[114] In the last chapter we have been primarily concerned with the decline of the royal government south of Poitou and Burgundy as we traced the civil wars and invasions of this period and the rise of certain noble families to positions of hereditary, de facto independence in regions they had originally controlled as subordinate officials appointed by Carolingian monarchs. This, however, leaves unanswered the important question of how these developments changed the governmental system as established in the reigns of Charlemagne and Louis the Pious. What has become of government in Southern France and Catalonia by the year 900?

At this point it might be well to reiterate that in the early ninth century the Carolingians entrusted their government to counts who could be assisted in performing their duties by certain subordinates -- viscounts and vicars -- and who were often supervised by special officials or missi sent out by the monarchs. Such a system was in theory the one established everywhere in the Carolingian Empire and implemented by capitularies and other decrees issued by the monarchs. How did this system actually work out in practice, though, in our regions? What use was actually made of viscounts and vicars in Southern France and Catalonia during this period?

Here we note an interesting fact. We have little evidence of the use of such subordinate officials by counts during the reign of Louis the Pious. In fact, clear evidence exists for two only -- one of Thouars in Poitou, who is mentioned in 833,⁽¹⁾ and a Viscount Stephen who presided over a court [115] in Narbonne in 834 which considered a case involving the aprisia of John of Fontjoncouse.⁽²⁾ We hear more of them, however, during the early years of the reign of Charles the Bald. In a charter given a church in the Pallars-Ribagorça region by Count Frédélon about 850, we find a reference to viscounts and other officials of the court.⁽³⁾ In 855 there is a mention of a certain Viscount Lambert in Angoulême.⁽⁴⁾ In 859 another charter from Roussillon mentions a Viscount Richelme.⁽⁵⁾ Four years later, after being called a fidelis by Charles the Bald, a certain Isembert is called a viscount and missus in the Narbonne area.⁽⁶⁾ We also find several references to missi during these years. Count Bernard of Auvergne refers to himself as a missus in presiding over a court in a case involving a serf of the royal fisc in 851,⁽⁷⁾ and Count Saloman is called a missus in a case involving the archbishop of Narbonne in 862.⁽⁸⁾

After 870, judging by our sources, royal missi disappear, but we find viscounts appearing throughout the Midi and Catalonia. In Provence and the Valley of the Rhone, for instance, where we find counts mentioned at Vienne in 842⁽⁹⁾ and at Marseille in 845,⁽¹⁰⁾ but no viscounts, by 870 such subordinate officials begin to appear. The first reference to them dates from this year when we find two viscounts assisting Count Gerald in a court held at Vienne.⁽¹¹⁾ In 902 we find another assisting Count Hugh of Arles in a similar capacity in the same city.⁽¹²⁾

[116] In Septimania at Nîmes we find a viscount who is presiding over a court in 876, another acting in the same capacity in 892 and still a third present at a court headed by Count Raymond of Toulouse in 898.⁽¹³⁾ In Béziers too, where Appolonius was count as late as 872, viscounts appear also. A certain Rainald, called a fidelis in 881 by King Carloman by 897 seems to be serving as viscount and in the same year is succeeded by his son-in-law, Boson, who bears the same title.⁽¹⁴⁾ In 875 in the Narbonne Carcassonne region we find a certain viscount, Berold, who is presiding over a court⁽¹⁵⁾ and a little later in 883 a viscount, Sunifred, present at a tribunal headed by Count Oliba II.⁽¹⁶⁾ In 898 in the same region Viscount Aton of Albi presides over another court as a special missus or representative of Count Eudes of Toulouse -- a court in which Count Oliba II this time is referred to as a missus. At about this same time we also find a viscount at Toulouse.⁽¹⁷⁾

In Roussillon and the rest of Catalonia we again find such officials. In the former region in 875 there is record of a court held with a missus of Count Bernard of Gothia presiding.⁽¹⁸⁾ Some four years later we find a reference to a Viscount Franco, called a fidelis of Count Guifred of Barcelona-Cerdanya.⁽¹⁹⁾ In 898 we learn of a Viscount Ermenico who is assisting at a court presided over by Count Sunyer II of Ampurias-Roussillon.⁽²⁰⁾

We have no proof that viscounts existed in Auvergne during these years. But in 876 we first hear of a viscount of Limoges, a certain Hildebert, who is called a fidelis in a charter given him by Charles the Bald.⁽²¹⁾ In the same year we again learn of a viscount at Thouars in Poitou.⁽²²⁾ A little later in 898 we find two viscounts in the Limousin, Ademar and Gausfred, who presided over a court held near Limoges.⁽²³⁾ In the same [117] year we hear for the first time of a viscount at LePuy, the first of the celebrated Polignac family.⁽²⁴⁾ To the west in the Angoulême region we learn that Count Vulgrin, sometime between 869 and 878, has appointed a certain Viscount Ramnulf as castellan of the fortress he has built to protect this region from Viking attacks.⁽²⁵⁾ A charter of 879 shows him still holding this position in Angoulême.⁽²⁶⁾ Without in any way denying the possibility that a number of other viscounts were placed in office in the Midi and Catalonia during this period, what seems clear is that we do have proof of their existence by 900 at Vienne, Nîmes, Béziers, Narbonne, Carcassonne, Toulouse, Roussilon, Ampurias, Ausona, Angoulême, the Limousin, Poitou, the Albigeois, and Velay.

Why do we find such viscounts in so many regions of the Midi and of Catalonia at a time when they seem to be rare in the rest of the Carolingian Empire? No final response to this question can be given, but a plausible answer seems to be that during this period single individuals came to control many counties or honores. Under the circumstances counts and marquises like Bernard of Septimania, Bernard of Gothia, or Frédélon and Raymond of Toulouse needed help in governing the regions entrusted to them. True, they could use counts for the task and subordinate such counts to themselves by special ties of vassalage or fidelitas. Such a system seems to have been used in the Quercy-Limousin region by Raymond of Toulouse about 860 where he appears to have subordinated the counts of Turenne to his authority.⁽²⁷⁾ It was certainly attempted in Carcassonne where in 872 Count Oliba II did homage for his county to Count Bernard of Toulouse.⁽²⁸⁾ Such subordination may have continued down to 898 and even afterwards.⁽²⁹⁾

But the difficulties arising from such subordination of one count to another were very real, as the murder of Count Bernard in 872⁽³⁰⁾ and the revolt of Count Asnar Galindo in Urgell in 838⁽³¹⁾ dearly showed. It therefore seems probable that counts who controlled a large number of [118] honores generally preferred, when possible, to entrust power to subordinates who, bearing the title of viscounts, were far enough down the ladder not to attempt to usurp the authority entrusted to them.

Therefore, it was to a viscount that Bernard seemed to have entrusted Narbonne in 834 and viscounts whom the counts of Toulouse used in governing the trans-Pyrenean lands of Pallars and Ribagorça down to 872⁽³²⁾ and a Viscount Richelme whom Marquis Humfrid appointed as a missus in Roussillon in 862.⁽³³⁾ After 870 similar magnates, controlling a large number of counties, seem to have used viscounts more extensively. At a time when Count Gerald of Vienne and Count Hugh of Arles controlled large areas in the Valley of the Rhone we find viscounts appearing there,⁽³⁴⁾ just as the family of the counts of Toulouse seem to have established them in Nîmes, Béziers, Narbonne, Toulouse, Albi, and the Limousin.⁽³⁵⁾ The counts of Carcassonne, who controlled both Carcassonne and Razes, seem to have established such officials in Carcassonne,⁽³⁶⁾ while Marquis Bernard established them in Razès⁽³⁷⁾ and Count Guifred seems to have been responsible for their appointment in Ausona and other portions of his domains.⁽³⁸⁾ It was probably Duke William the Pious of Aquitaine, who controlled many counties, who established a viscount of Le Puy in 898,⁽³⁹⁾ just as it was Count Vulgrin, controlling several counties, who seems to have established the first viscounts in Angoulême.⁽⁴⁰⁾ Even the viscounts of Thouars may have been made necessary by the special authority exercised over a border region of Aquitaine by Rannoux II during [119] this period.⁽⁴¹⁾ Viscounts in the Midi and Catalonia, then, were not created by usurpation, but were initially a response to a governmental need of the period.

While there seems to be little doubt that the initiative in appointing viscounts came from the counts, the dukes, and the marquises, it also seems clear that Carolingian rulers, down to the time of Eudes, were not content to allow them to be responsible only to their immediate superiors. They, therefore, seem often to have established dose ties with such viscounts. That is the meaning of the fact that Isembert is called a fidelis by Charles the Bald in 859 five years before he appears as a viscount at Narbonne.⁽⁴²⁾ It explains why the same monarch in granting land to Hildebert, viscount of Limoges in 876, refers to him also as his fidelis.⁽⁴³⁾ Five years later in 881. Carloman refers in like terms to a certain Rainald, who we learn is serving as viscount of Béziers a little later.⁽⁴⁴⁾ While Carolingian rulers were willing to let their counts make use of such subordinates, they seem to have wished to tie them directly to the monarchy also by ties of vassalage or fidelitas, at least down to 887. After this date, such a relationship seems largely to have disappeared.

This leads to the final and difficult question as to whether or not the charge of viscount had, like that of count, become an hereditary one by the end of this period. It certainly rapidly became so in the next century, as viscounts followed the example set for them by their superiors, the counts. Down to 900, however, we can be sure that viscounts passed their position or honor on to their heirs in two cases only. One is in Béziers where Viscount Boson succeeded his father-in-law in 897.⁽⁴⁵⁾ The other is in Angoulême where Viscount Ramnulf was succeeded by his sons soon after the death of Count Vulgrin in 886.⁽⁴⁶⁾

When we turn from consideration of viscounts to the less important subordinates of the counts, the vicars, we find ourselves facing greater difficulties. Our documents make it clear that in some portions of the Midi like Auvergne, Velay, the Limousin, and neighboring Poitou -- that is to [120] say in Aquitaine -- counties were divided into districts known as vicaria.⁽⁴⁷⁾ South of these areas of Southern France, however, we do not at first find such vicaria mentioned, and we find that almost everywhere south of Poitou and Burgundy evidence of actual vicars exercising their charges is rare. In 845, we find one presiding over a court at Marseille.⁽⁴⁸⁾ Two are mentioned in a charter from Angoulême dating from 855⁽⁴⁹⁾ At about the same time they may have been used under the title of centenarii by the counts of Toulouse in Pallars and Ribagorça.⁽⁵⁰⁾ After 870 references to them are slightly more numerous in our sources. We find them at Vienne in this year attending a court presided over by Count Gerald.⁽⁵¹⁾ They are found at Nîmes in 876,⁽⁵²⁾ and one is mentioned in the civitas of Limoges in 884.⁽⁵³⁾ About 881 a Vicar Fredanci seems to have served as a castellan of a frontier fortress in Ausonia as a fidelis of Count Guifred.⁽⁵⁴⁾ This is all. Useful minor officials they seem to have remained, but not too widely distributed until well into the next century.

Even rarer than the vicars are the vassi dominici, judging by our sources. We find them mentioned only twice after 828, once in 852 attending a court in Narbonne presided over by Count Oldaric⁽⁵⁵⁾ and again in 870 in the Toulousain at a court held by Count Bernard of Toulouse.⁽⁵⁶⁾ Then they seem to have disappeared. Did they actually cease to exist, or did they become vassals of the counts instead of vassals of the Carolingian rulers and continue to attend the courts of the counts in this capacity after 887? It is difficult to say. All we can be sure of is that there is only one indication of such a survival of vassi dominici in this new role. In a court held [121] at Nîmes in 898 presided over by Count Raymond I of Rouergue we do find two men, called vassi, in attendance.⁽⁵⁷⁾ Elsewhere, all is silence for the Midi and Catalonia.

To sum up, then, the governmental system of the Midi did change during the years from 828 to 900 from what it had been under Charlemagne and Louis the Pious. Not only did noble families establish themselves as de facto hereditary rulers of important regions, but in the process the use of viscounts became more widespread, along with some use of vicars in a minor but important subordinate capacity. As this happened vassi dominici disappeared though some of their functions may have been taken over or transformed into those of vassi beholden to local counts instead of to Carolingian and Capetian monarchs who lived north of the Loire.

A discussion of officials and governmental administration found in Midi and Catalonia after 828, however, and of the changes which took place in their character, is still incomplete without an examination of two other important governmental phenomena. The first is the institution of a system of co-comital authority in certain regions south of Poitou and Burgundy. The second is the new power exercised in government by women. Both are a feature of the next century. Both already seem to have appeared by the year 900.

It is difficult to see how either of these developments owes anything to previous, prevailing Carolingian governmental practices. Carolingian monarchs always gave an honor or a group of honores to a single official whom they bound to them by the oath and ceremony of homage and fidelitas. They do not appear to have allowed women to exercise local governmental functions, since such charges always induded the responsibility of military service at the pleasure of the monarch. Yet soon after 870 we find both developments in the Midi. These two phenomena seem to have appeared at the moment when public office in our regions was transformed into hereditary familial right. As this happened, counts, who held their authority still in theory from late Carolingian monarchs, began privately to associate with themselves brothers or sons who bear the title of counts and who take a share in governing their domains. It is in Catalonia and the Carcassonne region that we first find this system in effect. Here certain counts like Oliba II [122] of Carcassonne, Sunyer II of Ampurias-Roussillon, and Guifred of Barcelona-Gerona-Cerdanya-Urgell legally owed their office to Carolingian monarchs. Yet we find, associated with them as counts after 870, other members of their families who seem to share authority with them. Thus, along with Oliba II we find his brother Acfred I and then, at the turn of the century his sons, Bencio and Acfred II.⁽⁵⁸⁾ In Ampurias-Roussillon Count Dela, a brother of Sunyer II, is mentioned in our documents.⁽⁵⁹⁾ In the honores he controls Count Guifred seems to have associated with him his brothers Count Miró the Old and Count Radulf.⁽⁶⁰⁾ Nearby in Pallars and Ribagorça after the death of Count Ramón, late in the century, two of his sons, Isarn and Llop, serve as co-counts of Pallars and two others, Bernard and Miró, as co-counts of Ribagorça.⁽⁶¹⁾

This system cannot be considered a purely Spanish one confined to regions under the influence of the Christian kingdoms of the Iberian peninsula, since it is also found in the rest of the Midi. When the counts of Toulouse recovered their authority from the slipping grasp of Bernard Plantevelue and his family, they instituted a similar system in their domains, where Count Eudes associated with himself his son Raymond of Rouergue as count over a portion of his honores and a little later another son, Ermengaud, as well. In Western Aquitaine after Count Vulgrin died in 886, his sons, William and Aldouin, ruled his counties of Angoulême and Périgord in common and transmitted this tradition to their heirs.⁽⁶²⁾ By 898, in the Limousin we find the same practice with two viscounts, Ademar and Gausfred ruling a part of this region in common.⁽⁶³⁾ By 900, then, in the Limousin, in Angoulême, in Languedoc, and in Catalonia and the Spanish March a new governing system had made its appearance.

One reason for this development may well have been a purely practical [123] one -- the need of governing large areas or multiple counties with inadequate governmental machinery or personnel. One possible answer to this problem was the use of viscounts, as we have mentioned. But an even better method was to designate a brother or a son, or both, as co-rulers and so assure one's family of direct control of regions which might otherwise slip away. In Catalonia and Carcassonne-Razès such a co-ruler's rights were often limited to his own lifetime and do not seem to have been inherited by his issue.⁽⁶⁴⁾ In the rest of the Midi this does not appear to have been so.

Another consideration, however, seems more important -- a subtle but real change in the nature of power exercised by counts after 870. As they made their authority private and hereditary, they came to think of it as a private family possession and so subject to traditions of the equal division of family property among all heirs -- as Roman and Visigothic law provided. Naturally, therefore, all heirs shared in a county or counties as they shared in other family property. Such a change emphatically did not make authority feudal in the commonly accepted sense of the word, but it did make it familial and caused it to be shared among a count's family and heirs in a new way.

It was this change which also seems to explain the new and different role played by women in the governing system of the Midi and of Catalonia. Earlier in the century women in these regions were almost always associated with their husbands in private charters and even possessed considerable property in their own name. In a public capacity, however, they appear to have been unimportant and subordinate to their husbands or sons who, as counts and viscounts, exercised the real authority. As public office became a private family matter, however, women's position changed. Since in the ninth century, like the twentieth, women tended to outlive their husbands, women, as wives and particularly as widows, often became the heads of these families who exercised rights as rulers over the counties and honores of their regions. As such, they began to share authority in a new way with their husbands and their sons. A Countess Ermissende of Carcassonne or a Countess Guinelda, wife of Count Guifred of Barcelona-Cerdanya, now appear in contemporary documents as equal to their spouses and begin to exercise a certain political authority in their own right.⁽⁶⁵⁾ Thus in 883 Countess Berteiz of Toulouse in her [124] own right signed a charter which left to the abbey of Vabres considerable property.⁽⁶⁶⁾ In 899 Countess Guillaumette of Melgueil acted with a similar authority in dividing her lands between the Church and her heirs.⁽⁶⁷⁾ The way was opened toward an era in which women like Countess Garsinde and Viscountess Arsinde and others were to dominate so much of the Midi in the next century. More important, then, than the de facto hereditary independence of noble families or the appearance of viscounts or the disappearance of vassi dominici was this important change from public authority to family control which not only led to the use of co-counts, but to a special and real authority exercised by women -- an authority all but unknown north of the Loire.

These important changes in the way in which honores or counties were held and in the administrative officials who ruled them south of Poitou and Burgundy, were not, curiously enough, reflected in the legal system. Down to 900, at least, the system of law and of courts continued in a form like that established by the earlier Carolingian rulers. Courts, judging from our documents, continued to be held in most regions and were presided over by counts, or their legal representatives, the viscounts, the vicars, and the missi. Everywhere we find in attendance at these tribunals the same judices, scabini, and boni homines whom we have met with earlier, obviously representing the principal local landowners or magnates of the region. In areas south of Carcassonne, in addition, we find present those special saiones or wise men versed in the law which Visigothic procedures required.⁽⁶⁸⁾

Courts of such Carolingian-type are found meeting at Marseille in 845, at Vienne in 870, at Nîmes in 876, 892, and 898 and at Béziers in 897.⁽⁶⁹⁾ Our records concerning them are even more numerous for western Languedoc where we find them held at Narbonne in 855⁽⁷⁰⁾ and at Carcassonne in 862, 873, 875, 883, and 898.⁽⁷¹⁾ In nearby Roussillon we find them [125] meeting in 832, 858, 865, and 868,⁽⁷²⁾ and in Confluent in 874.⁽⁷³⁾ Though our records for Aquitaine concerning them are less extensive, we do find one held in the Toulousain in 870,⁽⁷⁴⁾ another in Rouergue in 878,⁽⁷⁵⁾ a third in Auvergne in 851,⁽⁷⁶⁾ and a series of others in the Limousin in 870, 887, and 898.⁽⁷⁷⁾ One which may vary slightly from the accepted pattern is to be found meeting in Angoulême in 868,⁽⁷⁸⁾ as well.

In addition to such secular tribunals we find evidence of other courts in these regions. These are ecclesiastical ones presided over by bishops and abbots. They were in line with Carolingian practice which, in granting immunities to abbeys and churches, gave them the right to establish courts for those subject to their authority. The earliest example of such a court in our sources is one held in the Limousin over which Bishop Stodile of Limoges presided and in which a dispute between a nearby abbey and a vassus of the bishop was handled.⁽⁷⁹⁾ The next is a tribunal presided over by Archbishop Frédol of Narbonne in 865 to consider a dispute over land between a priest and the abbey of Caunes.⁽⁸⁰⁾ Finally we hear of a third, a court of the abbot of Arles in Vallespir held in 875 and concerned with a similar property dispute.⁽⁸¹⁾ Several points concerning these courts should be of interest to us. In the first place, all three seem to concern themselves with cases involving disputes and individuals subject to their legal authority. They are not, then, courts which conflict with the jurisdiction of secular tribunals. Secondly, they seem remarkably like the secular courts of the period. In each case the presiding officer seems to be assisted by the same type of boni homines or judices we meet in secular courts, and the procedure seems the same except that many of those present are clerics instead of laymen. Except for this, however, one might be dealing with any typical Carolingian court.

[126] In fact there is only one case in the Midi in which we seem to find evidence of a different sort of legal procedure. This is in the record of the settlement of a dispute between the abbot of Aniane and a nearby landowner some time between 829 and 840. This settlement was not apparently the result of the meeting of a court, but rather the result of an informal gathering which took place between the parties concerned in the castle of the abbey. The resulting agreement is called a conventia and seems to be the first example of this sort found in the Midi⁽⁸²⁾ -- a precedent perhaps for the common use of this type of settlement of disputes in later centuries.

Not only did courts remain much what they had been in the earlier Carolingian period, but so did the law which was used in them, though there is some evidence that the legal concepts of this law were not always clearly understood by the parties concerned -- particularly in the case of Roman and Visigothic law. Nevertheless, personality of the law continued in use. South of Narbonne and Carcassonne it was Visigothic law which seems to have been used almost exclusively and indeed our sources show it specifically invoked by name at tribunals held in Narbonne in 855,⁽⁸³⁾ in Carcassonne in 862,⁽⁸⁴⁾ and in Roussillon in 865 and 868.⁽⁸⁵⁾ It is also mentioned in a court in Nîmes held in 898.⁽⁸⁶⁾ Salic law, however, did not die out and continued to be used by certain individuals. Documents dating from 845 mention it at Marseille,⁽⁸⁷⁾ and in Auvergne in 859⁽⁸⁸⁾ and at Nîmes in 898.⁽⁸⁹⁾ Roman law is also referred to at Marseille in 845⁽⁹⁰⁾ and in the Limousin in 841, 851, 865, and 871.⁽⁹¹⁾ It is difficult, however, to be sure that we should attach too much legal significance to such references to various types of law, except in those areas where Visigothic procedures were omnipresent. Elsewhere whatever law is said to be used, the courts seem very similar one to another. Procedures differ little and findings seem still to be rendered in the same way by lay officials and churchmen who were assisted by judices and boni homines in [127] attendance. From the standpoint of law the world of the Carolingians still lingered on relatively unchanged during this period south of Poitou and Burgundy.

While we can come to some definite conclusions as to changes which took place in the administrative and judicial side of government in our regions, the military aspects of it are more difficult to analyze with any precision. Yet it is dear that in these turbulent years in the Midi and Spanish March the military side of its governmental system must have been its most important responsibility. Moslem invasions of Provence, Catalonia, and other parts of the Spanish March, and Viking raiding activities in Gascony, Aquitaine, and even, for a brief period, in the Mediterranean, account for this fact. So do the civil wars which took place between Charles the Bald and Pepin II from 840 to 865 and the struggle over Provence and the Valley of the Rhone which lasted down to 887 or 890. Nor was this all. There is ample evidence that these turbulent decades abounded in private quarrels between magnates which tended to be settled by forceful means.

Ademar of Chabannes, for instance, comments on the fact that the nobles of Aquitaine, or the duces as he calls them, were so busy fighting each other during the period 845-852 that they had no time to deal with Viking invaders.⁽⁹²⁾ Later on, in 866, we hear of a battle to the death between Count Ermenon of Périgord and Count Landry of Saintes in a castle in Angoulême.⁽⁹³⁾ The men of Bernard Plantevelue in 872 kill Bernard of Toulouse.⁽⁹⁴⁾ In fact, the election of Boson as king of Provence in 879 seems to have been due to the desire of churchmen of the Rhone Valley to restrain disorders.⁽⁹⁵⁾

Even where neither civil war nor invasions had any direct impact we find evidence in our documents of continual struggle. In 859 and 860 we find Archbishop Rodulf of Bourges complaining of the presence of evil men in the Limousin -- "infestorum malorum hominum."⁽⁹⁶⁾ A little later in the Rouergue the same phrase is used about 894, when Viking invasions had ceased to be a problem.⁽⁹⁷⁾ In charters from the Auvergne region [128] which give land to the abbey of Brioude during this period we also find that a surprising amount of this property was acquired by conquest.⁽⁹⁸⁾ Charters mentioning this fact are to be found dating from the years 858, 859, and 860, then again in 870 and 873 and finally in an unbroken sequence in 881, 882, 883, 888, 890, 891, 892, 893, 895, 898, 899, and 900.⁽⁹⁹⁾ The same thing is revealed for the Lyonnais in 857,⁽¹⁰⁰⁾ for the Vienne region in 870 and 875,⁽¹⁰¹⁾ and for the Valence area in 889.⁽¹⁰²⁾

Unfortunately when we try to discover specific evidence of the military system in use during this period we find little really pertinent information. The Manual of Dhuoda with its general references to fidelitas helps little.⁽¹⁰³⁾ Nor do the generalities of Carolingian capitularies. It seems probable that Carolingian monarchs like Pepin II, Charles the Bald, and Carloman, regularly summoned into battle their fideles, the counts, the vassi dominici and the important aprisio holders, and that counts did the same for their vassi. We hear of the men of Bernard Plantevelue who killed Bernard of Toulouse in 872.⁽¹⁰⁴⁾ We learn of Bernard of Septimania invading Catalonia and being killed by the Goti of Barcelona in 850.⁽¹⁰⁵⁾ We find a reference to the fideles who followed Count Guifred and Count Miró between 878 and 898 and helped repopulate the plain of Vich and [129] garrison its castles.⁽¹⁰⁶⁾ But we can tell little of how they were mustered into service and how they fought.

This is particularly true of those who battled invaders. We learn that Count Séguin of Bordeaux was captured and killed while fighting the Vikings in 845,⁽¹⁰⁷⁾ that Count Rannoux of Poitou fled from them in 852,⁽¹⁰⁸⁾ that Count Turpio in 863 was slain while opposing them,⁽¹⁰⁹⁾ and that Toulouse was able to beat them off a little later.⁽¹¹⁰⁾ That is all. Just so we learn of the failure of the archbishop of Arles in a campaign against Moslem pirates in the 850's⁽¹¹¹⁾ and of the success of Count Gerald of Vienne in 869⁽¹¹²⁾ without having any details which might help us in getting a true picture of the military organization of the areas involved.

What such information, slight though it may be, seems to emphasize is the fact that the military organization which existed during this period was rather ineffective. The failure to destroy Moslems established in Provence or to expel Vikings from Aquitaine, like the difficulties experienced by Carolingian rulers in overcoming the resistance of Count Gerald and King Boson in Vienne, seems to reflect a failure in military organization. So does Count Guifred's defeat and death in battle against the Moslem lords of Lérida in 897 or 898.⁽¹¹³⁾ In contrast to the disciplined warriors whom Charles Martel and Pepin led south between 718 and 768 or those whom William, count of Toulouse, led against Barcelona in the campaign of 801-803,⁽¹¹⁴⁾ the fighting men of the Midi seem disorganized and poorly led during this period. Except for some early successes in Ausona and Berga by Counts Miró and Guifred,⁽¹¹⁵⁾ as a matter of fact, we do not know of a single military operation of any consequence involving warriors of Southern France and Catalonia during these years which was a success.

[130] There is still one more way, however, in which we may gain some knowledge of the military aspects of the government of this period. That is by an analysis of information concerning castles and other fortifications. By examining the occupation of old castles and the building of new ones and by studying their distribution we can get some indication of military organization. We can also learn something by examining what we can discover of castle ownership and the power of castellans over the surrounding countryside.

Our first problem concerns evidence of the existence of castles and other fortifications in various parts of the Midi and Catalonia between 828 and 900. By the end of the century it is clear that there were a number of such fortified centers in the Rhone Valley and in Provence. We find mention of a castle at Tournon in 855⁽¹¹⁶⁾ and two castella at Vienne between 855 and 860.⁽¹¹⁷⁾ In 887 we hear of the fortress of Châtel in Maurienne which guarded the clusae leading to Italy.⁽¹¹⁸⁾ Seven years later there is a reference to a Castle of Inchas near Vienne.⁽¹¹⁹⁾ In addition to such castles, Lyon is mentioned as having walls in 892 and Vienne in 849 and 899.⁽¹²⁰⁾ In Provence we learn of a castrum located at Venasque in 862 and one built in the Camargue in 869.⁽¹²¹⁾ Others are referred to at Tournon in 879 and Incia at about the same time.⁽¹²²⁾ After Moslem attacks in 859 Arles seems to have been fortified, while in 898 Avignon possessed walls.⁽¹²³⁾

To the east in Septimania the castle of the Arena of Nîmes was again being used in 876.⁽¹²⁴⁾ So was the nearby castle of Anduze mentioned between 876 and 897.⁽¹²⁵⁾ Aniane still possessed its castle in Lodève about 840⁽¹²⁶⁾ and we have every reason to believe that the nearby castles of Melgueil [131] and Substantion were still occupied.⁽¹²⁷⁾ Near Béziers we hear of the castra of Mesoa and Turres,⁽¹²⁸⁾ while Narbonne possessed city walls half of whose turres Charles the Bald gave to the archbishop of the city in 844.⁽¹²⁹⁾ Near Carcassonne we find a castle of Minerbe in 873, one at Alzonne in 888, and evidence of fortifications at the abbey of Montolieu by 898 if not before, since it is now called the Castrum Mallesti.⁽¹³⁰⁾ In Razès in 900 there was a castrum of Por near the abbey of Saint-Jacques de Joucan.⁽¹³¹⁾

To the south in Roussillon our documents speak of a castle of Corbi in 832, a rocca at Fruesendi in 854 and several other roccas, which Marquis Humfrid had "in beneficio" and which were given to Count Sunyer in 862.⁽¹³²⁾ Perhaps one of these was the castle of Saint-Stephen of the rocca of Ponus whose castellan is referred to in a charter dating from 865.⁽¹³³⁾ In Catalonia proper there is a mention of the castle of Vellosa in Gerona in 834 and 844 with the nearby castellum fractum and in the same pagus one at Monte Aspero which dates from at least 889.⁽¹³⁴⁾ Barcelona was protected by walls throughout the period and so was the important fortification of Tarrassa referred to in documents dating from 844 and 874.⁽¹³⁵⁾ Ampurias had at least one castle, that given to a certain Stephen by Charles the Simple in 899.⁽¹³⁶⁾

Further to the west the activity of Count Guifred and his brothers in advancing their frontiers to include Ausona and the marches of Berga resulted in the building of a large number of castles. One in Besalu called Castellaris is mentioned in 871, one at Montgrony in 887 and a third at Tarabaldi in 892⁽¹³⁷⁾ -- all in Ausona. Others at Manressa, Cardona, Besaura, Kastellionem, Casserès, and many other places date from this [132] period also.⁽¹³⁸⁾ So do the many small turres or castles which we find scattered throughout Pallars and Ribagorça.⁽¹³⁹⁾

To the north in Aquitaine we find Toulouse possessing walls in 844⁽¹⁴⁰⁾ and capable still of resisting Norse attacks two decades later, while on the Gascon borders there is a record of at least one castle, that of Cerrucium which before 847 was given to Astronove by Pepin II.⁽¹⁴¹⁾ Other regions of Aquitaine contained a number of castella which the Vikings sacked between 844 and 848.⁽¹⁴²⁾ Among them perhaps was the castle of Rancogne in Angoulême mentioned in 866.⁽¹⁴³⁾ We also know that some time between 869 and 877 Count Vulgrin raised two castles in the same region and in the Saintonge to thwart Viking raiders, the fortresses of Marillac and Matas.⁽¹⁴⁴⁾ A little later, soon after 886, Count Aldouin rebuilt the walls of Angoulême.⁽¹⁴⁵⁾

To the east in Quercy we find a reference to a castle of Cosilacum near Cahors which seems to have been in existence in 844 and one in Castellione overlooking the Dordogne.⁽¹⁴⁶⁾ In 859 the castle of Turenne in the Limousin was occupied.⁽¹⁴⁷⁾ So was one at Castellacus mentioned in 885 and one at Ségur in 888.⁽¹⁴⁸⁾ In the same year in nearby Poitou we find reference to a tower at Château Larchier.⁽¹⁴⁹⁾ In our documents concerning Auvergne we find a reference to a castrum of Mozérat in 864⁽¹⁵⁰⁾ and ten years later in 874 to the fortifications of the abbey of Brioude and to a castrum at Aurillac as well.⁽¹⁵¹⁾ As early as 825 Le Puy had a castrum built by Count Berengar.⁽¹⁵²⁾

[133] This incomplete list of castles and fortified civitates dating from this period -- incomplete because of a lack of documentary materials -- does seem to show certain things. It makes clear that there was a certain concentration of castles in three regions, in Provence and Eastern Languedoc near the mouth of the Rhone, in Western Aquitaine, and in Catalonia. It suggests that this concentration was the result of reactions to invasion, or in the case of Catalonia a response to armed advance into frontier areas. A Count Vulgrin in Angoulême or local magnates in Provence and the Nîmes area, or a Count Guifred in Catalonia seem to have been responsible for the fortifications of such regions, often in response to outside pressure. For other parts of the Midi, like the Carcassonne region, the Middle Rhone Valley, or Auvergne and the Limousin, no such explanation is possible. Here the occupation of old castles and the building of new ones must be regarded rather as a response to the insecurity which prevailed during these years. This seems particularly true of fortifications which protected abbeys like Aniane, Brioude, and Montolieu.

Furthermore, some indication of defense matters can also be gained from an analysis of castle holding. In most cases we are ignorant as to who actually controlled a particular castle or fortified city. In other cases we can be more precise. A large number of them, perhaps even a majority of them, were under the direct control of the authorities, such as the counts and other representatives of secular government. The castles of Inchas and Châtel⁽¹⁵³⁾ in the Upper Rhone Valley region, those of Turenne and Ségur⁽¹⁵⁴⁾ in the Limousin, the castles of the Arenas and Anduze⁽¹⁵⁵⁾ in the Nimois, and those built by Vulgrin in Angoulême⁽¹⁵⁶⁾ and Count Guifred in Ausona⁽¹⁵⁷⁾ are cases in point. Others built on land which belonged to the royal fisc seem to have been controlled, at least initially, by castellans [134] tied to Carolingian rulers by ties of vassalage or fidelitas. Such seems to have been the case of Mesoa and Turres near Béziers held "jure beneficiario" from Charles the Bald, Cerrucium on the Gascon border which Charles the Bald gave to Astronove, the rocca of Fruesendi in Roussillon which Sunald and Riculf held as a benefice in 854 and other roccas in Ampurias held "ad beneficium" from Charles the Bald down to 862.⁽¹⁵⁸⁾

A surprising number of fortresses, however, either were originally or during the course of the century became the possessions of archbishops, bishops, and abbeys. The archbishops of Narbonne were given one half of the turres of Narbonne, the bishop of Maurienne the castle of Châtel, the bishop of Gerona the castle of Vellosa.⁽¹⁵⁹⁾ The abbots of Aniane controlled the castrum of Montecalmense about 840; the abbey of Moissac, that of Cerrucium after 847; Lagrasse, Alzonne after 888; Saint-Paul de Fontclara, one-third of Monte Aspero in 889.⁽¹⁶⁰⁾ Brioude's fortifications after 874 seem to have belonged to the abbey, as did Montolieu's in 898.⁽¹⁶¹⁾ Both church and private individuals then seem in this period to share in a military defense system which one would tend to think of as one belonging to the secular authorities alone -- just as had been true earlier -- a proof, perhaps, of the ineffectiveness of the governmental system established in these regions.

What, however, did control or ownership of such a castle imply? Did castellans of the period tend to have authority over territory which surrounded their fortresses and to levy special dues upon it, as they were to do later on? Did mandamenta already exist? In at least two cases there is definite indication that they did. In 881 we hear of the appendici or terms of the castle of Torello⁽¹⁶²⁾ in Ausona and in 887 the fortress of Châtel in the Alps is mentioned along with "villis villaribus que subjectis eodem castro."⁽¹⁶³⁾Did this mean that dues of a special nature were levied on the surrounding countryside? Perhaps so. In 848, on the plea of Count Appolonius of Agde, Charles the Bald forbade the levying by the laity of certain dues called "mansionaticum, portaticum, salinaticum et hospitalicum" -- this in a region close to the fortresses of Mesoa and [135] Turres.⁽¹⁶⁴⁾ In 874 the bishop of Barcelona asked the help of the same ruler against what he called the insolence of the castle of Tàrrassa and two Goti who were usurping lands belonging to his church.⁽¹⁶⁵⁾ Near Carcassonne in 883 a certain Ermenard, in returning land he had taken unjustly from the abbey of Saint-Hilaire retained a receptum over this property belonging to the monastery.⁽¹⁶⁶⁾ Already then, though opposed, it would seem, by important elements in the population, mandamenta were being formed around the castles to be found in these regions.

Examination of castles then, of their occupation, their building, their control, and their distribution -- as well as evidence of exactions which were levied beyond their walls -- seems to show the same changes revealed in other aspects of military organization or in the administration of honores. It shows us a Midi and a Catalonia in transition from a Carolingian governmental system to something quite different in which local nobles, churchmen, and private individuals more and more tend to exercise power which was once the prerogative of the Carolingian government. Much of the old Carolingian system still remained, particularly the courts and the legal system, but much had already begun to be profoundly altered in spirit and in substance towards a new and unusual system of government which we will be examining in later chapters.

1. Chartes poitivines (833-1160) de l'abbaye de Saint-Florent de Samaur, ed. P. Marchegay, no. 33, in Archives historiques du Poitou, II (1873), 46. On the lack of viscounts in Poitou until late in the ninth or early in the tenth century see M. Garaud, "Les circonscriptions administratives du comté de Poitou et les auxiliaires du comté au X^e siècle," in Le Moyen Age, LIX (1953), 11-13.

2. Catalunya Carolingia, II, 442-444. See also the mention of a viscount's court in 791 in Hist. Gén. de Lang., II, no. 10. In both these years Narbonne formed part of a large honor or a group of honores ruled by a single count. In 791, for instance, it was probably ruled by Duke William of Toulouse. In 834 it was one of the honores of Bernard of Septimania. Hence a viscount in this important city.

3. See the charter in F. Valls-Taberner, Els origines dels comtats de Pallars i Ribagorca, p. 5. On the possible spuriousness of this charter see G. Tessier, "A propos de quelques actes toulousains du IX^e siècle," in Recueil de travaux offerts à M. C. Brunel, II, 566-580.

4. Cartulaire de l'église d'Angoulême, ed. I. Nanglard, I, no. 52.

5. Hist. Gén. de Lang., II, no. 150.

6. Cart. de Carcassonne, II, 598, mentions an Isembert as a fidelis of Charles the Bald in 859. A little later he is called a viscount and missus (Cart. de Carcassonne, I, 71-72).

7. Grand Cart. de Brioude, no. CXIV.

8. Catalunya Carolingia, II, 325-326.

9. Cartulaire de l'abbaye de Saint-André-le-bas de Vienne, ed. U. Chevalier, appendix no. 2, pp. 211-213 (hereafter cited as Cart. de Saint-André -le-bas).

10. Cart. de Saint-Victor, no. 26.

11. Recueil des chartes de l'abbaye de Cluny, ed. A. Bernard and A. Bruell, I, no. 15, pp. 18-20.

12. Cart. de Vienne, no. 1, appendix 1, pp. 219-221.

13. Cart. de Nîmes, nos. 1, 5, 8.

14. Hist. Gén. de Lang., V. nos. 4, 17, 18.

15. Cart. de Carcassonne, IV, 70-71.

16. Hist. Gén. de Lang., V, no. 5.

17. Cart. de Carcassonne, I, 73-74, and Odo, Vita Sancti Geraidi Aureliacensis, II, 28, in Migne, Patrologia latina, CXXXIII, cols. 685-686.

18. Hist. Gén. de Lang., II, no. 180.

19. R. de Abadal i de Vinyals, Els Primers Comtes Catalans, pp. 78-79.

20. El Archivo Condal de Barcelona en los Siglos IX-X, ed. F. Udina-Martorell, no. 9.

21. Cart. de Saint-Etienne de Limoges, no. 84.

22. Garaud, "Les circonscriptions administratives du comté de Poitou," pp. 13-15.

23. Cartulaires des abbayes de Tulle et de Roc-Amador, ed. J. Champéval, in Bulletin de la société scientifique, historique et archéologique de la Corrèze, VI (1893), no. 276.

24. A. Jacotin, Preuves de la maison de Polignac, I, 21.

25. Ademar de Chabannes, III, 20, p. 138, and "Aquitaniae Historiae Fragmenturn," in ibid., pp. 197-198.

26. Cart. d'Angoulême, no. 40.

27. According to a charter of 860 Raymond I controlled the abbeys of Solignac and Beaulieu (Cart. de Beaulieu, no. 1). Yet in 854 Gotfred was still count of Turenne (ibid., no. 186).

28. Cart. de Carcassonne, V, 228.

29. Ibid., I, 73-74.

30. Annales Bertiniani, ed. E. Waitz, anno 872.

31. Abadal, Els Comtats de Pallars i Ribagorça, pp. 107-108.

32. Catalunya Carolingia, II, 442-444.

33. Valls-Taberner, Els origines des comtats, p. 5.

34. Cart. de Saint-André-le-bas, appendix no. 2, pp. 211-213.

35. On a viscount at NImes see Cart. de Nîmes, no. 1. On one at Béziers see Hist. Gén. de Lang., V. no. 17. On one at Narbonne see ibid., no. 5. On one at Albi see mention of Aton, viscount of Albi, who served as missus of Count Eudes at Carcassonne in 898 (Cart. de Carcassonne, I, 73-74). See Abadal, Els Primers Comtes Catalans, pp. 76-81, on the organization of Ausona and the establishment of viscounts by Guifred. On one in the Limousin see Cart. de Saint-Etienne de Limoges, no. 84.

36. See mention of Viscount Berald in 875 in Cart. de Carcassonne, IV, 70-71, and in 883 in Hist. Gén. de Lang., V. no. 5.

37. Catalunya Carolingia, II, 442-444.

38. See Abadal, Els Primers Comtes Catalans, pp. 76-81, on the organization of Ausona and the establishment of viscounts by Guifred.

39. See E. Delcambre, "Géographie historique du Velay du pagus, au comté et au baillage," in Bibliothèque de l'Ecole des Chartes, XCVIII (1937), 37-41.

40. Ademar de Chabannes, III, 20, p. 138. The "Aquitaniae Historiae Fragmenturn," in ibid., pp. 197-198, says that Gislebert was given the castle of Marillac at the same time that Ramnuif was given Matas and made viscount. It also calls Ramnuif the "fidelissimus amicus" of the count.

41. Note that when Rannoux II established a viscount at Thouars he had claims to a wider region than Poitou proper and is called "dux maximae partis Aquitaniae" in 889 according to Annales Vedastes, ed. B. de Simson, anno 889, p. 335.

42. Cart. de Carcassonne, II, 598.

43. Cart. de Saint-Etienne de Limoges, no. 84.

44. Hist. Gén. de Lang., V. no. 4.

45. Ibid., no. 18.

46. See Ademar de Chabannes, III, 23, pp. 144-145, on the restoration of Viscount Ramnulf's honores to his heir by Count Bernard, son of Count Vulgrin.

47. On such vicaria in Auvergne see H. Doniol's introductions to the Cart. de Brioude and the Cartulaire de Sauxillanges. For Velay see Delcambre "Géographie historique du Velay." For the Limousin see Deloche's introduction to the Cart. de Beaulieu. For Poitou see M. Garaud, "Les origines des Pagi poitevins du moyen-âge," in Revue historique du droit français et étranger, XXVI (1948); "La construction des châteaux et les déstinées de la vicaria et du vicarius Carolingiens en Poitou, in ibid., XXXI (1953); and "Les circonscriptions administratives du comté de Poitiers et les auxiliaires du comté au X^e siècle," in Le Moyen Age, LIX (1953).

48. Cart. de Saint-Victor, no. 26.

49. Cart. d'Angoulême, no. 52.

50. Valls-Taberner, Els origines des comtats, p. 5.

51. Chartes de Cluny, I, no. 15.

52. Cart. de Nîmes, no. 1.

53. Cart. de Saint-Etienne de Limoges, no. 3.

54. Abadal, Els Primers Comtes Catalans, pp. 88, 98.

55. Hist. Gén. de Lang., II, no. 139.

56. Ibid., no. 174.

57. Cart. de Nîmes, no. 8. On the possibility that these vassi were royal vassi dominici see F. Ganshof, "L'origine des rapports féodo-vassiliques," in I problemi della Civiltà Carolingia, p. 43 n.

58. See the mention of co-counts Oliba II and Acfred of Carcassonne-Razès in Marca hispanica, I, appendix 39, and Hist. Gén. de Lang., II, no. 184. See also Abadal, Els Primers Comtes Catalans, pp. 13-24, 244-246.

59. See Abadal, Els Primers Comtes Catalans, pp. 18-24, 199-201, on the co-counts of Ampurias-Roussillon.

60. See Hist. Gén. de Lang., II, no. 184, and Abadal, Els Primers Comtes Catalans, pp. 13-24, 234-246, 249-250.

61. On documents dating from as early as the late ninth century showing such co-counts see Abadal, Els Comtats de Pallars i Ribagorça.

62. See Chapters I and II, Section IV on the power of the counts of Toulouse and their method of governing their honores. Ademar of Chabannes says that the heirs of Count Vulgrin "communem haberunt totum honorem" (Ademar de Chabannes, III, 23, pp. 144-145).

63. About those viscounts (who were perhaps brothers) who jointly presided over a court see Cart. de Tulle, no. 276.

64. See Abadal, Els Primers Comtes Catalans, pp. 199-201, on the way this system worked in Catalan counties.

65. See ibid., pp. 241-246. For mention of Countess Guinelda after 885 see especially Marca Hispanica, I, appendices 53, 404.

66. Hist. Gén. de Lang., II, nos. 164, 203.

67. Cart. de Maguelonne, I, no. 3.

68. See the excellent article on the survival of Visigothic law into this period by A. Gouron, "Les étapes de la pénétration du droit romain au XIIe siècle dans l'ancienne Septimanie," in Annales du Midi, LXIX (1957), 104-111.

69. Cart. de Saint-Victor, no. 26; Chartes de Cluny, I, no. 15; Cart. de Nîmes, nos. 1, 5, 8; and Hist. Gén. de Lang., V. no. 18.

70. Cart. de Carcassonne, IV, 192-193.

71. lbid., I, 71-72 (862), 73-74 (898); IV, 70-71 (875); Cros-Meyrévielle, Documents, no. 17 (873); and Hist. Gén. de Lang., V, no. 5 (883).

72. Hist. Gén. de Lang., II, no. 80 (832), 150 (858), 169 (868); Cart. roussillonnais, no.1 (865).

73. Hist. Gén. de Lang., II, no. 173.

74. Ibid., no. 174.

75. Ibid., no. 201.

76. Grand Cart. de Brioude, no. CXIV.

77. Cart. de Beaulieu, no. 27 (870), 164 (887); Cart. de Tulle, no. 276 (898).

78. Cart. d'Angoulême, no. 37. This court, which seems to have had no presiding officer, appears more like the informal assemblies found in the next century than a Carolingian tribunal.

79. Cart. de Saint-Etienne de Limoges, no. 3.

80. Hist. Gén. de Lang., II, no. 165.

81. Ibid., no. 175.

82. Cart. d'Aniane, no. 55.

83. Cart. de Carcassonne, IV, 192-193.

84. Ibid., I, 71-72.

85. Cart. roussillonnais, no. 1; Hist. Gén. de Lang., II, no. 169.

86. Cart. de Nîmes, no. 8.

87. Cart. de Saint-Victor, no. 26.

88. Cart. de Brioude, no. 43.

89. Cart. de Nîmes, no. 8.

90. Cart. de Saint-Victor, no. 26.

91. Cart. de Beaulieu, no. 20 (841), 186 (865); Cart. de Saint-Etienne de Limoges, no.3 (851), 8 (871).

92. Ademar de Chabannes, III, 17, p. 134.

93. Ibid., III, 19, pp. 136-137, and Chronicum Aquitanicum, ed. G. Pertz, anno 866, in Mon. Ger. Hist. Scriptores, II, 252-253.

94. Annales Bertiniani, anno 872.

95. Cart. de Grenoble, no. 70, pp. 265-267. No nobles seem to be listed in this charter.

96. Cart. de Beaulieu, nos. 1, 24.

97. Ibid., no. 76.

98. The phrase ex conquestu in charters does not necessarily mean conquest, but land acquired by means other than gift or inheritance. Such land, however, if found in large amounts, as in this region, shows an instability in the pattern of landholding, some of which was certainly due to actual loss of land by the weak and seizure of it by the powerful. On the meaning of conquestu see Cart. de Brioude, no. 126, which mentions land "per parentes vel per conquestem."

99. Cart, de Brioude, nos. 13, 19, 43, 56, 60, 85, 102, 119, 126, 208, 225, 227, 271, 277, 282, 311, and Grand Cart. de Brioude, nos. CCCVIII, CCCLXXII. Fortunately we have some corroborative evidence of disorders in the very region where the abbey of Brioude was located. In a rare narrative, written at about this time concerning Gerald, the self-appointed count by Aurillac, St. Odo of Cluny mentions brigands in the forests, the disturbed state of the countryside, and private wars being waged. Odo, Vita Sancti Geraldi Aureliacensis, in Migne Patrologia latina, CXXXIII.

100. Cart. de Savigny, no. 19.

101. Cart. de Vienne, no. 106, 109.

102. Cartulaire de Saint-Barnard de Romans, ed. U. Chevalier, no. 4.

103. This document, dating from the early part of the ninth century, may still be true of its later years. See E. Bondurand, L'education Carolingienne: Le Manuel de Dhuoda, XV, 89-91.

104. Annales Bertiniani, anno 872.

105. Ibid., anno 848, and Chronici Fontanellensis fragmentum, ed. G. Pertz, anno 849, in Mon. Ger. Hist. Scriptores, II, 43. See also Abadal, Els Primers Comtes Catalans, pp. 181-182.

106. El Archivo Condal de Barcelona, no. 38.

107. Ademar de Chabannes, III, 17, p. 134.

108. Ibid., III, 18, pp. 136-137.

109. Ibid., III, 19, p. 138.

110. See J. Calmette, "Le siege de Toulouse par les Normands," in Annales du Midi, XXIX (1917).

111. See L. Levillain, "Giraut, Comte de Vienne," in Le Moyen Age, LV (1949).

112. E. Duprat, in Bouches du Rhône, II, 136.

113. P. Ponsch, "A propos de la date du décès de Wilfred le Velu," in Études Roussillonnaises, 11(1952), 132-134.

114. See Chapter II, Section I, and Chapter I, Section II, on Carolingian successes in these campaigns.

115. Abadal, Els Primers Comtes Catalans, pp. 191-198. Particularly melancholy is the record of the failure to stop Moslem invasions of Catalonia in 852 and 856.

116. R. Poupardin, Recueil des actes des rois de Provence (855-928), p. 20.

117. Ibid., p. 11.

118. Chartes de Maurienne, no. 1.

119. Cart. de Grenoble, no. 27.

120. On the walls of Lyon see ibid., no. 81. On the walls of Vienne and the castle of Castellione beyond the walls mentioned in 849 see Cart. de Vienne, no. 4. On the walls of the city mentioned in 899 see Cart. de Grenoble, no. 20.

121. Poupardin, Recueil des actes des rois de Provence, p. 27; Duprat, in Bouches du Rhône, p. 275.

122. Poupardin, Recueil des actes des rois de Provence, pp. 38, 57.

123. Duprat, in Bouches du Rhône, pp. 132-134; Poupardin, Recueil des actes des rois de Provence, p. 38.

124. Cart. de Nîmes, no. 1.

125. Ibid., no. 2.

126. Cart. d'Aniane, no. 55.

127. At least the castle of Substantion is mentioned in 1005 ("Cartulaire de Psalmodi" [unpublished], fol., p. 16).

128. Hist. Gén. de Lang., II, no. 105.

129. Ibid., no. 115.

130. Cros-Meyrévielle, Documents, no. 17; Cart. de Carcassonne, II, 217; 1, 73-74.

131. Cart. roussillonnais, no. 3.

132. Hist. Gén. de Lang., II, no. 80; Catalunya Carolingia, II, 348; Cart. roussillonnais, no. 1.

133. Catalunya Carolingia, II, 121-124, 127-130.

134. Ibid., pp. 13-15, 423-425.

135. Ibid., pp. 423-425, 431-433.

136. Hist. Gén. de Lang., V. no. 25.

137. Catalunya Carolingia, II, 176-178; El Archivo Condal de Barcelona, nos. 4, 8.See also the castle of Torello mentioned first in 881 in Abadal, Els Primers Comtes Catalans, p. 88.

138. El Archivo Condal de Barcelona, no. 9, and Abadal, Els Primers Comtes Catalans, pp. 87-90.

139. On the large number of turres scattered through these regions see Abadal, Els Comtats de Pallars i Ribagorça.

140. Cart. de Saint-Sernin, no. 3.

141. Hist. Gén. de Lang., II, no. 131.

142. Ademar de Chabannes, III, 17, p. 134.

143. Ibid., III, 19, p. 136.

144. Ibid., III, 20, and "Aquitoniae Historiae Fragmentum, in ibid., pp. 197-198.

145. Ademar de Chabannes, III, 23, pp. 143-144.

146. Cart. de Beaulieu, no. 34; Hist. Gén. de Lang., II, no. 121.

147. Cart. de Beaulieu, no. 33.

148. Ibid., nos. 55, 166; G. Tenant de la Tour, L'homme et la terre de Charlemagne à Saint Louis, pp. 300-301.

149. Cartulaire de l'abbaye de Saint-Cyprien de Poitiers, ed. Redet, in Archives historiques du Poitou, III (1874), no. 400, p. 271.

150. Cart. de Brioude, no. 176.

151. Ibid., no. 132.

152. Hist. Gén. de Lang., II, no. 67.

153. Cart. de Grenoble, no. 27 (originally the castle of Inchas belonged to King Louis of Provence-Burgundy); Chartes de Maurienne, no. 1 (originally Châtel belonged to King Boson).

154. During this period the castle of Turenne probably belonged to the counts of Turenne (See Cart. de Beaulieu, no. 33). On Ségur, the central castle of the viscounts of Limoges, see Tenant de la Tour, L'homme et la terre, pp. 300-305.

155. In 876 the castle of the Arenas seems to be in the possession of the viscounts of Nîmes (Cart. de Nîmes, no. 1). By the late ninth century the fortress of Anduze belonged to Viscount Frédélon of Anduze (ibid., no. 2).

156. See Ademar de Chabannes, III, 20, p. 138, and "Aquitaniae Historiae Fragmentum," in ibid., pp. 197-198, on the circumstances of the building of these castles by Count Vulgrin.

157. On the castles in Ausona and elsewhere in Catalonia built by Count Guifred see Abadal, Els Primers Comtes Catalans, pp. 80-110.

158. Hist. Gén. de Lang., II, no. 105, 131; Catalunya Carolingia, II, 348, 357- 358.

159. Hist. Gén. de Lang., II, no. 115; Chartes de Maurienne, no. 1; Catalunya Carolingia, II, 12 1-124.

160. Cart. d'Aniane, no. 55; Hist. Gén. de Lang., II, no. 131; Cart. de Carcassonne, II, 217; Catalunya Carolingia, II, 13-15.

161. Cart. de Brioude, no. 132; Cart. de Carcassonne, I, 73-74.

162. See Abadal, Els Primers Comtes Catalans, pp. 78-79, 88.

163. Chartes de Maurienne, no. 1.

164. Hist. Gén. de Lang., II, no. 133.

165. Catalunya Carolingia, II, 431-433.

166. Cros-Meyrévielle, Documents, no. 19, pp. 24-25.