Lucy . . .You Got Some 'Splainin' To Do!

Well, one comes across some strange things in genealogy . . . . Here's one:

[In the following scenario, a single initial represents both a first name and a surname.] S was born in 1932. My initial research discovered an S-jr born in 1965 in the same county as S. I could not determine a spouse for S or a mother for S-jr. Some weeks later, I discovered that S-jr's mother was M. Yet no record of marriage for M and S turned up. Nonetheless, I was prepared to assume that M and S were S-jr's parents. That assumption seemed confirmed months later when I found a record of marriage for M and S, although the marriage took place in 1974, some nine years after S-jr's birth. That wasn't all that strange, after all, and I recorded S, M, and S-jr as a family.

Things got weird this week when I for the first time looked into Ancestry.com's new Texas Birth Index [Ancestry.com subscription required to view database]. This database lists births from 1903 through 1997. The data is from the Texas Department of State Health Services. [My original source of data about this family came from county files that were not as complete as the state records are]. In the Texas Birth Index, S-jr's mother is M, but his father is H! What's more, H and M have another child also with S's surname! While it's not unheard of for a woman to give a child her husband's surname when the child is not her husband's, it's downright weird to give a child a surname of someone other than the man named as father on the birth certificate.

The other twist here is that in 1964, the year before S-jr was born, according to the Texas Birth Index, S had a child with J, a woman he never apparently married. That was the same year H and M had their first child with S's surname.

What's going on here?