> I found the Willie Abernathy yesterday, but had not found the time to check if the age was right.
Let's just say that it looks close enough to put that lead very clost to the top of a "follow up now" list.
> The fact Willie is a step-child probably adds to the problem.
A little, yes. You're going to have to find both of his mother's marriages, for one thing.
> In the 1900 Ellis Co. census, Willie is listed as William and his name is spelled with a "bb" instead of "b".
Unusual variations (mis-spellings) of the surname now number well over sixty variations just from census records alone. "Abbernathy" isn't all that uncommon although it does happen less frequently than some.
> A quick search of some of the more popular sites did not show anything on Rachel Alice Keel either.
She's going to be a bit more of a problem for you and I caution you not to ignore the possibility that the line between Tennessee and Alabama appears, from a genealogical point of view, about as fixed as the line between one county and another often do. Families liviing close to that line are often listed on one census in one state as being born in that state when in fact they may have been born in the other state. And on a later census in that other state, birth locations could be reversed. One more caution to recall: back then, census enumerators weren't paid very highly, nor held to a very high standard of accuracy. How could they, when some of the people they were trying to count couldn't read or write or even say how old _they_ happened to be? Sometimes when an enumerator asked the illiterate 'lady' of a household in really rural areas how old a particular child of hers was, they sometimes got back the equivalent of "Tater diggin' time t' year t' barn burnded."