THE spelling of surnames can cause confusion among those researching their family histories.

People delving into early 19^th century parish registers for the first time are sometimes adamant, in the face of overwhelming evidence to the contrary, that these people cannot be their ancestors, because ‘our family has always spelled Thompson with a ‘p’, and this lot are Thomsons’.

It doesn’t take long for them to realise that before a certain point in history, somewhere perhaps in the 19^th century, nobody cared how names were spelled.

In fact, I have come across documents with the same surname spelled differently each time the person is mentioned.

Researching a family in Cornwall recently with the relatively unusual surname Gerrans, I quickly came across Gerrens, Gerrance, Gerans and Gerens. They were all the same family, and the right people.

Looking into my own family in Devon in the 18^th century, Shimell morphed into Shomill, Schimmel, Shymall and many others. The truth is: you cannot read anything into it.

The other problem, and one which family history researchers generally come across much sooner, is errors in transcription.

I went a long time without realising that one county record office held key documents about my one of my ancestors – they had been entered into their online catalogue with a ‘k’ where there should have been an ‘h’; a tiny error which merely turned my family from Shimell to Skimell, but which hid those records from my searches for years.

It’s the same with transcriptions of census records. Researchers can search on websites by entering the name of their ancestors, yet sometimes they find them, and sometimes they don’t.

How many people have traced their ancestors back from 1911 to 1901 to 1891, then suddenly they disappear, only to reappear in earlier censuses living in the same place.

I’ve heard people inventing all sorts of imaginative explanations for this; searching emigration records on a hunch that their forebears might have gone to Australia, or Canada, decided they didn’t like it, and came home again.

But, of course, they were at home all the time, oblivious to the fact that 150 years later, someone transcribing the scrawly Victorian handwriting of the census enumerator would misread Shimell and enter Trimmel, or Shirrel, or something stranger.

So when consulting any database – whether Ancestry’s census records, the archive catalogues on A2A, Google books, or any other, I always try as many variations of the same surname as I can think of.

And if it’s appropriate I use whichever wild cards are accepted by that database.

For example, searching on Ancestry for someone with the surname Shimell, I’ll enter Shim* in the surname field, which finds all those surnames starting with the characters Shim.