Name structures tend to be quite simple: a single surname, with the exception of true patronyms, but even these are primarily single-element bynames. There is a chaotic mixture of true patronyms, inherited English-style surnames, patronyms treated as fixed surnames (even to the extent of gender confusion), and the optional adoption by women of a husband's surname (even when it appears to be a true patronym), or even his given name as a surname.
Despite the efforts of English beaurocrats, surnames were not fixed either in the sense of being inherited unchanged, or in the sense "one person, one surname". There is only marginal evidence for the use of ad hoc personal nicknames or occupational identifiers, but the use of true patronyms still continues strongly.
Looking particularly at the dynamics of "married names", we see that while they are common, they not only aren't the rule, but are considerably less common than the alternative. When both the husband and wife's name are known, only in one third of the cases does the wife appear to have taken a "married name". When the names of her children are known, in only a third of the cases does her surname match theirs (and, presumably, their father's). When a woman's parents' names are known, in only one quarter of the cases is it probable that she bears her husband's name. When a brother's name in known, in nine times out of ten, her name matches her brother's and not her husband's. The first two cases are probably the best guideline to frequency, as these are the cases where we know that the woman is, in fact, married. So as a rule of thumb, only a third of married women adopt some form of their husband's name as a surname -- and some of them are known alternately by what is presumably a "maiden name".
Editted and published by Arval Benicoeur