1. We have a copy of his death certificate stating that he died June 11, 1841 and was born December 25, 1872.

Georgia Nell Scalf Wheelock gave information from her Bible records about her uncle Arch, brother of her father George. George was was born in 1872. George had no twin.

Arch himself, states May of 1870 on the 1900 census. Also consider that Arch did not give the information on his death certificate and that might be another factor.

Here is some of what is known:

Washington County, TN - 1870 Census
District 13
Scalf, Jesse 37
Clarissa 31
Laura 9
John W. 8
Louisa 6
Elmira J. 4
Archibald 3 mo. <-would make him born in 1870
Brown, Margaret 68----(Clarissa's mother)

Washington County, TN - 1880 Census
District 13
Scalf, Jesse 45
Laura L. 19
John W. 17
Louisa 15
Archibald 10 <---- validates 1870
George 8

Washington County, TN - 1900 Census
District 13
Arch is married and has his own family

Scalf, Archibald 30     May 1870
Laura 27 Feb. 1873   (Laura Hulse (marriage records of Wa. Co.)
Thomas 7 Nov. 1892
John 6 Dec. 1893
Ada 4 July 1895
Sarah 2 Aug. 1897

Arch is buried in Atichson Co. Missouri (Rock Port)
in the Linden Cemetery.

He does have a headstone:
Arch J Scalf
b. Dec 25, 1872   (I believe this to be in error also!)
d. June 11-1941

I would assume that 1870 is the correct year, even though the death certificate is a legal document and will probably take precedence over it. He could not have been on the 1870 census if he were born in 1872, so, it's probably going to be up to the individual as to what they want to accept.

The confusion between December 25, 1870 and May 1870 is a little more involved, and there are several possibilities, from census taker phobia, to dodging ex-wives. Margaret is right about earliest information being best and I think an excellent resolution to the question is to reread that Washington County 1870 census and get the date the census was taken. We suspect it will show the census was taken in August or thereabouts and if so, I at least would have my hat on the May date instead of December, regardless of what was put on what record later. I think that approach would make a lot of sense.

I don't have a method by which we can gain a great deal of confidence about the day of the month.

For me, I say Archibald was born in 1870 and the census check will confirm the month. Nothing else makes any sense.

Thanks to Margaret Fleenor for her valuable contribution to this posting.