Alex Robinson (ar @t cheerful.com) of Coventry writes: The focus of my research is my great grandfather John Robinson.  He was born between 1827 and 1830, the range of dates reflect the varying ages given on census entries, one of his marriages and his death record (Watford, Dec. 1893 3a 385).  His birthplace is not in Hertfordshire and is variously described as Middlesex or Paddington.  The 1871 census records him living at 230 High Street, Watford with his wife Emma.  She is described at that time as aged 32 and born in Hemel Hempstead, Herts.  Her death record shows she died in Watford in 1873 aged 35 (Dec. 3a 250).  These two facts suggest she was born around 1838.  This is all we know about her.  What I would really like help with is finding John and Emma's marriage record.

We have tried using BMD and LDS to locate a marriage between John Robinson and Emma prior to 1871.  We have searched from 1854 (when Emma would have been 16) through to 1871 when they were living together in Watford.  John spent the rest of his life in Watford and died there, but we do not know what he did between being born in Paddington and living with Emma in 1871 in Watford (his occupation then being described as a Master Boot Maker), but we feel it is reasonable to assume that he met and married Emma in Hertfordshire, most likely between Hemel Hempstead and Watford.  This is of course an assumption.  None of the marriages that we have found between a John Robinson and an Emma meet all the above criteria.  We have contemplated making a trip to Hertfordshire to examine local records, but are not sure where such a quest would best be carried out.  (I have checked the 1861 census record for 230 High Street, Watford and John was not living there then.)

The first thing to say is that if a couple were living together they were, for the purpose of the census, man and wife, whether or not there had been a formal marriage ceremony. I can also see no reason why, knowing the social views on marriage at the time, a couple living together should reveal the fact that they were not married. If they were not married everyone in Watford would have treated Emma as John's wife (OK some might know and said "tut tut" behind their backs) so she would have been buried as Mrs John Robinson, and he would be treated as a widower on remarriage. This would be particularly true if his marriage certificate to Dorcus shows that he could not read or write (many people could not sign their own names).

The second thing is that the 1881 census shows John and Dorcus Robinson with young children - none old enough to be Emma's, so we can't use a child's birth certificate to establish her maiden name. Emma died in the December quarter of 1873 and you don't say what the death certificate says. If she died in childbed (a common cause of death) the child may have been born alive but not survived - but there is no Robinson infant birth/death registration in Watford at a suitable date. If the death certificate shows there was an inquest (unlikely) there may be an account of what happened in the local newspaper at the time. Otherwise the local newspaper is unlikely to say anything useful.

A search using FreeBMD shows are large number of "John Robinson" who married an "Emma" or who were recorded on the same page of the register as another marriage involving a register. None are close to Hemel Hempstead, but a few are in the London area (where John was born and where Emma may have gone into service) - and and I agree with you that there are no obvious matches (but see below).

The marriage indexes are full of errors - see A Comedy of Errors. If they were married Emma may be correctly indexed under her maiden name - but what was her maiden name?

Let me show you a technique that sometimes works - and can at least come up with some candidate names. It is by no means certain  - and to complicate matters Emma may have been a widow , or taken the surname of a stepfather. Basically it involves selecting likely maiden names and trying to select the more likely or eliminate the least likely.

Stage 1 is to search the Ancestry census files for all females unmarried called Emma born in Hemel Hempstead between 1837 and 1841 in the 1851 and 1861 censuses The resulting surnames are:

It is therefore likely that Emma's maiden name is one of the following:

Adams, Allen, Biggerstaff, Cross, Daniels, Davis, Dell, Dolt, Edmonds, Freeman, George, Jenkins, Lacey, Peak, Martin, Matthews, Orchard, Parker, Redup, Smith, Tarbox, Townsend, Ward, Young

I would have hoped for better agreement between 1851 and 1861 but, for example, some may have been recorded as being born in Boxmoor or Apsley in one of the censuses.

Stage 2: A check of the 1871 census shows the Emma Dolt and Emma Smith were not married in 1871. So we now have:

Adams, Allen, Biggerstaff, Cross, Daniels, Davis, Dell, Dolt, Edmonds, Freeman, George, Jenkins, Lacey, Peak, Martin, Matthews, Orchard, Parker, Redup, Smith, Tarbox, Townsend, Ward, Young

Stage 3: One can then check FreeBMD for women called Emma who married in Hemel Hempstead area between 1854 and 1871. Many of these will be of the wrong age, or not born in Hemel Hempstead. We get the following surnames - with names matching the above list highlighted.

Stage 4: Marriage registers have two weddings per page - so the FreeBMD should have one or two couples for each page reference. However it is not possible to know who married who. In addition quite a few pages ado not fit this pattern - and this usually means that there is an index error.

 Adams, Allen, Biggerstaff, Cross, Daniels, Davis, Dell, Dolt, Edmonds, Freeman, George, Jenkins, Lacey, Peak, Martin, Matthews, Orchard, Parker, Redup, Smith, Tarbox, Townsend, Ward, Young

Stage 5: Checking the later censuses for married women called Emma born in Hemel Hempstead I found an Emma Munn (widow) who was born in Hemel Hempstead in 1841 - which is presumably the Emma Ward marrying Joseph Munn in 1861 - so we can eliminate her. A search for Emma Rowe married to William was less successful.

 Adams, Allen, Biggerstaff, Cross, Daniels, Davis, Dell, Dolt, Edmonds, Freeman, George, Jenkins, Lacey, Peak, Martin, Matthews, Orchard, Parker, Redup, Smith, Tarbox, Townsend, Ward, Young

Stage 6: We know that one of the highlighted names must be the wrong Emma and it may be that both are the wrong Emma! At this stage, with two highlighted candidates, the easiest way forward may be to concentrate on these and only come back to other non-eliminated names if the marriage certificates prove negative.

Stage 7: Is one more likely than the other? If John married Emma Tarbox in 1860 one might expect children by 1871, while if Emma Allen married John in 1870 (Jan-Mar quarter) there is far less chance of there being a child. For this reason I would start by purchasing the Emma Allen marriage certificate. However there is a complication in that the FreeBMD index suggests that the reference number (3a 403) is incompatible with Hemel Hempstead - and in this case I would definitely request the certificate from the Hertford Registrar. The reason is that the Hertford copy of the marriage certificate is less likely to have an manual copying error than the central registers and we know that errors have occurred at some stage at the copying/indexing stage. In addition you are more likely to get a personal service.

If this fails you could try buying the Tarbox marriage certificate - and there is also the possibility of eliminating more names from the full list by various searches of FreeBMB (including death records) and the Ancestry census returns. For instance you could check all the Emma that married John Robinson (wherever married) to see if the bride's surname matches a name on the list. You could also check names on the list to find out where they married.  The real question is how far you want to go in searching for a step great grandmother who had no children.