One of the good things about having a subscription to Ancestry (other websites are available) is that your search for one person may bring up details about another . . . which you'd got wrong.
So there I was, looking for a John Culpin and I saw John Arthur Whatford Culpin, a direct relation, in the 1901 census. So I had a quick gander at my database, saw that I didn't have an entry for him at this census and duly noted it down. Then I realised that his wife was down as Sarah, whereas I had her as Selina.
Now, I accept that some of the transcribing leaves a bit to be desired (but some of the enumerators' writing was appalling) but I don't think anyone could make "Sarah" out of "Selina" so I looked again at the particular marriage entry that I had for John & his wife.
OK, so I made a mistake. It wasn't Selina Denison at all, it was Sarah Jennings. That made a lot more sense. And I then went on to find them in the 1891 census and therein discovered another son of John & his first wife. Result!
Memo to self (again) - don't assume!