New Arrivals    Books    Archival Products   Charts   Newsletters   Upcoming Events   Contact Us  

Popular Categories

   Australia & New Zealand
   British Home Children
      - Acadie, Acadian
      - New Brunswick
      - Newfoundland & Lab.
      - Nova Scotia
      - Ontario
      - Prince Edward Island
      - Quebec
      - Western Canada
      - First Nations, Metis
      - Military - Before 1920
      - Loyalists / UEL
      - Pioneers' Stories
      - Genealogy How-To
   England & Wales
   Ireland & Northern Ireland
   United States
      - American Revolution
   more countries...
   Genealogy How-To

   Archival & Acid-Free
   Conservation How-To

   Charts, Forms
   Gift Certificates

Popular Authors

   Thomas MacEntee
   Paul Milner
   Chris Paton
   Ron W. Shaw
   Gavin K. Watt

Popular Publishers

   Global Heritage Press
   Institute of Jesuit Studies
   MacDonald Research
   OGS - Ottawa Branch
   Unlock The Past

Search by topic, title, author or word:

News & How-To
Formerly branded as

Articles, press releases,and how-to information for everyone interested in genealogy and history

Subscribe to our free newsletter

[an error occurred while processing this directive]

Article posted: October 4, 1999
By: Fawne Stratford-Devai   Biography & Archived Articles

When I first started the English and Welsh Roots Column in the Gazette I promised guest authors who were experts in a variety of English and Welsh family history topics. I am very pleased to welcome author Michael Whitfield Foster. The incompleteness and inaccuracies of the central birth, marriage and death indexes for England and Wales have been known for many years by frustrated family historians. Earlier issues of English and Welsh Roots provided information on the use of these indexes to finding civil registration records (part 1: Gazette 26) and the importance of Michael's findings to the outcry for easier and cheaper access to English Civil Registration Records (part 2: Gazette 27). Michael's timely book "A Comedy of Errors" sets down very clearly the problems with the registration system and the inaccuracies and incompleteness of the central indexes. Michael has continued his work examining the indexes and offers his new findings and observations to Gazette Readers. I am very pleased Michael has chosen English and Welsh Roots as the place to pass along this new information and I look forward to many more such articles from him.

Fawne Stratford-Devai

A Comedy of Errors Continued...
The Marriage Records of England and Wales
by Michael Whitfield Foster

Some more reflections on the GRO indexes and records
Since publishing my book "A Comedy of Errors" in 1998 I have been continuing to study the marriage indexes. I have some new examples and results that extend some of the findings of the book and a few additional thoughts that I feel it may be useful to share. Hence this article.

The stated objectives of the Civil Registration system
Now that official attention is being turned to the future of the civil registration records it seems appropriate to look again very briefly at the aims of the system and how far they have been achieved.

When the Civil Registration system came into existence in the third quarter of 1837, the new Registrar General wrote to the clergy to explain its aims. They were to be responsible for the registration of most marriages. He urged their cooperation and care in operating the new system so that copies of marriage entries (can be) available "without difficulty" and may be "obtained promptly, easily and with very trifling cost". To this end the legislation passed in 1836 included the requirement that "the Registrar General shall cause indexes of all the certified copies to be made and kept in the General Register office".

How far have these intentions been realised in practice?
Nothing in the legislation defined the form and scope of those indexes. The system of volumes and page numbers that we see in the indexes of births, deaths and marriages reflects the general register office's way of filing and accessing the records. The indexes themselves were designed by the GRO, not specified by the legislation.

I am now more than ever convinced that one of our greatest problems is that the design of the indexes was inadequate for the stated objective of prompt and easy access to the records. For the huge numbers of the more common names in the indexes we know only too well how impossible it can be to decide which entry is the one we need. With more and better detail in the indexes those decisions could often be made reliably or at least the number of possibilities could be reduced to a feasible number. In practice the inadequacy of the index system can be just as intractable a problem as the many errors and omissions that I described in "A Comedy of Errors". The objectives of prompt and easy access have simply not been met.

My time in the GRO in 1998 gave me an opportunity to test this in practice. I had once looked for the birth of my great grandmother Emma Cox and found three entries in the Eton district in a single year. When I was at the GRO I spent some time looking for a David Evans in my family. There can be dozens of these names in a single year. With the benefit of immediate access to the actual GRO films I was able to take film after film off the racks and check all the likely entries until I found the right one. That is an option totally unavailable to researchers. A common name can be as complete a bar to further search as the many errors and omissions. I would never have found my David Evans through the unaided use of the GRO index.

It is particularly unfortunate that the first Registrar General opted to index the marriages by volume and page number alone. Although that was enough to allow his staff to locate a requested marriage within the four marriages on a page, some unique identifier like adding a 1,2,3 or 4 would have transformed the value of the index for users. It would have identified each marriage uniquely. Countless faulty or fruitless attempts to match marriage partners in the indexes would have been avoided, as would many wasted certificate purchases. Both my files and the FreeBMD files would identify all marriage partners. We can now have two Mary Browns indexed to one page - she may be one person included twice or may be two different people - both situations occur within the indexes and we can't now tell. If James William Hurst signed the marriage register as James Hurst he is indexed twice. A better index would distinguish them as one person or two. I'm very conscious that my studies at the GRO have given me an invaluable understanding of the system when I am looking up a name in the index, something that most researchers sadly can't have. My hope is that "A Comedy of Errors" will have gone a long way towards filling that gap.

The Office for National Statistics has recently initiated a process of public consultation on the future of the GRO record system. Leading British genealogists have been echoing the findings of "A Comedy of Errors" that the only valid basis for the records lies with the original registers. Going back to those records would make it possible to create new and effective indexes with even more identifying features. If that breakthrough is achieved then research will be transformed and my research will have been more than amply rewarded.

1881 marriages
The research for "A Comedy of Errors" contained a great deal of work on the records of the first twenty years or so of the marriage records and relatively less for later years of the century.

I have been keen to take a better look at the later period and I have now extended my original incomplete computer coverage of marriages for the March quarters of 1881 and 1891. In particular I am aiming for a complete data entry of the 1881/Q1 index. This is an index with just over 77000 entries, no small task. It would take a very long time without a good stint of work each and every day. Nevertheless the work is tending to confirm that the problems of the earlier years were still happening. For an idea of the size of this index, think of printing the 1881/Q1 entries at a little over 60 names per page - more than two reams of paper if printed single-sided.

The 1881 data files are about two thirds complete as this article is written. It felt rather apt that the half-way mark was passed when entering the "Jordan" surnames. "Crossing the Jordan" took on a new meaning ! As the data entry takes place, I regularly sort the entries into volumes and into page sequences within volumes. With ten percent of a file completed there is generally a clear structure of the districts within the volumes though there are still many large gaps in the sequence of pages. With forty per cent completed the structure was very well defined though still with some noticeable gaps. On the other hand there were already many pages for which three or four names had dropped into place. With over fifty per cent completed, gaps between pages are few and there are runs of many pages where the entries are complete. Past the sixty percent mark and the coverage begins to look very good.

It is only when a complete quarter has been extracted and sorted, however, that it becomes possible to deal with entries that are clearly wrong. Some fiche is so poor or the typeface so damaged that the page references are uncertain or easily misread. These generally show up as misplaced entries in the final files and many problems can then be resolved. An error is sometimes an actual error by the indexer or in the typing of the fiche and some of these errors can be resolved as well. Thus the completed and amended files become much better resources than the original indexes They also reveal the multiple index entries because they fall together when sorted by page reference, however far apart they may have been in the alphabetical index. I look at some of those below. It is only when data entry for the quarter is complete that any gaps or obvious errors can be indentified. In fact it is only the completed files that will make it possible to pass any judgement on the completeness or accuracy of the material.

These exercises are a little like working on a jigsaw puzzle. The first half of the pieces begin to create a fragmentary picture. The second half gradually fill in the missing details.

The 1881/Q1 entries are particularly useful because virtually all those married in that quarter will be in the 1881 census as man and wife. It may well be possible to use the LDS CDs of the census to identify most of the 1881/Q1 marriages. I have found a few without difficulty. With many of those marriage entries on computer, I may well be able to identify a bride for a known groom. If you find Albert Smith and Elizabeth in the census and believe they married in 1881/Q1, then I may be able to identify Elizabeth for you. In fact, to extend the options, my major files contain all the marriage entries for the March quarters of 1844, 1849 and 1856, nearly 200,000 names, and I will willingly try to find likely marriage partners within those files.

Unfortunately not everything is straightforward. I found one bridegroom in the Camberwell district who was likely to be of interest to a friend of mine. I also had entries for two matching brides. I can't find the groom or either potential bride in the 1881 census CDs. It is possible that the couple emigrated immediately after the marriage, or perhaps poor work by the enumerator has irretrievably disguised them.

At this stage in the work, my files now contain nearly 300,000 entries, roughly the same total as the FreeBDM project

Corrupted entries in the indexes
There is strong evidence that many marriage entries apparently missing from the marriage index are not actually missing but are wrongly entered. This recent work on the 1881 index suggests very strongly that the problem of badly written records remained very much alive. It is ironic that Michael Armstrong was able to write in Family Tree Magazine that even his own wife's much more recent index entry is wrong. Similar problems continue to be reported by users of the IGI and the LDS CDs of the 1881 census.

One useful window into this situation is the way that the bad writing forced the indexing clerks to index two or even three guesses at a name. In 1881, for example, an entry for the Shropshire district of Kington, on page 723 of volume 6a, appears three times :

    Henry Coombes, Henry Coomby, Henry Coventry

and a few pages later in the Kington record the clerk felt obliged to give Elizabeth Mary Bufton and Buxton. In Stockport, on page 33 of volume 8a, we have :

    Jane Athom, Jane Athorn, Jane Attrom.

Alderbury, volume 5a, page 241, gives us Augusta Elizabeth Bonny, Bown and Bowns. In the district of Erpingham the indexer has struggled with a name and came up with both Herricorn and Herrieven. In the Pontypool district one can only wince at the double entry of John Keef and John Kffe. These variations demonstrate the continuing poor quality of the writing and leave us with the conviction that the indexers must have mis-interpreted many of the names confronting them. I could quote many other examples. The 1881 census comes to our rescue with George Herricorn or Herrieven, showing George Herrieven. It is likely that the 1881 census will resolve many of these doubtful names for us (and probably provide some new surprises as well).

Variations of index entries like these are valiant attempts to be helpful but they do have a serious down side. They create variant entries that are wrong and can easily mislead another researcher into ordering a wrong certificate or making a wrong assumption. Someone looking in the GRO index for Henry Coventry in Shropshire will have no idea that his real name could be Henry Coombes, or vice versa. (If the index had identified individual marriages, as I suggested above, then this would not have been such a problem).

In 1881/Q1 we have James Hall and James Hill in Leigh, Lancashire, vol 8c p210. In Bury in the same volume we have George Hill and George Holt (p440). Which of these paired names is right and which is a trap ? We can't tell, but only one name is the right one.

As my file of index entries grows, an occasional page reveals a little secret. Page 807 of volume 1b is one of the pages for the district of Holborn (quite a large district in 1881) and I have five names referenced to that page. These are :

    Bridget Boland
    Samuel Brookes
    John Burnes
    Mary Ann Daniel
    Martha Ann Daniel

From 1852 there were two marriages to each page. Thus we now know that Mary Ann and Martha Ann are one person. It is possible that she used both names but it is much more likely that one or the other is wrong. From what I've seen in the marriage records at the GRO it is likely that one name is in the panel of details and one is the signature and is the name more likely to be correct. It is also very possible, and I have seen this in cases where a vicar has submitted the same marriage in error but differently in two quarters, that the difference is a copying error by the vicar. Whatever the background, the indexer has happily entered two different names in the index. The original register might perhaps give us the right answer.

(As a matter of interest, the immediate verso page, namely page 808 which belongs to the same parish within the Holborn district, has the four names Paolo Bergonzi, Amelia Eldridge, James Horrigan and Maria Miglionini. Normally we can't tell who married whom but in this case most would happily put money on the assumption that Paolo married Maria and Amelia married James).

These multiple entries are only one side of the coin. The other side of the coin is that the same poor writing has simply led to mistaken names and hence mis-written index entries. Such errors are not just predictable variants on the true name but they can be corruptions in very unpredictable ways, often far away in the alphabet from the true place. I am often asked for help in finding a missing name but the fact is that a corrupted name can be so different as to be extremely hard or even impossible to find, even hard to recognise if you happen to be looking straight at it.

Because my computer files of marriages are sorted into districts, if you believe that a lost marriage took place in one of "my" quarters and in a particular district, then I can examine that district for a possible outlandish corruption of your missing entry. It may be a slim chance but that can be better than no chance at all. One thing I learned at the GRO is that the indexers sometimes seemed to lose track of the current page number or district name for the space of several entries but never seemed to correct those mistakes when they put themselves right. I have quoted examples in "A Comedy of Errors".

In 1881/Q1 the indexer working through volume 8d (Lancashire) indexed the final page of the marriages in the Ashton district (which was page 624) and then began the next district, Oldham, beginning on page 625 of that volume. The very first index entry for the Oldham marriages is a bridegroom named Ashton Ashton ! I would love to check this one in the GRO film. Is it real or is it just a wild or frivolous slip by the indexer ? I have often suspected that the early indexers worked in pairs, one reading and one writing. It would help to explain some of the erratic happenings in the indexing. One well known genealogist has recently exposed a wild and totally fictitious family entry in the LDS CD of the 1881 census, obviously put in completely as a joke entry by one of the transcribers. I have wondered whether this has sometimes happened in the GRO indexes.

To summarise, we have the double tragedy that many entries are hard or impossible to pin down because of the inadequate design of the indexes and many others are corrupted and therefore lost.

I have said that the design of the GRO record system is an arbitrary one and has seen many changes in detail over the years.

I mentioned in "A Comedy of Errors" that I was surprised to find no entries for the district of Shrewsbury in my computer entries for 1881 and 1891. Now that those entries have risen to over 80,000 names I can be quite sure that Shrewsbury has disappeared. I don't know when it disappeared or for how long, but I do know that the Superintendent Registrar for Shrewsbury can only suggest that the Shrewsbury entries could have been filed by the GRO with those for Atcham. The numbers of marriage entries now seem to confirm that assumption. If you have been looking for Shrewsbury entries, try Atcham instead ! The question remains,"How often did this sort of thing happen ?"

In England and Wales there were rather over 600 separate and distinct Superintendent Registrars' districts. It was only when their separate quarterly data arrived at the GRO that these copies of the marriage records were put together into numbered volumes and given a structure. As far as I know, there is no published listing of GRO volume structures by districts such as the Page Range Tables that I was able to include in "A Comedy of Errors", based though they were on less than two per cent of the marriage index for the 1800s. I do know that some districts reappeared in the indexes after a period of absence. Perhaps they were there as superintendents' districts all the time, with the superintendents unaware that they were not appearing in the GRO indexes.

There were also many boundary changes (parishes being moved from one district to another) and some districts were merged and others subdivided. I have never seen any complete record of which parishes belonged to each Superintendent Registrar.

One can see that there are many unseen and often unsuspected traps for researchers. One might have hoped that the GRO would have documented and published all these structural changes if entries were to be found "without difficulty". Sadly no one thought to do it.

Big and small districts
This Shrewsbury discussion leads to another curiosity and emerges very clearly from my computer records. This is the fact that the districts of the Superintendent Registrars varied so enormously in size. Some were so small that it is hard to see why they should have existed at all. None of this is visible to the average user of the indexes.

Of the hundreds of examples that could be taken from my files, here are summary details for some Yorkshire districts in 1844 and 1856 to illustrate some of the differences. In 1844 the pages held up to four marriages and in 1856 two marriages.

Entries Pages
Total Number
of Pages

The structure of odd-numbered (recto) and even-numbered (verso) pages is explained in "A Comedy of Errors" but, very briefly, it reflects the gathering together of pages for individual parishes, some having entries on one side only and some on both.

Bradford, a major district with large parishes, submitted pages predominantly full on both sides. Hence the large number of verso pages and an average of nearly eight entries per page (four marriages) in 1844 and nearly four entries in 1856. The 648 index entries for Bradford in 1844/Q1 included at least 14 doubly indexed names and three or four likely aka entries.

York, though a city district, appears to contain more small parishes, hence fewer verso pages with data on them.

At the other extreme we have Patrington with marriages only on recto pages in these two quarters. Each page therefore represents a separate parish or register office. Many pages have only one marriage. Please don't write to tell me I have 19 and 21 entries. The 19 entries represent 9 marriages and one person indexed twice.

Some districts have changed dramatically. Billesdon, for example, recorded 38 entries in 1856/Q1 (19 marriages) but the first third of my 1881/Q1 file found just one Billesdon entry and is likely to finish up with very few marriages in the quarter. A second Billesdon entry popped up after processing 27000 names. These two entries happened to be on verso pages 32 and 34. Therefore I expected marriages on pages 31, 33 and 35 and indeed one entry for page 35 appeared at the 42% mark in the file. I can perhaps expect about 12 more Billesdon names in the rest of the index.

An almost identical example is the Superintendent Registrar's district of Easingwold in volume 9d. 1856/Q1 produced 36 names (18 marriages) but again the first third of 1881/Q1 has just one name. A second Easingwold entry popped up early in the second third of Q1. Again it is unlikely that there will be more than three or four marriages for the whole quarter.

The tiny district of Church Stretton has just three entries in the first 40% of the 1881/Q1 index. By contrast there are districts such as Liverpool, West Derby and Kensington with more than 400 entries each at the same stage in the extraction. There is huge variation in size and, incidentally, the size differences also carry differences in internal structures and densities.

I hope this little picture conveys something meaningful. These are just a few examples out of many. I'm painfully aware how hard it is to paint the right picture of these structures in a few short paragraphs. I have to admit to being fascinated by them.

Those "a" entries and what they can tell us
In "A Comedy of Errors" I discussed briefly the index entries that have page numbers such as 34a. Such entries have appeared in the indexes from their beginnings. In the handwritten indexes they are usually interlined, squeezed between the other entries and often almost illegible. In this note I am putting forward some new findings from the work on the 1881/Q1 index, including what I see as some new conclusions.

By way of an example of such an entry from the typed 1881/Q1 index :
    Harden Ann Kidderminster 6
    c 215
    Harden George Edwin Colchester 4
    a 386a
    Harden George Lawrence Kensington 1
    a 76
    Harden William Norwich 4
    b 235
Here George Edwin's marriage was added to the GRO records after the quarter's data had been assembled and indexed. At that time there was already a sheet in the GRO bundle with page numbers 385 and 386 and with marriages on both sides of it. George's marriage could be for that same parish or could be from a parish lying immediately after it alphabetically. Either way his marriage had failed to get to the GRO with the original quarterly returns. We can't tell when it arrived without looking at the GRO film in Southport.

There are quite a lot of these "a" entries. From what I have seen in the GRO records, they are either references to corrected entries or refer to entries that have been added to the record (having been previously omitted). For example my book describes the London synagogue that missed the marriages for a whole month when sending in data for one quarter. The omissions were sent in later and added as a set of pages 258a, 258b etc.

Most corrections arise from some form of representations from GRO customers, sometimes marriages missing from the GRO system and very often entries that had errors such as a wrongly stated name. My understanding is that no corrections have normally been made by the GRO of their own accord. This suggests that the "a" entries can only relate to a small part of the total errors and omissions. There are likely to be far more errors and omissions that are still undetected and uncorrected.

Have you noticed that these "a" pages are always even numbered pages ? There is a good reason for this. It is because they contain entries that arrived at the GRO after the main body of returns had been assembled into volumes and had been numbered throughout as sequential pages. Sometimes the additional entries have arrived many years or even decades later. I have found entries for the early 1800s that arrived in the 1900s.

When the page numbering took place, each volume of collected sheets was numbered consecutively and completely from beginning to end. For example volume 12 in the early records (Essex and Suffolk) begins with Billericay in Essex and finishes with Woodbridge in Suffolk and there are no gaps in the page numbers. Because many parishes were small, pages were often empty on the reverse (even-numbered) side. If an additional page comes in later, where can you put it in relation to the page numbers ? Wherever you put it in the collection of sheets, it always follows an even-numbered page and precedes an odd-numbered page. All you can do by way of numbering is to relate it to the preceding even page. Hence all those pages such as 512a, 274a and so on. These sit in the GRO records immediately after the correspondingly numbered original pages.

So far so good but this is only part of the story. In this latest work on 1881/Q1 I have found that these "a" entries can tell us much more. In the typed index material of this quarter most of the "a" references are included in the typing - they were additions at some point in time before the typing took place. BUT there are many additional entries written by hand at the bottom of the index sheets and these also include a large number of entries without the "a" or "b" in the page number. This is a highly significant distinction : -

    All the "a" entries in the index represent entries that have come into the system after the quarterly data submissions took place. They have the "a" in their page number because the GRO has had to insert them into an already complete numbering system. They are marriages sent in by clergy or registrars that were either missed from their original submissions or have since been found to be in error. Thus in one sense or another all these items can be described as corrections to "clergy errors".

    The handwritten additions to the index sheets that do not have an "a" or "b'" as part of the page number are references to marriages that were already in the quarterly data but were mis-indexed, missed or lost at some stage in the indexing process. Their page numbers are part of the original page numbering given when each quarterly volume was put together.

Readers of "A Comedy of Errors" will recall my findings that many names present in the GRO marriage films were missed from the indexes for various reasons, or were misindexed. Very many seem to have been lost in the process of typing the indexes. This category of handwritten additions to the indexes can therefore be described as corrections to "GRO errors" of one sort or another.

In the work on the 1881/Q1 index I have added up all the correcting entries as follows :

    1. Entries in the typed index with "a" or "b" etc   307
    2. Handwritten additions with "a" or "b" etc        164
    3. Handwritten additions without "a" or "b" etc    128

Do you notice one missing category ? If there are corrections (1) to "clergy errors" in the typed index with "a" or "b" etc, then there must also be many corrections of "GRO errors" in the typed index. We can't know how many because they don't show up like the ones with "a" on the end. If, as a rough approximation, they are in the same ratio as the handwritten ones then perhaps they are about 200 - 250 in the quarter, bringing the overall total to around 800 amendments in an index of less than 80,000 entries.

There is one possible clue relating to this unidentified category. For persons who have a second given name the handwritten footnotes almost always give only the initial of the second name. By contrast most such names in the typed part of the index (for 1881) are typed in full. Some, however, are typed with an initial and I suspect that these are ones that may have been added long after the index was written out but before it was typed. Even access to the GRO films would not give us the answer on this one.

These figures, then, give us a very rough idea of the number of errors corrected in the index we are using. They also confirm that errors were rampant throughout the 1800s. My belief is that this is only a small proportion of the total number of errors because :

    a) My understanding is that amendments are largely the result of customer action.

    b) The startling high level of errors that I described in "A Comedy of Errors" were undetected and uncorrected errors.

    c) The errors I found in the work at Southport were many times great than the errors I could detect from studying my files of sorted index entries.

    d) Many researchers continue to report that entries can't be found in the GRO indexes, or they report finding events via local registrars that can't be seen in the GRO indexes.

But there is yet another discovery to comment on here.

The "a" entries (the corrections to "clerical errors") come in pairs. Each pair is a reference to a bride and groom, a marriage record added to the system or amended. This has been apparent through all my work.

When I looked in detail at the "non-a" entries I fully expected most of them to come in pairs as well. Not so. Of all the 128 "non-a" entries as footnotes in the 1881/Q1 index, only two entries match up as a pair. All the rest are single unpaired names. What I strongly suspect, though I can't verify it until 1881/Q1 is all on computer, are that many of these are corrections of names lost from the index in the typing process. My work on the Taunton index, reported in "A Comedy of Errors", indicated substantial losses in the typing, perhaps around two per cent. The local Taunton index shows many brides and grooms where the GRO index has one party only. The completed 1881/Q1 index will show whether many of these footnote items slot into place as missing parts of marriages.

One example I have already noticed is in the district of Swansea, volume 11a. For page 680 I have found four names. Three of these (Frederick George Davies, Catherine Hopkins and William John) are all in the typed index. One name (Lucy Johns) appears as a hand-written footnote. The likely explanation is that she was missed when the index was typed and was added some time later because of a customer's query.

Within this group of handwritten entries I can certainly say that more and more are looking like corrections of mis-read spellings in the original indexings. Examples are Sledmoor, Guppy and Colton in the footnotes for Hedmoor, Guffy and Cotton in the index as typed. (The GRO has not seen fit to cross out the original entries, apparently keeping them as possible traps for the unwary). The typed index includes a William Limall in the district of St Albans. Later in the index is a late footnote amending this to William Minall. It would be interesting to follow this one through to the GRO records because it suggests a copying error by the clergyman in his quarterly return rather than an indexing error. It also shows how wildly wrong the entries can be.

One other radical amendment is to the index entry for George Inison in the district of Chelmsford. He has been amended in a late footnote to George Mison. Many other footnote entries may yet prove to be names lost from the typed index but I can't yet be sure All such amendments are very probably submitted by users..

Even some of the carefully added handwritten additions are in fact wrong. In 1881/Q1 I have seen wrong district names and wrong volume numbers among them. There is Kingston for Kington, and confusion between Mitford and Mutford. Such carefully written wrong data just seem to emphasise, to my mind, that a genealogical record in the hands of clerks who are not genealogists is bound to be liable to error. They shout out at me as soon as I see them. Another incorrect handwritten footnote is for William A Morrell, Shiffnal, vol 10a p819. Shiffnal is in volume 6a rather than volume 10a, and page 819 is a valid Shiffnal page in that volume. On the other hand page 819 in volume 10a is in the district of South Shields. Did the clerk give the wrong volume or the wrong district name ? It turns out that the answer lies within the index. When my file was sorted into page sequence the hand-written William A Morrell was found to lie next to William Allen Mowell who is indeed indexed as Shiffnal vol 6a page 819. Therefore the clerk's error was in writing 10a instead of 6a. The comedy of errors goes on and on.

One district with a lot of "a" entries in 1881/Q1 is the district of Epsom in Surrey. I now project that I will have just over 40 marriages for this district for the whole quarter but I already have evidence that at least six of those marriages were added to the index by way of "a" entries. These are marriages that are additional to those submitted to the GRO via the quarterly returns. I have never seen such a spread of additions in any district. Why so many late corrections for Epsom ? Three possible reasons suggest themselves :

    1. Epsom clergy had omitted more returns than other clergy.

    2. The superintendent registrar was better organised than others and had worked on improving the Epsom data.

    3. A keen researcher with family in Epsom district had uncovered omissions and reported them for correction.

I incline strongly towards the third reason. Reason (1) may also apply to some extent but Epsom in 1881 was probably a district of prosperous and hopefully well organised parishes. These added entries come from several different parishes because they are scattered through the GRO pages. If I had convenient immediate access to the GRO marriage films it would be very simple to look at these marriages to see exactly when they arrived at the GRO. They certainly arrived at very different times because some are within the typed material and some are added as footnotes (virtually ruling out reason 2). Even when you obtain a marriage certificate this valuable little piece of information about the timing of the return isn't included. There is no reason why it shouldn't be. Appendices 6 and 7 of "A Comedy of Errors" illustrate this part of the GRO record system.

The number and scatter of the marriages that failed to be included in these Epsom quarterly returns is disconcertingly high, especially for 1881. There must have been many other districts with as many omissions but with fewer that have been repaired. You may well be tempted to conclude that Epsom is a fairly typical district in terms of missing marriages and that many other districts have just as many omissions but very few corrections.

The Montgomeryshire district of Llanfyllin is also a strange one. Here a group of some seven marriages has been added by way of the handwritten footnotes, i.e. long after the index was first created. They are all indexed as pages 244a to 244d. Page 244 with two marriages was previously the last page for Llanfyllin with Holywell entries beginning on page 245. On the basis of the normal structure of the records we would expect these insertions to be those for the Llanfyllin Register Office. Hardly a good advertisement for the system. Furthermore we seem to have five entries for p244a of which one needs to refer to p244d. Finally one of the footnote entries is for a Rachel Davies who has been wrongly inserted in the footnote as Llanfyllin instead of Llandilofawr where she actually belongs.These insertions are all valuable but such errors are unfortunate.

It isn't very surprising that GRO performance seems to have seen little change since it was set up. Of all bureaucracies it is pre-eminently the one whose processes and structures have been shackled by legislation and it has always been very conscious of the need to conform to that legislation. The British High Commissioner to New Zealand mentioned in a recent talk that "bureaucracies dislike change". Perhaps that is also part of the story.

The memory of bureaucracies is often quite short. Staff come and go. Even the most junior employee of the GRO in the 1830s would have long retired from service by 1881. Many bureaucrats continue to make the same mistakes as their predecessors, especially in a system that is essentially doing the same job decade after decade. "Quality control" was a concept little known in the 1800s. Much of the work would have been unbelievably tedious to anyone without a genuine interest in it. In the 1800s the number of BDM certificates ordered by the public was probably quite small. Most of the records laboriously compiled would have remained largely untouched, hardly a situation to keep those early GRO clerks on their toes. The large number of the additional entries is clear enough evidence that the quarterly returns remained significantly incomplete during the 1800s.

As one footnote to this little discussion of amendments, let me also stress that the whole basis of the GRO record as a set of "certified copies" by the clergy required that any amendments to the GRO records, even decades later, had to be in the form of special copies made by the current incumbent of the parish concerned. The GRO itself couldn't simply make an amended or additional entry, even if there were blank spaces in the existing sheets.

As a second footnote, the GRO marriage films actually have far more "a" pages than we see in the indexes. Access is often via a scrawled note on the original such as "see p64a".

As a third but more encouraging footnote, most (but not all) of these "a" pages contain a single marriage. Therefore they can be good targets for the game of "matching the page reference".

Marriage entered twice ?
One practice that the Registrar General tried to stamp out in his booklet for the clergy at the end of the 1800s was the double recording of marriages that were performed first in a register office and then in a church setting. He was not against the repeated ceremony but ruled that the second ceremony was not to be entered in the register. My book shows several examples of such twice-recorded marriages picked from 1844. They would appear in the national marriage statistics as additional marriages.

In 1881/Q1 there is a striking repeated entry in the index for a Walter Edward A Goeters, once in Liverpool (vol 8b, page 174) and once in West Derby (vol 8b, page 759). It seems exceedingly likely that this is the same person and that I will eventually find a similar double entry for some young lady. This is not a case of a register office wedding and a church wedding. One can tell from the structure of the page numbering that neither wedding was in a register office. It is remotely possible that this may be a bigamous marriage, in which case the brides would not be the same. I've noted a few other entries that may well be for double ceremonies.

One other name that I noted in passing was Ellen Bosson which is indexed twice under Barnstaple, once as page 662 and once as page 703. I was nearly half way through the alphabet when I came upon John Jago with both those same references. Without doubt this is the same couple with two distinct marriage references in two different parishes. Neither is in a register office. Was this one of those repeated ceremonies or was one a "failed" ceremony that should never have got in the quarterly returns ? Only the GRO film will tell us the answer.

The Registrar General, however, didn't seem to object to repeated marriages if they were both church ceremonies and I find this curiously inconsistent. In his booklet of 1901 to the clergy he stressed that both such marriages were to be as much subject to banns, licence or certificate as any ordinary first ceremony. For a church ceremony following a register office ceremony it was only necessary to present the marriage certificate from the first ceremony (no banns being needed) but in this situation NO register entry was to be made. For repeat church ceremonies both parish registers would record the marriage and apparently, from email correspondence, this is still the case today. No reason is given for the two different positions.

Spotting these repeated marriages is often a matter of luck but I have found one other certain one in 1881/Q1. This is in the district of Weobly, where my file shows John Davies and Frances Griffin with reference 6a 665 and again with reference 6a 677. Page 677 is the last page in the Weobly entries (page 679 is the start of the Bromyard district) and this makes it virtually certain that page 677 comes from the register of a registrar and not from a parish register. I discussed this internal structure of the data in "A Comedy of Errors". The inference is that John and Frances married in the register office and later repeated the ceremony in a parish church. Thus this malpractice was still taking place late in the 1800s. For anyone wanting the true marriage certificate, the higher page number is almost certain to be the one. The other is not a true marriage in the eyes of the Registrar General.

Finding the Davies/Griffin marriage was largely chance. Since that time I have been noting other entries in passing that look as if they may be part of a repeated marriage ceremony (I have half a dozen at this point) but I can't be sure of confirming such situations until the quarter is completely captured. Even then it can be a time-consuming and only partly successful search. There may be plenty of districts with two Alfred Evans's for example but you may look in vain for evidence of two matching brides. An unusual name is a good lead. In "A Comedy of Errors" I have William Hagon Goble as an example from Brighton in 1844 and Whitfield Eddy from Carlisle.

In the entries for the district of Carnarvon we can see Jane Jones and William Jones with reference 11b p469 and a second time with reference11b p494. Welsh names, however, recur so terribly often that it can be rather rash to make too many assumptions.

In the Registrar General's eyes most of these double records are wrong. They create an apparently legal record of a non-legal ceremony and they also distort in a small way the national marriage statistics. From the point of view of genealogists, however, they are of enormous interest and add something special to any family history. It would be ideal if they can be flagged with a special symbol in the marriage index, perhaps a consideration for any new indexing process. The big difficulty, though, still lies in actually spotting the double records. They may well be in a different district. They may sometimes be in different quarters. Any ideas ? Does anyone have any examples ? I had a phone call the other day from someone who had found just such an apparent double entry in 1871 and wanted to find out if there was a reason for it. Both appear to have been church ceremonies. Her own researcher had never met such a situation before, but then not many people would find and ask for two matching references.

What "a" puzzle !
I mentioned above that there was confusion between Mitford and Mutford in two of those handwritten "a" entries. Frame 80 of the 1881/Q1 fiche has these two additions as footnotes :-
    Henry Harris Mutford 4b 422a
    Henry Harris Mitford 4a 950a
Page 950a of volume 4a is a Mutford reference and page 422a of volume 4b is a Mitford reference. A simple slip of the pen ?

Not quite, it seems. Imagine my surprise in looking through my computer file for volumes 4a/4b to find that Henry Harris in Mutford matches up with Susanna Chilvers on page 950a and that his namesake in Mitford matches up with Susanna Chilver on page 422a. Both hese marriages appear to have been church ceremonies. (The clerks got Mitford and Mutford right for Susanna's entry).

The chance of spotting such a repeated marriage in two different volumes is minimal. If the repeat ceremony happened in a different quarter, and this must surely have happened, then discovery would be almost impossible. I begin to suspect that these repeat marriage entries are far more numerous than I had thought. It also makes one wonder how many family trees may contain the wrong one of a pair of marriage entries and hence a wrong marriage date.

There is an even more striking aspect of the two marriage entries for Henry and Susanna(h), namely the fact that both these entries were handwritten additions that were put in at some time after the original indexes were typed, i.e. a long time after the event. My research outlined in "A Comedy of Errors" has pointed to marriages that have failed to arrive at the GRO by way of the quarterly returns. What can we say about two marriages for one particular couple in two different parishes and in two different registration districts, both of which failed to reach the GRO ? They are two separate records in two separate parish registers. If missing marriages in general were few, then the chances of one couple being missed twice would have to be infinitesimal. To my mind this is another straw in the wind suggesting that the problem is a far bigger one than has ever been suspected.

The future of the records
Preliminary discussions about the future of the records and what can be done to improve them have raised the issue of the possible cost of a major revamp.

I can only say that all the research on the system, both as reported in "A Comedy of Errors" and as I have continued to find in recent months, paints a picture of problems created by inadequacies of the system. I believe this article extends that evidence. The stated objectives of the system have been frustrated by errors and omissions at all stages in the compiling and indexing of the records, compounded by the absence of effective controls and checks on what was happening. Against this background of default it seems totally necessary for a solution to be developed that will remedy the faults of the past by going back to the primary records and creating new indexes. Anything less would be a frustrating waste of effort that will be regretted for all time. My work continues to emphasise the extent of the problems. A revamp seems the only appropriate remedy for a system that has been so inadequate for so long. It is the only way of achieving the objectives that were stated so emphatically by that first Registrar General.

Surnames as given names
Surnames adopted as given names are sometimes a great help in research. My own middle name of Whitfield has been in our family for at least eight generations. I am happy to search my files for particular surnames used as middle names. It may give you a new line of enquiry. Please, though, not names that are common as surnames, like Hall,Williams, Parker and the like.

When I have finished the 1881/Q1 records I intend to make some numerical comparisons with the 1856/Q1 data. I believe there can be much to be learned from such a study. For example the relative numbers of variant index entries may tend to indicate whether there has been any significant change in the quality of handwriting over the 25-year period.

Relative numbers of odd and even pages and overall density of the records may throw a little light on average sizes of parishes.

I shall give more thought to this once the present work is completed.

Michael W Foster
19 Khouri Avenue
Wellington 6005 New Zealand


Rodger Goodger of the Bucks email discussion list offered the following 1881 Census CD tip on September 17th which I thought Gazette Readers would find most interesting:

If those of you with the CD-Rom for the 1881 Census have not already discovered this, it may be helpful to mention that on looking up one of our entries the husband was missing from the household. We knew from the earlier 1881 Census on fiche that he WAS in the house on Census night but his name was the last entry on a page, with his wife and family appearing on the following page.

The CD transcription lists them as separate households, and we found this out by using the "neighbours" button. There he was in the previous entry on his own!. So check the neighbours if someone is missing.

About Fawne Stratford-Devai
Fawne Stratford-Devai's work on Land Records and early Ontario records is well known in the genealogy community. A published author of several Canadian and UK research books, she has also contributed articles to the Ontario Genealogical Society's newsletter "Families" as well as writing for the online family history newsletter the "Global Gazette". Biography Inc. 1992-2017
Sign up for our free newsletter!   |   Unsubscribe from our newsletter

History & Master Roll
Jessup's Loyal Rangers

Loyalists - American Revolution