How can I submit corrections for geographical errors?

How can I submit corrections for geographical errors?  The city of New Britain, Connecticut is indexed as being in New Haven county, while the city is entirely within Hartford county.  (The document images say Hartford county.)  There are also several districts in Danbury, Connecticut that are indexed as Unincorporated Fairfield County, even though they are within Danbury town limits - and there are NO unincorporated areas of any county in Connecticut.  All six New England states - plus New York, New Jersey, and Pennsylvania - have every square inch incorporated as a city, town, Reservation, or miscellaneous subdivision [buels, gores, etc], so none of these states should have any district indexed as unincorporated county.  Please advise, thank you.

  • Hi Julie,

    Trying to duplicate your effort as I'm always interest in geographical errors that might creep into my own databases on the census locational tools.

    I assume this if for material on the National Archives 1950 census site.  I assume it is not for the 1950 census maps that were used by the census bureau and which can be found on the Archives catalog.  I assume this is not for the original ED definitions that the Archives links to (the original T1224 films).

    I'm a little confused then as to what you want to correct.  The 1950 census site has a location index.  You can choose a state, and then choose from a list of areas within that state.  The choices seem to be cities within that state that have their own ED prefix (usually given to cities of over 50,000), or a county name.  You are right the Archives index incorrectly shows "New Britain, New Haven" as the choice term for the city.  It should be New Britain, Hartford.  But there is nothing in the index that goes below the large city or entire county names within a state.  That is, you won't see a city name of Danbury with 22,067 (I also see on one of my resources there is also a Danbury town with 30,337 people in 1950) on the list of locations you can choose.  Where are those errors you want to correct if they are not on the location index box at   ??

  • I also don't understand when you say: "All six New England states - plus New York, New Jersey, and Pennsylvania - have every square inch incorporated as a city, town, Reservation, or miscellaneous subdivision [buels, gores, etc], so none of these states should have any district indexed as unincorporated county."  Tell me why ED 60-28 in New York State (randomly chosen) has "unincorporated" in its original ED filmed definition.  It's in Westchester County.  To check that....go to the 1950 Archives census page.  Choose New York, and then put in 60-28 for the Enumeration District (ED) number as your only search terms.  The Archives transcribed definition shows the area is unincorporated within the county of Westchester.  If you click on the "(View Original ED Description)" link on the page you will see the census bureau's original filmed description of the ED that the Archives transcribed.  Note that 60-28, and below it on the filmed page 60-29 are "unincorporated" areas.

  • Julie, Thanks for pointing out the New Britain error.  We'll get it corrected.

    Can you point out the EDs in Danbury, Connecticut that are indexed as Unincorporated Fairfield County, even though they are within Danbury town limits?

    Claire Kluskens, NARA

  • Hi, Joel -

    Thanks for your response!  I see what you mean - on this site, the larger cities are given their own headings, separate from the counties they're in.  This does become a problem searching on, which I understand you have no control over - part of what I'm trying to figure out is where the disconnect is; I assume they received the data from the National Archives?  The way this site is set up, the New Britain/New Haven error won't throw researchers off course, whereas on Ancestry, if you did a search for, say, everyone with the last name Kowalski in Hartford county, then no residents of New Britain will be included in the search results.  I get that they're a private company and you guys are a government agency, so their issues are not under your purview; I'm just trying to figure out where the glitch happened in the data feed.

    As for the Westchester county situation you mentioned: Herein lies the madness of New York's geographical heirarchy!    In NY state, everything is part of a city, town, or reservation, but then they have an additional level of incorporated villages.  If a place is incorporated as a village, it is a subset of the town from a census perspective, but an entirely separate location from a vital records perspective.  (The village gets its own Hall of Records for births, marriages, and deaths.)  So New York folks will call an area "unincorporated" if it is not part of a village, but that area is still part of an incorporated town.  See the second page of ED 60-28: It says Westchester county, and then Courtlandt town underneath.  Courtlandt has two incorporated villages - Croton-on-Hudson and Buchanan - as well as areas that are not part of a village.  So a New Yorker will call this area unincorporated, meaning it is not contained within a village, but that area still belongs to a town.  So it was an error of omission for the census-taker to leave the "incorporated place" field blank on the cover page.  "Unincorporated" means something very different in the 9 northeastern-most states, than it does in the whole rest of the country, which have areas that truly belong to nothing below a county level.

    BTW, this whole "village" system in NY drives me nuts, it's so needlessly complicated!  And Westchester is the worst county for this; the Ryes and the Mamaronecks and overlaps and discontinuous areas are a *huge* pain in the ...

    Thanks for your attention on this!  Take care,


  • Hi Claire,

    Thanks for your response!

    The EDs that are part of Danbury town are 42, 43, 44, 45, 46, and 47; all are called either Germantown, Hayestown, or Beckettville.

    Three of the EDs - 9, 30C, and 49, the Green Hotel, State Teachers College Dormitory, and the penitentiary - don't have a Danbury heading, although it is mentioned in the description that they are Danbury.  So they're relatively easy to find on this website, but they're not searchable as Danbury on  (As above, I totally get Ancestry is not part of the National Archives, but if they did receive their data from the N.A., it looks like the missing headings caused these districts to get indexed on *their* end as unincorporated; I will be contacting them as well.)

    EDs 152A, B, and C are all New Fairfield town; The description about improper districting is very confusing, but I clicked through all the images, and they are all marked as New Fairfield, so I think the dispute is just about the district boundaries within the town.

    154A and B, which have no description, are both Newtown.

    156 has the same issue as 9, 30C, and 49 - there is no Newtown heading, although the town name is mentioned in the description.

    221 and 222 have Georgetown in the heading, which is not incorporated as a separate town; it's a neighborhood [with its own Post Office] that is spread across small portions of Redding, Wilton, and Weston.  I don't know how you would parse out which pages are part of which towns.

    Thanks for looking into this!  Best wishes,


  • Julie:  Our goal in providing the Enumeration District description transcriptions was to follow the Census Bureau's original ED definitions as closely as possible.  We did that with institutional EDs that you mention:  CT EDs 1-9, 1-30C, 1-49, 1-156, which were written by the Census Bureau in a format used in many places, such as Ohio ED 4-25 "(Ashtabula Hotel (In Ashtabula city, Ward 4, Ashtabula township))."  It is true that the Census Bureau wrote institutional ED descriptions differently in large cities in particular, such as Ohio ED 89-71 "(Akron city - That part of Ward 2 (Tract A - 3 - part), comprising Akron Hotel in Block 7))."  However, we are not going to rewrite the Bureau's ED descriptions to make them conform to a particular pattern.

    We can update the descriptions of CT EDs 1-152A, 1-152B, and 1-152C to indicate they are part of New Fairfield town since that information is apparent from the schedules and the original description of CT ED 1-152.  Same with 1-154A and 1-154B for Newton the same reason.

    As far as CT EDs 1-42, 1-43, 1-44, 1-45, 1-46, 1-47, 1-221, and 1-222, for Germantown, Hayestown, Beckettville, and Georgetown, they are acceptable as written because we followed the Census Bureau's written descriptions.  The Census Bureau made an effort nationwide to separately enumerate unincorporated villages and the like.  For insight into their rules and procedures for establishing, numbering, and describing EDs in the 1950 census, see 17FLD-Mapping 100, "How to Establish Enumeration District Boundaries (September 1949)" (National Archives Identifier 214451610) which is digitally available here:  The Census Bureau chose to not include "Danbury Town" in the descriptions for 42, 43, 44, 45, 46, 47, or the appropriate town for 221 or 222, and NARA isn't going to rewrite them to include the town identification. 

    Other websites can do what they wish.  For example, most of the points you raise for the EDs mentioned above are handled in the way you want at

    We do appreciate you bringing to our attention our blatant error of ascribing New Britain to New Haven County instead of Hartford County.  I'm not sure how that happened during our metadata review before opening day.  I brought the error to the attention of the 1950 census website team and I anticipate that the correct information will appear online in a week or two in our 1950 Census Website,  Thanks!

    Claire Kluskens, NARA

  • Hi Julie,

    I disagree with your interpretation and your generalizations.  You indicate "So New York folks will call an area "unincorporated" if it is not part of a village, but that area is still part of an incorporated town.  See the second page of ED 60-28: It says Westchester county, and then Courtlandt town underneath".  Show me that Cortlandt Town (note spelling) is "incorporated".  I see at  that an incorporated area "means the area that was within the corporate limits of St. Croix Falls, Osceola, North Hudson,.... or Prescott".  I treat northeastern state's "towns" not as population centers with corporate limits but more like other state's townships which are geopolitical divisions.  I see at  that "The Town of Cortlandt, New York is located in the northwestern corner of Westchester County. With a population of almost 43,000 citizens, Cortlandt includes two incorporated villages, Buchanan and Croton-on-Hudson, and several hamlets including Montrose, Crugers and Verplanck".  With your view we would have have an incorporated area (Buchanan) within an "incorporated" town of Cortland.   The Census Bureau 1950 enumeration district definition of ED 60-28 of an unincorporated area within the town of Cortlandt within the county of Westchester, NY appears to me to accurately describe geographically the area in question.

  • Hi Claire,

    Thanks for the links and info.  I can't speak to any issues of population study (as in, *why* a place would be categorized in a particular way), and I certainly wouldn't request omitting or altering any of the original descriptions in the historical record.

    I checked my notes, and I have at least one person in my research for each of those districts, so I'll use them for an example: Adam Piskura in '42, Sebastian Mocci in '43, Colombo Melillo in '44, Andrew Repko in '45, Fred Fiore in '46, and Richard Aldreany in '47.  All of them are listed in the Danbury City Directory for 1950 as being residents of Danbury.  (The directory does note when a Danbury businessperson is, say, a resident of Bethel, or elsewhere.)  And all of them are on a 1950 census sheet with "Connecticut/Fairfield/Danbury" in the upper left corner, so presenting these districts as CDP's (under a Danbury heading) would be supported by the original documents.

    The 1950 city directory also notes that Danbury has a consolidated city/town government, for what that is worth.  (I don't know if such distinctions are taken into consideration by the Census Bureau.)

    Ultimately I would say that of course it is proper form to present all images of all descriptions exactly as they are, but as far as organizing these images on a website for public viewing/research, it is misleading to present these districts as being separate from Danbury.  However these districts are defined from an institutional perspective, people looking for their [ancestor/research subject/whoever] - known to have lived in Danbury in 1950 - are going to be thrown off by the omission.  Which I say as a Danbury native (and I have ancestors recorded there for the 1900 through 1950 censuses.)

    All of this I present respectfully.  I do sincerely appreciate the time and attention that so many people have given to this collection.

    Best regards,


  • Hi Joel,

    As I mentioned to Claire, I would not make the case for removing or redefining any descriptions from the original records; I can't speak to why the Census Bureau classifies locations one way or another for the purpose of population study.  I just mean that as far as organizing the documents on a website for public research, it is misleading to put ED 60-28 (for example) under the heading of "other places" rather than Cortlandt town.

    Let's say that someone was looking for the record of someone who lived on Crompord Road in 1950.  (That's my best guess for the street name on the first page of 60-28.)  They would search by drilling down from NY state, to Westchester county, to Cortlandt town, and then find that none of the ED's identified as Cortlandt included Crompord Road.  So they would think this area was not enumerated at all.  To use another Westchester example, ED 60-292 is also categorized as an "other place."  It is located within the town of Pelham, which is in the opposite corner of the county, over 30 miles away.  The other "other places" are in Eastchester, Greenburgh, and Ossining, all spread across the height of the county.

    By all means, present the scanned images exactly as they are, and the Census Bureau can consider Crompord Road to be miscellaneous Westchester county for whatever reason they want, but from a user standpoint, it is more specific and accurate to categorize an ED [of any stripe] as a sub-set of one town [if it is indeed within one], rather than as a spot that could be anywhere in any part of the county.

    Again, I present all of this respectfully. I'm sure all pockets of this country have different geographical quirks that can clash with the uniform metrics needed for a national census.  I have been doing research in and about the NYC metro area for 20+ years, and I've gotten to know its hierarchy (however maddening) really well; I wouldn't attest to any other region.  Like, Virginia has a Richmond county that is nowhere near the independent city of Richmond, and the city's districts got lumped into the county's, but I couldn't *begin* to unravel that knot!

    Thank you for the links, and for taking the time to read all this.

    Best wishes,