excusable because he shared it with his generation, but a writer who now broached that topic without having looked into Gaius would barely escape the reviewers alive. The certainty that he will find far more chaff than wheat is no good ground why the investigator should shrink from threshing his full crop of authorities and winnowing them. The man who has "just enough of learning to misquote" is in no worse plight than the historian who knows just enough about a question to let a vital document elude him. This task of getting at the raw material often implies familiarity with several languages, with epigraphy (or inscriptions), with paleography (or old manuscripts), and extensive travel. It involves a certain acquaintance with the contents of great national collections like the Bibliothèque Nationale at Paris, and the British Museum, of great collections like the Vatican and the Bodleian which belong to ancient institutions, and of great series like the Monumenta Germaniæ Historica, the Documents Inédits and the Rolls Series. And besides having a tolerable notion of where the huge warehouses are situated and what they contain, one should command a stock of bibliographical lore which will guide him through the retail shops, that is through the many special subdivisions of classical, medieval and modern, political, ecclesiastical and social history. According to common definition a scholar either knows a thing or can at once find it out. He does not fulfil this second condition unless he cuts a wide swath in his reading and has studied the classification of books with a view to ascertaining the cardinal features of their contents. It goes almost without saying that any writer whose aim is proof of a particular point or judgment of a particular character, must exhaust the bibliography connected therewith. There is more immediate necessity of reminding those who would read intelligently, and yet cannot devote their whole attention to history, that a degree of bibliographical training is essential. For instance, the blunder of mistaking a partisan or biassed work for gospel is both fatal and frequent. One can readily understand a beginner's belief that Thiers has said the last word on the French Revolution or

Sir Walter Scott on Napoleon, or Mr. T. P. O'Connor on Lord Beaconsfield. Unless he has heard of others he may probably suppose that the first dogmatic book he meets is a just and final authority. In the matter of sources and recent literature alike, bibliography is the handmaid of history.1

After accumulating the data from which our information will finally be drawn, we confront the vexed questions presented by their diverse qualities. At this stage we must seize as an axiom the principle that honest and impersonal criticism is the essence of historical scholarship. No document should expect to escape it, and those which cannot meet the test of examination from without and within may seek in vain a court where favouritism will restore the confidence which defect has cost them. At least that is the high standard of historians like Ranke. Human weakness will hinder its universal attainment, but the process of levelling up to complete candour is much more stimulating than that of levelling down to carelessness and prejudice. We may assume, then, the need of criticising each authority before using it for purposes of demonstration. By Canadian post-office regulations articles are scheduled for convenience into classes which range from first to fifth. In this way things completely alien to each other are jostled together and pay at the same rate. The bulbs of plants and mortgage deeds are enumerated under one head, flour and razors under another, microscopic glasses and oil under a third. Similarly, in classifying historical sources, convenience may require that a single division shall include materials of the most varied nature. No attempt can be made here at distributing documents into small groups and supplying an elaborate system of cross-references which shall connect them. But one

can show roughly how criticism works by making certain large

1 Any serviceable comment upon the bibliography of English history would run into a separate article, or, very likely, into a separate volume. Limits both of scope and space exclude it here. In supplementing these selections readers should consult Gardiner and Mullinger's Introduction to the Study of English History. Part II., by Mr. Mullinger, estimates the authorities, while Mr. Gardiner's general sketch supplies a framework upon which portions of individual knowledge can be hung. The latest edition is preferable.

categories, each of which shall have a clearly marked character and fill a clearly marked place in the scale of values. The further analysis which specialists would require lies outside the range of the present introduction.

In assorting records according to their merits, let us first consider the most trustworthy. As ignorance and misrepresentation are deadly foes of historical accuracy, we should seek and prize the documents from which they are excluded. Now both these vices abound wherever they have a chance, and, if we would see events through the "dry light" of strict truth, we must suspect that they lurk beneath the surface unless tokens of knowledge and good faith are strong. Fortunately one class of evidence is untainted: that, namely, in which writings tell their own colourless story without ornament either from friend or foe. Take the whole body of public and official instruments; statutes, charters, legal notices, bulls, parliamentary writs and returns, memorials, articles of impeachment, and in fact the original text of any important measure, or even of any important statement. No room for distortion exists in such cases, because, once granted that the words have been honestly preserved, the writing proves itself and is final. Roger of Wendover cites the terms of Magna Carta;1 but supposing the original parchment lost, and that he had given only a general account of the promises which John made the barons, we might doubt whether his abstract were a fair As we have the very charter we can feel certain of the conditions which were extorted from the king, and can, besides, use them as the solid basis of inferences concerning feudal life at that period. Magna Carta was signed nearly seven centuries ago, and has long since passed out of the controversial stage. For the sake of showing how precious is the indentical form of any document, attention may be called to Extract 97, "No. 45 of the North Briton". Quite apart from whether Wilkes's charges against the ministry were just or not, the commotion which he raised is a piece of history in itself. Had his article perished, except for references in the pamphlets of his enemies or defenders, we could at best form an imperfect 1 With a few variations from the precise text.

one.

idea of his attitude. As it is, we know the exact nature of his attack on Grenville, and "No. 45," however violent and onesided its language, is in this sense an untainted document.

In a single volume of selections (and especially in a short one like this which covers nearly 1900 years) the quality of each excerpt is designedly kept high. There is little or no room for inferior material. An editor must not waste his space in illustrating the different kinds of poor sources simply because they exist in profusion. Accordingly the reader will find among the following passages numerous examples of those sound and final authorities which have just been described. Constitutional history is always dependent on official records such as the Charter of Liberties (No. 19); John's charter to Dunwich (No. 28); the writ of Edward I. summoning knights and burgesses to the Parliament of 1295 (No. 34); Somerset's Edict against Religious Innovations (No. 58); the preliminary charges against Strafford (No. 71); William and Mary's act against papists (No. 83); and Napoleon's Berlin Decree (No. 110) which provoked the English “Orders in Council". In a closely allied class may be put the Bishops' oath of allegiance to Henry VIII. as Head of the Church (No. 56); the proclamation of James I. on the subject of sports (No. 69); and Anne's Speech to Parliament at the time of the legislative union between England and Scotland (No. 87). Distinct aspects of social life or economic conditions are disclosed by the statutes of a religious guild at Abbotsbury (No. 11); the manumission of a villein (No. 33); and the edict of Edward II. concerning the price of food (No. 36). The bull of Gregory XI. (No. 41) explains the papal disposition towards John Wyclif; and Jenner's memorial to Parliament (No. 108) proves that the discoverer of vaccination appreciated the value of his services to the race. Other citations from original texts belong in the same class. Though tinged with the spirit of faction, they vouch for a prominent person's stand on some memorable occasion. Richard of York's manifesto to the burgesses of Shrewsbury (No. 46) indicates the limit of his claims in 1452; the violent tract (No. 88) which was provoked by the South Sea catastrophe voices the wrath of ruined investors;

and Junius's letter to the Duke of Bedford (No. 98) reveals the gall and wormwood of that celebrated pen. Outside political quarrels, the paragraphs from Lyly's Euphues (No. 65) preserve the high-flown style of Elizabethan gallants, while Nehemiah Wallington's account of Edgehill (No. 73) is charged with the Roundhead's belief that God fought on his side and protected him amid dangers.

Speeches count among documents, for the spoken word once correctly reported is stereotyped into a bit of literature. Before shorthand abbreviations were introduced, a slow speaker's utterance was so rapid that it outstripped the hearer's pen, and the torrent of words poured out by a fast speaker was baffling. Hence Hansard's verbatim fidelity was wanting throughout the Middle Ages, and, in fact, till a recent period. What tricks memory will play may be seen by a comparison of the four contemporary reports which profess to give the address of Urban II. at the Council of Clermont. Did the discrepancies relate solely to words one could reconcile them without difficulty, but the versions vary quite as much in structure as in detail. A part of one is printed in No. 17. The oration which Tacitus ascribes to Galgacus just before Agricola's victory over the Caledonians (No. 3) shows a wider divergence still from the right line. It is purely imaginary, and one cannot conceive that a barbarian should have used such compact and polished language. Froissart probably reproduces the substance of John Ball's arguments, dressing them up in his own terms (No. 42). The remaining selections of this sort comprise three royal speeches (Nos. 61 A, 61 B and 87); two from the bench (Nos. 68 B and 81); and two by the illustrious orators, Burke and Fox (Nos. 99 and 105). In each instance the text verges closely upon exact precision, and thus we can hold these speakers responsible for their words with a strictness which would be unreasonable in the cases first cited. Against the personal tone which marks Elizabeth's speech to Parliament (No. 61 B) it is worth while to set the officialism of Anne (No. 87), and with Burke's splendid judgment must be coupled his faith that principles have a place in politics.