I have also visited the Somerset Record Office, the Bristol Record Office, the Wiltshire Record Office, the Society of Genealogists, and the Public Record Office to get access to records. Also included is information received from several people with whom I have corresponded. The data is currently in Family Tree Maker for Windows, and has been converted to HTML via the Microsoft Access database using a program I wrote in Visual Basic 4 that reads the GEDCOM file from FTM.
The records are based on separate individual families that are not linked across the generations. Hence, the same person could easily appear twice in these records: once as a parent and once as a child. The method of access to the records has been designed to permit users themselves to explore the inter-generational links using indexes of fathers and children. In these records the father's birth year was imputed as discussed below.
Although the concept of not providing links across the generations appears to involve more work (both for me and you), it was eventually adopted as the solution to a difficulty I faced not infrequently. In this, I sometimes would find a couple having children baptized in a town, but when I looked back a generation in that town I would find two possible candidates in the form of children baptized with that given name a few years apart. In a linked genealogic system, I would have to identify one of those children as the eventual father in the later family. Prior to about 1840, that was frequently impossible to do reliably, since there just isn't sufficient data around to make the determination. Consider the following example. Two families: John and Mary, and James and Ann, were having children in the same town at about the same time. John and Mary have John baptized in 1763. James and Ann have their John baptixed in 1765. In 1789, a John marries Hannah in the same town. Now, which of John(1763) and John (1765) was the husband of Hannah? Without further information, who knows? A look at the other children's names might provide a clue, but a lot of speculation might be involved. A similar situation exists where a family crops up in a village where there were none of that surname before, indicating in-migration. How, to make links from families in the previous generation in other places?
The only solution I could see, is for me not to make the links, but
to allow the user to do so. However, to provide some guidance, I did provide
some suggested links, with a code indicating
degree of likelihood. Note that this solution works well for the early
years, when there is little ancillary information, but becomes a little
ungainly in later years, when census and marriage records provide direct
information on parentage and places of birth. However, in my mind it was
the best way to go.
For the early years (roughly pre-1760, when no other information on age is available either from the censuses or burials), the father's probable birth year is imputed from (marriage year - 25), or (first-born baptism year - 26), following the guidelines of the Silicon Valley PAF User's Group. Similarly, the mother's birth year was set at two years after the father's, again using the same guidelines. Later, when the people started to turn up in census records it was possible to be a bit more exact (though census records aren't necessary that trustworthy - again, judgment is needed). Census or other information was used if I did not have a baptismal record. Burials are entered only in those cases in which I felt sure that the information was either explicit or was sufficiently definite. Marriage entries are included if they seemed reasonable (names agreed, and date of marriage was reasonably close to baptism of first child).
Notes are attached to most individual records. These supply information I have gleaned from various sources, including wills, parish records, and other documents. The notes also discuss doubtful points and related issues. For the fathers, the notes include a section providing hints on which family he might have belonged to in the previous generation. For each child, each record contains a section suggesting possible links to families in the next generation. In both cases, the number of asterisks indicates the perceived strength of the link: *** meaning almost certain, ** meaning some doubt, * your call!
To save time in downloading files, I have purposely not provided lines for information I do not have. For instance, if I have no data on a person's burial or death, I have left out the lines for death and burial. Birth years for children are not shown if I have a baptismal record, but are if the information comes from some other source.
Note that I can provide information only on what I have seen. Later families will be incomplete, because I typically stopped searching at the 1851 census. Hence, families having children at that time will often be shown with only the children born before 1851. A similar situation occurs with baptismal records. I have not searched every parish register from start to end (though have done it with some). Hence, individuals could be missed for this reason. My suggestion is to use these records as a starting point, and do your own verification. I am willing to be corrected, and even more willing to receive new information!
This information I tried to put into order, first by organizing into families, and later by trying to link the families into generations. Later, as the information became available, and as I contacted and was contacted by others, I followed those families into other counties and countries. From all of this I created a report, which is currently in the LDS library in Salt Lake City. It is from this report that the information in these pages is derived.