> As the database gains more families, I was trying to avoid several Marys > in the index that are not traceable to a family. Same problem if I > identify Mary as Mary Unknown. > > Keith Nuttle If you have several Marys not traceable to family then you have several Marys not traceable to family. There's no getting around it. I think you have several separate issues. One is data & another is presentation. From a data point of view good /database/ practice would be to leave a null which is distinct from an empty string or a string filled with spaces and simply means that the data isn't known. It's possible your application's underlying database engine won't support this. And probable that you won't be able to find out! Nevertheless the closest you'll be able to get to this is to leave the surname empty if you don't know it - with luck this will be treated as null and is the closest you'll get if it isn't. The presentation issue depends on the application you're dealing with. If display, reports etc. are configurable you might be able to make the application default to some other name if the birth name isn't known. There may be other options. Gramps, for instance, has provision for multiple names and the ability to specify one of several types against each name and to designate one as the preferred name which will be used in displays and reports. Using such a scheme could initially set the preferred name as Mary Jones and flag it as a married or unknown type, switching the preferred name to her birth name if and when you discover the correct surname. Another facility with Gramps is the provision of a field called "Call name". You could set this up as "Mary MiL of Jim Jones" and add the field to the display. Even if you leave the preferred name as the baptismal name with a blank this will still allow you to see which Mary is which. This approach would also allow you to distinguish between people with the same name, for instance "Jim Jones I", "Jim Jones II" etc. The biggest issue which another posting has touched on is the fact that expecting a surname, especially one with a standard spelling, is historically and culturally naive. I have several lines where surnames are very variable in spelling even in recent times and I'm pretty sure my own surname started with a single "d" in the middle. In earlier times there were all sorts of patronymics and other epithets in use. What do you do with an individual who was described in the witness list of one deed as "Edmund the tanner" and in another, drawn up on the same day, as "Edmund the barker"? A capable genealogical program should be able to handle such situations gracefully without twisting either the data or the program out of shape. -- Ian Hotmail is for spammers. Real mail address is igoddard at nildram co uk Ian Goddard <goddai01@hotmail.co.uk>