happy mingling of blurring and spreading of the lines in some places and a lack of impression in others. Retouching is unavoidable, even at the present time and-after twenty years of progress; *-and when photo-lithography was still a "newly discovered process" it was necessarily often resorted to. But the moment hand and eye get a chance to intervene in reproductive work, humanum est errare. † In the cross-bars of the letter e and the ƒ and long s, and in respect of battered type and punctuation, the Staunton Folio is, in spots, admittedly defective.

The fourth and latest reprint is the reduced photo-lithographic reprint, miscalled "fac-simile," in small octavo, published by Chatto & Windus in 1876, with an Introduction by the veteran, Mr. Halliwell-Phillipps, which, as I take it, is the extent of his connection with the enterprise. Its cheapness, 8s. 6d., is its chief discommendation, for it has been an inducement to put a defective text in the hands of many who are thereby led to suppose themselves “ on a level with the envied possessors of the far-famed original." The typography is small enough, and the blurring and general sloppiness of the workmanship bad enough, to make its use perilous without a magnifying-glass and a safety-valve. Neither the title-page nor the Preface give any indication of the original copy from which it has been "reduced." The late Mr. Ingleby, who, with all his virtues and exceeding geniality, indulged at times in needless causticity,

* Mr. Kell

the printer of this text-who put on stone the transfers in lithographic ink supplied to him by Mr. Praetorius-states that he has done his very best with the (often faulty) transfers supplied to him. He has lost all his profit, and more, by paying for cleanings and corrections by hand. The Museum copy of the Quarto is bad in some pages, and the negatives required more painting out of letters printedthrough, and more cleaning of the transfers than the price of the book would (in the fotografer's opinion) stand. In this work, good transfers from the negatives are all in all.-F. J. F. (Introduction to Praetorius fac-simile, R. & J., 1599.)

Henceforth, all sheets will be passt for press by the Editor as well as the Lithographer. (Fore-words to Griggs fac-simile, Ven. & Adonis, 1593.) Halliwell-Phillipps' Preface, p. xi.

spoke of it as "a reduced reproduction of Mr. Staunton's Folio; " and added, "But why is Mr. Staunton not mentioned? ""* As to this assertion, I am inclined to suspend judgment, for there are peculiarities in the Chatto & Windus phototext which do not seem referable to the Staunton.

The reduced type reprint of Booth is so easily attainable at moderate cost that few students, even of the "common" class, will care to be without it. Of the four complete reprints thus far vouchsafed to us, it is probably as safe to refer to the Booth text for the decision of a disputed reading as to any known original of the 1623 Folio, short of the Grenville copy in the British Museum; which, as one of the most perfect of the Third Period copies, and as the standard of the National Library, may be regarded as invested with a pre-eminence of its own to which all conflicting texts must yield. Mr. Furness-than whom no more impartial and capable judge in such matters exists-writes me the following: "It is my settled conviction, founded on an experience of twenty years, that Booth's Reprint is the very best reprint of so large a work that the world has ever seen, or is likely to see. Considering the variations in the copies of F1, I doubt if a single misprint can be fastened upon Booth. Because his reprint differs, no matter how widely, from my original, I shouldn't think of imputing an error to him." It is proper to add that the large-paper copies of Booth's Reprint, which were first issued in three parts, are the most trustworthy, for I have found in the small-paper copies, later issued, in which the serial character of the three parts is abandoned, indications of a deterioration of the presswork inseparable from the printing of a large edition.

Reference to a fairly authentic text is especially necessary if there is to be any serious effort to compare parallel texts, such as of the Folio with the reprints of the earlier Quartos. Readyprepared parallel texts are not generally accessible. With the exception of the side-by-side reprints of the Folio and First Quarto of Henry V. edited by Dr Nicholson and published in 1877 by the New Shakespeare Society, and the lately printed.

*Ingleby: The Man and the Book, pt. 1, p. 114, note.

parallel Folio and "Pied Bull" texts of King Lear, prepared by Professor Vietor, of Marburg, and published by Whittaker in London, I know of none which bring the Folio and the early Quartos into juxtaposition in the manner pursued in our own Society's "Bankside Edition." The New Shakspere Society announces as in course of preparation parallel texts of the Folio and Quartos of Richard III., of 2 and 3 Henry VI. with the Contention and the True Tragedy, as well as the long delayed Four-text Hamlet, begun by the lamented Teena RochfortSmith and now finishing by Mr. Furnivall; and a number of others are lavishly suggested. We have also many parallel texts of different Quartos (without the Folio), such as Mr. Sam. Timmins's exquisite reproduction of the Devonshire Hamlets (the Quartos of 1603 and 1604), printed by Mr. Josiah Allen at Birmingham, and the New Shakspere Society's Romeo and Juliet, in 1874. The German critics appear to be especially fond of the parallel study of the Quartos.

Until our own Bankside Shakespeare shall furnish a critically edited uniform and exact parallel version of all the plays found in quarto and folio form, the ordinary student who seeks to compare a disputed passage or to collate for himself an entire play, has no help for it but to spread out his Folio fac-simile and such Quarto text as he may find, side by side with some modern text as a guide, and peg away.*

It is just here that the loss of time creeps in, of which I so bitterly complain. The problem is, to find the parallel passages. Not one of the complete Folio texts has a standard linenotation, to facilitate cross-reference. The few Quarto texts that are line-numbered, follow no co-ordinated system. Each editor adopts his own no-system, and, in the mass, confusion necessarily results. It is worse than the simple absence of a Canon-it is an ever-present conflict and discord.

For a century and a half, from Rowe to the Cambridge Editors, Shakespeare's plays remained without a convenient

*Since Mr. Adee read this Paper, the N. Y. Shakespeare Society has issued three volumes of the Bankside Edition, with the annotation first suggested by him in this Paper.

scheme of reference numeration. The primitive citation of act and scene was alone available. The labor of finding a hurriedly sought-for line or word in a scene of six or seven hundred lines in length, such as 1 Henry IV., II., iv. (602 lines), or Hamlet, II., ii. (634 lines), is appalling, and in the good old days of Mrs. Cowden Clarke's Concordance, before Schmidt or the Globe were, the provocation was often strong to relinquish the search. There were, to be sure, some few editions of separate plays with numbered lines, but they were either school-editions, often expurgated, or Continental reproductions, and no two texts of the same play by different editors would agree. It was not until Clark and Wright, after numbering the line of the Cambridge edition for convenience of collating the footnotes, conceived the happy thought of numbering the lines of the Globe edition also, that a popular reference Shakespeare lay at hand. Of course, the Globe and Cambridge texts do not agree -the difference in the length of the type-lines and the varying spacing break up the arrangement of all parts where prose is used (as is seen in Hamlet II., ii., which counts 581 lines in the Cambridge edition and 634 in the Globe). But the convenience of the natty little single volume of the Globe, with its creamy paper, its singularly clear-faced type, and its cheapness, joined to the natural craving of the human mind for some sort of a canon of uniformity, caused the Shakespearian world to seize upon the Globe as a standard of reference; and the scholar, the professor, and the casual critic have accepted, by a sort of lex non scripta, the citation of act, scene, and line of the Globe, even where it is wrong.* And yet, in spite of the alacrity with which the Globe edition was admitted as the line-numbered standard, and the practical universality of its use, not a single modern edition follows throughout the numeration of the Globe. Take, for example, Hamlet's "dull and muddy-mettled rascal," which, according to the Globe: is II., ii., 594—we run it to earth in a few other editions professing to be numbered for

"The line-numbers are those of the Globe edition, even where they are wrong, as they once or twice are." F. J. Furnivall, Forewords to Grigg's fac-simile Quarto 1, Love's Labor's Lost, 1598, p. xv.

"convenience of reference" and find that its number is: in the Cambridge edition and in Furness's Variorum (which follows the Cambridge), 541; Leopold, 575; Rolfe's, 552; Clark and Wright's Clarendon Press series, 548; Richard Grant White's Riverside edition, 535; and Tschischwitz's, 593. Is there not a touch of satire upon the critical faculty of the professional critic, to find that, in three editions by the same editor, W. G. Clark, the same line has three different line numbers? It will probably soon have four, for I learn that the Macmillans are about to issue a "Jubilee edition" in three volumes, "substantially following the text of the Globe."

Having ascertained the line-number of a particular phrase, according to the Globe standard, this is merely a guide to its approximate position in some other edition, if the latter happen to be line-numbered too. But it is not even a guide to the neighborhood, if the editor mounts some pet hobby and adopts a notation of his own, like Karl Elze, who numbers his Hamlet by paragraphs of from 12 to 20 lines continuously throughout the play, which makes our dull and muddy.mettled rascal "peak" in paragraph No. 100. Elze says of this system, "The division into paragraphs is transferable to all editions without any the least difficulty, so that in time a uniform mode of reference may be adopted by the students of Shakespeare in all parts of the habitable globe. It is true that numbering by paragraphs does not enable the student or reader to refer to a particular line, yet the average length of the paragraphs (numbering between twelve and twenty lines) is so convenient that the eye will catch in an instant the passage or word referred to." Leo's earlier Coriolanus is divided into 255 sections coincident with the natural pauses or transitions of the dialogue.* Elze's later Hamlet has 241 paragraphs. Craik has broken up Julius Cæsar much more generously, having allotted 795 paragraphs to it, on the general plan of numbering each speech, whether of a single word or of many lines.‡

*F. A. Leo. Coriolanus, London, 1864. (Gives a photo-lith. of Fi.) Karl Elze. Hamlet, London, 1882.

Craik. The English of Shakespeare, London, 1878.