

Tables showing a selection of the larger institutions, arranged by type or theme, can be accessed from the Useful Links & resources bar on the right-hand side of this page.

Shipping is the umbrella term in the 1921 Census of England & Wales for all vessels enumerated on census night, irrespective of their size and function, with the exception of the Royal Navy overseas. The official figures record that 5,241 such vessels were enumerated and an additional 1,760 barges and boats on inland waterways.

Enumerators and their equivalents (such as customs officers) were required to distribute schedules to all vessels, whether British or not, during the week leading up to census night, and then to collect them as soon as possible afterwards. Note that they were also required, in the days after census night, to keep a look-out for vessels returning to port which might have been at sea and not been enumerated elsewhere. Enumerators and customs officers were expected to remain alert for such vessels for up to two weeks after census night.

Generally, men whose vessels were in their home port on census night tended to sleep at home and will appear in the regular household schedules rather than on shipping schedules. This is of course also true of mariners who were between voyages. Contrarily, men whose vessels were away from home often slept on board to avoid unnecessary lodging expenses, and therefore are much more likely to be found in the shipping schedules. This also means that a shipping schedule does not necessarily record all its crew.

Note that not all merchant marine and fishing fleet boats would have returned to harbour in time to have been enumerated as part of the 1921 Census. This is a feature common to all the decennial census returns from 1841 to 1921 – those whose work was at sea tend to be under-recorded.

The merchant service and fishing vessels are enumerated upon census returns with the S and SS schedule type code. An S schedule is sufficient for small craft (and larger vessels not heavily manned) with up to 10 persons on board. The SS schedule allows for up to 70 individuals to be enumerated. S and SS are used both for England and for Wales without distinction (there were no shipping schedules printed in the Welsh language).

Pages 1 and 3 of the six-page SS schedule have the schedule type code SS printed towards the bottom left. Some masters of vessels mistook this for a part of the form asking for the name of the vessel, and wrote it in there. For example, see page 1 of the SS schedule for the Norwegian vessel SS Grontoft (Grøntoft) and others in piece RG 15/25675.

The shipping returns for the Channel Islands and Isle of Man are broadly similar to the S schedules but have the schedule type codes AS and MS respectively.

A few vessels in the 1921 Census of England & Wales are enumerated with schedule type code D, which is the 1921 Census of Scotland shipping schedule. Such vessels were in Scottish ports just before census weekend, collected their Census of Scotland returns there, but completed and handed them in to English or Welsh enumerators or customs officers having sailed to a port south of the border in the meantime. Examples of such Scottish shipping schedules can be found returned in various ports, such as the Birkenhead piece RG 15/17185.

Shipping schedules are similar to those used for households. However, they ask for Crew or Passenger, rather than for Relationship to Head of Household.

Royal Navy vessels and establishments are not returned on shipping schedules such as the S and SS. They have the schedule type code NM (armed forces schedules, in common with the Army and Royal Air Force). The Royal Navy overseas (including in Ireland) can be found in a sequence of piece numbers from RG 15/28142 to 28152. However, Royal Navy ships in English and Welsh docks and ports can be found with other shipping returns. For example, piece RG 15/4070 (for Medway RD in Kent) contains the Royal Navy returns (schedule type code NM) for HMS Thruster and HMS Tintagel as well as merchant shipping (using schedule type code S). The official figures record that 534 Royal Navy vessels were enumerated in the 1921 Census.

Generally, the pieces (bound volumes of census schedules) for each Registration Sub-district (SD) are arranged in Enumeration District (ED) order, starting with ED 1. Shipping tends to appear at the end of the collection for the Sub-district. Some pieces containing only shipping are not given an ED number but appear simply as “ED Shipping” after the run of numbered EDs. However, this is not always the case. For example, piece RG 15/17540 is numbered as ED 16 of the Litherland SD and contains shipping schedules. Shipping returns can, therefore, be found in many places, and it should be noted that canal barges and other vessels (including house-boats) on inland waterways will usually be enumerated using the schedule type code S census returns.

Lighthouses and lightships will usually be returned on schedule type code S. For example, the Plymouth Breakwater Lighthouse and Eddystone Lighthouse are enumerated in piece RG 15/10568, which is the ED Shipping for the Yealmpton Sub-district. Similarly, the Flatholm Lighthouse is enumerated in piece RG 15/26463, which is the ED Shipping for Central Cardiff Sub-district. The Corporation of Trinity House was responsible for the enumeration of lighthouses.

The British armed forces are well covered by the 1921 Census. This includes the Army, the Royal Air Force (formed in 1918) and the Royal Navy (including the Royal Marines).

A special section of the 1921 Census is devoted to the armed forces overseas. There are 35 pieces (volumes) for the Army overseas, 11 pieces for the Royal Navy and one piece for the Royal Air Force. Those for the Army and the Navy are grouped geographically. For instance, there are nine pieces for the Army in Ireland (right across the island of Ireland), and two for the Royal Navy in the Mediterranean. There are pieces covering the British Army in India (RG 15/28116 to 28124 inclusive), the Army of Occupation in Constantinople (RG 15/28110), the British Army of the Rhine in Germany (RG 15/28138) and the peace-keeping British Upper Silesian Force (RG 15/28139). The extensive presence in the Middle-east, especially Mesopotamia (mostly modern Iraq) and Palestine, is also reflected in the census returns.

A table giving details of the armed forces overseas pieces can be accessed from the Useful Links & resources bar on the right-hand side of this page.

The Army presence in Ireland is particularly extensive, as suggested by the fact that it fills nine whole volumes (pieces RG 15/28125 to 28133 inclusive). It should be borne in mind that the Irish War of Independence was ongoing and in fact the truce was not signed until 11th July 1921, three weeks after census night. This means that you may find British Army units at unexpected locations across Ireland as well as at the large barracks and camps such as the Curragh.

It is possible to search these overseas volumes exclusively if you know or suspect that an ancestor was serving abroad with the forces in the summer of 1921. To do so, select Armed Forces Overseas from the Country dropdown on the advanced search screen. However, do not forget the armed services which were in England & Wales at the time of the census. You will find many soldiers in army barracks, for example, across the country, sailors aboard ships in home ports, and airmen at their bases.

Note that the Army Reserve had been mobilised in April 1921 during the state of emergency declared by Lloyd George. At the same time, a new Defence Force was formed, in which men enlisted for 90 days’ emergency service – which period would extend beyond the taking of the census on 19th June 1921. Defence Force units may be found in various locations, including drill halls and army camps and hutments.

All armed forces establishments are included. This includes temporary camps, training sites, military hospitals, detention barracks and married quarters.

The main omission is the armed forces from all three services which were stationed in Scotland, or at sea in Scottish waters, in June 1921. These would have been enumerated in the 1921 Census of Scotland and therefore do not appear in the 1921 Census of England & Wales. You should look for such service personnel in the 1921 Census of Scotland.

The armed forces had special census returns (schedules) which carry the code NM (assumed to stand for “naval and military”). The NM schedule was used both within the country and overseas. Two types of NM schedule were printed, one for domestic use within England & Wales and one for service overseas. There are several page types within each NM form. The main part of the return deals, of course, with serving officers and other ranks or ratings. However, there is often a section headed the “Return of all persons OTHER than serving officers and other ranks or ratings” – in other words, civilians, including wives and family. Note that the wives and children are therefore on different pages and not linked to the servicemen to whom they are related.

The NM form is broadly similar to the regular census schedules completed by householders all around the country. However, there are some differences. Firstly, the Relationship to Head column is omitted in the main section of the return, as of course officers and men are usually not related to one another; however, this column does appear in the “OTHER” section in which it is relevant. Secondly, the schedule records details of rank or rating and the branch or arm of the service (whereas the normal household return asks for occupation and employment details). Thirdly, there is a column for Language Spoken, which specifically enquires after Gaelic or Welsh.

There appears to have been no explicit instruction on how service personnel should be ordered on a census return. However, generally speaking, as is the custom in the forces, officers and men are listed by rank and, within those of equal rank, by seniority (length of service not age). Other than that, men may be arranged by company or other grouping, or (in the case of other ranks) simply alphabetically by surname.

Note that the NM form does not ask for service number although in a few cases a keen senior NCO has included this extra detail, which is a wonderful bonus for researchers.

Three Gibraltar 1921 Census household schedules were used in lieu of an NM schedule for the crew of the Royal Fleet Auxiliary “Bacchus”, a stores ship which had docked in Gibraltar for three days. The return was rightly rejected by the Registrar of Births in Gibraltar and forwarded, via H.M. Dockyard in Gibraltar, to the Census Office in London. This curious adapted return can be found in piece RG 15/28151.

Veterans. The 1921 Census was, of course, undertaken soon after the end of the Great War. It is therefore replete with ex-service personnel. There are many army pensioners, for instance, and disabled ex-servicemen – some of whom used the vehicle of their census schedule to bemoan their fate, especially when unemployed or living in constrained circumstances in the “land fit for heroes” – and an increase in both widows and orphans. Interesting returns include ED 17 in piece RG 15/3868 for the Queen’s Hospital for Facial Injuries in Sidcup, ED 25 in piece RG 15/3110 for Queen Mary's Convalescent Centre in Epsom, ED 42 in piece RG 15/631 for St Dunstan’s Hostel for Blinded Soldiers & Sailors, and ED 77 in piece RG 15/524 for the Y.M.C.A. Hostel for Servicemen. There are many Ministry of Pensions Hospitals, such as the Ministry of Pensions Special Surgical Hospital (ED 42 in piece RG 15/245) and the Ministry of Pensions Hospital in Knotty Ash (ED 51 in piece RG 15/17893). The 1921 Census is therefore a great resource for those researching surviving veterans of the Great War, as well as those still actively serving in the armed forces in summer 1921.

A selection of pieces of relevance to armed forces research are shown in the England & Wales institution tables, which can be accessed from the Useful Links & resources bar on the right-hand side of this page. These tables are arranged by type or theme. Those of special interest to military historians include the tables for army & navy, Ministry of Pensions Hospitals, nautical training and Royal Air Force.

The 1921 Census of England & Wales of course contains all surviving census schedules for Wales. However, there are some details which are worth emphasising.

Schedule types. There were two types of household schedule printed for the 1921 Census in Wales. Firstly, there were English-language schedules, which largely resemble the schedules for England. These have the schedule type codes W, WW and WWW, depending upon the size of household or establishment. Secondly, there were Welsh-language schedules which ask for exactly the same information. These have the schedule type codes C and CC (for Cymraeg), again depending upon household size. Welsh-language household schedules had been distributed according to need ever since the 1841 Census.

In areas of Wales where there was Welsh-language proficiency, enumerators were given stocks of both the W and the C code schedules, and distributed them according to need during the week before census day. This means that in many Enumeration Districts (EDs) where Welsh was spoken alongside English, you will see a mixture of both schedule type codes. It is possible that there are a few EDs where only Welsh was spoken (or Welsh was preferred as the first language of the bi-lingual) and you will see only schedule type code C and CC returns.

Welsh language. All the schedule type codes used in Wales – both C and W types – ask whether the occupants speak English, Welsh or both. The census had required this information in each census since the 1891 Census. This is the only difference between the schedules distributed in Wales and those across the border in England.

Although not necessarily invariably the case, a majority of the schedule type code C schedules will have been completed in the Welsh language and a majority of the schedule type code W schedules in English. All completed schedules were despatched by the local registrar to the Census Office in London. There, a small team of trained officers went through the Welsh returns and, in most cases, translated Welsh-language text into English. Where this has been done, you will see their distinctive green-ink notes and codes. Where an English translation exists for an occupation given in the Welsh language, we have indexed the English version for convenience and consistency of searching across the census.

If you want to search only Welsh-language returns, you may do so. Choose schedule type codes C and CC from the schedule type dropdown. You will then only receive search results from the Welsh-language forms.

If you want to search only self-declared Welsh-language speakers, you may do so. Choose options from the faceted search field called Language Spoken. You may need to exercise some care, as Welsh speakers appear in different guises. They should be returned with the four dropdown options Both Welsh & English; Cymraeg (Welsh); Welsh; and Y Ddwy (Both Welsh & English). They should not appear under Saesneg (English), which ought to relate only to monoglot English-speakers (such as a relative or lodger) described as such by Welsh-speaking heads of household. Do not use the dropdown option Both, as this should appear in non-Welsh contexts only.

The official figures for the 1921 Census are 155,989 speakers of Welsh only and 766,103 speakers of both English and Welsh in Wales.

Note that Welsh speakers in England are not counted; there is no column on the household schedules for England (E, EE and EEE) asking for language spoken. For example, there were Welsh-speakers in the lower Ceiriog Valley (Dyffryn Ceiriog) in St Martin’s SD (known as Llanfarthin in the Welsh language) within Oswestry RD in Shropshire. For those interested, the piece range for St Martin’s SD is RG 15/12417 to 12422.

Border issues. The 1921 Census uses the administrative geography of the civil registration system – the network of Registration Districts (RDs), each of which is divided into Registration Sub-districts (SDs). It is important to note that these RDs and SDs sit within registration counties which do not necessarily observe administrative or ancient county or even country boundaries (the same is of course true on a smaller scale of civil parishes and ecclesiastical parishes). Some examples:

• Piece RG 15/12404 is for Llansilin SD within the Oswestry RD of Shropshire in England. Part of Llansilin SD lies within Wales but the Welsh border is ignored. This piece includes Llangedwyn & Llancadwaladr which are in Denbighshire, and Rhyd-y-Croesau which is situated right on the border.
• Pieces RG 15/27856 to 27870 are for the SDs of Beaumaris, Llanfairpwll and Llanidan. All three are on the isle of Anglesey. For civil registration purposes, however, these SDs are within the RD of Bangor, which is on mainland Carnarvonshire.

A good way to familiarise yourself with the many such border-issue discrepancies is to refer to our table of Welsh civil parishes in the 1921 Census – see the Useful Links & Resources bar to the right-hand side of this page.

Those interested in the social and cultural history of Wales might like to look at our GoogleMap showing 50 Welsh cultural figures in the 1921 Census. This is also accessible via the Useful Links & Resources bar to the right-hand side of this page.

The 1921 Census of England & Wales includes the offshore islands of the Isle of Man and the Channel Islands.

The Channel Islands returns are arranged by island, civil parish and by numbered Enumeration District, and not – as England & Wales returns are – by Registration District, Registration Sub-district and Enumeration District.

However, for ease of searching across the entire collection, the Channel Islands civil parishes are treated as Registration Districts for our search purposes. You will therefore find:

• Channel Islands as a country in the Country dropdown
• Alderney, Guernsey, Jersey and Sark as counties in the County dropdown
• The islands’ civil parishes as Registration Districts in the Registration Districts dropdown

Channel Islands household schedules have the schedule type code A or, in the case of households of 11 or more occupants, AA. These Channel Islands returns are similar to those for England & Wales in most respects, although the address panel on the front gives the different enumeration geography used on the Islands.

Channel Islands shipping schedules have the schedule type code AS. Note that these are used for all shipping on the Islands on census night in 1921 irrespective of where a vessel was officially registered. In other words, a boat from, say, Devon or Pembrokeshire which was in harbour in Guernsey on 19th June 1921 would be enumerated on an AS schedule (unless it had brought with it an S schedule collected elsewhere within the days preceding the census). Channel Islands shipping can be seen in, for example, pieces RG 15/28007 (Guernsey) and 28057 (Jersey). Lighthouses are also enumerated. For example, the Gaskets and Alderney Lighthouses can be found in piece RG 15/28061 and Sark Lighthouse in 28064.

There are also two pieces in the Armed Forces Overseas section which relate to the Channel Islands. These are piece RG 15/28140 for Guernsey and Alderney, and piece RG 15/28141 for Jersey. Piece RG 15/28140 covers Fort George (Guernsey), Fort Albert (Alderney) and Castle Cornet (Guernsey). Piece RG 15/28141 covers St Peter's Barracks, Fort Regent, the Military Hospital and Government House (Jersey).

Note that none of the Channel Islands census schedule types asks after whether French (or the local languages such as Guernésiais or Jèrriais) is spoken.

The 1921 Census of England & Wales includes the offshore islands of the Isle of Man and the Channel Islands.

The Isle of Man returns are arranged by civil parish and by numbered Enumeration District, and not – as England & Wales returns are – by Registration District, Registration Sub-district and Enumeration District.

However, for ease of searching across the entire collection, the Isle of Man civil parishes are treated as Registration Districts for our search purposes. You will therefore find:

• Isle of Man as a country in the Country dropdown
• Isle of Man also as a county in the County dropdown
• Isle of Man civil parishes as Registration Districts in the Registration Districts dropdown

Isle of Man household schedules have the schedule type code M or, in the case of households of 11 or more occupants, MM. These Isle of Man returns are similar to those for England & Wales in most respects, although the address panel on the front gives the different enumeration geography used on the Isle. Additionally, they also ask whether Manx is spoken. The Census Office had required this information in each census since the 1891 Census.

Isle of Man shipping schedules have the schedule type code MS (short for Manx shipping). These again enquire about the Manx language.

Note that these MS schedules were used for all shipping on the Isle of Man on census night in 1921 irrespective of where a vessel was officially registered. In other words, a boat from, say, Lancashire which was in port in Douglas on 19th June 1921 would have been enumerated on an MS schedule (unless it had brought with it an S schedule collected elsewhere within the days preceding the census). Examples of Manx shipping returns can be seen in piece RG 15/28099.

The 1921 Census of England & Wales was taken on 19th June 1921. This was only three weeks before the truce ending the Irish War of Independence (11th July 1921) and six months before the Anglo-Irish Treaty was signed (6th December 1921). The Partition of Ireland had already been enacted (3rd May 1921) before the census was taken, and the 1921 Treaty created the new independent Irish Free State.

Against this backdrop, there was no 1921 Census of Ireland.

Therefore, Ireland is not, of course, included in the 1921 Census of England & Wales. However, there are some returns from Ireland. These comprise:

• British Army returns in piece number range RG 15/28125 to 28133.
• Royal Navy vessels in Irish ports or Irish waters in piece RG 15/28143.
• H.M. Coastguard Service for stations all across Ireland in piece RG 15/28149.
• Royal Air Force in piece RG 15/28153.

In addition, there are of course many Irish, and persons of Irish descent, resident in England & Wales, or serving in the British armed forces, who therefore are enumerated in the 1921 Census of England & Wales. The place of birth field is of particular value if you are trying to research Irish roots back from England or Wales and need to find the place of origin within Ireland – usually, the 1921 Census will give town or village and county.

Scotland is a separate jurisdiction and holds its own decennial census. It is not included in the 1921 Census of England & Wales.

However, there were of course many Scots, and persons of Scottish descent, resident in England & Wales, or serving in the British armed forces, who therefore are enumerated in the 1921 Census of England & Wales.

Note that all Scots serving in the British Army, Royal Navy (including Royal Marines) and Royal Air Force should be included in the 1921 Census of England & Wales if in England & Wales, in Ireland or overseas on census night. Only those Scottish soldiers at army barracks in Scotland, or Royal Navy officers and ratings aboard vessels in Scottish ports or Scottish waters, will not be included. You should look for such service personnel in the 1921 Census of Scotland.

Scottish merchant marine and fishing vessels in English & Welsh ports, and any in the Channel Islands or the Isle of Man, will be included in the 1921 Census of England & Wales.

Note that there are in fact some 1921 Scotland Census returns in the 1921 Census of England & Wales. A number of vessels collected Scottish shipping schedules (Scottish schedule type code D) while in harbour in Scotland a day or two before census night, and then handed these in, completed, to harbour masters in England & Wales after census night, having sailed south in the meantime. Similarly, there are at least two Scottish household schedules (Scottish schedule type code A) which were delivered to householders in Scotland who then travelled to England over the census weekend and handed in their completed schedule to an enumerator in England. The 1921 Census of Scotland comprises enumerator books; the original householder schedules were not retained, meaning that these few found within the 1921 Census of England & Wales may be the only completed examples surviving anywhere.

Examples of Scottish schedule type code A (household) can be found in pieces RG 15/19118 (Chorlton) and 20792 (Blackpool).

Examples of Scottish schedule type code D (shipping) can be found in pieces RG 15/17185 (Birkenhead) and 25487 (Tynemouth).

The 1921 Census of Scotland will be made available at Scotland's People in 2022.

The following terms are used when talking about the 1921 Census of England & Wales.

Series. This is The National Archives’ archive series. There are two main series in the 1921 Census. These are RG 15 and RG 114. RG stands for the Registrar General, being the government department which created the records. RG 15 is the main collection of census returns for 1921. RG 114 holds the so-called Plans of Division – in other words, the description of the way in which the area to be covered by the census was divided up into manageable chunks (the Enumeration Districts). The Registration District maps are from a third series, RG 18.

Piece. This is archive piece. A piece is a unit within a series. For example, RG 15 is divided up into pieces numbered from 1 to 28156. Each piece is a bound volume of census returns, with the exception of the last three pieces in the series (RG 15/28154 to 28156), which are boxes of fragments. Pieces have numbers only and are not named. The piece number is written in pencil on the front cover board into which the piece is bound (often scrawled in inelegant handwriting right across it). The piece number was also given on a white label stuck to the lower spine of the piece – part of this can sometimes be glimpsed on the image of the front cover board.

Enumerator. Enumerators were engaged on temporary contracts to prepare the census for their Enumeration District, to distribute the schedules to the households within it, to ensure they were completed properly, and then to collect and check them afterwards and bundle them up for the Sub-district Registrar. Female enumerators were employed for each census since the 1891 Census and there would have been plenty in 1921.

Schedule. This is the correct word for a census return or form. On the front of the schedule is the address panel (completed by the enumerator) and the instructions on how to fill in the schedule. The schedules were numbered on the back (by the enumerator). The rest of the information on the back of the schedule is completed by the householder and signed by them towards the bottom-right. Annotations in green ink – usually numerical codes – were added by employees of the Census Office when processing the census.

Enumeration District (ED). This is the basic administrative unit of the census, the smallest division, being the area that could be covered by an enumerator in a single day. In urban areas, an Enumeration District could be a very compact geographical area of just a few densely-populated streets. In rural areas, an Enumeration District would be a significantly larger area (e.g. of farmhouses, isolated cottages and hamlets) that could be covered on foot or bicycle by an enumerator. Enumeration Districts are numbered but not named. Enumeration Districts are created for the census and have no life outside or beyond it.

Registration Sub-district (SD). This is the next level up from Enumeration District (ED). Each Sub-district (SD) is divided up into EDs. There may be just a few EDs in a Sub-district or there may be very many. For instance, the Brilley SD in Herefordshire contains five EDs, while the Rhondda SD has 84 EDs. Sub-districts are both named and numbered. For example, Princes Risborough SD is SD 3 of three within the Registration District of Wycombe. Unlike the EDs, the SDs exist outside the census, and are the divisions of the Registration District used in the civil registration of births, marriages and deaths.

Registration District (RD). Also (and more properly) known as a Superintendent Registrar’s District. This is the next level up from Registration Sub-district (SD). Each Registration District (RD) is divided up into one or more SDs. Usually, there are just a few SDs within an RD but sometimes there are many more – for example, there are 11 SDs within West Derby RD and 13 SDs within West Ham RD. Registration Districts are both named and numbered. For example, Rochford RD is RD 195 and Halifax RD is RD 496. Numbering runs from RD 1 (Paddington RD) to RD 634 Holyhead. Some RD Nos have alpha suffixes, e.g. there is an RD 3A for Hammersmith RD and RD 3B for Fulham RD. Note that there is no RD 13 or RD 376 in 1921.

As with SDs, the RDs exist outside the census and will be familiar to family historians who have searched the indexes to the civil registers of births, marriages and deaths created by the General Register Office.

County. Everyone knows what a county is but in fact administratively there are different types of county. The 1921 Census used the so-called registration counties, those used in the civil registration of births, marriages and deaths. These may be different from the ancient counties, which may date back to Anglo-Saxon times; from parliamentary counties, which are used for electoral registration; and administrative counties, which were established for local government purposes in 1888. Note that the Census Office used registration counties for structuring the census itself but administrative counties when reporting on its findings from the 1921 Census.

Parish. Two different sorts of parish feature in the 1921 Census. The first is the ecclesiastical parish, which here is Anglican (Church of England or Church in Wales, as applicable). This is the one with which family historians are familiar from parish registers of baptism, banns, marriage and burial. The oldest of the ecclesiastical parishes are known as ancient parishes, and usually acquired secular as well as religious functions. These so-called mother churches often developed daughter chapelries, many of which had grown into ecclesiastical parishes in their own right by the time of the 1921 Census. The second type of parish in 1921 is the civil parish, which is an entirely secular, administrative district used in local government. Where “parish” is used in 1921 Census documents without any further qualification or comment, civil parish will usually be meant.

The 1921 Census most closely resembles the 1911 Census and superficially they are similar. In both cases, the original household schedules survive (unlike for the previous England & Wales censuses for 1841 to 1901 inclusive), so we get to see what was actually written by householders, in their own handwriting and with their own signatures.

However, despite the similarities there are in fact many differences.

The archival materials for the 1911 Census include Enumerator Summary Books (archive series RG 78 at The National Archives), in which the heads of household are shown at their addresses. These have not survived for 1921. However, on the other hand, for 1921 there are the Plans of Division (series RG 114), which show how the country was divided up into manageable units (Enumeration Districts), and also, in the majority of cases, helpful Registration District maps (from series RG 18).

A 1921 Census schedule is a differently-sized landscape document to the 1911 Census schedule. The original schedules for 1921 measure approximately 550 mm x 250 mm, compared to 530 mm x 315 mm for 1911. The size is not as consistent as you might expect, and a variation of +/- 5 mm can be found. A minority of schedules are of a much lighter weight than the others, upon paper not dissimilar to that used in post office and trade directories of the period. H.M. Stationery Office Press in Harrow, Middlesex produced at least two major print-runs, in November 1920 and February 1921. The position of the printed text upon the paper varies. Finally, guillotining of the paper was sometimes wayward, creating a wave along the top edge.

For the 1911 Census, the Census Office had printed a small quantity of foreign-language forms – at least in German and Yiddish – as well as in Welsh. Only Welsh-language census forms (schedule type codes C and CC) were printed in addition to English for 1921.

The 1911 Census was taken on 2nd April 1911, whereas the 1921 Census was postponed from its originally intended 24th April date to 19th June 1921. This will have affected the distribution of the population, and there was evidence of population inflation in popular seaside resorts such as Blackpool, Eastbourne, Margate and Southend.

The 1911 Census shows the household address twice, once on the address panel on the front of the schedule (in the enumerator’s handwriting) and once on the back under the signature (in the householder’s hand). For 1921, the address is given only on the front, in the address panel, completed by the enumerator.

The 1911 Census schedule includes the following features not found in 1921:

• Fertility data – length of marriage, and the number of children born, still alive and now deceased
• Infirmity – a record of, for example, deafness or blindness

The 1921 Census schedule includes the following features not found in 1911:

• Age given in years and months, instead of just in years
• Marital status – this now includes Divorced as a permissible condition
• Orphanhood – showing for minors whether either or both parents are deceased
• Education – showing whether attending part-time or full-time
• Employment information – giving name of employer and address of workplace
• Minor dependency grid – showing number and ages of children and step-children up to and including the age of 15 years, whether or not resident in the household and included on the household return

There are some differences in coverage too. The 1921 Census of England & Wales includes the British armed forces across Ireland, whereas these were not included in 1911 (as they were being covered by the 1911 Census of Ireland). 1921 also includes the Coastguard Service in Ireland, likewise missing in 1911 for the same reason. Finally, the Royal Air Force was enumerated in 1921 whereas, of course, it did not exist at the time of the 1911 Census!