1. The Earliest Collections

THE worship of the Virgin Mary, originating in the East, slowly penetrated into western Europe during the interval from the fifth to the eleventh century.1 The festivals of the Virgin began to be celebrated; pictures were painted and images were carved; churches were dedicated to her; and prose and poetry, profane2 as well as sacred, honored her. The absence of details about the life of the mother of God in the gospels was an advantage, rather than a disadvantage, to the growth of the Mary cult. Accounts of her parents, her childhood, her marriage, her purification, and the miraculous deeds she performed both before and after her assumption multiplied.3

At first the miracles of the Virgin were told singly by churchmen to point a moral or to celebrate a festival, related to pilgrims visiting particular shrines, or incorporated into the records of monasteries. As early as Gregory of Tours (c. 538-594) eastern tales of this nature begin to be narrated by ecclesiastical writers in the West. In his Liber miraculorum, the first book of which is entitled In gloria martyrum, he tells, among a number of others, six Mary legends of eastern origin.4 It is, however, not much earlier than the eleventh century that tales of western origin make their appearance in written form in significant numbers.5

The earliest known collections of Mary legends originating in the West are those which gathered about the churches of France dedicated to the Virgin Mary. They date in their present form from the beginning of the twelfth century, although they were probably begun much earlier. Such are those compiled for the Mary churches of their respective towns by
 [[ Print Edition Page No. 4 ]] 
John of Coutances6 at the very beginning of the twelfth century, by Guibert de Nogent (d. 1124) and Herman of Laon7 (compiled c. 1150), by Hugo Farsitus of Soissons,8 and by the abbot Haimon writing for the Mary church of St. Pierre-sur-Dive in Normandy.9 Rocamador in Guienne,10 Chartres,11 and Fécamp12 also had their collections made by anonymous compilers in the twelfth century.

By the eleventh century the popularity of the Virgin Mary had outstripped that of all the saints and was growing in ever-widening circles. Sinners as well as the righteous could rely upon her to get them into heaven, if only they had that mediaeval virtue of loyalty. She had become the Mother of Mercy13 whose privilege it was to mitigate the Justice of God. Saturday was in many places a day reserved for the recitation of the office of the Virgin.14 The network of Cluniac monasteries and the pilgrimage routes offered new and broader avenues along which Mary legends could make their way. When Odo caused Mary to be honored as the Mother of Mercy at Cluny,15 his example was followed in many monasteries. As a consequence not only did the number of Mary legends increase rapidly, but the same legends came to be known far from their original home. And thus there began the second stage in the history of Mary legends in the West, the collection of tales gathered from far and near, of universal rather than merely local interest.

The oldest of these collections, according to Mussafia,16 is the group of legends which make up the first seventeen numbers of the compilation printed by Bernhard Pez in the eighteenth century.17 It begins with the tale of ‘Hildefonsus’ and ends with that of ‘Murieldis,’ hence it is known as
 [[ Print Edition Page No. 5 ]] 
HM.18 The origin of the collection is obscure, and the authors are, for the most part, anonymous. Six of the seventeen legends have lost the names and places with which they were originally associated; the others come variously from Italy, France, Spain, and Germany. The subjects are gathered from the lives of saints and from monastic records of various sorts. Some of them were not in their first form Mary legends at all. The saint for whom they were originally written is still there, but he has been subordinated to the Virgin Mary, who now usurps the leading role.19 This ancient series of tales exists in more or less complete form in almost all the great collections, although the order in which the legends are told is frequently altered. John of Garland’s Stella maris includes eleven of the seventeen.

Next in order of their development, and also of the eleventh century, are the legends of the Elements-series.20 The four anecdotes which comprise it have been rewritten by their unknown compiler so that each illustrates the Virgin’s dominion over fire, earth, air, or water.21 Although it is English in origin and has been found as a unit only in collections which can be identified with England,22 the single legends appear very commonly in the great collections of northern France. John of Garland tells two of them, omitting ‘Childbirth in the Sea’ and ‘Jew of Bourges.’23 They are not told in series, nor with their original significance.

The third series in point of time, belonging to the twelfth century,24 is the collection which forms the third book of two manuscripts, MSS British Museum Cotton Cleopatra C x and Toulouse 482. Related to it also is MS Oxford Balliol 240. There are seventeen numbers, beginning with the legend of ‘Toledo’ and ending with ‘Saturday,’ hence the collection is designated TS.25 Mussafia believes that this series also is English in origin,26 although the tales are gathered from many different sources. All three of these early collections appear together in MS Cotton Cleopatra C x27 of
 [[ Print Edition Page No. 6 ]] 
the twelfth century. The four legends of the Elements-series make up the first book;28 the HM series, the second book; and the TS group, the third book.29

Of primary importance in the history of Mary legends is the so-called Pez collection.30 Probably French in origin and a product of the twelfth century, it was transported to Germany to become the foundation upon which Latin collections in that region were built.31 The first seventeen of the forty-three legends is a complete HM series, told in the same sequence in which they appear in the Cleopatra-Toulouse-Oxford manuscripts.32 Two of the four Elements-series are narrated in the same versions as in Cleopatra-Toulouse-Oxford.33 Nine are TS legends, and one more, ‘Conception,’ is a different version of a TS tale. To these, fourteen new legends have been added from the constantly increasing number of incidents recorded in other sources. That TS was one of the principal sources of Pez is indicated by the fact that the nine miracles of Pez which are identical with those of TS follow in Pez in exactly the same order as in TS.34 The Pez
 [[ Print Edition Page No. 7 ]] 
collection is, therefore, derived from three sources: HM, the Elements-series, and TS, to which other legends have been added. These four collections are the foundations upon which were built, not only the French collections of the twelfth and thirteenth centuries, but those in all western Europe.

2. The Collections of the Parisian Monasteries

At the close of the twelfth century the Mary cult was in full bloom, and the century which followed has been properly called‘the century of Mary.’35 The Ave Maria was as frequently and universally on the lips of all classes as any other prayer. The Salve Regina, the favorite anthem of the pilgrims and crusaders, was sung regularly in the evening before the image of the Virgin, not only in the churches of the religious orders, but by laymen as well.36 The four great Mary festivals already covered the entire year, one for each of the seasons, when the Conception of the Virgin and the Visitation were added in the twelfth century.37 Relics of the Virgin, her shift, her girdle, her sweat, and her milk accumulated. Images of the mother with the Child in her lap were countless, and the great Gothic cathedrals were rising in her honor.

Much of the ever-growing popularity of the Mary cult in the twelfth century is to be attributed to the new monastic orders, the Cistercians, Praemonstratensians, and the various groups of canons regular. In the thirteenth century the mendicant orders joined them in propagating the worship of the Virgin. Robert of Molesme, the founder of the Cistercian order, it is said, had been particularly devoted to her by reason of a vision seen by his mother before he was born.38 Alberic, his successor, had decided that the order should be dedicated to her, and that its houses should be under her invocation. According to legend the Virgin bestowed upon him a white mantle, and it was for that reason that the Cistercians gave up their black habit for a white one. All their churches were dedicated to the Virgin, and each had its Mary altar, before which the office of the Virgin was celebrated every Saturday.39 The seal of the Cistercian monasteries
 [[ Print Edition Page No. 8 ]] 
pictured the Virgin crowned. It is understandable, then, that in the late twelfth or early thirteenth century the collection of Mary legends became a particular project of the Cistercian order.40 The Dominicans and the Franciscans also were greatly devoted to the Mother of God. Tradition tells how the Virgin herself urged St. Dominic to found his order, and the preaching of the mendicants spread Mary legends to the far corners of Europe.41

The larger place given to Mary in the calendar of the church and in the hearts of the people called for greater efforts on the part of the compilers of Mary legends, for it was particularly on the Virgin’s days that the story of her life and her marvellous deeds was read in monasteries and churches. They were not content merely to tell and retell the anecdotes of the HM and TS collections, and to search the old sources, monastic records and saints’ lives, for occasional fresh tales. Entirely new and more varied sources were sought. Even history was forced to make its contribution. The Virgin became the heroine of incidents in the chronicles in which previously she had had no place; and then succeeding chroniclers retold the incident with the legendary accretion as sober history!42 The chance narratives of the ecclesiastical writers of the twelfth century were gathered into the growing compendia, and single miracles were culled from local collections.43 The quickening interest of the twelfth century in the matter of ancient Rome prompted the transformation of some highly pagan tales into Mary legends.44 Even folk tradition may have played a considerable part in the shaping of Mary legends, though the evidence is hard to lay hold of.45

Some new and entertaining anecdotes were selected from the growing collections of sermon stories, especially in the thirteenth century. Mary legends had always been used in sermons, and the collection of sermon stories ran parallel with the gathering of Mary legends. It was natural, therefore, that the compilers of one series should borrow from the other. Some rather undignified anecdotes written to entertain popular audiences were, consequently, rewritten with great freedom to include the Virgin and added to the number of Mary legends.46 Equally diverting are the Mary legends used by preachers as sermon stories. The stately old miracles, composed for the most part for leisurely monastic audiences, had to be
 [[ Print Edition Page No. 9 ]] 
made briefer and livelier to appeal to impatient lay crowds. This appropriation of Mary legends by popular preachers grew until in the fourteenth century collections of Mary legends were made primarily for the use of preachers, and many of them had become unrecognisable.47

The more frequent contacts between the East and West beginning with the twelfth century and the increasing travel in the West itself stimulated by the pilgrimages and crusades, the growth of the mendicant orders, and the development of commerce opened other new sources. More legends about miraculous images came back from the East as well as new eastern versions of the old stories that had been told by Gregory of Tours. Even merchants were credited with the transmission of legends about Mary images.48 Travellers in western Europe itself carried tales from one place to another by word of mouth.49

Still other Mary legends came into existence in a purely literary fashion. New details were fitted into favorite old themes until whole cycles of Mary legends having the same theme may be distinguished.50 Westerners grown accustomed to visualizing the Christ as a child sitting in the lap of the Virgin transformed eastern tales about crucifixes into Mary legends with a few strokes of the pen.51 Latin legends, it seems, were translated into the vernacular and back again into Latin with the homely details and new incidents which they had acquired in the process.52

As a result of the growth of the Mary cult in all its aspects, the collections of Mary legends rolled up like huge snowballs in the late twelfth and early thirteenth centuries. The Pez collection of the twelfth century comprises forty-three legends. A similar collection about a generation later counts seventy-seven legends, even if the separate miracles of Soissons are not given numbers.53 A compendium made in the same region about 1200 numbers one hundred and five legends,54 and another by Alfonso el Sabio in the Galician vernacular in the thirteenth century, more than three hundred and fifty.55 The manuscript which John of Garland used at Ste. Geneviève in 1248 must have included a large number, for he calls the sixty-one legends
 [[ Print Edition Page No. 10 ]] 
he tells, ‘a few.’56 In all, more than two thousand separate legends originating in the middle ages have been counted, and there are other collections still to be studied. They come from Iceland and from Ethiopia, from Spain and Russia and the middle East; and they are written not only in Latin, but in the vernaculars.

Some of these numerous new anecdotes about the Virgin, originating in the twelfth and thirteenth centuries, gained as wide currency in Europe as those of the HM and TS series. Others got no father than the collections in which they first appeared. A considerable number, however, circulated freely within fairly large areas, more or less coterminous with language boundaries.57 There developed, consequently, in the twelfth and early thirteenth centuries, Latin collections which may be identified as specifically Anglo-Norman, or French, or German in origin.58 At a later date, however, even the latter part of the thirteenth century, the distinctions between regional collections have broken down.

Identification of anonymous regional collections is made in several different ways. The most important is the differentiation of tales characteristic of certain regions whose author or place of origin can be determined. The legend of the ‘Maid of Arras,’ for instance, occurs only in collections which can be traced to northern France. Certain redactions of some legends, especially the ancient TS legends, are characteristic of certain areas. There are, for example, two particular versions of the ‘Milk’ miracle which are found only in collections originating in England and another which is characteristic of collections French in origin.59 The differences between versions of English and French provenance appear clearly in the manuscripts of John of Garland’s work. His collection belongs to the series made in northern France, and the individual who wrote some of the commentary in the British Museum manuscript is familiar only with the redactions of English origin, or with similar English stories.60 Because the first vernacular collections were translations from Latin, they too are useful in determining the regional origin of particular Latin collections.61

Paris with its two great monasteries, Ste. Geneviève and St. Germain-des-Prés, was already an important monastic center, when in 1108 William of Champeaux founded the Canons of St. Victor. They lived under the rule connected with the name of St. Augustine, as did the canons regular of Ste. Geneviève when the monastery was reformed. A particular feature of the Austin canons — they included also the Praemonstratensians founded by St. Norbert in 1120 — was their devotion to the Virgin Mary.62 To the city on the Seine in the twelfth century also came merchants, students, and pilgrims. The road which ran from north to south across the island was one of the great pilgrimage routes. Paris was, therefore, very early a center of the new life in Europe which contributed to the growth of the Mary cult and the collection of Mary legends. Although the narratives themselves offer little proof of the specific place where the collection was made, there are strong indications that the monasteries of Paris did play a leading part in the development of the great French collections.

There are to-day in the Bibliothèque Nationale a number of manuscript collections, obviously related one to another, which were made in northern France during the twelfth and thirteenth centuries. One group in particular connects itself with the monastic and educational centers about Paris. The first and oldest of these is MS Paris Bibliothèque Nationale 14463, written in scripts of the twelfth century. It once belonged to the monastery of St. Victor, hence Mussafia designates it SV.63 The Sorbonne possessed two very similar collections.64 The second is a larger collection, MS Bibliothèque Nationale 12593 of the thirteenth century.65 Since it came formerly from the library of the monastery of St. Germain-des-Prés, it is called SG. Several Paris manuscripts, with no clue as to their origin, are related also to this series of Latin collections, especially MSS Bibliothèque Nationale 17491 of the thirteenth, 2333A of the fourteenth, and 18134 of the late thirteenth or early fourteenth centuries. No similar compilation has been identified as the possession of the third great monastery of Paris, Ste. Geneviève. One did exist there in the thirteenth century, nevertheless, for John of Garland used it in 1248 as the source of the Stella maris. The monastery of Ste. Geneviève, then, had in the thirteenth century a collection very similar to SV and SG, manuscripts which are known to have been the property of the two other great monasteries of Paris. It is, therefore, reasonable to suppose that these two collections were in the thirteenth
 [[ Print Edition Page No. 12 ]] 
century lying on the shelves of their respective monasteries. If they were also originally compiled there, then an important part in shaping the great collections of northern France could be attributed to the Parisian monasteries, for all three collections are important links in the history of Mary legends in northern France in the twelfth and thirteenth centuries.

The Collection of the Monastery of St. Victor

The collection of St. Victor, SV, compiled in the latter half of the twelfth century,66 numbers seventy-seven legends, aside from the entire local collection of Soissons. They fall into two major groups. The first sixty-six, SV¹, are legends, many of them originating in northern France, which became widely disseminated in France and Spain. The Sorbonne collections and two manuscripts of the Brussels Library follow SV only to this point.67 The second series, SV² (nos. 67-77), are a miscellaneous group, among which are interpolated the De transitu beatissime virginis of Melito of Sardis and a commentary on St. Bernard’s sermon Missus est angelus (fols. 72^v-89). Eight of the legends (nos. 67 and 71-77) may have belonged to a single small collection, for they are found together in MS British Museum Royal 6 B x with an introductory paragraph.68 This small collection also originated in northern France, for three of the tales were first told by an abbot of Capelle, one by King Louis of France (probably Louis VII, 1137-1180), and another by a monk of Noyon. It was, perhaps, appended to SV¹ by the monks of St. Victor. Their additions did not, however, become popular enough to be told in the great collections.69

It is the first series, SV¹, therefore, that is significant in the history of the collections made in northern France. In it are included all the miracles of the HM series, three of the Elements-series, and all except one of the TS legends.70 The compilation is not done, however, as in the Cleopatra-Oxford-Toulouse manuscript, simply by adding one series after another in the same order. The Paris compiler put the tales together as one would a mixed bouquet — several from one old source, a few from another, and then some new ones to give variety to the compilation.71 The first seven
 [[ Print Edition Page No. 13 ]] 
numbers of SV¹ are the first seven HM legends, except that no. 7 precedes no. 6. A second series of eight from HM is interpolated after no. 15, although they do not follow in the same order as in Pez. The two remaining HM anecdotes are nos. 39 and 59 of SV¹. The TS tales are treated in the same manner, as are also the remaining Pez legends. The following shows the relationship between TS, Pez, and SV¹:

	TS	Pez		SV1
1.	Toledo	—		41
2.	Foot Cut Off	18		65
3.	Musa	—		31
4.	Mother of Mercy (Sicut iterum)	—		32
5.	Libia	20		42
6.	Gethsemane	21		43
7.	Mary Image Insulted	—		26
8.	Drowned Sacristan (Nonus)	—		—
9.	Devil in Beasts’ Shapes	23		37
10.	Complines	29		27
*11.	Milk: Monk Laid Out	30		28
12.	Three Knights	—		60
13.	Eulalia	32		52
14.	Mead	—		9
15.	Conception	—		10
16.	Leuricus	38		24
17.	Saturday	42		25
	Pez
24.	Son Restored	24		47
25.	St. Dunstan	25		48
26.	St. Dunstan	26		49
27.	Pilgrim in the Sea	27		50
28.	Light on Masthead	28		51
31.	Jew of Bourges	31		66
33.	Jew Lends to Christian	33		53
34.	Hours Sung Daily	34		54
*35.	Love by Black Art	35		36
36.	Abbess	36		55
*37.	Bonus	37		38
39.	Monk Drowns: Friend Prays	39		56
40.	German Nobleman	40		57
*41.	Uncompleted Confession	41		30

The table illustrates clearly a relationship between SV¹ and Pez. Is SV¹ derived from the completed Pez? Or did the compiler of SV¹ have available some collection or collections which may be regarded as the ancestor
 [[ Print Edition Page No. 14 ]] 
of Pez? The latter seems to be the case. In the first place, the St. Victor collection includes tales which belong to the TS series, but which are omitted in the Pez collection (TS 1, 3-4, 7, 12, and 14). Secondly, the version of TS 15, ‘Conception,’ which SV¹ presents is that of the Cleopatra-Oxford-Toulouse collection, not that of Pez 19. In the third place, the arrangement of the legends of SV¹ indicates that the compiler had access to a manuscript which presented TS in the form in which it is found in Cleopatra-Toulouse, not in Pez.72 Note especially the three series TS 3-4 = SV 31-32; TS 14-15 = SV 9-10; TS 16-17 = SV 24-25. Moreover, three of the Elements-series are included in SV¹ in the versions of Cleopatra-Toulouse.73 It appears, then, that SV¹ and Pez share a common source, that one of the sources that the compiler of St. Victor used in the making of his collection included not only HM, TS and the Elements-series, but also the miracles which were eventually to become associated with them to form the Pez collection. The compendium which was their common source was probably French in origin, possibly Parisian.74

These four common compendia account for forty-nine of the sixty-six legends of SV¹. Although some of them may well have originated in northern France, they spread so widely over western Europe that they have lost much of their value as a means of distinguishing the collections of northern France from others. It is the seventeen (or eighteen) legends75 of SV¹ still unaccounted for that characterize the collections made in this region. For many of them it is their earliest appearance in a collection of Mary legends, and the narrative is embedded in the considerable introductory matter which belonged to them in their original form. They were probably extracted from their original sources by some monk of Paris:

11. Mother of Mercy (Pseudo-Anselm). Meminimus et meminisse delectabile est.

12. Foot Cut Off: Grenoble. Dilectis in Christo fratribus . . . Cum nuper in territorio Grannopolitano.

13. Bridegroom: Transported To a Remote Region. Audiant adhuc quos audire delectat. . . . Iuvenis erat predives. S.M. 35.76

14. Milk: Tongue and Lips Restored. Clericus quidam vite secularis. S.M. 1.

15. Mouth of Hell. Sanctimonialis quedam sancte enutrita. S.M. 39.

29. Monk Dies Suddenly in Burgundy. Olim erat cognitus alter quidam monachus (versified).

 [[ Print Edition Page No. 15 ]] 

33. Sacristan Worthy to Kiss Hands and Feet of Mary. Erat in quodam sancte dei genitricis . . . cenobio frater.

34. Boy Devoted to the Devil. Inter innumera dei genitricis . . . insignia. . . . Fuit quidam nobilis affluensque. S.M. 52.

35. Excommunicate Absolved by Foolish Servant of Mary. Preiudicatis quippe nonnullis in seculo eundi ad penas. . . . Erat quidam mire sanctitatis presbyter.

40. Will for Deed. Fuit quidam miles nobilitate et dignitate conspicuus. S.M. 55.

44. Bread. Spiris est locus famosus. S.M. 49.

45. Chaste Empress. Quam mercedem legale coniugium. . . . Hic siquidem de quo loqui incipimus imperator. S.M. 14.

58. Maid of Arras. Memoriam gloriose virginis. . . . Erat igitur in suburbio civitatis Atrebatensis. S.M. 33.

[58a. Poor Man Strikes Stone. Quam magna sit. . . . Quidam pauperculus homuntio Walterus nomine in vicinia Aquicinensis]77

61. Rich Man and Poor Widow. Fuit ecclesia cuius parrochie presbiter preerat (versified).

62. Incest. Quam secreta regis abscondere. . . . Erat Rome vir quidam nobilis (versified).

63. Stepmother and Stepson. Chiviachus villa est episcopii Laudunensis.

64. Wife and Mistress. Est preterea aliud quiddam relatu breve exemplo. . . . Ex relatione Atrebatensis episcopi mulier quedam fuerat.

The whole of Hugo Farsitus of Soissons is included in SV¹, preceding no. 58.

Four of the eighteen legends characteristic of SV¹ are taken more or less literally from two ecclesiastical writers of the eleventh and twelfth centuries, both of whom wrote in northern France. The first, no. 11, included among the works of Anselm of Canterbury, is by Maurilius, archbishop of Rouen, 1055-1067.78 Three, nos. 12, 63, and 64, come from the De laude sancte Marie of Guibert de Nogent, chs. 10-12.79 Another tale, no. 44, is much the same anecdote as that told by Gautier de Compiègne, De miraculis beate virginis Marie,80 but the diction is so altered and the details so different that it can scarcely have been taken directly from that work. Two of the legends, nos. 58 and 64, originated at Arras; one at Anchin near Douai, no. 58a; two, nos. 12 and 29, in Burgundy; and one, no. 44, at Speyer. The Speyer legend is, however, probably not of German authorship, for the author remarks of a phrase which he quotes, ‘as German boys say.’

The trail of four of the legends which are new in SV¹ leads back to
 [[ Print Edition Page No. 16 ]] 
Rome, and suggests an earlier period. The legend of ‘Chaste Empress’ appears to be from an ancient collection connected with Roman emperors in which the Virgin at first had no part. In the tradition of three of the others, ‘Bridegroom: Transported,’ ‘Boy Devoted to the Devil,’ and ‘Incest,’ early popes are important characters.81 It is possible, therefore, that the compiler of SV¹ had available also an ancient collection of legends from Italy. The ‘Milk’ legend, no. 14, is a redaction of one of the TS miracles. The origin of the remaining five legends is unknown. All of the eighteen are related again and again in the collections made in northern France and Spain in both Latin and the vernaculars. With some exceptions they are not so frequently told elsewhere in the twelfth and thirteenth centuries. John of Garland tells nine of them in the Stella maris.82

The Collection of St. Germain-des-Prés

Analysis of the St. Victor manuscript suggests the way in which collections of Mary legends grew in the twelfth and thirteenth centuries. Each compiler copied and rearranged the old legends. He rejected some, perhaps because they did not suit his fancy, because a certain number were needed for the requirements of the services, or because he had a certain limited amount of parchment to fill. In any case few collections were made without adding some legends from a new source. So it was with the monk who brought together the collection of St. Germain-des-Prés, MS Paris Bibliothèque Nationale 12593 (SG) of the thirteenth century.83 The legends, therefore, fall into two series: SG¹, nos. 1-80, and SG², nos. 81-105.

The compiler of the collection of St. Victor had been satisfied with the old prologue, beginning Ad omnipotentis Dei laudem, which stood before the HM series. The compiler of the legends of St. Germain-des-Prés, embarking upon a more ambitious project, writes a new one, beginning Quoniam gloriosissima virgo virginum. In the last paragraph he announces his intention to bring together the miracles of the Mother of God, those performed ‘in different times and in diverse places, upon diverse persons of either sex, of different age, and of different condition and status; whatever is to be found in the books of the saints or scattered about in the writings of the faithful.’84

The writer of this prologue intended to do a good deal of original research in ecclesiastical literature, but, none-the-less, the nucleus of his collection is SV¹. Of its sixty-six legends, sixty are repeated in SG¹ in almost
 [[ Print Edition Page No. 17 ]] 
identical words, and another, ‘Jewish Boy,’ is told in a different version.85 It was, however, not the manuscript of St. Victor itself that the compiler of St. Germain-des-Prés used, but another similar to it which he, perhaps, found in his own monastery. A legend obviously belonging to the SV¹ series, but lacking in the St. Victor manuscript, ‘Poor Man Strikes Stone,’86 is no. 75 of his collection. Either he, or a compiler who preceded him, was, moreover, strongly influenced by a twelfth-century collection which Mussafia designates as APM,87 so-called because it exists to-day in three manuscripts, MSS British Museum Arundel 346 (twelfth century), Paris Bibliothèque Nationale 18168 (twelfth century), and Montpellier 146 (twelfth to thirteenth century), as is shown by the table on the following page.

The first twenty-six legends of SG¹ in general follow the sequence of APM, except for Sicut iterum which is inserted between HM 4 and 5, TS 16 (‘Leuricus’) which is added, and TS 3 (‘Musa’) which is deferred. As in SV, however, Pseudo-Anselm is interpolated after ‘Conception.’ The first twenty-five legends of SV¹, on the other hand, show no marked relationship to APM, except the series‘Mead’- ‘Conception’ (nos. 9-10). Nevertheless, beginning with no. 26 and continuing through no. 43, SV¹ is tied more closely to APM than is SG¹ (see nos. 31-32, 42-43, 28-30, etc.). The APM sequence ought to be older than the sequence of SV¹ 1-25, because it reproduces HM 1-17 intact and certain series of TS in order (TS 12, 14-15, 1-6, 11, to which SG adds TS 16). Could there have been at St. Germain-des-Prés a copy of SV¹ older than the manuscripts of St. Victor, the Sorbonne, and the Brussels collections which reproduced the first twenty-six legends more nearly in APM sequence? It is difficult to explain the marked parallels between SV¹, SG¹, and APM in any other way. There are two suggestions of such a collection in the Sorbonne manuscripts themselves. They follow SV¹ almost exactly except that (1) ‘Three Knights’ (TS 12) is placed after Pseudo-Anselm (no. 10) in approximately the same place as in SG¹ and (2) ‘Theophilus’ has been deferred, as in SG¹, until later in the collection (no. 27). The original home of SV¹, then, may well have been the old monastery of St. Germain-des-Prés rather than the newer

 [[ Print Edition Page No. 18 ]] 

	APM			SV1				SG1				Sorbonne
1-17.	HM, 1-17 (ending ‘Murieldis’)	1-7,		18, 39, 23, 16, 17, 19-22, 59	1-4,		6-18 (ending‘Murieldis’)	1-7,		13, 23, 18, 32, 12, 14-17, 60
18.	Jew of Bourges P 31	66			26,		but a different version	67
19.	Three Knights TS 12	60			19			11
20.	Mead TS 14	9		8. Theophilus	21		50. Theophilus	8		27. Theophilus
21.	Conception TS 15	10		11. Mother of Mercy (Pseudo-Anselm)	20		22. Mother of Mercy (Pseudo-Anselm)	9		10. Mother of Mercy (Pseudo-Anselm)
					23			24
22.	Toledo TS 1	41			24		25. Leuricus TS 16	64
23.	Foot Cut Off TS 2	65			103			21
24.	Musa TS 3	31
25.	Mother of Mercy (Sicut iterum) TS 4	32			5			22
					37			45
26.	Libia TS 5	42			57			46
27.	Gethsemane TS 6	43
*28.	Milk: Monk Laid Out TS 11				78			35
		28			—			36
*29.	Monk Dies Suddenly	29
*30.	Uncompleted Confession P 41				41			37
		30			—			—
*31.	Wife and Mistress	—			43			41
*32.	Love by Black Arts P 35	36			—			43
*33.	Bonus P 37	38			27			33
34.	Mary Image Insulted TS 7	26

 [[ Print Edition Page No. 19 ]] 

St. Victor. The re-ordering of the legends of SG¹ beginning with no. 27 is to be explained by the desire of the compiler to group together legends with similar themes.88 A series of five, nos. 27-31, are about Mary images; nos. 34-35 tell of sinners saved by repeating the prayer O intemerata; nos. 45-46, of cities saved by the Virgin; and nos. 48-56, of Mary festivals and Mary in the liturgy.

The basic source of SG¹ is the twelfth-century collection SV¹, or an earlier compilation, of which SV¹ is a redaction. The addition of seventeen legends distinguishes SG¹ from SV¹, legends which together may be used to differentiate the Quoniam (the first word of SG’s prologue) collections and their descendants from others made in northern France.

29. Bridegroom: Ring on Finger. In antiquis temporibus factum de ymagine genitricis — Erat quedam ecclesia in qua imago. S.M. 8.

30. Orleans. Omnis cetus fidelium audiat hoc miraculum — Quoddam municipium est Aurelianensi proximum civitati quod Avenon nuncupatur. S.M. 26.

31. Saracen and Mary Image. Contigit Sarracenum quendam habuisse ymaginem. S.M. 7.

32. Unchaste Monk Warned by Widow. In territorio Cameracensi quoddam cenobium est.

33. Mare. De matre misericordie matre domini mirabilia multa narrantur — In vicinia Remensis urbis est quoddam cenobium regule Premonstratensi deditum. S.M. 29.

34. O intemerata: Devil as Servant. Homo quidam erat nobilis divitiis potens et honoribus.

35. O intemerata: Buried Outside the Churchyard. Aliud quoque beneficium gloriose virginis de eadem oratione — Erat quidem adolescens nequam.

45. Constantinople Saved. In diebus Theodosii predecessoris Leonis pape.

46. Chartres. Temporibus Karoli regis Francorum qui cognominatus est Simplex anno ab incarnatione domini 806. S.M. 32.

47. Sardenay. Tempore quo Greci terram inhabitabant.

48. Nativity. Sancte dei genitricis semper que virginis Marie navitatis — Solitarius quidam sancte vite fuit. S.M. 15.

49. Purification. Temporibus Iustiniani Augusti quinto decimo imperii ipsius anno. S.M. 30.

52. Thread in Lip. Dominus noster Ihesus Christus dei filius annuntiatus — Erat quidem tunc temporis in Gallia in suburbio scilicet Noviomince urbis quedam puellula Heremburgis nomine.

53. Origin of Antiphony. O Maria virgo, pia maris stella. Hanc suprascriptam antiphonam docuit s. dei genitrix monachos in quodam cenobio — Laicus quidam inibi degens.

54. Sight Restored. Est responsorium quod quidam asserunt in ecclesia non debere cantari — Responsorii . . . istius . . . auctor exitit quidam Romanus nomine Victor.

 [[ Print Edition Page No. 20 ]] 

56. St. Luke’s Portrait of Mary. Dum beatissima dei genitrix moraretur cum apostolis.

58. Barns Filled. Monasterium est valde magnum in Iherusalem. S.M. 6.

Of the seventeen tales which are new in SG¹, five are from monastic or communal records of northern France, nos. 30, 32, 33, 46, and 52. Six are anecdotes connected with the liturgy, nos. 34-35, 48-49, and 53-54. Two are ancient tales about Mary images, nos. 31 and 47. One is from Gregory of Tours’ In gloria martyrum, ch. 10, no. 58. The other three, nos. 29, 45, and 56, have nothing in common except that they originate in Rome or mention early popes. This is true also of nos. 53 and 54, as well as four other legends told in SV¹ and repeated in SG¹, except for ‘Incest’ which is lacking only because by some accident it was versified.89 These legends probably belonged to an ancient collection of Roman origin brought to France. A further hint of the existence of such a collection comes from the British Museum gloss of the Stella maris, written by some one who had seen the collection of Ste. Geneviève. John of Garland tells all of the legends except ‘Constantinople,’ and the author of the gloss remarks, Cum maximo affectu compositus illo modo totaliter de miraculis beate virginis, prout in registro Rome et etiam alibi comprobantur.90 There seems to be no other explanation for this statement, except that a particular group of legends in the manuscript of Ste. Geneviève were said to come from an ancient papal ‘register.’ Moreover, some of the anecdotes, together with others of the same nature, are related in MS Paris Bibliothèque Nationale 5268 of the twelfth century. One of them, a tale about Leo the Great, begins, Romanorum testimonio didicimus et in commentariis pontificalibus.91 These two phrases, it should be noted, cover both types of ancient legends among the Quoniam anecdotes.

The legends of the first series of the collection of St. Germain-des-Prés, therefore, clearly point to northern France as their home. Some come from collections already in existence there, probably including one of Roman or Italian origin. Others are the result of the search of the works of ecclesiastical writers, of chronicles, and of monastic and communal records promised by the compiler in his preface. It is in the second series of the collection that he most specifically carries out his intention to tell tales ‘of diverse persons of either sex, of different ages and different conditions and status.’ These, as well as the seventeen tales of SG¹, probably were his own additions. In SG² he groups the legends according to the social status of the individuals about whom they are told. Numbers 81-90 are about priests
 [[ Print Edition Page No. 21 ]] 
and monks; nos. 91-92, about conversi; nos. 93-95, about clerks; no. 96, about a nun; and nos. 97-105, about laymen. The laymen include an emperor, a knight, a servant, an impoverished man, and a blind man. Two of the tales are about boys. It may be inferred that the compiler was a monk, for he puts even the conversi of the monastery before the clerk and the nun, to say nothing of the emperor and the knight.

Of the twenty-five tales of SG² only three, nos. 101-103, are well-known in France, and even they are not the usual versions. The others suggest some region farther from Paris. Does SG² offer any clue as to the home of the compiler and the origin of his legends?

A first examination of the series, SG², suggests Germany or at least the Rhineland as the home of these tales. The scene of two of the legends is the Rhineland, no. 100 at Strassburg92 and no. 93 at Trier.93 A third occurred in a monastery in Germany.94 Two themes particularly popular in German lands predominate in the collection, the ‘Drowned Sacristan’ theme,95 and the theme of the ‘Bread’ legend.96 In fact this particular ‘Bread’ legend seems to be the point of departure for all others of German origin. Two of the three popular legends, nos. 101 and 102, are from thirteenth-century collections which exist to-day in German libraries,97 and one of them, no. 101, is a version seldom found in the French collections. Mussafia reports three collections (MSS Copenhagen Thott 26, 12th-13th centuries; and Leipzig 821 and 819, 13th century) which are related to SG. The Copenhagen collection and the second Leipzig manuscript repeat only the first legend of SG², no. 81. The first Leipzig collection includes six, nos. 81, 88-90, 96, and 101.98

The first series of the collection of St. Germain-des-Prés presumes northern France as its home, and the second series strongly suggests Germany or the Rhineland. In view of these facts, was the compiler a German monk who brought SG² with him to Paris? Or was he, perhaps, a French monk who had visited Germany? The author of the ‘Bread’ legend, no. 104, writes at the close of his narrative, ‘This, the lord Louis, abbot of St. Peter at Châlons-sur-Marne,99 is a witness, was related to him by the abbot of
 [[ Print Edition Page No. 22 ]] 
that very monastery [where the incident occurred], a very truthful and pious man, when he was studying in Germany in that same monastery.’100 The author says also that the abbot Louis frequently told the anecdote to his monks in his [the author’s] hearing.101 The legend which follows the ‘Bread’ miracle and closes the series, no. 105, comes also from the region of Châlons-sur-Marne. It is a tale of a young boy whose soul was rescued by St. John because he repeated the prayer O intemerata to Mary and St. John the Evangelist. The prayer was found upon his pillow when he died. Gonterus (fl. 1156), abbot of Chalade, a Cistercian monastery near Clermont-en-Argonne founded in 1138,102 heard it at a chapter meeting of Cistercian abbots, to which a priest brought the scroll with the prayer on it. Gonterus wrote down the prayer.103

The collection of St. Germain-des-Prés could, therefore, have been made by a monk who had come to Paris at the end of the twelfth or the beginning of the thirteenth century from the region east of Paris toward the Rhineland bringing with him a collection from a monastery of that region. If the compiler of SG² was the author of the ‘Bread’ miracle he was a monk of the Benedictine monastery of Châlons-sur-Marne.104 In the Benedictine monastery on the Seine he made another collection gathered from the sources available there and appended his earlier collection to it. The whole he refurbished with a prologue of his own, and it became the progenitor of a group of Quoniam collections, of which John of Garland’s source, the collection of Ste. Geneviève, was one.105

3. MSS Paris Bibliothèque Nationale 17491 AND 2333A

Nothing is known of the origin or the compiler of the large collection of MS Paris Bibliothèque Nationale 17491 (X)106 in a script of the thirteenth century. It could scarcely have been made earlier than the last
 [[ Print Edition Page No. 23 ]] 
quarter of the twelfth century, for it includes an incident recorded by the Cistercian monk, Helinand of Froidmont (d. after 1227), under the year 1161,107 and a tale told in his later years by the abbot Baldwin of the Praemonstratensian abbey of Belleval in Lorraine, who flourished in 1179.108 The work is constructed according to no discernible plan109 except to include under one cover as many Mary legends as possible. Among the more than ninety numbers are the following:

• 1.  The complete collection of Soissons by Hugo Farsitus,110 of Laon,111 and of Rocamador.112
• 2.  Gregory of Tours, In gloria martyrum, chs. 8-10.113
• 3.  A tale from Helinand’s Chronicon.114
• 4.  A legend from the Cistercian monastery of Savigny in the diocese of Avranches which is not a Mary legend.115
• 5.  Some verses, De transitu virginis Marie, attributed by their author to Melito of Sardis.116
• 6.  Sixty-nine of the eighty legends of SG¹.
• 7.  Ten versified legends.117
• 8.  Two unique legends.118

The collection ends with remarks on various incidents in the life of the Virgin, the Annunciation, Assumption, etc.

The work of the compiler of MS Paris 17491 (X) falls into two distinct parts, each of which closes with an unique legend.119 The first series, X¹, nos. 1-72, are entirely in prose; the second, X², nos. 73-82, are in verse, or partly in verse, and in some cases the details differ from the prose versions in other collections. The first collection, X¹, is derived either from SG¹ or from a common source. The prologues, Quoniam gloriosissima virgo,120 are identical; and not only are sixty-nine legends the same in both collections, but there are several long series which follow in the same sequence:

SG1	X1	SG1		X1
2	14	61		18
3	15	62		21
4	16	63		22
5	17121	64		26
		65		27
13	31	66		28
14	32
15	33	67		60
16	34	68
		69		61
29	67
30	68	70		40
31	68a	71		41
32	68b
33	68c	74		63
34	69	75		64
35	70

Of the seventeen legends which differentiate the collection of St. Germain-des-Prés from the earlier one of St. Victor, all but two122 are repeated in the first series of X in much the same order.123 Six legends, neither in SV nor SG, have been added:

7. Columns Raised. Beata dei genitrix et perpetua virgo . . . cuius basilica a Constantino ammirabili opere fabricata renitet. S.M. 13.

 [[ Print Edition Page No. 25 ]] 

8. Light in a Mary Church. Huius gloriose virginis reliquie in oratorio Marciacensis Averni territorii continentur.

9. Jewish Boy. Quid etiam in oriente actum fuerit . . . Iudei cuiusdam vicarii filius. S.M. 3.

11. Mary Relics. Huius itaque beate virginis reliquias quadam die super me in cruce aurea positas exhibebam.

65. Judas in Hell. Omnis qui de testimonio veritatis novit agere. S.M. 27.

65a. Christ Appears to Monk. Quoniam de visionibus studio aptare pennam cepimus quod in monasterio Saviniensi contigit sub venerabili patre . . . Serlone relatu dignum credimus. Unus fratrum cuius nomen novimus. S.M. 40.

The first four incidents are from Gregory of Tours’ In gloria martyrum, chs. 8-10.124 The other two are tales of visions seen by monks. The first is recorded under the year 1161 in the Chronicon of the Cistercian monk of Froidmont, how a novice in a monastery in England saw a vision of heaven and hell. The second incident occurred in the Cistercian monastery of Savigny when Serlo de Vaubadon was abbot, 1140-1153 (abbot of Clairvaux, 1153-1158), how Christ appeared to a monk on two different occasions when he was saying mass. If the two collections do not emanate from a common source, then MS Paris 17491 (X¹) would seem to be derived from the first series of the collection of St. Germain-des-Prés, because of the slightly later date, as compared with SG¹, of the legends which distinguish it;125 and because their common prologue is more at home in SG than in X.

The versified series, X², of MS Paris 17491, on the other hand, is derived either from the first series of the collection of St. Victor (SV¹) or from a common source. Of the ten numbers, seven are identical with those of SV¹, and the sequence is almost the same in both collections:126

The manuscript of the Bibliothèque Nationale, 2333A, in a script of the fourteenth century, is little more than a copy of MS Paris 17491. Four of the legends are omitted,127 and two have been added.128 The legend, ‘Theophilus,’ in verse replaces the prose version of X. Otherwise the legends follow in the same order.129

The growing tendency to versify the legends of the Quoniam collections is observed in MS Charleville 168 of the thirteenth century, a collection which has much in common with MS Paris 17491. The compiler of X², or some one who preceded him, versified a part of Melito of Sardis. He attempted also to turn two others into verse, but gave up after the first two lines. The compiler of the Charleville collection adds to the number, ‘Mare’ and ‘Unchaste Monk Warned by Widow,’ besides another legend which is unique.130

4. MS Paris Bibliothèque Nationale 18134

The fifty-nine legends of MS Paris Bibliothèque Nationale 18134 in a script of the second half of the thirteenth century or the beginning of the fourteenth131 may be divided into two distinctly different series: Q¹, nos. 1-26, except nos. 1 and 17; and Q², nos. 27-59 and nos. 1 and 17. The first series, Q¹, depends upon the first series of the collection of St. Victor, for the legends are identical in form and diction. They are told in almost the same order in both collections except that some have been omitted from Q¹.132 The prologue is the ancient Ad laudem prologue, as in SV.133

The second series, nos. 1, 17, and 27-59 (Q²), is one of the most interesting and mysterious series in all the collections of Mary legends in Latin. It is difficult to believe that they were originally put together by the same person who compiled Q¹, for the treatment of the two series is very different. The compiler of the first series has painstakingly followed his source, except for the omission of some of the legends and the rearrangement of others. In Q², not even the most venerable legends escape radical alteration in diction. Others are revised to make them briefer and more dramatic in the manner of the vernacular versions of the thirteenth century. The Latin style of the legends of Q² is more direct than that of the original versions, and the form is often cruder.

The thirty-five legends of Q², aside from nos. 41 and 58 which are Pez legends, may be divided into four groups according to the sources from which they appear to come:

1. Eight legends, nos. 17, 27, 35, 42, 54-56 and 59,134 are unique, or not often told in Latin collections made in northern France.

2. Fourteen legends, nos. 28-30, 34, 39, 43-47, and 49-52,135 are sufficiently similar in diction, detail, and sequence to the collection in the French vernacular of Gautier de Coincy136 that it must be concluded either (1) that Gautier de Coincy used Q² in the composition of his work or (2) that the author of Q² was following the text of Gautier de Coincy. Although it is not possible with the evidence at hand to conclude definitely that Gautier’s vernacular collection was the source of the fourteen legends of Q²,137 a study of some of the narratives in the next classification as well as the script of MS 18134 would indicate a date later than 1223 for the composition of Q², and therefore lead to the second conclusion.

3. In the third group are five legends, nos. 32-33, 36, 38 and 48,138 identified with a Latin collection referred to as the Mariale magnum, but not now extant.139 They are similar in theme to certain tales found in other collections where they are said to have been extracted from that enigmatical work.140 In addition to these five legends, said to be from the Marialemagnum, which do not appear in Gautier de Coincy’s collection, there are ten or eleven other Mariale magnum legends among those fourteen in the second group which are common to both Gautier de Coincy and MS 18134, nos. 28-30, 34, 39, 43-44, 46-47, 49, and probably 50.141 These numbers include all but one of the legends of MS 18134 in whose diction there are traces of both Gautier de Coincy and another source which Mussafia believed to be SG.142 But, the SG legends are the same versions as those found in Vincent of Beauvais and attributed to the Mariale magnum. The compiler’s second source, therefore, could have been, not SG, but the Mariale magnum. In view of this fact and of the appearance of the five legends of the Mariale magnum in Q², but not in Gautier de Coincy, it would seem that the author of Q² (provided he did use Gautier de Coincy as his source and not vice versa) was employing both Gautier de Coincy and the Mariale magnum as
 [[ Print Edition Page No. 29 ]] 
sources. When he came upon a legend told in both his sources, his instinct for the dramatic led him to rely chiefly upon the vernacular version of Gautier de Coincy. Legends which he found only in the Mariale magnum, he revised in the spirit of the vernacular examples he had before him.

Some clue as to the date of Q² of MS Paris 18134, aside from the script in which it is written, is furnished by several of the legends in this third group. The evidence, such as it is, supports the conclusion that the collection of Gautier de Coincy, although in the vernacular, is earlier than that of Q² in Latin. Three of the five legends of the Mariale magnum are among the narratives of the second collection of MS British Museum 15723 of the thirteenth century.143 The first series of this compendium depends upon Vincent of Beauvais. The entire collection is said to have been taken from the Mariale magnum.144 The details of three legends make it possible to date them. Ward places one not earlier than 1161,145 another not earlier than 1180,146 and a third not earlier than 1200.147 A fourth legend, appearing in the Mariale magnum, but not in Q², could have been told in its original form as late as 1223.148 It is not at all likely that legends with termini ex quo so late as 1200 could have been written down in Latin and then revised by the compiler of Q² in time for Gautier de Coincy to use them in the early 1220’s. Therefore, in the absence of other evidence, it may be concluded tentatively that Q² of MS Paris 18134 was composed from at least two sources, a Latin collection known by reputation as the Mariale magnum and the vernacular collection of Gautier de Coincy. As such it is a most unusual collection, for ordinarily vernacular collections were made by translation from the Latin, not Latin collections from the vernacular.

4. The fourth group of legends comprising Q² is a miscellaneous group of six, nos. 1, 31, 37, 40, 53, and 57.149 One of them, no. 40, how the Virgin rebuked a Cistercian monk who sang softly, has the same general characteristics as the legends of the Mariale magnum, and may well belong there.150 It exists today in the French vernacular in the Miserere of the Renclus of Moiliens,151 written about the end of the twelfth century. A second anecdote, no. 57, is the same tale as no. 60 of MS Rouen A 535, a collection closely related to the Mariale magnum.152 The first legend of the manuscript
 [[ Print Edition Page No. 30 ]] 
is another which has been identified in the same version only in a vernacular collection.153 A story similar to no. 37, how a widow taught her son to lay garlands upon the altar of the Virgin, is told by the Dominican preacher Herolt in the next century.154 The ‘Chorister,’ no. 31,155 and ‘Beatrice the Sacristan,’ no.53, could also have come to MS Paris 18134 by way of the Mariale magnum.156

It is probable that the original home of Q² was Soissons, or the vicinity of Soissons, for the trails which lead to and from the collection tend to converge on that region. Not only did the author probably use the Soissons manuscript of Gautier de Coincy, the monk of St. Medard at Soissons, as one of his sources, but a number of the legends told by him appear in a large compilation of Mary lore in the vernacular made not earlier than 1328 by an anonymous of Soissons, part of which is known as the Rosarius. There may have been a copy of the Mariale magnum available in the Soissons region about 1325, a possible date for Q² as indicated by the script of MS Paris 18134, because the Rosarius depends upon a Grant Marial as its principal source.157 The author of the Rosarius knew the work of the Renclus of Moiliens, whose vernacular poem, Miserere, yields the only parallel to one of the legends of Q² of MS Paris 18134.158

5. The Rouen Mariale

The title ‘Mariale’ is defined in the thirteenth century as a collection of materials in praise of the Virgin, a sort of anthology or summa of Mary lore. Such volumes were useful in the monasteries and churches as sources of readings for Saturdays and the celebration of the great festivals of the
 [[ Print Edition Page No. 31 ]] 
Virgin Mary.159 Some of the collections already described include scatered materials other than legends. The collection of St. Victor interpolates among the miracles, the De transitu of Melito of Sardis and a commentary on a sermon of St. Bernard. The compiler of MS Paris 17491 versified part of the De transitu, and he concludes the collection with references to scenes from the life of the Virgin.

The Public Library of Rouen has, among others, two collections of Mary legends, one of which is a copy of the other. The earlier, MS Rouen U 134160 of the thirteenth century, bears the title, Mariale, quod est de laude gloriose et perpetue virginis Dei genitricis Marie. Both belonged to the Benedictine monastery of Jumièges. The Rouen Mariale was not originally a single collection, for it has two prologues, one awkwardly following the other. The first, In libro Mariali, beginning Mirande virginis laudes qui miratur, mirandus est nemo, ends with the statement that the work is compiled‘from miracles and sayings of the fathers, ancient and modern.’161 In spite of this declaration, the thirteenth-century copy162 is made up of
 [[ Print Edition Page No. 32 ]] 
legends only, four of which are not Mary legends. A second prologue, Generalis prefatio, which belongs specifically to the series of Mary legends, follows immediately. It is the prologue of SG and MS Paris 17491, Quoniam gloriosissima virgo virginum.163

As the second prologue foreshadows, the Rouen Mariale is a descendant of the collections of St. Germain-des-Prés and MS Paris 17491, X¹. Of the sixty-four legends of the Rouen collections, the first forty-three, R¹, are the same, except for changes in diction and sequence, as those of the first series of MS Paris 17491. Among the anecdotes common to both are three of particular significance because they have been found only in Quoniam collections, including the Stella maris: the vision of ‘Judas in Hell’ from Helinand of Froidmont; following it, as in MS Paris 17491, that of the monk to whom Christ appeared while he was saying mass; and the legend of the fire in the church at Mare. The series from Gregory of Tours follows in the same order in both except that chapter 18 of the In gloria martyrum has been interpolated between the two legends of chapter 8, and chapter 10 has been omitted. Both are legends about Mary relics. The fact that the series in MS Paris 17491 (X¹) is chapters 8-10 without modification suggests that chapter 18 is a substitution made by the compiler of the Rouen Mariale or his source, and that X¹ of MS Paris 17491 is the earlier collection. That the Rouen Mariale was not copied directly from X¹ as it appears in the Paris manuscript 17491, however, is indicated by the presence of ‘Musa,’ which is omitted in X¹, in its appropriate place before ‘Columns Raised,’ as in MS Paris 2333A, clearly related to X¹.164 The Stella maris includes, not only the three distinctive legends common to MS Paris 17491 and the Rouen Mariale, but also three from the Gregory of Tours series.165

There is other evidence that MS Paris 17491 was not the immediate source of the Rouen Mariale. The Rouen collection uses prose versions of two legends,166 versified or partly versified, in the second series, X², of MS Paris 17491 and in the collection of St. Victor. Their presence here is significant, because they are part of a series which appears in versified redactions in the Parisian collections, with the exception of the Stella maris, and in prose versions in the Mariale magnum and the Stella maris.167 These facts imply not only that MS Paris 17491 was not the source of the Rouen Mariale, but also that the Rouen Mariale shares a common source with the Mariale magnum, or that its source was the Mariale magnum.

The second series of the Rouen Mariale, R² (nos. 44-64), provides more evidence of its relation to the Mariale magnum. The twenty-one legends include six which have not been found in other collections,168 and six which are more or less familiar.169 A third group of eight are characteristic only of the Mariale magnum and collections related to it:

44. Painter of Flanders.170 Caritati diligentium virginem matrem Christi Iesu Mariam incentivum aliquod superaddi cupientes, parvum sed pulchrum quid de pictore quodam loquimur, qui eidem Domine pro vite sue merito familiaris in partibus Flandrie sibi nomen fecerat. S.M. 54.

45. Unwilling to Deny Mary.171 Lapsis gravi peccatorum ruina . . . In Aquitanie itaque partibus in castello quodam milites duo principabantur.

46. Cisterian Monks at Their Field Work.172 Laborantis anima laborabat sibi . . . Nobilis quidam honeste in seculo vite timens ne in vacuum curreret.

47. Christ Image Broken by Brabantine Blasphemers.173 Iuxta castrum Radulfi est quedam abbatia, que Dolis vocatur.

48. Cistercian Monk Persecuted.174 Rei geste quam sequens perstringit lectio.
 [[ Print Edition Page No. 34 ]] 
. . . In quodam aiebat [Simon, abbas de Los] Cisterciensis [ordinis] monasterio fuit monachus quidam religiosus.

53. Monk Who Recited Five Psalms Daily.175 Quidam archiepiscopus Cantuariensis ecclesie.

60. Chaplain Whom Mary Chose.176 In territorio Lexoviensi quidam fuit iuvenis natalibus ortus non infimis.

62. Demons in the Form of Swine.177 Fuit ex eorum numero, quos Cartusiensis ordo suscipit laicus quidam humilis genere.

Besides the two prose legends of R¹ and the eight tales of R², eleven others178 of the narratives of R¹ are also identified with the Mariale magnum by the compilers of collections yet to be analyzed.179 Moreover, the sequence of certain of the legends of both R¹ and R² was determined either by the Mariale magnum or a common source.180

The Rouen Mariale, therefore, betrays evidence of two collections: (1) a Quoniam collection similar to, but not the same as MS Paris 17491 (X¹, but not X²); and (2) the Mariale magnum. It was not composed, however, simply by uniting two collections, for the sequence of the legends throughout must have been influenced by the Mariale magnum, or a collection much like it. On the other hand, the more detailed comparison of the Rouen Mariale with the collection of Vincent of Beauvais which follows does not justify the conclusion that the Rouen Mariale was descended from the Mariale magnum. It was, instead, probably derived from a single collection which was the common source of both. That collection, as yet unidentified, will be referred to at present as the Ur-Mariale.181

The title Mariale was, perhaps, borrowed from its source, as was also the title Mariale magnum, hence the first prologue of the Rouen collections, In libro Mariali. The concluding sentence of the same prologue, huius libelli series ex miraculis et ex dictis patrum veterum et recentium tota contexitur,
 [[ Print Edition Page No. 35 ]] 
makes a promise that is not really fulfilled in the Rouen Mariale,182 which is made up entirely of legends. It is fair to suppose that the source from which the legends were drawn did include matter of this sort. A study of the collections descended from the Mariale magnum yields evidence that it did.

The following table compares the Rouen Mariale with other collections:

6. The Mariale magnum

The history of Mary legends in France presents no more intriguing problem than the Mariale magnum, known at present only by reputation. It must have been a work of importance in its day, for it is cited by French compilers of Mary lore from the middle of the thirteenth to the fifteenth century. The most accurate information as to its nature and content comes from the collection of Mary legends which Vincent of Beauvais incorporates in the Speculum historiale.185 The work was probably completed by 1244, certainly by 1247.186 Since one of the legends which the compiler
 [[ Print Edition Page No. 37 ]] 
attributes to it mentions the year 1187,187 the Mariale magnum must have been compiled, as Vincent of Beauvais used it, in the sixty years between 1187 and 1247.

MS British Museum Additional 15723

The second collection of Mary legends in MS British Museum Additional 15723, in a script of the thirteenth century, names the Mariale magnum as the source of its legends.218 It begins with Vincent of Beauvais’
 [[ Print Edition Page No. 45 ]] 
note, obviously copied from the Speculum historiale. The series of Mary legends which follows agrees substantially with Vincent of Beauvais through chapter 119a,219 or well beyond the end of the series which the Dominican took from the Mariale magnum. Several of the legends of Vincent of Beauvais are missing, but none of those omitted is a Cistercian legend.220

The forty-three legends of MS Additional 15723, therefore, comprise two series, A¹, (nos. 1-32, with the exception of no. 20) and A², (nos. 34-43). The first series, A¹, was copied from Vincent of Beauvais, who took them from the Mariale magnum, a Cistercian collection. The second series, A², is also Cistercian. Ten of the eleven incidents occur in Cistercian houses of England or France, or are reported by abbots of Cistercian monasteries of those regions. Seven are from northern France.221 Eight of them are later retold by the abbot of a Cistercian house on the borderline between the diocese of Chartres and Paris.222 The anecdotes of A² are related in very circumstantial fashion including names and places that can be readily identified. Ward’s date for the collection, not earlier than 1200 and while Philip Augustus was still king of France, that is 1200-1223,223 may well apply to the Mariale magnum from which the compiler said he had the legends. It is impossible to believe that it applies to the Additional collection itself, because, except for certain additions of details, A¹ was clearly copied from Vincent of Beauvais; and the collection must therefore have been made after 1244 or 1247, the date of the completion of the Speculum historiale.

It is probable that the compiler of the collection was a monk of Clairvaux. When he reaches no. 19 of Vincent of Beauvais, how a knight of Aquitaine refused to deny Mary, he suddenly interrupts the narrative to tell an unique legend of which he is reminded by the similarity of the theme. The tale which is interpolated in this fashion, no. 20, is the story of two boys from the vicinity of Clairvaux who went to Paris to study. One of

 [[ Print Edition Page No. 46 ]] 

	Vincent of Beauvais	SG		X		S.maris		R		A	Gobius	Vendome
81a.	Columns Raised	—		7		13		7		1	1	—
81b.	Toledo	23		36		9		16		2	2	—
82.	Jew Lends to Christian	72		29		19		17		—	3	2
83.	Orleans	30		68		26		20		3	4	—
84.	Milk: Tongue and Lips Restored	39		25		1		24		4	6	5
85.	Childbirth in the Sea	—		27		—		25		5	9	3
86.	Abbess	80		38		2		26		6	11	6
87.	Bridegroom: Ring on Finger	29		67		8		—		7	13	7
88.	Pilgrim in the Sea	70		40		10		32		8	10	—
89a.	Light on the Masthead	71		41		—		32a		9	—	—
89b.	Thread in Lip	52		37		—		—		—	7	—
90-92.	Chaste Empress	79		—		14		34		—	—	—
93-95.	Incest	—		—		20		—		10	—	—
96.	Rich Man and Poor Widow	—		—		—		—		11	15	19
97.	Bonus	—		—		—		36		12	19	9
98.	Poor Man Strikes Stone	75		64		—		37		—	—	—
99a.	Jewess in Childbirth	—		—		37		—		—	27	—
99b.	Bread	28		58		49		38		13	—	—
100.	Wife and Mistress	69		61		—		—		14	12	—
101.	O intemerata: Devil as Servant	34		69		—		—		15	30	25
102-103.	Girl Named Mary	—		—		—		—		16	14	—
104a.	Painter	—		—		54		44		17	33	—
104b.	Blasphemer of Lausanne	—		—		—		—		18	34	—
105-106.	Unwilling to Deny Mary	—		—		—		45		19	31	10
107.	Monks at Their Field Work	—		—		—		46		21	20	—
108.	Electuary	—		—		—		—		22	—	—
109.	Cistercian Monk Persecuted	—		—		—		48		23	16	—
110a.	Blasphemers of Brabant	—		—		—		47		24	35	—
110b.	Unable to Injure Mary Image	—		—		—		—		—	—	—
111.	Jew of London	—		—		—		—		25	49	—
112.	Demons as Swine	—		—		—		62		26	32	—	[[ Print Edition Page No. 47 ]]
113a.	Mother of Mercy	5		17		16		—		—	—	—
113b.	St. Dunstan	67-68		60		—		35		27	—	—
113c.	Priest of One Mass	10		13		25		—		28	—	43
114.	Vision of St. Hugh of Cluny	—		—		—		—		—	—	—
115.	Boy Devoted to the Devil	62		21		52		—		29	46	—
116a.	Ebbo	7		43		61		—		—	54	48
116b.	Five Psalms	—		—		—		53		30	21	34
117.	Woman Revived for Confession	—		—		12		—		31	—	—
118.	Little Devil in Church	—		—		42		—		—	22	—
119a.	Nativity	48		6		15		5		32	—	—
119b.	Saracen and Mary Image	31		68a		7		14		—	—	—
119c.	Mary Image Insulted	27		4		18		15		—	—	—

 [[ Print Edition Page No. 48 ]] 

them fell in love and was commanded by the devil to deny Christ and the Virgin Mary in order to win the girl. His companion prevented him from doing so, and they both entered Clairvaux.224

The Mariale magnum is given by the compiler as the source of the entire collection of MS Additional 15723. In the face of the clear relationship to Vincent of Beauvais, it must be concluded that he used two collections: (1) the Speculum historiale for the legends of A¹ and (2) another Cistercian collection for A². Three of the legends of A², aside from the compiler’s word for it, show some relation to the Mariale magnum. Gobius attributes the last, no. 43, to the Mariale magnum.225 A series of two, nos. 34-35, are distinctly similar to nos. 51 and 34 of the second series of MS Paris 18134, which seems to have been derived from the Mariale magnum.226 Certainly the tales are similar in form and character to the Cistercian legends of A¹ and could well have belonged to that collection. They are similar, too, to that Cistercian legend which the compiler of the Speculum exemplorum said he had from the Mariale,227 probably the Ur-Mariale, which was the source of the Mariale magnum. There is, therefore, no reason to doubt the compiler’s veracity except that he clearly used Vincent of Beauvais’ versions of some of the legends. Why, if he had the complete Mariale magnum available, did he choose to use a second-hand source? The answer is that it saved him labor and parchment. Vincent of Beauvais had already reduced the bulky legends to a briefer form without changing the order or detracting from the narrative. He chose the shorter versions of Vincent of Beauvais, except for some details omitted by the encyclopedist, and then added the others at length from the Mariale magnum itself.

If the collection of MS Additional 15723 was taken from the Mariale magnum, and the evidence permits us to assume that it was, then the dates of the compilation of the Mariale magnum may be narrowed to the period between 1200 and 1247, dates which are not incongruous with what is known about other collections of Mary legends in northern France related to the Mariale magnum.

The Vendome Collection

Certain characteristics of another Cistercian collection of Mary legends now in the Public Library of Vendome228 lend support to the picture of
 [[ Print Edition Page No. 49 ]] 
the Mariale magnum just outlined and permit a further conjecture about the work. The Vendome collection, comprising sixty-five legends as it stands,229 was made at the Cistercian monastery of Vaux-Cernay, founded in 1128 on the borderline between the dioceses of Paris and Chartres. Believing it to be the work of Thibaud de Marley, abbot of Vaux-Cernay, Bouchet dates it between 1235 and 1247,230 the years of his incumbency.

Most of the introductory matter of the Vendome collection is almost identical with that of Vincent of Beauvais. It begins with excerpts from the same portions of the visions of St. Elizabeth of Schönau. Although the diction is very similar, the abbot was clearly not copying from Vincent of Beauvais, for he mentions details from the work which the latter omits. Nor could Vincent of Beauvais have been taking his excerpts from the Vendome collection for the same reason. They were using a common source which presented that portion of the work in full, presumably the Mariale magnum or the Ur-Mariale. Between the excerpts from St. Elizabeth, the compiler of Vaux-Cernay inserts a brief paragraph from the De transitu of Melito of Sardis quoted by Vincent of Beauvais at greater length. There are also two paragraphs about the girlhood of Mary from a work which John of Garland uses, but which Vincent of Beauvais does not employ in this connection, the Pseudo-Matthei Evangelium.231

The abbot of Vaux-Cernay, for the most part, relates Cistercian legends and anecdotes which are not frequently in other collections in northern France. Here and there are some popular ones very greatly abbreviated. Although he tells some familiar tales not among the first series of Vincent of Beauvais, the character of the narratives as well as the order in which they are told points to knowledge of the Mariale magnum or a collection related to it:

	Vendome	V1
Jew Lends to Christian	2	ch. 82
Childbirth in the Sea	3	85
Milk: Tongue and Lips Restored	5	84
Abbess	6	86
Bridegroom: Ring on Finger	7	87
Bonus	9	97
Unwilling to Deny Mary	10	105-106
Rich Man and Poor Widow	19	96
O intemerata: Devil as Servant	25	101

The Cistercian legends of the Vendome collection may be divided into two classes: (1) those which were related to the abbot of Vaux-Cernay
 [[ Print Edition Page No. 50 ]] 
⁁C0⁁and were written down by him232 and (2) those which he found among the records of Cistercian monasteries or in collections of Mary legends. Among the first group of five legends is one told to him by the abbot of Clairvaux.233 In the second class are several unique tales originating at, or in the vicinity of, Clairvaux,234 and a series of legends which are similar to a sequence in the second series (A²) of MS Additional 15723.235 It was possibly on a visit to Clairvaux, perhaps a chapter meeting, that the abbot of Vaux-Cernay heard the story which he says the abbot of Clairvaux told him. While he was there, it may be imagined, he took the opportunity to do some research in the library and among the monastery records. If he was actually Thibaud de Marley (abbot 1235-1247), he probably did not find the collection of the Additional manuscript there, for part of that was certainly copied from Vincent of Beauvais, whose work was not completed until 1244 or 1247. He might very well have seen there that other source used by the compiler of MS Additional 15723, which he said was the Mariale magnum. The great volume probably lay on the shelves of the library of the monastery of Clairvaux in the second quarter of the thirteenth century along with the earlier Ur-Mariale which will now be referred to as the Clairvaux Mariale. The likelihood is that they were both compiled there where St. Bernard left behind him a strong tradition of devotion to the Virgin Mary. The collection and the recording of Mary
 [[ Print Edition Page No. 51 ]] 
legends was probably something of an official project, pursued chiefly at chapter meetings when numbers of Cistercian abbots came together.236 This would account for the authority as well as the availability of the Mariale magnum.

Collections derived from the Mariale magnum had their descendants. The pseudo-Celestine collection,237 comprising twenty-seven legends, briefly told, depends upon the work of the abbot of Vaux-Cernay. It is included among the works printed by Telera in 1640 as those of Pope Celestine V (c. 1215-1296), though they are certainly not written by him. A clue as to the origin of the collection is given in the first legend, the only unique one. This tale, together with the attribution to an Italian, suggests Italy rather than France as the place where the collection was compiled. It tells how a priest, accused of writing a letter attacking the emperor, had his hand cut off and exposed before his own church dedicated to the Virgin Mary. She restored the member, and the emperor came on foot and sought the priest’s pardon.238

The legends follow in the same order as in the Vendome collection, although many have been omitted. The diction employed in telling them is not the same, and there are very minor variations in the details. None of the legends told on the authority of the abbot of Vaux-Cernay is included. The additional legends at the end of the pseudo-Celestine collection only serve as a reminder that the Vendome collection, as we have it to-day, is a fragment. A comparison of these two collections and that of MS Additional 15723 follows:

Ps-Celestine	Vendome	A
1	—	—
2	4	—
3	5	4
4	7	7
5	9	12
6	14	—
7	16	—
8	17	—
9	21	—
10	22	—
11	23	—
12	26	34
13	27	35
14	34	30
15	39	—
16	40	—
17	41	—
18	48	—
19	56	—
20	63	—
21	—	—
22	—	41
23	—	—
24	—	—
25	—	—
26239	—	21
		22
27	—	—

7. Relation of Caesar of Heisterbach and Pseudo-Caesarius to the ‘Mariale’ Collections

The ‘Mariale’ collections originated in the Cistercian monasteries of northern France. Collections were made in the convents of the same order in other regions, and because the nature of Cistercian monastic organization promoted frequent contacts between abbots as well as the exchange of books, some similarities among collections of various regions would naturally result. The seventh book of the Dialogus miraculorum of Caesar of

The ‘Mariale’ Collections

Heisterbach together with several Mary legends in the Libri VIII miraculorum and the anonymous collection attributed to him by Meister are Cistercian collections made in the region of Cologne.254 The first two were completed during the years 1223-1227, and the third, it would seem, not a great deal later.255

Neither the collection of Caesar of Heisterbach nor pseudo-Caesarius clearly belong to the series of collections which have been discussed in relation to the Stella maris, for they do not include the tales typical either of SV or the Quoniam series. They are, however, related to the ‘Mariale’ collections of Clairvaux. In the Dialogus Caesar of Heisterbach tells three or four256 of the legends that characterize the ‘Mariale’ collections; and in the Libri VIII miraculorum, a single one.257 The best clue we have to explain his acquaintance with them is a statement prefacing quite another tale about the Virgin, ‘We read in the Book of the Miracles of Clairvaux.’258 Could it have been the Clairvaux Mariale or possibly the Mariale magnum that he knew? It is difficult to give an affirmative answer. No one of his legends is exactly the same version as the corresponding ‘Mariale’ legends, even when allowances are made for differences in diction; and of the one, ‘Girl Named Mary,’ which is most similar in detail Caesar of Heisterbach says, ‘I heard it recently from a certain abbot of our order,’259 indicating
 [[ Print Edition Page No. 58 ]] 
an oral source. It seems more likely, therefore, that the collection of Clairvaux in which he read was an earlier collection than either of these. It may not have been limited to Mary legends and was, perhaps, the common source of some of the Cistercian legends in the Clairvaux compilations of Mary legends, Caesar of Heisterbach, and the Exordium magnum ordinis Cisterciensis of Conrad of Eberbach, a contemporary of Caesar of Heisterbach.

As for pseudo-Caesarius, it cannot be said with certainty that he did not know the Clairvaux Mariale or the Mariale magnum, although it seems very doubtful. He includes six characteristic ‘Mariale’ legends, aside from those which seem to come from Caesar of Heisterbach himself.260 One of them, no. 91, appears again in Gobius, Scala celi, where it is clearly attributed to the Mariale magnum. The ‘Painter’ is very much abbreviated. The other four do appear in MS Paris 18134, although the details differ radically from those of pseudo-Caesarius.261

If, on the other hand, pseudo-Caesarius compiled his collection at Heisterbach, there must have been available to him as to Caesar of Heisterbach,
 [[ Print Edition Page No. 59 ]] 
the presumably older Book of the Miracles of Clairvaux, a probable source of some of the Cistercian legends of the‘ Mariale’ collections made at Clairvaux.262 His acquaintance with them could then be attributed to a common source, as in the case of Caesar of Heisterbach.263 Whatever the relationship, the absence of the series of Mary legends typical of northern France in these two Cistercian collections from the region of Cologne, made in the years when the Mariale magnum was taking shape, serves to emphasize the separate identity in the first half of the thirteenth century of the ‘Mariale’ collections made in northern France. Caesar of Heisterbach’s reference to the Book of the Miracles of Clairvaux also confirms our thesis that the monastery of Clairvaux was an important center for the collection of Mary legends in the same period.

The Mariale collections, it is clear, were descended from the large compilations of northern France, specifically those with a prologue beginning Quoniam gloriosissima virgo virginum. Two representatives of this group existed in the thirteenth century in the monasteries of Paris, one at St. Germain-des-Prés and another at Ste. Geneviève, where John of Garland used it. A third, MS Paris 17491, cannot be traced to a particular source. One of these, or another very like it, must have been the link between the Parisian collections and the Mariale family, the source of the non-Cistercian legends of the Clairvaux Mariale. That collection, the evidence of the Stella maris suggests, was John of Garland’s source, the collection of Ste. Geneviève.

8. The Collection of Ste. Geneviève and the Stella Maris

John of Garland says in a note appended to the Stella maris that he versified the legends he tells from a collection which he found in the book-press of the monastery of Ste. Geneviève.264 Unlike those of St. Victor and St. Germain-des-Prés, the library of Ste. Geneviève was dispersed in the sixteenth century, so that not a single manuscript or printed book remained
 [[ Print Edition Page No. 60 ]] 
when the monastery was reformed by the Cardinal de la Rochefoucault in 1624.265 It is perhaps for this reason that no collection among those examined can be identified as that which John of Garland had before him. What has proved to be an important collection in the history of Mary legends in northern France can be studied only through the verses of the schoolmaster’s Stella maris. That it was a much more extensive collection than his may be inferred from his reference to his own tales as ‘a few.’266

The original of John of Garland must have been a summa of Mary legends current in northern France in the first half of the thirteenth century. Among the ‘few’ tales of the Stella maris are characteristic representatives of all the collections that have thus far been analyzed. The collection of Ste. Geneviève began with one or two prologues. The Quoniam gloriosissima virgo prologue, which characterizes the legends of St. Germain-des-Prés, MS Paris 17491, and the Rouen Mariale was certainly there. It is chiefly a theological discussion of the Virgin Mary as the immaculate mother of God and queen of heaven, ‘the one hope of mankind after God.’267 This had been the theme of John of Garland’s Epithalamium beate virginis,268 and it is the involved symbolism of the Epithalamium that he uses rather than the words of the Quoniam prologue. The author of the long gloss of the British Museum manuscript of the Stella maris treats the theme more fully in the manner of the Quoniam prologue.269

It is possible that the manuscript of Ste. Geneviève had also the Mirande virginis laudes prologue which appears before Quoniam gloriosissima in the Rouen Mariale, for John of Garland tells one of the legends which distinguishes the only other collection in Latin in northern France which used it, MS Bibliothèque Nationale 5268 (12th century), from all other collections investigated.270 Following the prologue, or prologues, the collection of Ste. Geneviève related the life of the Virgin as told in the Pseudo-Matthei Evangelium, excerpts from which are inserted at the beginning of the Vendome collection. John of Garland merely suggests the incidents, to which the Bruges gloss of the Stella maris adds more detail.271 It may
 [[ Print Edition Page No. 61 ]] 
have included also the De transitu of Melito of Sardis, for that treatise, or a part of it, appears regularly in other collections related to it.272

The legends of the Stella maris were, of course, derived from a single source, the manuscript of Ste. Geneviève, but in order to clarify the relation of their source to other collections of northern France, they have been divided into four groups: those legends which the Stella maris has in common with (I) the first series of St. Victor (SV¹), the first series of St. Germain-des-Prés (SG¹), and MS Paris 17491 (X); (II) the Rouen Mariale; (III) Vincent of Beauvais; and (IV) the second series of MS Paris 18134 (Q²).

I. The nucleus around which the collection of Ste. Geneviève, as well as many other Latin collections of northern France, was built is clearly the first series of St. Victor. Of the sixty-one legends of the Stella maris, thirty-seven273 are the same legends in the same versions as those of the twelfth-century SV¹. The only miracles of SV¹ conspicuously absent from the Stella maris are those which are versified.274 One of these, ‘Incest,’ John of Garland tells, but his version is not the same as that of SV¹.

The relationship of the collection of Ste. Geneviève to the first series of St. Germain-des-Prés, a Quoniam collection, is even closer. The Stella maris and SG¹ have in common, not only thirty-four of the thirty-seven legends of SV¹,275 but also nine more narratives. Forty-three of the sixty-one legends which John of Garland versified at Ste. Geneviève might have come, therefore, from the collection of St. Germain-des-Prés. One of the additional tales, ‘Mare,’ is significant, because it occurs, among the collections examined, only in those with the Quoniam prologue and one other, defective in the beginning, MS Charleville 168.276 The nine legends are arranged in the Stella maris in two well-marked series, nos. 6-8, 15, 26, 29-32, in much the same sequence as in SG¹, where they are numbers 29-33, 46-49 (with the exception of 47), 54 and 58.

It is more likely, however, that the collection from which the legends of Ste. Geneviève emanated was not SG¹, but a work which stands in closer relationship to the first series of MS Paris 17491. There are in the Stella
 [[ Print Edition Page No. 62 ]] 
maris and X¹, not only the thirty-seven tales of SV¹ and the nine additional narratives of SG¹, but also four more legends, not in SV¹ or SG¹. Fifty of the sixty-one legends of the Stella maris, then, may be accounted for with reference to MS Paris 17491 (X¹). The character of the four narratives of John of Garland’s collection, not in SV¹ or SG¹, indicates an intimate relationship between the collection of Ste. Geneviève and the two other Quoniam collections, MS Paris 17491 and the Rouen Mariale. Two of the tales have been found together in no other collections investigated:277‘Judas in Hell,’ told by the Cistercian monk of Froidmont, and ‘Christ Appears to Monk’ from the Cistercian monastery of Savigny. The two others, ‘Jewish Boy’ and ‘Columns Raised,’ are part of the series from Gregory of Tours reproduced in MS Paris 17491 and the Rouen Mariale.278

Comparison of the legends of the Stella maris with the second series (X²) of MS Paris 17491, however, makes it certain that the collection of Ste. Geneviève was not copied directly from MS Paris 17491. John of Garland did not use the versified redactions of X². He tells three of the legends, but his versions are those of the Rouen Mariale and Vincent of Beauvais in which they are related in prose.279 The collection of Ste. Geneviève, the Rouen Mariale, and the Mariale magnum, therefore, probably reach back into a tradition earlier than MS Paris 17491 or even the collection of St. Victor, where some of the legends appear in verse. The manuscript which John of Garland used at Ste. Geneviève could, and probably did, ante-date the present MS Paris 17491, and the first series of it, X¹, was older in its original form than the Ste. Geneviève collection.280

II. Analysis of the legends of the Stella maris to this point has disclosed
 [[ Print Edition Page No. 63 ]] 
incidentally certain important similarities between the Rouen Mariale and the manuscript of Ste. Geneviève, John of Garland’s exemplar. They shared the Quoniam prologue and one legend, ‘Mare,’ characteristic only of the collections it prefaced. The two Cistercian legends, furthermore, have been the means of distinguishing certain of the Quoniam collections from others. The work of St. Germain-des-Prés lacks both of them; MS Charleville 168 has only the first (‘Judas’); three collections only include them both: MS Paris 17491, the collection of Ste. Geneviève, and the Rouen Mariale. The collection of Ste. Geneviève has thirty legends altogether in common with the Rouen Mariale; and the larger collection, MS Paris 17491, relates forty of the same legends as are narrated in the Rouen compilation. Obviously the three collections are an intimately related group. The first series (X¹) of MS Paris 17491 was probably one of the sources of the collection of Ste. Geneviève. Where shall the Rouen Mariale be placed with reference to the other two?

The first two collections, X¹ and the collection of Ste. Geneviève, seem to be earlier in origin than the Rouen Mariale. Of the legends that can be dated, the most recent occur in the Rouen Mariale.281 The Rouen Mariale, then, probably descended from one of the two earlier collections. Comparison of the three collections suggests that the first series (R¹) of the Rouen Mariale might have been drawn either from MS Paris 17491 (X¹) or from the collection of Ste. Geneviève, with the probability in favor of Ste. Geneviève. Only three of the legends of R¹, except ‘Musa,’ lack in MS Paris 17491 (X¹): (1) chapter 18 of the In gloria martyrum of Gregory of Tours (no. 8); (2) ‘Chaste Empress’ (no. 34); and (3) ‘Bonus’ (no. 36). The two latter, however, are important in the differentiation of the Quoniam collections. They belong to the versified series of X². With the exceptions of MS Paris 17491 and MS Charleville 168, the Quoniam collections, including John of Garland and the Rouen Mariale, use the prose versions of most of the legends.282 Although John of Garland does
 [[ Print Edition Page No. 64 ]] 
not relate the first of the legends which differentiate X¹ from the Rouen Mariale, he does include chapter 21 of the same work of Gregory of Tours, ‘Christ Image Wounded.’283 The collection of Ste. Geneviève, therefore, may have included both chapters 18 and 21.284 The Stella maris lacks sixteen of the legends of the first series of the Rouen Mariale, but fifteen of them are the common tales of the Quoniam collections which ought to have been in the collection of Ste. Geneviève. The other is ‘Bonus,’ one of the series lacking also in MS Paris 17491.

The legends which clearly distinguish the collection of Ste. Geneviève from MS Paris 17491 (X¹) in its relation to the Rouen Mariale occur in the second series (R²) of the Rouen Mariale. It is, for the most part, made up of legends which are lacking in both the Stella maris and X¹, some of them local or unique. Only two legends commonly in other Quoniam collections are among them.285 The Stella maris includes one of these and MS Paris 17491 (X¹), both of them. The Stella maris, on the other hand, includes two additional legends of the second series, R², not to be found in any other of the Quoniam collections, nor in SV:286 the ‘Painter of Flanders’ and ‘Beirut.’287 The latter is not a Mary legend, either as told by John of Garland or by the compiler of the Rouen collection, and its presence is exceptional in collections of Mary legends. In other words, the correspondence between MS Paris 17491 (X¹) and the Rouen Mariale ends at no. 43 (R¹), while the similarity of the Rouen collection to the manuscript of Ste. Geneviève continues through the entire work. The lost collection of Ste. Geneviève was, therefore, nearer to the source of the Rouen Mariale than is MS Paris 17491 (X¹), a conclusion which is substantiated by a study of the legends of the Stella maris which are repeated in the collection of Vincent of Beauvais.

III. Vincent of Beauvais and John of Garland both extracted legends from collections using the Quoniam gloriosissima virgo prologue.288 Both
 [[ Print Edition Page No. 65 ]] 
their sources, the collection of Ste. Geneviève and the Mariale magnum, incorporated ‘Mariale’ materials of the sort described in the Mirande prologue of the Rouen Mariale, writings of ‘the fathers ancient and recent.’289 John of Garland, moreover, gathered from the manuscript of Ste. Geneviève fourteen of the thirty-three legends which Vincent of Beauvais, V¹, took from the Mariale magnum, and eight of the ten, V², which he took from the Clairvaux Mariale.290 Seventeen of the twenty-two (14+8) are the tales of the Quoniam collections, St. Germain-des-Prés and MS Paris 17491 (X¹). The other five have come to light in the same versions only in Vincent of Beauvais, or in collections known to be related to his or to the Mariale magnum:

Vincent of Beauvais (V²) selected, then, from the Clairvaux Mariale eight of the same legends which John of Garland took from the collection of Ste. Geneviève, two of them unusual legends, missing from other Quoniam collections. From the Mariale magnum he took fourteen of the legends of Ste. Geneviève, three of them distinctive. Nine of the fourteen legends common to the collection of Ste. Geneviève and the Mariale magnum, including one distinctive narrative (no. 44), are also in the Rouen Mariale. Presumably these nine legends, at least, arrived in the Rouen Mariale by way of its source, the Clairvaux Mariale. Seventeen of the narratives of Ste. Geneviève, then, can be certainly ascribed to the Clairvaux Mariale (9 in V¹ + 8 in V²), among them three significant ones. The number must actually have been much larger for two reasons: (1) John of Garland merely selected legends from a larger collection at Ste. Geneviève, and Vincent of Beauvais reproduced in summary form only a portion of the Mariale magnum. (2) The compiler of the Mariale magnum, the source of Vincent of Beauvais, had already eliminated many of the common
 [[ Print Edition Page No. 66 ]] 
Quoniam legends. These selective processes tended to reduce the number of legends common to Vincent of Beauvais and the Stella maris, upon whose fragments of their originals our evidence must depend. Even if there were no other confirmation, seventeen legends, therefore, seems a sufficiently large number upon which to base the conclusion that the collection of Ste. Geneviève was the source of the non-Cistercian legends of the Clairvaux Mariale. From the Clairvaux Mariale, the Rouen collection had thirty legends of Ste. Geneviève; and the Mariale magnum, the fourteen non-Cistercian legends related by Vincent of Beauvais (V¹). The eight legends of V², the Dominican compiler derived directly from that source.

The order of the narratives of Vincent of Beauvais, the Rouen Mariale, and the Stella maris supports the same conclusion. The Mariale magnum and the Rouen Mariale, it would seem, inherited it from the collection of Ste. Geneviève by way of their common source, the Clairvaux Mariale, and passed it on to Vincent of Beauvais.291 The similarity in sequence extends only through V¹, but V² is lacking the careful notation, ex Mariali magno . . . Explicit de Mariali magno. Item alia, which indicates that the compiler was using all the legends between those notes in their original order. In other words, the series V¹ is a portion of the Mariale magnum; the legends of V² are merely tales selected in no particular order from the Clairvaux Mariale.

The character of the Stella maris as a whole is further evidence of the relation of its source, the collection of Ste. Geneviève, to the Mariale collections. Besides the incidents from the Pseudo-Matthei Evangelium, a large amount of matter appropriate to a ‘Mariale’ is scattered throughout the Stella maris — medicine, physical geography, and astronomy ‘appropriated’ to the Virgin Mary,292 a description of the festivals of the Virgin,293 a sequence in her honor,294 and numerous other songs of praise.295 What John of Garland has composed, in fact, is a Mariale parvum, adapted not to the uses of monks in a monastery, as was his source, but to young students
 [[ Print Edition Page No. 67 ]] 
learning the seven liberal arts. As such it is a collection unique in the history of Mary legends.

IV. The hypothesis which makes the collection of Ste. Geneviève the source of the non-Cistercian legends of the Clairvaux Mariale explains the presence of the ‘Chorister’ (no. 43) in the Stella maris. Although it is missing from the Rouen Mariale and Vincent of Beauvais, it does appear in the second series, Q², of MS Paris 18134 and the Vendome collection which are both descendants of the Mariale magnum.296 Comparison of the texts of the ‘Chorister’ supports the conclusion that all three were descended from a common source.297

Three, or possibly only two, of the narratives of John of Garland’s Stella maris are unique. One, the story of the healing of the victims of ergotism at Notre-Dame de Paris,298 was obviously inserted to enable the author to pay a compliment to William of Auvergne, bishop of Paris. Another, the victory of the Parmese over Frederick II, was the latest piece of news to come to Paris from Italy.299 The third, a tale of a school-boy saved in a thunderstorm by the singing of Ave maris stella,300 may also be unique, though the theme is a common one. It could have been inserted by the author from memory, because he was preparing a lesson for school-boys.

Almost all the legends and much of the introductory matter, John of Garland took from his source, but these account for only two-thirds of the lines of the Stella maris. The author intended to use his collection, as he did many of the other things he wrote, in the schools. He tells us, therefore, that he has inserted lessons on various subjects, natural science, astronomy, and theology.301 These lessons are, none-the-less, subordinated to the purposes of a Mariale, or a work in praise of the Virgin Mary. Hippocrates is mentioned in support of the power of the Virgin to heal human disease.302 The author of the British Museum gloss says that it was John of Garland’s purpose to adapt the astrology of Martianus Capella to the Virgin Mary.303 The constellations, as John of Garland describes them,
 [[ Print Edition Page No. 68 ]] 
each in its own particular way, signify some aspect of the Virgin’s dominion or that of her Son.304 The constellation of Virgo, as described by Albumasar, for instance, pictures the Virgin Mary nursing her Child.305 The Stella maris demonstrates how even the secular learning of the schools was in the author’s classes made to serve the Mary cult.

The Stella maris is significant also in other respects, quite aside from its connection with the schools and the substance it gives to the personality of the valuable collection of Ste. Geneviève. John of Garland wrote, not as a monk, but as a layman interested in the art of writing poetry. So far as the evidence goes, he is the first lay compiler of Mary legends in northern France. For the first time he gives a collection of Mary legends the semblance of the literary form of the panegyric. His purpose is not that of the monk-compilers, to assemble a large number of Mary legends. The marvellous deeds of the Virgin are but examples of her power in heaven and earth, a power which the learning of the seven liberal arts, medicine, and theology confirms. The narratives, he merely suggests, and these are followed at intervals by songs in her praise.

In this respect the collection made by Alfonso el Sabio, King of Castile, 1252-1284, resembles the Stella maris. The Cantigas de Santa Maria306 was written in the second half of the thirteenth century in the Galician language. Of the four hundred and twenty-two canticles set to music, three hundred and fifty-three are Mary legends. The narratives are at times very brief, and songs in praise of the Virgin are interpolated at regular intervals. Although it could not be maintained that the author owes the particular form to the Stella maris, there are some important similarities between the two collections.

The legends of the Cantigas were gathered from many sources, most of them written sources.307 The author mentions a book of the miracles of Soissons,308 and he had also the Speculum historiale, sent him by St. Louis of France. One of the collections he used, he says had nearly three hundred legends in it.309 To add to these, he sent to shrines and had local compilations copied, so that he complains that he cannot write down half of those he knows.310 The legends fall into two groups, (1) those which are local, and (2) those widely-known in France and perhaps in Spain. Of the latter,
 [[ Print Edition Page No. 69 ]] 
Mussafia counts ninety-four.311 Although king Alfonso says that he used collections of many countries,312 all ninety-four are legends commonly found in the collections of northern France, especially the collection of Ste. Geneviève and the ‘Mariale’ collections.313 Of the sixty-one tales of the Stella maris, thirty-seven are included in the Cantigas, fifteen of them also in the Mariale magnum. Particularly significant is the treatment of the legends of Soissons. They are in the Cantigas, as in the Stella maris, scattered through the entire collection. In both the Cantigas and the Stella maris the festivals of the Virgin are named and described.314 In addition to the fifteen legends of the Mariale magnum which are common to the Cantigas and the Stella maris, there are twelve other ‘Mariale’ legends among the Cantigas. Clearly either the collection of Ste. Geneviève or one of the ‘Mariale’ collections, or both, was used in the composition of the Cantigas. Alfonso el Sabio is known to have visited Paris, where he could have seen both John of Garland and his source; and it is not likely that his agents overlooked Clairvaux in their search for Mary legends.

9. The Stella Maris in England

John of Garland’s collection seems to have been more popular in England than in France. To a Frenchman of the thirteenth century, they were but old familiar narratives; but to an Englishman of the same period many of them were new tales, while others were unfamiliar versions of anecdotes he already knew. One of the manuscripts of the Stella maris which remain, MS British Museum Royal 8 c iv, belonged in the middle ages to the monastery of Bury St. Edmunds. There is some reason to believe that it was copied there about 1300 by a monk whose name was Adolf, who supplied it with certain glosses in addition to those which his exemplar already possessed. A study of the text shows that two manuscripts were used in the construction of Adolf’s text, one from which he copied the text, and another which was used for corrections.315

A manuscript given to the library of St. Augustine’s Abbey at Canterbury by one of its monks, Thomas de Wyvelsberge, included both the Stella maris (or a part of it) and the Epithalamium beate Marie virginis of John of Garland.316 If he is the Thomas de Wyvelsberge mentioned in the
 [[ Print Edition Page No. 70 ]] 
register of the abbey in documents of 1242 and 1243,317 then his volume must have contained the earliest known copy of the Stella maris in England.

At the beginning of the fifteenth century John Boston, monk of Bury, listed a manuscript of the Stella maris among the books which he saw in the monastic libraries of England and Scotland. He uses the title mentioned in the colophon of the British Museum manuscript and quotes the incipit.318 It is possible that the manuscript he knew was the one which Adolf copied, the British Museum manuscript we know today; but, on the other hand, Boston does not give it the number (82) by which he usually designates a Bury book,319 nor does he mention other works in the same codex.320 The likelihood is, therefore, that it was not the Bury copy that he has in mind, but one belonging to another library.

At about the same time that Boston of Bury was noting the presence of the Stella maris in English monastic libraries, an anonymous Englishman was using it to produce a remarkable collection in English verse. A considerable fragment of his work is now in the British Museum, MS Additional 39996,321 in script of the first half of the fifteenth century. It includes eighteen legends, breaking off in the middle of the tale of ‘Julian the Apostate,’ number 24 of the Stella maris. There can be little doubt that the Englishman had before him the Stella maris of John of Garland, even though he stretches the narratives to many times the length of the original. The order in which he tells the tales is that of the Stella maris, except for slight alteration:

In many cases the Englishman’s tales are spun in alarmingly original fashion from the meager details of John of Garland. The materials with which he elaborates his narrative are not all gathered from thin air, however. He knows, in some instances, the complete versions of the stories;323 and in at least one case, he seems to have had another collection close at hand when he wrote.324 Only the first legend of MS Additional 39996, how a monk was tempted by the devil in the form of a woman and saved by his prayer to the Virgin has no connection with John of Garland.

In conclusion, therefore, it may be said that the collection of Ste. Geneviève, the source of John of Garland, was a large and significant compilation of materials about the Virgin Mary, chiefly Mary legends. It included not only those tales which had been gathered together in the monastery of St. Germain-des-Prés and by the unknown compiler of the first series of MS Paris 17491 (X¹), but also certain significant legends used in the construction of the Clairvaux Mariale. The Cistercian monk who compiled it had little to do but interpolate groups of Cistercian legends here and there, and his collection was complete. John of Garland, on the other hand, adapted the collection of Ste. Geneviève for use in his class-room. He wrote some verses merely suggesting the content of selected legends, and then he built his own ‘Mariale’ by the addition of matter from the arts’ curriculum, intended to show how all the seven liberal arts unite in worship of the Virgin Mary. His little work also gave new form to collections
 [[ Print Edition Page No. 72 ]] 
of Mary legends. The entire Stella maris is a panegyric in praise of the Virgin. The legends are merely cited as examples of her prowess over man and nature. Whether or not the Spanish king, Alfonso el Sabio, saw the Stella maris at Paris, he composed a work which used Mary legends in much the same way, not as pieces to be read in monasteries or told by preachers to point a moral, but as part of a literary composition glorifying the Virgin. Before the end of the thirteenth century John of Garland’s work was taken to England, perhaps by the author himself or by one of his students, and there became the source of a collection in English verse.

The table on pages 74-75 shows the relationship between the Stella maris and other collections.

The last quarter of the twelfth century and the first half of the thirteenth century was, as this study shows, a period of importance in the development of the Mary cult, one of the chief features of which was the collection and popularization of legends about the Virgin Mary. The collection of anecdotes of universal rather than merely local interest had got well under way by the beginning of the twelfth century with the compilation of HM, the Elements-series, and TS. They included not more than seventeen legends. Possibly the number had mystical significance in connection with the Mary cult, or it may have been determined only by the requirements of the services and other uses to which the collections were put.

Just when or where the Pez collection of forty-three legends came into existence is not certain, but it could not have been much earlier than the SV¹ collections, nor far distant from their home. To Pez, the compiler of SV¹, probably in the third quarter of the twelfth century, added another series of seventeen or eighteen tales, gathered chiefly from the ecclesiastical writers of northern France and from monastic and communal records in the same region. The number of copies of SV¹ and the collections descended from it at Paris in the thirteenth century325 make it probable that it was compiled in a Parisian monastery.

The compiler of the first collection with the Quoniam gloriosissima virgo virginum prologue was probably a monk of St. Germain-des-Prés, who did his work in the last quarter of the twelfth century. To SV¹, or a collection like it, he added another group of seventeen tales, nearly all of which originate in the region of northern France of which Paris is the center. Other monastic compilers used his to heap up even larger collections to the glory of the Virgin and to give greater interest and variety to the readings in the monasteries.

It was still another Paris collection, that of the monastery of Ste. Geneviève, which became the foundation upon which the Cistercians built the Clairvaux Mariale by the addition of legends gathered chiefly from their own monasteries. Obviously there was a limit to which the mere accumulation of materials of this sort could go, if they were to be used conveniently. The collection of St. Victor was already a large folio volume, and it included nothing but legends and two other short treatises about the Virgin. A collection of three hundred legends, such as Alfonso X mentions, if they were reproduced in full in an easily legible script, must have been an unmanageable volume. Some drastic sort of selection must take place when new tales were added, or the legends must be revised to make them briefer. Both these processes went on in the Cistercian monasteries. The Mariale magnum used only selected legends from the older collections, and the abbot of Vendome reduced the traditional tales to the briefest form possible. Even the Cistercian anecdotes were pruned of circumstantial details. The compiler of MS Additional 15723 to save space used the briefer versions of Vincent of Beauvais, but copied in full other legends of the Mariale magnum. It remained for the preachers of the Dominican and Franciscan orders who took the legends out of the monasteries and into the market squares of the towns to eliminate all the matter which encumbered the narrative.

In the earliest compilations, except in the Elements-series which has a unity of its own, little care had been given to the form of the collection itself. There was some attempt in Pez and SV¹ to distribute new legends among the old, probably to give some degree of novelty to each group as it was used in the monasteries. The compiler of the Quoniam collection of St. Germain-des-Prés used subject matter to some extent as a principle of organization in his first series (SG¹), and the social status of the individuals concerned in the second (SG²). The matter, other than Mary legends, was apparently incorporated into most of the collections of the twelfth and early thirteenth centuries without much thought as to its position in the collection. In the collections of Ste. Geneviève and the Mariale magnum, these materials were placed at, or near, the beginning of the collection and arranged with reference to the life of the Virgin: (1) accounts of her parents and her earthly life from the Pseudo-Matthei Evangelium, (2) her assumption as in Melito of Sardis and Elizabeth of Schönau, and finally (3) the miraculous deeds she performed after her assumption. Preachers varied the order of the tales to suit their need to illustrate a moral truth, grouping those from which a given lesson could be extracted under a single heading.

The collection of Mary legends was in northern France by the middle of the thirteenth century no longer confined to monks or cathedral clergy. John of Garland is the first known lay compiler of Mary legends in that

	Stella maris	MS Paris 14463	MS Paris 12593		MS Paris 17491		Rouen Mariale		Vincent of Beauvais	Vendome		MS Paris 18134	Gobius	Alfonso el Sabio	MS Additional 39996
1.	Milk: Tongue and Lips	14	39		25		24		84	5		29	6	—	6
2.	Abbess	55	80		38		26		86	6		39	11	7	—
3.	Jewish Boy	—	—		9		10		—	—		—	—	—	7
4.	Son Restored	47	66		28		31		—	—		—	—	21	8
5.	Devil in Beasts’ Shapes	37	64		26		28		—	—		30	—	47	9
6.	Barns Filled	—	58		10		11		—	—		—	—	187	10
7.	Saracen and Mary Image	—	31		68a		14		119b	—		47	—	46	—
8.	Ring on Finger	—	29		67		—		87	7		43	13	42	11
9.	Toledo	41	23		36		16		81b	—		—	2	12	—
10.	Pilgrim in the Sea	49	70		40		32		88	—		—	—	33	—
11.	Mead	9	21		45		—		—	—		—	—	23	12
12.	Woman Revived	—	—		—		—		117	—		—	—	—	2
13.	Columns Raised	—	—		7		7		81a	—		—	1	231	3
14.	Chaste Empress	45	79		—		34		90-92	—		22	—	5	4
15.	Nativity	—	48		6		5		119a	—		—	—	—	13
16.	Mother of Mercy	32	5		17		—		113a	—		—	—	—	—
17.	Libia	42	37		3		12		—	—		—	—	27	14
18.	Mary Image Insulted	26	27		4		15		119c	—		—	36	34	16
19.	Jew Lends to Christian	53	72		29		17		82	2		—	3	25	17
20.	Incest	—	—		—		—		93-95	—		—	—	17	15
21.	Beirut	—	—		—		63		—	—		—	—	—	—
22.	Clerk of Chartres	3	3		15		—		—	—		4	18	24	18
23.	Hours Sung Daily	54	73		59		39		—	—		—	—	—	—
24.	Julian the Apostate	70	102		2		—		—	—		—	—	15	19
25.	Priest of One Mass	39	10		13		—		113c	43		23	—	32	—
26.	Orleans	—	30		68		20		83	—		—	4	51	—
27.	Judas	—	—		65		1		—	—		—	—	—	—	[[ Print Edition Page No. 75 ]]
28.	Mal des Ardents at Paris	—	—		—		—		—	—		—	—	—	—
29.	Mare	—	33		68c		18		—	—		—	—	—	—
30.	Purification	—	49		5		4		—	—		—	—	—	—
31.	Sight Restored	—	54		12		23		—	—		—	—	—	—
32.	Chartres	—	46		23		22		—	—		—	—	—	—
33.	Maid of Arras	58	74		63		—		—	—		—	—	105	—
34.	Conception	10	20		66		—		—	—		—	—	36	—
35.	Transported	13	61		18		30		—	16		8	—	—	—
36.	Drowned Sacristan	2	2		14		—		—	—		3	—	11	—
37.	Jewess in Childbirth	—	—		—		—		99a	—		—	27	89	—
38.	Hildefonsus	1	1		—		—		—	—		2	—	2	—
39.	Mouth of Hell	15	40		19		—		—	—		9	—	58	—
40.	Christ Appears to Monk	—	—		65a		2		—	—		—	—	—	—
41.	Fire at Mont-St.-Michel	21	16		34		19		—	—		11	—	39	—
42.	Little Devil in Church	—	—		—		—		118	—		—	22	—	—
43.	Chorister	—	—		—		—		—	41		31	—	6	—
44.	Soissons	57a	—		56a		—		—	—		—	—	61	—
45.	Nose Restored	57b	—		56b		—		—	—		—	—	81	—
46.	Two Brothers at Rome	23	11		20		—		—	—		—	—	—	—
47.	Landmark Removed	16	12		53		—		—	—		—	—	—	—
48.	Charitable Almsman	5	6		49		55		—	—		6	—	—	—
49.	Bread	44	28		58		38		99b	—		21	—	139	—
50.	Pilgrim of St. James	18	9		35		—		—	—		—	—	26	—
51.	Foot Cut Off	65	24		46		—		—	—		—	—	37	—
52.	Boy Devoted to Devil	34	62		21		—		115	—		14	46	115	—
53.	Christ Image Wounded	—	—		—		—		—	—		—	37	—	—
54.	Painter	—	—		—		44		104a	—		—	33	74	—
55.	Will for Deed	40	42		54		—		—	17		20	—	45	—
56.	Ave maris stella	—	—		—		—		—	—		—	—	—	—
57.	Freed from Captivity	57c	—		56c		—		—	—		—	—	62	—
58.	Saturday	25	55		56		33		—	—		41	—	—	—
59.	Parma	—	—		—		—		—	—		—	—	—	—
60.	Theophilus	8	50		71		—		—	—		—	—	3	—
61.	Ebbo	6	7		43		—		116a	48		46	54	—	—

region. He, before the middle of the century, and Alfonso el Sabio in the second half gave their collections something resembling literary form. Their verses do little more than to suggest the familiar narratives, and songs of praise are introduced at intervals between groups of legends. Once the collection of Mary legends has emerged from the monasteries into the hands of the mendicant preachers, vernacular writers and laymen, not only the legends themselves, but the collections into which they were gathered take on such variety of form that relationships between them are difficult to establish.

Two centers in northern France emerge in the last half of the twelfth and the first half of the thirteenth century as scenes of important activity in the gathering and dissemination of Mary legends in the Latin language: the monasteries of Paris and the Cistercian monastery of Clairvaux. Of the collections made at Paris, that of the monastery of Ste. Geneviève was probably the largest and the most significant; and of those made at Clairvaux, the Mariale magnum was the best-known and most authoritative. If either collection has survived the middle ages, it has yet to be discovered. The Stella maris, nevertheless, provides us with a shrunken image of the former, and MS Additional 15723 with a somewhat more adequate reflection of the latter.

As for vernacular collections, with which this study has not been primarily concerned, the scanty evidence available in print at the present time suggests that Soissons was an equally important focal point for the production of collections in the French vernacular. At any rate, northern France, when its full history has been written, must be accorded a place of prominence in the development of the Mary cult in its popular and literary aspects.

1. The Authorship and Date of the Stella Maris

JOHN OF GARLAND himself acknowledges the authorship of the Stella maris. In the midst of an apostrophe to the Virgin interpolated between two legends he writes,

In hiis ridmis quasi cannis

Stridulis planctum Iohannis

Audi de Garlandia.1

His contemporaries knew it as one of his minor works. In the Ars lectoria John of Garland finishes the enumeration of his works with the line, Hiis scriptis alia poteram conjungere multa, and the author of the Bruges gloss comments, Ut Commentarium, Stellam maris, Assertiones fidei, Morale scolarium, Georgica spiritualia.2

The Stella maris is not, as has been suggested,3 to be identified with another work which John of Garland mentions as his own in the Morale scolarium,

Edita sunt annis in paucis Mira Iohannis,

Transversis pannis promptis ad bella tirannis.

Sarra molestatur ab Agar, dolet et lacrimatur;

Set pius armatur deus ut vindicta sequatur.4

The Morale scolarium was, in the first place, written in 1241,5 seven or eight years before the Stella maris. Moreover, the content of the Mira or Mirabilia, as John of Garland describes it, is in no way suggestive of the Stella maris, nor is there anything in the work that would make reference to it apt in the situation which he is discussing in the Morale scolarium.

The date of the work may be determined with considerable accuracy by means of internal evidence. The author, recounting a miracle which took
 [[ Print Edition Page No. 78 ]] 
place in the church of Notre-Dame at Paris, mentions a certain bishop Guillaume as a witness,

Est in templo virginali

Virgo, medicina mali

Ardentis Parisius.

Fertur ignis infernalis,

. . .

O Gwillelme, presul pie,

Qui conservans es Marie

Oves et ovilia.

Ista vides et testaris.6

Guillaume d’Auvergne was bishop of Paris from 1228 until his death, March 30, 1249.7 During his episcopacy there was a recrudescence at Paris of the plague of ergotism, commonly known in the middle ages as sacer ignis, or as John of Garland calls it, ignis infernalis. In an act of the bishop and the chapter of Notre-Dame it was ordered in March 1248 (1249) that the entrance before which the sick and dead were laid, especially those suffering with sacer ignis, should be illuminated by means of six lamps.8

Still another passage in the Stella maris itself confirms the date, March 30, 1249, as the terminus ad quem,

Ista mira quando scripsi,

Tunc scripture favet isti

Studium Parisius.

Hoc magister tunc Galterus

Pie rexit, prudens erus,

Pius cancellarius.9

Gautier de Château Thierry became chancellor of the University of Paris in 1244 and was made bishop of Paris after the death of Guillaume d’Auvergne, March 30, 1249.10 If the Stella maris was composed, then, when Gautier de Château Thierry was chancellor of the University and Guillaume d’Auvergne was bishop of Paris, it must have been written before March 30, 1249.

A terminus ex quo is provided by a reference in legend no. 59 to a battle between the Parmese and Frederick II,

De Frederico a Parmensibus superato et quo tempore factus liber iste

Dum Parmenses invaserunt

Fredericum, detulerunt

Virginis ymaginem.11

The battle mentioned was the siege of Victoria12 by Frederick II of February 12, 1248. Matthew Paris describes it in the Chronica majora, and there are numerous references to it in the annals of the north Italian towns.13 Nothing is said in these accounts about the miracle, although the victory is generally attributed to divine aid. It is Fra Salimbene, the gossipy Parmese chronicler, who not only reveals the source of John of Garland’s miracle, but also gives the clue as to the manner in which the news reached John of Garland in Paris and the length of time it took it to get there. The friar says that in the year his city was besieged by Frederick he went to France, first to the pope at Lyons, then to Paris where he stayed eight days from the festival of the Purification of the Virgin, and finally to Sens. As he was lying ill at Sens early in March, some French brothers from the Franciscan convent at Parma came to him telling the news of the victory. The letter which he mentions was written by the Parmese themselves to Pope Innocent IV at Lyons, a sort of official account of the saving of the city. Salimbene intended to include it in his Chronicle,14 but it has not been preserved. There is, however, in the Chronica majora of Matthew Paris a similar letter written by the podesta of Parma to the people of Milan through their podesta, telling of the same event.15 The story which this official account tells of the battle is the one which John of Garland puts into verse. Parma was saved by a picture of the Virgin painted on its standard, and the enemy suffered great losses. Doubtless the same French Franciscans who brought the letter to Salimbene at Sens had visited Paris as well. John of Garland might have had it from a sermon preached by one of the friars, or he may have seen the document itself.

It is possible to conclude, then, that the Stella maris was completed between February 12, 1248, the date of the siege of Parma and March 30, 1249, the date of the death of Guillaume d’Auvergne and the end of Gautier de Château Thierry’s chancellorship of the University of Paris. It could have been finished as early as March, 1248, for the letter which was the basis of one of the last three legends had reached Sens, and presumably Paris as well, early in that month.

2. The Manuscripts of the Stella Maris

There are but two known manuscripts of the Stella maris of John of Garland:16

I. MS Bruges 546 = B

The manuscript now in the Public Library of Bruges is the most significant of all the known manuscripts of John of Garland because of its early date, the richness of the marginal and interlinear glosses, and the bulk of the work of John of Garland it contains.17 It once belonged to the Cistercian monastery of Dunes in western Flanders, which was burned by the Gueux in 1560 and re-established in Bruges in 1624.18 De Poorter has deciphered an inscription on the first folio which names a certain knight, Henry Sumi, as an ancient owner.19

The volume, written in an irregular script of the second half of the thirteenth century on vellum, comprises 174 folios, many of which have been damaged. With the exception of the first number and the various brief items which fill in the blank spaces, it is devoted to the works of John of Garland.20 The Stella maris is written in three columns of thirty-eight to forty-four lines each. The text has been so frequently corrected that it is more accurate than the British Museum (M) manuscript, which was originally written with more care. Some of the corrections are in the book hand of the text, while others are in the semi-cursive of the gloss. Titles appear in faded red ink. Space has been left for rubric initial letters, but they have never been added.

Numerous interlinear and marginal glosses, probably by a single hand, accompany the text, some of them of considerable length. There is some evidence that John of Garland himself wrote much of the gloss. The note at the bottom of the first folio is clearly his own, written in the first person. Certain other comments on the legends themselves must have been made by some one who had seen and read the manuscript of Ste. Geneviève, John of Garland’s source. In the case of fourteen legends appropriate details are added by the gloss, especially accurate references to the scene of the incident,21 which would not have remained long in the memory of some
 [[ Print Edition Page No. 81 ]] 
one who was only generally familiar with the tales. The author of the gloss as a whole betrays a particular interest in the natural sciences, as does the Stella maris and many other works written by John of Garland.22 At the top of folio 88 is the only existing reference to an unidentified work of John of Garland, a Liber elegiarum.23

Other glosses must have been written for young boys learning the Latin language. They are intended to explain the meaning of difficult words, the gender of nouns, the particular conjugation to which a verb belonged, to point to an antecedent of a relative pronoun, or to identify a figure of speech. In the latter part of the manuscript some Latin words have been given French equivalents.24 The gloss is, therefore, evidence of the use of the Stella maris in the schools, the purpose for which the author composed it.

II. MS London British Museum Royal 8 c iv = M

The Stella maris is the only work of John of Garland in the British Museum manuscript25 of 210 folios, vellum in 4°, comprising miscellaneous numbers in hands of the thirteenth and fourteenth centuries. The volume consists of two collections which have been bound together since the fifteenth century.26 The first collection comprises twenty-four numbers, including the Stella maris, and bears the press mark of the abbey of Bury St. Edmunds, R. 42. The second is a medical collection with the Bury press mark, M. 35.27 The codex was one of seventeen manuscripts known to have been acquired by John, Lord Lumley, from the abbey in the late sixteenth
 [[ Print Edition Page No. 82 ]] 
or early seventeenth century. King James I purchased it for the Royal Library along with Lumley’s collection.28

John of Garland’s Stella maris, the seventh article in the first collection, begins on folio 16 and is concluded on folio 23^v with the colophon, Explicit liber magistri I. de G. de miraculis beate virginis. The script, a hand of the thirteenth century, perhaps about 1300, is more careful than that of the B manuscript. It is possible that the same scribe wrote both text and glosses, for they are contemporary hands, and some of the corrections of the text are in the semi-cursive hand of the gloss.29 The scribe who copied the commentary, and perhaps the text as well, has left his name on the top margin of folio 20^v at the close of some verses which are pen-trials.30 The first lines have been trimmed off the margin,

Est stilus equivocus, quia dat. . . .

Columpne medium qualitas Carmentis habet,

Cum ipsum carmen sit officialisque poete,

Est instrumentum quo scribit scriptor Adulfus.

Adolf was probably an Englishman, perhaps even a monk of Bury. Most of the gloss he must have copied from the text he had before him, for it is at some points identical with that of the Bruges manuscript. Some of the glosses were, however, composed by an Englishman who did not know the legends current in northern France. The legend ‘Ring on Finger’ is unfamiliar to him, for he names ‘Eadmundus’ as the hero, thus connecting it with a similar English tale told about Edmund Rich, archbishop of Canterbury in the reign of Henry III.31 Apropos of the disease of the tongue of the clerk of the ‘Milk’ legend, he writes a cancero devoratam, a remark which proves that he is thinking of ‘Twenty-three Plants in Flower,’ a legend of English origin, rather than John of Garland’s ‘Tongue and Lips Restored,’32 current in northern France. The English glosses which appear on several folios are probably his also.33 Two proper names in the text itself, ‘Willelme’ (l. 496) and ‘Walterus’ (l. 913) are spelled in English style. Several other spellings of the text and gloss probably indicate English origin. In five instances ‘e’ is used where ‘i’ would be expected:‘palleo’ for ‘pallio’ (l. 84); ‘incendeo’ for ‘incendio’ (l. 87); ‘vexit’ for ‘vixit’ (l. 87); and ‘descraseas’ for ‘discrasias’ (l. 117).

The text of the M manuscript is written in two columns of thirty-three to forty-four lines each with the exception of the first and last folios. The first has but one column and the last three. The chapters are marked by the use of crude red and blue initial letters. Most of the titles appear in the margin in the semi-cursive hand of the gloss, the first seven in hands which differ from that in which the others are written. Two kinds of corrections have been made. The scribe did some in book-hand presumably when he copied the text.34 Others were done in semi-cursive from a manuscript related to B, if not B itself, when the titles and glosses were written.35 Several peculiarities of spelling, besides those mentioned above, differentiate M from B. Occasionally ‘liquor’ is spelled ‘licor’ (l. 218), ‘quondam’ appears as ‘condam’ (l. 604), and ‘quedam’ as ’cuedam’(l. 868). The same peculiarity extends to the gloss and the titles of M: ‘pertorcuens’ (p. 89), ‘secuntur’(l. 42 and title, p. 89) and ‘licor’ (l. 219). In two cases ‘s’ is written for ‘c’: ‘Sedit’ for ‘Cedit’ (l. 391) and ‘selis’ for ‘celis’ (gloss to l. 148).

As for the relationship of M and B, the evidence is not entirely conclusive. There are some indications that a manuscript of the M type was used in the correction of B, and conclusive evidence that B itself, or a manuscript close to B, was used in the rubrication and probably the correction of M.

After the text of B had been completed the same scribe, or possibly another who was a contemporary, corrected it and wrote the titles with the help of a second manuscript. His second exemplar was not M, but it had some of the characteristics of M. Writing in semi-cursive, the scribe introduced four of M’s readings, all of them verbs, into the text of B as alternative readings.36 He created new and appropriate chapter divisions on folios 87^v and 88, writing the titles in the margin or squeezing them in between the columns.37 He made a sign opposite line 1005 on folio 88 to mark the omission in B of twenty-four lines which appear in M, and wrote the title which belongs to them between the columns.38 Using a pair of the same signs, he indicated the transposition of six lines of the B text to a more suitable position, the position which they occupy in M.39

The nature of these corrections suggests that the scribe’s second manuscript, one of the M type, was a later revision of the Stella maris made by
 [[ Print Edition Page No. 84 ]] 
the author himself. The twenty-four lines of M, omitted in B, have a unity of their own, one of a series of brief pieces in praise of the Virgin covered by the title, de eius magnificentia usque ad finem libri. The verses thus interpolated in M are, in all probability, from the pen of John of Garland. They are similar in form and theme40 to others included under that title and elsewhere in the Stella maris. He may well have written them after the completion of the Stella maris and incorporated them into a ‘second edition,’ which was the parent of the M text.41

A striking case of agreement in error proves that either B, or a manuscript of the B type, was used in the rubrication of M. When B was copied, there was a thin place, or perhaps a tiny hole, in the vellum of folio 86 exactly where the red initial belongs at the beginning of legend no. 34 (l. 538). The rubric was never made; nor was there a guide to the rubricator. The first word of the line in B, then, is [A]bbas. With use the weak spot wore into a hole with thin edges until it framed exactly the first letter of the line beneath on the next folio. The letter was ‘P,’ and the word to the casual reader even today appears to be Pbbas. Just so the rubricator of M copied it, quite oblivious of the fact that the correct letter ‘a’ had been written there for his guidance.42 The scribe of the M text has, moreover, sometimes failed to provide for a rubric where one is clearly needed. The rubricator of M may have had the assistance of B when he marked the beginning of those chapters with a red paragraph sign.43 Certain corrections and one alternative reading may also be the result of contact between B and M.

The M manuscript, except for the last three folios, is even more liberally supplied with interlinear and marginal glosses than is B. The character of the commentary as a whole is similar to that of B and sometimes identical.44 Several long glosses, however, distinguish it from B. A theological discussion on the first folio, similar to that of the Quoniam prologue prefixed to the collection of St. Germain-des-Prés is followed by more details from the Pseudo-Matthei Evangelium than John of Garland gives. Both these glosses, together with the reference to a‘register at Rome’ as the source of some of the legends, imply acquaintance with John of Garland’s source, the manuscript of Ste. Geneviève.

The author of the glosses of M, however, was primarily interested in grammar and the natural sciences, not in theology. He has written a
 [[ Print Edition Page No. 85 ]] 
rhetorical preface to the Stella maris, pointing out how the principles of the Clavis Compendii of John of Garland have been applied in the composition of the Stella maris.45 He knows also the Epithalamium beate Marie virginis and the Compendium grammatice of the same author, from both of which he quotes.46 The most interesting of the glosses are those on ‘spiritual astronomy,’ to use the words of the title of the chapter.47 These are but the counterpart of the glosses on Albumasar’s De naturis signorum in B.48

In fact, a comparison of the glosses of M and B leaves the impression of two students sitting in the class-room listening to their master lecture on the Stella maris who have decided to divide the labor of taking notes between them. Certain of the briefer remarks are identical, but the long glosses supplement each other. The M manuscript, it should be noted, has no French glosses after line 663; and B, none before line 725. At least, it can be said, that the bulk of the commentary of B and M must have emanated from the same general source. It is just possible that John of Garland himself wrote a great deal of the M gloss also to go with his‘second edition.’

3. The Editing of the Stella Maris

John of Garland’s collection of Mary legends is usually referred to as Miracula beate Marie virginis or Stella maris.49 The latter title has been adopted for this edition, because it is the author’s title, and he attached a particular significance to it. Just as the star of the sea, Virgo, shows itself in the north to sailors in danger, so his little book by means of the miracles of the Star of the Sea spreads forth the way of salvation amid the bitterness of the world to him and his students there in the studium at Paris.50

The two known manuscripts of the work have been used in the preparation of the edition. The Bruges manuscript is designated B and the British Museum manuscript M. Because it is earlier in date and more adequately corrected, the aim has been to present as accurate a picture of the Bruges manuscript, as is consistent with the making of a legible text. In cases of garbled words, omissions, and errors in B, the readings of M have been adopted. Lines 1006-1029 are lacking in B. They have been included in the text, nevertheless, for two reasons: because there is a device in B at the end of line 1005 which the scribe who made some of the corrections has
 [[ Print Edition Page No. 86 ]] 
used before to indicate an omission, and because the missing lines were almost certainly written by John of Garland.51

The titles are those of the Bruges manuscript with the exception of that which prefaces the verse prologue and one other.52 These have been borrowed from M. The assistance of M has also been required in arranging the sub-divisions of the text. Numbers have been given to legends only, even though other sub-divisions have titles. The spelling is that of the Bruges manuscript with a few exceptions, and these have been noted. No distinction has been made between ‘ci’ and ‘ti,’ between the long and the short ‘i,’ or between ‘u’ and ‘v.’

The prose preface which introduces the work as it has been edited is in reality a note at the bottom of the first folio of the Bruges manuscript. It has been used as a preface to the text for the following reasons: (1) It is indisputably by John of Garland himself. (2) It explains the circumstances under which the work was composed and serves as a sort of grammatical preface. (3) And it helps to solve the mechanical difficulty which results from the number and length of the glosses at the beginning of the work. The corresponding preface of the M manuscript has been placed in a parallel position as a gloss.

All the glosses are included as they appear in the manuscripts, except for such words as ‘dico,’ ‘id est,’ and ‘scilicet’ when they are not required to make the meaning clear. Interlinear notes are given without reference to their position on the folio; the location of others is indicated. Several of the verses quoted in the glosses have not been identified. Presumably they are from contemporary works.

The Latin texts of the original versions of the narratives which John of Garland used could not be reproduced in the notes because of their bulk. Instead, brief English summaries have been placed at the beginning of each of the notes on the legends. So far as possible the texts of MS Paris Bibliothèque Nationale 12593 have been employed for this purpose.53 The lists of parallel texts included in the notes take into account only those which are available in print, and they are rather strictly limited in scope. No attempt has been made, for instance, to include tales which are merely similar in theme or in a few details.