in the Mémoires de la Société des Antiquaires de Normandie, 1837, vol. xi. p. 179, and, after him, by M. Florent Richomme, in a pamphlet published at Falaise under the title of La Naissance de Guillaume-le-Conquérant à Falaise. There is no doubt that William was born in 1027 or 1028; M. Deville endeavours to fix the exact date to June or July, 1027. William was seemingly between seven and eight when Robert set out on his pilgrimage. "Habebat tunc,” says William of Malmesbury (iii. 229), "filium septennem." So Wace (14360);

"N'aveit encor que sol set anz,

Quant li Dus Robert se croisa
Et en Jerusalem alla."

seems to be fixed to January, Deville from the Departmental

Petit esteit, n'ert mie granz, The date of Robert's departure 1035, by a charter quoted by M. Archives at Rouen. It is granted by Robert on the Ides of January,

quo et Hierusalem petiturus ibi licentiam eundi a Deo et sanctis ejus petii." But it is argued that William was full eight years old when the news of his father's death reached Normandy, and when he was accordingly invested with the Duchy. William of Jumièges (vii. 44) calls him "fere sexagenarius, anno ducatûs in Normanniâ LII," at his death in September, 1087. This puts his birth in 1027, and his accession in 1035. Orderic (459 D) says that, at his accession, "tunc octo annorum erat," and again (656 C) William is made to call himself at that time "tenellus puer, utpote octo annorum.” It is therefore inferred that William attained the full age of eight years at some time after his father's departure, but before his death, or at least before his death was known in Normandy. For this purpose six months or thereabouts is allowed, and it is thus ruled that William was eight years old in June or July, 1035, and was therefore born in June or July, 1027.

I am not fully convinced by these arguments. The expression of William of Jumièges, "fere sexagenarius," would seem to imply that William was not fully sixty in September, 1087, and, if he succeeded in July, 1035, he would then be in the fifty-third, and not in the fifty-second, year of his reign. Orderic indeed (459 D) says that he reigned fifty-three years, but, as he succeeded in 1035 and died in 1087, he certainly did not reign fifty-three years full. And Orderic's chronology is very confused on the matter; in the passage (656 C) where William calls himself eight years old at his accession, he calls himself sixty-four years old at his death (“mala

quæ feci per LX quatuor annos"). This would put his birth in 1023, quite contradicting Orderic's other statement. Moreover the Chronicle of Saint Michael's Mount (Labbé, i. 348) calls him "septennis" at the time of his accession. It seems to me therefore that it is not safe to attempt to fix the date of William's birth so minutely as M. Deville does, but that it certainly happened in 1027 or 1028, and more probably in 1027.

M. Deville connects the birth of William with that siege of Falaise which made Robert submit to his brother Richard (see vol. i. p. 464). This, and the death of Richard, he places in August, 1027. But William of Jumièges (vi. 2) distinctly says that Richard died in 1028, after a reign of two years. Orderic (459 D), by making Richard reign a year and a half, might agree with M. Deville. Most of the Chronicles however make Richard die in 1026, the year of his accession. See the Chronicles of Fécamp (Labbé, i. 326), of Rouen (i. 366; cf. Duchesne, 1017 B), and of Saint Michael's Mount (i. 348). The authority of William of Jumièges is no doubt much the highest, but his chronology is inconsistent with M. Deville's view.

M. Deville has however done good service in bringing prominently forward the fact, which is commonly forgotten, that Robert, at the time of his first amour with Herleva, was not yet Duke of the Normans, but only Count of the Hiesmois, in which character Falaise was his capital. He has also well pointed out his extreme youth. Robert was the second son of Richard and Judith. The marriage contract of Judith, dated in 1008, is given in Martène and Durand's Thesaurus Novus, i. 123. Robert could therefore hardly have been born before 1010; he could have been only eighteen at the most at the time of the birth of William, and only twenty-five at the time of his pilgrimage and death. His brother Richard, the father of the Abbot Nicolas, must have been equally precocious. Edward the Third too was only eighteen years older than the Black Prince; but at any rate he was married.

III. That William was born at Falaise all accounts agree; but there is not the faintest authority for placing his birth in the present donjon. M. Deville says that the tradition is a very modern one. A room is shown as that where William "fut engendré et nâquit,” and a sufficiently absurd inscription commemorates the supposed fact.

But we have seen (see above, p. 174) that the

existing keep is, in all probability, of a later date than William's birth; and, if it did exist in Robert's time, and if William were born in the castle at all, it is far more likely that Herleva would be lodged at such a time in some other part of the building, and not in the keep. The keep was not the common dwelling-place of the lord of a castle, but only his occasional place of defence. See Mr. G. T. Clark, Old London, pp. 14, 39, 43.

But there is another statement which, if it be trustworthy, as it seems to be, puts it beyond all doubt that William was not born in the castle at all, but elsewhere in the town of Falaise. The local historian of Falaise, M. Langevin (Recherches Historiques sur Falaise, 1814, p. 134), says, on the authority of "les anciens manuscrits extraits du chartier" of Trinity Church, Falaise, that William was born in 1027, in that parish, in a house belonging to him-that is, seemingly to his mother or her father-in the old market-place, and that he was baptized in Trinity Church. See Richomme, p. 12, who follows Langevin. One would like to have the exact extracts from the manuscripts, and to know something of their date; but in any case they are better authority than a romantic modern story, which seems not even to be a genuine tradition.

IV. Most writers state, or rather assume, that William was the only child of Robert and Herleva. The lioness was bound to bring forth only a single cub. But Mr. Stapleton, who pried into every corner in Norman matters, has, in a paper in the Archæologia (xxvi. 349 et seqq.), brought some strong arguments to show that William had a sister by the whole blood, Adelaide or Adeliza, wife of Ingelram, Count of Ponthieu. This Adelaide was the mother of two daughters, one of whom bore her mother's name, and married Odo of Champagne; the other was Judith, the too famous wife of our Earl Waltheof. The elder Countess Adelaide has been commonly taken to be only a half-sister of William, a daughter of Herleva by her husband Herlwin. She appears to have been so looked on by the continuator of William of Jumièges (viii. 37), who calls the mother of Judith "soror uterina Willelmi Regis Anglorum senioris," words which he would hardly use of a daughter of Robert. Still Mr. Stapleton's case is very strong. It rests mainly on a charter, which Mr. Stapleton prints, granted to the College (afterwards Monastery) of Saint Martin of Auche (Alcis), near Aumale. Adelaide is there

distinctly called the wife of Ingelram and sister of William, and her daughters, Adelaide and Judith, are spoken of. After the death of her husband, she enriched the church of Saint Martin, and, while still young ("quum esset adhuc in juvenili ætate"), she had it hallowed by Archbishop Maurilius. Now Count Ingelram died in 1053, and Maurilius was Archbishop of Rouen from 1055 to 1069. Mr. Stapleton thinks that these dates better suit a daughter of Robert and Herleva, who must have been born between 1028 and 1035, than a daughter of Herlwin and Herleva, who could not have been born before 1036. There are also two statements which, though erroneous as they stand, point to the parentage argued for by Mr. Stapleton as their groundwork. Thus Orderic (522 C) makes Odo of Champagne marry a sister of William and daughter of Duke Robert. The two Adelaides, mother and daughter, are here confounded, but the fact that Duke Robert had a daughter is preserved. So Robert de Monte, under the year 1026 (Pertz, vi. 478), preserves the name of Aeliz or Adelaide, daughter of Duke Robert, though he makes her the child of another mistress and not of Herleva. This is doubtless an attempt to reconcile the existence of Adelaide with the belief that William was an only child.

The Norman writers, it must be remembered, know nothing, or choose to say nothing, of the marriage of Robert with Cnut's sister Estrith. See vol. i. p. 468. They look upon Herleva as Robert's only consort, lawful or unlawful. So William of Malmesbury, iii. 229; "Unice dilexit et aliquamdiu justæ uxoris loco habuit." But no writer asserts any actual marriage, except the Tours Chronicler in Bouquet, x. 284. He marries Herleva to Robert soon after William's birth; "Dux Robertus, nato dicto Guillelmo, in isto eodem anno matrem pueri, quam defloraverat, duxit in uxorem." He also transfers the story of Herleva from Falaise to Rouen. Possibly also some notion of a marriage may have floated across the brain of our own Knighton, when he said (2339) that William was called " Bastardus," quod ante celebrationem matrimonii

natus est."

66

The story of the Tours Chronicler cannot be true, as such a marriage would have legitimated William, and he could not have been known as William the Bastard. But Herleva might seem from William of Malmesbury's words to have been looked on

as something more than an ordinary concubine. It is strange that he should be the only writer who makes Herleva marry Herlwin during Robert's lifetime. His words (iii. 277) are, "Matrem, quantum vixit, insigni indulgentiâ dignatus est, quæ, ante patris obitum, cuidam Herlewino de Comitisvillâ, mediocrium opum viro, nupserat." But William of Jumièges (vii. 3) distinctly puts the marriage after Robert's death; " Postquam Hierosolymitanus Dux obiit, Herluinus quidam probus miles Herlevam uxorem duxit, ex quâ duos filios, Odonem et Robertum, qui postmodum præclaræ sublimitatis fuerunt, procreavit." According to Orderic (660 B), Herleva was the second wife of Herlwin, whose son Ralph by a former marriage was also promoted by William. The honours shown by William to his mother seem to have struck writers at a distance. Besides William of Malmesbury just quoted, the Tours Chronicle in the French Duchesne (iii. 361) says, "Matrem dum vixit honorifice habuit," and the Limousin writer William Godell (Bouquet, xi. 235) says, "Guillelmus Rex matrem suam, quamvis esset inferiori genere orta, multum honoravit." He goes on to mention the promotion of her sons.

Of the sons of Herleva, Odo and Robert, I need not speak here; but I may mention that she had also a daughter by Herlwin, named Muriel, who has naturally been confounded with William's other sister Adelaide. Wace says (Roman de Rou, 11145),

"Ki à fame avait Muriel,

Seror li Dus de par sa mere
E Herluin aveit à pere."

See Taylor's note, p. 102.

One would have thought that the story of Robert and Herleva was one which could never have been forgotten. Yet later writers did not scruple to provide the Conqueror with new and strange mothers. Thomas Wikes, the royalist chronicler of the thirteenth century (Gale, ii. 22), gives William the following wonderful pedigree. He was "natus ex nobilissimâ muliere Matilde, quæ fuit filia strenuissimi militis Richardi dicti Sanz-peur, filii Willielmi Lungespeye, filii Rolandi, qui fuit primus Dux Normannorum." And in an unpublished manuscript of the famous Sir John Fortescue of the fifteenth century (for a knowledge of which I have to thank the Right Hon. Chichester Fortescue), William is