[capitan_refugio]

I agree with you about the historical problems with the census and the utility of tax records (government seems to get very efficient when it comes to collecting taxes!). However, although the absolute numbers may be somewhat compromised, what was collected (in the absense of outright fraud) represents a reasonable statistical sample of the whole. As with any old database, you need to sift the wheat from the chaff.

[stand watie]

when the problem IS fraud (and it more frequently than not was outright fraud in the 1800s. that's why they don't pay census enumerators on piece work now & haven't since 1900!) you can NOT separate the wheat from the chaff. for that reason, i trust NO census records that cannot be conformed from another reliable source.

one other example: in 1890 29YO identical twins were listed ONE as fullblood Seminole and the OTHER as fullblood Cheyenne! that i had to laugh about!

it does however point out how incorrect the database was/is.

free dixie,sw