The above forms are curious, but not without parallel. Thus the descendants of Urse d'Abetot are spoken of as 'Ursini' in Heming's Cartulary. Æthelred of Rievaulx speaks of 'Poncii' and 'Morini' as present at the battle of the Standard; Gerald, in a well-known passage (v. 335), speaks of the 'Giraldidæ' and 'Stephanidæ', and Orderic, we have seen, of the 'Geroiani'.
The doubly influential character of this descent is well illustrated in this passage (quantum valeat) from the chronicle of St John's Abbey, Colchester.
Parcebatur tamen Eudoni, propter genus uxoris ipsius Rohaisæ: erat enim hæc de genere nobilissimo Normannorum, filia scilicet Ricardi, qui fuit filius Gilbert Comitis, duxitque Rohaisam uxorem, quæ erat soror Willelmi Giffardi, Episcopi Wintoniæ. Itaque, cum fratres et propinqui junioris Rohaisæ quoslibet motus machinaturi putarentur, si contra maritum ipsius aliquid durius decerneretur, sic factum est ut interventu predicti Episcopi, etc., etc.
This passage is, I believe, the sole evidence for the real parentage of Bishop William. It was clearly unknown to Canon Venables, who wrote the Bishop's life for the Dictionary of National Biography.
Like most of these 'foundation' histories, this document is in part untrustworthy. But it is Dugdale who has misread it, and not the document itself that is responsible for the grave error (Baronage, i. 110) that Eudo's wife was 'Rohese, daughter of Walter Giffard, Earl of Buckingham'. Here again, as in the Tirel case, the daughter of a Clare, by a Giffard, is converted into a Giffard. The error arose from referring the 'qui' to Eudo instead of to his father-in-law, Richard. The 'Historia' is perfectly consistent throughout in its identification of the younger Rohese, of whom it states that 'commorata est marito annis triginta duobus, cui ante habiles annos nupta est' (iv. 609).
In asserting under 'Clare' (Baronage, i. 208) that Eudo married the widow (not the daughter) of Richard, Dugdale relied on another and more inaccurate document (Mon. Ang., v. 269) which actually does speak of
Rohesia una sororum Walteri [Giffard secundi]—duas plures enim habuit—conjuncta in matrimonio Ricardo filio Gilberti, qui in re militari, tempore Conquestoris, omnes sui temporis magnates præcessit—
as marrying Eudo Dapifer after her husband's death. But we must decide in favour of the Colchester narrative: Eudo's wife was her daughter and namesake.
We see then that Walter Tirel was son-in-law to Richard de Clare, who had enfeoffed him in 'Laingaham' before 1086. Now this 'Laingaham' was Langham in Essex, just north of Colchester, which gives us an important clue. Walter's widow 'Adeliz' was in possession in 1130 (Rot. Pip., Hen. I) because, as we have seen, it was probably given her by her father 'in maritagio'. But her son Hugh held it under Stephen, and Anstis saw among the muniments of the Duchy of Lancaster a mortgage of it by Hugh to Gervase 'Justiciar of London'. I have not yet identified this 'mortgage', but the confirmation of it to Gervase de Cornhill by Earl Gilbert de Clare, as chief lord of the fee, is extant,[1] and its first witness is Earl Gilbert of Pembroke, so that it cannot be later than 1148, or earlier than 1138 (or 1139). Moreover in yet another quarter (Lansdown MS. 203, 15 dors.) we find a copy of a charter by this latter Earl Gilbert, belonging to the same occasion, which runs as follows: