Hugh, >The transcription quality is abysmal, and I am frequently only able to find "missing persons" by having a very good idea of where they are< .... and using a good bit of ingenuity! You mentioned the word bizarre! What is more bizarre than finding a widowed daughter living with her father transcribed with her father's surname? Look for George Robert born 1821 in Alverthorpe and you will find Annis Ainsley, widowed daughter, transcribed as Annis Robert although her children have the correct surname of Ainsley. This is not an isolated case - try some searches for yourself. (Hint - use 1837Online's advanced search in their 1891 census and look for any fairly common surname and select widow and daughter, then try and find them on Ancestry). Makes it look as though in the input program the surname field is linked to the relationship field, but Ancestry deny it. They also deny that the gender field was linked to the relationship field. Do a search for all the female Williams - there are thousands of them in all the censuses! Look down the list for a mother(-in-law) or a sister and you will invariably find the mother should be brother and the sister visitor or son. There are too many of them for it just to be coincidence! Jeanne Bunting