The registers of signatures of depositors contain a rich historical record of new account openings at the Freedman’s Bank. The original records are stored at the National Archives and have been photographed and transcribed by FamilySearch.org as a genealogical record. The transcribed records were generously made available to us through a research agreement with FamilySearch International.
The data was supplied in the form of a single, comprehensive data table in .csv format. The transcription is of high quality, though some occasional transcription errors may need to be corrected when using the data for analytical research. More importantly, a number of issues relating to the incomplete nature of the physical records also need to addressed when dealing with this data.
The records can be browsed via the FamilySearch.org website. Researchers who wish to use the full dataset should obtain permission FamilySearch International, upon which you may contact a member of the research team for the programs which clean the dataset and create valid record sequences. A log of important details about the records are provided here.
Description
The registers of signatures of depositors, also called ledger books, contain a rich historical record of new account openings at the Freedman’s Bank. The FamilySearch database contains the indexes to the ledger books for 29 of the branches in the Freedman’s Bank. Each page of the book contains eight entries, each of which contains blanks for account number, account holder name, date, birthplace, where brought up, residence, age, occupation, employer, wife or husband, children, father, mother, brothers and sisters, and remarks.
Although some branches are not available in the records and gaps in coverage exist, the records provide a substantial history of the Freedman’s Bank. The subsequently transcribed file contains each individual record and all of the descriptive data for each entry, which includes demographic information, as well as next of kin. This file has 480,597 entries, which includes both underlying account information and associate records for friends and family of account holders.
The transcription of the data provides a wealth of textual information about the depositors, but the transcription process contains a number of errors, which we fix through a replaceable sequence of edits that can be applied via fully programmatic computer code. In addition to fixing these errors, we also perform a substantial amount of algorithmic categorization of free-form textual records that facilitate large-scale analysis. Researchers who are are granted access to the raw data from FamilySearch can make use of this provided code to generate the cleaned and fully parsed data.
Details
FamilySearch Database Notes
-
Image collection is described by a unique number (digital_gs_number) but this does not uniquely divide up the collection of bank records. Some image collections may contain up to 3 different branches. Therefore, the image number “image_nbr” may not correspond to the “Image _____” field on the website UI. No direct correspondence appears to exist, and the UI direction appears to be handled internally by another data reference.
-
There appears to be 3 database errors in the image_id and system_addedsort_key, where these string variables which are concatenations of unique numeric variables in the data table have an image_nbr which disagrees. In each case the correct account number appears to be correct for image_nbr while the name appears to be correct for image_id/system_addedsort_key. None of the entries represent the primary account holder.
-
The following variables are constant for all entries, and therefore contain no additional information:
- event_type
- fs_collection_id
- fs_record_version
- fs_vis_status
- pr_is_principal
-
The following are nearly always missing:
- death_date_std
- death_year
- image_type
- birth_date_std
- birth_year
- mthr_is_principal
- spouse_is_principal
- fthr_is_principal
-
The following variables: digital_gs_number, image_nbr, unique_identifier create a unique sort key for the database which will sort the records in roughly chronological order
Record Ordering Issues
-
Note that while Memphis declares there to be three separate scan targets in Roll 24, they are actually fully contiguous, and in fact contain records 1996-1999 which are stated to be discontinuous by the archive sheet
-
Accounts in Tallahassee have a discontinuous numbering scheme. On July 18, 1871, Film Group 4098148 image nbr 669, the accounts renumber going from 1380 to 688 and then ending the book in 887. This appears to be a valid sequence. It is unknown why the accounts were renumbered.
-
A record for the Raleigh branch exists, but it contains only a single page of very unreliable records, and is largely unusable.
-
The Charleston branch contains a lengthy middle section of the book between image #139 and image #189 which is clearly devoted to miscellaneous housekeeping. It contains no valid dates anywhere and is full of transfers to and from other branches and banks which are generally not present in other branch account books.
-
Records from the Philadelphia branch represent only a small number of new accounts held only by organizations and not by individuals and are possibly from a parallel ledger record rather than the main book.
Odd lot records
-
Savannah 4098339_00619 through 4098339_00787 is a second target for film roll 10 and is a loosely ordered set of odd lot duplicate records. They are very sparse, and appear to all be duplicates of accounts recorded in the other sequential logbooks.
-
Washington 4098148_00009 through 4098148_00067 is the first target for film roll 5 and is a loosely ordered set of odd lot records. These records have very few gaps, so date splining should produce few problems.
Missing Record Issues
-
Charleston has a minor record gap from account #3824 to 3833 over the course of 1 to 3 missing days 12/2/1869 to 12/4/1869, over microfilm image_nbr 695 to 705 gap. This constitutes 9 missing records. This gap should be noted if this time period is of very specific interest.
-
Richmond has a minor record gap from account #1582 to 1591 over the course of 1 to 2 missing days 6/20/1971 to 6/21/1971, over microfilm image_nbr 457 to 468 gap. This constitutes 9 missing records. This gap should be noted if this time period is of very specific interest.
-
Records for Savannah which span across multiple rolls 8, 9, and 10 are fully contiguous with no gaps.
-
The National Archive declares a gap in records for Washington D.C. in the record of film roll 5 (digital_gs_number 4098148) from image_nbr 292 to 294. A renumbering takes place involving a change in new account numbers from 16303 to 20001, but this does not represent an actual gap in the records, which are contiguous
-
Richmond has a more significant gap of 21 to 23 days in record #3948 to 4005 from 11/20/1871 to 12/11/1871 over microfilm image_nbr 788 to 790. This gap of 57 accounts may be small enough that some estimation may be allowed under certain conditions.
-
A more substantial gap of 24 days and 209 records exists for Washington D.C. from 29 Apr 1871 to 23 May 1871. This gap may be small enough that some estimation may be allowed under certain conditions.
-
The Washington D.C. record book representing film roll 5, digital_gs_number 4098147, from image_nbr 12 to 231 appears to be an unusual supplemental account book to the Washington D.C. branch. Accounts here are valid, but the date span of this book does not represent a contiguous set of new accounts, and thus the full set of dates should not be taken as a valid sequence unless they overlap with the other existing records. Specifically, 12/31/1872 - 7/1/1874.
FamilySearch database discrepancies that may need to be corrected
-
Note that the FamilySearch data table erroneously indicates the Lexington, Kentucky bank_location as Louisville, Kentucky
-
Note also that the FamilySearch data table erroneously assigns the Shreveport and New Orleans records (they are correct on the website)
Valid Date Sequences
The archived physical indices contain gaps which must be accounted for when trying to examine a contiguous block. The gaps are not entirely obvious from the archival descriptions, so care must be taken when trying to form a sequence of records. Most branches have a reasonably long contiguous block of records leading up to the closure of the bank in July of 1874. The sequences are listed below for each available city.
| City | Seq | Roll | Block | First Acct | Last Acct | First Date | Last Date |
|---|---|---|---|---|---|---|---|
| Atlanta | 1 | 6 | 1 | 1 | 4517 | 15 Jan 1870 | 02 Jul 1874 |
| Augusta | 1 | 7 | 1 | 2167 | 6701 | 23 Nov 1870 | 29 Jun 1874 |
| Baltimore | 1 | 13 | 1 | 1 | 4 | 03 May 1866 | 03 May 1866 |
| Baltimore | 2 | 13 | 2 | 220 | 1484 | 15 Nov 1866 | 29 Sep 1868 |
| Baltimore | 3 | 13 | 3 | 1573 | 6768 | 24 Nov 1868 | 23 Jun 1874 |
| Beaufort | 1 | 20 | 1 | 2732 | 4707 | 20 Jun 1868 | 29 Jan 1872 |
| Beaufort | 2 | 20 | 2 | 5063 | 5988 | 14 Jan 1873 | 01 Jul 1874 |
| Charleston | 1 | 21 | 1 | 1 | 319 | 19 Dec 1865 | 17 Oct 1866 |
| Charleston | 2 | 21 | 2 | 2151 | 3824 | 07 Sep 1868 | 02 Dec 1869 |
| Charleston | 3 | 22 | 1 | 3833 | 6626 | 04 Dec 1869 | 25 Feb 1871 |
| Charleston | 4 | 23 | 1 | 6627 | 11103 | 25 Feb 1871 | 02 Jul 1872 |
| Columbus | 1 | 14 | 1 | 21 | 927 | 18 Aug 1870 | 16 Jun 1874 |
| Huntsville | 1 | 1 | 1 | 1 | 1698 | 16 Dec 1865 | 2 Jul 1874 |
| Lexington | 1 | 11 | 1 | 217 | 1976 | 21 Nov 1870 | 11 Apr 1874 |
| Little Rock | 1 | 3 | 1 | 153 | 1359 | 27 Feb 1871 | 15 Jul 1874 |
| Louisville | 1 | 11 | 1 | 1 | 1928 | 15 Sep 1865 | 28 Jan 1868 |
| Louisville | 2 | 11 | 2 | 5122 | 7333 | 01 May 1872 | 26 Jun 1874 |
| Lynchburg | 1 | 26 | 1 | 153 | 215 | 08 Jul 1871 | 22 Aug 1871 |
| Memphis | 1 | 24 | 1 | 1 | 1995 | 28 Dec 1865 | 01 Jul 1874 |
| Mobile | 1 | 2 | 1 | 777 | 2323 | 18 Jun 1867 | 10 May 1869 |
| Nashville | 1 | 25 | 1 | 4174 | 6189 | 23 Dec 1871 | 15 Jun 1874 |
| Natchez | 1 | 14 | 1 | 1 | 707 | 29 Mar 1870 | 18 Jun 1874 |
| New Bern | 1 | 18 | 1 | 1327 | 4157 | 30 Oct 1869 | 25 Jul 1874 |
| New Orleans | 1 | 12 | 1 | 5 | 1018 | 18 Jun 1866 | 11 Mar 1869 |
| New Orleans | 2 | 12 | 2 | 4365 | 8570 | 17 Jan 1872 | 29 Jun 1874 |
| New York | 1 | 17 | 1 | 1422 | 6942 | 25 Oct 1870 | 29 Jun 1874 |
| Norfolk | 1 | 26 | 1 | 3950 | 5415 | 04 Dec 1871 | 26 Jun 1874 |
| Philadelphia | 1 | 19 | 1 | 1 | 3004 | 06 Jan 1870 | 26 Jun 1874 |
| Richmond | 1 | 26 | 1 | 232 | 1582 | 18 Jul 1867 | 20 Jun 1870 |
| Richmond | 2 | 27 | 1 | 1591 | 3948 | 21 Jun 1870 | 20 Nov 1871 |
| Richmond | 3 | 27 | 2 | 4005 | 7691 | 11 Dec 1871 | 29 Jun 1874 |
| Savannah | 1 | 8 | 1 | 1 | 1137 | 10 Jan 1866 | 05 Aug 1868 |
| Savannah | 2 | 8 | 2 | 1298 | 4947 | 16 Nov 1868 | 17 Dec 1870 |
| Savannah | 3 | 9 | 1 | 4948 | 9868 | 17 Dec 1870 | 22 Oct 1872 |
| Savannah | 4 | 10 | 1 | 9869 | 14558 | 22 Oct 1872 | 01 Sep 1874 |
| Shreveport | 1 | 12 | 1 | 149 | 1320 | 11 Feb 1871 | 29 Jun 1874 |
| St. Louis | 1 | 16 | 1 | 223 | 366 | 06 Apr 1869 | 08 Oct 1869 |
| Tallahassee | 1 | 5 | 1 | 1 | 887 | 25 Aug 1866 | 15 Jan 1872 |
| Vicksburg | 1 | 15 | 1 | 1157 | 8662 | 16 Jul 1868 | 29 Jun 1874 |
| Washington | 1 | 4 | 1 | 5 | 1553 | 28 Aug 1865 | 10 Apr 1868 |
| Washington | 2 | 4 | 2 | 3500 | 7197 | 25 Jan 1870 | 29 Apr 1871 |
| Washington | 3 | 4 | 3 | 7406 | 9316 | 23 May 1871 | 30 Dec 1871 |
| Washington | 4 | 5 | 1 | 3 | 456 | 28 May 1872 | 17 Jun 1874 |
| Washington | 5 | 5 | 2 | 14631 | 16303 | 31 Dec 1872 | 23 Aug 1873 |
| Washington | 6 | 5 | 3 | 20001 | 21397 | 25 Aug 1873 | 01 Jul 1874 |
| Wilmington | 1 | 18 | 1 | 1208 | 1343 | 03 Sep 1869 | 30 Oct 1869 |
| Wilmington | 2 | 18 | 2 | 5406 | 7266 | 12 Dec 1872 | 26 Aug 1873 |