In December, 1487, Caxton issued an edition of the Sarum Missal, though he was not himself the printer. The work was done for him by a printer at Paris named Guillaume Maynial, about whom but little is known. He is presumed to be a relation, son, perhaps, or nephew, of George Maynial, the partner of Ulrich Gering in 1480. He printed only three books, of which this Missal is the earliest, the other two being the Statutes and the Manual of the Church of Chartres, issued in 1489 and 1490. The only copy of this book at present known is in the library of Lord Newton. It is a folio, and when perfect should have contained 266 leaves, but of these 23 are now missing. The page is printed in two columns, with 39 lines to a column.

One point which gives this book a peculiar interest is, that in it is found for the first time Caxton's well-known device. It consists of his initials, divided by his merchant's mark, with a deep ornamental border at top and bottom. Many ingenious writers have attempted to read into this mark several items of information. The merchant's mark they say is not a merchant's mark at all, but the figures 74 significant of the time when he began to print. Two small ornaments shaped like an S and C stand for Sancta Colonia, where he learned the art of printing. The mark is, however, merely an ordinary merchant's mark, which in some shape or another all printers introduced into their devices, and the letters S C merely ornamental flourishes.

Another question has been raised as to whether this device was cut in England or in France, but it has no resemblance to French work, and is almost certainly a native production. As Mr. Blades justly remarks: "Caxton, desirous of associating his press more directly with this issue than by the colophon only, which many people might overlook, probably designed his mark for the purpose of attracting attention. He no doubt stamped this device on the last blank page of the books after they had been received from abroad and before putting them into circulation."

It seems not improbable that besides the Missal, Maynial printed for Caxton another service-book, the Legenda according to the Salisbury use. The existence of this book is known only from a few odd leaves, for the most part rescued from old bindings and preserved in different libraries, but it agrees in every respect typographically with the Missal. The type is identical, the number of lines and size of page the same, and everything points to the same printer. Perhaps some day a copy with the colophon may be found and our doubts on the subject set at rest.

About 1488 appeared a new issue of the Golden Legend. It is not an entire reprint of the first, but only of certain parts of it. It contains 448 leaves, being one less than the first issue, and of these 256 are reprinted and 192 are of the original edition. It is difficult to explain this reprinting, but it was probably caused by the destruction of a large part of the stock of the original issue. Caxton took the opportunity to make two improvements in the reprint. He compressed the quires signed X and Y, which contained the awkward number of nine leaves, into a single quire X of eight leaves, and instead of having a blank leaf at the end of the book he added the life of St. Erasmus. The parts of the book which are of the second issue may be readily distinguished from the first by the head-lines. In the first issue they are in the larger type No. 3; in the second, in the smaller type No. 5.

On the 14th of July, 1489, Caxton finished printing a translation of the work of Christine de Pisan, entitled the Fayts of Arms and of Chivalry. This translation, as he tells us in the epilogue, he undertook at the express desire of Henry VII., who himself lent him the manuscript with the original French text. It is not improbable that the identical manuscript which Caxton used is one which is now in the British Museum, and which formed part of the old Royal collection. It was written for John Talbot, Earl of Shrewsbury, who died in 1453, and by whom it was presented to Queen Margaret, and it agrees very closely in every way with Caxton's English version.

Considerable doubt has been thrown on the authorship of Christine de Pisan, but apparently unjustly. In the prologues of many manuscripts, and in Caxton's edition, the writer apologizes as a woman for treating of such warlike subjects, and appeals to the goddess Minerva, saying, "I am, as thou wert, a woman Italian."

A complete copy should contain 144 leaves, the first being blank, and over twenty copies are known. A perfect copy in the Cambridge University Library contains a manuscript note showing that it was bought in 1510 for three shillings and eight pence.

In 1489, also, Caxton issued two editions of an Indulgence of John de Gigliis, or rather a license to confessors, giving them power to grant indulgences to any Christian person in England or Ireland who should contribute four, three, two, or even one gold florin to assist a crusade against the Turks. These Indulgences are of peculiar interest, as they were printed in a new type of Caxton's, the smallest which he ever cut, and of which he never again made use. The first to draw attention to them was Archdeacon Cotton, who in the second part of his "Typographical Gazetteer" mentions one which he had found in the library of Trinity College, Dublin, and which he considered to be a product of the early Oxford press. Henry Bradshaw, the University Librarian at Cambridge, obtained a photograph of it, and at once conjectured from the appearance of the type that it must have been printed by Caxton. He immediately communicated this discovery to Blades, who, however, refused to accept it as the work of Caxton's press without some further and more convincing proof, and never even alluded to either the type or Indulgence in later issues of his book. The necessary proof was soon afterwards found, for Bradshaw discovered at Holkham an edition of the Speculum Vitae Christi, printed by Wynkyn de Worde in 1494, which had the side-notes printed in this type, and as De Worde inherited all Caxton's materials, this fount must have belonged to him.

The Statutes of the first, third, and fourth years of Henry VII. may also be put down to the end of 1489, for the fourth year of Henry VII. ended on August 21, 1489, and the Statutes would no doubt be printed at once.

With the exceptions just given, none of Caxton's books printed between May, 1487, and his death in 1491 bear any date, so that although all may be approximately dated, their exact order cannot be determined. One very common error in the method of arranging Caxton's books may be pointed out here, which arises from the method adopted by Blades. In his Life of Caxton the books are arranged according to types, which would be an excellent plan if the use of one type had been discontinued as soon as a newer one was made. This, however, was not the case, for several were often in use at one time, and thus Blades's system, though correct in one way, is very misleading to a superficial reader. For instance, Caxton started at Westminster with types Nos. 2 and 3, and both are used in his first books, but Blades puts the books in type No. 3 after all those in type No. 2, and thus the Sarum Ordinale, certainly one of the earliest books printed in England, comes thirty-sixth on his list, and while one book with the printed date of 1481 is number 33, another with the printed date of 1480 is number 39. It will thus be seen that Blades's arrangement was not a chronological one, though most writers have made the mistake of thinking so, and have followed it as such, as may be seen, for instance, in the list appended to Caxton's life in the Dictionary of National Biography, which blindly follows Blades's arrangement without any reference to his system or mention of the types.

Two interesting romances were printed about 1489, the History of the Four Sons of Aymon and the History of Blanchardyn and Eglantine. The first was an extremely popular story both at home and on the continent; indeed, it still circulates abroad in the form of a pedler's chap-book, which perpetuates in a very mutilated state the story of Renaud, Alard, Richard, and Guichard, with their famous horse Bayard, on which all four rode at once. The early English editions of this book almost suffered extermination. The earliest edition of which a complete copy is known is that printed at London by William Copland in 1554. The colophon of this book speaks of an edition printed by Wynkyn de Worde in 1504, of which no trace remains except perhaps some fragmentary leaves in the Cambridge University Library; while of the edition printed by Caxton only one copy, and that imperfect, is known. It is in folio, and probably contained when complete 278 leaves. The unique copy, wanting some leaves at the beginning, was obtained by Lord Spencer from Triphook, the bookseller, and is now, with the rest of the Spencer Library, in Manchester.

The History of Blanchardine and Eglantine is also known only from an imperfect copy which was in the Spencer Library. It is impossible to settle what the correct collation may have been, as the book breaks off abruptly at leaf 102 and all the remainder is wanting. As, however, the last chapter of the work is just beginning on the last remaining page, it seems probable that only the last quire is missing. On the fly-leaf is a curious note in Lord Spencer's handwriting relating to its purchase. "This book belonged to Mr. G. Mason; at whose sale it was bought by John, Duke of Roxburghe. The Duke and I had agreed not to oppose one another at the sale, but, after the book was bought, to toss up who should win it; when I lost it. I bought it at the Roxburghe sale, on the 17th of June, 1812 for £215 5s." At the earlier sale the Duke had paid £21 for it. This book was undertaken at the request of Margaret, Duchess of Somerset, who brought to Caxton a copy of the French version, which she had long before purchased from him, commanding him to translate it into English.

During the last two years of his life at least half of Caxton's books were merely new editions of some of his earlier works, and therefore hardly call for much detailed notice. The Dictes or Sayings was reprinted for the third time, and the Directorium Sacerdotum, Reynard the Fox, and the Mirror of the World for the second.

Of the Directorium but one copy is known, which is in the Selden collection in the Bodleian. Blades remarks about it that it is "still in the original parchment wrapper as issued from Caxton's workshop." All evidence goes to prove that Caxton never made use of parchment or vellum as a binding material, and in the case of the present book it is quite clear, on close examination, that it has been made up from two imperfect copies, and that the binding is not earlier than the seventeenth century.

The Reynard the Fox is also unique, and buried in that almost inaccessible collection, the Pepysian Library at Magdalene College, Cambridge. It wants, unfortunately, the last two leaves, so that the colophon, if it had one, is wanting.

The Mirror is a fairly common book, and is an exact reproduction, though in different type, of the first edition. In the interval between the printing of the two editions one wood-cut had been lost or destroyed, so that the illustration for Chapter II., "Why God Made and Created the World," instead of being the correct picture of the Almighty with the globe in his hand (which Blades strangely calls "the figure of a philosopher"), is the inappropriate cut of the Transfiguration of Christ.

The Doctrinal of Sapience, a translation from a French version of the Manipulus Curatorum, was doubtless printed in the latter half of 1489, as the translation had been finished on the 7th of May of that year. The book itself is not of much interest, though one copy deserves special mention. It is preserved in the Royal Library at Windsor Castle, to which it was presented by a Mr. Bryant. It is printed throughout upon vellum, and contains three leaves found in no other copy. In the text of the book, Chapter 64 is not printed, but the following heading is inserted: "Of the neclygences of the masse and of the remedyes I passe over for it apperteyneth to prestes and not to laie men. C. Lxiiij."

In the Windsor copy this chapter is printed at the end of the book on three extra leaves, and ends as follows: "This chapitre to fore I durst not sette in the boke by cause it is not convenyent ne aparteynyng that every laye man sholde knowe it."

In June, 1490, Caxton finished the translation of two books, The Art and Craft to know well to die and the Eneydos. The first is not a translation of the complete book, but merely a small abridgment, running to thirteen printed leaves in folio. Blades mentions only three copies, and curiously enough makes no mention of the peculiarly fine one which belonged to Lord Spencer, though he made a careful examination of all the Caxtons at Althorp.

The Eneydos is not, as might be expected from the name, a translation of Virgil's Aeneid, but is more in the nature of a romance founded upon it. Caxton's version was translated from "a lytyl booke in frenshe, named Eneydos," probably the work called Le Livre des Eneydes, printed at Lyons in 1483 by G. le Roy. The most interesting part of the work is the prologue, for in it Caxton sets out at length his views and opinions on the English language, its changes and dialects. He notes that it was rapidly altering. "And certaynly our langage now used varyeth ferre from that whiche was used and spoken when I was borne." While some were anxious to preserve the old style, others were equally wishful to introduce the new. "And thus bytwene playn rude and curious I stande abasshed, but in my judgemente the comyn termes that be dayli used ben lyghter to be understonde than the olde and auncyent englysshe."

In order to make the style as good as possible, Caxton obtained the assistance of John Skelton, lately created a "poeta laureatus" at Oxford, who revised the work for the press.

A second edition of the Speculum Vitae Christi and the Liber Festivalis belong probably to 1490. The latter book is not a reprint of the first edition, but another version, and is reprinted from the Oxford edition of 1486.

The last five books printed by Caxton are theological or liturgical. The Ars Moriendi, a unique little quarto of eight leaves, was discovered in a volume of early tracts in the Bodleian by Henry Bradshaw, and is described by Blades in the second edition of his book. He there states that no other edition in any language is known; but it was certainly reprinted by Wynkyn de Worde. The Fifteen Oes, a little quarto containing fifteen prayers, each commencing with O, is known from a unique copy in the British Museum. The book was no doubt intended as a supplement to the Sarum Book of Hours, but no edition agreeing with it typographically is known. It differs from all other of Caxton's books in having wood-cut borders round each page of text. It also contains a beautiful wood-cut of the Crucifixion, one of a series intended for a Book of Hours. No doubt Caxton possessed the set, and we find it later on in the hands of Wynkyn de Worde.

The Servitium de Transfiguratione Jesu Christi and the Commemoratio lamentationis Beatae Mariae Virginis are special issues of new services to be incorporated into the Breviary. The first contains 10 leaves, and is one of the very few books in which Caxton introduced printing in red. The only copy known, bound up with a unique tract printed by Pynson, and some foreign books, was formerly in the Congregational Library, London, but was purchased by the British Museum in 1862 for £200. The Commemoratio, a quarto of 34 leaves, is known only from the unique copy, wanting two leaves, presented to the University of Ghent by the learned librarian, Dr. Ferdinand Vander Haeghen. This little book was purchased for a trifle at a sale in Ghent and remained unrecognized for many years, until M. Campbell of The Hague identified it as a production of Caxton's press.

The book generally considered to have been the last printed by Caxton consists of three treatises printed with separate signatures. These are the Orologium Sapientiae, the Twelve profits of Tribulation, and the Rule of St. Benet.

A writer in the British Museum speaking of these three books, says that they "are in most of the known copies bound together, and have been usually treated as a single volume under the title, probably dating from the eighteenth century, A Book of Divers Ghostly Matters. There is, however, no reason to suppose the connexion to be due to any other cause than similarity of subject and form, combined with nearly simultaneous publication."

No doubt this idea commends itself to the Museum authorities, since they possess only one of the three portions, ruthlessly abstracted by a thief some years ago from a perfect copy in a private library, but unfortunately it is quite incorrect. The compiler distinctly speaks of the books having been printed together, and on account of their treating different subjects, his wish that the compilation should be called the Book of Divers Ghostly Matters.

When complete the book consisted of 148 leaves in quarto. It contains, at the end of the second tract, a wood-cut which belongs to the series specially cut for the Speculum Vitae Christi, though it was not used in it.

The number of books actually printed by Caxton in England, counting separate editions, is ninety-six, and with the three printed at Bruges and the Missal makes altogether one hundred genuine Caxtons. Blades describes ninety-nine books, but amongst these he includes two which were certainly printed at Bruges after Caxton had left, and three printed by Wynkyn de Worde after Caxton's death, so that the number of genuine books which he describes is ninety-four. The finest collection is now, as is right, in the British Museum, which by judicious purchases in recent years has quite outstripped any possible rival.

Five more books remain to be described, which although not printed by Caxton himself, were printed with his types, and have therefore often been ascribed by different writers to his press. These are the Life of St. Katherine, the Chastising of God's Children, the Treatise of Love, the Book of Courtesy, and the third edition of the Golden Legend.

The first of these books is a small folio of 96 leaves, and contains, besides the Life of St. Katherine of Siena, the Revelations of St. Elizabeth of Hungary. The type used is a modification of Caxton's type No. 4*, recast on a slightly smaller body and with several new additions. Unlike Caxton's books which were made up in quires of eight leaves, this has been made up in quires of six. Another point which distinguishes it and the remaining books from Caxton's work is the introduction of several remarkable capital letters. These were obtained along with a fount of type and some wood-cuts from Godfried van Os, apparently about the year 1490, when he moved from Gouda to Copenhagen. The fount of type was not used until 1496, and then only for one book.

The Chastising of God's Children, a folio of 48 leaves, printed in Caxton's type No. 6, is notable as being the first book issued at the Westminster press with a genuine title-page. It is printed in three lines, and runs as follows: "The prouffytable boke for mannes soule, And right comfortable to the body, and specyally in adversitee and trybulacyon, whiche boke is called The Chastysing of goddes Chyldern."

Why so obvious an improvement as a title-page never commended itself to Caxton it is hard to say. It could not have been for want of examples, for, introduced in Germany as far back as the year 1468, they had at any rate during the last ten years of Caxton's life been in common use abroad. Even the London printer, William de Machlinia, had prefixed one to an edition of the Treatise on the Pestilence, by Canutus, Bishop of Aarhaus, which he printed about the year 1486. Of the Chastising, about twelve copies are known.

The Treatise of Love is also a folio of 48 leaves, and agrees typographically with the Chastising; indeed, the two were often bound together, and are quoted by Dibdin as two parts of one book. The introduction tells us that it was translated in 1493 from French into English by a person "unperfect in such work," but no mention is made either of the original author or the translator. It was most probably printed also in 1493, for at the end of that year De Worde introduced his own type and ceased the use of Caxton's for the text of his books. At the end his first device is found, consisting of Caxton's initials and mark, much reduced in size, in black on a white ground, and apparently engraved on metal. Blades quotes four copies of this book, all of them perfect, but does not mention the copy in the University Library at GÖttingen, and there are probably at least two other copies in private libraries in England.

Of the Book of Courtesy, which, like the earlier editions, was in quarto, nothing now remains but two leaves printed on one side in the Douce collection at the Bodleian. These two leaves, which have been used at some time to line a binding, are waste proof of the beginning and end of the second and last quire of the book, which probably consisted, like the earlier edition, of 14 leaves. On the last page, under the colophon, "Here endeth a lytyll treatyse called the booke of curtesye or lytyll John. Enprynted atte westmoster," is De Worde's device printed upside down, the reason no doubt for the rejection of the sheet.

The last book, the Golden Legend, is a small, thick folio of 436 leaves, with a number of illustrations which had been used in previous editions. The colophon is reprinted verbatim from the first edition, with the simple alteration of the date and regnal year. It ends, as do those of the preceding editions, "By me William Caxton," a circumstance which gives Blades the opportunity of remarking on the carelessness of Wynkyn de Worde. "This is only another instance," he writes, "of the utter disregard of accuracy by Wynken de Worde, who has here reprinted Caxton's colophon, with the date only altered, and thus caused what might have been a puzzling anomaly."

This is, I think, hardly fair criticism. The book is the largest which Caxton translated, and the words "By me William Caxton" may apply quite as much to the translation as to the printing, and it is no doubt that De Worde retained it as applying to the former. As Caxton was but recently dead, and well known to every one, he could not possibly have intended to signify that he was the printer.

One point in connexion with this book is curious. How was it that this third edition was printed when the stock of the earlier edition was not exhausted? Caxton, by his will, bequeathed a certain number to the churchwardens of St. Margaret's, to be sold for the benefit of the church, but these were not exhausted even by 1498, when a fourth edition was printed. In 1496 Caxton's son-in-law received twenty, and a number still remained in possession of his daughter.

A solution of this difficulty has occurred to me, which, though it may be considered as improbable, is by no means impossible. This is, that the "legends" mentioned in the various documents were not copies of the Golden Legend at all, but were copies of the Legenda of Salisbury use, which, as pointed out on page 71, were probably printed for Caxton. Being a book printed specially for the use of the clergy in church, such a bequest would be very suitable. In 1496 these "legends" were valued in the law-court at thirteen shillings and four pence apiece, but the twelve copies sold by the churchwardens of Westminster between 1496 and 1500 gradually decreased in price from six shillings and eight pence in the first year to five shillings in the last.

Considering the number of Caxton's productions that are now known to us only from mere fragments, it is probable that many have disappeared altogether. Amongst these may be reckoned one of considerable importance, the Metamorphoses of Ovid.

In the introduction to the Golden Legend Caxton writes: "Whan I had parfourmed and accomplisshed dyvers werkys and hystoryes translated out of frensshe into englysshe at the requeste of certeyn lordes, ladyes and gentylmen, as thystorye of the recuyel of Troye, the book of the chesse, the hystorye of Jason, the hystorye of the myrrour of the world, the xv bookes of Metamorpheseos in whyche been conteyned the fables of ouyde, and the hystorye of godefroy of boloyn ... wyth other dyuers werkys and bookes, etc."

These, like all Caxton's translations, were done for the press, so there is every reason for believing that the Ovid also was printed. Fortunately we have further evidence, for in the Pepysian collection at Magdalene College, Cambridge, is a manuscript on paper bought by Pepys at an anonymous auction, which contains the last six books of the Metamorphoses, with the following colophon: "Translated and fynysshed by me William Caxton at Westmestre the xxij day of Apryll, the yere of our lord. M. iiijc iiijxx. And the xx yere of the Regne of kyng Edward the fourth."

Though the point can never be settled, it is not unlikely that this manuscript has preserved for us a genuine specimen of Caxton's own writing, not, of course, the ordinary current hand, but the book hand used in copying manuscripts. At that time there was still a prejudice amongst the nobles against printed books, so that the presentation copy to the patron generally took the form of a neatly written manuscript.

There is another interesting point to be noticed about this manuscript. It contains the autograph of Lord Lumley, who inherited the library formed by the Earls of Arundel. Now, William Fitzalan, Earl of Arundel, was one of Caxton's patrons, so that it seems extremely probable that this manuscript was presented to him by Caxton himself.

Another translation of which no trace remains is mentioned in the prologue to the Four Sons of Aymon. The only known copy of Caxton's edition is imperfect, and wants the earlier part containing this prologue, but it occurs in full in the later edition printed by William Copland in 1554, from which the following quotation is taken: "Therefore late at the request and commandment of the right noble and virtuous Earl, John Earl of Oxford, my good singular and especial lord, I reduced and translated out of French into our maternal and English tongue the life of one of his predecessors named Robert Earl of Oxford tofore said with divers and many great miracles, which God showed for him, as well in his life as after his death, as it is showed all along in his said book." What this romance may have been is difficult to say, but it probably refers to the favourite of Richard the Second, the Duke of Ireland, who was killed in France while engaged in a boar-hunt.

Caxton, like all other printers at that time, numbered bookbinders amongst his workmen and issued his books ready bound. Every genuine binding from his workshop is of brown calf, ornamented with dies. His general method of covering the sides of his bindings was to make a large centre panel contained by a framework of dies. This panel was divided into lozenge-shaped compartments by diagonal lines running both ways from the frame, and in each of these compartments a die was stamped. The die most commonly found has a winged dragon or monster engraved upon it. The framework was often composed of repetitions of a triangular die pointing alternately right and left, also containing a dragon. This die is interesting, not only because the use of a triangular die was uncommon, but because it was an exact copy of one used by a London binder of the twelfth century.