For more on the Indian Census Rolls, see the National Archives pamphlet M-595. Although the paid Ancestry and Ancestry-owned Fold3 web sites originally were the only places to find the images online, all Indian Census Rolls are now freely available at Access Genealogy. There are many more Navajo-related records than the Leupp Agency records.

It then became clear to me that I needed a way to link the same people in all the censuses. The spelling of their names and clans varied, and the names sometimes changed as a person progressed in life. So some method of connecting them -- beyond what a lineage-linked database supported -- was needed. So I began transcribing the complete censuses into separate spreadsheets for each year, starting with the 1915 census.

The 1915 census had 850 people on 34 pages. The 1916 census had 1,442 people on 71 pages. So just under 600 people appear for the first time in the 1916 census. While there were some new births (and one death) since the 1915 census, these additional people were mostly entire households that were not included in the 1915 census. The 1917 census had 1,441 people on 49 pages, so that it is very similar to the 1916 census. Only 3 people not in the 1916 census appear in the 1917 census, and two families with a total of 8 people appear in the 1916 census but not in the 1917 census.

I have not yet completed the clan name standardization for the censuses. I initially created a standard name, based on one of the 1916 spellings. But I ultimately want to use the standard spellings and translations from the 1983 book "Saad Ahaah Sinil: Dual Language: A Navajo-English Dictionary: Revised Edition", edited by Martha Austin and Regina Lynch with revisions by Regina Lynch (see http://maiidinebizaad.weebly.com/uploads/6/8/5/7/6857372/saad_ahaah_sinil.pdf ). However the first priority is the transcription, so that the clan standardization and analysis will come later.

Some observations for tracing family history

• Many names include a word indicating gender, Dine or Hosteen or Hoska for males, Asdzan or Dzan or -- at the end -- BiXXX (e.g. Bi-ma, Bi-asdzan, Bitsi) for females. A name can contain both, such as Dine-XXX-bi-ma, mother of the male Dine-XXX and thus a woman.
• Many wives and widows are not identified with their own names. Instead they are identified only as "XXX Bi-asdzan" (XXX's wife) or as "YYY Bi-ma" (YYY's mother). So the actual names of these women are not in these records. Their names must be found in other sources - if there are other sources. Probate records may be useful in this regard.
• The "Bi-ma" reference are actually quite helpful within the census family for connecting a child and the mother. Other relationship suffixes are also helpful for connecting people. For example, some couples appear with the husband as "AAA Badani" (AAA's in-law) and the wife as "AAA Bitsi" (AAA's daughter); so in this case AAA is the father of "AAA Bitsi" and the father-in-law of her husband "AAA Badani". Of course, the actual names for both the husband and wife are not known when they are referred to in this way, so that other sources must be sought to find their actual names.
• Spelling did not count. Just because a name is written down in some particular spelling does not mean that it was always spelled that way. The censuses were made by English speakers who wrote down the Navajo names as best they could figure out how to spell them. Even the same census taker sometimes wrote the same name (person or clan) with different spellings on different pages or in subsequent censuses. So do not lock on to some particular spelling in the censuses as being the one true spelling because there was no one true spelling in most cases.

Three different systems of assigning identification numbers to individuals help to confirm that, despite significant differences in name spelling, records frefer to the same person.

1. Present-Last Census Numbering In the Indian Census Rolls prior to the 1930's, each person was assigned a sequential number as they were enumerated. The rolls included the person's sequential number from the prior census, which was almost always a different number. In this way, a person could be tracked from one census to the next, even if the census taker spelled their names differently or if the person was referred to in a different way in one census than in the prior census.
2. Person-Specific Census Numbering Apparently starting in the 1928 census (which does not survive in the records available), these numbers are in a column headed "Allotment, Annuity, and Identification Numbers". The Dawes Act of 1887 actually set up the concept of allotments. But these numbers in the 1928 and later Indian Census Rolls seem merely to have been unique identification numbers and not tied to allotments. In the Leupp censuses, these are 6-digit numbers, with the 1929 census being the first in which they appear, although they clearly seem to have been assigned in the 1928 census, since they are not in the 1927 census.
3. A-number Allotment Numbering This is the system I understand the least. I am not sure when it was begun, apparently sometime in or after the 1940's. I write these as the letter A followed by the number - 6 digits, sometimes with a parenthetical qualifier.

Here are the spreadsheets I have thus far completed.

• 1915 (850 people on 34 pages)
• 1916 (1,442 people on 71 pages)
• 1917 (1,441 people on 49 pages)
  A separate worksheet "1916 peopel not in 1917" has the two households with a total of 8 people who appeared in the 1916 census but not in the 1917 census. People who had appeared in the 1916 census and died prior to the 1917 are so noted -- without a number assigned -- in the census, so that these 8 people are not included for other unknown reasons.
• All censuses - Combines the completed censuses into a single spreadsheet so that each person has one line to trace them easily through all censuses in which they appeared.

The combined spreadsheet is useful for tracing an individual or a specific family over the years. The spreadsheets for the individual years let you see who their neighbors were each year and how the community as a whole changed over the years.

Having the censuses transcribed into the spreadsheets allows us to easily look at the "big picture" from different perspectives.